摘要:

CHEMICAL SOCIETY REVIEWS VOLUME 18,1989 0 Copyright 1990 CAMBRIDGE THE ROYAL SOCIETY OF CHEMISTRY CONTENTS PAGE OF POLYOXOMETALLATES AND TUNGSTENPHOTOCHEMISTRY OF MOLYBDENUM AND/ORVANADIUM.By E. Papaconstantinou 1 THE CHEMISTRY SULPHIDES. By R. Michael Paton 33OF NITRILE STRUCTURAL IN DNA: THE FORMATION STRUCTURESISOMERIZATION OF CRUCIFORM IN SUPERCOILEDDNA MOLECULES. By David M. J. Lilley 53 RECENTDEVELOPMENTSIN THE SYNTHESISOF WI~O-~NOSITOLPHOSPHATES.By David C. Billington 83 SULPHONYLTRANSFERREACTIONS. By Isobel M. Gordon, H. Maskill, and Marie- Frangoise Ruasse 123 SYNTHETICAPPLICATIONSOF Neil G. Connelly ORGANOTRANSITION-METALREDOX REACTIONS.By 153 CENTENARY LECTURE. CHEMICAL OF SCIENCEMULTIPLICATION CHIRALITY: AND APPLICATIONS.By R.Noyori 187 REACTION AND EXTREME IN THE STUDYOFBRANCHING KINETICISOTOPEEFFECTS REACTIONMECHANISMS.By Alf Thibblin and Per Ahlberg 209 PROPERTIES COENZYMESTRUCTURAL OF ORGANOCOBALT B1 MODELS. By L. Randaccio, N. Bresciani Pahor, E. Zangrando, and L. E. Marzilli 225 OXIDATIVE OF METHANE CATALYSTS. By G. J. Hutchings, COUPLING USING OXIDE M. S. Scurrell, and J. R. Woodhouse 251 REDOX CATALYSTS AND CHLORINEHETEROGENEOUS FOR OXYGEN EVOLUTION.ByAndrew Mills 285 PHOTOELECTRON OF UNSTABLE WITH RELEVANCEULTRAVIOLET STUDIES MOLECULES TO SYNTHESIS, CHEMISTRY, By Nicholas P. C.QUANTUM AND SPECTROSCOPY. Westwood 317 RHONE-POULENC LECTURE. THE ORIGINOF THE SPECIFICITYIN THE OF OLIGOSACCHARIDESRECOGNITION BY PROTEINS.By R. U. Lemieux FRS 347 BIOSYNTHESIS ALKALOIDS. By David J. Robins 375OF PYRROLIZIDINE MELDOLA MEDAL LECTURE. REDOX RESPONSIVE RECEPTORMACROCYCLIC CONTAINING TRANSITIONMOLECULES METALREDOXCENTRES. By Paul D. Beer 409 CHEMISORPTION PATHWAYS THE ROLEOFAND REACTION AT METAL SURFACES: SURFACEOXYGEN. By M. W. Roberts 451 THE ISOKINETIC By W. Linert and R. F. Jameson 477RELATIONSHIP. 1989 Indexes 507

ISSN:0306-0012    

DOI:10.1039/CS98918FP001

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. SOC.Rev., 1989,18,33-52 The Chemistry of Nitrile Sulphides By R. Michael Paton DEPARTMENT OF CHEMISTRY, UNIVERSITY OF EDINBURGH, WEST MAINS ROAD, EDINBURGH EH93JJ 1 Introduction 1,3-Dipolar cycloaddition reactions are one of the most widely used and versatile preparative methods in heterocyclic chemistry.' With more than a dozen classes of 1,3-dipoles and numerous types of unsaturation able to fill the role of dipolarophile the scope for synthesis is immense. Particular attention has been paid to the nitrilium betaines (Scheme 1). Nitrile RC'A\ A98 + --A=B RCNA\ 11 B RCEN-X 11 B N\x/ N'X / / B 'X Scheme 1 oxides (RC=N+-O-) have a long and varied history with the parent fulminic acid * and benzonitrile oxide being first reported in 1800 and 1886 respectively.Their chemistry has been explored4 in great depth: they are formed with ease from readily accessible precursors, they have been identified as transient intermediates in a variety of reactions, and cycloadditions involving numerous dipolarophiles have been accomplished. Moreover, utilizing their facile stereo-specific cycloaddition to alkenes and manipulation of the resulting 2-isoxazolines (4,5-dihydroisoxazoles) they are a key component of a novel and versatile approach to natural product synthesis which is attracting widespread current interest. Nitrile imines (RC=N +-N-R)4b and ylides (RCGN+-C-R~),~although possess-ing a less extensive literature, are nevertheless readily generated, have a wide range of cycloaddition reactions, and are often the intermediate of choice for the construction of five-membered heterocycles incorporating C=N-N and C=N-C.Nitrile ylides, imines, and oxides all feature prominently in the recent two-volume monograph ' (edited by Padwa) entitled '1,3-Dipolar Cycloaddition 'I .3-Dipolar Cycloaddition Chemistry', ed. A. Padwa, Wiley, 1984. E. Howard. Phil. Trrms. R. SOC.London, 1800, 204. S. Gabriel and M. Koppe, Ber., 1886, 19, 1145. ((I) C. Grundmann and P. Griinanger, 'The Nitrile Oxides', Springer-Verlag, 1971; (b) P. Caramella and P. Griinanger, Ref. I, ch. 3.'((I) A. P. Kozikowski, AM. Clietn. RLIS.,1984, 17, 410 (h) V. Jager, I. Muller, R. Schohe, M. Frey, R. Ehrler. B. Hafelle. and D. Schroter. LWI. Hrterocycl.Cliem., 1985,9, 79. H.-J. Hansen and H. Heimgartner. Ref. 1, ch. 2. The Chemistry of Nitrile Sulphides heat + RC 3N -co,,-s“S CO2Me +-DMAD RC N-S + RNQc02Me 121 (a,R =Ph I Scheme 2 Chemistry’, which has deservedly become the definitive work on the subject. In contrast nitrile sulphides (RCrNf-S-), the only other nitrilium betaine for which cycloaddition reactions are known, have received much less attention and their chemistry has not previously been reviewed. By analogy, they should be uniquely well suited for the synthesis of five-membered heterocycles incorporating C=N-S. Various of these (eg. 1,2,4-thiadia~oles,~isothiazoles *) show useful biological activity, and generally applicable synthetic methods are therefore in demand.The nitrile sulphides have the potential to satisfy this need. Nitrile sulphides are also of interest from a theoretical point of view as they represent a rare case of an N-sulphide. Whereas N-oxides are commonplace the only comparable N-sulphide is the recently detected dinitrogen sulphide (N2S).’-’ Since they were first reported’’ in 1970 there has been a steady, though not heavy, flow of papers devoted to their chemistry. It is the purpose of this review to gather together this literature, to assess what has been achieved, and to identify future applications. Also included is a brief discussion of the closely related but largely unexplored nitrile selenides (RC =N+-Se-). 2 Discovery The first clear evidence for the existence of nitrile sulphides was provided in 1970 by Franz and Black.’ ’ Stimulated by a report l2 that oxathiazolone (la) decomposes on heating to benzonitrile, sulphur, and carbon dioxide they surmised that this process might involve initial decarboxylation to benzonitrile sulphide followed by loss of sulphur (Scheme 2).To test this hypothesis they repeated the thermolysis in the presence of dimethyl acetylenedicarboxylate ’(a) F. Kurzer. Ah. Heterocy./. Cheni., 1982. 32. 285; (h) J. E. Franz and 0. P. Dhingra, in ‘Comprehensive Heterocyclic Chemistry’, ed. A. R. Katritzky and C. W. Rees, Pergamon, 1984, Ch. 4.25. D. L. Pain, B. J. Peart, and K. R. H. Wooldridge, Ref. 7h, 4.17. C. Wentrup. S. Fischer, A. Maquestiau, and R. Flammang, J. Urg. Clmcm.,1986,51, 1908.lo H. Bender, F. Carnovale. J. B. Peel, and C. Wentrup, J. Am. Clmem. Soc.. 1988,110,3458. ” J. E. Franz and L. L. Black, Tetrulirclron Lett., 1970, 1381. A. Senning and P. Kelly, Actu Chmm.Sund., 1967,21, 1871. Paton RCONH, heat + -=7RC=N---S 7RCN I 11 I t -HC1RCONHSCCL 3 “ri*\B N.s’ Scheme 3 (DMAD), a well-established dipolarophile. From the reaction mixture they isolated isothiazoledicarboxylate (2a), the anticipated [3 + 21-cycloadduct. Since this initial report, nitrile sulphides have been generated from several sources and other cycloaddition reactions have been identified. Although all have so far proved to be too unstable to isolate at ambient temperature, they have been detected spectroscopically using matrix isolation techniques (uideinfra).3 Methods of Generation As nitrile sulphides are transient intermediates prone to decomposition it is necessary for synthetic applications for them to be generated in situ in the presence of the dipolarophile. In this section the various sources will be described and the most efficient identified. There are two general approaches involving either thermal or photolytic reactions. Of these the former group are more useful synthetically, while the latter are suitable for matrix isolation and spectroscopic investigations. A. Thermal Methods.-The most widely used routes are based on cycloreversion of five-membered heterocycles which already contain the C=N-S unit, par- ticularly 1,3,4-oxathiazo1-2-ones.However, acyclic precursors such as thioacyl diphenylsulphimides and alkyliminosulphur difluorides have also been employed.(i) From 1,3,4-0xathiazol-2-ones. Decarboxylation of oxathiazolones (l), the reaction in which benzonitrile sulphide was originally identified, remains the method of choice and has been used for the generation of numerous substituted analogues. The precursors usually have a good shelf-life and are readily prepared from the corresponding carboxamide, either in two stages using perchloromethyl mercaptan followed by formic acid l2 or triethylamine,13 or directly by treatment with chlorocarbonylsulphenyl chloride l4 (Scheme 3). The former route is l3 E. Miillbauer and W. Weiss, Belgian patent, 680644 (1966), British patent 1079348 (1967); Chem.Ahsrr., 1968, 68, 69000. I‘M. M. Kremlev, A. I. Tarsenko, and I. V. Koval, Vop. Khim., Khim. Tekhnol., 1973, 30,40; Chem. Ahstr., 1974,81, 3837. The Chemistry of Nitrile Sulphides assumed to involve hydrolysis of an intermediate dichlorooxathiazole (3). Various substituents can be accommodated including phenols,15-' ester~,'~-'~ nitriles,2092' and alkenes 18,22,23 in addition to simple alkyl and aryl groups. Nucleophiles, e.g. primary and secondary amines, which react with oxathiazolones 24-26 must be excluded. Thermolysis of the parent oxathiazolone (1; R = H)27 affords hydrogen cyanide and isothiocyanic acid rather than formonitrile sulphide. Nitrile sulphides are generated conveniently for cycloaddition reactions by heating the oxathiazolone with an excess of the dipolarophile at 11&-160 "C in an inert solvent (eg.xylene, chlorobenzene). Nitriles and sulphur as by-products are commonly observed even with reactive dipolarophiles; in the absence of dipolarophile the nitrile is formed quantitatively. The decarboxylation involves a thermally allowed ,2, + ,2, + .2, process. The reaction is accelerated by electron-donating substituents indicative of development of partial positive charge in the transition state at the 5-position of the heterocyclic ring.20 Lewis acids such as BF3 also accelerate decomposition but result in lower adduct yields.28 (ii) From 1,4,2-Dithiazo/-5-ones. Closely related to oxathiazolones (1) are dith- iazolones (4) and thiones (5)in which the ring oxygen is replaced by sulphur. By analogy, these are potential sources of nitrile sulphides by extrusion of SCO and CS2 respectively.Thiones (5) have been prepared from thioamides with thiophosgene-carbon disulphide29 or better using perchloromethyl mercaptan 30*3' (Scheme 4). Thione to ketone conversion is achieved by treatment with mercury(1r) acetate,30 KMn04,32 or benzonitrile oxide.30 Dithiazolones (4) are more stable than the corresponding oxathiazolones fragmenting only slowly (75-200 h) to nitrile, sulphur, and C02 in refluxing mesitylene. In the presence of DMAD isothiazoles are formed in good yield, implicating nitrile sulphides as intermediate^.^^ The more forcing conditions l5 A. Senning and J.S. Rasrnussen, Acta Chem. Scand., 1973,21,2161. l6 P. A. Brownsort and R. M. Paton. f.Chem. Soc., Perkin Trans. I, 1987, 1339. P. A. Brownsort, PhD Thesis, University of Edinburgh, 1987. l8 P. A. Brownsort, R. M. Paton, and A. G. Sutherland, Tetrahedron Lett., 1985,26,3727. l9 R.K. Howe, T. A. Gruner, and J. E. Franz, J. Org. Chem., 1977.42, 183. 2o R. K. Howe, T. A. Gruner, L. G. Carter, L. L. Black, and J. E. Franz, J. Org. Chem., 1978,43,3736. R. K. Howe and B. R. Shelton, f.Org. Chem., 1981,46,771. 22 W. Beck, E. Leidl, M. Keubler, and U. Nagel, Chem. Ber., 1980, 113, 1790. 23 R. M. Paton, 1. Stobie, and R. M. Mortier, Phosphorus Sulfur, 1983, 15, 137. 24 G. Westphal, A. Weise, and A. Otto, Z. Chem., 1977, 17,295. 25 A. Rajca, D. Grobelny, S.Witek, and M. Zbirovsky, Synthesis, 1983, 1032. 26 M. C. McKie, PhD Thesis, University of Edinburgh, 1988. 27 (a) B. Bak, 0. J. Nielsen, and H. Svanholt, f. Mol. Spectroscop., 1977, 68 169; (6) B. Bak, J. J. Christiansen, 0.J. Nielsen, and H. Svanholt, Acta Chem. Scand., 1977,31A, 666. 28 R. K. Howe and J. E. Franz, J. Org. Chem., 1974,39,962. 29 H. Behringer and D. Deichmann, TefrahedronLeft.,1967,1013. 3o D. Noel and J. Vialle, Bull. SOC.Chim. Fr., 1967, 2239. 31 D. J. Greig, M. McPherson, R. M. Paton, and J. Crosby, f.Chem. SOL'.,Perkin Trans. 1, 1985, 1205. 32 M. S. Chauhan and D. M. KcKinnon, Can. J. Chem., 1976,54,3879. 33 D. J. Greig, R. M. Paton, J. G. Rankin, J. F. Ross, and J. Crosby, Tetrahedron Left.. 1982, 5453. Paton cL3CSCIRCSNH 2 -*CA heat I -sco Scheme 4 required and less straightforward access to the dithiazolones makes them a less attractive source.The reaction between dithiazolethione (5) and DMAD takes a different course yielding dithiolethione (6) rather than nitrile sulphide-derived products. Direct interaction between thione (5) and the dipolarophile in a cyclo-substitution reaction is the most likely mechanism. (iii) From 1,3,4-Oxathiazoles and 4,SDihydro- 132,4-thiadiazoles. 1,3,4-Oxathiazoles (7),34 prepared at 135 “C from nitrile sulphides and activated carbonyl com-pounds (e.g. C13CCOCC13, CF3COPh), decompose slowly at higher temperatures (ca. 160 “C) regenerating the original carbonyl compound and nitrile sulphide, which can be trapped with alkynes and nit rile^.^' The rate of this retro-1,3- dipolar cycloaddition is critically substituent-dependent, being rapid for R’ = R” = CC13 and slow for R’ = H, R” = CC13.Benzonitrile sulphide is also a likely intermediate in the reported 36 thermal decomposition of the fluorene-8-spiro derivative (7; R = Ph, R’R’’= C13H10), which was prepared by cyclization of fluorenethione S-benzoylimide. Similar behaviour is observed 37 for 4,5-dihydro-1,2,4-thiadiazoles(8), the cycloadducts of nitrile sulphides with imines (Scheme 5). Prolonged heating at 160 OC results in cycloreversion to the nitrile sulphide which can be trapped with DMAD and ethyl cyanoformate (ECF). With few alternative routes available to such oxathiazoles and dihydro- thiadiazoles they offer little advantage synthetically.34(a)R. M. Paton, J. F. Ross, and J. Crosby, J. Chem. SOC.,Chem. Commun., 1979, 1146; (b) A. M. Damas, R. 0.Could, M. M. Harding, R. M. Paton, J. F. Ross, and J. Crosby, J. Chem. SOC.,Perkin Trans. I, 1981,2991. 35 (a)R.M. Paton, F. M. Roberton, J. F. Ross, and J. Crosby, J. Chem. SOC.,Chem. Cornmun., 1980,714; (6) R. M. Paton, F. M. Robertson, J. F. Ross, and J. Crosby, J. Chem. SOC.,Perkin Trans. 1, 1985, 1517. 36 (a) E. M. Burgess and H. R. Penton, J. Am. Chem. SOC.,1973, 95, 279; (b) J. Org. Chem., 1974, 39, 2884. 37 R. 0.Could, R.M. Paton, J. F. Ross, M. D. Walkinshaw, and J. Crosby, J. Chem. Res., 1986, (S) 156, (M)1372. The Chemistry of Nitrile Sulphides R'RT =O ArCH=NPh Ph +-RrNXHRCZN-S NLS Ar (7) Scheme 5 r" F Scheme 6 S S0 4.-PhS /\II +-ArCN=SPh2 4-b ArC=N-SPhz +ArC=N +ArC=N-Su Scheme 7 (iv) From (AIky1imino)sulphur D$uorides. (Benzy1imino)sulphur difluoride, prepared from benzylamine and SF4, on heating with NaF and 18-crown-6- polyether at 130°C in the presence of DMAD yields isothiazole (2a).38-40 The reaction is believed to involve 1,3-elimination of two moles of HF to form benzonitrile sulphide as illustrated in Scheme 6. Acetonitrile sulphide and trifluoroacetonitrile sulphide have also been generated by this appr~ach.~' (v) From N-Thioacyl Diphenylsulphinimides. Thermolysis at 50-70 "C of N-thioaroyl diphenylsulphimides, prepared by treatment of diphenylsulphimide with methyl dithiobenzoates, affords the corresponding nitriles together with diphenyl sulphide and sulphur.In the presence of electron-poor alkynes, isothiazoles are formed 42-presumably via thiazirine and nitrile sulphide intermediates (Scheme 7). B. Photochemical Methods.-Benzonitrile sulphide has been invoked 43-53 as a transient intermediate in photofragmentation reactions of various five-membered heterocycles incorporating C, N, and S (Scheme 8). In each case the process is believed to involve extrusion of a small inorganic fragment (C02, COS, CS2, NZ, N20, HNCO) forming the unstable 47c-antiaromatic thiazirine (9), followed by rearrangement to the corresponding nitrile sulphide. Spectroscopic evidence for these intermediates is presented in Section 5.38 J. R.Grunwell and S. L. Dye, Terrahedron Let[., 1975, 1739. 39 M. J. Sanders, S. L. Dye, A. G. Miller, and J. R. Grunwell, J. Org. Chem., 1979,44, 510. 40 M. J. Sanders, Diss. Absfr.,1979,40B, 1181; Chern. Abstr., 1980,92, 5782. 41 M. J. Sanders, and J. R. Grunwell, J. Org. Chem., 1980,45,3753. 42 H. Yoshida, H. Taketani, T. Ogata, and S. Inokawa, Bull. Chem. SOC.Jpn., 1976,49, 3124. Paton (1)X = 0 (10) (SIX = s I (11)X = 0,Y = co (91 (12)X = 0,Y = cs (13)X = 0,Y = CNH 1hV (1L)X = C0,Y = co (1S)X = C0,Y = s (161X = CS,Y = S -xy IE = C02Me) (17) X = Y = N (18) X = N0.Y =N Scheme 8 In the presence of DMAD the expected isothiazoles (2) are formed; yields are generally poor, ranging from 5% for dithiazolethione (16) to 29%for mesoionic oxathiazolone and are invariably lower than those for the thermal methods described above.Nitrile sulphide-derived products are also formed on 43 H. Gotthardt, Terrahedron Lerf., 1971,1277. 44 A. Holm, N. Harrit, K. Bechgaard, 0. Buchardt, and S. E. Harnung, J. Chem. Soc., Chem. Commun., 1972, 1125. 45 H. Gotthardt, Chem. Ber., 1972, 105, 188. "A. Holm, N. Harrit, and N. H. Toubro, J. Am. Chem. Soc., 1975,97,6197. 47 I. R. Dunkin, M. Poliakoff, J. J. Turner, N. Harrit, and A. Holm, Tefrahedron Left., 1976,873. 48 A. Holm, N. Harrit, and N. H. Toubro, Tetrahedron, 1976,32,2559. 49 A. Holm, N. Harrit, and I. Trabjerg, J. Chem. Soc., Perkin Trans. I, 1978, 746. 50 A. Holm and N. H. Toubro, J.Chem. Soc., Perkin Trans. I, 1978, 1445. 51 A. Holm, J. J. Christiansen, and C. Lohse. J. Chem. Soc., Perkin Trans. I, 1979,960. 52 H. Gotthardt, F. Reiter, and K. Kromer, Liebigs Ann. Chem., 1981, 1025. 53 N. Harrit, A. Holm, I. R. Dunkin, M. Poliakoff, and J. J. Turner, J. Chem. SOC.,Perkin Trans. 2, 1987, 1227. The Chemistry qf Nitrile Sulphides photolysis of 3-ethyl-5-phenyl-l,2,3,4-thiatriazolium By-pro-tetrafluor~borate.~~ ducts are commonplace: e.g. isothiocyanates from thiatriazoles (17) and ethyl benzoylformate from mesoionic compound (10).These are formed in addition to benzonitrile, the expected photodecomposition product of benzonitrile sulphide. Thermolysis, rather than photolysis, of compound (10) with DMAD yields the isomeric isothiazole-3,4-dicarboxylate(19) resulting from cycloaddition to the thiocarbonyl imine Photofragmentation of 3,4-disubstituted 1,2,5- thiadiazoles affords nitriles and sulphur consistent with nitrile sulphide formati01-1;~~however, attempts to trap the 1,3-dipole proved unsuccessful. C.Miscellaneous Methods.-Nitrile sulphides have been proposed as inter-mediates in the oxidative dimerization of thioamides to 3,5-disubstituted- 1,2,4- thiazdiazole~.~~-~~Although nitriles are often formed as by-products and such thiadiazoles are formally 1,3-dipolar cycloadducts of nitrile sulphides and nitriles, in most cases there is scant evidence for the reaction proceeding in this way. Indeed, the low dipolarophilicity of unactivated aromatic nitriles suggests that nitrile sulphides are not involved and that alternative oxidative pathways are more likely.4 Reactions Whereas some nitrile oxides can be isolated and characterized at room temperature all nitrile sulphides reported so far are thermally unstable and can only be detected using low-temperature matrix isolation techniques. Rapid desulphuration to the corresponding nitrile occurs in solution. However, genera- tion in situ in the presence of suitably activated dipolarophiles allows 1,3-dipolar cycloaddition reactions to be accomplished. Optimum yields are obtained by using an excess of dipolarophile and maintaining a low concentration of the nitrile sulphide thus minimizing decomposition, a process for which the mech- anism has not been firmly established but which is known to be higher than first order.The range of established cycloaddition reactions is limited compared with the other nitrilium betaines. Dipolarophiles examined to date and described below comprise alkynes, alkenes, nitriles, imines, carbonyl compounds, and phospha- alk ynes (Scheme 9). The mechanisms of desulphuration and cycloaddition are discussed in Section 6. 54 H. Gotthardt, Tetrahedron Lett., 1971, 1281; Chem. Ber., 1972, 196. 55 T. S. Cantrell and W. S. Haller, J. Chem. Soc., Chem. Commun., 1968, 977. M. Bahadir, S. Nitz, H. Parlar, and F. Korte, J. Agric. Food Chem., 1979, 27, 815.''M. T. M. El-Wassimy, K. A. Jorgensen, and S.-0. Lawesson, Tetrahedron, 1983,39, 1729. 58 M. Machida, K. Oda, and Y.Kanaoka, Tetrahedron Lett., 1984,25,409. 59 Y. Takikawa, K. Shirnada, K. Sato, S. Sato, and S. Takizawa, Bull. Chem. Soc. Jpn., 1985,58,995. Paton (261 Ph Rrox"' NxS R" (301 (71 I8 1 Scheme 9 formed in yields as high as 96%,20 and other electron-poor alkynes (e.g. dibenzoylacetylene 42) react similarly. Alkyl propiolates give a mixture of 4-and 5-carboxylates, the isomer ratio depending on both the nitrile sulphide and its source. For benzonitrile sulphide generation from oxathiazolone (1a) affords nearly equal amounts of the two regioisomers, whereas there is a slight preference for the 4-substituted product using benzyliminosulphur difluoride as precursor,39 possibly due to protonation of the dipolarophile by hydrogen fluoride, the co- product.Of examples examined so far only trifluoroacetonitrile sulphide shows pronounced regioselectivity, the 5-carboxylate predominating (5 :1). This general lack of regioselectivity contrasts with the behaviour of nitrile oxides for which 5-carboxylates are often the major product^.^ Isothiazole-4-carboxylates can also be prepared by thermal decarboxylation of isothiazole-4,5-dicarboxylates.20*62 The yields of cycloadducts decrease as the alkyne becomes less electron poor: e.g. ca. 50% for ethyl phenylpropiolate,20 traces for ethyl butynoate,20 and only nitrile and sulphur are formed with diphenyla~etylene.~~ ''R. K. Howe and J. E. Franz, J. Chem. SOL..,Chem. Commun., 1973,524. 61 (a) R. M. Mortier, R.M. Paton, and I. Stobie, J. Chem. Soc., Chem. Commun., 1983, 901; (b) R. M. Mortier, R. M. Paton, G. Scott, and I. Stobie, Br. Polymer J., 1987, 19, 303. 62 M. C. McKie and R. M. Paton, J. Chem. Res., 1987, (S)245, (M),2051. 41 The Chemistrj-of' Nitrile Sidphides E (21 1 Scheme 10 B. Cycloaddition to A1kenes.-Whereas the cycloaddition of nitrile sulphides to acetylenes described above is one of several synthetic approaches to isothiazoles the corresponding reaction with olefins is the only method developed so far for 2-isothiazolines (4,5-dihydroisothiazoles). The process is most efficient for electron- poor alkenes, although norbornene-type unsaturation also shows adequate reac- ti~ity.~~,~~ With dialkyl fumarates 2-isothiazoline-trans-4,5-dicarboxylates(20) are formed in SO-SO% The stereochemistry of the products is evident from their 'H n.m.r.spectra which show couplings for H(4)-H(5) of ca. 4 Hz, comparable with those observed for analogous 2-isoxazoline-trans-4,5-dicarboxylates.For diethyl 3-phenyl-2-isothiazoline-~r~ns-4,5-dicarboxylatethe structure has been confirmed by X-ray cry~tallography.~~ N-Phenylmaleimide 41963 and maleic anhydride 38 afford the expected cis-adducts with &(4)H(5) 11 Hz. In contrast dialkyl maleates yield trans-isothiazolines (20),identical to those formed from fumarate ester~.*~,~~ The isomerization may occur in an initially-formed cis-adduct (21) (Scheme 10) or in the dipolarophile prior to cycloaddition. The corresponding reaction of nitrile oxides with maleate esters also give trans-products at elevated temperatures; the cis-adduct, which can be isolated at room temperature, isomerizes on heating.65 The yield of cycloadduct, as with acetylenes, is dependent on the nature of the dipolarophile.Acrylate esters afford a mixture of 2-isothiazoline-4- and 5-carboxylates with little regio~electivity.~~.~~No reaction is observed with tetraethyl ethenetetra~arboxylate,~~presumably because of steric effects, or with the double bond of tetracyanoethene; instead 1,2,4-thiadiazoles are formed by cycloaddition to one or more of the nitrile groups.63 P-Nitrostyrene and 3-nitrostyrene are similarly unreactive. Norbornene-type double bonds, which are strong dipolarophiles towards nitrile 63 R.K. Howe and J. E. Franz, J.Org. Chem., 1978,43.3742. 64 J. F. Ross, PhD Thesis,,University of Edinburgh, 1981. 65 A. Rahman and L. Clapp, J. Org. CIiem., 1976,41, 122. Paton N H R 0 0 (22 1 (231 (2L1 +-Ar ArCrN-S Na Scheme I1 oxide^,^ also react with nitrile sulphides. em-Adducts (22) have been isolated from norbornene itself 62 and from dimethyl 5-norbornene-cis,endo-2,3-dicarb-oxyla te. 63 The reaction of nitrile sulphides with alkenes can be used as a route to isothiazoles. 2-Isothiazolines are readily dehydrogenated (DDQ63 or NaOCl 64) and, in some cases, isothiazoles are the isolated products from alkene-nitrile sulphide reactions. For example, isothiazoloquinones, (23) and (24), are formed from p-naph thoquinone and p-benzoquinone respectively, presumably by in situ oxidation by excess quinone of an intermediate isothiazoline.66 Ethyl x-chloro- and P-pyrrolidinyl-acrylates both afford isothiazoles via cycloaddition followed by elimination of HCl and pyrrolidine re~pectively.~~ 3-Arylisothiazoles (25) unsubstituted at both 4- and 5-positions, formally adducts between nitrile sulphides and acetylene, are produced using nor-bornadiene which acts as an acetylene equivalent.By analogy with nitrile oxides and imines the process is presumed 62 to involve the cycloaddition-cycloreversion pathway illustrated in Scheme 11. Isothiazoles (25) can also be prepared by thermal extrusion of C2H4and C02 from 3-arylisothiazole 3-and 4-carboxylates and 4,5-dicarboxylates under flash vacuum pyrolysis conditions.62 C.Cycloaddition to Nitri1es.-Reaction of nitrile sulphides with nitriles provides a versatile preparative route to 1,2,4-thiadiazoles [(26), Scheme 91, an important class of heterocycles showing a range of biological activity.' Whereas traditional methods such as oxidation of thioamides are well suited for symmetrically substituted derivatives,' e.g. 3,5-diphenyl- 1,2,4-thiadiazole is readily prepared from thiobenzamide, the nitrile sulphide cycloaddition approach allows controlled introduction of different substituents at the 3- and 5-positions.'6~'9~21~23,28~61~67~68 ''R. M. Paton. J. F. Ross, and J. Crosby, J. Chem. Soc., Ciiem. Cornmun.. 1980, 1194. 43 The Chemistry qf' Nitrile Sulphides S0 t-HO H HO OH BZO 062 ( 27 1 (281 (29 1 Nitriles are surprisingly reactive towards nitrile sulphides.For most dipolaro- philes the yields of cycloadduct obtained from nitrile sulphides are much less than those from nitrile oxides. Indeed, cycloadditions involving nitrile sulphides generally require strongly activated dipolarophiles. Towards nitriles, however, the reactivities of these two nitrilium betaines are more comparable although still favouring nitrile oxides. As with other dipolarophiles the yield of cycloadducts is strongly substituent-dependent and is greatly increased by electron-withdrawing groups. Ethyl cyanoformate (ECF),'6*23*28*61 and trichloro- ~-ketonitriles,~~.~~ acetonitrile "are the most reactive, readily affording 50-95% of 5-substituted 1,2,4-thiadiazoles.ECF is a particularly useful trap for nitrile sulphides; not only is it highly reactive but it also often affords crystalline adducts and its moderate boiling point (ca.115 "C) facilitates its removal from the reaction mixture. Simple arenenitriles,' 932 dichloroacetonitrile,26 aryl thio- and seleno-cyanates,68 1328*35 and ethyl cyanoacetate 28 give substantially smaller amounts. 1,2,4-Thiadiazoles unsubstituted at the 5-position (26; R' = H), formally the cycloadduct of HCN, are accessible by decarboxylation of their 5-carboxylate derivatives.28 Symmetrically substituted thiadiazoles are sometimes observed 28*64 as by- products. For example 4-chlorobenzonitrile sulphide and benzonitrile yield 3,5- diphenyl-1,2,4-thiadiazole(26, R = R' = Ph) (7%) in addition to the expected 3-(4- chlorophenyl)-5-phenyl compound (26; R = 4-ClCsH4, R' = Ph).Its formation is attributed 28 to the presence of benzonitrile sulphide arising either by sulphuration of benzonitrile by S, (n = 1-8) or by sulphur atom transfer between 4- chlorobenzonitrile sulphide and benzonitrile. Recent applications include the synthesis of the cytotoxic marine natural product dendrodoine (27) from dimethylaminoformonitrile sulphide and indole-3- ~arbonitrile,~~and the ribovasin analogue (28) 70 via the protected P-D-ribo- furanosylcarbonitrile sulphide (29) and ECF. D. Cycloaddition to Carbonyl Compounds.-l,3,4-Oxathiazoles [(7), Scheme 91, the adducts of nitrile sulphides and carbonyl are a rare class of heterocycles accessible only with difficulty by other ' The cycloaddition occurs only when the carbonyl group is activated by electron-withdrawing 67 D. J.Greig, M. McPherson, R. M. Paton, and J. Crosby, P/mphorus Sulfur, 1986, 26, 151. D. J. Greig, D. G. Hamilton, M. McPherson, R. M. Paton, and J. Crosby, J. Chem. Soc., Perkin Trans. I, 1987. 607. h9 I. T. Hogan and M. Sainsbury, Tetrahedron, 1984.40,681. 70 D. K. Buffel, B. P. Simons. J. A. Deceuninck. and G. J. Hoornaert, J. Org. Chetn., 1984,49, 2165. 44 Paton groups, e.g. hexachloroacetone, chloral, x,a,a-trifluoroacetophenone34 and methyl benz~ylformate.~The oxathiazoles, like the oxathiazolones from which they are formed, have a planar heterocyclic ring with a localized C=N double bond.Dehydrochlorination of chloral-derived compound (7; R’ = H, R” = CC13) affords dichloromethylene derivative (7; R’R” = CCI2), which is formally the adduct between a nitrile sulphide and C=O of dichloroketene. All the oxathiazoles are themselves thermally labile and cyclorevert 35 on heating to nitrile sulphide and carbonyl fragments (Scheme 5). Subsequent decomposition of the nitrile sulphide gives the corresponding nitrile quanti- tatively. E. Cycloaddition to 1mines.-Schiff bases formed from aniline and para-substituted benzaldehydes cycloadd to nitrile sulphides affording 3,4,5-trisubstituted-4,5-dihydro-1,2,4-thiadiazoles[(8), Scheme 91. Yields for the few reported examples 37 are very low (<13%) even when the para-group is electron withdrawing. Although various 4,5-dihydro- 1,2,4-thiadiazoles have been known for many years, e.g.Hector’s base and Dost’s keto compound,73 previous examples have sp2-hybridization at C(5) and a near planar heterocyclic ring. In contrast, for the nitrile sulphide-derived analogues (8) both C(5) and N(4) are sp3-hybridized resulting in a fold in the ring about the S(l)-N(4) vector. Their thermolytic behaviour closely parallels that of the 1,3,4-oxathiazoles described above. On heating they undergo retro- 1,3-dipolar cycloaddition to nitrile sulphides and imines (Scheme 5). This system has not been studied in depth and the imines used so far have been of low reactivity.By inclusion of electron-attracting groups in the dipolarophile component greater yields may be expected. F. Cycloaddition to Phosphaa1kynes.-t-Butylphosphaacetylene (Bu’C=P) is a sufficiently strong dipolarophile to cycloadd to several nitrilium betaines including benzonitrile sulphide.74*75 Thermal decarboxylation of phenyloxa- thiazolone (1a) in the presence of Bu‘CrPaffords 5-t-butyl-3-phenyl- 1,2,4-thiazapha- sphole [(30), Scheme 91 (82%), the first example of this class of heterocycle. Traces of compound (30) are also formed from mesoionic precursor (lo), accompanying the 3-t-butyl-5-phenyl isomer, which is the expected mesoionic cycloaddition G. Intramolecular Cycloaddition Reactions.-Intramolecular 1,3-dipolar cycloaddi- ”(a) N.Haake, B. Eichenauer, and K. H. Ahrens, Z. Naturforsch., 1974,29B, 284; (b)W. S. Levchenko, L. V. Budnik, V. N. Kalinin, and A. A. Kisilenko, Zh. Org. Khim., 1982, 18, 2549; Chem. Abstr., 1983, ’* 98, 125987. H. C. Gibbard and R. M. Paton, unpublished observations. 73 (a) A. R. Butler, C. Glidewell, and D. C. Liles, Am Crystalfogr., 1978, 34B, 3241; (b) A. F. Cuthbertson, C. Glidewell. H. D. Holden, and D. C. Liles, J. Chem. Res., 1979, (S) 316, (M) 3714 and references therein. 74 W. Rosch and M. Regitz. Synthesis, 1987, 689. 75 W. Rosch, H. Richter, and M. Regitz, Chem. Ber., 1987, 120, 1809. The Chemistry of Nitrile Sulphides N--s -2H . 0 (311 (32 1 Scheme 12 tion reactions find widespread application for the construction of fused hetero- cyclic systems and examples have been reported 76 for most 1,3-dipoles including the nitrilium betaines.Whereas nitrile oxides have been examined in great detail, particularly for natural product synthesis,'" it is only recently that the first cases involving nitrile sulphides have been described.'* o-Cinnamoyloxybenzonitrile sulphides (3 l), which incorporate an adjacent 1,3-dipole and activated dipolarophile, were generated by thermal decarboxylation of the corresponding oxathiazolones. They yield isothiazolocoumarins (32), pre- sumably uia the intramolecular cycloaddition-dehydrogenation pathway illu- strated in Scheme 12. The same fused isothiazoles (32) are also formed directly from benzonitrile sulphides bearing o-arylpropiolyloxy substituents.A similar approach has been used to form 3-phenylisothiazolo[4,3-c]quinoline-4-(5H)-ones from phenylpropiolamidobenzonitrile sulphide. H. Polymeric Nitrile Su1phides.-The cycloaddition reactions of polymeric 1,3- dipoles are an ideal means of preparing polymers bearing pendant heterocycles. Examples include nitrile imine~,~~ and nitrile sulphides.61 Free radical- azide~,~' initiated polymerization of a-alkenyloxathiazolones (33) with styrene or methyl methacrylate (MMA) affords 61,79 oxathiazolone-containing polymers (34). Ther-mal decarboxylation generates polymer-bound nitrile sulphides which are effici- ently trapped6' in the presence of reactive dipolarophiles such as DMAD, ECF, and chloral (Scheme 13).An alternative route to such polymer-linked heterocycles involves initial cycloaddition of a-alkenylnitrile sulphides 23 to form adducts (35) with polymerizable substituents, and subsequent co-polymerization with styrene or MMA. The latter method has the advantage of yielding products free from contamination by nitriles which, unlike the corresponding monomer reactions, are necessarily retained in the polymer when the former approach is used. Bis-oxathiazolones [eg. 1,2-bis(2-oxo-1,3,4-oxathiazol-5-yl)-ethane,a potential precursor of succinonitrile sulphide] have also been utilized 'O to vulcanize styrene-butadiene rubbers. It is doubtful that the process involves cycloaddition between nitrile sulphides to the alkene units. Whereas the latter are known to l6 A.Padwa, Ref. 1, Ch. 12. "J. K. StiIIe, Mcrkrornol. Cfrem.,1972, 154, 49. H. L. Cohen, f.Po1j.m. Sci., Polyni. Chem. Ed., 1981, 19. 1337. R. M. Paton, I. Stobie, and R. M. Mortier, J. Polym. Sci.. Polym. Lett. Ed, 1983, 21, 145. J. Crosby, British patent 1509406 (1975); Chem. Ahsfr.. 1977.86, 156824. Paton -c02-(351 Scheme 13 form 81 bis-isoxazoline crosslinks with dinitrile oxides, it is unlikely that they would be sufficiently good dipolarophiles to react with nitrile sulphides. A mechanism involving vulcanization by a reactive form of sulphur resulting from decomposition of nitrile sulphides is more probable, a hypothesis supported by the observation that monofunctional analogues such as compound (la) also effect crosslinking.5 Matrix Isolation and Spectroscopic Detection The thermal instability of nitrile sulphides prevents spectroscopic investigation at room temperature. They have however been detected when generated photolyti- cally at cryogenic temperatures using matrix isolation techniques. Irradiation (400-470 nm) of 4-phenyl-l,2,3-oxathiazolylium-5-olate(10) at 85 K in diethyl ether-isopentane-ethanol (EPA) and isopentane-methylcyclohexane (MPH) glasses or a PVC film results in formation of benzonitrile sulphide as one of the primary photopr~d~~t~.~~~~~~~~~~~Its U.V. spectrum shows characteristic h,,,. values at 335 (E = 6.7 x lo3 1.mol-' cm-'), 324, 313, 295, and 240 nm. In EPA the spectrum remains unchanged on warming to 140 K at which point the glass melts and the nitrile sulphide decomposes to benzonitrile (70%).In PVC traces of benzonitrile sulphide are still detectable at room temperature. Fragmenta- tion to nitrile and sulphur is also accomplished by irradiation at 335 nm, one of the absorption maxima. The same U.V. spectrum is obtained on photolysis of 5-phenyl- l,Z3,4-thiatriazole (17),46349 its 3-oxide (18),49and oxathiazolone (la).49 Benzonitrile sulphide has also been detected by i.r. spectro~copy.~~ After irradiation of mesoionic compound (10) in PVC a peak is observed at 2 185 J. Crosby, R. M. Paton, R. A. C. Rennie, and J. Tanner, British patent. 147691 (1973); Cliem. Absir., 1975.82, 141047. The Chemistry of Nitrile Suiphides path (a) t-b ArCrN-S -0CRR' (a,RR'=O ; b ,R= H,R'= CC13; path(b) DMAD C, R= R'=CCl,;d,R= Ph,R'=CF3) DMAD Scheme 14 cm-'.It is assigned to benzonitrile sulphide as it appears before benzonitrile, it is efficiently photolysed at 335 nm, and it behaves during warm-up in PVC as benzonitrile sulphide does when monitored by U.V. spectroscopy. The value of 2 185 cm-' compares with 2 288 cm-' for benzonitrile oxide,82 2 228 cm-' for benzonitrile N-~henylimide,~~and 1 926 cm-' for benzonitrile meth~lide,'~ suggesting a nitrilium betaine-like structure (RCrN+-X-) comparable to those assigned to nitrile oxides and imines, rather than the allene-like skeleton of nitrile ylides (RC-=N+=CH2). Formation of benzonitrile sulphide from compounds (la), (lo), (17), and (18) is believed45-53 to proceed via thiazirene (9) (Scheme 8), and evidence for the existence of this unstable species in PVC at 10-15 K has been pre~ented.~' Although it has not been detected directly its presence is implied by the formation of benzonitrile sulphide on heating (204 140 K) a sample of thiatriazole (17) which had previously been photolysed at 10 K.6 Mechanistic Aspects A. Evidence for Nitrile Sulphides.-Evidence for the existence of nitrile sulphides is based on spectroscopic measurements of matrix-isolated samples at cryogenic temperatures (Section 5) and trapping reactions in solution. Formation of isothiazoles and 1,2,4-thiadiazoles respectively from reactions of oxathiazolones (36a) with DMAD and ECF, both well-established dipolarophiles, 82 R.H. Wiley and 8.J. Wakefield, J. Org. Chem., 1960,254 546. R3 N. H. Toubro and A. Holm, J. Am. Chent. Sw., 1980,102,2093. 84 (a)0. L. Chapman and J.-P. Le Roux, J. Am. Chem. Soc., 1978, 100, 282; (h) E. Orton, S. T. Collins, and G. C. Pimentel, J. Phj~s.Chem., 1986, 90, 6139. Paton strongly indicates nitrile sulphides as intermediates. It does not however, prove that they are involved and alternative pathways to the observed products must be considered. Mechanisms involving zwitterions can be discounted in view of the isolation of both regioisomers from reactions with propiolate esters.20 However, cycloadduct formation is consistent not only with intermediate formation of nitrile sulphides (Scheme 14, path a), but also with a mechanism (path b) involving direct interaction between the dipolarophile and precursor (36) to form adduct (37), followed by elimination of COZ. That path (a) involving a discrete nitrile sulphide intermediate is operative rather than path (b) is demonstrated both by kinetic measurements and by competition experiments.Howe et ~1.~'established that the rate constant for the consumption of oxathiazolone (36a, Ar = Ph) is first order and independent of concentration of DMAD. Furthermore, the rate constants for the formation of isothiazole and benzonitrile, the by-product of the reaction, are also first-order and equal to the rate constant for the disappearance of the oxathiazolone. Complementary evidence is provided by competition experiments performed by Ross et Thermolysis of four sources, oxathiazolone (36a, Ar = 4-MeOCsH4) and three oxathiazoles (36b-d, Ar = 4-MeOC6H4), in the presence of ethyl propiolate as trapping agent afforded the same regioisomeric ratio (1.33 & 0.02) of isothiazole-4- and 5-carboxylates.These results imply generation of a common intermediate (4-MeOC6H4C=N +-S-) from each source. The involvement of nitrile sulphides as discrete intermediates in these reactions is thus firmly estab- lished. B. Theoretical Treatment.-Frontier molecular orbital (FMO) theory has fre- quently been used to rationalize the reactivity of nitrilium betaines and the regioselectivity of their cycloaddition reactions.' For nitrile ylides the process is regarded as dipole-HOMOclipolarophile-LUMO controlled, i.e.type I in the Sustmann clas~ification.~~ Nitrile imines and oxides are categorized type I1 where both sets of frontier orbitals (dipole-HOMO-dipolarophile-LUMO and dipole- LUMO-dipolarophile-HOMO) must be considered. Nitrile sulphides have not been examined rigorously from a theoretical point of view. However, some rationalizations and predictions have been made on simple electronegativity grounds and on the basis of CND0/2 SCFMO calculations. Since sulphur is little more electronegative than carbon, nitrile sulphides should have LUMO and HOMO energy levels only slightly lower than those of nitrile ylides. By analogy they are expected to react readily with electron-poor acetylenic esters with predominantly dipole HOMO control.The observed 2o order of reactivity for such esters (DMAD > ethyl propiolate > ethyl phenyl- propiolate > ethyl but-2-ynoate) is consistent with this hypothesis. CND0/2 SCFMO calculations 39 performed on benzonitrile sulphide, with geometry optimization of the N-S bond and standard geometry for the rest of the molecule, lead to broadly similar conclusions. The energy difference between the ''R.Sustrnann, Puw Appl. Clicvn.. 1974, 40. 569. The Chemirtr?, of' Nitrile Sulphides 0.225 0.140 -0.226 -0.138'1.052 1.140 1.678 Figure 1 HOMO of benzonitrile sulphide and the LUMO of electron-deficient acetylenes is calculated to be ca. lev less than the alternative dipole-LUMO-dipolarophile-HOMO interaction.Nitrile sulphides are therefore expected to show more dipole-HOMO control than the corresponding nitrile oxides. Explanation and prediction of regioselectivity are necessarily less precise. The CNDO/2 calculations 39 indicate that the largest orbital coefficient is associated with sulphur in the HOMO but with carbon in the LUMO. Cycloaddition to propiolate esters is thus expected to afford predominantly isothiazole-4-carb- oxylates under dipole-HOMO control and 5-carboxylates under dipole-LUMO control. Whereas benzonitrile oxide reacts with methyl propiolate to give mainly isoxazole-5-carboxylate (ca. 4 : 1),4 the same alkyne with benzonitrile sulphide affords a 50/50 mixture of regioisomers. Although some lack of selectivity can be attributed to the higher temperatures used for nitrile sulphide reactions, these results are also consistent with the CND0/2 prediction of increased dipole- HOMO-dipolarophile-LUMO interaction for nitrile sulphides compared with nitrile oxides.Ab irzitio calculations have been performed 27h*5 on the parent, but as yet unidentified, formonitrile sulphide (HCNS). Geometry optimization indicates a linear structure with the bond lengths and atomic charges shown in Figure 1. The estimated N-S bond order is 1.0. It is predicted to be less stable than the known isomeric isothiocyanic acid (HNCS) (-490.153 44 us. -490.212 38 hartrees). The calculations also suggest a value of ca. 5.4 D for the dipole moment. C. Desu1phuration.-The thermal and photochemical behaviour of nitrile sul-phides is very different from that of nitrile oxides.The latter rearrange to the isomeric is~cyanates,~ with oxazirenes and acyl nitrenes as likely intermediates. In contrast photolysis of nitrile sulphides at cryogenic temperatures or allowing the sample to warm above 100 K results in desulphuration. The mechanism of this process has been the subject of some debate. Ab initio MO calculations 27b*51 suggest that extrusion of a singlet sulphur atom, S('D), is endothermic (ca. 100 kJ), but formation of atomic sulphur in the ground state, S(3P),is exothermic (en. 40 kJ) although spin-forbidden. The rate of decay is strongly concentration-dependent. The process appears to be unimolecular in dilute ethan01,~' but substantial deviations from first-order kinetics occur in non-polar solvents and at higher concentration.Flash photolysis studies in ethanol at 10 -30 "C enabled decomposition rates and relative stabilities of several arenenitrile sulphides (4-XC6H4CNS) to be examined. Half- life times varied from 0.21 ms for X = Br to 0.87 ms for X = MeO. A Hammett p-value of 2.2 was obtained for a limited range of substituents (X = H, Me, MeO) indicating stabilization of the nitrile sulphides by electron-donating groups. Paton The following observations provide evidence for higher order reactions. Firstly, isolated samples are stable only in an inert matrix.53 In EPA glass the compound is consumed as soon as the glass melts (140 K); in PVC it is still detectable at room temperature.Secondly, the yields of 1,3-dipolar cycloadducts are re-markably dilution-dependent; for example, the yield of 3,5-diphenyl- 1,2,4-thiadi- azole from benzonitrile sulphide in neat benzonitrile increases from 14% at reactant ratio 1:lO to 74% at 1:lOO. Moreover, portionwise addition of the precursor 21 or slow delivery by means of a syringe pump 62 also raises the yield substantially. These results are not consistent with a unimolecular process as the sole mechanism for the decomposition of nitrile sulphides. Reactions that may account for the deviations from first-order kinetics include bimolecular collisions of nitrile sulphides and scavenging by sulphur atoms or short sulphur chains (Scheme 15). These observations are of considerable significance for synthetic applications. By maintaining a low concentration of nitrile sulphide, respectable yields of cycloadducts can be achieved for less reactive dipolarophiles.7 Nitrile Selenides Much less is known about nitrile selenides (RCrN+-Se-) than their sulphur analogues. There is good evidence 86*87 for their existence at low temperatures (<100 K), but so far attempts to trap them as 1,3-dipolar cycloadducts, e.g. with DMAD, have failed. The approach used to detect these short-lived species was very similar to that successfully employed for isolating and studying nitrile sulphides. On photolysis of 3,4-diphenyl-l,2,5-~elenadiazole(38a) at 85 K in PVC film or EPA glass, or at 20 K in a nitrogen matrix, a transient was observed with A,,,.at ca. 255, 325, 360, and 390 nm. An i.r. peak at 2 200 cm-I in N2 (2 190 cm-' in PVC) was also detected and similar spectra were obtained from the isomeric 3,Sdiphenyl- 1,2,4- selenadiazole. The thermal and photochemical behaviour of the transient is consistent with benzonitrile selenide. On heating above 100 K or irradiation (300 or 360 nm) selenium was deposited and the signals attributed to benzonitrile selenide replaced by those of benzonitrile (Scheme 16). Under similar conditions 1,2,5-selenadiazole (38b) and its dimethyl derivative (38c) also afforded short-lived species. These were tentatively assigned to the parent formonitrile selenide (HGN +-Se-) and acetonitrile selenide.The failure of 86 C. L. Pedersen and N. Hacker, Trtruhedron Let(., 1977,3981.''C. L. Pedersen, N. Harrit, M. Poliakoff, and 1. Dunkin, Acrn Chem. Scand., 1977,31B,848. The Chemistrj-(?f'Nitrile Sulphides (38) (a,R= Ph; b,R=H; c,R=Me) Scheme 16 trapping reactions, even with a reactive dipolarophile, can be attributed to nitrile selenides being less stable and more prone to fragmentation than the sulphides. 8 Conclusion The cycloaddition reactions of nitrile sulphides described in this review provide a route to several classes of heterocycles which are accessible only with difficulty by other means. Although additions to most of the common dipolarophiles have been reported, some simple systems have yet to be examined. These include thiones and azo compounds, both of which are reactive towards other nitrilium betaines.There is thus scope for the synthesis of further unusual sulphur- nitrogen heterocycles. The limited reactivity of nitrile sulphides, the high temperatures usually required, together with their tendency to desulphurate are restrictions on their more widespread use. There is therefore a need for alternative sources which retain the easy access and good shelf-life of current precursors, particularly oxathiazolones, but can be used under less forcing conditions. Acknowledgements. I wish to acknowledge the contributions of my co-workers: P. A. Brownsort, D. J. Greig, M. C. McKie, M. McPherson, J. F. Ross, and I. Stobie; to Dr J. Crosby for introducing me to nitrile sulphides and for valuable discussions; and to SERC and ICI plc for financial support.

ISSN:0306-0012    

DOI:10.1039/CS9891800033

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. Soc. Rev., 1989, 18, 53-83 Structural Isomerization in DNA: The Formation of Cruciform Structures in Supercoiled DNA Molecules By David M. J. Lilley DEPARTMENT OF BIOCHEMISTRY, THE UNIVERSITY. DUNDEE DD14HN 1 The Structure of DNA As the repository of genetic information for all life-forms on this planet for billions of years, DNA is arguably the most important macromolecule in the cell. The field of molecular biology has undergone explosive expansion in the past 30 years, much of which can be traced back to the elucidation of the structure of DNA; the heuristic value of this model has been of incalculable value in the development of molecular genetics. Thus the potential importance of structural studies on DNA is hard to overemphasize. While the general architecture of DNA has been known since the fifties -a right-handed double helix with hydrogen bonding between bases to generate sequence-specific base-pairing -the challenge of the last decade has been twofold. First to go into the structure in greater depth, asking how local sequence can alter the structure at the level of the single base-pair, and second to understand how certain proteins recognize and bind specific sequences contained within long DNA molecules.These two questions are probably closely related. Very significant progress has been made in answering both questions, and I shall attempt to review the progress made in one aspect of the former. 2 Structural Variation in DNA It has long been known from X-ray fibre diffraction studies that DNA can exhibit structural polymorphism as a function of base composition and sequence, ionic composition, and water activity.Two main right-handed helical structures (or, more accurately, families of structures) were described,*v3 the A-and B-structures depicted in Figure 1. Differences between the two are obvious to a glance. The B- structure is relatively slim, with two grooves of about equal depth (which is to say that the helix axis is located at the centre of the structure), where the planes of the aromatic base-pairs lie approximately normal to the axis. By contrast, the A-structure is more wide and squat, with grooves of unequal depth and a pronounced tilt of the base-pairs relative to the helix axis. In general the B- structure is favoured at high water activity, and is therefore the more important structure in solution, and thus to biology.The most important interactions holding the structures together are probably ’ J. D. Watson and F. H. C. Crick, Nnturr, 1953, 171,737.- 738. ’W. Fuller, M. F. H. Wilkins, H. R. Wilson, and L. D. Hamilton, J. Mol. Biol., 1965, 12, 60-80.’R. Langridge, D. A. Marvin, W. E. Seeds, H. R. Wilson, C. W. Hooper, M. F. H. Wilkins, and L. D. Hamilton, J. Mol. Biol.. 1965, 2, 38-64. Structurul Isomerixtion in DNA A B Z Figure 1 Helical riuriants of DNA structure. Two major families qf right-handed DNA are knoLt7n (A and B), Mhile Z-DNA is left-handed, Comparing the two right-handed conforma- tions. NY> .see [hut the A-fbrrn has a pronounced tilt, leading to a u>ider, squater structure compurcd to the B;fi)rm.The base-pairs of the ATJi,rm exhibit a pronounced displacement fiom the hclicul ii.vi.v, getwating a hole ch-n the centre of the structure, bvhile in B-DNA the asis pa.ss~~.sthrough the cvntre of the base-pairs. In general the B-structure is more hydrated, id lo~vritig \twter actiaitj. fuiwurs the A-structure. The structures are also .sequence-ikpc~nrknt.M’ifh GC-rich serpcnces tending to adopt the A-conformation. Z-DNA is N lefi-hrtiiiled hc1i.r bused on u CpG rlinuckeolide rc)pcating unit. It is ,formed by alteriiating purino- pj~rimiclincscJyutvitvs,piirticukarlj-at high salt concentrations. These pictures wwe adapted /rom Dickerson M’ith pcvmi.v.vion R.E.Dickerson, Sci. Am.. 1983., 249,94 I I1 Lilley those between the base-pairs, the stacking. Thus a model of the base-pairs alone may provide a good description of a given DNA ~tructure,~ in which the important parameters are base tilt (rotation about the short axis of the base- pair), roll (rotation about the long axis of the base-pair) and slide (translation along the long axis). In older literature the conformation of the deoxyribose phosphate backbone is emphasized, particularly that of the sugar pucker with the A-structure based on C3'-endo conformation and the B structure based on C2'- endo conformation, but this has been replaced with a view in which the energetic preferences of the base-pair interactions are accommodated by the backbone. The advent of single-crystal X-ray diffraction analysis of oligonucleotides of defined base-sequence at near atomic resolution has provided a rich new source of data on DNA conformation at the level of the single nucleotide, and revealed that DNA can indeed be structurally quite adventurous.Dickerson and colleagues solved the first structure of a piece of B-DNA,6 of self-complementary sequence CGCGAATTCGCG. One great lesson from this structure has been that at this level of resolution DNA exhibits considerable microheterogeneity in structure,' in which torsion angles may vary considerably from nucleotide to nucleotide. Calladine has had some success in accounting for these variations based on cross-strand steric clash between purines of successive base-pairs caused by the propeller twist of base-pairs (the deviation between the planes of the two bases comprising a given base-pair), and this is a demonstration of the importance of the base-pair interactions in determining the final structure.Dickerson has recently tabulated 35 separate crystal structure analyses which have been performed on synthetic oligonu~leotides,~ which shows the wealth of structural data now available. Unfortunately, relatively few of these are B-DNA, a consequence of the low water activity inherent in the crystallization process. One major surprise which emerged from the early crystallographic studies was that the handedness of DNA is not immutable. Rich, Wang, and colleagues solved the structure of CGCGCG and showed that the crystal contained a new left-handed form of DNA, which they called Z-DNA." This is based on a repeating dinucleotide unit in which the bases are paired with conventional Watson-Crick hydrogen bonding, but where the conformation about the guanine glycosyl bond is syn and the corresponding deoxyribose pucker is C3'-endo.Z- DNA formation is not restricted to alternating (CG), sequences, although these undergo the B-Z transition most readily. In general alternating purine-pyrimi- dine sequences are required, although some deviation from this can be tolerated. (TG),.(CA), forms Z-DNA quite readily," but (AT), has not been observed in 'C. R. Calladine and H. R. Drew, J. Mol. Biol.. 1984, 178,773---782.'R. Wing, H. R. Drew. T. Takano, C. Broka, S. Tanaka, L. Itakura, and R. E. Dickerson, Nature, 1980. 287,755 758. R. E. Dickerson and H. R. Drew, J. Mol. Biol.,1981,149,761-786. C. R. Calladine, J. Mol. Biol.,1982, 161, 343-- 352. R. E. Dickerson in 'Unusual DNA Structures'. ed. R. D. Wells and S. C. Harvey, Springer-Verlag, 1988, pp. 287 -306. lo A. H.-J. Wang, G. J. Quigley. F. J. Kolpak, J. L. Crawford, J. H. van Boom, G. van der Marel, and A. Rich. Nature, 1979. 282,680 -686. D. B. Haniford and D. E. Pullyblank, Nafure, 1983.302, 632 -634. Structural Isomerization in DNA the Z-conformation to date. Formation of Z-DNA may be facilitated by certain base modifications, such as 5-MeC and 8-BrG, and the nature of the counterion is critical.'2 Z-DNA exhibits chemical reactivities not found in B-DNA,13*14 and is highly immunogenic,' both poly- and mono-clonal antisera having been raised.The formation of Z-DNA may be regarded as a particularly dramatic example of sequence-dependent structural polymorphism, but it is by no means unique. The cruciform '6-'8 is perhaps an even more disruptive structural variant, which requires a sequence having twofold symmetry, i.e. an inverted repeat. Such a sequence may in principle form intra-strand base-paired helices, it. hairpin-loop structures, on each strand, to form the cruciform structure. These are discussed in much greater detail below, where they form the main theme of this review. Curved (sometimes called bent) DNA sequences have recently attracted a great deal of interest.Certain sequences, such as those found in the kinetoplast mini- circle DNA of trypanosome mitochondria, possess a strong intrinsic curvature," resulting in altered physical properties such as anomalously slow electrophoretic migration through polyacrylamide gels. This is called sequence-directed curvature of DNA, to distinguish it from the bending of DNA which results from the application of a force, as can occur on binding of a protein such as the CAMP- dependent activator protein 2o of Eschericia coli. Sequence-directed curvature is associated with short runs of oligo-dA.oligo-dT, repeating every 10 to 11 base-i.e. in phase with the helical repeat of B-DNA. It is known from recent crystallographic studies 23,24 that such sequences may adopt a novel geometry (sometimes called B'), characterized by large propeller twist of the A-T base-pairs, such that the thymine O(4) can form bifurcated hydrogen bonds to N(6) protons of the conventionally base-paired adenine, and the cross-strand adenine of the adjacent base-pair.This results in a narrowed minor groove. At the junctions between the A,, tract and the normal DNA there is a large roll angle, and this is a possible origin of the curvature. When added in phase with the DNA helix a large overall bend is the result, such as that found in Crithidia for example. In retrospect, we can now say that the structure adopted by the "J. H. van de Sande, L. P. McIntosh, and T. M. Jovin, EMBO J., 1982,1,777 -782.l3 K. Nejedly, M. Kwinkowski, G. Galazka, J. Klysik, and E. Palecek, J. Biomol. Struct. Dyn., 1985, 3, 467----478. l4 B. H. Johnston and A. Rich, Cell, 1985,42,713 -724. *'E. M. Lafer, A. Moller, A. Nordheim, B. D. Stollar, and A. Rich, Proc. Natl. Acad. Sci. USA, 1981,78, 3546 3550. l6 M. Gellert, K. Mizuuchi, M. H. ODea, H. Ohmori, and J. Tomizawa, Cold Spring Harbor Symp. Quant. Biol., 1979,43, 35 40. D. M. J. Lilley, Proc. Natl. Acad. Sci. USA, 1980,77,6468--6472. lRN. Panayotatos and R. D. Wells, Nature, 1981,289,466 470.~ l9 J. C. Marini, S. D. Levene, D. M. Crothers, and P. T. Englund, Proc. Natl. Acad. Sci. USA, 1982, 79, 7664 -7668. 'O H.-M. Wu and D. M. Crothers, Nature, 1984,308,509--513.'' S. Diekmann and J.C. Wang, J. Mol. Biol., 1985, 186, 1--11.''H. S. Koo, H.-M. Wu, and D. M. Crothers, Nature, 1986,320,501--506. 23 H. C. M. Helson, J. T. Finch, B. F. Luisi, and A. Klug, Nature, 1987,330,221-226.~ 24 M. Coll, C. A. Frederick, A. H.-J. Wang, A. Rich. Proc. Natl. Acad. Sci.USA, 1987.84, 8385-8389. Lilley Dickerson dodecamer CGCGAATTCGCG ’ is really a B’ conformation because of the central sequence. Other sequence-specific DNA polymorphs exist, though in general these are less well characterized than Z-DNA, cruciforms, or curved DNA. An example is the structure adopted by sequences in which there is an asymmetry of purines on one strand and pyrimidines on the ~ther.~’-~’ At low pH and moderate levels of supercoiling (see below), these form a novel structure characterized by sensitivity towards a number of enzymes and chemicals.Probing data are consistent with a novel structure comprizing a looping of the pyrimidine strand in triple-helix formation with half of the purine ~trand,~~,~~ although it remains possible that other structures may be adopted under some conditions. The C-G-C triads formed in the triplex structure require the participation of protonated cytosine in a Hoogsteen base-pair with guanine, accounting for the pH-dependence of the structure. Thus we can see that DNA structure is profoundly sequence-dependent, and that structural perturbation on a major scale is possible. 3 Helix Opening in Relaxed DNA For many of the biological functions of DNA, such as transcription (synthesis of RNA from the DNA template) and replication, some helix opening is a necessary prerequisite, separating the paired bases to allow for the genetic ‘reading’ required.How easy is this process in physical-chemical terms, and what can be done to facilitate the opening process? The most direct way to observe transient opening of base-pairs in the double helix is to look for a chemical reaction occurring at the atoms which remain completely protected when the base-pairing is undisrupted. The simplest reaction is the exchange of the hydrogen-bonded imino protons with solvent water, a reaction catalysed by proton acceptors such as Tris base, ammonia, or hydroxyl ions. Originally this was observed by studying the exchange of tritium from fully tritiated DNA fragment^.^' The reaction is visualized as occurring in two stages. Closed -!%+ Open “..Exchange3H ‘k,l ‘H The first stage is the opening of the base-pair, characterized by the opening and closing rate constants, k,, and kCl.Once the base-pair is open, proton exchange occurs with a rate constant kex.Two extreme situations may be considered, depending on the relative rates k,, and k,l.If k,, + k,l the measured exchange rate is the base-pair opening rate. Exchange will occur with each opening. This situation may be brought about by extrapolating to infinitely high concentrations of exchange catalysts. 25 J. M. Nickol and G. Felsenfeld, Cell, 1983,35467----477. 26 E. Schon, T. Evans. J. Welsh. and A.Efstratiadis, Cell, 1983,35, 837-848. ” D. E. Pullyblank, D. B. Haniford, and A. R. Morgan, Cell, 1985,42, 271 280. ”S. M. Mirkin, V. I. Lyamichev, K. N. Drushlyak, V. N. Dobrynin, S. A. Filippov, and M. D. Frank-Kamenetskii. Nature, 1987,330,495 497. 29 Y. Kohwi and T. Kohwi-Shigematsu, Proc. Natl. Acad. Sci. USA, 1988,85,3781 -3785.’’ J. J. Englander and P. H. von Hippel, J. Mol. Biol., 1972.63, 171 -177. Structural Isomerization in DNA If k,, 6 k,, the exchange is rate-limiting, and many openings are required to achieve exchange. In early work employing tritium exchange the former condition was assumed, leading to the estimation of rather high values for the opening probability (i.e. kop/kCl),and long lifetimes for both open and closed states.Recently, Gueron and colleagues have re-examined this process 31 using ‘H n.m.r. This technique has two major advantages over tritium exchange. First, imino proton resonances of specific bases may be assigned unambiguously, and thus one may study the exchange rate, derived from the resonance linewidth, of a particular base proton. Second, the method permits the measurement of faster exchange rates. Gueron et al. studied the exchange of the thymine imino proton in self-complementary oligonucleotides, notably in dCGCGATCGCG. They observed catalyst-induced changes in the rate of imino proton exchange, a feature not previously observed. This enabled them to obtain individual base- pair lifetimes, around 10 ms at 15 OC, and also the probability of opening, around These values differ by orders of magnitude from those reported earlier in similar compounds, although Frank-Kamenetskii had arrived at similar values using a reinterpretation of chemical reactivity data.32 Earlier investigators were misled by the existence of proton exchange even in the absence of added catalyst.Gueron et showed that the rate of proton exchange was unrelated to base- pair lifetime, and exchange occurred from the same open state as exchange by added catalyst. This was most probably the proton-acceptor nitrogen of the complementary base. We may conclude that in linear DNA the base-pairs open singly and transiently, and the bulk properties of the DNA will be well represented by a fully base-paired molecule.However, as we shall see, the situation is quite different for a supercoiled DNA molecule. 4 DNA Supercoiling Most natural DNA molecules are constrained into circles or loops, whereupon the topology adds a new dimension to the structural properties of the molecules.33 While this is true for complex loops of chromosomes in bacteria34 or higher cells,35 the properties are most easily studied in small circular DNA molecules. Perhaps the simplest such circles for study are the bacterial plasmids, which are small (typically 2 to 6 kbp*), easily manipulated by established techniques of modern molecular genetics, and isolated from the cell as supercoiled species. DNA supercoiling is based upon the fundamental topological property of 31 M.Gueron, M. Kochoyan, and J.-L. Leroy, Nuture, 1987.328,89 92. 32 M. D. Frank -Kamenetskii in ‘Siruciure mrl Motion: Memhrones. Nucleic Acid.7 uncl Proteins’, ed. E. Clementi. G. Corongiu, M. H. Sarma, and R. H. Sarma, Adenine Press, 1985, pp. 417 432. 33 J. Vinograd. J. Lebowitz. R. Radloff, R. Watson, and P. Laipis, Proc. Nail. Acad Sci. USA, 1965, 53, 1104 1111. 33 A. Worcel and E. Burgi, J. Mol. Biol., 1972. 71. 127 147. 35 J. R. Paulson and U. Laemmli. Cell, 1977.12,817 828. * DNA molecules are conveniently measured in terms of the number of base-pairs which they comprise. 1 kbp = 1000 base-pairs. linear 21 0 basepair DNA molecule,20 helical turns J Negative Relaxed Positive supercoiling supercoiling Lk 19 Lk 20 Lk 21 ALk -1 ALk 0 ALk +1 CJ -0.05 00 CJ +0.05 Figure 2 Topoisomeric DNA circles and the linking number.We take a lineur DNA molecule comprising e.uactlj. 20 turns qf double helix, and ligate it into a circle. If this is carried out in the absence of any torsional.force, the resulting circle will also contain 20 turns. and ki’ill be torsionally relaxed (central circle). The tM’o strands of the DNA will mutually interlink twenty times, thereby dejning a linking number of 20. Isomers qf the relased species will be generated if torsion is applied before ligation. Removal ?f turns will generate topoisomers of reduced linking number (negative supercoiling, left), while addition of turns vyill increase the linking number @ositive supercoiling, right).If this experiment cr1a.Y carried out in practice, a distribution of topoisomers Liwld be obserued. corresponding to thermul population of I( Boltzmann distribution linkage, a property that is exhibited by two interwound circles. Two circles can be linked n times, where n is an integer which measures the minimum number of times one circle pierces the plane of the other. The linking number (written as Lk, L. or X, depending upon the author) is invariant so long as the integrity of both circles is maintained. Circular DNA exhibits linkage because the two strands turn around each other once in every ten or so base-pairs. Thus in a covalently closed circular DNA molecule the linking number is approximately N/10.5, where N is the number of base-pairs. The phenomenon of supercoiling in DNA arises from differences in the linking number, illustrated by the imaginary experiment depicted in Figure 2.In this experiment we have taken a linear DNA molecule comprizing exactly twenty turns, and used an enzyme to join it into a circle. If no torsional force is applied the resulting circular species also contains 20 turns, and hence the linking number is 20. Since this molecule is directly derived from the torsionally relaxed linear species, we refer to this as the relaxed circle, and the linking number is denoted by Lk’. If we perform this experiment Structural Isomerization in DNA again, but apply torsion to the molecule prior to ligation,* we may change Lk. Thus we may add or substract turns of DNA, resulting in a circular molecule which is over- or under-wound. We define a linking difference ALk = Lk -Lk' which is just the number of turns removed or added in the above experiment.We may also define a further parameter o,the specific linking difference or superhelix density, which is independent of size CJ = ALklLkO Thus in the above experiment, if one turn was removed before ligation, Lk = 19, ALk = -1 and o = -0.05. Most natural DNA molecules are underwound, or negatively supercoiled, in this way, with typical superhelix densities of o = -0.06. It should be noted that although Lk is necessarily integer, this restriction does not apply to Lko and hence to ALk, because most DNA molecules will not contain an exact number of helical turns under a given set of conditions (helix pitch is a function of environmental conditions such as temperature and ionic strength).It will be noted that the circular forms derived in the circularization experiment are isomers which result from topological differences, and are therefore called topoisomers. These will differ in their free energy, to an extent which arises from the torsional force required to underwind the DNA helix. This is a quadratic function 36-38 1050RTbLkZAGs = N where AGs is the free energy of supercoiling, R the gas constant, T the absolute temperature and N the size of the molecule in base-pairs. These energies are significant. The plasmid pBR322, which is commonly used for cloning DNA sequences in bacteria, has a free energy of supercoiling of about 100 kcal mol-' at o = -0.05, an energy greater than that of the C-C covalent bond.Living cells have evolved an array of enzymes to manipulate the topology of DNA molecules. Enzymes which can change the linking number of a topoisomer are called topoisomerases, and these are present in all cells.39 In bacteria two main topoisomerases are known. Topoisomerase I 40 has a relaxing activity, "R. E. Depew and J. C. Wang, Proc. Null. Acuil. Sci. USA, 1975,72,42754279.''D. E. Pullyblank. M. Shure, D. Tang, J. Vinograd, and H.-P. Vosberg, Proc. Narl. Acud Sci. USA, 1975,72,4280 4284.''D. S. Horowitz and J. C. Wang. J. Mol. Bid., 1984, 173, 75- 91. 39 M. Gellert. Ann. Rec. Biochmi., 1981,50, 879 910.40 J. C. Wang, J. Mol. Bid, 1971.55, 523 533. * Ligurion refers to the covalent joining of two ends of DNA molecules, i.p. the formation of two new covalent phosphodiester linkages, using an enzyme (called ligase) and ATP. The joined molecules might be completely base-paired ('blunt' ended) or have overhanging cohesive ends, e.g. XXXG YYYCTTAAS'. Lilley which increases the linking number of a negatively supercoiled substrate until it is relaxed. It is the product of the topA gene,41 and works by acting first as a nuclease which is transiently covalently linked to the DNA as a phosphate ester while the DNA swivels,42 followed by a re-ligation process. The process is passive in the sense that no energy is required. The other main topoisomerase is DNA gyra~e,~~a tetrameric protein of two different subunits encoded by the unlinked genes gyrA and gyrB.44-46 It introduces supercoiling using the coupled hydrolysis of ATP as an energy source.The cellular activities of the two enzymes is regulated to maintain the level of supercoiling in uiuo at a fixed level. 5 DNA Supercoiling and Structure DNA supercoiling modifies both DNA structure and dynamics. The change in structure can be very simply demonstrated experimentally by electrophoresis of a mixture of topoisomers of plasmid in an agarose gel, shown in Figure 3. Each topoisomer migrates as a well-resolved species, because it has a shape which is different from topoisomers of different linking number. Changes in linking number are coupled to alterations in helical twist (the path of the strands about the helical axis, Tw)and deformation of the helix path in space from a plane (writhing, Wv),according to47 Lk = Tw -k Wr or ALk = ATw -F Wr where ATw is the deviation in twist from the equilibrium value.Thus supercoiling results in changes in helix twist, or axial writhing, or both. The exact partition between twisting and writhing in DNA circles of several thousand base-pairs, and the shape of the writhing molecules, are the subject of some debate, and will depend on a number of environmental factors. Both toroidal and interwound forms are possible, and while the former has been advocated on the basis of solution X-ray scattering measurement^,^^ electron microscopic images of supercoiled DNA are normally interwound, and these are generally considered to be the form adopted in solution.In addition to influencing the overall tertiary structure of the molecules, supercoiling may determine local structure, arising from a coupling between local 4’ R. Sternglanz, S. DiNardo, K. A. Voelkel, Y. Nichimura, Y. Hirota, V. Becherer, L. Zumstein, and J. C. Wang, Proc. Nail. Acad. Sci. USA, 1981,78,2747-2751. 42 R. E. Depew, L. F. Liu, and J. C. Wang, J. Biol. Chem., 1978,253,511-518. 43 M. Gellert, K. Mizuuchi, M. H. O’Dea, and H. A. Nash, Proc. Natl. Acad. Sri. USA, 1976,73, 3872 3876. 44 M. J. Ryan, Biocliemisiry, 1976, 15, 3769 -3777. ”W. A. Goss, W. H. Deitz, and T. M. Cook, J. Bacteriol., 1965,89, 1068--1074.46 W. L. Staudenbauer, Eur. J. Biochem., 1976,62,491 497. 47 F. B. Fuller, Proc. Nutl. Acucl. Sci. USA, 1971,68,815--819. 48 G. W. Brady, D. B. Fein, H. Lambertson, V. Grassian, D. Foos, and C. J. Benham, Proc. Natf. Acad. Sci. USA, 1983,80,741 -744. Structural Isomerization in DNA sc R s2 b Nb SIb Figure 3 Topoisomers qf a 3 657 bp circular DNA species. This small plusmid M'USpartiullj?relaxed with a topoisomerase which interconverts topoisomeric species, changing the linkingnumber, und the product was examined by electrophoresis in an ugarose gel. On the left (heSC) is shown the unrelaxed negatively supercoiled DNA, obtuined by extraction directly ,from bacterial cells. This DNA is supercoiled to about CY = -0.06, but the distrihurion of' topoisomers is unresolved under these conditions, S1; supercoiled monomeric plasmid DNA.N; monomer which has undergone strand breakage. S2; supercoiled dimer. After incubation with the topoisomerase a ladder of species is generated (lane R. right) in which euch band is (I single topoisomer dtfyering in unit linking number from its direct neighbours. The topologicul d$ferences between topoisomeric species result in alteration to the overall shapc. of' the molecules, and hence the rates of migration in the gel geometric factors and the global topology of the molecules. Take the example of cruciform formation, shown diagrammatically in Figure 4. An inverted repeat of n base-pairs contributes a Tw of n/10.5 to the total twist of the molecule in its unperturbed state, but after adopting cruciform geometry this drops to approxi- mately zero.Thus there is a negative ATw of n/10.5 turns due to cruciform formation, and the molecule may relax by this amount. The unwinding which accompanies Z-DNA formation is even greater, almost by a factor of two. To return to our earlier example of pBR322, formation of a 30 bp cruciform will relax approximately 3 turns, corresponding to a lowering of free energy by 21 kcal mol-' at o = -0.06. Therefore even though the formation of the cruciform may be quite unfavourable in linear DNA, provided the free energy of formation Lilley A AA T B AG CG CG TA AT AT CG GC AT TA CG CG TA GC AAA CAA llT GTT CG AT GC GC riAT TA CG GC TA TA AT GC GC TC TA T Figure 4 Cruciform estrusion by inverted repeats in supercoiled DNA. (A) The sequence of an inverted repeat found in the E.coli plasmid ColEl drawn in its crucifOrm conformation. Note that as a result of the tw!ofold symmetry of the sequence, the individual strands can base- pair within the same strand, generating stem-loop, or hairpin, structures. (B) Negativesupercoiling stabilizes cruciform structures. This scheme illustrates the relaxation of superhelical stress which may be obtained on cruciform .formation (or the formation of anystructure leading to a local reduction in tMiist-Z-DNA Mwuld he another example). If the change in thefree energy of supercoiling resulting.from this relaxation is greater than that of ,fi,rmation of the cruciform, a stable cruciform results is less than 21 kcal mol-’ this will be compensated by the gain in that of supercoiling, and the cruciform will be stable. In fact, the free energy of cruciform formation has been measured for a number of sequence^^'-^^ and lies in the range of 12 to 18 kcal mol-’.The same kind of analysis applies to any local structural variation which is accompanied by a negative twist change, including the formation of Z-DNA and melted ‘bubbles’. This shows why the opening of the DNA helix by RNA polymerase is facilitated by negative supercoiling in general, where about 12 bp are required to become unpaired on formation of the open Multi-state conformational equilibria are possible in supercoiled DNA mol- ecules.They can arise in two different ways. First, a circular molecule may have more than one sequence which can undergo a conformational transition. Since a ‘‘ M. Gellert. M. H. O’Dea. and K. Mizuuchi, Proc. Nd. Acad. Sci. USA, 1983,80,5545 --5549. SO A.J. Courey and J. C. Wang, Cell, 1983.33,817 -829. ” D. M. J. Lilley and L. R. Hallam, J. Mol. Biol., 1984, 180. 179- 200. s2 D. R. Greaves. R. K. Patient, and D. M. J. Lilley, J. Mol. Biol., 1985, 185,461- -478. ’3 L. H. Naylor, D. M. J. Lilley, and J. H. van de Sande. EMBO J., 1986,52407-2413. ”J. Siebenlist, Nature. 1979, 276,651- -652. T sequence will reduce the overall level of supercoiling in the molecule, this will change the probability of the second transition occurring.Such cases have been observed experimentally for Z-forming sequences, and have been treated within the framework of statistical thermodynamic^.^^ Second, a given sequence may itself be capable of forming more than one alternative structural form. The sequence (GT),(AC), for example might exist as B; Z; or cruciform-DNA, depending on the length, level of supercoiling etc., and be described by a multi-dimensional phase diagram. We have observed similar two- state conformational equilibria for (AT),,tracts in supercoiled molec~les,~~ which may exist either as cruciform structures or an unwound and chemically reactive uniform helical variant of B-DNA. This equilibrium is a sensitive function of salt concentration and temperature.Not only does negative supercoiling increase the structural repertoire of DNA, it also increases its dynamic character, resulting in enhanced chemical reactivity for example. As we shall see, cruciform structures may be extruded in supercoiled DNA by at least two kinetic pathways. Whatever the details of the mechanism however, it is clear that many base-pairs must be broken during this process and that therefore helix openings must be possible in supercoiled DNA on a scale much greater than that found in linear molecules. To this extent therefore, supercoiling is a kind of 'activation' of DNA, generating new structural and dynamic aspects to the character of the double helix. 6 The Cruciform Structure Cruciform structures were first recognized experimentally at the beginning of this decade,I6-' twenty-five years after their first theoretical des~ription.~'*~~ The crucial role of supercoiling had been overlooked in earlier studies, and this was the first example of the now-familiar stabilization of structural variants by negative supercoiling.Cruciform formation has subsequently been demonstrated for many inverted repeat sequences in a variety of supercoiled plasmids and phage. Conventional physical methods (spectroscopy, diffraction, etc.) cannot in general be applied to investigation of supercoil-stabilized features because the feature of interest, e.g. the cruciform, is normally a very small part of the entire circular plasmid molecule.We have therefore developed new methods with which to investigate these structures. We introduced the procedure of nuclease cleavage on supercoiled DNA This probing method takes advantage of some feature of the novel structure which can be differentially recognized by an enzyme or chemical. In the case of the cruciform there are two such features, the single-stranded loops and the four-way junction. Most probes select the former as their target, being either single-strand specific enzymes (such as S 1 nuclease), or single-strand selective chemicals (such as bromoacetaldehyde). Some probes of the junction 55 R. J. Kelleher, M. J. Ellison, P. S. Ho, and A. Rich, Proc. Nut/. Acud. Sci. USA, 1986,83,6342-6346. 56 J. A. McClellan and D.M. J. Lilley,J. Mol. Bid, 1987, 197, 707--722. "J. R. Platt, Proc. Natl. Acad. Sci. USA, 1955,41, 181 183. 58 A. Gierer, Nuture. 1966,212. 1480 1481. Table 1 En:j*me and chemical probes bt,hich haw been employed in the study of cruciform structures Probe Target Result Reference S1 nuclease cleavage 17,18 Micrococcal nuclease cleavage 59,60 Ba131 nuclease cleavage 51 P1 nuclease cleavage 61 Mung bean nuclease cleavage 62 Bromoacetaldeh yde etheno-adduct 63 Osmium tetroxide cis-diester 64 Glyoxal etheno-adduct 65,66 Bisulphite deamination to dU 67 Diethyl pyrocarbonate carbethoxylation 68,69 Resolvase cleavage 70-74 e.g. T4 endonuclease VII have impressive structural selectivity, notably the Holliday junction * resolving enzymes.A tabulation of probes for cruciform structures is presented in Table 1. The structure of the cruciform is defined by the two features indicated above- the loops and the four-way junction. Chemical probing experiments 67 have indicated that the optimal loop size lies between four and six nucleotides. Detailed analysis of diethyl pyrocarbonate modification patterns in the ColEl cruciform loop have suggested to US" that the loop possesses a well defined structure, probably involving base stacking. Studies of isolated DNA hairpins by Hilbers and co-workers75 have indicated that bases may stack, as if to continue the helical structure, on the 5'-side of the loop. The four-way junction is formally equivalent to the Holliday ''junction, and "C.Dingwall, G. P. Lomonossoff, and R. A. Laskey, Nucleic Acids Res., 1981.9,2659 2673. "D. M. J. Lilley. ColrlSpring Hirrbor Synip. Quiint. Bid. 1982,47. 101 112. 'l D. B. Haniford and D. E. Pulleyblank, Nucleic Acids Rex, 1985, 13,4343-4363. "L. C. Sheflin and D. Kowalski, Nuckic Acids Res.. 1984, 12. 7087 7104.~ '3 D. M. J. Lilley, Nucleic Acids Res., 1983. 11, 3097---31 12. '4 D. M. J. Lilley and E. Palecek, EMBO J., 1984.3, 1187 1192. 6s G. W. Cough, Pl1.D. Thesis. University of Dundee, 1986. 'b D. M. J. Lilley in 'The Role of Cjdic Nucleic, Acid Aclrlucts in Carcinogenesis and Mutagenesis'. ed. B. Singer and H. Bartsch, IARC Publication No. 70, Lyon, 1986. "C. W. Cough. K. M. Sullivan, and D.M. J. Lilley. EMBO J.. 1986.5. 191 196. "J. C. Furlong and D. M. J. Lilley. Nuc,leic Acids Rrs., 1986, 14, 3995-4007. "P. M. Scholten and A. Nordheim, Nucleic Ac,iil.s Ra., 1986, 14, 3981 3993. "' K. Mizuuchi, B. Kemper. J. Hays. and R. A. Weisberg, Cell, 1982,29. 357 365. " D. M. J. Lilley and B. Kemper. Cell, 1984.36.413 422. iz B. deMassey, F. W. Studier, L. Dorgai, F. Appelbaum, and R. A. Weisberg, Cold Spring Harbor Symp. QI~LI~I.Biol., 1984. 49, 7 15 726.''L. Symington and R.Kolodner. Proc. Ncitl. Acacl. Sci. USA, 1985,82,7247 7251. l4 S. C. West and A. Korner, Proc. Nail. Acnrf. Sci. USA, 1985, 82. 6445- -6449. "C. A. G. Haasnoot. C. W. Hilbers, G. A. van der Marel. J. H. van Boom, U. C. Singh, N. Pattabiraman, and P. A. Kollman, J.Bioniol. Strucr. Djm, 1986. 3, 843 857.~''R. Holliday, Genet. Rex, 1964. 5, 282-304. * Hollirlaj, junctions. Recombining DNA molecules initiate the process by strand exchange, leading to a four-way helical junction as first proposed by Holliday in 1964. The interested reader must consult any good molecular biology text under the subject of homologous genetic recombination. its study is important to a full understanding of homologous genetic recombina- tion, where great progress has been made very recently. Chemical probing with bisulphite and diethyl pyrocarbonate indicates that under normal circumstances the junction is probably fully ba~e-paired,'~.'~and this is supported by n.m.r. studies of an isolated junction.77 lnvestigation of a pseudo-cruciform construct '' revealed a strongly position-sensitive retardation of gel migration akin to that seen for kinetoplast DNA,20 and led us to propose that the cruciform junction introduces a pronounced bend into the molecule.In the presence of magnesium ions. the junction adopts a more stable and compact configuration. Using an extension of this method Cooper and Hagerman 79 studied the relative disposition of the helices, concluding that the junction is asymmetric, and not, for example, tetrahedral. We have extended these studies" using a related method, and shown that the junction has the shape of an X, formed by co-linear stacking of arms in pairs. Two isomeric structures are possible, and the choice of helical partners is determined by the local sequence in the junction itself.The resolution of a given junction by enzymes such as T4 endonuclease VII depends on relative isomer stability, and thus the products of recombination is likely to depend on local sequence at the point of resolution. Ion binding is also of fundamental importance. In the absence of ions, the structure of the junction is quite different-the helices are no longer stacked in pairs, but are instead fully extended in a square arrangement, resulting in thymines at the junction becoming reactive to osmium tetroxide. 7 Cruciform Extrusion Processes-Two Mechanisms In general the process by which cruciform structures are extruded in supercoiled DNA is far from facile. The first kinetic studies of cruciform extrusion" revealed significant kinetic barriers, as might have been predicted in view of the substantial reorganization of DNA structure which must occur.It was found that the extrusion reaction could be very slow, even at elevated temperature^.^^,^^ On the other hand, work on different inverted repeats showed that extrusion could occur rather more We made a comparative studys4 of cruciform extrusion kinetics in two .plasmids, pColIR315 and pIRbke8. pColIR315 contains a 440 bp insert from natural E. coli plasmid ColEl,*' which includes a 31 bp ColEl inverted repeat and about 100 bp and 300 bp of left and right flanking DNA respectively. 11 D. E. Wemmer. A. J. Wand. N. C. Seeman. and N. R. Kallenbach, Biochemis/r~;1985, 24.5745 5749. 78 G. W. Gough and D. M. J. Lilley. Nature. 1985, 313. 154 156. 19 J. P. Cooper and P. J. Hagerman. J. Mol. Biol., 1987, 198, 71 I 719. 80 D. R. Duckett, A. 1. H. Murchie. S. Diekmann, E. von Kitzing, B. Kemper and D. M. J. Lilley. Cell, 1988,55,79-89. HI K. Mizuuchi, M. Mizuuchi. and M. Gellert. J. Mol. Biol., 1982, 156, 229--243. HZ I. Panyutin, V. Klishko. and V. Lyamichev, J. Biomol. Sfruc.1.Dyn., 1984, 1. 1311 1324. H3 R. R. Sinden and D. E. Pettijohn, J. Biol. Cizetiz.. 1984, 259. 6593 6600. 84 D. M. J. Lilley. Nucleic Acids Res., 1985, 13. 1443 1465. 8s M. Bazaral and D. Helinski, J. Mol. Biol., 1968.36. 185 194. 66 Table 2 Kinetic characteristics of C-type and S-type crucijorms C-type S-type Example ColE1 pl Rbke8 Occurrence rare common [NaCI] optimum 0 mM 50 mM Ek3 I80 kcal mol-' 40 kcal mol-' AHf 180 kcal mol-' 40 kcal mol-' ASs (at 37 "C) 400cal deg-' mol-' 60 cal deg-' mol-' pIRbke8 was constructed 86 by cloning synthetic oligonucleotides into the BuinH1 restriction site of pAT153 to generate a perfect 32 bp inverted repeat (termed bke), flanked by vector sequences.Figure 5 summarizes kinetic data obtained for the two plasmids. The properties are strikingly different, in two main respects: NuCl Dependence. The ColEl cruciform of pColIR3 15 extrudes maximally in the absence of added salt, and the extrusion is strongly suppressed with the addition of NaCl. By contrast, the bke cruciform of pIRbke8 fails totally to extrude in the absence of added salt, and exhibits a maximal extrusion rate at 50-60 mM NaCI.Temperature Dependence. pColIR3 15 exhibits a marked temperature dependence for cruciform extrusion. The rate constant for the extrusion reaction increased by a factor greater than 2000 in an 8 degree temperature interval. The reaction is thus characterized by an enormous Arrhenius activation energy (E,) in the region of 200 kcal mol-'. The temperature dependence for the extrusion of pIRbke8 is rather lower, with an E, of about 40 kcal mol-'. However, despite the much larger activation barrier for pColIR3 15, extrusion proceeds at lower temperatures, indicating that the extrusion of the ColEl cruciform must have a large entropy of activation. These kinetic properties are highly contrasting, and constitute members of two classes of cruciform, termed C-type (for ColE 1) and S-type (for Salt-dependent).The only natural member of the C-type class is ColE1, and its subclones such as MostpColIR315 84 and the deletant ~A03.~~ sequences behave as S-type cruciforms, including PUC~,~~ PAC 102,50 pOCE 12,83 with lower temperature dependences, a requirement for salt and extrusion occurring at higher tempera- tures. The kinetic properties of the two classes are summarized in Table 2. The strongly contrasting kinetic properties of C-and S-type cruciforms suggests the existence of alternative mechanistic pathways for the extrusion process. In fact, from an entirely theoretical standpoint,86 we had previously proposed two possible pathways, which are presented schematically in Figure 6.In the upper mechanism, a large region of DNA is unpaired in the probable "D. M. J. Lilley and A. F. Markham. EMBO J.. 1983,2,527 533 StructuraI Isomerization in DNA C-type S-type -1 -1 Ea 180 kcal mol Ea 40 kcal rnol 3-2-1-IIII Rate Temperature Temperature 3--2-0 mM -50 rnM 1--/ J 28 32 36 40 28 32 36 40°C 6 Arrhenius Plot Ink 101-3.15 3.20 3.25 3.30 3.35 T("K)/lOOO Figure 5 The kinetic properties of cruciform extrusion in supercoiled DNA-a summary of two kinetic classes. Salt (upper) and temperature (lower) dependences .for extrusion of typical C-type (left) and S-type (right) plasmids. In these experiments the extrusion process is allowed to proceed,for a short time, after which the relative degree of extrusion is measured, which is proportional to the rate qf extrusion.C-type extrusion is characterized by cruciform ,formation at low salt concentrations, with a very high temperature dependence, while S-type estrusion requires the inclusion of salt and proceeds with a low'er temperature dependence. In other experiments the extent of extrusion is measured as a function of time in order to calculate rate constants for the process. Rate constant.yJor C- and S-type extrusion have been measured as a ,function of temperature, and activation energies calculated from Arrhenius plots like that shown at the bottom. T pica1 Arrhenius activation energies are 200 kcal mol-'.for C-tjye extrusion and40 kcal mol- II?.for S-type extrusion transition state, which then undergoes intra-strand base-pairing to form the fully extruded cruciform.In the lower path, the extent of initial disruption is smaller. A relatively short proto-cruciform is formed as an intermediate, followed by branch Lilley Avcooperativeopening / large bubble C-type extrusion inverted repeat fully-extruded cruciform centralopening S-type extrusion branch\ /migrationinfm-strand pairing c1 central bubble proto-cruciform Figure 6 Two mechanisms for cruciform extrusion. The inverted repeat, represented by the thicker line, is shown in the une-ytruded form on the left. We believe that C-type cruciforms (top) initiute the estrusion process with a coordinate opening of many base-pairs to form a large bubble.An intra-strand reassociation then forms the mature cruciform structure. The extrusion of S-type cruciforms (lower), is initiated by a smaller opening event. Intra-strand pairing generates a smaller proto-cruciform, which may undergo branch migration. Base- puiring is transferred from unextruded sequence to the growing cruciform stems in a multi- step process, to ,form the fully extruded structure. The principal differences between the two mechanisms lie in the size of the initial opening and the degree of tertiary folding in the transition state migration (sequential transfer of base-pairing to the growing cruciform stem) to the completely extruded cruciform.The transition state is harder to identify in this pathway, but is likely to be intermediate between the initial unpairing and the proto-cruciform. We have proposed 84 that the C-type cruciforms extrude via the top pathway, while the S-type sequences proceed oia the lower mechanism. The major differences between the two mechanisms postulated may be sum-marized: 1. There is much more disruption of base-pairing in the upper pathway. 2. The transition state in the lower mechanism resembles the forming four-way junction, rather than a simple melted ‘bubble’. The upper pathway would be expected to be characterized by large values of enthalpy and entropy of activation, and to be facilitated by low ionic strength (phosphate repulsion reduces the melting temperature of DNA).The reduced disruption in the lower pathway will lower the enthalpy and entropy of activation, but phosphate-phosphate repulsion is likely to be significant in a more structured transition state. The energy of the activated complex will therefore be reduced by cation binding, in a manner analogous to the Structural Isomerization in DNA requirement for cation binding in forming the compact conformation of the complete cruciform. Experimentally, we have found that C- and S-type cruciforms have exactly these kinetic properties. We have examined the S-type mechanism in more detail. We took pIRxke/vec, a typical S-type molecule closely similar to pIRbke8, and studied the extrusion as a function of the nature of the cation present.87 We observed that ions vary greatly in the efficiency with which they promote the extrusion.The effects of cations may be divided into four classes: 1. Extrusion rate maxima 50-60 mM, forming a distinct optimum. The monovalent ions, mainly Group la metals and (CH3)4N+. 2. Extrusion rates levelling off by 100-200 pM. The divalent ions of Group IIa, together with Mn2+ and (poorly) Co2+. 3. Peak of maximum rate at 15-50 pM. [Co(NH3)6] (111) and polyamines. 4.Ineffective at all concentrations. Most transition metals are completely ineffective in promoting extrusion. Despite the different efficiency with which the different ions facilitate cruciform extrusion, the reaction mechanism is unchanged.Thus, Arrhenius activation energies for the reaction in the presence of Na' and Mg2+ are the same within experimental error. Our conviction is that the observed kinetic consequences of using various cations results from differential stabilization of the transition state, from which we may deduce some aspects of the structure of the activated com- plex. 1. Groups 1 to 3 show that the optimal ion concentration is reduced as charge increases, reflecting ionic binding. 2. Polyamines are very effective, while basic amino-acids are totally ineffective. The distribution of positive and negative charges on these ions must affect binding. 3. Many transition metals are totally ineffective in promoting extrusion. Most of the ions in this group are soft, binding preferentially to nitrogen or exocyclic keto-substituents, i.~.they will be better at binding bases than phosphate groups.In fact there is a good correlation between the ability of ions to promote S-type cruciform extrusion, and their position in the Irving-Williams series.88 Clearly phosphate binding is required. 4. Within the la and IIa metals, there is an excellent correlation between the rate of cruciform extrusion and ionic radius, shown graphically in Figure 7. Ion binding to normal DNA, which is effectively a cylindrical polyelectrolyte, is well treated using electrostatic consideration Counterions reduce phosphate charge by acting as a screening cloud, and indeed it has not proved 87 K. M. Sullivan and D.M. J. Lilley, J. Mol. Biol.,1987. 193. 397 404. 88 H. Irving and R. J. P. Williams, Noiurc. 1948. 163, 746 747.''G. S. Manning. Q. Rm. Bioph>.s.. 1970, 11, 179 246. ')" B. H. Zimm and M. Le Bret, J. Biomol. Siruci. 41~7.. 1983. 1. 461 471. Lilley 4k x 10 0.6 0.8 1.0 1.2 1.4 1.6 1.8 ionic radius (A) Figure 7 The rate of S-type cruciform extrusion depends upon the ionic radius of the cation present. Rate constants for cruciform extrusion at 35 "C were measured in the presence of 60 mM group Ia metal chlorides, and these are plotted against the ionic radius. possible to localize ions in crystal structures of double-stranded DNA. In contrast, tRNA has a number of high affinity binding sites for Mg2+,91-93 generated by the tertiary folding of the molecule, and we now believe very similar ion binding is very important in the formation of the four-way junction.80 The dependence of extrusion rate on ionic size therefore implies selective ion binding,94 suggesting the formation of a structure which contains electronegative clefts capable of such ion binding.This is further evidence for the transition state having significant four-way junction character, as proposed earlier. In a different study of the S-type extrusion reaction, we have constructed a series of variants of the typical S-type molecule pIRbke8, in which one or two mutations (mainly A.T to G.C or vice uersa), have been introduced into the symmetrical unit of the bke inverted repeat.95 The results are summarized graphically in Figure 8.The mutations exhibiting the most significant alterations to extrusion rates were those in which the base changes are close to the sequence dyad. Thus the half-time for the process at 37°C can change from 70 min in pIRbke8 (central sequence AGAATTCT) to less than 1 min for ATATATAT or more than 29 h for AGCCGGCT. With a few special exceptions, sequence changes further from the centre of the inverted repeat only alter the rates by a factor of two or less. This suggests that only the central region of the inverted repeat becomes altered in the formation of the transition state, confirmed by a '' A. Jack, J. E. Ladner, D. Rhodes, R. S. Brown, and A. Klug, J. Mol. Bid., 1977, 111,315-328. 92 S. R. Holbrook, J. L. Sussman, R.Wade Warrant, G. M. Church, and S.-H. Kim, Nucleic Acid Res.. 1978,8,281 I 2820.~ 93 G. J. Quigley, M. M. Teeter, and A. Rich, Proc. Natl. Acad. Scr. USA, 1978,75, 64 68. 94 S.-K.Tam, and R. J. P. Williams, Srrucr. Bonding, 1985,63, 103-151. 95 A. I. H. Murchie and D. M. J. Lilley, Nucleic Acids Res., 1987, 15, 9641- 9654. Structural Isomerizationin DNA FASTER TGTGGATCCGGTACCAGAATCTGGTACCGGATCCTCT ACACCTAGGCCATGGTCTAAGTCCATGGCCTAGGAGA Figure 8 Rates of S-type cruciform extrusion depend on the sequence at the centre of the inverted repeat. Rate constants for cruciform extrusion were measured at 37 "Cfor a series of closely related mufant sequences related by one or two bases in the symmetric unit. In this representation the bars indicate the ratio of the extrusion rate of the mutant sequence to that of the original sequence.Those above the sequence are rate enhancements, while those below indicate slower rates. In the sequence bold typeface denotes bases comprizing the parent inverted repeat sequence. Mutant sequences are named to indicate the altered position (numbered from the 5' end of the inverted repeat) and the new base at this location. Corresponding mutations are always present that preserve the twofold symmetr.v similar Very recently Courey and Wang 97 have studied S-type cruciform extrusion as a function of supercoiling, from which they were able to deduce that the initial opening corresponds to about 10 bp. These results are fully consistent with the S-type mechanism shown in Figure 6.Even more subtle changes may be made at the centres of these inverted repeats, using base methylati~n.~~ For example, the bke sequence is symmetrically disposed about the sequence GAATTC, which is a substrate for the EcoRI methylase. Incubation with this enzyme and S-adenosyl methionine results in methylation at the N(6) of the second adenine of the sequence on both strands, and we find that this small change alone results in a threefold rate enhancement for S-type cruciform extrusion. A related inverted repeat containing a central AGCT sequence could be modified using AluI methylase to introduce 5-Me cytosine on both strands, and this resulted in a threefold reduction in cruciform extrusion rate. These results are fully explicable in terms of the proposed model for S-type extrusion, 9h G.Zhang and P. R. Sinden, J. Biol. Chem., 1989,205, 593 -602. 97 A. J. Courey and J. C. Wang, J. Mol. Biol.. 1988,202,35 43. 98 A. I. H. Murchie and D. M. J. Lilley, J. Mol. Biol., 1989, 205, 593 602 Lilley since methylation of A, which interferes with A.T base-pairing in the central region, leads to facilitation of the initial opening, while methylation of C, which stabilizes the helix by improved stacking, makes this harder. No data which seriously challenge the mechanistic models described for cruciform extrusion have emerged, and we are confident that these are a good description of the physical processes involved. 8 Contextual Influence on the Kinetic Character of Cruciform Extrusion We have described two contrasting mechanisms by which cruciform extrusion may occur, and yet we have given no indication as to what determines which pathway any particular sequence takes.Comparison of the base sequences of the ColEl and bke cruciforms provides no clues to the origins of the differences- both are 5&60% (A + T), and neither has additional sequence motifs such as purine-pyrimidine alternation or polypurine tracts. We should recall that the unusual behaviour is that of ColE1; most cruciforms are S-type. In fact, there is one respect in which pColIR315 is quite abnormal. The ColEl sequences which flank the inverted repeat on both sides are over 80% (A + T). We were forced to consider that this very (A + T)-rich base composition might be responsible for the C-type kinetic behaviour of the ColEl cruciform.We therefore examined the possibility that the (A + T)-rich ColEl sequences somehow influence the cruciform extrusion pathway, and confer C-type kinetics on the adjoining inverted repeat .' The following series of experiments were performed: 1. Using standard methods of recombinant DNA technology, we deleted the ColEl cruciform from pColIR3 15 (generating a plasmid called pColIRAxba), and replaced it with a new inverted sequence which was very similar to the inverted repeat of pIRbke8 (the central 20 bp were identical). Thus we constructed a plasmid in which a bke-like cruciform resides in the context of the (A + T)-rich ColEl sequences.Cruciform extrusion in the new plasmid (pIRxke/col) exhibited maximal rates at 0 mM NaCl, with an E, of 215 kcal mole-', i.e. properties typical for C-type cruciform extrusion. 2. We also performed the experiment in reverse. We cloned oligonucleotides to generate a ColEl inverted repeat in the BurnHI site of pAT153, the location at which the bke inverted repeat of pIRbke8 normally resides. The resulting plasmid (pIRCol/vec) exhibited typical S-type extrusion kinetics, i.e. maximal extrusion at 50 mM NaCI, with an E, of 50 kcal mole-'. These results show that the kinetic class, and thus the mechanistic pathway, of cruciform extrusion is determined by sequences which lie outside the inverted repeat. The sequence of the inverted repeat itself seems to be of secondary importance in this selection.The (A + T)-rich ColEl flanking sequences appear to confer C-type extrusion kinetics on whatever sequences are placed next to ')') K. M. Sullivan and D. M. J. Lilley, Cd, 1986,47, 817 827 Structural Isomerizution in DNA them, although we may well discover inverted repeat sequences refractory to their influence. We have termed them C-type Inducing Sequences. Manipulation of these sequences has revealed the following proper tie^:"^^' Oo Only a single sequence is required. Either the right- or left-hand ColEl sequences function as independent inducing sequences, and some other (A + T)-rich sequences, e.g. a 200 bp fragment from Drosophilu, may replace them both. However, not all (A + T)-rich sequences are active. Where an inverted repeat is flanked by one C-type and one S-type sequence, the dominance is determined by the salt concentration, i.e.the kinetics are C- and S-type at 0 mM and 50 mM NaCl respectively. Polarity can be unimportant. The left-hand side 100 bp ColEl sequence can confer C-type kinetics on inverted repeats placed at either end in different con- structs. The effects may be modulated over significant lengths of DNA. We have observed effects transmitted over 100 bp. 5. The C-type inducing effect may be blocked by insertion of the (G + C)-block GCCCCGGGGC between the element and the inverted repeat. The same sequence on the far side of the inverted repeat does not prevent C- type extrusion.We have also cloned random Suu3A restriction fragments from the plasmid pBR322 in the same interposed location. We found that many fragments block C-type extrusion, while others allow it to proceed with varying efficiency. Overall we concluded that the ability of a given sequence to transmit the effect depends on base composition and length. We define a transmitting sequence as one which does not block the effect of the inducing sequence (it may even augment the effect, i.r. extrusion may proceed at lower temperatures), but which by itself cannot act as an inducing sequence. 6. By systematic deletion analysis, using Ba13 1 exonucleolysis and re-cloning, we have identified a region of 30 bp of very (A + T)-rich DNA in the ColEl left-hand side flanking sequence (termed co130), which is very import- ant for C-type induction.We have cloned a synthetic oligonucleotide of the same sequence, which is sufficient to confer C-type kinetics on an adjacent inverted repeat. Furthermore, systematic sequences related to co130 may act as inducing sequences, and we have obtained C-type extrusion using a piece of DNA as short as 12 bp. 9 Origins of C-type Induction-A Dynamic View of Supercoiled DNA We have demonstrated that the sequences which control the selection of the pathway of cruciform extrusion lie outside the DNA directly participating in the transition. How can the (A + T)-rich elements affect the entire kinetic character at a remote location? (A + T)-rich DNA is structurally polymorphic and "") K.M.Sullivan.A. 1. H. Murchie. and D. M. J. Lilley. J. Bid. Chet~i.,1988. 263, 13074 13082 Lilley dynamic.56.101-109 We propose that these sequences are responsible for a coordinate destabilization of a large domain of the supercoiled DNA, thus increasing the probability of large-scale base opening in the inverted repeat itself. We have made a number of predictions based on this model, which have been tested experimentally. 1. If helical instability in the inducing sequences is responsible for cruciform extrusion by the C-type mechanism, perhaps stabilization of these regions will reduce their effect. We have employed the DNA-binding antibiotic distamycin 'lo to stabilize the (A +T)-rich sequences. Distamycin binds to runs of successive AT base-pairs in the minor groove,24 and stabilizes the double helical structure.' ' ' We have observed the complete suppression of C- type induction by inclusion of 5 pM distamycin into the extrusion buffer.' l2 Moreover, helical stabilization of pColIR315 by a combination of 50 mM NaCl and 3 pm distamycin resulted in kinetics which were indistinguishable from typical S-type kinetics.2. Conversely, if C-type extrusion requires a domain of reduced DNA stability, perhaps this can be approximated in a normally S-type molecule by the use of a helix-destabilizing agent. We observed that quasi-C-type extrusion (extrusion at relatively low temperature, in the absence of added salt) may be induced in normally S-type sequences (pIRbke8 for example) by inclusion of helix- destabilizing solvents such as 4004 dimethyl formamide into the extrusion buffer.' l2 The same solvent concentration also reduced the temperature required for C-type extrusion.Helical stability in inducing sequences may also be further reduced by methylation of A, and we have found that this may considerably reduce the temperature at which C-type cruciform extrusion pro- ceeds. 3. If the inducing sequences possess a reduced helical stability, perhaps this might be manifested as a hyper-reactivity towards chemical attack. The ColEl inducing sequences (in pColIRAxba) are chemically reactive towards bromo- acetaldehyde, glyoxal, and osmium tetroxide 3-examples are presented in '01 S. Arnott and E. Selsing, J.Mol.Biol., 1974;88, 509 521. I01 A. Klug. A. Jack, M. A. Viswamitra, 0. Kennard, Z. Shakked, and T. A. Steitz. J. Mol. Biol., 1979. 131.669 680. A. Mahendrasingham, N. H. Rhodes. D. C. Goodwin. C. Nave, W. J. Pigram. W. Fuller. J. Brahms, and J. Vergne. Nirrrrre. 1983,301. 535 537. 'O' F. Eckstein and T. M. Jovin, Biocheniistrj~,1983. 22.4546 4550. I0 i D. J. Patel. S. A. Kozlowski, D. R. Hare, B. Reid, S. Ikuta. N. Lander, and K. Itakurd, Biochemisfrj.. 1985.24.926 935. 'Oh D. J. Patel, S. A. Korlowski, M. Weiss, and R. Bhatt, Bioc,i~eniistrj~.24, 936 944. J. A. McClellan, E. Palecek. and D. M. J. Lilley, Nuc,lric Acids Rex, 14. 9291 9309. 'Ox J. W. Suggs and R. W. Wagner. Nucleic AciA Res., 1986, 14, 3703 3716. I OY M. J. Lane.S. Laplante, R. P. Rehfuss. P. N. Borer, and C. R. Cantor, Nucleir, Acids REX, 1987, 15, 839 852. II(1 C. Zimmer. K. E. Reinart, G. Luck. U. Waehnert, G. Loeber, and H. Thrum, J. Mol. Biol., 1971, 58. 329 338. I ' K. E. Reinart. E. Stutter, and H. Schweiss, Nucleic Acid.? Re.?.. 1979,7, 1375 1392.~ I '' K. M. Sullivan and D. M. J. Lilley, Nucleic. Acid7 Rex, 1988. 16, 1079 1093. J. C. Furlong. K. M. Sullivan. A. I. H. Murchie. G. W. Gough, and D. M. J. Lilley, Biocirmiisrrj,. 1989,28,2009 20 I 7. 75 Structural Isomerization in DNA A Bromoacetaldehyde B Glyoxal C Osmium tetroxide oso, bP DP 430 500 A colR 427,417,413 370 227 Xbal COIL ,":i 2' 311 181 249 83 200 60 151 Figure 9 C-type inducing sequences are chemically hyper-reactive in supercoiled DNA.The reuctivity towurds a number of single-strand selective chemical probes of the sequences in and uround the inverted repeat has been studied in supercoiled plasmid molecules. Two plasmids huve heen studied: pColIR3 15 contains the ColEl inverted repeat, extruded as a cruciform .structure, in its (A + T)-rich Junking context, while in pColIRAxba the inverted repeat has been deleted. Thus pColIRAxba allows the study of the (A + TI-rich flanking sequences in [he ubsence of cruciform,forniution. (A) Bromoucetaldehyde: Supercoiled pColIR3 15 and pColIRAxba were reacted with hromoucetaludeh"vd~~at 37 "C, complete cleavage with the restriction enzyme BamHI, ,fi)llo~wiby incubation with S1 nucleuse to cleave uny base adducts.A.fter radioactive luhelling. the products were anulysed by gel electrophoresis in agarose and uutoradiography. Supercoiled pColIR3 15 and pColIRAxba were also cleaved with S1 nuclease followed by BumHI. The distunce.s of the cleavages introduced into the molecule .from the unique BamHI site cun be obtained by comparison with the marker DNA ,fragments (phuge cpX174 DNA clcuwd wilh HinfI) at the lefi (sizes indicated on left). Note that S1 nuclease cleavage qf pColIR3 15 1eud.s to one sharp hand, corresponding to cleavage on the single-stranded cwcifbrm loop. By comparison, bromoucetaldehyde modification of this plasmid results in reaction oucr a wide region of the,flanking sequences. Sicpercoiled pColIRAxba was uncleaved by Sl nuclease (there is no cruciform to cut in this molecule), but the (A + T)-rich flanking Lilley sequences remain stronglj.reactive to modijication by bromoacetaldehyde. Thus these sequences are intrinsically reactive to this probe, tiahether or not the cruciform is present. (0)Glyosal: E.uperiments uvre carried out analogous to those using hromoacetaldehyde. Once again the flanking regions of supercoiled pColIR3 15 and pColIRAxba are chemicallyreactitre, irrespective of the presence or absence of the cruciform. (C) Osmium tetra-uide: Linear and supercoiled pColIRAxba Miere reacted with osmium tetro.uide, and the DNA (after restriction cleavage and radioactive labelling) Mias reacted with hot piperidine, which results in strand scission at osmium-cis-diesteriJied thymines.The DNA was electrophoresed on a sequencing gel, tt3hich resolves to the level of a single nucleotide. Bycomparing the positions of strand cleavage with the four chemical sequencing reactions (left) 1i.e can deduce the position of reactive thymines in the sequence. The position of the XhaI site, i.e. the position from which the ColEl intrerted repeat was deleted in pColIRAxba, is indicated on the left. While the reaction of linear pColIRAxba has not resulted in any thymine modiJication, the situation is very diflerent in the supercoiled molecule. All the thymines in the sequence to the 5' (called colL) side of the XbaI site are strongly reactive. This is precise!,? the region v.hich has been identijied by manipulation of the sequences to be the most important in conferring C-type cruciform e.utrusion on inverted repeats located at the XhaI site Figure 9.The left- and right-hand side ColEl sequences are independently reactive, as are other inducing sequences. In many ways the properties of the chemical reactivity parallel those of the C-type cruciform extrusion kinetics. Both reactivity and C-type extrusion are suppressed by addition of either salt or distamycin for example. Reactivity of these elements is dependent upon the presence of negative supercoiling, the threshold for which is salt-dependent. The temperature dependence of the chemical reactivity is revealing. The halogenoacetaldehydes require temperatures above 30 OC (this can be reduced by inclusion of a helix-destabilizing solvent like dimethyl formamide) and give profiles with a cooperative appearance, while osmium tetroxide may react at 5 "C and is less cooperative. It may be noted that the adduct formed by the halogenoacetaldehydes (principally 1,N6-ethenoadenine) prevents A.T base-pairing, and may therefore be regarded as a true probe of the absence of Watson-Crick base-pairing. By contrast, osmium tetroxide adds across the 5,6-double bond of thymine, and does not interfere directly with base-pairing.However, the required out-of-plane direction for electrophilic attack will normally be hindered by base-stacking, and thus we believe that reactivity may indicate unstacking events in the (A + T)-rich DNA.On this basis, we see temperature as a means of dissecting the helix opening process, with transient unstacking events at low temperatures, and cooperative opening events as the temperature is raised. The latter openings are responsible for the reactivity towards bromoacetaldehyde, and are probably closely similar to the events required for C-type extrusion. Recent evidence suggests that we may be able to uncouple the initial opening and subsequent cruciform formation steps of the C-type mechanism by changing ionic conditions. Thus we find that we can interconvert almost at will between the two mechanisms by perturbing DNA helical stability, and that we can observe the opening in the (A + T)-rich sequences directly with chemical probes.A propen-sity for coordinated opening in these sequences is also revealed by a theoretical Structural Issomerization in DNA C-type pCollA315 Temperature ("C)3 1.0 . 62 P(i)Probabilityof base-opening 0.5 Sequence PSfl EcoRl BamHl Safl v 1, T v ColE 1 IR S-type h 65 plRbke8 Temperature ("C) 1 .o 62 Sequence -800 -400 0 Psil ECOR I SaA 1 1 -Y I bke IR Figure 10 Calculated helix opening probability projiles for typical C-type and S-type sequences. Statistical thermodynamic helix-coil theory was used to calculate helix opening Probabilities [P(i)] .for each base-pair along the PstI to SalI sequences of pColIR315 (A) and pIRbke8 (B), over the temperature range 335 to 338 K in intervals of 0.2 degrees.These have been stacked vertically for each sequence. The ordinates are the sequences of the DNA molecules, measured in base-pairs. Inverted repeats are indicated by jilled boxes, and the (A + T)-rich ColEl sequences by an open region. Note the large region of cooperativemelting calculated for pColIR3 15, comprising the entire ColEl sequence including the inverted repeat, and the absence of this exfect in pIRbke8 approach. We have employed a statistical mechanical approach to calculate the helix opening probabilities of the DNA sequences used in these studies,' l4 based on DNA helix-coil theory. We find that the sequences which are experimentally observed to be C-type inducing sequences have high predicted probabilities of F. Schaeffer, E.Yeramian, and D. M. J. Lilley, Biopolymers. 1989,in the press. Lilley opening as cooperatively melting units at relatively low temperatures. This can be seen clearly in the profile of base-opening probability for pColIR315 shown in Figure 10. In contrast, the S-type sequences are predicted to be quite stable helical structures at the same temperatures. We have carried out calculations of this type for most of the constructs examined experimentally, with an almost perfect correlation. Moreover, the agreement is quantitative. We can correlate the temperature calculated for 50% opening probability in the inverted repeat with the experimental temperature required for 50% extrusion in a 5 min incubation, with a correlation coefficient of 0.93. Thus it seems possible that the mechanism of C-type induction is a striking example of telestability effects,' ' where helical instability is a cooperative property in an entire domain of DNA.We cannot rule out a more dynamic component of the process, whereby fluctuations arising in the (A + T)-rich elements are transiently mobile,' 16,' l7 although the likely lifetime of soliton-like states * in DNA makes diffusion over significant distances rather improbable. We find there is a good correspondence between the character of supercoiled DNA revealed by the study of cruciform extrusion, chemical reactivity, and the application of statistical thermodynamics. In a sense, we may regard the experiments on cruciform extrusion as simply a probe of melting or premelting events in torsionally stressed DNA.The dynamic view of openings in DNA is very different from the rather static picture for linear DNA emerging from the n.m.r. experiments, and illustrates the 'activation' of DNA structure brought about by negative supercoiling. 10 Other (A + T)-rich Sequences-Some Outstanding Questions Cruciform extrusion by the C-and S-type mechanisms explains a great deal of experimental data. However, there are several sequences which do not readily fall into these categories, and which provide interesting tests for the mechanistic framework developed. A. Alternative Adenine-Thymine Sequences.-A number of years ago we de-monstrated cruciform formation in supercoiled plasmids containing alternating (AT)" sequence^,'^ such as pXG540.This molecule contains a section of Xenopus luevis sequence cloned from the xT1 globin gene, that includes the 68 bp sequence (AT)34. Cruciform structures formed from (AT),, sequences (for an example see Figure 11) have a number of interesting properties, compared to other inverted repeats. First, they have relatively low free energies of formation. Second, they exhibit anomalously low twist changes on extrusion, implying a local underwind- 'I5 J. F. Burd, R. M. Wartell, J. B. Dodgson, and R. D. Wells, J. Bid. Chem., 1975,250,510!--5113. 'I6 S. W. Englander, N. R. Kallenbach, A. J. Heeger, J. A. Krumhansl, and S. Litwin, Proc. Natl. Acad Sci.USA, 1980.77. 7222 7226.~'" H. M. Sobell, T. D. Sakore, S. C. Jain.A. Bannerjee. K. K. Bhandary, B. S. Reddy, and E. D. Lozansky. Cold Spring Horhor Symp. Quont. Biol.,1983,47,293--314. * Some have tried to apply the solid-state physics concept of solution state to DNA molecules, in which transient conformational fluctuations behave like self-reinforcing waves that move along the double helix at sonic velocity. Structured Isomerizcition in DNA Figure 11 Formation of a cruciform by an alternating adenine-thymine sequence revealed by two-dimensional gel electrophoresis. The plasmid pXG540 contains the sequence (A T)34,an inverted repeat. A sample of pXG540 comprising a range of topoisomers from (3 = +0.005 to -0.08 M~USelectrophoresed from a single circular well on an agarose gel. After electrophoresis the gel u~s soaked in buffer containing an intercalating compound (1.8 pg/ml chloroquine)which shijts the supercoiling such that the middle of the topoisomer range is now relaxed, while the more underlinked topoisomers remain negatively supercoiled (but reduced in superhelix density).The gel M'US then turned 90" and electrophoresed once again, in the presence of' the same intercalator Concentration. The directions of the two electrophoresis dimensions is shown by the arrows. The topoisomers marked R1 are close to being fully relaxed in the absence of intercalation, and therefore exhibit the slowest mobility in the jrst dimension. In the absence of structural transitions, the topoisomers would move down and left from this point, turning around to migrate rightwads at the point at which the topoisomers arc' appro.ximate1.t: relaxed in the second dimension (R2), i.e.in the presence of the chloroquine. A topology-depL.ndL.nt structural transition, generating a local twist change, cuuses a discontinuity in the path of the topoisomers, because the formation of the structure at a threshold level of supercoiling results in a relaxation, and hence upward mobility shift, of the topoisomers (see Figure 4). This change does not occur in the second dimension, because of the relative relaxation caused by the chloroyuine intercalculation. Thus between topoisomers A and Lille-v B there is u pronounced discontinuit?, in the migration of the supercoiled species. Counting fiom R, there is ALk of -9 .for the midpoint of the transition, and counting buck the transition corresponds to a AT,!, of -5.5.From this we may calculate the free energy of ,fi)rmution yf the (AT)34cruciform structures. Species N is the open circular and species L the linear,jbrm of pXG540 ing of the helical DNA before cruciform formation. Third, there appears to be no kinetic barrier for the formation of these cruciforms. Even at low temperatures these sequences undergo cruciform extrusion as soon as negative supercoiling is restored by removal of intercalating drugs. Are (AT),, sequences a new kinetic class of cruciform, different from either C- or S-type? There may be two reasons for the unusual kinetic character of the (AT),, cruciforms. First, the ground state for the reaction may be perturbed.The twist discrepancy on cruciform formation, together with an observed chemical reactivity of the tracts, have led us to propose that (AT),, sequences possess an abnormal structure, characterized by a susceptibility to torsional deformation at every second base-pair step. Second, the (AT),, sequences are expected to be extremely easily denatured. Dinucleotide stability constants measured in thermal melting experiments l1 indicate that (AT),, tracts are the least stable of any DNA helix, and our statistical thermodynamic calculations l4 indicated that the (AT)34 tract of pXG540 will undergo a cooperative melting ten degrees lower than the ColEl flanking sequences. Thus perhaps in these sequences we have an example of DNA which is simultaneously inverted repeat and particularly effective inducing sequence, such that extrusion is already rapid at the lowest temperatures at which we can work.In this model, the (AT),, sequences are not in a new class, but are a special sub-class of C-type cruciforms. B. Very (A + T)-rich Flanking Sequences. While the ColEl sequences are certainly (A + T)-rich, it is nevertheless possible to find sequences which are even richer. However, such sequences are not necessarily better inducing sequences. An (A + T)-rich 200 bp fragment of Drosophilu melanogaster DNA was cloned adjacent to the bke inverted repeat."' We found that this successfully induced C-type cruciform extrusion in the absence of added salt, but the temperature required was several degrees higher than that for the ColEl sequences.However, the Drosophilu sequence was rather more (A + T)-rich, and contained long A,, runs in some places. This was the first indication that base composition was not a completely reliable guide to the ability of a sequence to act as a C-type inducing sequence. This observation was extended when we examined a series of DNA fragments cloned from Dictjjostelium, which were still more (A + T)-rich. To our surprise, these sequences failed completely to cause C-type cruciform extrusion in an adjacent inverted repeat. Moreover, the effect was dominant over normally functional inducing sequences, for when an inverted repeat was flanked by a ColEl sequence on one side, and a Dictyostelium sequence on the other, no cruciform extrusion was obtained in the absence of salt.The present working hypothesis is that these extremely (A + T)-rich 118 0.Gotoh and Y. Tagashira, Biopo!,wiers, 1981. 20, 1033-1042 81 Structural Isonzerization in DNA sequences adopt some conformation which actively prevents cruciform extrusion by the neighbouring inverted repeat. Further experiments are required to obtain a better understanding of this system. 11 In conclusion DNA is potentially a highly polymorphic molecule from a structural point of view, and this polymorphism is accentuated by negative supercoiling. Not all the available structural variants may be exploited in biology, but the study of these features leads to an enhanced understanding of the structural properties of DNA in general.The cruciform is just such an example. To date, no biological function has been ascribed to a cruciform structure, yet it is an extremely interesting entity from several points of view. Structurally it is an excellent model for recombination intermediates and for the study of recombination nucleases. The transition between unextruded and extruded conformations is now probably the best understood structural transition which DNA undergoes. In the course of these studies we have discovered new ways to examine the dynamic properties of supercoiled DNA, and revealed hitherto unsuspected long-range interactions between sequences. The combination of cruciform kinetic studies with chemical probing and theoretical methods is proving a new description of the physical chemical properties of DNA under superhelical stress, and is set to provide some general models of the dynamics of local sequences in supercoiled DNA. Acknowledgements. I am extremely grateful to my co-workers and collaborators: Alastair Murchie, Karen Sullivan, Judy Furlong, and Francis Schaeffer. I thank the MRC, SERC, CRC, Wellcome Trust, and the Royal Society for financial sup- port.

ISSN:0306-0012    

DOI:10.1039/CS9891800053

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chetn. SOC.Rev., 1989, 18,83-122 Recent Developments in the Synthesis of myo-InositolPhosphates By David C. Billington MERCK SHARP AND DOHME RESEARCH LABORATORIES, NEURO-SCIENCE RESEARCH CENTRE, TERLINGS PARK, EASTWICK ROAD, HARLOW. ESSEX, CM202QR 1 Introduction This review aims to present the current state of the art in inositol phosphate synthesis. A brief introduction to cell signalling is provided and the importance of the inositol phosphates in this area is outlined. This is followed by an introduction to the stereochemical properties of myo-inositol and a description of the problems posed by the synthesis of the inositol phosphates. An outline of the methods developed recently to overcome these problems is then followed by the bulk of the review which is concerned with the synthesis of the individual inositol phosphates, arranged according to their degree of phosphorylation.A. Intercellular Signalling.-Communication between cells is essential for the maintenance of life processes in complex organisms. Nature uses a wide variety of chemical signals to pass information from one cell to another, and these signals may be classified according to their function as neurotransmitters, hormones, growth factors, etc. These chemical messengers are received by specific receptors located on the surface of the receiving cells and the incoming signal is then converted into a response inside the target cell by a number of different mech- anisms.’ One class of receptors either contains or is closely linked to an ion channel which spans the plasma membrane, Figure 1.Stimulation of the receptor leads to alterations in the ability of the ion channel to allow the ions for which it is specific to pass across the cell membrane. A second class of receptors are membrane-spanning enzymes called tyrosine kinases. Binding of the extracellular messenger to the receptor leads to activation of the enzyme which then phosphorylates specific tyrosine residues in target proteins inside the cell. This system is used by many growth factors and hormones such as insulin. A third class of cell surface receptor exists which have no intrinsic activity as either ion channel control systems or as enzymes. These receptors are ‘coupled’ via a class of proteins known as ‘G proteins’2 to the enzymes or ion channels through which they evoke their responses.Activation of the receptor causes its associated G protein to release guanosine diphosphate (GDP) and bind guanosine triphosphate (GTP). In this form the G protein regulates the activity of the ion channel or enzyme ’Molecular Mechanisms of Transmembrane Signalling‘. ed. P. Cohen and M. D. Houslay, Elsevier, Oxford, 1985.’R. L. Rawls, Clicwi. Eq. Neits. 1987. 26. INTRACE L L ULAR 1 4v R1 is an ion channel linked Receptor. R2 is a membrane spanning tyrosine kinase. R3 is a receptor linked via a G protein[(;] to an enzyme[E] which produces second messengers. Figure 1 which forms the next link in the chain. In the case of receptors which are coupled to enzymes, these enzymes act to change the intracellular levels of ‘second messenger’ species.The classical example of this type of receptor system produces the second messenger cyclic adenosine monophosphate (CAMP) from adenosine triphosphate (ATP) uiu the G protein regulated enzyme adenylate cyclase. The alteration in intracellular levels of CAMP produced on stimulation of this type of receptor causes the observed overall cellular response. In recent years overwhelming evidence has been obtained for a new widely distributed second messenger system involving the hydrolysis of inositol * phospholipids. A detailed description of this complex system, which is still under intensive investigation, is far outside the scope of this review.Only a very brief overview of current ideas is given here in order to place the synthesis of the inositol phosphates in context, and the interested reader is directed to the many recent reviews available on the biochemistry of the PI * Throughout this review ‘inositol’ refers to the nijw stereochemistry unless otherwise indicated N. N. Osbourne. A. B. Tobin, and H. GhaLi, N~JLI~O~~W~.‘J. Altman, ,%‘[///w. Rc.r.. 198X, 13, 177. 19x8. 331. 119.’J. L. Marx. Scicwcx,, 1987, 235. 074. ” C. W. Taylor, 7‘rcwd\ P/it/rm/dSci.. 1987. 8. 79; A. H. Drumrnond, ihicl.. 1987, 8. 129.’M. D. Hotislay. Tro~itbBioc./ic,/ii.Sci.. 1987, 12. 133. R. H. Michell. Nc//i/rc,,19x6, 324. 613. ‘’ C. P. Downes. 7’rcvitl.s Ncirrochcwi. Sci.. 19x6, 394. “I 1’.W. Miijerus. T. M. Connolly, H. Deckmyn, T. S. Ross, T. E. Bross. H. Ishii, V. S. Bansal, and D. B. Wilson, Sc,icvic~c,.19x6. 234. IS 19. I I R. H. Michell, No/io.c, 19x6. 319. 176. C. W. Taylor and J. E. Merritt, 7>ewt/.\ P/icir./ritrcd. Sci.. 19x6. 7, XX; S. R. Nahorski and I. Batty, ihicl.. 1 YXO, 7, x3. 84 Billington INTRACELLULAR EXTRACELLULAR 1,4 1IP, myo-INOSITOL \ NOSITOL- II-I-P HO"' Figure 2 B. Inositol Phosphates as Second Messengers.-In 1975 Michell suggested that the receptor-controlled hydrolysis of inositol phospholipids could be directly linked to changes in calcium levels within cells. It is now known that receptor- stimulated hydrolysis of the phospholipid phosphatidylinositol 4,5-bisphosphate (PIP2) occurs via a G protein-controlled enzyme, phospholipase C (PLC), releasing two second messengers, diacylglycerol (DAG) and D-inositol 1,4,5- trisphosphate (1,4,5-IP3), Figure 2.3-6 Diacylglycerol acts as a second messenger by binding to and activating protein kinase C, and also serves as a precursor to the metabolites of the arachidonic acid cascade.1,4,5-IP3 binds to specific receptors on the endoplasmic reticulum and stimulates the release of calcium from intracellular storage sites. The complex metabolic cycle that converts 1,4,5- IP3 into free inositol, which is used for the resynthesis of PIP2, has been an area of intense biochemical interest during the past five years. It is now clear that two pathways exist for the metabolism of 1,4,5-IP3, Figure 2.The first pathway begins with a specific dephosphorylation giving inositol 1,4-bisphosphate (1,4- IP2). This bisphosphate is subsequently sequentially dephosphorylated, mainly uiu inositol 4-phosphate (I-4-P) to free inositol. This pathway probably serves only to terminate the 1,4,5-IP3 signal. The second, recently discovered, pathway begins with a specific phosphorylation of 1,4,5-IP3 giving inositol 1,3,4,5-tetrakis- phosphate (1,3,4,5-IP4). Recent experiments have suggested that this 1,3,4,5-IP4 may have a messenger role of its own, affecting the influx of calcium into the cell from the external medium. The tetrakisphosphate is then metabolized via a I.' R. H. Michell. Biochitii. Biopi1Jx Acra, 1975, 415, 81. Recent Developments in the Synthesis of myo-Inositol Phosphates second trisphosphate, inositol 1,3,4-trisphosphate (1,3,4-IP3), and two further bisphosphates, inositol 1,3- and 3,4-bisphosphates (1,3- and 3,4-IP2) to inositol monophosphates.These monophosphates are then converted into free inositol as before. Enzymic pathways also exist which convert 1,3,4-IP3 into a second tetrakisphosphate, 1,3,4,6-IP4, which is believed to be sequentiaIly de-phosphorylated once more to give free inositol. The relative importance of these latter pathways remains to be established. Inositol polyphosphates also exist in other systems, for example avian red blood cells contain 1,4,5,6-IP4 and 1,3,4,5,6- IPS, whose function and metabolism are not clear at the present time.The rapidly expanding complexity of the metabolic cycle, coupled with the fact that only minute amounts of these metabolites were available from natural sources, stimulated intense worldwide interest in the chemical synthesis of the inositol polyphosphates during the mid- 1980s. Many laboratories have subse- quently contributed methodology and techniques such that most of the naturally occurring metabolites have now been synthesized and are available in quantities sufficient to allow the isolation of the individual enzymes of the PI cycle. The availability of these substrates will prove crucial to the next generation of biological studies of this fascinating fundamental cell-signalling system. C. myo-Inosito1.-An optically inactive cyclohexanehexol was first isolated in 185014 and given the name ‘inosit’ by Scherer.This became a generic term for cyclohexanehexols with the suffix 01 added in English usage. There are nine isomeric inositols: allo, (+)-chiro, (-)-chiro, cis, epi, muco, myo, neo, and scyllo, Figure 3. The naming and numbering of the inositols and their derivatives pose many complex problems for which several different solutions have been advocated and used. IUPAC have published the ‘1967 IUPAC/IUB tentative rules for cyclitols’.’ For a basic guide to nomenclature, and a concise coverage of inositol phosphate chemistry up to 1980 the reader should consult the excellent book by Cosgrove. ’ 6o A recent review on nomenclature and stereochemistry is also avail- able.’ 6b Fortunately the only isomer of inositol which we are concerned with in the context of receptor signalling is myo-inositol, Figure 3.myo-Inositol has a single axial hydroxyl at C-2, and a plane of symmetry between C-2 and C-5, Figure 4. To emphasize this, the two-dimensional representation in Figure 4 will be used throughout this review and the inositol carbon atoms will be numbered anticlockwise, with C-1 at 2 o’clock. Incorporation of a substituent at C-2 or C-5 leads to an optically inactive compound (plane of symmetry retained), whereas substitution at C-1 (enantiotopic to C-3) and/or C-4 (enantiotopic to C-6) leads to a pair of enantiomers (plane of symmetry lost). Thus inositol 2-phosphate, 5-phosphate, and 1,3-bisphosphate are achiral, whereas inositol 1-phosphate, 4-phosphate, and 1,4-bisphosphate all exist as pairs of enantiomers, Figure 5.The l4 J. Scherer, Ann., 1850,73, 322. Is IUPAC Information Bulletin, 1968.32, 51.’’ (a) ‘Inositol Phosphates. their Chemistry, Biochemistry and Physiology’, D. J. Cosgrove, Elsevier, Oxford, 1980 (h) R. Parthasarathy and F. Eisenberg, Biocltem. .I.,1986,235,313. Billington H*J$&HO HO OM OH OH ailo ( + 1-chiro (-I-chiro HO HO%OH OH cis epi muco 'OMOHHO "W0" neo scyllo Figure 3 H HO$ 6 H 'OHOH HO'*4 .'OH OH my0 -Inositol (1) I I t = Symmetry planeI I Figure 4 enantiomers of inositol 1-phosphate shown in Figure 5 exemplify the fact that it is equally correct to number the inositol carbon atoms clockwise, with C-1 at 10 o'clock.The normal convention is to choose the alternative which leads to the lowest overall number count. As outlined above a C-1 at 2 o'clock anticlockwise Recent Developments in the Synthesis of myo-lnositol Phosphates (+) -and (-1 -Inositol 1 -phosphate HOQ HO'. *'OH OH Inositol 2 -phosphate Inositol 5 -phosphate Figure 5 system will be followed in this review, and any examples which require the opposite numbering system to generate the correct chemical name will be indicated in the schemes by numbering of the ring atoms. The symmetry properties of myo-inositol have been exploited in a number of synthetic strategies. D. Synthesis of Inositol Phosphates-The synthesis of the inositol phosphates poses three main problems: (i) the synthesis of a suitable selectively protected inositol derivative; (ii) phosphorylation in an efficient manner with a reagent bearing suitable phosphate-protecting groups (this is a particularly serious problem for polyphosphates containing vicinal diols where cyclic phosphate formation is a major side reaction) and (iii) deprotection without migration of phosphate substituents to adjacent free hydroxyl functions. Although considerable methodology was available for the synthesis of selectively protected inositols, it was not until the mid-1980s that the problems of efficient polyphosphorylation, and deprotection without migration were solved. An additional requirement is for efficient procedures for the resolution of suitable synthetic intermediates, allowing the preparation of optically pure inositol phosphates.(i) Synthesis of Protected Inositols. Due to the ready availability of pure myo-inositol, most syntheses begin with the parent cyclitol. Reaction of inositol (1) with cyclohexanone,' or more efficiently with a cyclohexanone precursor such as l-ethoxycyclohexene,'8*'9 in the presence of an acid catalyst gives a mixture of three bisacetals [(2)-(4)] which may be separated by crystallization and chromatography, Scheme 1. Each of these bisacetals gives the monoacetal (5) on mild hydrolysis of the less stable transacetal. ",' Due to the conformational constraints imposed on the inositol ring by the bisacetal groups, each of the free "S.J. Angyal, M. E. Tate, and S. D. Gero, J. Client.Soc., 1961,4116. lRR. Gigg and C. D. Warren, J. Client.Soc. (C), 1969,2367. l9 D. E. Kiely, G. J. Abruscato, and V. Baburao, Carholqdr. Res., 1974,34, 307. Billing ton \(*) (5) Scheme 1 hydroxyl groups in (2), (3), and (4)may be selectively manipulated under suitable conditions,20 providing access to a series of inositol derivatives having five hydroxyl groups differentially protected. The selective hydrolysis of the trans acetal, coupled with the possibility of selective reactions at specific hydroxyl groups in (2), (3), and (4)*"before hydrolysis of the less stable acetal has led to the widespread use of (2), (3), and (4) in synthesis. Other acetals such as the isopropylidene and cyclopentylidene derivatives have also been used.2 Recently the mono orthoformate of inositol (6) has been isolated and characterized,22 Scheme 2.The orthoformate (6) provides' B derivative in which positions 1, 3, and 5 are simultaneously protected. In addition, the normal axial/equatorial relationship of the remaining free hydroxyls is reversed, Scheme 2. The spatial juxtaposition of the axial hydroxyls in (6) allows highly selective alkylations to be performed at these positions,23 Scheme 3. (ii) Phosphorylation Methodology. Two successful general strategies have been P. J. Garegg. T. Iversen, R. Johansson, and B. Lindberg, Carholzydr. Res., 1984, 130, 322 and references therein. J. Gigg. R. Gigg, S. Payne, and R. Conant, Curbolzydr.Res., 1985, 142, 132; ibid., 1985, 140, ~143. 22 H. W. Lee and Y. Kishi, J. Org. Chem., 1985,50,4402. Recent Developments in the Synthesis of myo-lnositol Phosphates Hoq:c(I:..-DMSO 100 OC H OHO 4 OH OH OH HoQklH HO"Ope"HO" OH Scheme 2 H H NaH ' R-X DMF H O HO q HO HO OH OR R = benzyl , allyl, etc. Scheme 3 developed for the phosphorylation of polyhydroxy-inositol derivatives. The first involves reaction of alkoxide anions with the readily available, crystalline reagent tetrabenzylpyrophosphate(7) (TBPP),23,24 Scheme 4.This method gives good results for polyhydroxy compounds incorporating up to four phosphate esters in a single reaction and in high yield. Vicinal diols are efficiently phosphorylated, and the benzyl protecting groups are readily removed by hydrogenolysis, without phosphate migrati~n.~~.~~ 23 D.C. Billington and R. Baker, J. Clrem. Soc.. Clzem. Commun., 1987. 101 1. 24 Y. Watanabe, H. Nakahira, M. Bunya, and S. Ozaki, Te~rahedronLet(., 1987,28,4179. 25 D. C. Billington, R. Baker, J. J. Kulagowski, and I. M. Mawer, J. Cliem. Soc., Clrem. Commun., 1987, 314. Billington 0 0 II II RO-M+ + (BnO)2P-O-P(OBn)2 0II +R-O-P(OBn)2 0II+ M+6P(OBn)2 Scheme 4 R-OH + X-P /OR' 'OR' R-0- p/OR' 'OR' R = e.g.CH ,CH2CN 0 S !,OR' 'ORR -0- R-0-P II,OR 'OR ' (10) (11) Scheme 5 The second phosphorylation strategy has developed from the use of PI1' species for the synthesis of polynucleotides.Reaction of the polyol with a Prrl reagent (8) with displacement of either NR2 or halogen gives an intermediate phosphite (9).26-28This phosphite may then be oxidized (e.g. using mCPBA) to the protected phosphate26 (10) or to the thiophosphate (S8 in pyridine) (ll), Scheme 5.29 Numerous variations on this theme have been used for the synthesis of inositol polyphosphates, with varying degrees of success, and these are discussed in full below. (iii) Deprotection. Migration of phosphate groups via cyclic intermediates [e.g. (1 means that strongly acidic or basic conditions need to be avoided, Scheme b. in addition, methods which lead to free ring hydroxyls adjacent to protected phosphate esters can also lead to migration. These problems are avoided by using benzyl esters as protecting groups for phosphate, in conjunction with benzyl ethers for protection of the ring hydroxyls. Hydrogenolysis rapidly cleaves the phosphate esters giving free phosphates which are not prone to migration during the slower hydrogenolysis of the benzyl ether on the ring.23,25 Cyanoethyl groups are also suitable as phosphate protecting groups in conjunction with benzyl ethers of the ring hydroxyls, as they may be deprotected without 2b K.-L.Yu and B.Fraser-Reid, Tetrahedron Lett., 1988,29,979. 2' A. M. Cooke. B. V. L. Potter,and R. Gigg, Tetrahedron Lelt., 1987,28,2305. 28 A. M. Cooke, R. Gigg, and B. V. L.Potter, Biochem. Soc. Trans., 1987,15,904. 29 A. M. Cooke, R. Gigg, and B. V. L. Potter, J.Chem. Soc., Chem. Commun., 1987, 1525. Recent Developments in the Synthesis of myo-Inositol Phosphates Ht or OH-OH J mixture of isomers Scheme 6 migration using alkali metals in liquid ammonia, with concomittant deprotection of the benzyl ether^.^^,^' A number of other successful strategies have been developed, which will be apparent from the individual syntheses described below. (iv) Resolution of Protected Inositols. Three resolution methods having general applicability have been reported for protected inositol derivatives. All of these methods rely on conversion of the racemic inositol derivative into a pair of diastereomeric esters, followed by separation of the diastereomers and regenera- tion of the separate enantiomers of the protected inositol, Scheme 7.While each of these methods (formation of orthoesters of D-mannose, formation of men- thoxyacetates, and formation of camphanate esters) is successful in certain cases, the most generally useful method available at the present time appears to be that involving the formation and separation of camphanate ester~,~’.~’ Scheme 7(c). The camphanate route appears the most attractive for the following reasons: (a) It uses (-)-camphank acid chloride, a stable crystalline reagent, readily available in high optical purity from a number of commercial sources. 30 (a)J. Gigg, R. Gigg, S. Payne, and R. Conant, J. Gem.Soc.. Perkin Trans. I, 1987, 1757; (b) T. Desai, J. Gigg, R. Gigg, and S. Payne, ‘Studies directed towards the synthesis of inositol phosphates of biological interest’, (Plenary Lecture) Glycolipids in molecular recognition and membrane organiza- tion, University of Sheffield, September 12th, 1988.Billington R*-X ($1 Protected inositol Diasteremeric derivatives Method (a) R* A Separated diastereoisomers CH CH3 enantiomeric protected inositols Method (c) R* = &; Scheme 7 (b) The camphanate esters formed are normally readily separated by chromato- graphy on silica gel, or by recrystallization. (c) HPLC analysis may be used to determine the diastereomeric purity of (d) The camphanate esters are often highly crystalline and suitable for X-ray Recent Developments in the Sjwthesis of myo-lnositol Phosphutes the intermediate esters, and thus the enantiomeric purity of the final pro- ducts. analy~is,~ thus allowing the absolute configuration of the individual en-antiomers to be established.(e) The route has been successfully used for a wide variety of inositol derivative^,^' many of which have not as yet been converted into inositol phosphates, and are thus outside the scope of this review. 2 Synthesis of Inositol Monophosphates There are four possible inositol monophosphates. Inositol 1-phosphate and inositol4-phosphate exist as pairs of enantiomers, while inositol 2-phosphate and inositol 5-phosphate are meso compounds with a plane of symmetry through C-2 and C-5. Both of the optically active monophosphates have been synthesized as pure enantiomers, in addition to racemic syntheses.A. Inositol 1-Phosphate.-The dextrorotatory myo-inositol monophosphate, isolated by alkaline hydrolysis of phospholipids, was shown to be D-( +)-myo-inositol 1-phosphate by Pizer and Bal10u.~~ This was followed by a number of syntheses of both the racemic compound and of its individual enantiomers. The absolute configurations of ( +)-and (-)-inositol 1-phosphate were established by synthesis of the (-) form from galactinol which was known to have the absolute configuration shown, Scheme 8. Galactinol was perbenzylated to the nonabenzyl derivative (1 3) which on hydrolysis gave the pentabenzyl inositol (14). Phosphorylation and deprotection then gave the laevorotatory inositol 1- phosphate (1 5).The naturally occurring dextrorotatory isomer is therefore the enantiomer of this material. i.e. (16). The isolation of (16) from natural lipids on a useful scale has been described, but it is not possible to isolate useful quantities of the enantiomer (15) from natural source^.^ 1332 (i) Syntheses qf Racemic Material ( +)-(19). Diacetyl-l,2-anhydroconduritol(17), Scheme 9, may be converted into the protected phosphate (18) by epoxide opening using dibenzylphosphate followed by acetylation of the free hydroxyl Permanganate oxidation of (18) followed by deprotection gives a mixture of inositol 1-phosphate (19) and inositol 4-phosphate (20). These isomeric materials may be separated by crystallization to give pure (+)-( 19) in very low yield.Benzylation of the monoacetal (9,see Scheme 1, followed by acidic hydrolysis gives 3,4,5,6-tetra-O-benzyl inositol (21), Scheme 10. Direct phosphorylation of (21) with POC13 gives a mixture of protected inositol 2-phosphate (22) and inositol 1-phosphate (23).' These isomeric phosphates may be separated by careful crystallization and (-I-)-(19) is obtained by deprotection of (23). 31 F. L. Pizer and C. E. Ballou, J. Am. Chem. Soc., 1959. 81. 915; C. E. Ballou and L. I. Pizer, J. Am. Chem. Soc.. 1960, 82, 3333. 32 C. E. Ballou, Biochetn. Prep., 1962,9,99. 33 N. Kurihara, H. Shibata, H. Saeki. and M. Nakajima, Liebigs Ann. Clzem., 1967,701,225. Billington Galactose-oQ OH H (BnO)4Galactose-pOBn Bn ____) HO" 'OH BnO' 'OBn OH OBn Galactinol (13) H2° 3 :H'fJ 2 steps4 HO" Bn0°' "OBn OH OBn (15) (-1-Inositol 1-phosphate (14) OP03H2 HO"HoQ.oH OH (16) (+) -lnositol 1 -phosphate Scheme 8 Q0 --•2 steps AcO" AcO'* OAc OAc (17) (18) 3steps HOAOPO,H, HO * 'OH OH (19) (3-Inositoi 1 -phosphate (20)(+)-Inositol 4-phosphate Scheme 9 Recent Developments in the Synthesis of myo-Inositol Phosphates HOQ 0-f‘6~10 BnOQ- 0-f6H10 - BWQ OH H HO-‘ **OH BnO‘* *‘OBn BnOO* *‘OBn OH Bn OBn (9421) ___) B n BnO’* O e oy2 steps OBn OBn (24) (25) OBn 1 HOQ OP03H2 + HOQ PO3H2 HO‘* *‘OH H0‘-**OH OH OH OBn (27)Inositol (2, -(19) (26) 2 -phosphate NB Scheme 10 96 Billing ton -.PhOH d (19) + (27) mixture - (2 8) (29) (26) (30) Scheme 11 A more selective approach is to exploit the higher reactivity of the equatorial alcohol in (21), for example by reaction with allyl bromide giving mainly (24).'* Benzylation of the remaining free hydroxyl group followed by removal of the allyl group gives (25), which may be phosphorylated using diphenylchlorophos- phate to give the pentabenzyl diphenylphosphate (26). Hydrogenolysis of (26) using palladium to remove the benzyl ethers, followed by platinum to remove the phenyl phosphate esters, gives (19) contaminated with 2.5% of the isomeric 2-phosphate (27).25 The isomeric product (27) presumably arises via formation of a cyclic phenylphosphate ester (29) from the intermediate diphenylphosphate (28) formed in the two-step deprotection procedure, Scheme 11.This problem may be overcome by transesterification of the fully protected compound (26) using the anion of benzyl alcohol to give the dibenzyl phosphate (30).25This dibenzyl ester may be deprotected without phosphate migration 25 using a single hydrogenolysis over palladium, as the benzylphosphate esters are cleaved much faster than the benzyl ethers, to give a free phosphate which is not prone to migration under these conditions, Scheme 11. (ii) ( -)-lnositol 1 -phosphate (1 5) and ( + )-inasitof 1 -phosphate (1 6). The synthesis of (15) from galactinol has been described ab~ve,~' Scheme 8. The optically active natural product quebrachitol (31) has also been used as a starting material in the synthesis of (15), Scheme 12.Treatment of (31) with cyclohexanone gives the bisacetal (32) which may be converted into tosylate (33).34,35 Boron trichloride removes both the acetal and methyl protecting groups, giving (34). Benzoylation of (34) gives (35) which on treatment with NaF in ""-dimethyl- 34 S. D. Gero, D. Mercier, and J. E. G. Barnett, Mefhorh Carbohyrlr. Chem., 1972,6,403. 3s D. Mercier. J. E. G. Barnett, and S. D. Gero, Tetrahedron, 1969, 25, 5681; S. D. Gero, Tetrahedron Lett., 1966, 591; D. Mercier and S. D. Gero, ibid., 1968, 3459. Recent Derelopments in the Synthesis of myo-Inositol Phosphates OH ___3 ___) CH,O" CH,O" OH (31) (32) (33) TsO.. 6:TsO ..OBzl 0 BzlO" .'OBzl HOM*6 J 0 Bzl OHII Bzl = CPh (35) (34) 3 steps Bzlo= 'OH OBzl OBzl OH (37) (36) (15) (-1-Inositol 1 -phosphate Scheme 12 formamide undergoes intramolecular benzoyloxy displacement of the tosyl ester giving a mixture of the desired optically active 2,3,4,5,6-pentabenzoyl inositol (36) and the isomeric 1,3,4,5,6-meso material (37). Recrystallization affords pure (36) which on phosphorylation with diphenylchlorophosphate, followed by deprotection by hydrogenolysis of the phenylphosphate esters and subsequent basic hydrolysis of the benzoyl esters yields pure (-)-inositol 1-phosphate (15). A number of approaches have succeeded in resolving protected inositol derivatives, allowing the synthesis of the individual enantiomers of inositol 1-phosphate.Conversion of the racemic pentaacetate (38) into its acid oxalate gives (39), Scheme 13, which is amenable to resolution using salt formation with chiral base^.^^.^' Unfortunately to obtain good yields of the resolved materials, salt formation with quinidine is required to obtain one enantiomer and with (-)-x-phenylethylamine to obtain the other enantiomer. Multiple recrystallization of the diastereomeric salts is also necessary. The enantiomeric pentaacetates (40) and (41) obtained on hydrolysis may then be converted into the enantiomers of 36 J. G. Molotkovsky and L. D. Bergelson, Tefraherlron Left.,1971,4791. 37 J. G. Molotkovsky and L. D. Bergelson, Clzem. Phys. Lipids, 1973, 135. Billington I I "OAc AcOO.y- 'OAc OAc OAc OAc (40) (41) 1!.steps0:H203P 0 HO'.11 steps OH (-1 -(15) (+I -(16) Scheme 13 inositol 1 -phosphate (1 5) and (16) by standard methods. Conversion of racemic protected inositols into diastereomeric mixed orthoesters using mannose de-rivatives has been used to resolve a number of useful intermediate^.^^.^^ Treatment of the racemic pentabenzyl derivative (25) with the orthoester of D-mannose gives the diastereomeric mixed orthoesters (42) and (43) (Scheme 14), which may be separated by a mixture of chromatography on alumina and 3x V. I. Shvets, B. A. Klyashchitskii, A. E. Stepanov, and R. P. Evstigneeve, Tetrahedron, 1973,29, 331. 3') A. E. Stepanov. 0.0.Tutorskaya, B. A. Klyashchitskii, V.I. Shvets, and R. P. Eustigneeva, Zh. Obs. Khim., 1972. 42, 709; S. P. Kozlova. I. S. Pekarskaya, B. A. Klyashchitskii. V. I. Shvets, and R. P. Eustigneev, ihid., 702; B. A. Klyashchitskii, V. V. Pimenova, A. I. Bashkatova. E. G. Zhelvakova. S. D. Sokolov, V. I. Shvets, R. P. Evstigneeva, and N. A. Preobrazhenskii, ibid. 1970, 40. 2482. Recent Developments in the Synthesis of myo-lnositol Phosphates BnOAH X -0Et x0QBn + BmQ BnO“ ’0 Bn BnWO ‘OBn 0Bn OBn (42) (43) 1. Separate 2. Hydrolysei 1?R ~0 Bn and0 BnoQ,n = H20R CH3+ BnW “0 Bn Brio'. 0Bn OBn (+I -(44) (-1 -(45) CH CH3 y= & Scheme 14 Billington 0 -Hog-"HOI (48) 0II *,OJ+Ok HO" OH (27) ( 50) Scheme 15 recry~tallization.~~Mild acidic hydrolysis regenerates the pure enantiomers (44) and (45). The racemic alcohol (25) has also been resolved via conversion into the diastereomeric camphanate esters (46) and (47), Scheme 14, via treatment with commercially available (-)-camphanic acid chloride.*' The camphanates were separated by chromatography, and their diastereomeric purity confirmed by HPLC. Single crystal X-ray structure analysis of one of these esters allowed the absolute configuration of the camphanates, and thus of the inositol 1-phosphates derived from them to be assigned.This confirmed the original assignment of structure (16) to the dextrorotatory isomer of inositol 1-phosphate. Phosphoryl- ation of (44)and (45) with diphenylchlorophosphate, followed by transesterifica- tion with the anion of benzyl alcohol and deprotection by hydrogenolysis give the pure enantiomers of inositol 1-phosphate (15) and (16).25 B.Inositol 2-Phosphate.-An optically inactive inositol monophosphate was isolated from the acidic hydrolysis of wheat bran in 1912 and subsequently assigned the structure (27).40,41 It is now clear the penultimate product of the acidic, alkaline, or enzymic hydrolysis of inositol hexakisphosphate (phytic acid) is inositol2-phosphate (27). Acetobacter suboxydans selectively oxidizes myo-inositol at the 2-position giving scyflo-inose (48)40 and this intermediate has been used to prepare (27), Scheme 15.41 Acetylation of (48) followed by hydrogenation of the carbonyl group gives 40 T.Posternak. Hela. Chin?.Arrci, 1941.24, 1045. 41 B. M. Iselin, J. Ant. Cheni. Sor., 1949. 71, 3822. 101 Recent Developments in the Synthesis of myo-Inositol Phosphates (51) 2 steps1 0 HO' BnO#*Y =OBn OH 0Bn (27) (52) Scheme 16 the pentaacetate (49). Phosphorylation with diphenyl chlorophosphate then gives (50) and subsequent deprotection gives inositol2-phosphate (27).41 A more conventional approach 42 uses the previously described selective alkylation of the tetrabenzyl diol (21), Scheme 16. Benzylation of (21) gives predominantly the pentabenzyl alcohol (51). Phosphorylation using di-phenylchlorophosphate under vigorous conditions, followed by transesterification then gives the fully protected compound (52).A single hydrogenolysis then removes all of the protecting groups, giving pure inositol2-phosphate (27). The previously described orthoformate (6) may be selectively dibenzylated at the 4- and 6-positions by taking advantage of the chelation-controlled reactions of mono anions of (6).43Thus, sequential treatment of (6) with 1 eq. of sodium hydride and 1 eq. of benzyl bromide, followed, after alkylation is complete, by a second treatment with 1 eq. of base and alkylating agent, gives (53) in good yield. Phosphorylation of (53) with sodium hydride/tetrabenzylpyrophosphate gives (54) which on deprotection gives inositol 2-phosphate (27), Scheme 17. C. Inositol 4-Phosphate.-Alkaline hydrolysis of brain phospholipids gives a second optically active inositol mon~phosphate,~~ in addition to the dextro- rotatory 1-phosphate, in minor quantities, and this is therefore one enantiomer of inositol4-phosphate (20).42 J. J. Kulagowski and D. C. Billington. unpublished observations. 43 D. C. Billington, R. Baker, J. J. Kulagowski. I. M. Mawer, J. P. Vacca. S. J. de Solms, and J. R. Huff, J. Clinii. Sw.. Pd-iti Truns. I. in the press. 44 C. Grado and C. E. Ballou, J. Bid. Cheni., 1961, 236.54. Billing ton (6) (53) HO" OH (27) Scheme 17 (i) Racemic Material (+)-(20). The majority of the racemic syntheses reported use the biscyclohexylidene acetal (3), Scheme 1 as starting material. Selective protection of the more reactive 3-hydroxyl group in (3) is possible, and the mono benzoate ester (55a),4s mannose orthoester (55~),~~ and benzyl ether (55b)25 have all been prepared and used as intermediates, Scheme 18.Phosphorylation of the selectively protected intermediate (55) at the free 6-hydroxyl gives the corresponding fully protected phosphate (56a-c). As C-4 and C-6 of myo- inositol are enantiotopic, the deprotection of these intermediates (56a-c) using standard methods gives ( )-inositol 4-phosphate, exploiting the symmetry inherent in the parent inositol. A more recent synthesis uses the chelation- controlled phosphorylation of the monoanion of orthoformate (6)to generate the protected phosphate (57), Scheme 19, in a single step.23,43 Deprotection by hydrogenolysis of the dibenzylphosphate esters followed by acidic hydrolysis of the orthoformate group, gives (&)-inositol4-phosphate(20) in excellent yield.(ii) (+) and (-) Inositol 4-phosphate (60 and 61). Attempts to separate the diastereomeric D-mannose orthoesters formed from ( )-(3) [i.e. the diastereomers of (55c), Scheme 181 have not succeeded to date.39 In contrast, treatment of the alcohol (55b) with commercial (-)-camphank acid chloride gives a mixture of diastereomeric camphanate esters (58) and (59), Scheme 20, which are readily separable by crystallization and chromatography.2s Hydrolysis of the camphan- ate esters gives the free enantiomers of (55b), which may then be converted into the enantiomers of inositol 4-phosphate (60) and (61) by the same methods used for the racemic series.25 The absolute configurations of these enantiomers have not been confirmed by physical means to date.45 S. J. Angyal and M.E. Tate, J. Client. Soc., 1961,4122. 103 Recent Developments in the Synthesis of myo-Inositol Phosphates 07'6~10 O--f6H 10 "QOHd' 0'-RoooH Lo LO C6H 10 'GHl0 (c) R = mannose or thoest er 1 2-3 steps C6H10 (20)(2)-Inositol 4 -Phosphate (56) (a) R = Bzl , R'= Ph (b) R =En, R'=Bn (c) R = mannose or t hoester R'= Ph Scheme 18 -____j HO'' D. Inositol5-Phosphate.-A single synthesis of inositol 5-phosphate (65) has been rep~rted,~' starting from the rather inaccessible 2-amino-2-deoxy-neo-inositol (62),46*47Scheme 21. The amino inositol (62) (obtained either by ~ynthesis,~~ or by hydrolysis of the antibiotic hygromycin A 47), protected as its pentaacetate 46 G.R. Allen, J. Am. Clteni. Soc., 1956, 78, 5691. J. B. Patrick, R. P. Williams, C. W. Waller, and B. L. Hutchings, J. Am. Chem.Soc., 1956. 78, 2652; R. L. Mann and D. 0.Woolf, ibid., 1957.79, 120. 104 Billington 7'6" 10 %*lo07c6H 10 4-$-0.-BnOo + 0'"OQOH 0' ''OR RO'. '\ LO LO '6*10 '6"lO 046H 10 (2)-(55b) (58) (59)A R= i2 steps 2 stepsI " AND HOQ H HO".!+'*OP03H2 H,O,PO'' =OH OH OH (60) (611 (+) and (-1 Inositol 4 -phosphate Scheme 20 (63), gives the protected inositol (64) on treatment with nitrous acid. This con- version of NH2 into OH proceeds exclusively with inversion of stereochemistry. Phosphorylation of (64) with diphenylchlorophosphate, followed by deprotection using standard methods gave inositol5-phosphate (65).45 3 Synthesis of Inositol Bisphosphates A.Inositol 1,3-Bisphosphate.-The symmetric nature of the orthoformate (6) has been exploited in the synthesis of inositol 1,3-bi~phosphate.~~ Exhaustive benzyl- ation of (6) gives the fully protected intermediate (66), Scheme 22, which on acidic Recent Developments in the Synthesis of myo-Inositol Phosphates H O 6 -Ac0fJkc -AcOQ, HO'' *'OH0'' Ac Ac 0' I OHNH* NHZ (64)(62) (63) / HO''HoQ:H OPO,H, (65) Inositol 5-phosphate Scheme 21 (70)Inositol 1,3 -bisphosphate Scheme 22 Billington o~c6H10 o-? H'o HoQ0' *'OH ::.$f0 Lo C6H 10 046 10 (31 (4) 2 steps 2 steps H2°3p00 '2'3''Q OH HofJ;03HO" HU' * * OPO,H 2 H,O,PO' .*OH OP03H2 OH OH (71) (73) (f)-Inositol 4,5 -bisphosphate (t)-Inositol 3,4 -bisphosphate (72) (2)-Inositol 1,4 -bisphosphate Scheme 23 hydrolysis gives 2,4,6-tri-O-benzyl inositol (67).As the 1 -and 3-hydroxyl groups in (67) are less sterically hindered than the 5-hydroxyl, phosphorylation with diphenylchlorophosphate gives an 8:2 mixture of the 1,3-(68) and 1,5-(69) bisphosphorylated products. The desired 1,3-isomer (68) may be isolated by crystallization, and deprotection using lithium in liquid ammonia at -78 "C gives inositol 173-bisphosphate (70).23,43 B. Inositol 1,4-, 3,4-, and 4,5-Bisphosphates.-(i) Racemic Syntheses. Phosphoryl-ation of the 3-biscyclohexylidine acetals (2), (3), and (4)of myo-inositol, Scheme 1, leads to syntheses of inositol 45, 1,4-, and 3,4-bisphosphates (71), (72), and (73) re~pectively.~~Scheme 23.The original method used employed diphenylchloro- phosphate as a phosphorylating agent.45 However, the use of a phosphite reagent, followed by oxidation to the phosphate, and deprotection has been reported recently, and is probably more efficient.48 The phosphite method also allows the synthesis of phosphorothioates by oxidation of the intermediate PI" species with Sg in pyridine in place of mCPBA, and has been applied to tetrabenzyl inositols [e.g. (74)] in addition to inositol bi~acetals.~~.~' Scheme 24. 4R M. R. Hamblin, J.S. Flora, and B. V. L. Potter, Biochem. J., 1987, 246, 771. 49 M. R Hamblin, B. V. L. Potter, and R. Gigg, J. Chem. Soc., Chem. Commun., 1987,626. 5" M. R. Hamblin, B. V. L. Potter, and R. Gigg, Biochem. Soc. Trans., 1987, 15,415. Recent Developments in the Synthesis ?f myo-Inositol Phosphates OBn -* HO" OH (74) I H2O PHOQC0.. (CNCH2C H20)2PO' OPO3H2 -(71) 0P(OCH ,CH 2CN l2 II0 Scheme 24 Note the need to number compounds (71) and (72) clockwise in order to obtain the lowest number count (i.e. 43-and 1,4-bisphosphates, not 5,6- and 3,6-bisphos- phates). (ii) Individual Enantiomers. The individual enantiomers of all three of the above bisphosphates have been prepared from the resolved antipodes of the three bisacetal~.~~The acetals were resolved using the orthoesters of D-mannose as seen previously for the enantiomers of inositol 1-phosphate, Scheme 14.38,39 The optically active bisacetals were then phosphorylated as before using di-phen ylchlorophosphate. A more recent approach,s2 uses the conversion of the bisacetal (3) into its diastereomeric biscamphanate esters by treatment with 2 equivalents of (-)-camphanic acid chloride.These diesters may then be separated chromatogra- phically and basic hydrolysis gives the enantiomeric bisacetals (cJ Scheme 20). Phosphorylation and deprotection gave the enantiomers of inositol 1,4-bisphos- phate. 4 Synthesis of Inositol Trisphosphates As previously outlined D-inositol 1,4,5-trisphosphate is now well established as a fundamental intracellular second messenger, directly involved in the mobilization of calcium from cellular stores.The problems inherent in the synthesis of this 51 V. N. Krylova, N. 1. Kobel'kova, G. F. Oleinik, and V. I. Shvets, Zh. Org. Khim., 1980, 16,62. 52 J. P. Vacca, S. J. de Solms, J. R. Huff, D. C. Billington, R. Baker, J. J. Kulagowski, and 1. M. Mawer, manuscript in preparation. Billington type of polyphosphate are highlighted by the fact that the first synthesis of 1,4,5- IP3 was not reported until late 1986. A number of more efficient synthetic approaches have been published following this first disclosure, and isomeric trisphosphates including the 1,3,4-trisphosphate and 2,4,5-trisphosphate have also been prepared.A. Inositol 1,4,5-Trisphosphate.-The first reported synthesis was of the D-( -)-isomer, (-)-(84). Benzylation of the bisacetal (3) gives the fully protected inositol (79, Scheme 25.53 Selective hydrolysis of the less stable trans acetal, followed by allylation of the free hydroxyl groups generated gives (76), which on acidic hydrolysis gives (77). Racemic (77) may be resolved by conversion into its diastereomeric monomenthoxyacetyl derivatives (78) and (79) (reaction occurring only at the less hindered equatorial OH group), which are separable by crystallization and chromatography. Hydrolysis of the desired isomer (78) then gives the optically active diol (80). Selective allylation of the more reactive equatorial OH group in (80) [cf: Scheme 10, (21)-(24)] gives (81), which on benzylation and cleavage of the allyl protecting groups gives the triol (+)-(82).Phosphorylation of (+)-(82) with dianilidochlorophosphate gives the fully protected inositol 1,4,5-trisphosphate (83) which was sequentially deprotected to ( -)-inositol 1,4,5-trisphosphate (84). The phosphorylation/deprotection strategy used gave only low yields of the desired product D-( -)-(84). A more efficient synthesis of (-)-(84) would be possible by using one of the more recently reported phosphorylation/deprotection strategies, e.g. phosphoryl-ation of the alkoxide with tetrabenzylpyroph~sphate,~~’~~or reaction of the triol with N,N-diisopropyl dibenzylphosphoramidite, followed by oxidation and deprotection.26 Both of these strategies have been shown to be generally applic- able.The selective allylation approach used above is based on the original work of Gigg et a1.18 who subsequently reported a modified route to the racemic form of ( )-(84), starting from the bis-isopropylidene acetal (85),54 Scheme 26. Benzylation of (85), followed by acidic hydrolysis of the less stable trans acetal, allylation of the resulting diol, and a second acidic hydrolysis cleaving the cis acetal, gives the diol (86), as before. Treatment of the diol (86) with tributyltin oxide gave the cyclic dibutylstannylene deri~ative,~ which on treatment with allyl bromide gave a high yield of the desired tri-0-ally1 derivative (87). Benzylation of the free hydroxyl group in (87), followed by cleavage of the allyl groups then gives the racemic triol ( k)-(88).54 Reaction of (88) with CIP(OCH2CH2CN)N(CHMe2)2 followed by displacement of the diisopropyl amine with cyanoethanol gave the phosphite (89) (cf: Scheme 24 for the use of these reagents).Oxidation of phosphite (89) to the protected trisphosphate with t-butylhydroperoxide, followed by deprotection gave ( f)-inositol 1,4,5-trisphosphate ( +)-(90).27*28Alternatively s3 S. Ozaki. Y. Watanabe, T. Ogasawara, Y. Kondo, N. Shiotani, H. Nishii, and T. Matsuki, Tetrahedron Lett.. 1986,27, 3157. 54 J. Gigg, R. Gigg, S. Payne, and R. Conant, J. Clieni. Soc., Perkin Trans. 1, 1987,423. 55 S. David and S. Hanessian, Tetrahedron, 1985,41,643. Recent Developments in the Synthesis of myo-Inositol Phosphates $ + ROQ Bn n "OBn BnO'/ n 'O- E I ' -A 2 steps HO" OH (82) 0 P(NHPh1, (-1 -(84) (83)(-1 -Inosito1 1,4,5 -trisphosphate Scheme 25 Billington OH 4 steps 2steps A/O' 'OBn -v (87)2 steps (88) BnOfiOP(OCH2CH2CN l2 "OBn OP(OCHzCH2CN12 2J \ OP03H2 H,O,PO""Q.OH OPO,H, (90) (2)-Inositol 1,4,5 -trisphosphate (91) -Inositol 1,4,5 -trisphosphorothioate Scheme 26 Recent Developnients in the Synthesis of myo-Inositol Phosphates treatment of (89) with s8 in pyridine followed by deprotection gave the trisphosphorothioate (f.)-(91).29 In a related approach, Scheme 27, selective protection of the more reactive hydroxyl group in bisacetal (4) as the benzyl ether gives alcohol (92).56 Conversion of this alcohol into its diastereomeric camphanate esters with (-)-camphanic acid chloride, followed by separation of the diastereomers [(93) + (94)] and hydrolysis of the less stable trans acetal in the desired isomer, gives the optically active diol (-)-(95).Basic hydrolysis of the ester group then provides the optically active triol (+)-(96). This triol may be efficiently phosphorylated using KH/THF/tetrabenzyl pyrophosphate, and deprotection by hydrogenolysis, and acidic hydrolysis of the remaining acetal gives (-)-inositol 1,4,5-trisphos-phate, (-)-(84). The value of tetrabenzyl pyrophosphate as a phosphorylating agent is underlined by the ca. 60% overall yield of (-)-(84) from the triol (96).A less conventional approach 57 uses the cyclopentylidene acetal (97). Scheme 28, obtained from the previously described diol (21) (see Scheme 10) by acetalization and removal of the benzyl ethers using Na/NH3. In a three-step sequence triol (98) is obtained by silylation of (97) with t-butyldimethylsilyl chloride, reaction of the crude product with 9-chloro-2,7-dibromo-9-phenyl-xanthene, and subsequent desilylation, in some 30--40% yield. Phosphitylation of (98) and oxidation to the phosphate gives the fully protected trisphosphate (99). Deprotection of (99) then gives racemic inositol 1,4,5-trisphosphate, (f.)-(90). Repetition of the synthesis using enantiomerically pure diol (21) (obtained from the racemic diol by the method of StepanovS8) gave the optically active inositol 1,4,5-trisphosphate (-)-(84). An interesting and very short synthesis of the racemic trisphosphate ( _+ )-(90) has recently been reported.59 Phosphitylation of the tetrol (100) [obtained by benzoylation of bisacetal (3) and subsequent hydrolysis of both acetals], with 3.3 eq.of dimethyl chlorophosphate, followed by acylation of the crude product and oxidation of the phosphite products to phosphates with H202 gave the protected trisphosphate (101) in ca. 94% yield and 95% purity, Scheme 29. This selective phosphitylation of the equatorial hydroxyl groups thus avoids the need for a selective protection strategy. Deprotection with HBr in acetic acid and ester hydrolysis gave (+)-inositol 1,4,5-trisphosphate ( -&)-(90) of ca.95% purity. B. Inositol 1,3,4-Trisphosphate.-The racemic triol (104) required for the synthesis of inositol 1,3,4-trisphosphate (105) has been prepared by essentially the same strategy as seen for the 1,4,5-trisphosphate intermediates, Scheme 26, and has formed the basis of several synthetic approaches. Bis-allylation of acetal (3), followed by selective hydrolysis of the less stable trans acetal and benzylation of the resulting diol gives (102), Scheme 30. Hydrolysis of the cis acetal in (102) followed by regioselective allylation of the more reactive equatorial hydroxyl s6 J. P.VaCCd, S. J. de Solms, and J. R. Huff, J. Am. Cliem. Soc., 1987, 109, 3478. ”C. B. Reese and J. G. Ward, Terralietlron Leir..1987, 28, 2309. ”A. E. Stepanov, B. A. Klyashchitskii, V. I. Shvets, and R. P. Evstigneeva, Bioorg. Khim., 1976,2, 1627 59 J. L. Meek, F. Davidson, and F. W. Hobbs,J. Am. Chem. Soc., 1988, 110,2317. Billing ton ___I) LO ‘GH1O 1 + HO’*QR OH ‘OBn ** OBn BnO“ (-1 -(95) C6H10 (93) (94) OH 2 steps H20,PO’* OH OP03H2 (+I -(96) (-1 -(84) (-1 -Inositol 1,4,5-trisphosphate Scheme 27 Recent Developments in the Synthesis of myo-lnositol Phosphates 2 steps 3 steps "OBn OBn (97) (98) J n n-C5H8 HZo3p0Q 2 steps HO" \'OPO,H~ OPO,H, (?)-(90) 6P (OR'),'d (2)-Inositol 1,4,5 -trisphosphate (f)-(99) Br R= R'= -CHzCHzCN Ph Br Scheme 28 group gives alcohol (103). Benzylation of (103) followed by cleavage of the ally1 groups gives trio1 (104).This route has been followed with both the biscyclohexyl- idine acetal (3) 6o and the analogous bisisopropylidine acetal.54 Phosphoryl-ation of the alkoxide of (104) with tetrabenzylpyrophosphate (TBPP) gave the fully benzylated trisphosphate 61 which was cleanly deprotected to ( +_)-inositol 1,3,4-trisphosphate ( i-)-( 105). Alternatively 6o phosphitylation with CIP(OCH2CH2CN)2followed by oxidation of the trisphosphite to the protected trisphosphate, and subsequent deprotection also gave pure (+_ )-(105). A second approach involves the use of the bisacetals (2) and (4),62Scheme 31. 6" C. E. Dreef, G. A. van der Marel. and J. H. van Boom, J. R. Nefh.Ciiem. Sac.. 1987,106, 161.61 S. J. de Solms, J. P. Vacca, and J. R. Huff, Tefrahrdran Lett., 1987, 28,4503. 62 S. Ozaki, M. Kohno, H. Nakahira. M. Bunya, and Y. Watanabe, Chem. Letf., 1988,77. 114 Billing ton OH 3 steps HO" OH (100) 'OBzl (CH 0)'8PO'Oy*'OBZl OP(OCH,),Id OP03Hz Scheme 29 By reversing the order of incorporation of the benzyl and p-methoxybenzyl protecting groups, both (2) and (4) may be converted into the racemic diol (107), via selective hydrolysis of the trans acetals in (106a) and (106b) and subsequent protection of the diols. Cleavage of the remaining acetal gives diol (108) which may be resolved via its monomenthyloxyacetyl derivative (cf: resolution of (77) in Scheme 251 to give the required enantiomer of (108). Selective incorporation of a methoxymethyl group at the equatorial 1-position followed by benzylation gives the fully protected enantiomerically pure intermediate (109) which on removal of the PMB and MOM groups gives the optically active triol (1 10).Phosphorylation using TBPP as before then gives D-inositol 1,3,4-trisphosphate (111). A similar strategy using allyl ethers as temporary protecting groups has been used to obtain racemic triol (104) from ketal (2).26 C. Inositol 2,4,5-Trisphosphate.-Diol (1 12) has been prepared from the iso- propylidene acetal corresponding to acetal (3), Scheme 32, by benzylation, hydrolysis of the less stable trans acetal, allylation of the resulting diol, and hydrolysis of the remaining acetal [cf: preparation of (86) from (85), Scheme 26].54 Selective benzylation of the equatorial hydroxyl group in (112) and cleavage of the allyl groups gives (1 13) which, on phosphorylation with NaH/TBPP, followed by hydrogenolysis to remove the benzyl ethers and esters, gives ( +)-inositol 2,4,5-trisphosphate (1 14).61 A formal synthesis of D-inositol 2,4,5-trisphosphate has also been reported 63 in which an intermediate diol corresponding to (112) was resolved via its mono- menthoxy acetate derivatives as outlined previously, Scheme 25.This synthesis has not been successfully concluded to date due to problems with phosphorylation technology. 5 Synthesis of Inositol Tetrakisphosphates. The first reported synthesis of inositol 1,3,4,5-tetrakisphosphate23 takes advantage of the highly specific chelation-con h3 Y.Watanabe. T. Ogasawara. N. Shiotani, and S. Ozaki, Tetruliedron Lett., 1987,28,2607. Recent Developments in the Synthesis of myo-Inositol Phosphates 3 steps (102) 2 stepsI 2 steps H 0' -'OBn OEn (104) (103) \2 steps OH -(105) (21 -Inositol 1,3,4 -trisphosphate Scheme 30 trolled alkylation reactions of orthoformate (6), Scheme 33 (described previously Scheme 2, Scheme 17, and Scheme 19). Selective mono-allylation of (6) gives (115), which on benzylation gives (116). Isomerization of the ally1 group to the enol ether followed by acid hydrolysis of the enol ether and orthoformate groups gives tetrol (1 17). Phosphorylation of (1 17) with TBPP/NaH/imidazole, gave the decabenzyl tetraphosphate (1 18) in high yield, which on hydrogenolysis gave ( & )-inositol 173,4,5-tetrakisphosphate(1 19).An essentially identical synthesis Billing ton 07c6H10 Hloc6c7$oH0-’ OH (2) (4) 07c6H10 Hloc6<oQoBn0--OBn (106a) (106 b) 2 steps stepi07c6H10 pM BO*#z/; ““g‘0Bn PMBO** OBn O8n (1071 (107) 2. Resolution PMBOfJ OH Zsteps+ PMBO~J MOM PMBO‘ “OBn PMBO** “OBn OBn OBn OBn (-1 -(I081 (109) (-1 -(111) (-1-Inositol 1,3,4 -trisphosphate Scheme 31 117 Recent Developments in the Synthesis of myo-Inositol Phosphates OH OH 2 steps Bno*Bn $0 // (113) 2 steps H 2 03PO” OP03H2 (t)-(114) (-1 -Inositol 2,4,5 -trisphosphate Scheme 32 using a benzyloxymethyl ether in place of the ally1 group for temporary protection was completed concurrently and has subsequently appeared.61 Phos- phitylation of (1 17) with N,N-diisopropyl dibenzyl phosphoramidite, followed by oxidation of the tetrakisphosphite to the tetrakisphosphate and subsequent deprotection also gives (1 19) in high yield.26 In a more conventional approach, Scheme 34,64acetal (3) was selectively benzoylated (cf.Scheme 18) using benzoyl imidazole/CsF giving (55a). Benzylation of the free hydroxyl in (55a) using benzyltrichloroacetimidate and a catalyst, followed by selective hydrolysis of the less stable trans acetal, benzoylation of the resulting diol, and subsequent hydrolysis of the remaining cis acetal gave diol (120). Conversion of diol (120) into its diastereomeric monomenthoxy acetates followed by separation of the diastereomers (121) and (122) gave optically pure intermediates. The desired diastereomer (121) was then benzylated to give the fully protected inositol (123), which on hydrolysis of the four ester groups gave the optically pure tetrol (124).Phosphorylation using TBPP followed by hydrogenolysis gave D-inositol 1,3,4,5- tetrakisphosphate (125). Recently the enantiomers of tetrol (117) [i.e. (124) and (130)] have been rep~rted,~’prepared from the monosilyl orthoformate (1 26). Monobenzylation of (126) gives the alcohol (+)-(127), which may be resolved via conversion into the diastereomeric carbamates (1 28) and (1 29), Scheme 35. Separation of the diastereomers is possible only after desilylation, and then benzylation of the free hydroxyl group, followed by hydrolysis of the carbamate and orthoformate groups gives the enantiomeric tetrols (124) and (130).An 64 S. Ozaki, Y. Kondo, H. Nakahira, S. Yarnaoka, and Y. Watanabe, Trtrcri~e~fronLefi., 1987, 28,4691. 65 G. Baudin, B. I. Glanzer, K. S. Swaminathan, and A. Vasella, Hehi. Chim.Acici, 1988,71, 1367. Billington J Bn203PO~ ~OP03Bn2 H O O H Bn,O,PO' *'OBn H0" .'OBn OP03Bn2 OH (118) (t)-(117) OP03H, (f)-(119) (f)-Inositol 1,3,4,5 -tetrakisphosphate Scheme 33 alternative approach 65 involves diesterification of (126)and subsequent desilyl- ation to (131).Treatment of (131) with pig liver esterase gives a high yield of the optically active ester (-)-(132) in 95% e.e.Dibenzylation of (-)-( 132),followed by ester and orthoformate hydrolysis then gives tetrol (130).Taking advantage of the symmetry properties of myo-inositol, (-)-( 132)may also be converted into the enantiomeric tetrol (124).Protection of (-)-(132) as its bistetrahydropyranyl (THP) ether, ester cleavage, and benzylation of the resulting free hydroxyl group gives (133).Hydrolysis of the THP ethers followed by selective benzylation at the less hindered equatorial hydroxyl group gives (1 34), which on orthoformate cleavage gives (124).Phosphitylation, oxidation 65,26 and deprotection of these enantiomeric tetrols then gives ( +)-and (-)-inositol 1,3,4,5-tetrakisphosphate. An alternative approach uses the migration of a benzoyl group from the equatorial position in (100) to the neighbouring axial position to obtain tetrol (135) in modest yield,59 Scheme 36.Phosphitylation of (135)with CIP(OMe)2, followed by oxidation gives the protected tetrakisphosphate (1 36). Acidic hydrolysis of the phosphate esters, followed by basic hydrolysis of the benzoate esters then gives racemic inositol 1,3,4,5-tetrakisphosphate(1 19). Recent Developments in the Synthesis of myo-Inositol Phosphates 07c6H10 07‘6 10 BzlHoeoG 4 steps 0‘ OH LO OH LO OBzl ‘GH1O c6H10 (120)(31 (55a) iROQ Bzl +J BzlOQ R-BzlO“ B d *‘OBtl BzlO’* \‘OBn OBzl 0021 OBzl (123) H203PO OP03H2 2 steps HO‘’ “OBn H203PO0’ $OH OPO3H2 (124) (125) D -Inositol 1,3,4,5 -tetrakisphosphate Scheme 34 6 Conclusions From the above account it is clear that the major problems posed by the synthesis of inositol polyphosphates have been solved.Efficient protection Billington iPL 3 steps\ /OTHP THPO''O@OBn steps H:poBn Scheme 35 Recent Developments in the Synthesis of myo-Inositol Phosphates OH _______) HO'@ HO" OH OH (100) (135) 2 stepsI 2 steps (CH31203POf~ P03(CH3)2 (CH3),0,PO'@ 'OBzl OP03H, OP03(CH3)2 cf, -(119) (136) Scheme 36 strategies, phosphorylation methods and deprotection techniques all exist. The next challenge to chemists in this area will be to design and synthesize molecular mimics with selective actions on specific enzymes in the PI cycle.These compounds hold the promise of establishing the details of this fundamental process, and may lead to improved therapies for a number of disease states.

ISSN:0306-0012    

DOI:10.1039/CS9891800083

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. SOC.Rev., 1989,18,123-151 Sulphonyl Transfer Reactions By Isobel M. Gordon and H. Maskill CHEMISTRY DEPARTMENT, UNIVERSITY OF STIRLING, STIRLING FK9 4LA Marie-Franqoise Ruasse INSTITUT DE TOPOLOGIE ET DE DYNAMIQUE DES SYSTEMES DE L’UNIVERSITE PARIS VII (CNRS). RUE GUY DE LA BROSSE, 75005, PARIS 1 Introduction Although organic sulphur compounds have a long history and many important biological compounds have sulphur-containing functional groups, their study has remained largely peripheral to main-stream organic chemistry.’ Sulphur exists in a number of oxidation states in organic compounds and, within each, may have different coordination numbers. We shall be concerned in this review with substitution reactions in solution of the particular type of sulphur(v1) compounds shown in equation 1 where X-and Y-are nucleofuge and nucleophile, re- spectively.0 0 II II Ar(R)-S-X + Y-(HY) =Ar(R)-S-Y 8 + X-(HX) (1)8 In this overall transformation, an arenesulphonyl (or alkanesulphonyl) group may be regarded, conceptually though not necessarily mechanistically, as under- going transfer as a Lewis acid (ArS0; or RS02f) from one Lewis base (X-) to another (Y-). Reactions of equation 1 are used as methods of enzyme modification,’ and, given the recent and continuing redirection of chemical investigations towards biological systems, it is not surprising that we are currently witnessing a significant development of interest in their mechanisms. We shall consider in this review how such reactions are currently believed to take place, and how their mechanisms relate to those of better known reactions of other types of compounds, e.g.SN~and SN2 substitutions of alkyl halides and arenesulphonates (Scheme 1) and substitution reactions by addition-elimination of carboxylic acid derivatives (Scheme 2). Nucleophilic substitutions at saturated carbon were among the first to be investigated in mechanistic detail and are still discussed using the terminology described by Ingold.2 Since Ingold’s seminal work, there have been innumerable ‘‘Organic Sulphur Compounds’, Pergamon Press, Vol. 1, ed. N. Kharasch, 1961; Vol. 2, ed. N. Kharasch and C. Y. Meyers, 1966; ‘Organic Chemistry of Sulfur’, ed. S. Oae, Plenum Press, New York and London.1977; ‘The Enzymes’, 3rd Edn. ed. P. D. Boyer, Academic Press, New York, 1970, Vol. 1, Ch. 2, by E. Shaw; E. Ciuffarin and A. Fava, Progr. Phys. Org. Chem., 1968,6,81. C. K. Ingold, ‘Structure and Mechanism in Organic Chemistry’, Cornell University Press, 2nd Edn., 1969. Sulphonyl Transfer Reactions I I Y-C-+ x-Y-C-+ x-I I HJY-t inter mediate TS R-C8 /OH R-c //O + HX +R-C YH a ‘x A “x ‘vY X = CI, OR’, 02CR’ YH = H20, R”NH2, R”OH Scheme 2 investigations directed at elucidating the finer details of both sN1 and SN~ extremes, and also at the more difficult problem of characterizing those borderline reactions which appear to show some features of both SNl and sN2 mechanisms. These are now recognized as reactions in which a postulated intermediate in the sN1 mechanism is so reactive, i.e.short-lived, that there is doubt as to whether it really exists at all.3 The most obvious comparisons for sulphonyl transfer, however, are substitution reactions of carboxylic acid derivatives, i.e. acyl transfer, Scheme 2. This is a process of enormous importance and interest since it occurs in so many biological reactions catalysed by enzymes. Consequently, it has attracted attention over many years and is reasonably well under~tood.~The initial nucleophilic attack to give the tetrahedral intermediate may be reversible, as also may be the departure of the nucleofuge in the final step, depending upon the ’W. P. Jencks. Chem. SOC.Rev., 1981. 10,345. M. L. Bender, Chem.Reti., 1960, 60, 53; W. P. Jencks, ‘Catalysis in Chemistry and Enzymology’, McGraw-Hill, New York, 1969; C. Walsh, ‘Enzymatic Reaction Mechanisms’, Freeman, San Francisco, 1979. 124 Gordon, Maskill, and Ruasse ’0 *OH I //”H20 + Ph-c// Ph-C-OH Ph-C + H$ ‘OE t I ‘OE t OEt PhCb2H + EtOH PhC02H + EtOH Scheme 3 particular reaction. Also, either step may be catalysed by acids or bases, and either may be rate determining, again depending upon the natures of the reactants and the reaction conditions. In the case of the hydrolysis of esters, with either acid or base catalysis, the reversible formation of an intermediate was established by demonstrating exchange of isotopically labelled oxygen between recovered starting material and the water of the aqueous medium, Scheme 3.5 Some stable analogues of these tetrahedral intermediates have been isolated and others have been observed spectroscopically in chemical reactions.’ If the nature of the acyl transfer is such that there is no free-energy barrier separating the product of the initial nucleophilic attack from the product of the departure of the nucleofuge, then no intermediate exists and the mechanism has become an enforced concerted (SN2-type) displacement, e.g.the base-catalysed hydrazinolysis of acetylimidazolium cation in Scheme But whereas the 4.398 transition state of the SN2 reaction of alkyl halides in Scheme 1 involves a penta- coordinated carbon atom and is trigonal bipyramidal,’ that in the acyl transfer of Scheme 4 has a tetra-coordinated central carbon atom and a (distorted) tetrahedral structure.A non-enforced concerted mechanism for acyl transfer has also been identified by Williams and his colleagues lo in the transfer of the acetyl group between phenolate nucleophiles. In these reactions, the transition state still involves a tetra-coordinated central carbon, but its structure is less ‘intermediate- like’ than in the enforced concerted process. M. L. Bender, J. Am. Chrm. Soc., 1951, 73, 1626; M. L. Bender and R. J. Thomas, ibid.,1961, 83,4189; M. L. Bender, H. Matsui, R. J. Thomas, and S. W. Tobey, &id., 1961. 83, 4193. G. A. Rogers and T. C. Bruice, J. Am. Chem. SOC.,1974, 96, 2481; F. Khouri and M. K. Kaloustian, ihid.,1979,101, 2249; J.Hine, D. Ricard, and R. Perz, J. Org. Chem., 1973.38, 110.’B. Capon, A. K. Ghosh, and D. M. A. Grieve, Arc. Chem. Res., 1981, 14, 306; B. Capon, M. I. Dosunmu, and M. de N. de Matos Sanchez, Adt.. Phys. Org. Chem., 1985,21,31. M. I. Page and W. P. Jencks, J. Am. Chem. Soc., 1972,94,8828. J. Chandrasekhar, S. F. Smith, and W. L. Jorgensen, J. Am. Chem. Soc., 1985,107, 154. lo S. Ba-Saif, A. K. Luthra, and A. Williams, J. Am. Chem. Soc., 1987,109,6362. Sulphonyl Transfer Reactions 0 IIB-+ H-NN: /H + C /\CH3 ImH* 0 .. II BH + ImH+ /""-"\NH2 CH3 Scheme 4 H20 H20Ar-C -2 Ar-CZO + F-+ Ar-CO2H + H30*F-'F Ar = p-Me2N-C6HL Scheme 5 'f-+ Ar-S02-X ( k' . \r-l<f% Ar-SOZ-Y + X-k-1 (21 Scheme 6 Unimolecular mechanisms of acyl transfer have also been identified both in the gas phase l1 and in solution, e.g.Scheme 5.12 2 Nucleophilic Substitution at Four-coordinated Sulphur(v1) The stepwise mechanisms corresponding to SN1 and addition4imination (SAN) processes (equation 2 and Scheme 6 respectively) can be regarded as limiting cases for the sulphonyl transfer from X to Y in equation 1. Between these stepwise extremes, there is the concerted one-step SN2alternative (Scheme 7). These three conceptually distinct routes for the overall transformation of equation 1 are represented in the free-energy reaction map of Figure 1 from which contours have been omitted for ~larity.~.' 3714 J. K. Kim and M. C. Caserio, J. Am. Chem. Soc., 1981,103,2124.B. D. Song and W. P. Jencks, J. Am. Chem. SOC.,1987,109,3160. l3 R. A. More OFerrall, J. Chem. SOC.(B),1970,274. l4 H. Maskill, 'The Physical Basis of Organic Chemistry', Oxford University Press, 1985. Gordon, Muskill, and Ruusse 0 ( 51 Scheme 7 ArI /ArY-s+ x-y-'0 "0'. 0 n-n0 < 0lan Ar -Ar Y-S bondY-\,s-x formation t&X) I '0 0)' Figure 1 Free-energy reaction mapfor the SN1,SN2, and SAN mechanismsfor the sulphonyl transfer reaction of equation 1 Y -(HY) ArS02Y + X-(HX) (2)ArSOzX &ArSOt X- ---+ k-1 k2 A. The SNlMechanism.-There appeared to be the classic evidence in the earlier literature for an SN1 reaction of compounds such as N,N-dimethylsulphamoyl chloride in equation 3, i.e.product formation but no rate effect upon solvolysis caused by added nucleophiles,' but this was subsequently shown to be in error. These are SN2 reactions.* Me2NSO2C1+ 2Hz0 3Me2NS03H+ H30+ + C1-(3) H. K. Hal1,J. Am. Chem. SOC.,1956,78, 1450. l6 0.Rogne,J. Chem. SOC.(B), 1969,663. Sulphonvl Transfer Reactions sozcI (1) a. X = 2,4,6-trimethyl b, X = 4-methoxy c, X = 24-dimethoxy In principle, and on the basis of results for aroyl halides,12 there ought to be an sN1 reaction mode available to arenesulphanyl chlorides (1) if substituents in the arene ring are sufficiently electron-supplying. There have in fact been claims that the 2,4,6-trimethyl compound (la) reacts via an SNl me~hanism,'~.'~but here also the evidence has been di~puted.'~,'~ Tonnet and Hambly l8 predict, however, that, whereas the 4-methoxy compound (lb) does not undergo hydrolysis by an SNl mechanism, the 2,4-dimethoxy compound (lc) should.B. The SAN Mechanism.-The credibility of the SAN mechanism, Scheme 6, for the reactions of equation 1, like that for the SN1, has waxed and waned over the years. The intermediate in this route, (2) if it exists, will be trigonal bipyramidal, and stable analogues are known.20 Its electronic structure could include up to twelve electrons in the valence shell of the sulphur atom through the use of d- orbitals.21 In one limiting case of this mechanism, the formation of the intermediate is rate determining (k2 % k-1) and, at the other, the rate-determining step is its decomposition (k-1 % k2).14 Isotopic labelling studies comparable with those carried out by Bender in the acyl transfer reactions indicated that no exchange takes place between starting material and the aqueous solvent in the hydrolysis of phenyl tosylate.22 However, whilst a positive result would have implicated a mechanism with a reversibly formed intermediate, the negative result does not necessarily disprove such a mechanism.The SAN mechanism involves an intermediate (2) in which nucleo- phile and nucleofuge occupy apical positions. Consequently, there must be pseudo-rotation within this reversibly formed intermediate for there to be interchange of apical and equatorial ligands and hence isotopic exchange, Scheme 8. In contrast to the otherwise similar phosphorus compounds, it appears "R.V. Vizgert, Russ. Chem. Rev., 1963, 32, 1; E. Tornmila and P. Hirsjarvi, Actu Chem. Srunri., 1951, 5, 659. M. L. Tonnet and A. N. Hambly, Aust. J. Chem., 1971,24, 703. l9 0.Rogne, J. Chem. SOC.(B), 1968, 1294; 1970, 1056. 2o E. F. Perozzi, J. C. Martin, and I. C. Paul, J. Am. Chem. Soc., 1974,96,6735. 21 F. P. Ballistreri, A. Cantone, E. Maccarone, G. A. Tomaselli, and M. Tripolone, J. Chem. Soc., Perkin Trons. 2, 1981,438. 22 C. A. Bunton and C. F. Frei, J. Chem. SOC.,1951, 1872; S. Oae, T. Fukurnoto, and R. Kiritani, Bull. Chem. SOC.Jpn., 1963,36,346. Gordon, Maskill, and Ruasse Ar--02-OPh + *OH-Ar-SO>-OPh + OH-It It r *OH OH proton transfer I/O-\ Ar-S and pseudo -rotation1Ar-y\o I ,O-i OPh OPh I J ArSC$-+ PhOH (Ar = p-CH3C6H4, 0) = ”01 Scheme 8 that sulphonyl groups do not allow easy pse~do-rotation,~’,~~ so even if the SAN mechanism does operate, isotopic exchange would not be observed.When the rates of reaction of benzenesulphonyl halides with three nucleophiles (aniline, n-butylamine, and hydroxide) were studied in aqueous acetonitrile, the relative rates for chloride, bromide, and iodide nucleofuges were found to be almost the same.25 However, compared with these three, fluoride as a leaving group varied according to the nature of the nucleophile, e.g. kn/kF ratios for hydroxide, n-butylamine, and aniline were 4.6, 4.2 x lo’, and 1.65 x lo5, respectively. These leaving group mobilities were explained using the SAN mechanism of Scheme 6.When X = C1, Br, or I, the attack of the nucleophile, kl, is rate determining (k2 $ kl)and the overall rate is not much influenced by the S-Hal bond strength. The mechanism is altered by changing the leaving group to fluoride whereupon k2, the decomposition of the intermediate, becomes rate limiting (kPl % k2). The substantial positive Hammett parameter, p = 2.79, from the alkaline hydrolysis of three para-substituted benzenesulphonyl fluorides in aqueous dioxan was presented by the same Italian group as further evidence of this mechanism.26 The SANmechanism was also favoured, by another Italian group, from a study of the reactions of thiophen-2-sulphonyl halides (3) in methan01.~’ The sensitivity 23 R. Tang and K.Mislow, J. Am. Chem. SOC.,1969,91,5644. 24 E. T. Kaiser, AN. Chem. Res., 1970,3, 145. 25 E. Ciuffarin, L. Senatore, and M. Isola, J. Chem. Soc., Perkin Trans. 2, 1972,468. 26 E. Ciuffarin and L. Senatore, Tetrahedron Left., 1974, 1635. 2’ E. Maccarone, G. Musumarra, and G. A. Tomaselli, J. Org. Chem., 1974, 39, 3286; A. Arcoria, E. Maccarone, G. Musumarra, and G. A. Tomaselli, ibid., 1973,38, 2457. Sulphonyl Transfer Reactions (3) X = F,CI, Br of these compounds towards a range of substituted anilines was claimed to be similar to those of benzenesulphonyl halides and the authors concluded that an addition-elimination mechanism operates in both cases. This group also investi- gated the reaction kinetics of the parent compounds (3) plus three 5-substituted derivatives with anilines in a wide range of pure solvents (protic, aprotic, and deuteriated) and in mixed solvents at 25 0C.28They found that two single- parameter correlations (using the dielectric constant of the solvent) best described the solvent effects: one for the protic solvents with a positive slope, and another for aprotic solvents with a negative slope.They discussed their results in terms of the mechanism of Scheme 6 with rate-determining bond making (k,) in protic solvents, and bond breaking (k2)in aprotic ones. However, later publications from Italy reporting the reactivity of substituted derivatives of (3) with nucleophiles in water and methanol-a~etonitrile,~’and the kinetics of hydrolysis, methanolysis, and ethanolysis of a range of alkane-, alkene-, arene-, and heteroarene-sulphonyl halides ’’ were interpreted in terms of sN2 mechanisms with ‘tight’ or ‘loose’ transition states.These mechanisms could become SNl-like or SANmechanisms according to the nature of the nucleophile, ring substituent, leaving group, and solvent. Such a description is much more in accord with a wide range of other evidence described in the next section. (4)(R = menthyl) Scheme 9 C. The &2 Mechanism.-In 1969 the isotopically labelled chiral sulphonate (4) was reported to react with p-tolyl Grignard reagent to give the chiral sulphone with inversion of configuration at sulphur, Scheme 9.30 This and other stereochemical evidence were entirely consistent with the bimolecular displace- ment reaction of Scheme 7 that had been developed from earlier kinetics3’ and 28 A.Arcoria, V. Librando, E. Maccarone, G. Musurnarra, and G. A. Tornaselli, Tefmhedron,1977, 33, 105; A. Arcoria, F. P. Ballistreri, and G. A. Tomaselli, ihid., 1978,34, 2545. 29 A. Arcoria, F. P. Ballistreri, G. Musurnarra, and G. A. Tornaselli, J. Chem. SOC.,Perkin Trans 2, 1981, 221. ’O M. A. Sabol and K. K. Andersen. J. Am. Chem. Soc., 1969,91,3603. 31 C. G. Swain and C. B. Scott, J. Am. Chem. SOC.,1953, 75, 246; R. B. Scott and R. E. Lutz. J. Org. Chem., 1954, 19, 830; R. B. Scott and H. L. McLeod, hid., 1956, 21, 388. Gordon, Maskill, and Ruasse the (negative) isotope scrambling res~lts.~~,~~ During the sixties, Hambly and co- workers in Australia published an important series of papers on this subject.' t~~~ From their studies of the rates of solvolysis of alkane- and arene-sulphonyl halides in solvents of varying polarity, they concluded that in all cases except, perhaps, one, an SN2 mechanism was taking place in which there are continuous changes in the structure of the transition state according to the nature of the substrate and the polarity of the solvent.When the hydrolysis of substituted benzenesulphonyl chlorides in aqueous dioxan containing mole fractions of water above 0.9 was investigated in a Hammett-type study, substituents were found to have effects opposite from those that they had had in the media of lower polarity, i.e.the p-value changed sign.18 A similar phenomenon was subsequently reported by Tomaselli and his colleagues.21 Hambly's interpretation, based also on AC and AVi results, was that under the better ionizing conditions the S-Cl bond stretching in the transition state is appreciably ahead of the S-0 bond formation. Consequently, the reaction is facilitated by electron-releasing substituents which stabilize the developing positive charge on sulphur and the Hammett p is negative. In the less ionizing conditions, S-0 bond formation is ahead of S-Cl bond rupture, and a positive p indicates the ability of electron-withdrawing groups to stabilize the increased electron density on sulphur in the transition state. [For 2,4,6-trimethylbenzenesulphonylchloride, (la),' and presumably other compounds with better electron-releasing substituents in the benzene ring, an SN1 mechanism may intervene in the more ionizing media as had been proposed earlier,' 'see Section 2A above.] About this time, Rogne in Norway was publishing his important work on sulphonyl chlorides.He observed that benzenesulphonyl chloride underwent hydrolysis in neutral and alkaline conditions, and faster direct nucleophilic substitutions with aniline, azide, thiosulphate, and fluoride to give stable products.' As no reaction was detected with the softer thiocyanate, bromide, and iodide, polarizability is less important in determining nucleophilic reactivity towards the hard sulphonyl electrophilic centre. With pyridine, nitrite, and acetate, the kinetics, stoicheiometry and solvent deuterium kinetic isotope effects of the reaction with arenesulphonyl chlorides were consistent with nucleophile catalysis of the hydrolytic process.It was established from the study of the rates of hydrolysis of substituted benzenesulphonyl chlorides with substituted pyridines that this catalytic reaction occurred by an initial rate-determining formation of the unstable sul-phonylpyridinium intermediate and that the rate was increased by electron- donating substituents in the pyridine and electron-attracting substituents in the benzene ring.33 The substituent effects were correlated by Bronsted and Hammett equations and the Pnucand p parameters were found to be interrelated and sensitive to the reactivity of the system.32 F. E. Jenkins and A. N. Hambly, Aust. J. Chem., 1961, 14, 190,205; R. Foon and A. N. Hambly, ihid., 1962, 15, 668, 684; M. L. Tonnet and A. N. Hambly, ibid., 1970, 23, 2427; R. Foon and A. N. Hambly, ihid., 1971, 24, 713. 33 0.Rogne, J. Chem. SOC.(B),1970,727; J. Chem. Soc., Perkin Trans.2, 1972,489. Sulphony 1Transfer React ions A similar study of the rates of reactions of substituted benzenesulphonyl chlorides with a series of substituted anilines in methanol yielded very similar results.34 Again, the values of the Pnuc and p parameters were interpreted as indicating that the reaction was a direct displacement mechanism with a tighter or looser transition state depending on the substituents present. The retarding effect of one or two ortho-methyl substituents in the aniline was found to decrease with increasing electron-donating ability of the substituents in the ring of the benzenesulphonyl chlorideY3’ consistent with the notion that the transition state became looser when these types of substituents were present in the sulphonyl chloride.Rogne obtained evidence in agreement with his view that substitution at sulphonyl sulphur was direct and concerted when he compared the rates and activation parameters for the reaction of imidazole with substituted ben-zenesulphonyl chlorides in protic and aprotic media.36 Both the enthalpy and entropy of activation decreased in going from protic (methanol) to aprotic (acetonitrile) media, but p (and therefore the electronic characteristics of the reaction centre) remained unchanged. To separate the solvent effect on AHS into initial and transition state components, the heats of solution of the reactants in the two solvents were measured.It was shown that transfer of the activated com- plex from methanol to acetonitrile was substantially exothermic. This result, it was claimed, indicated an activated complex in a concerted process having a charge distribution as shown in (5) in Scheme 7 rather than a more polar one such as those immediately leading to or from the intermediate in the alternative SAN mechanism. Rogne supported his interpretation when a much less exothermic or actually endothermic enthalpy of transfer from protic to aprotic media was established for the activated complex in the reaction of imidazole with benzoyl chloride.37 This reaction was known to proceed uia a polar tetrahedral inter- mediate.In 1983 Lee and Koo~~published an investigation of the reactions of substituted benzenesulphonyl chlorides with substituted anilines in a range of methanol-acetonitrile mixtures. Their approach was similar to Hambly’s and Rogne’s in the 1960s and ’70s. The reactions of benzenesulphonyl chlorides with a series of substituted anilines yielded linear Hammett plots (and consequently linear Bronsted plots) with large negative slopes (p,,, = -2.0--2.9). The reactions of anilines with the series of substituted benzenesulphonyl chlorides yielded positive Hammett parameters which were numerically smaller (p~= 0.5 + 1.0).This led the authors to conclude that more negative charge is transferred from the nucleophile than is developed on the sulphur in the formation of the activated complex and, therefore, that there must be a partial transfer of electron density onto the nucleofuge which is beginning to depart in 34 0.Rogne, J. Chem. Soc. (B),1971, 1855. 35 0.Rogne, J. Chem. Soc., Perkin Trans. 2, 1972, 472. See also 0.Rogne, J. Chem. Sor.. (B), 1971. 1334, and J. F. Bunnett and J. Y. Bassett, J. Org. Chem., 1962, 21, 2345. 36 0.Rogne, J. Chem. SOC.,Perkin Trans. 2, 1973,823 & 1760. 37 0.Rogne, J. Chem. SOC..Perkin Trans. 2, 1975, 1486. ’* I. Lee and 1. S. Koo, Tetrahedron, 1983,39, 1803.132 Gordon, Maskill, and Ruasse xoz Ar --SO2 -OAr' Ar-S02-OX + Ar'O-Scheme 10 the activated complex. Again, an addition-elimination mechanism was ruled out. Lee and Koo went on to increase the solvent ionizing power to its maximum (8&90% methanol v/v) where the leaving group ability of the chloride should be at its greatest. Instead of finding, as they expected, less N-S bond formation in the transition state (because less nucleophilic assistance would be required) they found pnuc(and hence Pnuc) and ps exhibited maximal behaviour indicating an activated complex with a much shorter N-S bond and slightly longer S-Cl bond. It appears, therefore, that an increased extent of S-Cl bond cleavage in the more ionizing solvent is actually compensated by increased N-S bond formation. Similar work but using a series of oxygen nucleophiles with benzenesulphonyl chloride was reported in 1980 by Banjoko and Ok~uiwe.~~Rates with substituted benzoate anions were slower than with anilines but, in other respects, the results were similar and consistent with concerted bimolecular substitutions at sulphur.Evidence that intermolecular sulphonyl transfer between two oxygen bases takes place by a concerted mechanism was provided by Williams' study of the reactions of aryl arenesulphonate esters with oxyanions, Scheme With p-nitrophenolate (pKa p-N02C6H40H = 7.15) as nucleofuge and nucleophiles ranging from acetate (pK, CH3C02H = 4.76) to trifluoroethoxide (pK, CF3CH20H = 12.4), a linear Bronsted-type plot was obtained with no dis-continuity at PKa = 7.15.There was, therefore, no evidence of the stepwise mechanism also shown in Scheme 10 in which the rate-determining step changes as nucleophile and nucleofuge invert their relative base strengths. j9 0.Banjoko and R. Okwuiwe, J. Org. Chem., 1980,454966. 40 P. D.'Rozario, R. L. Smyth, and A. Williams, J. Am. Chem. Soc., 1984,106,5027. Sulphonyl Transfer Reactions 0-0- PhCH2- +/N\\ (Y' = CF~CHZO-; HY = imidazole) N-0 -S02C6HqC H3 (6) Path\ [PhCHi N20 OTs-] + Solvolysis products Scheme 11 Analogous results were reported from France by Monjoint and Ruasse41 for the aminolysis of p-toluenesulphonyl imidazole and corresponding imidazolium cations, in which sulphonyl migrates from one nitrogen to another. As the base strength of the nucleophile ranged over 10 pK, units from below to above that of the nucleofuge, linear Bronsted plots were obtained with very similar p parameters (-9.5) in all cases.Furthermore, the effect of the nucleofuge as expressed by the ljlg is virtually independent of the nature of the nucleophile. A more obscure leaving group, N-benzylazoxy anion, is displaced in the substitution reaction at sulphur shown in Path A of compound (6) in Scheme 11.42 This reaction in aqueous trifluoroethanol takes place only with harder nucleophiles such as the conjugate base of the solvent, CFJCH~O-, and imidazole. Softer nucleophiles, e.g. thiocyanate, simply compete with the solvent as a trap for the electrophilic intermediates generated in the alternative reaction mode involving heterolysis of (6) in the opposite sense, Path B.The reaction of compound (7) in Scheme 12 in aqueous solution with a specific base-catalysis rate law was interpreted earlier in terms of a rate-determining reaction of the deprotonated substrate. This has had to be rejected, however, since in "0 enriched water, the "0 turns up in the tosylate anion which does not exchange oxygen with the solvent under the conditions of the reaction.43 Furthermore, "0 was not incorporated into recovered starting material. Nucleophilic attack by OH-at the sulphonyl group of (7) now seems the more probable reaction path, as shown in Scheme 12. There remains the intriguing possibility that fragmentation of the conrplex anionic nucleofuge (8) is concerted with its departure in this SN2 process (or, less likely, the SAN alternative).41 P. Montjoint and M.-F. Ruasse, Tetruhrdron Lett., 1984, 25, 3183; Buff.SOC.Chim. Fr., 1988, 356. 42 H. Maskill, J. Chem. Soc., Chem. Commun., 1986, 1433; H. Maskill and W. P. Jencks, J. Am. Chem. Soc., 1987, 109,2062. 43 H. A. J. Holterman and J. B. F. N. Engberts, J. Org. Chem., 1977.42, 2792. 134 Gordon, Maskill, and Ruasse r. d. s. ArS02CHZO-SO2A; & ArSOzCHO-SO2Ar' Further reaction + H+ (7) 1-H+ (Ar' = p-CH3CcHh)1 ArSOy + CH2O Scheme 12 3 Nucleophilic Substitution at the Sulphonyl Groups of a-Disulphones and Sulphonic Anhydrides From the study of the hydrolysis of diary1 a-disulphones (9) and arylsulphinyl aryl sulphones (10) in acidic aqueous dioxane, equations 4 and 5, Kice and his co-workers 44,45 have been able to make a quantitative comparison of the influence of reaction variables on nucleophilic substitution at sulphonyl and sulphinyl sulphur.ArS02S02Ar + H20* ArS03H + ArS02H (4) (9) ArSOS02Ar + H20 2 ArS02H (5) (10) Despite the fact that nucleophilic attack at sulphinyl was found to be lo4 faster than at sulphonyl, the reactions were otherwise very similar. They have virtually the same dependence on aryl group structure [p = 3.5 for hydrolysis of (9) and J. L. Kice, Adv. Phys. Org. Clzrm., 1980, 17, 65. 45 J. L. Kice and G.J. Kasperek, J. Anz. Cizem.SOC.,1969, 91, 5510 J. L. Kice, Inl. J. &/fur Ciiem., 1971, 6,3. Sulphonyl Transfer Reactions 0 0II II Ar-S-0-S-Ar II II 0 0 (111 (121 (13) 0 0 0 It It fAr-S-0-S R Ar I1 II 0 0 (15) (161 3.4 for (lo)], and both reactions have very large negative entropies of activation (-158 and -155 JK-' mol-') and substantial solvent deuterium kinetic isotope effects (kH/kD= 2.3 and 2.7). The authors proposed SN2-like mechanisms involving intramolecular partial rate-determining proton transfers to the depart- ing ArS0; group for both (9) and (10) via activated complexes (1 1) and (12). As compounds (9) and (10) differ only in the oxidation state of the sulphur, it was possible from these reactions to make direct comparisons of their sus-ceptibilities towards different nucleophiles.For sulphonyl sulphur 46 the order is F-% AcO-% C1-> Br-> H20, whereas for attack at ~ulphinyl~~ it is Br- > C1-> AcO-> F-B H20. Hydrolysis of sulphonic anhydrides (1 3) 48*49 and (14)49 have also been investigated in aqueous acetone or aqueous dioxane and the results link those for the K-disulphones 44--46 to those from other arenesulphonyl systems discussed above. For the hydrolysis of (13) in aqueous acetone, p = 2.5 compared with p = 3.5 for (9) in aqueous dio~ane.~~*~~ The deuterium solvent kinetic isotope effect for (13), kH/kD-1.2, is also much smaller than the value of 2.3 for (9),45 a result which must rule out proton transfer in the rate-determining step for the reactions of (13). Evidently, when the leaving group is changed from ArS02 to the much less basic ArSO;, no proton transfer is necessary for departure of the nucleofuge, and a transition state such as (15) was proposed for the hydrolysis of (13). In none of these reactions was there any evidence which required other than a concerted SN2 mechanism.Not surprisingly, mixed anhydrides such as (1 6) 46 J. L. Kice, G. J. Kasperek, and D. Patterson, J. Am. Clzem. SIC.,1969, 91, 5516: J. L. Kice and E. Legan, ibid., 1973,95, 3912. 47 J. L. Kice and G. Guaraldi, J. Am. Chem. SOC.,1968,90,4076. 48 N. H. Kristensen, Acfa Chem. Srcmd., 1966,20, 1955; 1967,21, 899. 49 R. M. Laird and M. J. Spence, J. Cltem. SOC.(B), 1971, 1434. Gordon, Maskill, and Ruasse a sulphene Scheme 13 react via nucleophilic attack at the carbon electrophilic centre with the arenesulphonate anion as a very good leaving gr~up.~~,~’ 4 Sulphonyl Transfer via Sulphene Intermediates So far, we have considered the overall reactions of equation 1 in comparison with SN1and SN2 reactions of alkyl halides and arenesulphonates, and SANreactions of carboxylic acid derivatives. However, a proton may be abstracted from a carbonium ion in the El process which often accompanies the SN~,and from the P-carbon of an alkyl halide or arenesulphonate in either an Elc~or E2 reaction.Furthermore, an &-proton of an acyl chloride or anhydride may also be abstracted in E2 or ElcBprocesses. It is not surprising, therefore, that alkanesul- phony1 halides and related compounds also undergo elimination reactions, Scheme 13.51 But in these reactions, the products, sulphenes (17), are not stable; they are themselves reactive electrophiles so they intervene in stepwise elimination-addition reactions which may also be seen as overall sulphonyl transfer processes and hence part of the broader topic covered in this review.The intermediate sulphene (17) with general formula R’R2C=S02 may be regarded as the sulphonyl analogue of a ketene or as a derivative of sulphur trioxide formally obtained by the replacement of one oxygen by a substituted methylene group, R’R2C. In 1964 King and Durst put forward evidence for the intermediacy of a sulphene in the reaction of phenylmethanesulphonyl chloride with triethylamine in deuterated isopropanol, with the observation that monodeuterated isopropyl phenylmethanesulphonate accounted for 90% of the yield.52 The formation of such a product requires deuterium to be incorporated in an irreversible process.Further evidence was produced by Truce and his colleagues53 when they reported in the same year that methanesulphonyl chloride reacts immediately with triethylamine in methanol but not with weaker bases. Moreover, with [O-2HJ-methanol, singly labelled 2HCH2S03CH3 was obtained. An elimination- addition mechanism was implicated. Investigations of these types of mechanisms with sulphonyl species were carried out mainly by two research groups led by ”R. M. Laird and M.J. Spence, J. Chem. Soc. (B),1970,388; 1971,454. ” G. Opitz, Angew. Chem., Int. Ed. Engl.. 1967, 6, 107; T. J. Wallace, Q. Rm Chem. Soc., 1966, 20. 67; J. F. King. ACC.Chem. Rex. 1975.8, 10. 52 J. F. King and T. Durst, J. Am. Chem. Soc., 1964,86,287. s3 W. E. Truce, R. W. Campbell, and J. R. Norell, J. Am. Chem. Soc., 1964,86, 288. Sulphonyl Transfer Reactions B + ArCHZSOzOAr' -2 Ar CHS020Ar' kl + B H+ (18) k-1 Ar tHS020Ar' k2 -A r'0- [Arc= SO*] ROH+ ArCH2S020R Scheme 14 -51 I 2 4 6 8 10 PKa ArOH Figure 2 Plot of log(k,H//M-' s-') vs. pK, of the conjugate acid of the substitutedphenoxide leaving group in the OH --induced hydrolysis of uryl phenylmethanesulphonates,PhCH2S020Ar,25 OC, aqueous solution. Taken from reference 55 King in Ontario and Williams in Kent.Their work was complementary and both came to the conclusion that, while the second-order kinetics of the de-hydrohalogenation of alkanesulphonyl chlorides with tertiary amines (Scheme 13) were consistent with a concerted E2 elimination mechanism, sulphene formation from the reactions of aryl arylmethanesulphonates (18) could be stepwise (Elcs, Scheme 14) or E2 depending upon the particular reaction. Either way, the sulphene intermediate reacted rapidly with the alcoholic solvent. King and his colleagues studied the reactions of the esters in deuterated solvents and observed the number of deuteriums incorporated into recovered starting material and into the products of the reacti~n.~~?~~ A different approach was used by Williams who, in part of a major investigation, obtained a discontinuous Bronsted plot, Figure 2, for the alkaline hydrolysis (and aminolysis) of aryl phenylmethanesulphonates (18; Ar = C6H5), Scheme 14.55 The large s4J.F. King and T. W. S. Lee, J. Am. Cliem. Soc., 1969, 91, 6524; J. F. King and R. P. Beatson, Tetrahedron Lett., 1975, 973. s5 A. Williams, K. T. Douglas, and J. S. Loran, J. Chem. Soc., Chem. Commun.. 1974, 689; M. B. Davy, K. T. Douglas, J. S. Loran, A. Steltner, and A. Williams, J. Am. C1iem. Soc.. 1977.99, 1196. Gordon, Maskill, and Ruasse (Base = OH-or pyridine) Scheme 15 Hammett [p(o-) = 5.41 and Bronsted (PIg = -2.4) parameters were consistent with phenolate character in the transition state of the rate-determining step for compounds to the right of the break in Figure 2 where specific base catalysis was observed.The rate, therefore, was very sensitive to the phenolate structure, and departure of the leaving group was taken to be rate limiting. As the leaving group became a weaker base, i.e. as ArO- became a better nucleofuge (towards the left of the plot in Figure 2), the gradient changed abruptly and the rate became independent of the leaving group, and the reaction became subject to general base catalysis. The deduction that proton abstraction had now become rate determining was confirmed when large primary deuterium kinetic isotope effects were observed, e.g. kH/kD= 6.0 with fluoride as nucleofuge, for those substrates to the left of the break in the Bronsted-type plot in Figure 2.Williams and his colleagues have also investigated the effect of substituents upon the base-induced elimination reaction of arylmethanesulphonates with good phenolate leaving groups. A concerted elimination but with an un-symmetrical transition state was proposed for the reaction of 2,4-dinitrophenyl arylmethanesulphonates (1 9) with pyridine and hydroxide in aqueous solution, Scheme 15.56 Although there is not a large development of charge at the alpha carbon in the transition state of the first step of this reaction, C-H bond cleavage is more advanced than S-0 fission, i.e. it is an ElcB-like E2 process. (Under the alkaline conditions of the reaction, the sulphene ArCHS02 will yield the corresponding arylmethanesulphonate ArCH2SO;).Kice also extended his work on substitution at diary1 a-disulphones (Section 3) 56 S. Thea and A. Williams, J. Chern. SOC.,Perkin Trans 2, 1981, 72; S. Thea, M. G. Harun, N. Kashefi-Naini, and A. Williams, ibid.,p. 78. Sulphonyl Transfer Reactions R1R2CHS02-Nuc + R’R~CHSO; d.;f R’ R2 CH-S02-S02CHR’R2 + Nuc-/ L20 (20) Nuc-H + [R1R2C=S02] + R’R2CHSOf OL-L*O1 Scheme 16 to study the reactions of nucleophiles with dialkyl a-disulphones (20) in aqueous di~xane.~~.’When di-n-butyl a-disulphone reacted with piperidine, morpholine, glycine ethyl ester, and acetate in 60% dioxane 40% D20, the products contained one, and only one, deuterium bonded to carbon. Either an irreversible ElcB or a concerted E2 mechanism was taking place in an overall elimination-addition (e.a.) reaction, Scheme 16.Only with weakly basic, highly nucleophilic azide did the reaction proceed by direct substitution (d.s.). But even with azide, the preferred route switched from direct substitution to elimination-addition when the hydrogens alpha to the sulphone groups became more acidic as in PhCH2S02- S02CH2Ph. From a comparison of the rates of elimination from a-disulphones RCH2S02- S02CH2R with the rates of base-catalysed hydrogen exchange with the solvent in the corresponding trifluoromethyl sulphones RCH2S02CF3, Kice and his colleagues concluded that the eliminations were either ElCB or very Elcn-like E2 processes. This view was fully in accord with observed variations in the rates of eliminations from RR’CHS02S02R’’ with changes in R and R’.A. Modified Su1phenes.-The rate constant for the hydrolysis of 2,4-dinitrophenyl 3,5-dimethyl-4-hydroxybenzenesulphonate(2 1) was found to have the dependence on pH shown in Figure 3.58 This effect could be attributed either to an associative mechanism (Path A, Scheme 17) where nucleophilic attack by hydroxide on (21) is inhibited as (21) undergoes ionization, or to a dissociative mechanism where the step which is rate limiting is the breakdown of the conjugate base (22) (elementary rate constant kt in Path B, Scheme 17). From an 57 L. 0.Farng and J. L. Kice, J. Am. Chem. Soc., 1981, 103, 1137. 58 S. Thea, G. Guanti, A. Hopkins, and A.Williams, J.Atn. Chrm. Soc., 1982,104, 1128. Gordon, Maskill, and Ruasse I 1 4 6 8 10 12 14 PH Figure 3 Dependence on pH of the hydrolysis of 2,4-dinitrophenyl 3,5-dimethyl-4- hydroxybenzenesulphonate, (21), in dioxane-water, 25 "C.Taken from reference 58 OH 0--Me J -H' L + ArO-7 Ka Me*Me Me*Me SO2 SO* S Path o// \oI 1 OAr OAr (21) (221 H20, fast Path A J OH + ArOH Me*Me SO? ( Ar = 2,t-dinitrophenyl) Scheme 17 estimate of the second-order rate constant for path A on the basis of the known reactivity of 2,4-dinitrophenyl benzenesulphonate, the authors concluded that only path B via the sulphoquinone intermediate was ~perating.'~ This is the first sound evidence for such intermediates even though they had been proposed earl- ier.59 A study of the reaction pathways of other aryl esters of ortho-and para-59 T.Zincke and R.Briine, Chem. Ber., 1908,41,902; W. L. Hall, J. Org. Chem., 1966,31,2672. Sulphonyl Transfer Reactions OH OH S020Ar I S020Ar &YOH\/ SOZOAr (231 (25) 0 (26) (27) (281 hydroxyarenesulphonic acids (23), (24), and (25) has also been made.60 As before, the observed second-order rate constants for hydroxide attack on the ionized hydroxyesters were many orders of magnitude larger than those for hydroxide attack on the corresponding 0-methylated compounds indicating that the same mechanism was not operative for both. The positive entropies of activation for the hydroxy compounds are also good evidence that a simple bimolecular process is not involved in these reactions.Esters (23), (24), and (25) also appear, therefore, to hydrolyse under basic conditions through ElcB mechanisms via sulphoquinone intermediates (26), (27), and (28). There does appear to be, however, an additional route for hydrolysis of (23)-(25) under alkaline con- ditions through hydroxide attack at the sulphonyl residues of their ionized forms since a more complex rate law was observed for these compounds than for (21). Anionic sulphene intermediates were invoked to account for the alkaline hydrolytic reactivity of aryl (methylsulphony1)methanesulphonates (29), i.e. substrates in which the acidity of the wCH~is so enhanced by the second sulphonyl group that they yield a dianion in basic aqueous solution.61 As for compounds (23)--(25), two hydroxide-induced mechanisms were indicated by the logk-pH profile and the results were fitted to the rate law 60 S.Thea, G. Cevasco, G. Guanti, A. Hopkins, N. Kashefi-Naini, and A. Williams, J. Org. Chem., 1985, 50, 2158. 61 S. Thea, G. Guanti, A. R. Hopkins, and A. Williams, J. Org. Chem., 1985, 50, 5592; S. Thea, G. Guanti, and A. Williams, J. Chem. SOC.,Chem. Commun., 1981, 535. Gordon, Maskill, and Ruasse Kh 2-C H3S02C H2 S020Ar CH3S02CS020Ar (29) k, J-ArO-k; J-ArO-[CH3S0zCH=SO2] [C H 3 SO2C=S023 kH20 H20 fast J+OH-fast 1+H20 CH~SOZCH2 S03H C H3S02 C HzSOj CH 3SO2CH2SO, + ArOH Scheme 18 where kH20,k,, and kb are indicated in Scheme 18.Williams and Thea and their colleagues were able to extract reactivity parameters for both k, and kb from their comprehensive results and show that an ElcB mechanism is required. At pH -10 the reaction is principally uia k, whereas, at higher alkalinity (pH -12-14), the route uia kb is required through the previously unknown anionic sul- phene. Related to this is the investigation of the hydrolysis of aryl sulphamates (30) (Scheme 19). Previous work 62 had demonstrated the intermediacy of the neutral sulphonylamine CH3N=S02 in the hydrolysis of aryl (N-methy1)aminosulphon- ates through an ElcB mechanism. Compounds (30) were found to be many orders of magnitude more reactive than the N,N-dimethyl derivatives which could react only by direct bimolecular attack at the sulphonyl By a strategy parallel with that employed in the investigation of (29) in Scheme 18, the joint Kent-Genova group present convincing evidence that both specific base- catalysed routes of Scheme 19 are required as well as an uncatalysed route [presumably uia nucleophilic attack by water at the sulphonyl group of (30) itself].At low, intermediate, and high pH, the predominant paths are those via kH20,k,, and kb respectively. 5 Bifunctionality and Alternative Reactivity Simple sulphones RS02R’ are not readily susceptible to nucleophilic attack. The electrophilic reactivity of the sulphone group in thiirane- 1,l-dioxide (3 l), however, is enhanced by the prospect of ring-opening and release of strain.Two possible modes of reaction are indicated in Scheme 20 nucleophilic attack at a ring carbon and protonation of oxygen to give a 2-substituted sulphinic acid 62 A. Williams and K. T. Douglas, J. Chem. SOC.,Perkin Trans. 2, 1974, 1727. S. Thea, G. Cevasco, G. Guanti, and A. Williams, J. Chem. SOC.,Chem. Commun., 1986,1582. Sulphonyl Transfer Reactions 2-HNS020Ar NSOZO A r [HN=S021 [N=S02]H20 H2°kIHzNSOsH + HzNSOi H2 N SO3 ArOH Scheme 19 , Y-= OH- tPath A (321 /\CH2-CH2 (31I Path B1 , Y-= OH-(Path A), and nucleophilic attack at sulphur and protonation on carbon to give an ethanesulphonic acid derivative (Path B). King’s group have re-examined the reaction of (3 1) with aqueous barium hydroxide following an earlier report 64 that the product is barium 2-hydroxyethanesulphinate (32), presumably via Path A.The later workers e~tablished,~~ however, that the main product is, in fact, barium ethanesulphonate (33) via Path B. They demonstrated that a common penta-coordinated intermediate (34) in a stepwise process accounts for both the major reaction shown in Scheme 20 (Path B) and also the minor formation of ethene and, presumably, sulphite (the last step of the Bordwell mechanism for the Ramberg-Backlund reaction). For reasons mentioned earlier (Section 2B), failure 64 G. Hesse, E. Reichold, and S. Majmuder. Chem. Ber., 1957.90, 2106. 65 J. F. King, J. H. Hillhouse, and K. C. Khemani, Can. J. Chem., 1985, 63, 1. 144 Gordon, Maskill, and Ruasse OH 'S O-I+/-/\CHZ-CH~ Y-L20 CHz=CH-SOz-CI d [Y-CHz-CH=S02] Y-CHz-CHL-SO3L (35) -a-(36) CH2=CH-S02-Y Scheme 21 to detect the intermediate by '*O exchange in recovered starting material does not disprove this mechanism.Just as thiirane- 1,l-dioxide has two possible sites of direct nucleophilic attack, so also do 1-alkene-1-sulphonyl chlorides, e.g. (35) in Scheme 21. The nucleophile could bond at the sulphur or, in a vinylogous alternative, at carbon-2 to give (36). In addition, a third reaction mode is conceivable for (35): elimination-addition via a cumulated sulphene intermediate (37). It appears that, for (35) itself in aqueous hydroxide, 90% of the reaction is to give ethenesulphonate anion via direct nucleophilic displacement at sulphur.66 However, the formation of small amounts (-10%) of 2-hydroxyethanesulphonate at both high and low pH indicated a second route.In D20, the absence of deuterium in the ethenesulphon- ate showed that it could not have been formed through the cumulated sulphene, and the single deuterium bonded to carbon in the 2-hydroxyethanesulphonate was consistent with it having been formed exclusively through (36; Y = OH). The addition of a tertiary amine, e.g. a pyridine, leads to a faster reaction and an altered product distribution (Scheme 22). In most cases, the betaine (38) is the major product (-80%) and the salt (39) the minor one, the exact ratios depending upon the substituent in the pyridine and the pH of the medium.66 The simplest mechanism to account for the betaine is attack of the pyridine at C-2 of (35) to form the cationic sulphene (40) in either a concerted or stepwise proce~s.~Formation of monodeuterated betaine in D20 supported this mech- anism.66 Four possible routes to the salt were considered, although the one via a cumulated sulphene [(37) in Scheme 211 was easily ruled out as no deuterium was found in the salt when the reaction was carried out in D20. General base catalysis was also dismissed as pyridines substituted in position 2 gave a reduced ''J.F. King, J. H. Hillhouse, and S. Skonieczny, Can. J. Chem.. 1984,62, 1977 ''A. Le Berre, A. Etienne, and B. Dumaitre, BUN.Soc. Chim. Fr., 1970,954. Sulphonyl Transfer Reactions X + CHz=CHSOzCI (35) [P;~-CH~CH=SO~ ci-] --+ (40) \ base I CH2 C H2SO2 C I (411 Scheme 22 rate, and no solvent deuterium kinetic isotope effect was observed.Vinylogous catalysis was established (rather than catalysis via direct attack of the nucleophile at sulphur) when the intermediate sulphene (40) was generated from another source, (41). In reactions under identical conditions, betaine (38) and salt (39) were produced in the same ratios from (41) and from (35) plus pyridine, implicating a common intermediate, i.e. (40). This reaction appears to be the first well-supported example of vinylogous nucleophilic catalysis. 6 Intramolecular Effects The hydrolysis of 2-hydroxyethanesulphonyl chloride (42) under alkaline condi- tions takes place largely through an intramolecular SN2-like route uia the transient p-sultone (43) which then suffers ring opening, Scheme 23.68In the presence of tertiary amines, e.g.pyridine, a minor extent of reaction (15-20%) occurs via the hydroxymethylsulphene as was shown by isotopic labelling in D20.Pyridine also intercepts the p-sultone. Reactions in alcoholic media are closely similar. Sultones more stable than (43), which bear the same relationship to sulphonate esters as lactones do to carboxylic esters, have been studied and compared with their acyclic analogues and the corresponding phosphorus compounds. Com- pound (44) with X = H reacts in alkaline solution (Scheme 24; Y-= OH-) with kl almost lo6faster than its acyclic analogue, phenyl phenylmethane~ulphonate.~~ This compares with ratios of up to lo7 in the phosphorus series. 2-Hydroxy-5-nitrotoluene-a-sulphonicacid sultone (44; X = 5-NO2), although having benzylic hydrogens next to the sulphonyl group, did not react via an ElcB 68 J.F. King and J. H. Hillhouse, Cun. J. Chem., 1983,61, 1583. Gordon, Maskill, and Ruasse HOC H2CH2S02CI H20,Pyr -HCI1 0H-+ HOCH~CH~SO~ Scheme 23 x X (44) Scheme 24 mechanism, but via direct nucleophilic attack at sulph~r.~~.~ Suppression of the ElcBroute was ascribed to stereoelectronic effects. The sulphene that would have been formed by proton abstraction and expulsion of the leaving group would have been constrained by the five-membered ring to be in an unstable perpendicular configuration rather than the preferred planar one.On the basis of the Pnuc and the solvent deuterium kinetic isotope effect in aqueous solution, sultone (44) appears to react with nucleophiles by general base catalysis and direct nucleophile catalysis routes via transition states with advanced S-Nuc bond f~rmation.~’ A study of the reverse reaction was also made (k-1, Scheme 24 with Y -= PhO -).70 Intramolecular nucleophilic participation (rather than general base catalysis) by the phenoxide of (45;X = 5-NO2) occurs to form the sultone which then goes on to suffer hydrolysis itself. The equilibrium constant (K = 1.4 x lo5 M-’) for the reaction as written in Scheme 24 was derived from forward and reverse rate constants.The stability of the sulphonate compared with the sultone was a further indication of ring strain in (44),especially as there is a translational entropy loss on going from two species on the left to one on the right of the equation as written. The way was then open to attempt a Bronsted-type study of the effect on the reactivity of the system in Scheme 24 of substituents X and of substituents in Y-.71 As normally employed, the absolute values of the Bronsted parameters Pnuc or Plg are of restricted use. Empirically, they indicate the extent to which structural changes in a reactant, e.g. substituents, affect the rate of the reaction, 6y T. Deacon, A. Steltner, and A. Williams, J. Chem. Soc., Perkin Trans. 2, 1975, 1778. ’* C.R. Farrar and A. Williams, J. Am. Chem. Soc., 1977,99, 1912. ”T. Deacon, C. R. Farrar, B. J. Sikkel, and A. Williams, J. Am. Chem. Soc., 1978,100,2525. Sulphonyl Transfer Reactions Table 1 Reaction of Scheme 24 ArO-+ 5-nitrosultone PWl) 0.81 P(k-1) -1.03 PhO-+ substituted sultones -0.85 0.84 P-C H3C6 H4S020A r + NH3 p-CH,C,H,+SO,NH; + ArO-( P-CH3C6HbS02NH2 + ArOH) Scheme 25 i.e. how they affect the transformation of reactant into activated complex. If we wish to interpret this parameter mechanistically, it is much more useful to know how the effect of the structural changes on the rate process compares with the effect of the same changes upon the overall equilibrium, i.e. the transformation of reactant into product.72 Since, in the present context, the effects of substituents upon both kl and k-1 were determinable, i.e.P(k1) and P(Ll), the effects of substituents upon the equilibrium are determinable, i.e. P(eq) = P(k1) -P(k-1). And, in a comprehensive investigation, this was done both for substituents X and for those in the phenolate Y-as shown in Table 1. The sensitivity of the equilibrium as written (left to right) (Scheme 24) towards changes in the pK, of the conjugate acid of the phenolate (45) is, therefore, Plg(eq) = Pl,(kl) -Pnuc(k-1) = -0.85 -0.84 = -1.69. In the same way, the sensitivity of the equilibrium as written towards the pK, of the conjugate acid of the external nucleophile Y-in Scheme 24 is Pnuc(eq) = Pnuc(kl)-Plg(k-l) = 0.81 -( -1.03) = 1.84.The Leffler parameters for the equilibrium as written, therefore, are alg = Plg(kl)/Blg(eq)= -0.85/-1.69 = 0.50, and, correspond- ingly, anuc= Pnuc(k1)/Pnuc(eq)= 0.81/1.84 = 0.44.These results indicate consider- able extents of bond cleavage and formation in the transition state, consistent with a highly symmetrical electronic structure around sulphur. For the ammonolysis reaction of Scheme 25, Pl,(kl) = -1.08 but Pnuc(k-1) results are not available so the data cannot be used in the Leffler sense. However, by using the comprehensive results for the related reaction of Scheme 24, and literature results for Pnuc(kl) of similar reactions but using arylamines as nucleophiles, Suttle and Williams73 conclude that the reaction of Scheme 25 is also best described as an SN~process.One of the more strongly argued cases for an SAN mechanism is the hydrolysis of a sulphonamide (46), nucleophilically catalysed by an intramolecular car-boxylate, Scheme 26.74The authors separated the effects of substituents upon the ’* J. E. Leffler, Science, 1953, 117, 340; A. Williams, Arc. Chenz. Rex, 1984, 17, 425; S. Thea and A. Williams, Chem. SOC.Rev., 1986, 15, 125. 73 N. A. Suttle and A. Williams, J. Chem. Soc., Perkin Trans. 2, 1983, 1563. 78 T. Graafland, A. Wagenaar, A. J. Kirby, and J. B. F. N. Engberts, J. Am. Chem. Soc., 1979,101,6981. 148 Gordon, Maskill, and Ruasse 0. xg&o-0 Scheme 26 overall kinetic process of (46) into an effect through the carboxyl (p = -0.54) and an effect through the sulphonamide (p = -0.58).These reaction constants indicate that both the carboxyl and the sulphonyl have lower electron densities in the activated complex than in the reactant. Furthermore, results for substituents in the leaving group of (47) lead to a three-point Hammett plot with p = -0.76 which, though nothing like as negative as the value for protonation of anilines (p = -2.89), does indicate that the nitrogen also suffers a reduction of electron density upon formation of the activated complex. These results are presented as indicating a stepwise mechanism as shown in Scheme 26 with breakdown of the intermediate (48) as the rate-determining step. However, if this reaction is actually concerted, then, structurally, the transition state must closely resemble (48).The mixed anhydride (49) will proceed to react very rapidly to give the final hydrolysis prod~cts.~~,~~ The intramolecular reaction of Scheme 27 bears some resemblances to that of Scheme 26 but, for the former, an intramolecular S~2-type of mechanism is indicated. The hydrolysis of the ortho-substituted compound (50) under alkaline Sulphonyl Transfer Reactions (53) Scheme 27 conditions was found to be lo3 times faster than that of the para-isomer, so the propinquity of the amide residue and the sulphonyl group clearly facilitates the reaction in some way.” Kinetic analysis of the reaction according to the pseudo-first-order rate law allowed comprehensive reaction parameters for the two routes to be obtained.The major route (k,) is indicated in Scheme 27 and involves rate-determining expulsion of the phenolate anion from the conjugate base of the amide function of the reactant (51). All the evidence could be accommodated by this being an intramolecular concerted process. The minor route (kb) was less well characterized but appears to involve specific intermolecular facilitation of the departure of the phenoxide from (51) by OH-, but to give the same intermediate that is obtained in the intramolecular route. Previous work had implicated compound (52; R = CH3) as an intermediate in the hydrolysis of 2-acetamidobenzenesulphonyl fluoride (although, in that reaction, the amide function also suffered hydrolysis). 76 Virtually all the oxygen of the amide in (53) was found to be derived from the aqueous solvent by an l80 labelling study as required by this mechanism.(The product, once formed, would not be expected to undergo oxygen exchange.) When ammonia was used as a nucleophilic trap, indeed, (52) was intercepted, but there was no rate increase. 7 Conclusions The evidence available for the reactions of equation 1 overwhelmingly suggests that the mechanisms are concerted bimolecular displacements at the sulphonyl group. There are indications, however, that an SN~mechanism may be induced if substituents in the arene group are sufficiently electron supplying. The use of ”S. Thea, G. Guanti, A. R. Hopkins, and A. Williams, J. Org. Chem., 1985,50,3336. 76 M. E. Aberlin and C.A.Bunton, J. Org. Chem., 1970,35, 1825. Gordon, Maskill, and Ruasse highly ionizing weakly nucleophilic media should help in the characterization of such a process. At the other extreme, there is evidence that some reactions may be adequately explained by an SAN mechanism through a penta-coordinate intermediate, but there is none, so far, that actually requires such a mechanism. Here again, it should be possible to build into the substrate certain features which, under the appropriate reaction conditions, cause the SAN mechanism to be the most favourable reaction. At this extreme, however, the evidence necessary to confirm the SAN mechanism may be more difficult to acquire. Acknowledgements. It is a pleasure to acknowledge helpful discussions with Dr. P. Monjoint and Professor A. Williams.

ISSN:0306-0012    

DOI:10.1039/CS9891800123

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chern. Soc. Rec., 1989. IS, 153-185 Synthetic Applications of Organotransition-metal Redox Reactions By Neil G. Connelly SCHOOL OF CHEMISTRY. UNIVERSITY OF BRISTOL, BRISTOL, BS8 ITS 1 Introduction The importance of electron-transfer reactions in organotransition-metal chemistry has only been fully recognized in the past few years despite the discovery of the ferrocene-ferrocenium couple as early as 1952. There are several major reasons for this recognition. First, there are now excellent examples of the redox activation of metal or ligand centres leading to the synthesis of otherwise inaccessible species. Such syntheses, for example of metal carbonyl cluster derivatives cirr electron transfer catalysis,2 are often fast, efficient, and stereo- or regioselective.Second, detailed mechanistic studies have shown that 17- and 19- electron compounds play important roles in a wide range of organometallic reactions; single-electron transfer to or from diamagnetic precursors provides a major route to such radicals. Third, the electronic structures of important classes of complex can be probed by X-ray structural studies of redox-related pairs of complexes, an approach pioneered by L. F. Dah14 for compounds with metal- metal bonds. Many of these topics have been recently described in general reviews5 of organometallic electrochemistry or in more specific accounts of electron-transfer catalysis,2-6 redox-induced structural change^,^ areneiron complexes as electron reservoirs,* the coordination chemistry of anion radicals,' and electrochemical techniques of particular use to the synthetic chemist." However, the synthetic applications of redox-active organometallics have not been specifically reviewed.' J. A. Page and G. Wilkinson. J. Am. Chern. Soc., 1952, 74, 6149.'D. Astruc. Angew. Cheni.. Int. Ed. Engl., 1988, 27. 643; and references therein. D. R. Tyler. Progr. Inorg. Chem., 1988, 36, 125. K. A. Kubat-Martin, M. E. Barr. B. Spencer, and L. F. Dahl, Organometaliic.~,1987, 6. 2570: K. A. Kubat-Martin, B. Spencer, and L. F. Dahl. Organometallics, 1987,6. 2580 and references therein. 'N. G. Connelly and W. E. Geiger, Ah. Orgunonzet. Clwni., 1984, 23. 1; W. E. Geiger and N. G. Connelly, Ad:. Organotiiet. Chenz., 1985, 24, 87; J. C. Kotz.in 'Topics in Organic Electrochemistry', Plenum Press, 1986; J. K. Kochi. J. Orgnnonief.Chern., 1986, 300. 139: P. Lemoine, Coord. Cliem. Rrc.., 1988, 83, 169; C. J. Pickett, in 'Electrochemistry', A Specialist Periodical Report. ed. D. Pletcher, The Royal Society of Chemistry, Vol. 8. p. 81, 1983 and Vol. 9, p. 162, 1984. M. Chanon, ALT.Clienz. Res.. 1987, 20. 214. 'W. E. Geiger, Progr. Inorg. Chem., 1985, 33. 275.'M. Lacoste. M. H. Desbois, and D. Astruc, Neu, J. Cheni., 1987. 11, 561; D. Astruc, Ace. Clienz. Rex, 1986, 19. 377. 'I W. Kaim, Coord. Chenz. Rer.. 1987, 76. 187. W. E. Geiger. in 'Inorganic Reactions and Methods', Vol. 15, ed. J. J. Zuckermann. Verlag Chemie. 1986. p. 110; W. E. Geiger and M. D. Hawley. in 'Physical Methods of Chemistry', Vol.2, ed. R. W. Rossiter. J. Wiley and Sons, 1986, Chapter 1. I53 Synthetic Applications of' Organotransition-metal Redox Reactions It is the aim of this article, therefore, to redress this imbalance. Most of the examples given are from the author's own work but key references are provided to lead the reader into related areas of interest. Our entry into the area of organometallic electrochemistry effectively began in the early 1970s with the unexpected synthesis" of the stable, paramagnetic alkyne complex [Cr(CO)z(q-PhC =CPh)(q-CgMe6)][PFs] from [NO][PF6] and [Cr(Co)2(q-PhC=CPh)(~-c6Mes)l (rather than the diamagnetic nitrosyl [Cr(Co)(No)(q-PhC_cPh)(~-c6Me6)]'); the isolation of salts of other radical cations such as [Cr(Co)2(PPh3)(q-C6Me,)l +,12 [V(C0)3(PPh3)Cp] (Cp =+ q5-C5H5),' [Mn(CO)(dppe)Cp]+ (dppe = Ph2PCH2CH2PPh2),14 and [Fe(CO){ P(OMe)3} 2(q-C4Ph4)] + (C4Ph4 = tetraphenylcyclobutadiene) soon followed.These particular complexes were somewhat disappointing in showing little enhanced reactivity towards nucleophiles or radicals. However, they helped to show that 17-electron cations could be widely anticipated in organometallic chemistry, and the methods developed for their synthesis and isolation provided the experimental basis for all of our subsequent studies, some of which are featured in Sections 2-6 below. 2 The Complex Chemistry of the 17-Electron Cation trans-[Fe(C0)3(PPh3)2]+ The chemistry of trans-[Fe(CO)3(PPh3)2]+ (1 +) provides a simple but effective example of the activation of a metal carbonyl complex by one-electron oxidation.Prepared l6 as a green, crystalline [PF6] -salt from trans-[Fe(CO)3(PPh3),) (1) and AgPF6, (1') contrasts with the 17-electron cations noted in Section 1 in being the precursor 16,1 to a wide range of other carbonyl derivatives (Scheme 1). The reaction with NOz is particularly noteworthy in giving [Fe(C0)2(NO)(PPh3)2]+ uia the intramolecular linking of coordinated CO and NO2 and the elimination of C02. The reactions of (1 ') with iodide and iodine, coupled with kinetic studies using stopped flow i.r. spectroscopy, implicate the radical cation in the oxidative elimination l6 of (1) with 12 to give [FeI2(CO)3(PPh3)] (Scheme 2). In addition, the route by which two equivalents of (1+) and one of 1-yield (1) and [FeI(C0)3(PPh3)2]f [i.e.steps (iii) plus (iv) in Scheme 21 is also followed with other diamagnetic reactants. Thus, [S2CNMe2] -and the adduct CS2-PPh3 give the iron(1r) dicarbonyls [Fe(CO)2(S-S)(PPh3)2]z (S-S = S2CNMe2, 2 = 1; S-S = S2CPPh3, Z = 2) l7 respectively. The cyclic voltammogram of [Fe(C0)3(AsPh3)2] (2) is notably different from that of (1) in showing a completely irreversible oxidation wave in CH2C12. Despite the obvious instability of the cation (2') its e.s.r. spectrum is readily "N. G. Connelly and G. A. Johnson, J. Organomet. Chem.. 1974,77,341. l2 N. G. Connelly, Z. Demidowicz, and R. L. Kelly, J. Chem. Soc., Dalton Trans., 1975, 2335. N. G. Connelly and M. D.Kitchen, J. Chem. Soc., Dalton Trans.. 1976, 2165. l4 N. G. Connelly and M. D. Kitchen, J. Chem. Soc., Dalton Trans., 1977,931. I5 N. G. Connelly, R. L. Kelly, and M. W. Whiteley, J. Chem. Soc., Dalron Trans., 1981, 34. l6 P. K. Baker, N. G. Connelly, B. M. R. Jones, J. P. Maher. and K. R. Somers. J. Chem. Soc., Dalton Trans., 1980, 579. 17 P. K. Baker, K. Broadley, and N. G. Connelly. J. Cii(w.So(.. Dollon Truti.~..1982, 471. 154 Connelly L L = PPh3, X2 = halogen, 0-0=1,2 diketone Scheme 1 [Fe(C0)3L2] + I2 -[Fe(C0)3L2]+ + 1. + I-(0 [Fe(C0)3L2] + 1. -[Fe(C0)3L2]+ + 1-(ii) [Fe(C0)3L2]+ + I--[FeI(C0)3L2] (iii) [Fe(C0)3L2]+ + [FeI(CO),L2] -[Fe(C0)3L2] + [FeI(C0)3L2]+ (iv) [FeI(CO)3L2] + I--[Fe12(C0)3L] + L (v>+ L = PPh3 Scheme 2 detected when AgPF6 is added to (2) in CH2ClZ, even at room temperature.Here, the silver(1) ion most probably acts as a 'non-innocent' oxidant, giving the diamagnetic adduct [Ag(Fe(C0)3(AsPh3)2}2]+ (cf [Ag(Rh(CO)(PPh,)Cp}2]+, Section 3C) which slowly dissociates to give a steady-state concentration of (2+), together with silver metal and (2). The 'non-innocence' of chemical oxidants and reductants is a recurrent theme throughout organometallic electrochemistry and the possibility of reactions which compete with the desired redox process (e.g. adduct formation, substitution, etc.) should always be borne in mind. The stable and readily accessible cation (1') has also lent itself to detailed spectroscopic * and electrochemical studies 9,20 which have confirmed the proposed l6 structure and provided evidence that the substitution reactions of 17-electron cations generally occur via associative pathways.I' M. J. Therien and W. C. Trogler, J. Am. Chem. Soc., 1986,108,3697. l9 R. N. Bagchi, A. M. Bond, C. L. Heggie, T. L. Henderson, E. Mocellin, and R. A. Seikel, Inorg. Chem., 1983,22,3007. M. J. Therien, C.-L. Ni, F. C. Anson, J. G. Osteryoung, and W. C. Trogler, J. Am. Chem. Soc., 1986. 108,4037. Synthetic Applications of Organotransition-metal Redox Reactions -e [Co(CO)(PPh3)Cp] [Co(CO)(PPhj)Cp] ' L L I 0.5 0.0 -0.5 Potential / V vs. 5.c.e. Figure 1 Multiple-scan cyclic voltammogram of [Co(CO)(PPh3)Cp] (3) in the presence of PPh3 in thf (Reproduced fromJ.Chem. SOC., Dalton Trans., 1983, 121) 3 The Redox Chemistry of [M(C0)2(q-C5R5)] Derivatives (M = Co or Rh; R = H, Me, or Ph): Synthetic and Structural Consequences A. Introduction.-Although the chemistry of (1 ') (Section 2) provides an excellent example of the oxidative activation of a metal carbonyl, the compounds [M(C0)2-,L,(q-C5R5)] (M = Co or Rh; L = P- or As-donor; n = &2; R = H, Me, or Ph), which are isoelectronic with (l), are more versatile in showing that the variation of M, L, n, and R can drastically affect the reactivity of the 17- electron cations [M(C0)2 -,Ln(q-C5R5)] +. There are also added bonuses with this set of complexes in that [M(C0)2(q-C5R5)] can be reduced to 19-electron radical anions, and X-ray and e.s.r. spectroscopic studies provide a detailed picture of the bonding in the three-membered electron-transfer series [M(C0)2 -,-Ln(q-C5R5)JZ(Z = 1,0, and -1). B.Cobalt Complexes.-The oxidative electrochemistry of [M(C0)2(q-C5RS)] (M = Co or Rh, R = H or Me) in CH2C12 is complex and depends on the electrode material (Pt or Hg); only in the case of [CO(CO)~(~-C~M~~)]+is the Connelly primary redox product stable enough (at -20 "C) to be detected by e.s.r. spectroscopy.21 However, the monocarbonyl [Co(CO)(PCy3)Cp] (Cy = cyclo-hexyl) is reversibly oxidized at room temperature and [FeCp2][PF6] gives the stable, green crystalline salt [Co(CO)(PCy3)Cp][PFs] whose reactions with halogens, 1,2-diketones, and nitric oxide parallel those of (1 +) (Scheme 1).By contrast, the chemical oxidation of [Co(CO)(PPh3)Cp] (3) gives the carbonyl substitution product [CO(PP~~)~C~]+, yields of which are nearly quantitative in the presence of added PPh3. The mechanism of this reaction (Scheme 3) is revealed by the multiple scan cyclic voltammogram (Figure 1) of (3). On the first sweep, a partially reversible oxidation wave is observed at 0.25 Vj-coupled with a product wave which subsequent sweeps show to be reversible (Eo = -0.51 V). The latter corresponds to the reduction of [Co(PPh3)2Cp]+ so that initial one-electron oxidation of (3) to (3+) is followed by rapid substitution to give the bis(phosphine) cation. The dicarbonyls [Co(CO)z(q-C5RSI] (R = H or Me) react only very slowly with PPh3 at room temperature, but in the presence of [FeCp2]+ rapid oxidative +substit u tion again occurs giving [Co(PPh3)2Cp] and [Co(CO)( PPh 3)(q -C5Me5)]+.In all of these reactions, substitution is facilitated by the positive charge generated by one-electron oxidation.This charge not only renders the metal centre more susceptible to nucleophilic attack but also reduces x-back donation to the x*(CO) orbitals, thereby weakening the Co-CO bond.22 The oxidative chemistry described above provides not only a synthetic route to new compounds but also evidence to support molecular orbital (m.0.) schemes for the ring-metal bonding in [M(C0)2(~-CsRS)]23 and for the x-bonding of phosphine ligands to transition metals.24 X-Ray structural studies 2s by McKinney and co-workers on the redox-related pair [Co(PEt3)aCp]" (2 = 0 or 1) show that the C5Hs ligand is q2,q3-bound to the metal ('allyl-ene' bonding) in the neutral complex (Figure 2a) but more symmetrically bonded (ie.closer to qs-CsHs) in the cation. Just such a structural change is expected on the basis of m.0.calculations23 which show that the highest occupied molecular orbital (h.o.m.0.) for the neutral complex involves overlap of the el+ ring orbital (Figure 2b) with the metal d,, orbital. One- electron oxidation results in partial loss of anti-bonding character in bonds C(2)-C(3) and C(4)-C(5), loss of bonding character in bonds C( l)-C(2), C(3)-C(4), and C(l)-C(5), and, therefore, in a more symmetrical ring. t Unless stated otherwise, all potential data quoted in this review are relative to the saturated calomel electrode and are taken from cyclic voltammetric measurements at a platinum bead electrode in CH2C12 containing [NB@][PF6] as base electrolyte.Under these conditions, the E * value for the ferrocene- ferrocenium couple is 0.47 V. For reversible processes, Eo values {taken as [(EP)~~ are+ (E~)~~,j]/2) given; for irreversible processes the oxidation peak potential, (Ep),,, or reduction peak potential, (EP)~~~, at a scan rate of 200 mV s-', are given. T. Gennett, E. Grzeszczyk, A. Jefferson, and K. M. Sidur, Inorg. Chem., 1987,26, 1856. 22 K. Broadley, N. G. Connelly, and W. E. Geiger, J. Chem. SOC.,Dalton Trans., 1983, 121. 23 L. R. Byers and L. F. Dahl, Inorg. Chem., 1980, 19, 277; D.L. Lichtenberger, C. H. Blevins, 11. and R. B. Ortega, Organomelallics, 1984,3, 1614. 24 D. S. Marynick, J.Am. Chem. SOC.,1984, 106,4064. 25 R. L. Harlow, R. J. McKinney, and J. F. Whitney, Organometallics, 1983,2, 1839. 157 Synthetic Applications qf Organotransition-metal Redox Reactions +el el -(a) ( b) (d Figure 2 (a) The 'allyl-me' bonding mode for the c-yclupentadienyl ligand, and (h) and (c)the el+ and el-ring orbitals involved in the bonding of' the ring to the metal in [M(C0)2(C5R5)]comple.xes Further inspection 26 of McKinney's structural data showed that oxidation is also accompanied by a shortening of the P-C2HS bonds, from 1.846(3) to 1.829(3)A. Again, the change supports m.0. calculations 24 which suggest that the n-acceptor orbital of a phosphine ligand, PX3, is a hybrid of P 3d and P-X o* components (rather than simply the P 3d orbital which was often held to be too high in energy to participate in n-bonding).Thus, decreased back-bonding in the cation results in depopulation of the P-X G* orbital. The reductive electrochemistry of [CO(CO)~(~-C~R~)] is not as extensive as that based on oxidation but nevertheless also leads to insight into bonding. The cyclic voltammogram of [Co(CO),Cp] shows an irreversible reduction wave, and reaction of the dicarbonyl with sodium amalgam gives the paramagnetic anion [(Co(CO)Cp)2] p.27 By contrast, the C5Ph5 analogue is reversibly reduced and the mononuclear, 19-electron anion [Co(C0)2(q-C5Ph5)] -is readily produced at room temperature either electrolytically or chemically using sodium naph- thalenide; the thermal stability of this anion probably derives from the sheer bulk of the C5PhS ring (cone angle ca.203"). An analysis of the frozen solution e.s.r. spectrum 28 of [Co(CO)2(q-C5Ph5)]- showed that the C5-ring is again symmetrically bound to the cobalt atom (cf: [Co(PEt3)zCp]+), and this result is also compatible with the m.0. bonding scheme for [M(C0)2(q-CSRs)] (M = Co or Rh).23The 1.u.m.o. of the neutral dicarbonyl involves overlap between the el- ring orbital (Figure 2c) and the metal dyz orbital. Thus, half population of this l.u.m.o., ilia one-electron reduction, leads to a shortening of bonds C(2)-C(3) and C(4)-C(5) and thence to a more symmetrical C5-ring.C. Rhodium Complexes.-The cyclic voltammogram of [Rh(CO)(PPh3)Cp] (Figure 3) is very different from that of the cobalt analogue (3) in showing29 an irreversible oxidation wave at ca. 0.45 V coupled with a reversible product wave centred at 0.05 V. Thus, it is unsurprising that the chemical oxidation of ''A. G. Orpen and N. G. Connelly, J. Chern. Soc., Chmi. Commun., 1985, 1310. N. E. Shore, C. S. Ilenda, and R. G. Bergman, J. Am. Chem. Soc., 1977,99, 1781. N. G. Connelly, W. E. Geiger, G. A. Lane, S. J. Raven, and P. H. Rieger. J. Am. Chrm. Soc., 1986, 108, 6219. 29 N. G. Connelly, A. R. Lucy. J. D. Payne, A. M. R. Galas, and W. E. Geiger, J. Chem. Soc.. Dalton Trans., 1983. 1879. Connelly 1 I L 1 1 0.8 0.4 0 Potential/ V vs s.c.e.Figure 3 Multiple-scan cyclic voltammogram of[Rh(CO)(PPh3)Cp] in CHzClz (Reproduced from J. Ckem. Soc., Dufton Trans., 1983, 1879) [Rh(CO)(PPh3)Cp], with [FeCpz]+ or [N~C~H~F-P]+,does not lead to the isolation of the radical cation [Rh(CO)(PPh3)Cp] but rather to a secondary + product, namely the dicationic fulvalene complex [Rhz(CO)2(PPh3)2(q5 :~7’~-C1oH8)l2+ [(4) Scheme 41. The latter is easily reduced, by sodium amalgam or electrolytically, to [Rh2(CO)2(PPh3)2(q5 :q’5-Cl~Hs)] (5), and both (4)and (5) have been structurally characterized 30 (Figure 4). Compound (4)has the metal- metal bond required to account for the observed diamagnetism, but it is long [2.930(2) A] and the strain in the structure is also reflected in the bending and twisting of the CIOH8 bridge. By contrast, the two Rh(CO)(PPh3) groups of (5) are trans-disposed with respect to a planar fulvalene ligand.Both (4)and (5) show one reversible wave in their cyclic voltammogram (the dication is reduced and the neutral molecule is oxidized) at a potential (0.05 V) identical to that of the product wave in Figure 3. The peak separation of 38 mV observed for this wave was thought 30 to indicate simultaneous one-electron transfer at two non-interacting metal centres and to be incompatible with the structural differences between (4) and (5). However, this interpretation was erronedus and the electrochemical data are consistent with an ECE process (E = 30 M. J. Freeman, A. G. Orpen. N.G.Connelly. I. Manners. and S. J. Raven, J. Chem. Soc., Dalton Trans., 1985.2283. 159 Synthetic Applications of Organotransition-metal Redox Reactions I d (a 1 I b) Figure 4 The structures of (a) the dication [Rhz(CO)2(PPh&(qJ: q'5-Cl~Hs)]2+ (4),and (b) the neutral molecule [Rh2(CO)2(PPh3)2(q5:q'5-Cl~H8)](5); hydrogen atoms are omitted for clarity (Reproduced from J. Chem. SOC., Dalton Trans., 1985,2283) Scheme 4 electrochemical, C = chemical) where the intervening chemical step, namely cis -trans isomerization. is fast and re~ersible.~' 31 R. Moulton, T. W. Weidman, K. P. C.Vollhardt, and A. J. Bard, lnorg. Chmi., 1986,25, 1846. Connelly L= PPh, JQQi.1.1 I 1 Rh'r"y -/ '\\ L co oc L L co \/ +2 e-Rh+-Rh+ --2e-(4) (5) Scheme 5 A fulvalene complex related to (4),namely [Rh2(PPh3)4(j~-CloHs)]~+,was prepared by McKinney using AgBF4 to oxidize [Rh(PPh3)2Cp].Surprisingly, then, the reaction between AgPF6 and [Rh(CO)(PPh3)Cp] gives 29 the adduct [Ag{Rh(CO)(PPh3)Cp}2][PF6](6) despite the fact that the silver(1) ion is a strong enough oxidant to bring about electron-transfer and the formation of (4).As in the case of [Fe(C0)3(AsPh3)2] (Section 2), Ag' acts as a 'non- innocent' oxidant. However, in contrast to [Ag(Fe(C0)3(AsPh3)2) 23 +,the adduct (6) is an air-stable, crystalline solid also notable in behaving as a controlled source of the transient radical cation [Rh(CO)(PPh3)Cp] +.Thus, (6) and NO gas gives [Rh(NO)(PPh3)Cp] [cJ: the coupling reaction between (1 ')+ and NO; Scheme 11; no reaction occurs between nitric oxide and neutral [Rh(CO)(PPh3)Cpl.The proposed mechanism for the formation of (4)from [Rh(CO)(PPh3)Cp] is shown in Scheme 5. Intuitively, it should involve C-C rather than Rh-Rh coup- ling of [Rh(CO)(PPh3)Cp]+, assuming the positive charge to be localized on the metal centre. However, the oxidation of [Rh(CO){P(OPh)3)Cp], either elec- trolytically or with [FeCpz] +,yields the dimer [Rh2(C0)2{ P(OPh)3)2CpJ2+ which has an unsupported metal-metal bond (2.814 A) and unlinked cyclo- 32 R.J. McKinney. J. Chem. Soc., Chem. Commun., 1980,603. Synthetic Applications of Organotransition-metal Redox Reactions pentadienyl rings.33 Thus, it is more likely that the metal-metal bond of (4)is formed before the cyclopentadienyl rings are coupled; how the CI0H8 ligand then results (i.e.whether proton or hydrogen atom elimination occurs) is unknown. The monomeric radical cation [Rh(CO)(PPh3)Cp]+ has so far proved too reactive to be detected spectroscopically. By contrast, the pentaphenylcyclo- pentadienyl analogues [Rh(CO)L(q-C5Ph5)]+ (L = PPh3 or AsPh3) have been characterized 34 at room temperature in CHzCl2 by both i.r. and e.s.r. spectro- scopy. The remarkable stability conferred by the C5Ph5 ring (cj [Co(CO)2(q-CsPhs)] -,Section 3B) has synthetic consequences in that oxidative substitution reactions, similar to those described for [CO(CO)~(~-C~R~)] (R = H or Me, Section 3B) and dependent on the longevity of intermediate radical cations, can be carried out with [Rh(C0)2(q-CsPhs)] [but not with the C5R5 (R = H or Me) analogues)].The reaction of [Rh(C0)2(q-C5Ph5)] with P(OPh)3 is particularly noteworthy in requiring only a catalytic amount of [FeCpz]' as oxidant and in directly yielding the neutral complex [Rh(CO){P(OPh)3)(q-C,PhS)I. In the presence of the phosphite ligand the oxidation of [Rh(C0)2(q1-C5Ph5)] [(Ep),, z 1.0 V] is effected by [FeCpZ]', the irreversible loss of CO presumably providing the driving force to overcome the thermodynamically unfavourable dif- ference in potentials. Once formed, the long-lived cation [Rh(CO){ P(OPh),)- +(q-CsPhs)] (Eo = 0.67 V) can also oxidize [Rh(CO)*(q-C5Ph5)], thereby continuing the catalytic cycle and giving [Rh(CO){ P(OPh)3)(q-C5Phs)] as the final product.Many other examples of this important type of reaction, namely electron- transfer catalysis, are now known, particularly in carbonyl cluster chemistry. Such reactions, however, have been recently reviewed and will not be dealt with further here. Note, however, a new redox-induced pathway35 by which the two-electron oxidation of a dianion such as [Fe5C(CO)14]2- in the presence of a Lewis base, L, gives neutral substituted clusters such as [FesC(CO) I&]. 4 Redox-induced Isomerization Of the structural changes induced by electron-tran~fer,~ the reductive conversion of non-conjugated into conjugated dienes in [Co(q4-diene)(q-C5R5)] 36 and the oxidative isomerization of Group VI and VII metal carbonyl derivatives have been studied in most detail.This Section concentrates on the latter and in particular on how the isomerization process might play an important part in syntheses involving octahedral carbonyl complexes of manganese and rhenium. The oxidatively-induced isomerization reactions in question follow an ECEC pathway, represented by the 'square scheme' shown in Scheme 6 33 E. Fonseca, W. E. Geiger, T. E. Bitterwolf, and A. L. Rheingold, Organometalhcs, 1988, 7,567. 34 N. G. Connelly and S. J. Raven, J. Chem. Soc... Dalton Trans.. 1986. 1613. 3s S. R. Drake, B. F. G. Johnson, and J. Lewis. J. Chem. SOL..,Chem. Comrnun., 1988, 1033. 36 W. E. Geiger. T. Gennett, M. Grzeszczuk, G. A. Lane, J. Moraczewski.A. SalLer, and D. E. Smith, J. Am. Chem. SOL..,1986, 108, 7454. 162 Connelly Br Br (8) (a+) L = P (OPh),, P-P= dppm Scheme 6 for the conversion of cis,cis-[MnBr(C0)2(P(OPh)3}(dppm)] (7; dppm = Ph2PCH2PPh2) into trans-[MnBr(CO)z{ P(OPh)3)(dppm)] (8).37As will become apparent, however, the cross-reaction given by equation 1 must also be taken into account. The kinetic and thermodynamic parameters associated with square schemes can be quantified using either electrochemical or spectroscopic technique^.^^ For example, all of the rate and equilibrium constants for the reactions linking cis-and tr~ns-[M(CO)~(dppe)~]'(Z = 0 or 1, M = Mo or W) have been determined.39 The equilibrium constant for the cross-reaction depends on the redox potentials of the two couples involved, for example (7)-(7+) and (8)-(8+) in Scheme 6.These potentials may differ considerably [e.g. 0.31 and 0.89 V for the oxidations of (8)and (7)respectively] or be very similar [e.g. 1.06 and 1.10 V for the oxidation of mer- and fa~-[Mo(C0)~(P(OPh)3}~]; the basis for the 38 variations of redox potential with geometry has been discussed re~ently.~' Oxidative isomerization reactions have long been known to provide simple routes to otherwise inaccessible isomers. For example, the oxidation of cis,rner-[MnBr(C0)2{ P(OMe)3}3] 41 or cis,cis-[MnBr(C0)2L(P-P)] [L = P(OPh)3, 37 N. G. Connelly, K. A. Hassard, B. J. Dunne, A. G. Orpen, S. J. Raven, G. A. Carriedo, and V. Riera, J. Chrm. Soc., Dalton Trans., 1988, 1623.38 See, for example, A. M. Bond, S. W. Carr, and R. Colton, Organometallics, 1984,3, 541. 39 A. Vallat, M. Person, L. Roullier, and E. Laviron, Inorg. Chem., 1987,26. 332. 40 B. E. Bursten and M. R. Green, Progr. Inorg. Chem., 1988,36, 393. R. H. Reimann and E. Singleton, J. Chem. Soc.. Dalron Trans., 1973, 2658. Synthetic Applications of Organo transition-metal Redox Reactions P-P = dppm, dppe, etc.I4' followed by hydrazine reduction of the resulting manganese@) cations, gave trans,mer-[MnBr(C0)2( P(OMe)3) 3] and trans-[MnBr(C0)2L(P-P)] respectively. Although such reactions appear straight-forward recent studies have uncovered two catalytic processes which considerably complicate matters. The addition of small quantities of trans-[MnBr(CO)2(P(OPh)3)(dppm)] +(8+) to the neutral trans-compound (8) results43 in the quantitative formation of cis,cis-[MnBr(C0)2{P(OPh)3)(dppm)] (7).Although the equilibrium (equation 2) lies far to the right, the cross-reaction (equation 1) is thermodynamically favoured by 0.58 V and provides the driving force for the catalytic cycle. The conversion of trans-into cis-[Cr(C0)4(P(OMe)3)~] catalysed by trans-[Cr(C0)4(P(OMe)3)2]+ occurs similarly.44 In both of these examples the less thermodynamically stable isomer is converted into the more stable isomer. In the second process the conversion offac-[Cr(C0)3{ P(OMe)3)3] (EO = 0.57 V) to mer-[Cr(C0)3(P(OMe)3}3] (Eo = 0.42 V) is catalysed by mer-[Cr(C0)3(P(OMe)3)31 +.45 Here, although the cross-reaction (equation 3), as written, is thermodynamically unfavourable by 0.15 V, the catalytic cycle is driven by the rapid isomerization off&+ to mer+.mer' + fac emer + fac+ (3) The synthetic implications of these two different catalytic processes have not previously been considered and yet they may be dramatic. Consider, for example, the reduction of (8+). Here, the relative amounts of (7) and (8) produced may depend on the method used to carry out the electron-transfer reaction. Thus, controlled potential electrolysis, which will normally take 10 to 20 minutes for completion, leaves (8+) in contact with (8) thereby facilitating the catalysed formation of (7). By contrast, any method which will result in fast and complete quenching of (8+), for example rapid addition of a soluble reductant to a well- stirred solution of the cation, should result in higher yields of (8). Note also that the all-important cross-reaction is second order so that dilution may attenuate the catalytic cycle and lead to an increased proportion of (8).Such effects may explain the observation that in warm alcohols NaBH4 reduction of trans,mer- [MnBr(C0)2L3]+ [L = PMezPh, P(OMe)3, P(OEt),, or P(OMe)2Ph] gave the neutral cis,mer-isomer in all cases but that hydrazine in a mixture of CH2C12 and 40-60 petroleum ether gave trans,mer-[MnBr(C0)2L3] [L = P(OMe)3 or P(OEt)31.4 42 F. Bombin, G.A. Carriedo, J. A. Miguel, and V. Riera, J. Chern.SOC.,Dalton Trans., 1981,2049. 43 N. G. Connelly, S. J.Raven, G.A. Carriedo, and V. Riera, J. Chern. SOC.,Chem. Cornrnun., 1986,992. 44 A. M. Bond, R. Colton, and T. F. Mann, Organornetallics, 1988,7,2224. 45 A. M. Bond, R. Colton, and J. E. Kevekordes, Inorg. Chern., 1986,25,749. 164 Connelly The effects of isomerization in general and the attendant catalytic processes in particular may be relatively straightforward to test and control in the ground- state redox reactions so far described. However, they should also be considered in the excited state, particularly for photoactive rhenium carbonyl complexes such as [ReX(C0)3(N-N)] [X = halide; N-N = 2,2'-bipyridyl (bipy), o-phe- nanthroline (o-phen), etc.), even though they may be less easily quantified. The electron-transfer catalysed substitution of [Re(CO)3(NCMe)(o-phen)] + with PPh3 to give [Re(CO)3(PPh3)(o-phen)] +,46 and the catalysed reduction of C02 to CO mediated by [ReBr(C0)3(bipy)]47 both depend on the reductive quenching of a photogenerated excited state to give a strong reductant, for example [ReBr(C0)3(bipy)] -in the second case. In both reactions noted above, the reduction potential for the excited state was estimated from the known values for the ground-state reduction potential and the energy of the incident radiation.However, if isomerization follows photoexcitation a dramatic effect on reduction potential is possible, i.e. the excited state may be a much stronger or weaker reductant than assumed. In addition, any such isomerization reaction may well be likened to catalytic processes of the types described above.A reductively-induced isomerization process has been observed with manganese bipyridyl complexes related to the rhenium compounds mentioned above, but only in the ground state. For example, cis,cis-[Mn(C0)2(CNBut)2(N-N)]and+ fac-[Mn(CO)(CNBut)3(N-N)](N-N = bipy or o-phen) give the corresponding + cis,trans- and mer-isomers when treated sequentially with sodium amalgam and air.48 It is apparent from cyclic voltammetry that the mechanism of this isomerization is not the same as that for the oxidative reaction; a square scheme as simple as that in Scheme 6 seems unlikely. Nevertheless, e.s.r. spectroscopy shows that the reduction step generates radicals which are stable at room tempera- t~re.~'.~~ Mixed valence complexes are discussed in more detail in Section 6.However, it is appropriate to note here the redox chemistry of the binuclear cyanide-bridged cations (9+)-( 12') which undergo oxidative isomerization reactions similar to those relating the neutral mononuclear bromides (7) and (8). Each of the binuclear complexes is oxidized in two one-electron steps with trans-sites oxidized more readily than cis-sites, the latter with isomerization [as expected from the redox chemistry of (7) and (8)]. For (9+) (Eo= 0.61 and 1.22 V) and (10') CEO = 0.66 V and (Ep),, = 1.55 V] oxidation by [NO]' gives the mixed-valence dications (92') and (lo2+)for which both cyclic voltammetry and i.r. carbonyl spectroscopy suggest two non-interacting sites, i.e.trapped valence [MnzI3 +-containing cores. For (1 1+),however, where the two oxidation waves are closer in potential [Eo= 0.85 V and (Ep),, = 1.42 V] isomerization of (9'+) 46 D. P. Summers, J. C. Luong, and M. S. Wrighton, J. Am. Chem. Soc., 1981, 103, 5238. 4' C. Kutal, A. J. Corbin, and G. Ferraudi, Organometallics, 1987,6, 553. 48 F. J. Garcia Alonso, V. Riera, M. L. Valin, D. Moreiras, M. Vivanco, and X. Solans, J. Organomet. Chem., 1987,326, C71. 49 I. C. Quarmby, Ph.D. Thesis, University of Bristol, 1988. N. G.Connelly, I. C. Quarmby, F. J. Alonso Garcia, and V. Riera, unpublished results. Synthetic Applications of Organotransition-metal Redox Reactions 1+ L C0 L = P(OPh&, P-P = dppm seems to occur on formation of (112'), i.e.on oxidation of the trans-site only.51 The difference in potential between the two oxidation waves of the binuclear species [(P-P)L(CO)2Mn(p-CN)Mn(C0)2L'(P-P')]+ can be systematically varied by manipulating the donor-acceptor properties of the ligands [L,L'= PEt,, P(OPh)3, etc.; P-P, P-P' = dppm, dppe, Me2PCH2CH2PMe2, et~.],~~ and the oxidatively-induced isomerization process, readily detected by cyclic voltammetry and i.r. spectroscopy, should provide a simple probe of the mixed- valence properties of the [Mn2I3 +-containing oxidation products. 5 The Redox Activation of Coordinated Hydrocarbons A. C-H Bond Reactions.-Of the several examples now known of C-H bond cleavage induced by oxidation, three are applicable to a wide range of substrates.In each case the products isolated, usually cationic, are prone to nucleophilic attack at the coordinated hydrocarbon and are therefore precursors to many new organometallic and organic compounds. The first example involves the 02 oxidation of electron-rich substrates, most commonly 19-and 20-electron areneiron complexes. The reaction is complex, but well ~nderstood.~.~ In all cases, initial one-electron transfer gives the superoxide ion, 02-; subsequent steps depend on the coordinated arene and may lead to dimerization or to the formation of complexed organic peroxides. With benzylic C-H bonds, however, proton abstraction by 02-gives q5-benzyls, as shown in Scheme 7 for [FeCp('l16-C6Me6)]. 51 G. A. Carriedo, N.G. Connelly, M. C. Crespo, I. C. Quarrnby, and V. Riera, J. Clzem. Soc., Chem. Commun., 1987, 1806. 52 J. R. Hamon and D. Astruc, Organomefaffics,1988, 7, 1036. Connelly + 02-Scheme 7 [WMe2Cp2] + [CPh3]+ -[WMe2Cp2]+ + CPh3 [W(CH2)MeCp2] + + HCPh3 In the second example, C-H bonds (particularly those of metal alkyls) are activated using the trityl cation [CPh3] +.This ion is commonly used to abstract hydride ion from coordinated hydrocarbons but it is now apparent that it can also induce C-H bond cleavage via an oxidative EC mechanism. The reaction of [CPh3]+ with [WMe2Cp2] is shown in Scheme 8, with electron loss followed by a-hydrogen atom abstraction by the trityl radical; the EC mechanism is simply proved in that [WMe2Cp2]+ is isolable and reacts with CPh3 to give the same final product, namely the cationic ethylenehydride complex CWH(C2H4)Cpzl +.53 In this particular example the initial electron-transfer step is thermodynamically favoured.However, the same type of reaction also occurs 54 between [CPh3] + and [ReR(NO)(PPh3)Cp] (R = CH2Ph, CH2CHMe2, CHMePh, etc.), to give carbene or alkene cations such as [Re(CHPh)(NO)(PPh3)Cp] and [Re(q2- + Me2C=CH,)(NO)(PPh3)Cp] +.Here, electron-transfer is thermodynamically un- favourable by ca. 0.1 to 0.2 V, but the overall reaction is driven to completion by the very rapid, and irreversible, hydrogen atom abstraction. It should be noted that the EC mechanism does not always operate with the trityl cation. Where the difference in redox potential is too unfavourable, for example with [FeR-(CO) Cp], p-h ydride abstraction occurs rather than electron- transfer. 53 J.C. Hayes and N. J. Cooper, J. Am. Chew.Soc., 1982,104,5570. 54 G. S. Bodner, J. A. Gladysz, M. F. Nielsen, and V. D. Parker, J. Am. Chem. SOC.,1987, 109, 1757.''R. S. Bly, R. K. Bly, M. M. Hossain, G. S. Silverman, and E. Wallace, Tetrahedron, 1986,42, 1093. Synthetic Applications of Organotransition-metal Redox Reactions The use of [CPh3]' as an oxidant can also lead to the activation of the substituted cyclohexadiene ring of [Fe(q4-C6H7R)(q6-arene)].At room tempera- ture, the exo-R group is cleaved but at ca. -4O"C, the cation [Fe(q"-C6H7R)(q6-arene)]' is more stable and the CPh3 radical is able to abstract the endo-hydrogen atom to give [Fe(q 5-C6H6R)(q6-arene)] '.Subsequent nucleo- philic addition of R' -to the cation gives difunctionalized cyclohexadienes [Fe(q4-C6H6RR')(q6-arene)].56 The third oxidative route to C-H bond cleavage was after the reaction of [CPh3] ' with [Ru2(p-CO)(p-CH2)(p-dppm)Cp2](13), which had been designed to give [Ru2(p-CO)(p-CH)(p-dppm)Cpz]' (14') uia hydride abstraction, resulted only in poor yields of the carbyne cation in a mixture of products. The realization that [CPhJ]' might act as an oxidant prompted a cyclic voltammetric study of (13) the results of which are shown in Figure 5. Clearly an EEC, rather than EC, mechanism is operative in that the first wave (Eo = 0.24 V) is fully reversible (Figure 5b) whereas the second [(Ep),, = 0.65 V] is accompanied by a reversible product wave at 1.16 V.On the basis of the electrochemical results, (13) was treated with two equivalents of [FeCp2]' as oxidant, in the presence of an excess of lutidine as a proton scavenger, to give near quantitative yields of (14'). Figure 5c confirms that (14') is oxidized at 1.15 V, i.e. at the potential of the product wave in Figure 5a. Complex (14') is therefore formed (Scheme 9) by proton loss after double oxidatim of (13) to (132'). No C-H bond cleavage occurs after one-electron oxidation; the paramagnetic iron cation [Fe2(p-CO)(p-CH2)(p-dppm)Cp2]+ (15+) has been isolated and fully characterized by X-ray crystallography 59 and it does not lose H- on treatment with CPh3.Hydride abstraction or an EC mech- anism of the type shown in Scheme 8 has been assumed6' for the formation of +[Fe2(p-CO)(p-CH)(p-dppm)Cp~l(16') from [F~~(cI-CO>(~-CH~>(CL-~~~~)-Cp2] (15) but neither is correct; complex (15') gives (16') only when converted into (1 52 ') with a second equivalent of oxidant. The double-oxidation deprotonation (EEC) reaction can be used with a range of other hydrocarbon-bridged complexes,6 leading, for example, to the isolation of [RU~(~~-C=CH~)(CO)~CP~] It also results + from [Ru~(~~-CM~)(CO)~CP~]." in C-Me carbon-hydrogen bond activation in the metallacyclononatetraene s6 D. Mandon, L. Toupet, and D. Astruc, J. Am. Chem. Soc., 1986,108,1320. 57 N. G. Connelly, N. J. Forrow, B.P. Gracey, S. A. R. Knox, and A. G. Orpen, J. Chem. Soc., Chem. Commun., 1985, 14. 58N.G. Connelly, N. J. Forrow, S. A. R. Knox, K. A. Macpherson, and A. G. Orpen, J. Chem. Soc., Chem. Commun., 1985,16. 59 F. J. Collins, S. A. R. Knox, D. A. V. Morton, and A. G.Orpen, unpublished results. "See ref. 29 in B. E. Bursten and R. H. Cayton, J. Am. Chem. Soc., 1986, 108, 8241; C. P. Casey, M. Crocker, P. C. Vosejpka, P. J. Fagan, S. R. Marder, and M. A. Gohdes, Organometallics, 1988,7,670. 6' See, for example, N. M. Doherty, M. J. Fildes, N. J. Forrow, S. A. R. Knox, K. A. Macpherson, and A. G. Orpen, J. Chem. Soc., Chem. Commun., 1986, 1355; F. M. Collins, Ph.D. Thesis, University of Bristol, 1988. Connelly 0.0 0.5 1.0 1.5 V Figure 5 Cyclic uoftammograms of [Ru2(11-CH,)(~-CO)(~-dppm)Cp~l(13) (a) from 0.0 to 1.4 V and (b)from 0.0 to 0.5 V, and (c) of [Ru2(~-CH)(~-CO)(~-dppm)Cp~1+(14') from 0.0 to 1.4 V,in CH2C12 (Reproduced fromJ.Chem. SOC.,Chem. Commun., 1985,14) complex [Mo2(p-C8Me8)Cp2] (17); two equivalents of [FeCpz] give [M02(p- + C8Me7CH2)Cp2]+ (18).62 B. Migratory Insertions.-Probably the first example of oxidatively-induced C-C bond formation involved the synthesis of acetic acid esters, RC02Me, from [FeR(C0)2Cp] or [MoR(CO)~C~]and a strong oxidant (Ce4+ or Cu2+) in methan01.~ The mechanism of this reaction has been elucidated largely via chemical and electrochemical studies of the oxidation of species such as [FeMe(CO)(PPh3)Cp] (19). Once again, this topic has been recently reviewed 63 and will only be treated briefly here.62 S. G. Bott, N. G. Connelly, M. Green, N. C. Norman, A. G. Orpen, J. F. Paxton, and C. J. Schaverien, J. Chem. SOC., Chem. Commun., 1983,378. 63 M. N. Golovin, R. Meirowitz, M. M. Rahman, H. Y. Liu, A. Prock, and W. P. Giering, Organometallics, 1987,6, 2285; M. J. Therien and W. C. Trogler, J. Am. Chem. SOC.,1987, 109, 5127. Synthetic Applications of Organotransition-metal Redox Reactions /? l+13') P-P = dppm Scheme 9 -l+ I Moa II The neutral compound (19) only slowly undergoes migratory insertion with CO to give [Fe(COMe)(CO)(PPh3)Cp] (20) but the reaction is catalysed by one- electron oxidants such as [FeCpz]+ or Ag'. Scheme 10 shows the salient features of the redox-based mechanism.Initially, the formation of (20+) from (19') was thought to occur via CO addition to [Fe(COMe)(PPhS)Cp]+ (21+), a 15-electron intermediate. However, it now seems more likely that solvation of (19') gives the 19-electron complex [FeMe(S)(CO)(PPh,)Cp]+ (22'; S = solv-ent) which then rearranges to [Fe(COMe)(S)(PPh,)Cp] + (23 +); CO displacement of the solvent then yields (20'). C. C-C Bond Formation via Hydrocarbon Dimerizatiom-(i) General Considera- tions. The dimerization of a coordinated n-hydrocarbon can be induced by one- electron oxidation or reduction, as outlined in Schemes 11 and 12 for the simple, as yet hypothetical, examples of o-vinyl and q2-ethene complexes. Oxidation (Scheme 11) generates a 17-electron metal centre but a o-n Connelly [FeMe(CO)LCp] [FeMe(CO)LCp] + (19) (I?+)I1 -1 + +[FeMe(S)(CO)LCp] [Fe(COMe)LCp] (22 + ) 4, -L [Fe(COMe)(S)LCp] 3[Fe(COMe)(CO)LCp] ++ (23’) (20+ 1 L = PPh3,S = solvent Scheme 10 H /M-C -e -+M-C \\FH / \\F-H H H H H \C’ +M ‘M-II H-C M =17-electron fragment Scheme 11 rearrangement of the vinyl ligand allows the metal to return to the stable closed- shell configuration while the unpaired electron is placed on the a-carbon atom.Radical-radical coupling then gives the diamagnetic q’:q”-butadiene product. In Scheme 12, a 19-electron species is formed by metal-centred reduction. Then, a n--(r rearrangement can again regenerate an 18-electron configuration and provide the radical site for dimerization. Although these examples are hypothetical they are conceptually useful and helpful in rationalizing the oxidative and reductive dimerizations described in parts (ii) and (iii) below and the more complex redox reactions of coordinated cyclooctatetraene (cot) discussed in part (iv).(ii) Oxidative Coupling. The o-ally1 complexes [FeR(CO)zCp] (R = CH2CR’=CR2R3) react with AgPF6 to give binuclear q’ :q1’-hexa-1,5-diene complexes (Scheme 13) and a similar reaction occurs with [Fe(CH2C-CPh)- (CO),Cp] to give the q’ :q’’ complex of 3,4-diphenylhexa-1,2,4,5-tetraene.In Synthetic Applications oj Organotransition-metal Redox Reactions H‘ H. HH H M=l7-eIectron fragment Scheme 12 M7F’F/”” -e- ~ R3+‘7/c=c I\ R’ R2 R’ R2 M= fe (CO)2 Cp Scheme 13 both cases, the proposed mechanism is very similar to that of Scheme 11 but the reactions are also regiospecific in that coupling occurs between the more highly substituted termini of the three-carbon fragments.64 The iodosobenzene oxidation of the carbene cation [Fe(C=CHMe)(dppe)Cp] + similarly results (Scheme 14) in regiospecific dimerization giving the 2,3-dimethylbuta- 1,3-diene- 1,4-diylidene complex (24).Here, an oxidative deprotona- tion step of the type described in Section 5A is thought to precede the formation of [Fe(C=CMe)(dppe)Cp] which subsequently rearranges and then dimeri~es.~~ + The 17-electron alkyne radical cation intermediate was not detected during the 64 P.S. Waterman and W. P. Giering, J. Organomel. Chrm., 1978, 155, C47. 65 R. S. Iyer and J. P. Selegue, J. Am. Chem. Soc., 1987,109,910. Connelly CHPh c? +/\ L,OS, cc4 Ph (25)L = PMe, reaction but the analogous cycloheptatrienyl complexes [Mo(C-=CR)(dppe)(q 7- C7H7)]'(R = Ph, Bun, etc.) have been isolated and fully characterized. Moreover, these paramagnetic cations slowly dimerize to diamagnetic divinylidene dications, [MO~(CL-C~R~)(~PP~)~(~)7-C7H7)212+ , which are structurally similar to (24). As yet this molybdenum chemistry provides the only example where a paramagnetic monomer and the corresponding diamagnetic dimer are fully characterized.66 Although not a dimerization reaction, the formation of (25) from Ag' and [OS(C=CP~)~(PM~~)~]may also proceed via a mechanism related to that in Scheme 1 1, intramolecular C-C coupling following one-electron oxidation and rearrangement.6 (iii) Reductive Coupling.Although the reductive coupling of simple 7)2-alkenes (Scheme 12) has not been observed, the alkyne cations [Mo(NCMe)-(Mec~CMe)~(q5-L)] (26') react with magnesium amalgam or [Fe(C0)2Cp] -to give metallacyclononatetraene, [Mo2(p-C8Me8)L2] (1 7; L = Cp), or flyover, [Mo2(p-C6Me6)L2] (27; L = indenyl), complexes; 68 electrochemical studies show that the radicals (26) are formed in the first step.69 The reductive dimerization of cationic cyclic q5-dienyls has been long known 66 R. L. Beddoes, C. Bitcon, A. Ricalton, and M. W. Whiteley, J. Organornet. Chern., in press."J. Gotzig, H. Otto, and H. Werner, J. Organomet. Chem., 1985,287,247.''M. Green, Polyhedron, 1986,5427. 69 D. Pufahl, W. E. Geiger, and N. G. Connelly, Organornetullics. 1989,8,412. Synthetic Applications of Organotransition-metal Redox Reactions to give q4:q14-bidienes but recent studies of the cycloheptadienyl and substi- tuted cyclohexadienyl complexes [Co(q 5-C7Hq)Cp] + 70 and [Fe(C0)3(q 5-C6H6C02Me)]+ '' have finally provided good evidence for radical intermedi- ates. In addition, zinc reduction of the diallylamide complex (28) gives the lactam (29) (Scheme 15) in the first example of the intramolecular cross-coupling of a redox-generated radi~al.~ (iv) Redox-induced C-C Bond Formation and Cleavage in Cyclooctatetraene Complex Chemistry.One of the earliest 72 and subsequently most thoroughly Table 1 Cq'clic voltantmetric data,for the one-electron osiciution uf[ML,(q4-cot)] 'O W. E. Geiger, T. Gennett, G. A. Lane, A. Salzer, and A. L. Rheingold, Organomrtallics, 1986,s. 1352. "A. J. Pearson, Y.3. Chen, M. L. Daroux, A. A. Tanaka, and M. Zettler. J. Chem. Soc., Chem. Commun., 1987. 155.''N. G. Connelly, M. D. Kitchen, R. F. D. Stansfield, S. M. Whiting, and P. Woodward. J. Organomel. Chem., 1978, 155, (234. Connelly studied 73 examples of redox-induced C-C bond formation and cleavage is based on the oxidation of q4-cot complexes of Fe,74-77 RU,’~ CO,~~ and RhgO and the electron-transfer reactions of the resulting binuclear, C-C coupled products. All of these reactions, which are generally regio- and stereo-specific, can be rationalized in terms of the 18-17-18-electron and 18-19-18-electron cycles outlined in Schemes 11 and 12.All of the q4-cot complexes studied (Table 1) undergo irreversible one-electron oxidation at a platinum electrode, at potentials depen- dent on the metal and ancillary ligands. All undergo chemical oxidation, with oxidants chosen on the basis of the measured (Ep),, values in Table 1, to give dimeric dications via radical-radical coupling. The chemical oxidation of [M(CO)3(q4-cot)], with Ag+ or [N(C6&Br-p)3] + (M = Fe) or with [FeCp2]+ (M = Ru) gives high yields of the isomeric Complexes [Fez(C0)6(q5 :Tl’5-C16H16)]2+ (31) and [RU2(C0)6(Tl2,q3: ‘Tl’2,q’3- C16H16)]2+ (32) whose mechanism of formation is shown in Scheme 16.The first common intermediate in the two dimerizations, namely [M(C0)3(q4-cot)] + (30+), has not been detected but 6oCo y-irradiation of [Fe(C0)2L(q4-cot)] (L = CO or PPh3) in CFC13 at 77K results in the observation of e.s.r. spectra not inconsistent with the formation of [Fe(CO)2L(q4-cot)] f.81 In addition, the 17- electron cation [Fe(CO)(P(0Me)3}2(q4-tpcb)]+ (tpcb = tetraphenylcyclo-butadiene), which is a stable analogue of (30+; M = Fe), has been fully struct- urally characterized.” The metal-based paramagnetic cation (30’) rearranges to the allyl-based radical [M(C0)3(C8Hg)lf (33+) which then couples to give the final common intermediate (34). Two different rearrangement pathways then result in (31) and (32), which contain the q5-bicyclo[5.1.0]octadienyl and q2,q3-cyclooctatrienyl units favoured by iron and ruthenium respectively. Although the intermediate (34) has not been detected for iron or ruthenium, just such species are isolated, as dias tereomeric mixtures, when [Co(q 4-co t)(q -C Me 5)] and [R h(q 4-co t)Cp] are oxidized.Detailed electrochemical studies of [Rh(q4-cot)Cp] have shown that one-electron transfer is followed by a$rst-order reaction which precedes dimeriza- tion. It is difficult to envisage any such reaction other than the proposed 73 N. G. Connelly, in ‘Paramagnetic Organometallic Species in Activation Selectivity, Catalysis’, ed. M. Chanon, M. Julliard, and J. C. Poite, Kluwer Academic Publishers, Holland, 1989.74 N. G. Connelly, R. L. Kelly, M. D. Kitchen, R. M. Mills, R. F. D. Stansfield, M. W. Whiteley, S. M. Whiting, and P. Woodward, J. Chem. SOC.,Dalton Trans., 1981, 1317 75 N. G. Connelly, A. R. Lucy, R. M. Mills, J. B. Sheridan, and P. Woodward, J. Chem. SOC.,Dalton Trans., 1985,699. ”N. G. Connelly, M. J. Freeman, A. G. Orpen, J. B. Sheridan, A. N. D. Symonds, and M. W. Whiteley, J. Chem. SOC.,Dalton Trans., 1985, 1027. 77 N. G. Connelly, A. R. Lucy, R. M. Mills, M. W. Whiteley, and P. Woodward, J. Chem. SOC.,Dalton Trans., 1984, 161. ’’N. G. Connelly, P. G. Graham, and J. B. Sheridan, J. Chem. SOC.,Dalton Trans., 1986, 1619. 79 M. Gilbert, Ph.D. Thesis, University of Bristol, 1988; N. G. Connelly and M.Gilbert, unpublished re-sults. L. Brammer, N. G. Connelly, J. Edwin, W. E. Geiger, A. G. Orpen, and J. B. Sheridan, Organometallics, 1988,7, 1259. ”M. C. R. Symons and N. G. Connelly, unpublished results. “A. G. Orpen, N. G. Connelly, M. W. Whiteley, and P. Woodward, J. Chem. Soc., Dalton Trans., in press. Synthetic Applications of‘ Organotransition-metal Redox Reactions M (30) [31 ;M= Fe (CO)3 1 [32; M=Ru(CO)~I Scheme 16 isomerization of the metal-based 17-electron cation [Rh(q4-cot)Cp] + to an ally1 radical analogous to (33+). The dication [Rh2(q5:q’5-C16H16)CP2]2f(35) differs from (31) and (32) (and from [co2(q5:q’5-C16H16)(q-C5Me5)2J2+, see below) in slowly isomerizing thermally (Scheme 17), first to an q2,q3-bonded complex analogous to the ruthenium dimer (32) and then to the asymmetric species (36) which can only form from one of the two possible diastereomers shown.” Both (31) and [CO~(~~:~’~-C~~H~~)(~-C~M~~)~]~+(37) undergo redox reac- tions which result in further C-C bond transformations. Scheme 18 shows the sequence of two-electron steps by which the C16H16 hydrocarbon fragment of (31) is converted into three other isomers. The X-ray structures of 74[Fe2(CO)4{P(OPh)3)2(q5:q‘5-C16H16)]2+[the bis(tripheny1phosphite) de-rivative of (31)] and complexes (3QS3 (39), and (40),76also shown in Scheme 18, provide an insight into the origin of the regio- and stereo-selectivity of the redox- induced transformations.A more detailed discussion is given in reference 73 but, as an example, the formation of bond C(3)-C(3’) in (38) requires only a rotation of ca.30” about bond C(l)-C(l’), following reduction of (31), to align the appropriate carbon atoms for coupling. The twofold symmetry observed for the C16H16ligands in (31) and (38) to (40) led to the assumption that the redox reactions linking the various dimers involved simultaneous one-electron transfer at the two metal centres thereby leading to diradical formation. However, the cobaltocene reduction of (37) gives 83 H. A. Bockmeulen, R. G. Holloway, A. W. Parkins, and B. R. Penfold, J. Chem. Soc., Chem. Commun., 1976, 298; B. R. Penfold, personal communication. Connelly 1la (36) Scheme 17 highly asymmetric C16H16 complexes, (41) and (42), the therrnolysis of the former leading to a third isomer (43) (Scheme 19).The formation of (41) and (42) is perhaps more consistent, therefore, with a stepwise, ECEC, mechanism, as shown for the [FeCp2]+ oxidation of (43) to the dication (44) (Scheme 20).79 A description of the redox chemistry of cot complexes would be incomplete without reference to Geiger’s recent studies of M2(p-C&8) (M = Ru, Co, or Rh) complexes (Scheme 21) where reversible C-C bond transformations are observed. The two-electron oxidation of [Rh2(p-CsHg)Cp2] (45; M = Co or Rh) leads to the conversion of the ~l~:q’~-cot bridge into the twisted q5:q’’-form in the dication (45‘ +).84 A second two-electron oxidation, of [Ru2(p-CsHs)Cp2] (46) which is isostructural and isoelectronic with (452’), then results in complete C-C bond rupture to give the flyover complex (462+).85 Cyclic voltammetry shows that both two-electron oxidations, of (45) and (46),are irreversible but chemical reduction of the dicationic complexes gives the neutral precursors in near-quantitative yield.84 J. Edwin, W. E. Geiger, and A. L. Rheingold, J. Am. Chem. SOL..,1985,106,3052. 85 J. Edwin, W. E. Geiger, A. Salzer, U. Ruppli, and A. L. Rheingold, J. Am. Chem. SOC.,1987, 109,7893. 177 Synthetic Applications of Organotransition-metal Redox Reactions (31)I+2e-‘M 1’ 1 +2e- (39) 4’ 4 5 M-WFM 6 - 1‘ 1 7 M=Fe(CO), Scheme 18 Connelly + ,m, M**M +M (37b) Scheme 19 6 Mixed-valence Organometallics Much of the effort expended on understanding and classifying the properties of mixed-valence complexes has involved studies of the oxidation of binuclear metallocenyl complexes such as biferrocene, [Fe2(p-C 10H8)Cp2], and bis-(fulvalene) diiron, [Fe2(p-C IoH~)~]; the mixed-valence derivatives of these species contain [Fe2I3+ cores.However, the recent work of Astruc on binuclear electron-rich arene complexes 86 such as [Fe2(p-L)(~l-C5Me5)~] (L = biphenyl, (R= H or Me) has triphenylene, etc.) and [F~~(~-CIOH~)(~~-C~R~)~] con-siderably extended the range of mixed-valence (and other) states (e.g. [Fe2] + and [Fe2] ’) available for further detailed study. Other interesting mixed-valence organometallics can be prepared by the one- electron oxidation of the cyanide-bridged dimanganese complexes (9+) to (12 +), as described in Section 4, and of [Cr2(CO),L2(p-biphenyl)] where both trapped- valence (L = PPh3) and delocalized (L2 = p-dppm) bonding is observed.87 This section, however, highlights attempts to synthesize [Rhz] +-containing species R6 M.Lacoste, F. Varret, L. Toupet, and D. Astruc, J. Am. Chem. Soc., 1987. 109, 6504; M. H. Desbois, D. Astruc, J. Guillin, J.-P. Mariot, and F. Varret, J. Am. Chem. Soc., 1985, 107, 5280; M. Lacoste, D. Astruc, M.-T. Garland, and F. Varret, Organometallics,1988,7, 2253. ”N. Van Order Jr., W. E. Geiger, T. E. Bitterwolf, and A. L. Rheingold, J. Am. Clzem. Soc., 1987, 109, 5680. Synthetic Applications of Organotransition-metal Redox Reactions (44) M = Co(7-C5Me5) Scheme 20 (45;M = C0,Rh) (46;M = Ru,n=O) (4S2+;M = Ru) (45”.M =Co,Rh, n=2) Scheme 21 derived from ligand-bridged square-planar Rh’ complexes related to [Rht(CO),&-RNXNR)2] (47; R = aryl, X = N or CMe). The attraction of such species derives from their potential ability to bind n-acceptor ligands (alkenes, alkynes, etc.) at the coordinatively unsaturated low oxidation state metal centres Connelly R N/X\ R R N ON\NR Ioc' Ioc' C PPh,C C 0-0 0 (4 8) (471 while at the same time undergoing photo-induced intramolecular electron-transfer to a reactive excited state. The complex [Rh2(CO)2(PPh3)2(p-RNNNR)2](48; R = p-tolyl), prepared from (43; X = N) and PPh3 by thermal substitution, is oxidized in three, reversible, one-electron steps at a platinum electrode.88 These electrochemical results, taken with the observation that [Rh2(pcarboxylate)4] and related [Rh2I4+-containing tetrabridged species can be oxidized to mono-and di-cations,89 show that the basic face-to-face binuclear structure can exist in at least five different redox states (i.e.[Rh2I2+ to [Rhz]"'). Complex (48) is readily oxldized by [N2C6H4F-p] to the paramagnetic, + [Rh2I3+-containing cation (48'). X-Ray structural studies on (48') show a shortened metal-metal interaction C2.69 8, us. 2.96 8, for (48)] consistent with electron removal from an antibonding metal-metal orbital.88 A delocalized electronic structure for the cation is suggested both by the anisotropic, frozen- solution e.s.r.spectrum of (48+)and by m.0. calculations [which also predict an increased susceptibility to axial coordination when (48) is oxidized].90 In order to introduce electronic asymmetry within the [Rh2I3 core, thereby + enhancing the possibility of photoinduced intramolecular electron-transfer, the geometric structure of the dirhodium complex as a whole has been systematically modified. Modest changes may be introduced using the acetamidino-bridged complex (47; X = CMe). Its thermal reaction with PPh3 gives only the tricarbon yl [Rh2(CO)3( PPh3){ p-RNC( Me)NR) 2], but this compound undergoes oxidative substitution with ligands, L, in the presence of [FeCpz] (cJ: Section+ 3A) to give the asymmetric, but still delocalized, complexes [Rh2(C0)2L(PPh3)- {p-RNC(Me)NR}2] + [L = AsPh3, P(OPh)3, et~.].~~ A wide range of highly asymmetric, redox-active dirhodium complexes can be synthesized using [Rh2(C0)2(bipy)(j~-RNNNR)~](49) as the precursor (Scheme 22); formed from (47; X = N) and bipy, (49) differs from the disubstituted complex (48) in containing an Rh(C0)2 group which becomes substitutionally labile in (49+)91,92 88 N.G. Connelly, G. Garcia, M. Gilbert, and J. S. Stirling, J. Chem. Soc., Dalton Trans., 1987, 1403; N. G. Connelly, C. J. Finn, M. J. Freeman, A. G. Orpen, and J. S. Stirling, J. Chem. Soc., Chem. Commun.,1984, 1025. 89 See, for example, T. R. Felthouse, Progr. Znorg. Chem., 1982, 29, 73,and references therein.90 N. G. Connelly, A. G. Orpen, and P. H. Rieger, unpublished results. 91 N. G. Connelly and G. G. Herbosa, J. Chem. Soc., Chem. Commun., 1987,246. Synthetic Applications of Orgunotransition-metal Redox Reactions -e-I-* I ,N--~-?N Rh----Rh J Scheme 22 NNN = p-tolylNNNtoly1-p, N-N = 2,2'-bip?iridyl. S-S = S2CNMe2, L = PPh3 or P(OPh),; Rh ---Rh und Rh-Rh denote one- and two-electron metal-metal bonds in [Rht13+ and [Rh2I4+ cores respectively The cyclic voltammetric data in Table 2 show that the oxidation potential for a given dirhodium redox couple can be varied by design by as much as 2.0 V. Thus, bipy complexes with the [Rh2I4' core become chemically accessible. for example (51) and (52). As predicted by the m.0. calculations noted above, an increase in the oxidation level of the dirhodium core is accompanied by an increase in the susceptibility of the axial sites to coordination, hence the iodide- bridged structure of (51).92 The anisotropic e.s.r.spectra of the [Rh2] 3+-containing bipy complexes are still consistent with delocalized electronic structures in spite of the molecular asymmetry. However, the reaction of species such as (49') with molecular oxygen leads to the observation of e.s.r. spectra consistent with dirhodium superoxide complexes containing the localized Rh'Rh"'(O2 -) arrangement.93 92 T. Brauns, C. Carriedo, J. S. Cockayne, N. G. Connelly, G. Garcia, and A. G. Orpen, J. Ckem. Soc.., Dallon Trans., in press. 93 N. G. Connelly and A. C. Loyns, unpublished results.182 Connelly Table 2 Cyclic voltammetric data" jor dirhodium complexes Complex Couple (Eo/V) + ~ +[Rhz] +/3+ [~ +I4h ~[RhzI4+/' [Rh2(CO)2(PPh3)2(c1-L)21 0.19 1.29 1.58 +CR~~(CO>Z(~~PY)(CI-L)~I -0.25 0.9(1) --0.53 0.78C~~z(~~~~~~~3~(~~py~~cI-~~2lf 1.47 -1.07 0.2 1 1.25, 1.37d CR~~I(CO)(PP~J>(~~PY)(CI-L)ZI CRh2(CO)(S,CNMe~)(bipy)(cl-L)21+ --0.84 1.13 CR~ZC~(CO)(PP~~)(~~PY)(~~-L)~I -0.56+ -1.28 a All processes are reversible unless stated otherwise. L = [RNNNRI-, R = p-tolyl. Irreversible process; potential is (EP)~~ at a scan rate of 200 mVs-'. * Potentials for the oxidations of the tetranuclear dication (51) to the [Rh4]'+- and [Rh4]'0'-containing tri- and tetra-cations respectively (These spectra were originally attributed 91 to a concentration-dependant phenom- enon but clearly result from the presence of adventitious oxygen in the more dilute samples.) Axial coordination of O2 to [Rh2(a~)~] (ap = 2-anilinopyridinato) gives the Rh"Rh"'(Oz-) adduct [Rh2(ap)4(02)],94 but in the case of (49') the reaction with oxygen appears to result in carbonyl sub~titution.~~ The sequence (49) --+ (49') -(50) +(51) -(52) (Scheme 22) is also notable in linking five well-defined, isolable complexes in an ECEC mechanism for the oxidative elimination reaction of I2 with a Rh' centre.Such a one-electron transfer mechanism was proposed in Section 2 for the formation of [Fe12(C0)3(P- Ph,)] from [Fe(C0)3(PPh3)2] and 12, although in that case all the intermediates could not be fully identified.In the present example, the reaction at one rhodium centre is mediated by the second and the paramagnetic intermediates are stabilized by metal-metal bond formation.92 Other attempts to vary the electronic properties of species with two or more square-planar rhodium centres have involved the synthesis of paramagnetic A-frame complexes such as [Rh2(p-L')L4]+ [L' = 2,3-(NH)2CloH6 (53) or 1,8-(NH)~C~OH~(54); L2 = (CO)(PPh3) or q4-diene, diene = cyclooctadiene (cod), norbornadiene, et~.],~~ and the trinuclear [Rh3I4+-containing dications [55; R = H or Me, L2 = (CO)(PPh3) or Cation (54) is notable in containing an [Rh2I3+ core prone to axial coordination; reaction with NO gives the stable, diamagnetic nitrosyl adduct [Rh2{ p-1,8-(NH)2C1oH6)(NO)(q4-cod)2] + .95 7 Conclusions and Future Prospects The examples described in Sections 2 to 6 show that redox-based syntheses, involving 17-and 19-electron radicals as key intermediates, have and will have an important role to play in metal-carbonyl and -hydrocarbon chemistry.However, other important applications of redox-active organometallics are also y4 J. L. Bear, C.-L. Yao, F. J. Capdevielle, and K. M. Kadish, Inorg. Chem., 1988,27,3782. 95N.G. Connelly, A. C. Loyns, M. J. Fernandez, J. Modrego, and L. A. Oro, J. Chem. SOC.,Dulton Trans., 1989, 683. 96 N. G. Connelly, A. C. Loyns, M. A. Ciriano, M. J. Fernandez, L. A. Oro, and B. E. Villarroya, J. Chem. SOC.,Dalton Trans., 1989, 689.Synthetic Applications of Organotransition-metal Redox Reactions HN L’ IL \LL‘ (54) (53) 0 (55) coming to light. It is perhaps ironic that ferrocene, whose one-electron oxidation reaction was discovered nearly 40 years ago, features most heavily in such applications. Thus, the couple [FeCp2]/[FeCp2] is now a IUPAC recommended + internal reference potential standard,97 and plays an integral part in a com-mercially available biosensor for the determination of blood sugar levekg8 In addition the ferrocenyl group can be used as a redox-active substituent, for example in macrocyclic ionophores where shifts in the redox potential mirror the binding of group 1A metals.99 Finally, [Fe(q-CSMes)2] and tcnq (tetracyano- quinodemethane) give the soluble molecular ferromagnet [Fe(q-CsMes)2]-[tcnq] .O O Without doubt, chemical, electrochemical, spectroscopic, and structural studies 97G. Gritzner and J. Kuta, Pure Appl. Chem., 1982, 54, 1527; R. R. Gagne, C. A. Koval, and G. C. Kisensky, Inorg. Chem., 1980, 19, 2854. 98 A. E. G. Cass, G. Davis, G. D. Francis, H. A. 0.Hill, W. J. Aston, I. J. Higgins, E. V. Plotkin, L. D. C. Scott, and A. P. F. Turner, Anal. Chem., 1984, 56, 667; J. E. Frew and H. A. 0.Hill, Anal. Chem., l987,59,933A. 99 See, for example, P. D. Beer, H. Sikanyika, C. Blackburn, and J. F. McAleer, J. Organornet. Chem., 1988,350, C15. loo J. S. Miller, A. J. Epstein, and W. M. Reiff, Chem. Rev., 1988,88,201. Connelly of redox-active organometallics will continue to provide new and exciting discoveries.Their application should ensure the health and further development of organometallic electrochemistry. Acknowledgements. First and foremost I should like to thank my post-graduate and post-doctoral co-workers at the University of Bristol whose efforts form the back-bone of this review. I also acknowledge with gratitude invaluable collabora- tions with friends and colleagues in the UK, the USA, and Spain.

ISSN:0306-0012    

DOI:10.1039/CS9891800153

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. SOC.Reu. 1989,18,187-208 CENTENARY LECTURE Chemical Multiplication of Chirality: Science and Applications By R. Noyori DEPARTMENT OF CHEMISTRY, NAGOYA UNIVERSITY, CHIKUSA, NAGOYA 464-0 1, JAPAN 1 Introduction Chirality plays a central role in science and technology. A wide range of significant physical, chemical, and biological functions are generated through precise molecular recognition which requires strict matching of chirality. For a long time access to highly enantiomerically pure compounds, at least in a practical sense, was thought to be Nature’s monopoly and has indeed been accomplished by biological or biochemical transformations. Efficient creation of optically active organic molecules from prochiral compounds by chemical means, though it is challenging, has remained difficult, and only optical resolution and structural modification of naturally occurring chiral substances have provided complements in this respect.However, assiduous efforts made by synthetic organic chemists in the last two decades are converting the chemists’ dream into reality. In order to maximize synthetic efficiency, ‘multiplication of chirality’, namely, stereoselective production of a large quantity of a chiral target compound utilizing a catalytic amount of chiral source, is obviously desirable. Enantioselective catalysis using chiral metal complexes, among various possib- ilities, provides one of the most general, flexible methods for this purpose.’ Metallic elements possess a variety of catalytic activities, and permutation of organic ligands or auxiliaries directing the steric course of the reaction is practically unlimited.Accordingly, in principle, one can generate any dynamic properties at will through molecular architecture using accumulated chemical knowledge. To this end, creation of a single, highly reactive catalyic species possessing excellent chiral recognition ability is required. Besides the choice of central metals therefore, molecular design of the chiral modifiers is a particularly significant task. The efficient ligands must be endowed with a suitable function- ality; an appropriate element of symmetry; substituents capable of differentiating space either sterically or electronically; skeletal rigidity or flexibility (depending on the nature of the reaction), etc.-all of which contribute to accomplish highly enantioselective catalyses.* ’ For the present state of this subject, see R.Noyori and M. Kitamura, in ‘Modern Synthetic Methods 1989’, ed. R. Scheffold, Springer L’,rlag, Berlin, p. 115.’R. Noyori and H. Takaya, Chem. Scr., 1985,25,83. Centenary Lecture _____) %COOCzH5Cu* cat + N,CHCOOC,H, cis and trans Cu‘ cat: / +s-N+NH H5NHJ-?;;:””” COONa imipenem cilastatin Scheme 1 2 Discovery and Opportunities To our best knowledge, the first example of asymmetric synthesis from prochiral compounds catalysed by homogeneous chiral metal complexes appeared in the literature in 1966.3 A chiral Schiff base-Cu” complex was found to catalyse decomposition of ethyl diazoacetate in styrene to give cis-and trans-2-phenylcyclo- propanecarboxylates in <lo% e.e., proving the existence of a reactive Cu carbenoid placed in a chiral environment.The intermediary carbenoid was also trapped by racemic 2-phenyloxetane leading to optically active furan derivatives. Later extensive, systematic screening of the chiral Schiff bases resulted in a dramatic improvement of the optical yield of the cyclopropanation, allowing asymmetric synthesis of chrysanthemic acid derivatives in up to 94% e.e.4 This chemistry has been successfully applied to industrial synthesis of (S)-2,2-dimethylcyclopropanecarboxylic acid, a component of cilastatin which serves as an excellent inhibitor of dehydropeptidase-I increasing in uiuo stability of antibiotic imipenem (Sumitomo Chemical Co., Japan, and Merck Sharp & Dohme Co., USA) (Scheme 1).Among other asymmetric catalyses working in industry at this moment, perhaps the largest is a process involved in the synthesis of (-)-menthol (a) H. Nozaki, S. Moriuti, H. Takaya, and R. Noyori, Tetrahedron Lett., 1966, 5239; (b) H. Nozaki, H. Takaya, S. Moriuti, and R. Noyori, Tetrahedron, 1968,24,3655. T. Aratani, Pure Appl. Chem., 1985,57, 1839. Noyori R R = (CH,),C=CHCH,CH, (1) (Takasaga International Co., Japan). The key step is the Rh-BINAP catalysed enantioselective isomerization of diet hylgeranylamine to citronella1 diethylenam- ine proceeding in 9699% optical yield.6 The optical purity of the synthetic citronellol is much higher than that of the natural product; ca.80%. The technical refinement has led to an innovative catalytic process working on up to a 7 ton scale. Here, use of atropisomeric BINAP ligand has played a key role in the successful asymmetric catalysis. The fully aromatic diphosphine having an axial element of chirality was first prepared by optical resolution of the racemate through an optically active amine-Pd” c~mplex,~~’~ but is now obtainable more conveniently by resolution of its dioxide, BINAPO, with camphorsulphonic acid or 0-dibenzoyltartaric acid followed by reduction with trichl~rosilane.~A number of BINAP analogues can be prepared in such a way (Scheme 2). Olefinic double bonds are known to shift via a metal hydride addition- elimination mechanism or a x-allylmetal hydride pathway. However, the allylam- ine to enamine isomerization was revealed to occur uia a new, nitrogen-triggered mechanism (Scheme 3).* The nitrogen-coordinated allylamine-Rh complex+ causes four-centred hydride elimination from C(1) to generate a transient iminium-RhH complex.Delivery of the hydrogen from Rh to C(3) gives the enamine q2-and then q3-complexes. The latter, having an aza-ally1 structure, serves as the chain carrier in the catalytic cycle. The overall 1,3-hydrogen shift in the geranylamine occurs in a suprafacial manner from its s-trans-type conformer, as proved by the deuterium labelling experiments. The cationic Rh-BINAP complexes differentiate efficiently between pro-R and pro-S hydrogens at C( 1) through interaction with the adjacent nitrogen atom (Scheme 4).A transition-(a) A. Miyashita, A. Yasuda, H. Takaya, K. Toriumi, T. Ito, T. Souchi, and R. Noyori, J. Am. Chem. Soc., 1980, 102, 7932; (b)A. Miyashita, H. Takaya, T. Souchi, and R. Noyori, Tetrahedron, 1984, 40, 1245; (c) K. Toriumi, T. Ito, H. Takaya, T. Souchi, and R. Noyori, Acta Crystallogr., Sect. B, 1982, 38, 807; (d)S. Inoue, M. Osada, K. Koyano, H. Takaya, and R. Noyori, Chem. Lett., 1985,1007. (a) K. Tani, T. Yamagata, S. Otsuka, S. Akutagawa, H. Kumobayashi, T. Taketomi, H. Takaya, A. Miyashita, and R. Noyori, J. Chem. Soc., Chem. Commun., 1982, 600;(b) K. Tani, T. Yamagata, S. Akutagawa, H. Kumobayashi, T.Taketomi, H. Takaya, A, Miyashita, R. Noyori, and S. Otsuka, J. Am. Chem. Soc., 1984,106,5208.’(a) H. Takaya, K. Mashima, K. Koyano, M. Yagi, H. Kumobayashi, T. Taketomi, S. Akutagawa, and R. Noyori, J. Org. Chem., 1986, 51, 629; (b) H. Takaya, S. Akutagawa, and R. Noyori, Org. Synth., 1988,67, 20. H. Takaya, K. Tani, S. Otsuka, S. Inoue, T. Sato, and R. Noyori, to be published. Centenary Lecture uN (C2H5)2 diethylgeranylamine (s)-B INA P- Rh+ U C H O R, 96-99% ee citronella1 menthol (R)-BIN A P (S)-B INAP (S)-BI NAP-Rh' complex Scheme 2 state model in the Rh-(S)-BINAP catalysed reaction is illustrated by structure (1) (naphthalene rings are omitted for clarity). In principle, any donor groups including olefinic bond, heteroatom bases, carbanions, heteroanions, etc.are able to activate their adjacent C-H bonds through coordination to appropriate unsaturated transition metal centres. The resulting metal hydride complexes, depending on the situations, undergo unique chemical transformations. When racemic 4-hydroxy-2-cyclopentenone was exposed to 0.5 mol% of a cationic Rh-(R)-BINAP complex5 in THF at OOC, double bond isomerization occurred with 5 : 1 enantiomer discrimination to afford unreacted (R)-hydroxy enone in 91% e.e. in 27% yield and 1,3-cyclo-pentanedione in 61% yield (Scheme 5).9 3 Ruthenium-catalysed Asymmetric Hydrogenation Homogeneous asymmetric hydrogenation, discovered in 1968,' has been one of the most exciting subjects in organic chemistry in the last two decades and a M.Kitamura, K. Manabe. R. Noyori, and H. Takaya, Tetrahedron Lett.. 1987,28,4719. lo (a)L. Horner, H. Siegel, and H. Biithe, Angeiv. Chem., Inf. Ed. Engl., 1968, 7,942; (6) W. S. Knowles and M. Sabacky, J. Chem. SOC.,Chem. Commun., 1968,1445. Noyori m 191 Centenary Lecture (S)-BINAP-Rh' (R)-BINAP-R h'1 I Scheme 4 0 0 0 (R)-BINAP-Rh' HO HO 0 Scheme 5 number of impressive chemistries have been presented.' ' In addition, the catalysis is of practical significance. (S)-DOPA, a drug for the treatment of Parkinson's disease, has been prepared at Monsanto Co., USA," and VEB Isis-Chemie, DDR, by using hydrogenation of a (2)-(a-acetamido)cinnamic ester with soluble Rh complex catalysts possessing a chiral phosphine or phosphinite ligand. The same method was used for commercial production of (S)-phenylalani-ne, a component of the non-nutritive sweetener Aspartame (Anic S.p.A., Italy).' Thus a variety of natural and unnatural amino acids are now available in >90% e.e.by enantioselective hydrogenation but, unfortunately, the scope of the Rh- catalysed reaction is not very wide. For example, [Rh(binap)(CH30H)2]CI04 caused hydrogenation of dehydroamino acid derivatives (Scheme 6) with nearly perfect enantiosele~tivities,~~*~whereas optical yields of the reactions of geraniol or nerol with varying conditions did not exceed 70%.5dIn view of the general importance of hydrogenation in organic synthesis, we have been intrigued by the possibility of developing a catalyst system capable of adopting a wide range of olefinic substrates.In this context, recent invention of Ru-BINAP dicarboxylate complexes l4 extended the utility of asymmetric hydrogenation to a great extent (Figure 1). The Ru dicarboxylate complexes undergo ligand exchange reaction with a,@-or P,y-unsaturated carboxylic acids, resulting in highly enantioselective (80-100%) hydrogenation. ' Thus, with many substrates, the highest enantioselectivities 'I Pertinent reviews: (a) J. Halpern, in 'Asymmetric Synthesis', Vol. 5, ed. J. D. Morrison, Academic Press, New York,1985, Chapter 2; (6) K. E. Koenig, ibid., Chapter 3. l2 W. S. Knowles, J. Chem. Educ., 1986,63,222. l3 H. B. Kagan, Bull.Chem. SOC.Fr., 1988,846. l4 T. Ohta, H. Takaya, and R. Noyori, Inorg. Chem., 1988,27,566. T. Ohta, H. Takaya, M. Kitamura, K. Nagai, and R. Noyori, J. Org. Chem., 1987,52,3174. Noyori wCOOR’ R-fCOOR’ + H2 (S)-BIN A P-R h Rz NHCOR~ NHCOR~ R = Ar or H R\/(R)-BINAP-Rh ’ + “2 (S)-BIN A P-R h+ COOR’ NHCOR~ NHCOR~ R = Ar S Scheme 6 A-( R)-BINAP-R u A-(S)-Bl NAP-R u dicarboxylate dicarboxylate Figure 1 have been recorded. Methyl esters are inert to the hydrogenation. Alcohols are the solvents of choice. The sense and extent of the asymmetric induction are highly dependent on the substitution pattern of the substrates and reaction conditions, particularly the hydrogen pressure. Anti-inflammatory (S)-naproxen was prepared in 97% e.e.under a high-pressure condition. This method is also applicable to synthesis of a 1 P-methylcarbapenem precursor and some optically active methylated y-and 6-lactones (Scheme 7). Olefins containing certain neutral donor functionalities are also hydrogen- ated in a satisfactory manner.16 The Ru-BINAP catalysed hydrogenation of N-acyl-(2)-1-benzylidene-1,2,3,4-tetrahydroisoquinolinesin a mixture of etha-nol and dichloromethane leads consistently to (1R)-or (1s)-benzyltetra-hydroisoquinolines in nearly quantitative yield and in 95-100% e.e.” With Rh complexes such as [R h( binap)(cod)] C104 or [Rh( binap)( CH 3 OH)2]C104, the hydrogenation proceeded in lower optical yield (ca. 75%) and with opposite enantioselection.The asymmetric hydrogenation followed by removal or modifica- tion of the N-acyl groups gave tetrahydropapaverine, laudanosine, norreticuline (biogenetic precursor of morphine), tretoquinol (bronchodilating agent), efc. l6 For a review on stereoselective olefin hydrogenation directed by functional groups, see J. M. Brown, Angew. Chem.,Int. Ed. Engl., 1987,26, 190.’’ R. Noyori, M. Ohta, Yi Hsiao, M. Kitamura, T. Ohta, and H. Takaya, J. Am. Chem. Soc., 1986, 108, 71 17. Centenary Lecture "2 EINAP-RU *R3kc00HR2 R' R2 R' t-C, H (C H J2Si0 CH,O mcooHWCOOH nap roxe n intermediate of 1p-methylcarbapenem b Scheme 7 which became homochiral by single recrystallization. The reaction of the simple 1-methylene substrate affords, after deacylation, salsolidine in 96% e.e.This procedure is applicable to the synthesis of natural morphine, various benzo- morphan analogues such as metazocine and pentazocine, morphinans including dextromethorphan (anticough agent), etc. (Figure 2).l8 This discovery has thus realized a general asymmetric synthesis of isoquinoline alkaloids.' The Ru-catalysed hydrogenation of prochiral allylic alcohols exhibits un-precendented efficacy. Thus geraniol and nerol are hydrogenated in methanol containing a Ru-BINAP dicarboxylate complex to give (S)-or (R)-citronellol in 9699% e.e.'O Initial hydrogen pressure higher than 30 atm gave satisfactory results. Either natural or unnatural forms can be made flexibly by changing the chirality of the catalyst or geometry of the olefinic substrates.The enantiomeric purity of the synthetic citronellol exceeds the highest value of the natural product, 92%. The substrate/catalyst mole ratio is extremely high and, in certain cases, the efficiency of the chiral multiplication, defined as [major enantiomer -minor enantiomer] (in mole)/chiral source (in mole), approaches 48,500! Notably, in this hydrogenation, only allylic C(2)-C(3) double bonds are saturated and nonallylic C(6)-C(7) double bonds remain intact. Homoger- aniol was hydrogenated in 92% optical yield with the same enantioselection but the bis-homologue was inert to the standard reaction conditions. This hydrogenation is usable for the stereoselective synthesis of side chains of vitamin E and K1 (Scheme 8).M. Kitamura, Y. Hsiao, R. Noyori, and H. Takaya. Tetrahedron Lett., 1987,28,4829. l9 For synthesis cia stoicheiometric enantioselective alkylation, see A. I. Meyers, Aldrichimica Actu, 1985, 18, 59. 20 H. Takaya, T. Ohta, N. Sayo, H. Kumobayashi, S. Akutagawa, S. Inoue, I. Kasahara, and R. Noyori, J. Am. Chern. Soc., 1987,109, 1596,4129. Noyori::;“;aOCH, Ru(OCOR),( binap) t cH30qNHCH3O + Hz \ OCH, OCH, R = H, CH,, C,H, (R)-BINAP : 1R isomer (S)-BINAP : 1s isomer cH30wNHCH,O laocH3OCH, OCH, OCH, tetrahydropapaverine (S)-la u da n o s i n e tretoquinol CH30mNHcH30mNHCH,OcH30qNH mf OCH, -yocHzc6H5OCH, (S)-sal so Iid i ne ( R)-no r r e t ic u Ii ne R = H or OCHzC,H5 intermediate of morphine RO& morphine morphinans benzomorphans R = R’ = CH3: R = H; R’ = CH3: dextromet horphan metazocine R = H; R’ = CH,CH=C(CH&: pentazocine Figure 2 Chiral allylic secondary alcohols can be resolved efficiently by homogeneous hydrogenation catalysed by the Ru-BINAP diacetate complexes.2 The combined effects of intramolecular and intermolecular asymmetric induction give up to 76: 1 differentiation between the enantiomeric cyclic unsaturated alcohols.For instance, when racemic 3-methyl-2-cyclohexenol (Figure 3) was hydrogenated with the Ru-(R)-BINAP complex in methanol, at 46%conversion, R,R-configurated trans-3-methylcyclohexenol wac cbtained in 95% e.e.At 54% conversion, the slow-reacting S M. Kitamura, 1.Kasahara, K. Manabe, R.Noyori, and H. Takaya,J.Org. Chem., 1988.53,708 195 Centenary Lecture + HZ (S)-BINAP-Ru = d o “ (R)-cit ro ne Ilo I\/ ‘OH (S)-cit ronel lo I vitamin E 0 vitamin K, Scheme 8 6, b enantiomer was recovered in >99%e.e. A significant application includes a practical resolution of 4-hydroxy-2-cyclopentenone (Figure 3), an important building block for the three-component coupling prostaglandin synthesis.22 Homogeneous asymmetric hydrogenation of ketones has remained far less fruitful than the catalysis of olefinic substrates. Now, however, a variety of functionalized ketones can be hydrogenated with synthetically useful enantio- selectivities and in a predictable manner with the aid of RuX2(binap) [empirical formula; X = CI, Br, I; prepared by mixing Ru(OCOCH3)2(binap) and HX in a 22 (a) R.Noyori and M. Suzuki, Angew. Chem., Int. Ed. Engl., 1984. 23, 847: (b) M. Suzuki, A. Yanagisawa, and R. Noyori, J. Am. Chem. SOL..,1988, 110, 4718; (L.) Y. Motita, M. Suzuki, and R. Noyori, J. Org. Chem., 1989,54, 1785. Noyori H2 0 H2R)-BINAP-Ru J,c,x (S)-BIN A P-R u 9H R * RAC/X X, Y = heteroatom C = sp2 or nonstereogenic sp3 carbon % ee of Droduct % ee of product 00 96 /JIKN(cH3)2 LNuSC2H5 93 (c 312 93-96 R R = CH,, (CH,)zCH, (CH3)3C* C6H5 94-1 00 LOH 92RJJR R = CH, or C,H, 98 92 Scheme 9 1: 1 mole ratio] 23*24 or R~2Cl~(binap)~(C~H~)~N.~~The general sense of the asymmetric induction indicates that the key factor in the enantioface differentia- tion is the simultaneous coordination of the carbonyl oxygen and heteroatom, X or Y, to the Ru atom forming a five- and six-membered chelate ring, respectively.Some nitrogen- and oxygen-containing directive groups include dialkylamino, hydroxyl, siloxyl, keto, alkoxycarbonyl, alkylthiocarbonyl, dialkylaminocarbonyl, carboxyl, etc.22 To our surprise, halogen atoms were revealed to facilitate the carbonyl hydrogenation and to direct the stereochemical outcome. Thus o-bromoacetophenone gave the corresponding alcohol in 92% e.e. and 97% 23 R. Noyori, T. Ohkurna, M. Kitarnura, H. Takaya, N. Sayo, H. Kurnobayashi, and S. Akutagawa, J. Am. Chem. SOC.,1987,109,5856. 24 M. Kitarnura, T. Ohkurna, S. Inoue, N.Sayo, H. Kumobayashi, S. Akutagawa, T. Ohta, H. Takaya, and R. Noyori, J. Am. Chem. SOC.,1988,110,629. 25 T. Ikariya, Y. Ishii, H. Kawano, T. Ardi, M. Saburi, S. Yoshikawa, and S. Akutagawa, J. Chem. Soc., Chem. Commun.. 1985,922. Centenary Lecture RuX2[(R)-bi nap] RU O R e + H2 R uX,[( S)-b< Rin ap] R R = CH,, C,H,, n-C4H9, i-C,H, 98-100% ee R' = CH,, C2H5, i-C,H,, t-C4H, 93-100% yield X = CI, Br, I H2R u B rz[(S)-binap] OH 0 c1aoc2H5 0C2H5 R = CH,: carnitine intermediate for R=H: GABOB compactin Scheme 10 yield, although unsubstituted acetophenone and the rn-or p-bromo derivative failed to be hydrogenated in a satisfactory manner (Scheme 9). This method is particularly useful for enantioselective access to P-hydroxy carboxylic esters which serve as important intermediates for natural product synthesis.A wide variety of prochiral P-keto esters having flexible structures are hydrogenated consistently in nearly quantitative yields and with extremely high (up to 100%) enantiosele~tivities.~~Esters of methyl, primary, secondary, and tertiary alcohols as well as a-alkylated and a,a-dialkylated substrates were equally employable. Thus synthetic organic chemists no longer need envy bakers' yeast in this context (Scheme 10). This procedure allowed the first efficient chemical synthesis of GABOB and (R)-carnitine, a carrier of long-chain fatty acids through the mitochondria1 membrane.26 Hydrogenation of ethyl 4-chloro- 3-oxobutanoate aided by the (S)-BTNAP catalyst under the conditions effecting the reaction of 3-oxobutanoate in 99.4% optical yield (ethanol, room temperature, 100 atm, 1-0 h) afforded the desired (R)-hydroxy chloro ester in only <70% e.e.The inefficient enantiofacial differentiation is perhaps due to the competitive directing effect of the ester group and halogen atom present in the same molecule. However, a surprising chiral efficiency was obtained by the high-temperature, short-period reaction (100 "C, <5 min) affording the R enantiomer in 97% e.e. in 97% chemical yield. The same technique has been used for the synthesis of a component of compactin, an HMG-CoA reductase inhibitor (Scheme 10). Double stereodifferentiation is a powerful mechanism to enhance a degree of 26 M.Kitamura, T. Ohkuma, H. Takaya, and R. Noyori, Terrahedron Lef?..1988, 29, 1555. Noyori (R)-BINAP-Ru OHOH-&+m+H, 100% ee dllmeso = 99:l OH0 (R)-BINAP-RU + R+ OCZH5+ RAoczH5 RWOCzH5 + H2 NHBoc NHBoc NHBoc R = (CH,),CHCH,, C,H5CH2, threolerythro >99:1 cyclohexylmethyl, etc. statine Scheme 11 stereo~election.~’When prochiral, symmetrical P-diketones were subjected to the Ru catalysed hydrogenation, mixtures of dl- and meso-1,3-diols were formed (Scheme 11). The dl-isomers were dominant and their e.e.s were uniformly high. For instance, the reaction of 2,4-pentanedione catalysed by the (R)-BINAP catalyst proceeded by way of (R)-4-hydroxy-2-pentanone in 98.5% e.e., but the ultimate product was a 99: 1 mixture of (R,R)-2,4-pentanediol in nearly 100% e.e.and (R,S)-2,4-~entanediol. The minor (S)-hydroxy ketone intermediate was washed away by intramolecular 1,3-chirality transfer, giving the meso-diol, and the calculated R,R/S,Sratio in the dl-type diol was ca. 900: l.24Diastereoselective hydrogenation of N-protected y-amino, P-keto esters catalysed by the (R)-BINAP catalyst provides an efficient entry to statine, a key component of the aspartic proteinase inhibitor pepstatin.28 The efficiency of the catalyst-to-sub- strate chirality transfer (catalyst control, >33 :1) and the intramolecular 1,2-asymmetric induction (substrate control, 3 : 1) cooperate to form the natural threo series in >99 : 1 diastereoselectivity.A number of statine analogues are obtainable by this method using double asymmetric induction (Scheme 11). Thus the present Ru catalysed hydrogenation exhibits wider scope than reactions with any other chiral transition-metal complexes so far designed. A range of optically active compounds of either chirality sense are now accessible, providing a versatile tool in stereoselective organic synthesis. This homogeneous hydrogena- tion procedure is superior to the heterogeneous version and compares well with the biochemical transformations, whose yields and enantioselectivities are often variable. The hydrogenation method is clean, operationally simple, economical, 2’S. Masamune, W. Choy, J. S. Petersen, and L. R. Sita, Angew. Chem.,Int. Ed. Engl., 1985,24, 1.’* T.Nishi, M. Kitamura, T. Ohkuma, and R. Noyori, Tetrahedron Lett., 1988,29,6327. Centenary Lecture and hence is capable of conducting a large-scale reaction using high (up to 50%) substrate concentrati~n.~~.~~ Now one may raise questions: (1) What is the major difference between the Ru chemistry and well-studied Rh catalysed hydrogenation? (2) Why does BINAP ligand work so effectively? The mechanism of the Ru-BINAP catalysed reaction remains to be elucidated. However, d6 Ru" chemistry differs from d8 Rh' chemistry distinctly. First, Ru" complexes utilize higher co-ordination numbers, up to six in an octahedral structure, than Rh' complexes which normally have a square planar geometry. Second, reaction of a Ru" complex and with hydrogen generates Ru monohydride species 31 in contrast to the Rh promoted reaction occurring by way of the metal dihydride intermediate.' Such characteristics would reflect on the marked difference in reactivity-selectivity profiles in the hydrogenation.In the BINAP chemistry, the degeneracy caused by C2 chirality of the diphosphine minimizes the number of the diastereomeric reactive intermediates and transition states. Flexible atropisomeric skeletal backbone of BINAP can produce a conformationally unambiguous metal chelate ring without concomitant increase of strain energy.2 In addition, phenyl rings attached to the phosphorus atoms can suitably modulate stabilities of the intermediary com- plexes and transition states. Molecular structure of Ru-(8-BINAP dipivalate complex determined by single crystal X-ray analysis is given in Figure 4.14The whole structure approximates CZchirality.The dissymmetry of (S)-BINAP fixes the delta conformation of the seven-membered chelate ring containing the diphosphine and Ru. This cyclic structure is highly skewed and this geometry in turn determines the chiral disposition of the phenyl rings on the phosphorus atoms; two phenyl substitutents are oriented in axial directions and the others in equatorial directions. These equatorial phenyl rings exert profound steric influence on the equatorial co-ordination sites of Ru. Consequently, the bidentate ligation of the pivalate moieties to Ru occurs stereoselectively, leading to exclusive formation of the A diastereomer.This diastereomeric differentiation of the two sets of quadrant space sectors is made in such a way as to avoid nonbonded interactions between the sterically demanding equatorial phenyl substituents and the carboxylate ligands. This is merely a ground-state structure of a catalyst precursor but, whatever the detailed reaction mechanism is, such an argument should also be applicable to the transition state or intermediates. Actual chemical transformations take place at the oxygen coordinated sites, and we believe that this is the steric origin of the high level of enantioselection. Stability of the transition structure (1) in the Rh-BINAP chemistry is also understandable in such a way.8 29 R. Noyori, Chimia, 1988,42,215.30 For related work, see: (a) ref. 25; (b) H. Kawano, Y. Ishii, T. Ikariya, M. Saburi, S. Yoshikawa, Y. Uchida, and H. Kumobayashi, Tetrahedron Lett., 1987,28, 1905; (c) T. Tsukahara, H. Kawano, Y. Ishii, T. Takahashi, M. Saburi, Y. Uchida, and S. Akutagawa, Chem. Leff., 1988, 2055; (d) H. Kawano, Y. Ishii, M. Saburi, and Y. Uchida, J. Chem. Soc., Chem. Commun., 1988,81. 31 D. Evans, J. Osborn, J. A. Jardin, and G. Wilkinson, Nature, 1965,208, 1203. Figure 4 ORTEP drawings ofA-Ru[(S)-binap](OCO-t-C4H9)2 Centenary Lecture 6 R\ s+/XR-Zn-R Zn unreact ive reactive X = C, N, 0, halogen, etc. Figure 5 4 Asymmetric Alkylation of Carbonyl Compounds Enantioselective alkylation of aldehydes by organometallic reagents is a funda- mental problem in organic synthesis.Although there have been reports of several successful examples of this type of rea~tion,~~,~~ a high degree of enantioselection is achievable by using a stoicheiometric or even excess amount of chiral auxiliary. Certain ligands may accelerate the nucleophilic alkylation but the difference in rates of the catalysed and uncatalysed reactions is not large enough to lead to a practical asymmetric catalysis.32” In this context, dialkylzincs, the oldest or-ganometallic compounds, generate a variety of new, unprecedented chemistries, opening a novel domain of asymmetric catalysis. Monomeric dialkylzincs having a linear geometry are inert to carbonyl compounds but the structural modification by appropriate ligands or auxiliaries, forming a coordinatively unsaturated bent structure, increases the acceptor character of the Zn atom and donor property of the alkyl group, thereby increasing the reactivity toward carbonyl substrates (Figure 5).Here, some chirally well-designed auxiliary should also direct the stereochemical outcome in an absolute sense as well. Thus in the presence of a catalytic amount of ( -)-3-exo-(dimethylamino)isoborneol (DAIB), reaction of dialkylzincs and benzaldehyde in nonpolar solvents is accelerated markedly to give, after hydrolysis, the corresponding S alcohols in high (up to 99”i;;) enantiomeric purity (Scheme 1 2).34*35Various p-substituted benzaldehydes and certain a,P-unsaturated and aliphatic aldehydes can also be alkylated with a high level of enantioselectivity.Dimethyl-, diethyl-, and di-n-butylzinc are employable as alkylating agents. The catalytic cycle is illustrated in the scheme, where the DAIB structure is ~implified.~~Reaction of (-)-DAIB and dialkylzinc in a 1:l molar ratio ”(a) G. Soladie, in ‘Asymmetric Synthesis’, Vol. 2A, ed. J. D. Morrison, Academic Press, New York. 1983, Chapter 6; (h) J.-P. Mazaleyrat and D. J. Cram. J. Am. Chem. Soc., 1981, 103, 4585; (c) T. Mukaiyama, K. Soai, T. Sato, H. Shimizu, and K. Suzuki, J. Am. Chem. Soc., 1979, 101, 1455; (d) D. Seebach, A. K. Beck, S. Roggo, and A. Wonnacott, Chem. Ber., 1985, 118, 3673; (e) M. T. Reetz, T. Kukenhohner, and P. Weinig, Tetrahedron Lett., 1986, 27, 571 1. 33 R. Noyori, S.Suga, K. Kawai, S. Okada, and M. Kitamura, Pure Appl. Chem., 1988,60,1597. 34 M. Kitamura, S. Suga, K. Kawai, and R. Noyori, J. Am. Chern.Soc., 1986,108,6071. 35 Related works: (a) N. Oguni and T. Omi, Tetrahedron Lett., 1984, 25, 2823; (h)A. A. Smaardijk and H. Wynberg, J. Org. Chem., 1987,52, 135; (c) K. Soai. A. Ookawa, K. Ogawa, and T. Kaba, J. Chem. SOL..,Cfiern. Comrnun., 1987, 467; (a‘) S. Itsuno and J. M. J. Frechet, J. Org. C’hetn., 1987, 52, 4140; (e) P. A. Chaloner and S. A. R. Perera, Tetrahedron Lett., 1987, 28, 3013; (,f) E. J. Corey and F. J. Hannon. Tetrahedron Lett., 1987, 28, 5233, 5237; (g) K. Soai, A. Ookawa, T. Kaba, and K. Ogawa. J. Am. Chem. Soc., 1987, 109, 7111; (h) W. Oppolzer and R. N. Radinov, Tetrahedron Lett., 1988, 29, 5645.See also, D. A. Evans, Science, 1988. 240,420. 36 M. Kitamura, S. Okada, S. Suga, and R. Noyori, J. Am. Chem. Soc., in press. 202 Noyori Scheme 12 produces a single Zn chelate complex (2), which does not alkylate benzaldehyde but acts as catalyst precursor. Significantly, the alkylation proceeds uia a dinuclear Zn species (5) containing DAIB auxiliary, aldehyde ligand, and three alkyl groups. The resulting bridged alkoxide (6), upon exposure to benzaldehyde or dialkylzinc, undergoes instantaneous decomposition to the stable cubic tetramer (7), regenerating (3) and (4), respectively (Scheme 13). Under the catalysis conditions, complexes (2)-(5) are equilibrating on a soft energy surface, consistent with the fact that, when two different dialkylzincs are used, a statistical distribution of the possible products is obtained regardless of the order or way of mixing the two alkylzincs.Here, only relative reactivity of alkyl groups is important. Kinetic measurements and temperature effects on the en-(5) ---antioselectivity indicate that the alkyl transfer process, (6), is the turnover-limiting and stereo-determining step. ' The DAIB-aided enantioselective alkylation exhibits enormous nonlinearity in terms of optical purity of the chiral source and alkylation produ~ts.~~,~~~~~ Typically, when benzaldehyde and diethylzinc are reacted in the presence of 8 mol % of (-)-DAIB in 15% e.e. in toluene, (S)-1-phenylpropyl alcohol is produced with 95% e.e., a value close to 98% obtained using enantiomerically pure DAIB.The nonlinear effect is clear in Figure 6 which shows the e.e.s of (S)-products as a function of the e.e. of (-)-DAIB. Under certain conditions, turnover efficiency of the chiral catalyst system is >600 times greater than that of the coexisting achiral catalyst system. This unusual phenomenon is a result of a marked difference in chemical properties of the diastereomeric dinuclear com- plexes formed from dialkylzincs and DAIB. Reaction of equimolar amounts of dimethylzinc and enantiomerically pure (-)-DAIB affords a dinuclear chelate complex with Cz chirality, which dissociates readily to catalytically active monomeric species. By contrast, dimethylzinc and racemic DAIB generate a more stable, but much less reactive dinuclear complex possessing meso-Ci structure rather than a racemic mixture of the chiral complexes. Molecular structures of these complexes determined by single crystal X-ray analyses are illustrated in Figure 7.36 37 N.Oguni, Y. Matsuda, and T. Kaneko, J. Am. Chem. Soc., 1988,110,1877. 38 (a) C. Puchot, 0.Samuel, E. Dunach, S. Zhao, C. Agami, and H. B. Kagan, J. Am. Chem. Soc., 1986, 108,2353; (b) C. Agami. J. Levisalles, and C. Puchot, J. Chem. Soc., Chem. Commun., 1985,441. 203 Centenary Lecture hl. r N 204 Noyori 0 20 40 60 80 100 % ee of (-)-DAIB Figure 6 Correlation between the e.e. of the alkylation product and the e.e. of the chiral auxiliary. Reaction using 0.42 M (CzH5)zZn,0.42 M CsH5CH0,and 34 mM (-)-DAIBin toluene at 0 "C.A 0.47 M (CH3)zZn,0.49 M C6H5CH0,47 mM (-)-DAIB in toluene- & at 32 "C Alkyl transfer from the mixed ligand complex (5) is conceived to occur via a folded bicyclic transition state (8) featuring a tricoordinate structure of the migrating R group. The kinetic bias leading to the S-configurated alkoxide derives primarily from a nonbonded repulsion between the carbonyl substrate (Ar and H) and a terminal R group attached to ZnB atom. Organometallic chemistry of homo- or hetero-multinuclear compounds is increasing the synthetic importance, and the nonclassical dinuclear mechanism, which has been theoretically ad~anced,~' can provide reasonable explanations for various stereoselective reactions.In order to create a single reactive species, we designed binaphthol-modified Li/Mg binary organometallic reagents having empirical formula of (9) and found that they undergo stoicheiometric en-antioselective alkylation with aldehydes (Scheme 14) to give the corresponding secondary alcohols in high e.e.~.~~For example, 1-phenylpropyl alcohol was produced in up to 92% e.e. The possible transition state resulting in the SjS auxiliary/alcohol asymmetric induction is illustrated by the structure (10) (S = solvent; naphthalene rings are omitted in (10a)). Chiral reducing agent, BINAL- A planar bicyclic transition state has been proposed for reaction of methyllithium dimer and formaldehyde: E. Kaufmann, P. von R. Schleyer, K. N. Houk, and Y.-D.Wu, J.Am. Chem. Soc., 1985, 107,5560. Centenary Lecture C, reactive Ci unreactive Figure 7 ORTEP drawings of complexes formed from equimolar amounts of dimethylzincand (-)-DAIB (upper) and dimethylzinc and (-t-)-DAIB (lower) Noyori 1:1 THF-DME + "'K" -100 "C 0 HOR Li (9) Scheme 14 (R)-B INA L-H (S)-B INA L -HUnKRHO H 0 H OH R S Un = aryl, alkenyl, alkynyl, etc. R = alkyl, H[Li 0''OR' = YO'? I/Li R' (S)-B INAL-H R yun R yo;, t v AI---H oR"o-L\ S S H, exhibits exceptionally high enantioface-differentiating ability in the stoich- eiometric reduction of prochiral ketones having an aromatic, olefinic, or acetylenic substit~ent.~' With many carbonyl substrates, e.e.s greater than 90% are obtainable, where the enantioselection is governed primarily by electronic factors.Now a new model is presented to explain the general binaphthol/carbinol 40 (a) R. Noyori, I. Tomino, Y. Tanimoto, and M. Nishizawa, J. Am. Chem. Soc., 1984, 106,6709: (b) R. Noyori, I. Tomino, M. Yamada, and M. Nishizawa, J. Am. Chem. Soc., 1984,106,6717. Cen tenury Lecture SII S-Li-R configurational relationship (S/S or R/R), which is independent of the relative bulkiness of unsaturated and alkyl groups flanking the carbonyl moiety. In the (S)-BINAL-H reduction, the S-generating transition structure (11) is favoured over the diastereomeric R-generating structure (12), because the latter is configurational relationship (S/Sor R/R),which is independent of the relative bulkiness of unsaturated and alkyl groups flanking the carbonyl moiety.In the (S)-BINAL-H reduction, the S-generating transition structure (11) is favoured over the diastereomeric R-generating structure (12), because the latter is destabilized by the substantial n/n type electronic repulsion between a binaph- thoxyl oxygen and the unsaturated moiety. The oxygen/R steric repulsion in (1 1) becomes significant by increasing the bulkiness of R but cannot overcome the overwhelming electronic influence (Scheme 15). 5 Epilogue The development of homogeneous asymmetric catalysis using chiral metal complexes has provided the straightforward solutions to many challenging problems, proving the validity of the chemical conception.I would conclude that asymmetric catalysis is a four-dimensional chemistry which consists of two fundamental elements in Nature; chirality and circularity. High efficiency is obtainable by creation of ideal three-dimensional structures (x,y,z) coupled with appropriate kinetics (t).This is a frontier of organic chemistry full of promise. Acknowledgements. It is with much pleasure that I acknowledge my collaborators in the successful development of a range of highly stereoselective organic reactions which are described herein. Names of the individuals who made the sustained intellectual and experimental efforts are given in the references. I have particularly enjoyed collaborations in BINAP chemistry with the research groups headed by Professors H.Takaya (Institute for Molecular Science, Kyoto University), S. Otsuka (Osaka University), and Dr. S. Akutagawa (Takasago Research Institute). I am also grateful for the financial support of the Ministry of Education, Science, and Culture of Japan (Specially Promoted Research No. 62065005).

ISSN:0306-0012    

DOI:10.1039/CS9891800187

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. SOC.Rev.,1989,18,209-224 Reaction Branching and Extreme Kinetic Isotope Effects in the Study of Reaction Mechanisms By Alf Thibblin INSTITUTE OF CHEMISTRY, UNIVERSITY OF UPPSALA, P.O. BOX 531, S-751 21 UPPSALA, SWEDEN Per Ahlberg DEPARTMENT OF ORGANIC CHEMISTRY, UNIVERSITY OF GOTEBORG, S-412 96 GOTEBORG, SWEDEN 1Introduction Isotopes are of immense importance in the elucidation of reaction mechanisms as tracers and through their influence on reaction rates.’ Kinetic isotope effects are used for determining the degree of bond cleavage/formation in the transition state of the reaction. However, the interpretation of kinetic isotope effects are often complicated by short-lived intermediates. As will be discussed in this review, reaction branching may, under certain conditions, give rise to both unusually small and unusually large isotope effects.Extreme isotope effects of this origin are sometimes accompanied by anomalous temperature effects. These unusually large isotope effects and anomalous temperature effects from reaction branching may mimic those caused by tunnelling. The occurrence of reaction branching, inferred from measured extreme isotope effects, has been employed as evidence of short-lived intermediates. Accordingly, amplified and attenuated isotope effects have been used as probes of hydrogen- bonded carbanions in hydron-transfer reactions as well as probes of carbocations and ion pairs in solvolysis. The method has the potential to probe also other types of intermediates, i.e.free radicals. Below, effects from branching from a common intermediate, branching directly from the substrate, miscellaneous large isotope effects, and temperature effects that mimic tunnelling will be discussed together with some examples. 2 Branching from a Common Intermediate Let us consider the following simple type of reaction branching, where I denotes a short-lived intermediate: ‘See for example, (a) L. Melander and W. H. Saunders, Jr., ‘Reaction Rates of Isotopic Molecules’, Wiley-Interscience, New York, 1980; (b) ‘Proton-Transfer Reactions’, ed. E. Caldin and V. Gold, Chapman and Hall, London, 1975; (c) R. P. Bell, ‘The Proton in Chemistry’, 2nd Edn., Chapman and Hall, London, 1973; (d)R. P. Bell, ‘The Tunnel Effect’, Chapman and Hall, New York, 1980.(a) P. Ahlberg and S. Bengtsson, Chem. Scr., 1974, 6, 45; (b) A. Thibblin, S. Bengtsson, and P. Ahlberg, J. Chem. Soc., Perkin Trans. 2, 1977, 1569; (c) A. Thibblin and P. Ahlberg, J. Am. Chem. Sue., 1979, 101, 731 1; (d)A. Thibblin, Chem. Scr., 1983, 22, 182; (e) M. Olwegird, I. McEwen, A. Thibblin, and P. Ahlberg, J. Am. Chem. SOC., 1985, 107, 7494; (,f) P. Ahlberg and A. Thibblin, ‘Synthesis and Applications of Isotopically Labeled Compounds’, 1985, Proceedings of the Second International Symposium, Kansas City, MO, USA, Elsevier Science Publishers B.V., Amsterdam, 1986, p.p. 89-94, and references therein; (g)A. Thibblin, J. Phys. Org. Chem., 1988,1, 161. Reaction Branching and Extreme Kinetic Isotope Efects kA A e I k-A C The steady-state approximation applied to equation 1 yields the following relationships between the phenomenological and microscopic rate constants (kAB and kAC are the rate constants for reaction of A to B and C, respectively): Accordingly, the following expressions for the kinetic isotope effects are obtained (the prime denotes the heavy isotope): Thus, for example, the observed isotope effect on the reaction A to B is a product of two isotope effects, kA/kA,and k-B/k-B,,and a third factor.This latter factor depends on the relative magnitude of k-B and k-c and their isotope effects. If k -B/k-B, > kc/k-c,the following expression is valid: Thus we conclude: 2c-g Competition between two processes, which have different kinetic isotope effects and which follow a common (rate-limiting) step sensitive to isotopic substitution, results in an ampliJied observed isotope effect for the overall reaction which proceeds via the competing step with the largest isotope effect.The other overall reaction will show an attenuated isotope effect. If the isotope efect on thejrst step is substantial, the amplijkation may yield an unusually large overall isotope effect. A. Carbanionic Reactions.-The generalization may be exemplified in the following way using Scheme 1. Let the substrate be a carbon acid which is dehydronated by a base to give a hydrogen-bonded carbanion with a kinetic deuterium isotope effect of 7. For simplicity, we assume that internal return is not significant (k-l 4k2,k3). In addition, let us assume that the intermediate gives Thibblin and Ahlberg f-.j"' + Base Base + kl I k-1 + Base H(0)' + L-(C) I = Base H+, tightly hydrogen -bonded carbanion Scheme 1 two products, a rearrangement product formed by intramolecular rehydronation of the carbanion, also showing a kinetic isotope effect of 7, and another product formed uiu a reaction step that is insensitive to isotopic substitution.Accordingly, the subsfrate disappears with a primary kinetic isotope effect of (k!~+ k!c)/(k?~+ k&) = 7. However, the measured kinetic isotope effect on the rate of formation of the product B is enlarged owing to the competition, while the kinetic isotope effect on the rate of formation of C is decreased.The magnitude of the amplification and attenuation depends on the relative amounts of the two products. Let us assume, for example, that they are formed in equal amounts, i.e. k2 = k3. This implies that the measured isotope effects for the formation of the two products are kfB/kzB= 7 x 7 x [(l + 1)/(7 + l)] = 12 and k&/k& = 7 x 1 x [(l + 1)/(7 + l)] = 1.8. If C is the dominant product, i.e. kZ < k3, the measured isotope effects are much larger, -7 x 7 = 49 and -7 x 1 = 7, respectively. Competition from a common tightly hydrogen-bonded allylic carbanion as the cause of enlarged and attenuated kinetic isotope effects has been thoroughly investigated in our lab~ratories.~-~ The experimental results presented in these papers are accommodated in a mechanism in which a hydron is initially 'A.Thibblin and P. Ahlberg, J. Am. Chem. SOC.,1977,9!J, 7926. (a) A. Thibblin, Chem. Scr., 1983,22, 70; (b)A. Thibblin, J. Am. Chem. Soc., 1983, 105,853. 21 1 Reaction Branching and Extreme Kinetic Isotope Effects ( b) t1103s Figure 1 The time dependence of substrate and products in the reaction of 1-(2-acetoxy-2-propyl)indene (h-1-OAc)and 1,3-( *H2)-1-(2-acetoxy-2-propyl)indene(d-1-OAc),respectively,with quinuclidine in methanol at 20.00 "C 2b abstracted from the substrate to form a tightly hydrogen-bonded intermediate involving the carbanion and the conjugated acid of the hydron-abstracting base. In the next step, the carbanion is either rehydronated to form a tautomer or undergoes loss of a leaving group to form an olefin.Thus, enlarged and attenuated primary kinetic isotope effects have been used as a probe of a common tightly hydrogen-bonded intermediate in base-catalysed 1,3-hydron transfer competing with base-promoted 1,2-elimination reactions (Scheme 1 and Figure 1). Independent evidence in favour of an intermediate common to rearrangement and elimination has been also pre~ented.~ Thibblin and Ahlberg This mechanistic scheme implies an increase in the overall rearrangement isotope effect while the overall elimination isotope effect is decreased relative to the ionization isotope effect (cJ equation 6). Reversibility of the ionization step, i.e. k-l >> k2,k3,decreases the observed isotope effects.Accordingly, the isotope effects for the rearrangement and elimination as well as for the disappearance of the substrate are lowered. The largest isotope effect that we have measured for this type of system is a rearrangement kH/kDof ca. 904 and the smallest is an elimination kH/kDof ca. 1 (the value is very small owing to both substantial internal return and a large product ratio tautomer/olefin).2b There is controversy over the borderline between E2 and ElcB reaction^.^ What is the dependence of the mechanism on structure? Do base-promoted elimination reactions make use of both types of mechanisms in parallel or is there an exclusive switch of mechanism on crossing the borderline, ie. do the mechanisms merge on the borderline? Is a stepwise mechanism possible for substrates with such an efficient leaving group as chlorine? We have studied the reaction system shown in Scheme 2 in which the substrate 1 -C1 is reacted with pyridine in methanol.The base-catalysed 1,3-hydron transfer, i.e. the formation of 2-C1 shows an unusually large primary kinetic deuterium isotope effect, 14.6 f 1.0 at 30OC. The extreme isotope effect is proposed to originate from reaction branching in accord with Scheme 1, i.e. the base-promoted elimination and base-catalysed rearrangement of 1-C1 make use of at least one common hydrogen-bonded carbanion intermediate.2e The above generalization based upon equation 1 also implies that not only intramolecular rearrangement but also internal return to starting material may exhibit an amplified kinetic isotope effect when competing with a process from the intermediate that is less sensitive to isotopic substitution.Such a process may be exchange. Cram et al. concluded that the imine (3) ionizes with a kinetic isotope effect of ca. 3 at 75 "C to a potassium-ion-carbanion ion pair (Scheme 3).6 Approximate isotope effects of intramolecular return to starting material and collapse to rearranged imine (3') were calculated as 34 f 19 and 27, respectively. The results were discussed using a mechanistic model (Scheme 3) with azaallylic ion pairs as intermediates. It was concluded that these isotope effects are composite and represent combinations of large numbers of individual rate constants.2c However, the results are conveniently explained by the competition model (equation 1).The ion-pair collapses (k-a and k-& having substantial isotope effects, compete with exchange (ke)with a small isotope effect. Large isotope effects, originally attributed to tunnelling, have been measured by Caldin and co-workers in hydron-transfer reactions of (4-nitropheny1)ni- tromethane with bases containing the imine group in non-polar solvent^.^ For example (equation 7), an isotope effect of kH/kD= 45 was measured in toluene with tetramethylguanidine (BH).7" See, for example, A. Thibblin, J. Am. Chem. Soc., 1988, 110,4582, and references therein. R. D. Guthrie, D. A. Jaeger, W. Meister, and D. J. Cram, J. Am. Chem. Soc., 1971,93,5137.'(a) E.F. Caldin and S. Mateo, J. Chem. Soc., Chem. Commun., 1973, 854; (b) E. F. Caldin and S. Mateo, J. Chem. Soc., Furuduy Truns. I, 1975,71, 1876; (c)ihid., 1976, 72, 112. Reaction Branching and Extreme Kinetic Isotope Effects pyr idine -k12 CI h -1-CI h-2-CI (d -1- CI) (d - 2 -Cl) CH3 OCH3 Scheme 2 ArCD2N02+BH -ArCDNO; DBH+ (7) However, it has been pointed out that the kinetics are complex since the ion pair initially formed from the deuterium substrate gives several other ion pairs as a result of isotopic scrambling; these also return to the If these factors are not taken into consideration, an overestimation of the magnitude of the isotope effect will result since protonation competes favourably with deuteration of the anion in the ion pair.Kresge and Powell have measured a (a)0.Rogne, J. Clzem. Soc., Chem. Commun., 1977, 695; (b)0.Rogne, Acta Chem. Scan., Ser. A, 1978, 32,559;(c)0.Rogne, J. Chem. SOC.,Faraday Trans. I, 1978,74, 1254. Thibblin and Ahlberg HOR I I ROK+ IH?R I (H ID ??R II I 11 II III -'-b3-D(H) + R6; k-Q CH3 H(D) R6; + 3'-D(H) k, I'I -c HOR I I ROK + IH OR H HOR I' III1;: I 11I-+ 3-H + ROK -CH3: H(D) -ROk t 3'-HI I C6H5 Scheme 3 tritium isotope effect of 46 for the T-H exchange in the same system, corresponding to kH/kD= 14.9 A common intermediate has also been demonstrated for acid-catalysed H-D exchange and bromination of the bicyclic amidine in Scheme 4. The common intermediate in this case is not a carbanion but a ketene aminal formed by proton transfer from carbon in an amidinium ion to an amidine nitrogen." B.Solvo1ysis.-Many solvolysis reactions involve branching of the type presented by equation 1. Accordingly, an initial rate-limiting ionization is followed by partitioning of the carbocationic intermediate, ion or ion pair, to give substitution and elimination products, sometimes accompanied by internal return (equation 8).If the return occurs from a free carbocation (external return), the concentration of X -is involved in the rate constant and isotope-effect expressions (equations A. J. Kresge and M. F. Powell, J. Am. Chem. Soc., 1981,103,201. lo S. Lofis and P.Ahlberg, J. Am. Chem. Soc.. 1985,107,7534. Reaction Branching and E,utreme Kinetic Isotope Ejjects H D D D D D D H DBr-1 D CBrC131a+ CDCL, I I D D Scheme 4 k-A sOLef in 2-6). The following discussion holds also for a mechanism with a free carbocation as intermediate. If the substrate RX is isotopically substituted in the P-position, the ionization (kA)is slowed down owing to a secondary isotope effect. The dehydronation (k-c) of the ion pair, on the other hand, shows a primary isotope effect. The nucleophilic substitution (by the solve,nt or other nucleophile) is expected to show only a small inverse secondary isotope effect. The value should be close to unity since the transition state for the attack of the nucleophile is expected to have a structure very similar to the high-energy carbocationic intermediate. The same is valid for the return process.For the sake of simplicity, let us assume that internal return is not significant, i.e. k-A 6 k-B,k-c (equation 8). The isotope effect on the disappearance of the substrate is then equal to the ionization isotope effect /&/k?. However, equations 4 and 5 yield an isotope effect on the overall substitution reaction that is attenuated relative to /&/e;the overall elimination isotope effect k&/k& is enlarged. If the substitution reaction is much faster than the elimination reaction, a maximum overall elimination isotope effect kgc/kic = (kg/ka(kflC/kD& i.e. a product of a secondary and a primary isotope effect, is obtained.’ ‘-13 l1 A.Thibblin, J. Chem. Soc., Perkin Trans. 2, 1986, 320. l2 A. Thibblin, J. Am. Chem. Soc., 1987, 109,2071. l3 A. Thibblin, J. Php. Org. Chem., 1989, 2, 15. Thibblinand Ahlberg The measured isotope effects for the solvolysis of 9-(2-chloro-2-propyl)fluorene and its 2H6-analogue in 25 vol% acetonitrile in water (Scheme 5) were analysed in accord with the competition model.12 The protium compound was found to give a product ratio olefin/alcohol of 1.67. The ionization isotope effect was as large as ky/k? = 2.2, a value close to the expected maximum for a 'H6-compound. The overall elimination isotope effect was found to be relatively large for a solvolytic elimination reaction, k&/k& = 3.7.It was concluded that this isotope effect was composed of the ionization isotope effect, k?/k? = 2.2, the dehydronation isotope effect, kH4/kD4 = 2.8, and the isotope effect for the nucleophilic attack of water on the intermediate, kH 2/k? 2 -1: Moreover, quite in agreement with the proposed mechanistic model (Scheme 5), a drastic decrease in the solvent polarity decreases the amount of alcohol product and the overall elimination isotope effect decreases to 2.8; no change in the isotope effect (2.2) for the disappearance of the substrate occurs. An isotope effect as large as kH/kD6 = 5.7 at 25 "C has been measured for the acid-catalysed dehydration of 2-phenyl-2-propanol in 25 vol% acetonitrile in water.13 The isotope effect is large owing to the competition between slow dehydronation of the intermediate and a fast nucleophilic attack of solvent water. The corresponding reaction of the methyl ether shows similar behaviour.' A solvolytic system that is not accompanied by substitution but by external or internal return may also exhibit an enlarged overall elimination isotope effect owing to the competition between return (with a secondary isotope effect close to unity) and dehydronation of the carbocation intermediate (with a primary isotope effect) that follows an ionization with a substantial isotope effect.' Accordingly, Fry has simulated the observed isotope effect for an El reaction as a function of the amount of return.I5 C.Oxidation-Reduction Reactions.-Unusually large hydrogen isotope effects in oxidation reactions may arise from the following type of mechanism.An initially formed radical reacts with oxygen to give the product. A by-product is formed by a competing reaction of the radical with an inhibitor (equation 10). Both kl and kz involve abstraction of a hydrogen and are expected to show substantial isotope effects. The formation of the byproduct (k3), on the other hand, is not likely to be sensitive to isotopic substitution. Thus, more by-product l4 A. Thibblin, J. Am. Chem. SOC.,in press. l5 A. Fry, Chem. SOC.Rev., 1972,1, 163. l6 (a) S. G. Smith and D. J. W. Goon, J. Urg. Chem., 1969, 3127; (b)V. J. Shiner, Jr., W. Dowd, R. D. Fischer, S. R. Hartshorn, M. A. Kessick, L. Milakofsky, and M.W. Rapp, J. Am. Chem. SOC.,1969,91, 4838; (c) M. P. Jansen, M. K. Koshy, N. N. Mangru, and T. T. Tidwell, J. Am. Chem. Soc., 1981. 103, 3863; (d)X. Creary, C. C. Geiger, and K. Hilton, J. Am. Chem. SOC.,1983, 105, 2851; (e) X. Creary, J. Am. Chem. Soc., 1984, 106. 5568. Reaction Branching and E-xtreme Kinetic Isotope ELfects L = H or 2~ 9pL3C CL3 (6) Scheme 5 R x by -product is formed from the deuterated substrate. Accordingly, an unusually large kinetic isotope effect on the formation of the product results (cf: equation 4). A primary adsorptive isotope effect has been postulated as an explanation for the large isotope effect, kH/kD= 14.2 & 0.9, observed in the oxidation of benzyl alcohol by manganese dioxide.I7 The proposed reaction mechanism may be summarized as in equation 11. This rationalization of the unusually large isotope effect is valid only if the "I.M. Goldrnan, J. Org. Chem., 1969,34, 3289. 218 Thibblin and Ahlberg PhCHDOH + Mn02 / \ Oreact ive Hreactive k.i.e.71 A -I LO s k.i.e.1 C adsorption complexes (I) and (IT) are thermodynamically more stable than non- adsorbed benzyl alcohol. Oxidation reactions are usually mechanistically complex. It is plausible that the unusually large hydrogen isotope effects measured in some reactions of this type," e.g. the oxidation of formate ion by [(bpy)~(py)Ru'v(0)]2' (bpy is 2,2'-bipyridine and py is pyridine) exhibits an isotope effect of kscoo-jkDcoo-= 19 at 25 OC,l 8a are caused by branching.D. Solvent Isotope Effects.-The reaction branching shown in equation 1 may I%-also yield a large inverse solvent isotope effect. a-xmp-le, let I be an intermediate formed irreversibly from A; I may eiher by hyd lronated by the solvent LOS (or by buffer acid) (L = H or D) to giv: B or react to give product C. A substantial isotope effect on the hydronation sep and a nc sgligible isotope effect on the other product-forming step result in an nverse solve :nt isotope effect for the reaction A to C. Accordingly, equations 4 and 5 and the values in equation 12 yield %$s/k'!$s < 1 and %zs/E$s< 7; the extreme values are attained when B is the dominant product. (ci) L. Roecker and T. J. Meyer, J. Am. Cllern. Soc., 1985, 108, 4066; (h)D.Bethell and V. D. Parker. J. Clwm. Soc., Perkin Trans 2, 1982, 841; (c) J. A. Gilbert, S. W. Gersten, and T. J. Meyer, J. Am. Chem. Soc.. 1982, 104, 6812; (d) R. A. Binstead, B. A. Moyer, G. J. Samuels, and T. J. Meyer, J. Am. Ckem. SOL... 1981, 103, 2897; (e) R. Stewart and R. Van der Linden, Tefruhedron Leri., 1960, 28; (,f) L. Roecker and T. J. Meyer. J. Am. Chem. Soc., 1987, 109, 746. Reaction Branching und E.vtl-eme Kinetic Isotope Effijcts DA P Another example is afforded by the dehydronation of a substrate in a fast pre- equilibrium step to give a solvent-equilibrated carbanion intermediate (or a BH-carbanion complex that undergoes rapid hydrogen exchange), which is either rehydronated by the solvent LOS or another general acid, or reacts to product (equation 13), when a large inverse solvent isotope effect may result.This special case of branching as the cause of a large inverse solvent isotope effect has been discussed by Keeffe and Jencks." The mechanism shown in equation 13 accounts for large inverse isotope effects on the initial rate constant for formation of P in the following way (for simplicity, the formation of P from the intermediate is assumed to be insensitive to isotopic substitution). At the first part of reaction, before significant exchange with the solvent LOS has occurred, the reaction is retarded by the reversal of the first step, i.e. k-][BH] competes with k~.The reaction is retarded more in HOS than in DOS owing to a primary deuterium isotope effect on rehydronation, kHyS[BH] > kDyS[BD].This retardation is responsible for the enlarged inverse solvent isotope effect. Application of the steady-state approximation to equation 13 yields the following expressions for the rate constant and the solvent isotope effect: The second factor of equation 15 may have a substantial value. The magnitude depends principally on the ratio k!?S/kD?s and the magnitude of kz relative to kkys[BL]. The inverse isotope effect increases from a minimum value at zero buffer concentration to a maximum value at high buffer concentration. When the first step is not rate-limiting and [BH] = [BD] the observed isotope effect is given by Accordingly, it is possible to observe a very large inverse isotope effect since "J.R. Keeffe and W. P. Jencks. f.Am. Clrcwi. Soc., 1981, 103, 2457 Thibblin and Ahlberg t O z N e CHD-CHz-NR3 Scheme 6 the measured value is the product of a primary deuterium isotope effect (second factor) multiplied by the inverse solvent isotope effect on the ionization step which may attain a value of <2.20 Keeffe and Jencks have measured an inverse solvent deuterium isotope effect of 7.7 on the initial rate of the reversible ElcB reaction presented in Scheme 6.19 The hydronation of the carbanion in this system is much more effective with added buffer acid than with water. Accordingly, the measured isotope effect increases with enhanced concentration of acid (cJ equation 15). The data by More O’Ferrall and Slae 21 for the dehydration of 9-fluorenylmeth-anol in aqueous sodium hydroxide have also been analysed in a similar way; l9 an isotope effect k%:/k$F = 7.9 was obtained. Large inverse isotope effects have been measured in acid solution for the initial rate of cyclohexanone oxidation.22 Another large inverse solvent isotope effect, kD,O/kH,O = 3.5, measured for the glucokinase reaction,23 may also be caused by branching.A very large inverse solvent isotope effect has been reported for a photochemically induced double bond migration in a 2,3-unsaturated ester.24 3 Branching Directly from the Substrate Equation 17 constitutes another type of branching in which the substrate A reacts reversibly to compound B in competition with the reaction to the final product C.This type of branching may result in an apparently large isotope effect on kAC since, if there is an isotope effect on kAC but not on kAB,the isotopically substituted substrate yields a different amount of B than the unsubstituted; and, if only the ”D. A. Whey and E. R. Thornton, J.Am. Chem. SOC.,1975,97,3102. 2’ R. A. More O’Ferrall and S. Slae, J. Chem. SOC.B, 1970,260. 22 J. S. Littler, G. R. Quick, and D. Wozniak, J. Chem. Soc., Perkin Trans. 2, 1980, 657. 23 D. Pollar-Knight and A. Cornish-Bowden, Eur. J. Biochem., 1984,141,157. 24 M. J. Jorgensen, J. Am. Chem. Soc., 1969,91, 198. 221 Reaction Branching and Extreme Kinetic Isotope Effects k~~ A= 8 BA kAC 1 C (7b) Scheme 7 formation of the product C is recorded, an unusually large value for the kinetic isotope effect may result.An example: k& = 100 s-’, kAB = 10 s-’, kgA= 0.1 s-’, and k?c = 10 s-l result in an apparent isotope (kf~/k&,~~that depends on the time interval which is used. Let us assume that the times for formation of 60.65 mol% C from unsubstituted and isotopically substituted substrate (0.0100 s and 4.9 s, respectively) are used for calculation of the isotope effect. The measured value is then (kfc/k?&,s = 490.” Bordwell and co-workers have measured an unusually large kinetic isotope effect that is caused by this type of branching. They measured kH/kD= 340 for the dehydronation of 2-phenyl-1 -nitrocyclopentene with sodium methoxide in methanol.26 However, after careful analysis of the kinetic data they concluded that methoxide anion was added rapidly to the substrate and that the addition product slowly gave back the substrate; the latter underwent slow dehydronation to give product with a ‘normal’ isotope effect of ca.5.Another example of this type of branching is probably the following reaction system. The reketonization of (7) (generated by flash photolysis of the ketone) in a mixture of ether, isopentane, and ethanol has been asssumed to react by 1,5- hydrogen shift of sigmatropic type (Scheme 7).27 It was concluded that (7a) was the more stable of the two conformers; reaction occurs only from (7b) according to the rate expression given by equation 18. Deuterated enol reacts more slowly; koHbs/k?bswas measured as 3 at 300 K and 25 Obtained by computer simulation of the integrated rate expressions corresponding to equation 17.26 F. G. Bordwell, K. C. Yee, and A. C. Knipe, J. Am. Chem.Soc., 1970,92,5945. ”(a) K.-H. Grellmann, H. Weller, and E. Tauer, Chem. Phys. Leff.,1983, 95, 195; (h) U. Baron, G. Bartelt, A. Eychmuller, K.-H. Grellmann, U. Schmitt, E. Tauer, and H. Weller, J. Photochem., 1985, 28, 187. Thibblin and Ahlberg 180 at 140 K (corresponding to about 11 at 300 K). These isotope effects arise from a substantial isotope effect on k3; kl and k2 are not expected to show any significant isotope effects. The change in isotope effect with temperature was attributed to tunnelling since kZ was assumed to be much larger than k3 in the whole temperature inter~al.~’ However, the results may be explained in the following way without invoking any tunnelling.A fast interconversion of the conformers at 300 K is reasonable but it is possible that the equilibrium is ‘frozen’ at low temperature, i.e. (7a) is formed from (7b) but (7a) returns very slowly to (7b). Thus, the system is similar to that described in connection with equation 17. A trivial cause of unusually large kinetic isotope effects is that the reaction is accompanied by the formation of a by-product by branching directly from the substrate (equation 17, kgA= 0). Thus, if the rate of the side reaction is not sensitive to isotopic substitution, more of the by-product is formed from the more slowly reacting isotopically substituted compound.If this is not considered, a very small apparent rate constant for formation of the main product from the ‘heavy’ substrate may result which gives an ‘enlarged’ kinetic isotope effect. Thus, the rate constants given in connection with equation 17 and kgA = 0, and assuming that the times for the formation of 40 mol% C are employed for calculation of the isotope effect yield (k&/kfc)obs= 15.25 On the other hand, if the rate of disappearance of A is measured (with rate constant kobs= kAB + kAc), an attenuated isotope effect kHb,/k,Db,is obtained. 4 Miscellaneous Large Isotope Effects Other unusually large isotope effects in the literature include the following examples.There is no specific reason to invoke branching, but it should be of interest to reconsider these reactions with that possibility in mind. An unusually large isotope effect, kH/kD-27 at OOC, has been measured for the complex reaction of acid-catalysed cis-trans isomerization of a diene-iron tricarbonyl complex.28 Hydrogen atom abstraction from phenols by polyvinyl acetate radicals has been found to exhibit a deuterium isotope effect of 57 23 at 50 0C.29Another example of an unusually large isotope effect for a hydrogen abstraction has been reported for the 2-hydroxyphenoxyl radical, kH/kD-104 at 25 0C.30A kinetic deuterium isotope effect of 25 at 65 “C has been measured for hydrogen atom abstraction from cy~lohexene.~ The thermal oxidation of 4a,4b- dihydrophenanthrene and the analogous (*H12)-compound by oxygen exhibits kH/kD= 95 at -31 “C (corresponding to about 40 at 25 0C).32 An unusually large primary hydrogen isotope effect has been measured for the T.H. Whitesides and J. P. Neilan, J.Am. Chem. SOC.,1975,97,907. 29 J. Kardos, I. Fitos, J. Szammer, and M. Simonyi, J. Chem. SOC., Perkin Trans. 2,1978,405. 30 K. Loth, F. Graf, and H. H. Gunthard, Chem. Phys., 1976,13,95.’’ P. S. Engel, W.-K. Chae, S. A. Baughman, G. E. Marschke, E. S. Lewis, J. W. Timberlake, and A. E. Luedtke, J. Am. Chem. SOC.,1983,105,5030. 32 (a) A. Bromberg, K. A. Muszkat, and E. Fischer, J. Chem. SOC.,Chem. Commun., 1968, 1352; (b) A. Bromberg, K. A. Muszkat, E. Fischer, and F. S. Klein, J. Chem. SOC.,Perkin Trans.2, 1972, 588 and references therein. 223 Reaction Branching and Extreme Kinetic Isotope Efects ionization of 2-nitropropane with 2,4,6-trimeth~lpyridine.~~However, it has been concluded that the value kH/kD= 24 measured in t-butyl alcohol-water is too large since it is measured in the presence of oxygen.34 When the reaction was carried out under nitrogen, the rates were faster and the isotope effect was lower, kH/kD= 16.34High pressure (3000 bar) was found to decrease the isotope effect down towards 7. Also, Jarczewski and co-workers have measured unusually large isotope effects in hydron transfer from substrates activated by a nitro-gr~up.~~ 5 Temperature Effects that Mimic Tunnelling The unusually large kinetic isotope effects which have been reported in the literature have almost exclusively been concluded to be caused by tunnelling.' In some of these systems, the tunnelling hypothesis is also supported by studies of activation parameter^.'^^' However, it has been shown by simulation that temperature effects on activation parameters for branched reactions may mimic those that are expected to arise from t~nnelling.~~.~~ Accordingly, it has been concluded that both unusually large primary kinetic isotope effects and ratios of Arrhenius pre-exponential factors considerably smaller than unity may result from branching (equation 1) if the rate of the reaction under study increases more with temperature than the competing side reaction.The latter can be slower than the main reaction by a factor of ten and still give rise to this anomalous temperature effect.2Y It has also been concluded that branched reactions may result in constant or even increased isotope effect with tempera- ture.2g Acknowledgment.The Swedish Natural Science Research Council has provided financial support for this work. 33(a)E. S. Lewis and J. D. Allen, J. Am. Chem. Soc., 1964, 86, 2022; (b) E. S. Lewis and L. H. Funderburk, J. Am. Chem. Soc., 1967, 89, 2322; (c) R. P. Bell, E. S. Lewis, and J. K. Robinson. J. Am. Chem. Sor., 1968,90,4337;(d)R. P. Bell and D. M. Goodall, Proc. Roy. SOL'.,1966, A294,273. 34 N. S. Isaacs and K. Jivaid, J. Chem. SOC.,Parkin Trans. 2, 1979, 1583. 35 P. Pruszynski and A. Jarczewski, J. Chem. Soc., Perkin Trans. 2, 1986, 11 17. 36The internal-return mechanism may be considered as a special type of branching (one of the competing reactions gives back starting material).2g Also, this type of mechanism may give rise to activation parameters which mimic those of reactions with a large tunnel contribution (ref. 37). 37 H. F. Koch and D. B. Dahlberg, J. Am. Chem. Soc., 1980,102,6102.

ISSN:0306-0012    

DOI:10.1039/CS9891800209

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. SOC.Rev., 1989,18,225-250 Structural Properties of Organocobalt Coenzyme B12 Models By L. Randaccio, N. Bresciani Pahor, and E. Zangrando DIPARTIMENTO DI SCIENZE CHIMICHE, UNIVERSITA DI TRIESTE, 34127 TRIESTE, ITALY L. G. Marzilli DEPARTMENT OF CHEMISTRY, EMORY UNIVERSITY, ATLANTA, GEORGIA 30321, USA 1 Introduction Homolysis of the Co-C bond of coenzyme B 12 (5’-deoxyadenosylcobalamin, Figure 1) is generally accepted to be an essential step in the enzymatic reactions for which coenzyme BI2 is a cofactor.’-6 Cobalamins (Cbls) are a class of compounds with the cobalt(m) coordinated in the equatorial plane by four corrin N atoms and in the axial position by a 5,6-dimethylbenzimidazolenucleotide. If the nucleotide is missing, the molecules are called cobinamides.The seminal experiments and some of the key conceptual contributions in this field have recently been reviewed in this j~urnal.~ In that review, the rate of Co-C bond homolysis was estimated to be increased by z10’ on binding of coenzyme B12to the protein, and by a further factor of z105 by substrate binding. More recent studies document a 31013 increase in the rate of the coenzyme B12 Co-C bond cleavage in the holoenzyme compared to cleavage of this bond in the absence of enzyme in solution.’ The manner in which the enzyme accomplishes this rate enhancement remains a mystery, although it is likely that the enzyme must use its cofactor*enzyme intrinsic binding energy’ to weaken the Co-C bond or otherwise ‘trigger’ the Co-C bond homolysis step.The molecular level details of this process are of considerable interest. Specifically, B 2-cofactor-localized mechanisms that have been postulated to account for the enzyme-accelerated Co-C bond homolysis include: (a) a distortion in the corrin ring increasing the steric interaction with the adenosyl moiety (the ‘butterfly’ bending or upward conformational theory); 3,’-7 (b) a direct lengthening or angular distortion of the Co-C bond by the protein; and (c) an alteration in the position or Co-N bond ’ ‘B,:’. ed. D. Dolphin, J. Wiley, New York, 1982. J. Halpern, Science, 1985, 227, 869 and references therein; R. G. Finke, D. A. Schiraldi, and B. J. Mayer, Coord. C‘hem. Rev., 1984, 54, 1; B. T. Golding, J.R. Neth. Chem.Soc., 1987, 106, 342 and references therein. N. Bresciani Pahor, M. Forcolin, L. G. Marzilli, L. Randaccio, M. F. Summers, and P. J. Toscano, Coord. Chem. Rev., 1985,63, 1. J. M. Pratt, Chem. Soc. Rev., 1985, 161. B. P. Hay and R. G. Finke, J. Am. Chem. Soc., 1987,109,8012 and references therein. ’S. H. Kim, H. L. Chen, N. Feilchenfeld, and J. Halpern, J. Am. Chem. Soc., 1988, 110, 3120 and references therein. ’V. B. Pett, M. N. Liebman, P. Murray-Rust, K. Prasad, and J. P. Glusker, J. Am. Chem. Soc., 1987, 109.3207. Structural Properties of Organocobalt Coenzyme B12 Models length of the axial 5,6-dimethylbenzimidazole(bzm) ligand. U.V. spectral data have been used to suggest that the Co-bzm bond is broken during catalysis, giving the so-called base-off (benzimidazole-unprotonated) form of the coenzyme,' and several studies exist claiming that axial base-free 5'-deoxyadenosylcobin- amide (Ado-Cbi') is still a partially active c~factor.~ The adenosyl-cobalt bond in coenzyme B12 is relatively stable when compared to other alkyl-cobalt bonds.Coenzyme B12 and Ado-Cbi' not only exhibit the largest AH2 (Co-C bond homolysis activation enthalpy) for an alkyl-Cbl (2+ 27 kcal mol-') lo or an alkyl-Cbi' (27-32 kcal mol-'),lo respectively, but Ado- Cbi' also exhibits the largest AH2 value reported to date, 37.5 kcal mol-', when compared to all literature values (17-37.5 kcal m~l-').~*~.'~-''Since the essential, and perhaps only, role of coenzyme B12 in the holoenzyme is thought to be the homolysis of the Co-C bond,2*4,18-20comparison of the rates of thermally induced Co-C bond homolysis and Co-N bond dissociation rates to structures of both models and Cbls is of special significance and is the subject of this review.Particular emphasis will be placed on cobaloxime models which contain dioximate ligands in the equatorial plane (vide infra). 2 Cobalamins Cbls contain a corrin ring, which is similar in structure to a porphyrin ring with one bridge atom missing and with saturated P-positions on the five-membered rings A and D (Figure 1). Seven amide chains are attached to the corrin. One of the amides is linked by a 2-aminopropanol group to a 5,6-dimethylbenzimidazole nucleotide. These side chains, as well as the eleven ring-methyl substituents, limit the flexibility of the corrin and, when axially projected from the rings, serve to 'protect' the axial Co-C bond.The axial R ligand, coordinated to the upper or p-site of the corrin ring, is 5'-deoxyadenosyl in B12 coenzyme (Ado-Cbl), CN in vitamin B1 (CN-Cbl), and an alkyl group in alkylcobalamins (alkyl-Cbl). The benzimidazole residue is coordinated to cobalt in the lower or a-site. A second organocobalamin, Me-Cbl, is also an important cofactor for many enzymic processes. CN-Cbl serves as a metabolic precursor for both coenzymes. Their M. F. Summers, P. J. Toscano, N. Bresciani Pahor, G. Nardin, L. Randaccio, and L. G. Marzilli, J. Am. Chem. Soc., 1983, 105, 6259; M. F. Summers, L. G. Marzilli, N. Bresciani Pahor, and L.Randaccio, J. Am. Chem. Soc., 1984,106,4478. J. M. Pratt, Inorg. Chim. Acta, 1983.79, 27. lo G. N. Shrauzer and J. H. Grate, J. Am. Chem. Soc., 1981, 103,541. I' F. T. T. Ng, G. L. Rempel, and J. Halpern, J. Am. Chem. Soc., 1982,104,621. F. T. T. Ng, G. L. Rempel, and J. Halpern, Inorg. Chim. Acta, 1983,77, L165. l3 J. Halpern, F. T. T. Ng, and G. L. Rempel, J. Am. Chem. SOC.,1979, 101,7124. l4 (a) H. B. Gjerde and J. H. Espenson, Organometallics, 1982, 1, 435; (6) T. T. Tsou, M. Loots, and J. Halpern, J. Am. Chern. Soc., 1982.104,623. l5 R. G. Finke, B. L. Smith, B. J. Mayer, and A. A. Molinero, Inorg. Chem., 1983,22,3677. l6 A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1984,106,5197. Y.Ohgo, K. Orisaku, E. Hasegawa, and S.Takeuchi, Chem. Lett., 1986, 27. l8 S. Wollowitz and J. Halpern, J. Am. Chem. Soc., 1984, 106,8319. I9 R. G. Finke, W. P. McKenna, D. A. Schiraldi, B. L. Smith, and C. J. Pierpont, J. Am. Chem. Soc.., 1983,105,7592. 'O R. G.Finke and D. A. Schiraldi, J. Am. Chem. Soc., 1983,105,7605. Randaccio, Pahor, Zangrando, and Marzilli CONH, CONH2I I I II I CONH2 Figure 1 Atomic model of cobalamins biological aspects have been reviewed extensively,’ and the need for understand- ing the structural factors influencing the Co-C bond homolysis in Ado-Cbl has been empha~ized.~-~*~ A. X-Ray Structures.-The results of X-ray structure analyses of Cbls and related compounds were thoroughly reviewed in 1981.22 Surprisingly, the only structural results available then for Co-C bonded systems were for different forms of CN- Cbl in addition to the first determination of CN-Cbl done by Dorothy Hodgkin in 1950~,~~and the only organocobalamin structure was for Ado-Cbl.Since then, only a few more structure analyses have been reported. Among them, the Me-Cbl structural analysis finally appeared.23 Except for the axial group, the structure of this coenzyme is very similar to those previously reported, although there is some l1S. M. Chemaly, E. A. Betterton, and J. M. Pratt, J. Chem. Soc., Dalton Trans., 1987, 761 and references therein. ”J. P. Glusker in reference 1, Vol. 1, p. 23. 23 M. Rossi, J. P. Glusker, L. Randaccio, M. F. Summers, P. J. Toscano, and L. G. Marzilli, J. Am. Chem.Soc., 1985,107, 1729. Structural Properties of Organocobalt Coenzyme B12 Models Table 1 Coordination geometry in cobalumins R Co-N(ax) Co-C Co-N(eq) N(ax)-Co-C Co-Ca-CP (4 (4 (4 (") ("1 Re6 CN 2.03(2) 1.91(3) 1.91-1.94(2) 175.0(9) 176(3) U P(O)F(OMe) 2.09(2) -1.88-1.92(1) --b Me 2.19(2) 1.99(2) 1.88-1.97(2) 171(2) -C PO(OMe)2 2.20(1) -1.91-1.96(1) --b Ado (neutron) 2.24(1) 2.04(1) 1.82-1.96(1) 169.0(8) 121.0(8) d Ado (X-ray) 2.24(1) 2.00(1) 1.87-1.91(1) 173.0(5) 124.0(5) d (R)-DhP 2.27(2) 2.00(2) 1.84-1.95(2) 169.1(9) 120(2) e (V-DhP 2.36(3) 2.08(3) 1.89-1.98(3) 172(1) 114(2) e Ref. 25. Ref. 26. Ref. 23. Ref. 27. @ Ref. 24, Dhp = 2,3-dihydroxypropyl. disorder in the area of the phosphate group. A careful comparison of the benzimidazole orientation and the side-chain conformation in both Me-Cbl and CN-Cbl revealed a remarkable similarity, with the exception of the orientation of the side-chain terminal amide groups.The coordination bond lengths are given in Table 1, together with the axial N-Co-C and the CO-CCC-C~ angles. More recently, the crystal structures24 of two isomers (R)and (S) of 2,3-dihydroxy- propylcobalamin (Dhp-Cbl) gave the geometrical data reported in Table 1. The cobalamin portion of both molecules is very similar to that of Ado-Cbl. The alcohol residue, which adopts a similar conformation in the (R)-and (S)-isomers, uses in the latter a channel between c and D corrin rings to locate its CP-CH20H group so that the CP-OH residue can hydrogen bond to the carbonyl oxygen of the c-acetamido group.The ribose ring in Ado-Cbl nestles in this channel since it has a chirality corresponding to that of the (S)-conformer. No hydrogen bonding occurs in the (R)-isomer. The larger Co-Ca-Cp angle (Table 1) for (R) than for (S) has been attributed to the hydrogen bond and to less severe steric interactions in (S). On the other hand, the narrower Co-Ca-Cp angle cor-responds to a longer Co-C bond in the (S)-isomer. Although the experimental errors limit this comparison, it may be suggested that the response to the steric demands and to hydrogen bond formation takes place mainly through a lengthening of the Co-C bond in the (S)-isomer but through an opening of the Co-C-C angle in the (R)-isomer. Similar behaviour has been observed already in cobaloximes.The structure of a new crystalline form of CN-Cbl has been refined using diffractometer data.25 Some geometrical parameters are reported in Table 1. The overall structure, although more accurate, does not differ significantly from those already reported.22 The structure of a monocarboxylic acid derivative of CN-Cbl, in which an amide group (probably in chain b) is converted into a carboxylic 24N.W. Alcock, R. M. Dixon, and B. T. Golding, J. Chem. Soc., Chem. Conitnun., 1985, 603; R. M. Dixon, B. T. Golding, S. Mwesigye Kibende. and D. R. N. Rao, Philos. Trons. R. Soc. London, 1985, B311,531. 25 L. G. Marzilli, J. P. Glusker, N. Bresciani Pahor. L. Randaccio, and E. Zangrando. Abstracts XVIII Italian Inorganic Chemistry Meeting, E5, Como, 16 20 September 1985.228 Randaccio, Pahor, Zangrando, and Marzilli group, has axial distances26 very close to those of CN-Cbl, i.e. Co-C 1.91(1) 8, and Co-N 2.01(1) A. High-resolution neutron and X-ray diffraction data for Ado-Cbl were collected2' (Table 1) to obtain greater structural resolution than that of the early work.22 The c side-chain of the corrin is disordered between two extreme positions. Six of the eleven methyl groups appear to be ordered while the remaining five either undergo an extensive thermal libration or are statistically disordered. The details of the solvent structure analysis have also been reported.28 Comparison of the results reported in Table 1, although limited because of data accuracy, shows unequivocally that the Co-N(bzm) distance reflects the o-donor power of the trans-ligand.An increase of 0.3 8, occurs in the order CN < Me < Ado < Dhp. The Co-C bond length is particularly short for CN-Cbl, as expected on the basis of the covalent radius of a C(sp) compared with that of a C(sp3). On the other hand, the Co-Me bond length lies at the lower limit of the range found for the other alkylcobalamins of Table 1. The structures of two Cbls having P(0)(OMe)2 and P(O)F(OMe) as the axial ligand have been reported.26 The accuracy of these results (Table 1) allows an evaluation of the trans-influence of the two axial groups. The relatively long Co-N(axia1) bond in P(0)(OMe)2-Cbl is in agreement with evidence in solution that P(O)(OMe)2 is a strong trans-influencing group, similar to methyl, in both cobaloximes and in Cbk2l B.Kinetics and Co-C Bond Dissociation Energy.-Halpern2 showed that a reasonable estimate of the Co-C bond dissociation energy (BDE) can be obtained from AH* of the Co-C bond homolysis in solution of simple models. Since then, several determinations of Co-C BDE in Ado-Cbl and similar systems have appeared. In Ado-Cbl and related alkylcobalamins the Co-C bond homolysis can occur through both the base-on (a) and base-off (b) forms, which are in equilibrium (Scheme 1). A value of the base-on Co-C BDE of 26 & 2 kcal mol-' has been reported by Halpern,6 which is lower than that of 30 & 2 kcal mol-' reported by Fi~~ke.~' More recently, the latter authors have measured the Co-C BDE in AdoCbif.An estimate of 34.5 & 1.8 kcal mol-' was suggested, so that the removal of the trans-axial base strengthens the Co-C bond by about 4 kcal mol-', while the coordination of the axial base to Co increases the rate of the Co-C bond homolysis by a factor of lo2. As mentioned above, the Co-C(Ado) BDE in the complex natural cofactor Ado-Cbl is quite large and the precise value is the subject of controversy. Therefore, the factors influencing it and Cbl properties in general have been investigated extensively with model compounds such as those to be described next. These elegant evaluations of Co-C BDEs require exhaustive and extensive 26 J. Kopf, Habilitationsschrift, University of Hamburg, 1986 and references therein.'' H. F. J. Savage, P. F. Lindley, J. L. Finney. and P. A. Timmins, Acfa Crvstallorg.. Sc~ct.B, 1987, 43, 280. 2R H. F. J. Savage, Biophj~sics,1986.50, 967. 29 B. P. Hay and R. G. Finke. J. Am. Cheni. Soc.., 1986, 108,4820. Structural Properties of Organocobalt Coenzyme BIZModels R R R K experimental studies. The rates of decomposition of many alkyl Cbls have been evaluated in pioneering studies in Schrauzer's and Pratt's lab~ratories.~ Although we note that the effects of steric factors in Co-C bond cleavage have been re~iewed,~an abbreviated series relevant to our studies follows: Me < 5'-deoxyadenosyl < Et < Pri,CH2CMe3 In this review, we shall consider the question: How does the series compare with precise structural information on models and the limited structural data on Cbls? 3 Dioximates The most extensively studied model system, the cobaloximes containing the dioximate moiety, was reviewed in depth in 1985.3 Since this review was published, many structural and solution data have appeared on dioximate cobalt complexes, most of them having bis(dimethylglyoximato), (DH)2, as equatorial ligand (Figure 2).Some data have been reported for the bis(glyoximato), (GH)2, and bis(diphenylglyoximato), (PH)2, derivatives (Figure 2). About 50 new structures of organometallic complexes have appeared in the past four years. Structural and solution results for these complexes will be reported in this section and discussed together with those reviewed previously.A. Equatorial Moiety: X-Ray Results.-In our previous review a statistical analysis of the geometry of Co(DH)2 unit gave the bond length mean values for organocobaloximes reported, together with those of the dimethylglyoxime (DH2) in Table 2. The corresponding structural results of the new series of organo- Randaccio, Pahor, Zangrando, and Marzilli L R R'= H CO(GH)~ R' = Me Co(DH12 R' = Ph CO(PH)~ Figure 2 Table 2 Comparison of mean bond lengths (A) of equatorial ligands DHz and GH2 before and after ionization and coordination to cobalt n" Co(DH)zb DHzb no CO(GH)~' GHZb CO-N 306 1.8901(9) -20 1.884(9) -c-c 157 1.462(3) 1.474(3) 10 1.44(1) 1.453(1) C-N 302 1.301(1) 1.288(3) 20 1.30(1) 1.2849(8) N-0 306 1.3492(9) 1.410(3) 20 1.34(1) 1.3854(8)C-Me 306 1.501(1) 1.487(2) ---O*.*O 140 2.487(2) -10 2.49(1) -'Number of averaged measurements.'data from ref. 3. Present work. cobaloximes agree with these values within the experimental errors. The mean values for the Co(GH)2 moiety, together with those of the free ligand GH2, are reported in Table 2. Ionization and coordination to Co of the latter ligand provoke changes in the same direction observed in DH2, with a significant decrease only in the N-0 distance. Therefore, the delocalization in the Co(GH), plane is small, as suggested for the Co(DH), m~iety.~The Co-N bond lengths, which are very close in the two systems, confirm this hypothesis. B.Axial Ligands, X-Ray results.-The geometry of the L-Co-R fragment is reported in Table 3 together with the bending angle, X, between the two equatorial halves and the displacement, d, of the Co out of the four N-donor plane. Positive values for z and d indicate bending towards X or R and displacement towards L, respectively. The Co-L bond lengths reported in Table 3 confirm previous findings3 that this bond lengthens with the o-donor power of the trans-alkyl group (electronic trans-influence). For some series containing the same L ligand and different R groups, the electronic trans-23 1 n,w w Table 3 Structural parameters of the L-Co-R moiety and d and a values for cobaloximes having GH and DH as equatorial ligands R CO-L (A) CO-C (A) L-CO-R (") CO-C-Y (") a("> Ref: GH PY Me 2.064(3) 2.005(4) 178.0(2) -5.6 0.05 U Et 2.067(6) 2.020(7) 178.4(3) 119.9(7) 1.1 0.04 a Pr' 2.10 l(6) 2.084(9) 177.5(3) 116.2(8) 1.9 0.03 a Me 2.268( 1) 2.041(4) 179.9(0) -4.0 0.08 h Me 2.428( 1) 2.033(3) 178.8(1) -6.0 0.1 1 h DH NHzPh Me 2.129( 1) 1.992(2) 178.19( 7) -3.5 0.04 c Et 2.147(2) 2.030( 3) 178.3( 1) 117.8(3) 2.8 0.01 d CH2C(C02Et)2Me 2.143(2) 2.035(2) 170.86(8) 126.8(2) -3.3 -0.01 C CH20Me 2.169(3) 2.01 3(4) 176.3(2) 1 19.4(4) 2.7 0.03 d Pr' 2.177(2) 2.068( 3) 178.3(1) 114.2(2) -5.6 -0.02 d adamantyl 2.21 6(4) 2.158(4) 177.1( 1) 110.2(4) -10.8 -0.07 e (2-NH 2py)NH2 Et 2.156(3) 2.025( 3) 174.1( 1) 12 1.0(3) 4.0 0.01 .f (R)-PhCH(Me)NH2 (S)-CH(Me)C02Me 2.074(5) 2.08 3( 6) 174.5(2) --0.01 s,h PY CH(CN)Cl 2.017(2) 2.01 5(3) 177.0( 1) 1 17.0(2)Cl -7.0 -0.02 i 112.9(2)CN b NCHzCN 2.014(4) 2.024(6) 175.2(2) 115.1(5) 4.7 0.03 jCHCl2 2.045(2) 1.995(2) 176.52(8) 115.7( 1) -7.6 -0.02 k CHzCHzCN 2.050(5) 2.002(7) 176.4(3) 123.6(6) 2.7 0.04 i -CHzCHzCN 2.094(4) 2.005(5) 177.6(2) --1,m (R)-CH(Me)C02Me 2.049(5) 2.083(8) 176.6(2) n CH2Ph 2.056(3) 2.065(4) 177.1(1) 116.7(2) 5.0 0.04 .i --04-CNpy (S)-CH(Me)C02Me 2.076(6) 2.098( 7) 176.4(3) 114.2(6)Me 108.7( 5)C02Me 4- MezN Py adamantyl 2.102(3) 2.160(4) 178.8(1) 11 1.0(3) -10.5 -0.05 e 4-Mepy (R)-CH(CN)Me 2.054( 7) 2.05(1) 174.7(3) g1.2-MeJm CClzCN 2.025(3) 2.047(4) 175.6( 1) 114.1 (2)Cl -8.5 -0.02 P 112.6(3)CN CHzNOz 2.049(3) 1.999(3) 174.0( 1) 1143 1) 3.4 0.03 P 1,2-Me21m--conl.CH2CN 2.049( 3) 2.01 8(4) 177.1(2) 116.5(3) 6.0 0.06 P CHzCHzCHzCN 2.083(2) 2.023(3) 174.2( 1 ) 119.4(2) 5.9 0.06 P Me 2.086( 1) 2.001 (2) 176.62(8) -3.4 0.06 P Pr' 2.121(2) 2.096(3) 174.5(1 ) 114.4(2) 7.3 0.04 P Me3Bzm CH(CN)CI ,2.010(3) 2.000(5) 1 77.1(2) 126.1(3)C1 I .9 0.01 Y 1 18.2(3)CN CH(CN)CHzCN 2.03 l(3) 2.06 l(3) 175.3(1) 112.5(3)CN 3.1 0.01 r 116.5(3)CHzCN CHzNOz 2.013(3) 1.988(5) 176.6(2) 1 15.7(4) 4.8 0.05 S CHClz 2.043(2) 1.983(2) 173.72(8) 11 6.2(1) 1.5 0.01 k Me 2.060(2) 1.989(2) 176.8( 1) -4.7 0.06 S Pr 2.09 7( 2) 2.076(2) 176.41(8) 114.2(2) 4.0 0.06 S adamantyl 2.137(4) 2.179(5) 178.0(2) 1 10.4(4) -6.1 -0.03 Y Im Me 2.0 19( 3) 1.985(3) 177.2(1) -1.8 0.03 t 1-MeIm CHzCHzCN 2.037(3) 1.989(5) 177.0(2) 127.9(5) 4.9 0.05 P adamantyl 2.065(4) 2.154(2) 179.2(2) 1 1 1.2(4) -10.0 --0.06 U pyrrolidine CH2Ph 2.1 14(8) 2.128(8) 1 73.1(4) -1.5 0.02 V P(OMe)3 adamant yl 2.3 13( 1) 2.2 14(3) 178.91(8) 110.7(2) -7.3 -0.02 N' P(OW) adaman tyl 2.367(2) 2.199(6) 176.3(2) 11 1.2(5) -7.5 -0.01 M' PBu3 (R)-CH(CN)Me 2.3 1 6(2) 2.089(6) 176.5(2) 117.0(5)Me 2.1 0.03 x 11 1.9( 5)CN PEt2Ph (R)-CH(CN)Me 2.313( 1) 2.074(4) 174.4( 1) -1.8 0.0 1 .Y PEtPh2 (R)-CH(CN) Me 2.370( 1) 2.098(3) 1753 1) -5.1 0.05 Y PPh3 (R)-CH(CN)Me 2.4 10(3) 2.08( 1) 172.3(4) 116.4(9)Me 11.6 0.08 X 1 15.0(9)CN CN CN 1.91 l(7) 1.9 16(8) 178.8(2) 17737) Ref.34. 'Ref. 33. Unpublished results. L. G. Marzilli, F. Bayo, M. F. Summers, L. B. Thomas, E. Zangrando, N. Bresciani Pahor, M.Mari, and L. Randaccio, J. Am. Chem. Soc., 1987, 109, 6045. N. Bresciani Pahor, L. Randaccio, E. Zangrando, and P. A. Marzilli, J. Chern. Soc., Dalton 7rans., in press. L. G. Marzilli, M. F. Summers, E. Zangrando, N. Bresciani Pahor, and L. Randaccio, J. Am. Chem. Soc., 1986, 108, 4830. Ref. 40. mean values for two crystalline forms. E. Zangrando, N. Bresciani Pahor, L. Randaccio, J. P. Charland, and L. G. Marzilli, Organometallics, 1986, 5, 1938. j N. Bresciani Pahor, L. Randaccio, E. Zangrando, and L. Antolini, Acta Crystallogr., Sect. C, 1988, 44,2052. Ref. 35. ' The equatorial ligand is bis[( E,E)-l -phenyl-1,2-propanedionedioxirnato-N,N']. A. Uchida, Y. Ohashi, Y. Sasada, and Y. Ohgo, Acta Crystallogr., Sect. C, 1985, 41, 25. " Ref. 38. 'T. Kurihara, A.Uchida, Y. Ohashi, Y. Sasada, and Y. Ohgo, Sect. C, 1984, 40, 1557. Ref. 30. Ref. 36. 'Unpublished results. Ref. 31. ' V. Pattabhi, M. Nethaji, E. J. Gabe, F. L. Lee, and Y. Le Page, ActuActa Cr~~stalfogr., Crystallogr., Sect. C, 1984, 40, 1155. " N. Bresciani, Pahor, L. G. Marzilli, L. Randaccio, P. J. Toscano, and E. Zangrando, J. Chem. Soc., Chem. Commun., 1984, 1508. L'S. K. Tyrlik, A. T. H. Lenstra, J. F. J. Van Look, H. J. Geise, and R. A. Dommisse, Ac,/a Crystallogr., Sec/. C, 1986, 42. 553. N. Bresciani Pahor, L. Randaccio, E. Zangrando, M. F. Summers, J. H. Ramsden, P. A. Marzilli, and L. G. Marzilli, Organometallics, 1985, 4, 2086. T. Kurihara, A. Uchida, Y. Ohashi,w w Y. Sasada, Y. Ohgo, and S. Baba, Acfa Crystallogr.. Sect B, 1983, 39, 431.Y. Tomotake, A. Uchida, Y. Ohashi, Y. Sasada, Y. Ohgo, and S. Baba, Acra Crystallogr.,Sect. C, 1984,40, 1684. C. Lopez, S. Alvarez, X. Solans, and M. Aguilo, fnorg. Chim. Acta, 1987, 133, 101. Structural Properties of Organocobalt Coenzyme B12 Models 1.9r L =NH2Ph L = py L=Me3Bzm L=Me2Im -C! 1.91 --N3 2 .o; -CHZCN T -Cl Ca(C N)Cl 2 .OE CHZ CHzNOz Me -Me -C Hz C( Me (C92 Et 2 -EEt,CHzCMe3 -Me -CHzSiMe32.1c -Pri -Pri -adamantyl -Pri -Me 2.14 CH2C(Me)(COZEt l2 -adamantyl--Et -CH,OMe 2 .1e -Pri 2.22 -adamantyl c 0-L Figure 3 Diagram showing the trend of the Co-L distances (A) in cobaloximes for different alkyl groups trans to L = NH2Ph, py, Me3Bzm, Me21m. Available values for azido and chloro analogues are reported for comparison influence is clearly shown in Figure 3.For comparison, the Co-N distances trans to weak acido-donors are also given. An increase of about 0.2 A in each series is Randaccio, Pahor, Zangrando, and Marzilli 2 .15 1Me-Im Ipri Im Me3Bzm py 1.2-Mqlm 2-NHzpy Figure 4 Trend of Co-N distances for MeCo(DH)2L and PriCo(DH)2Lfor diffee, zt planar L ligands. The starred point corresponds to the 1-methylimidazole derivative, for k>hich the orientation (b)in Figure 5 is found. Im = imidazole apparent. The Co-py distances and the Co-N(axia1) distances in cobaloximes having different L ligands are linearly related.30 The Co-N(axia1) distances are also influenced by the bulk of the L ligand (steric cis-influence).For the same alkyl group with different planar N-donor ligands, the order of increasing Co-N(axia1) bond length is shown in Figure 4. This order is the same as that for the rate of dissociation of L (see below). The value of the C-N(axia1)-C angle has been used 31 to determine the steric effect of the N-ligand bulk on the Co-N(axia1) distance. This angle of ca. 105" in imidazole derivatives allows distances shorter than those found in py analogues, where it is about 120". However, when bulky side-groups are attached to an a-carbon atom of L, such as in 1,5,6-trimethylbenzimidazole(Me3Bzm) and 1,2- dimethylbenzimidazole (1,2-MezIm), the interaction with the equatorial moiety provokes a lengthening of the Co-N axial bond as well as a marked non-equivalence of the two Co-N(axia1)-C angles (see Figure 4).Finally, another factor has been claimed32 to influence the Co-N(axia1) distance. In all the 30 S. Geremia, Thesis, University of Trieste, 1988. 31 J. P. Charland, E. Zangrando, N. Bresciani Pahor, L. Randaccio, and L. G.Marzilli, to be published. 32 W. 0. Parker, E. Zangrando, N. Bresciani Pahor, L. Randaccio, and L. G. Marzilli, Inorg. Chem., 1988,27,2170. 235 Structural Properties of Organocobalt Coenzyme B12 Models cobaloximes, planar L have almost the same orientation with respect to the equatorial moiety, as shown in Figure 5a. The only exceptions are 1-MeIm-Co(DH)2Me and ~-M~I~CO(DH)~CH~CH~CN (1-MeIm = 1-methylimidazole), where orientations close to the ideal sketched in Figure 5b are found.The (b) orientation, which is probably dictated by crystal packing forces, is compatible with the intramolecular steric requirements when L is a non-bulky ligand such as 1-MeIm. In Figure 4,the starred point represents the value of 2.058(5) 8, in 1-MeImCo(DH);?Me. This value would be expected to be similar to that of 2.019(3) 8, in the imidazole analogue, but steric clashes in orientation (b) lead to the longer Co-N bond. It was already reported that in PR3Co(DH)2X (X = Me, CHZCN, Cl) the Co-P bond length is linearly related to the Tolman cone angle.3 The Co-P distances relative to the phosphines reported in Table 3 show a similar trend when R = (R)-CH(CN)Me. Axial CO-P~~ distances, measured in and CO-N~~ (GH)* complexes for different R groups, are not significantly different from those reported for (DH)2 analogue^.^ The Co-C distance is primarily influenced by the bulk of the alkyl group and increases for a given L by 0.2 8, going from Me to adamantyl derivatives, as shown in Figure 6.Data collected in Table 3 confirm also the previous findings3 that the Co-C bond length is minimally affected by the nature of the N-donor ligand. Furthermore, for N-donor ligands, good linear relationships are found when Co-C distances in py complexes with different R groups are plotted against the analogous distances in 1,2-Me2Im, Me3Bzm, and NH2Ph derivative^.^'.^^ When L is a P-donor ligand, a slight but significant increase of the Co-C bond is observed compared to N-ligands (Figure 6).Since the extent of the increase depends also upon the bulk of the alkyl group (see Figure 6), this lengthening has been attributed to the steric trans-influence exerted by L on the trans Co-R bond through the equatorial ligands. This influence is supported by the trend of the cc values shown in Figure 6, which vary 33 P. J. Toscano, T. F. Swider, L. G. Marzilli. N. Bresciani Pahor, and L. Randaccio. Inorg. Chem., 1983. 22, 3416. 34 N. Bresciani Pahor. L. Randaccio. E. Zangrando, and P. J. Toscano, Inorg. Chim. Acfn, 1985,96, 193. Randaccio, Pahor, Zangrando, and Marzilli ~ Me3Bzm R= Me 2 .oo---P(OMe)2Ph----P(0Me =\‘PChx, \P(0Me )Ph2R = CH2CMe3 2.05--PY -PMe3 2.10--PPh, R = adarnantyl 2 .15, -1MeImd NH2Ph ‘4Me2N-Py -Me3Bzm 2.20--P(OPri), CY co-c -10 -5 +5 +10 Figure 6 Trend of the Co-C bond length (A) and of the a angle (“) in cobaloximes having R = Me, CH2CMe3,and adamantyl and dfferent L ligands with the bulk of both L and R, although crystal packing effects may also play a role.We have assumed that the Co-C bond length is determined by the bulk of R and this appears evident from previous data3 and from data of Table 3. However, since many other structures with potentially bulky electron-withdraw-ing groups have now been determined, we observe that the substitution of H by C1 at Ca does not provoke any lengthening of the Co-C bond. This could be ascribed to an ‘effective’ bulk of C1 very close to that of H, because the C-C1 distance (1.8 A) is longer than the C-H distance (1 A).However, it could also be Structural Properties of Organocobalt Coenzyme B12 Models attributed to an electronic effect of the electron-withdrawing Ca substituents, such as C1, which could shorten the Co-C bond.35 Particularly short values of the Co-C bond lengths are also found for the CH2CHzCN group (2.00 A), when compared to those of Et (2.04 A). However, for other even stronger electron- withdrawing substituents, such as CN or COOMe, an increase of the Co-C distance with respect to Co-Me is observed, as for CH2CN (Table 3) or CH2COOMe.3 Further investigation is needed to clarify this problem. It was previously observed3 that the steric interaction of bulky R with the (DH)2 moiety also provokes severe distortions in the alkyl group geometry.For example, in CH2CMe3 derivatives the Co-CH2-C angle has values up to 130°, very far from the ideal tetrahedral value. Also, in Et complexes the C-C bond length is shortened, whereas the Co-CH2-CH3 angle is opened up to 120". It is also interesting to note that in adamantyl derivatives the lengthening of the C-C distances nearly parallel to the Co-Ca bond corresponds to the shortening of those involving CX as well as those of the top six-membered ring of adamant~l.~~ These deformations may be interpreted in terms of the anomeric effect.37 Comparison of ~YCO(DH)~R = Me, Et, Pr') with (GH)2 analogues does not (R show significant differences in Co-C bond lengths.34 C.Solid State Reactions.-Ohashi et al.have found 38 that the chiral -CH(Me)CN group in LCo(DH)2CH(Me)CN is both racemized and isomerized by X-ray exposure without degradation in crystallinity. Since the rate of isomerization is closely related to the packing of the R group in the crystal structure, these authors have defined a 'reaction cavity' for this group and have shown that the reactivity is correlated with the volume of this cavity.39 On the other hand, the chiral -CH(Me)CO*Me derivatives with L = 4-Clpy and py were racemized only at high ten~tie~atur-~. The single crystal X-ray structural analysis has shown that the 4-Clpy coniplex undergoes cooperative configurational and conforma- tional changes at high temperature.In the methanol-solvated py analogue, a rapid desolvation at room temperature is accompanied by conformational change of the chiral group to fill the space voided by solvent loss and at 343 K random inversion OCCU~S.~~*~~On the other hand, for L = 4-CNpy, (R)-and (S)-a-methylbenzylamine, no indication for such a reaction was obtained. Their crystal structures correlate with this ina~tivity.~' 35 Q. Chen, L. G. Mm4Ii, N. Bresciani Pahor, L. Randaccio, and E. Zangrando, Inorg. Chim. Aetn, 1988,144,241. 36 N. Bresciani Pahor, W. M. Attia, L. Randaccio, C. Lopez, and J. P. Charland, Acra Crystallogr., Sect. C, 1987,43, 1484. 37 A. J. Kirby, 'The Anomeric Effect and Related Stereoelectronic Effects at Oxygen', Springer Verlag, Berlin, 1983.38 T. Kurihara, A. Uchida, Y. Ohashi, Y. Sasada, and Y. Ohgo, J. Am. Chem. Soc., 1984, 106, 5718 and references therein. 3y Y. Ohashi, A. Uchida, Y. Sasada, and Y. Ohgo, Acta Crystallogr., Sect. B, 1983,39, 54. 40 T. Kurihara, A. Uchida, Y. Ohashi, and Y. Sasada, Acra Crystallogr., Sect. B, 1984, 40, 478; A. Uchida, Y. Ohashi, Y.Sasada, and Y. Ohgo, Acta Crystallogr., Secr. B, 1984.40.473. 41 Y. Sasada and Y. Ohashi, J. Mol. Struct., 1985, 126,417. Randaccio, Pahor, Zangrando, and Marzilli D. Kinetic Studies.-Rate and mechanistic studies on cobaloximes have been extensively reviewed and we report here only the most recent results obtained for organocobaloximes. It is now well established that for reaction 1: LCo(DH)zR + L’ ---+ L’Co(DH2)R + L (1) the rate-determining step is the dissociative first step in the mechanism given in reactions 2 and 3: CO(DH)~R+ L’ klL’CO(DH)~R (3) and kobs = k.The values of logk for LCO(DH)2R complexes, with different L and R groups, are given in Table 4. For a given L, logk. increases with the increase of the o-donor power of R by several orders of magnitude, e.g. for L = 4-CNpy, the rate increases by lo6 in going from CH(CN)CH2CN to CH(Me)Pr’. Interestingly, the trend of lo&, for different L ligands is very similar when R is varied. In fact, good linear relationships are found when logk for 4-CNpy is plotted against logk for the other neutral L ligand~.~’This similarity indicates that the changes in rate are not significantly affected by the nature of L when the R group is varied.On the other hand, for a given L, logk increases with the increase of the o-donor power of R and parallels the increase of the Co-L distance. Therefore, the kinetic trans-effect and the structural trans-influence have very similar trends. The values of logk for complexes with the same alkyl group and different planar L reflect the bulk of L. The trend of logk for the same R is 1-MeIm < Me3Bzm z py < 1,2-Me21m < 2-NH2py and is very similar to that of the Co-N(axia1) bond lengths (Figure 4). This suggests that the bulk of the L planar ligands affects both kinetics and ground state properties in a similar way. In addition, since logk values for 1-MeIm do not show irregularity, it may be concluded that the ‘anomalous’ orientation and the Co-N(axia1) bond lengths found in the structures of l-MeImCo(DH)zR, with R = Me and CH2CH2CN, are a conse- quence of the crystal packing, as suggested above.The logk values for LCO(DH)~R (R = CHC12, CHBr2, and different L) reported in Table 4 provide some additional inf~rmation.~~ First, substitution of H in methyl derivatives with C1 or Br provokes a significant decrease of k. The effect is greater in the Br derivatives, so that the donating ability of CHBr2 is smaller than that of CHC12. Furthermore, the noticeably increased bulk of 2-NH2py with respect to py is mainly responsible for the difference of about two orders of magnitude in the corresponding k value. Analogously, the larger dissociation rates for NH2Ph than py analogues can be mainly attributed to the longer Co-NH2Ph bond length.Structural Properties of Organocobalt Coenzyme B, 2 Models Table 4 Values of log kfur LCO(DH)~Rcomplexes" R/L 4-CNpyb anisidine' Me3Bzmd l-Meim DEA' CH(CN)Cl -4.85 * CH(CN)CH2CN -4.47 * CHzNOz -5.37* -3.42 CHBrz -3.28* (h) CHzCN -4.52 -3.77 -6.00 CHClz -3.08 * ' -2.36' -4.50 -4.40 -0.62 CH(CN)Me -2.88 * CHzCF3 -3.57 -2.96 -4.89 -4.84 -1.55 -0.04 CHzCOzMe -3.57 * -2.62 CH2I -2.79 -2.03 0.95 CHzBr -2.58 -1.76 -3.91 -5.19 -3.77 -0.82 1.02 CH2COMe -3.23 * -2.49 CHZCl -2.51 -1.46 -3.58 1.31 CHzCHZCN -1.59 * -0.8 1 CH2C(C02Et)zMe-0.51 0.03 -1.69 -3.24' -1.62 I .43 Me -1.39 -0.33 -2.38 -3.60 -2.27 0.54 2.12 CHzSiMe3 -0.37 0.15 -2.96 CH2Ph -0.48 0.45 -1.66 3.16 Et -0.02 0.89 -0.95 -2.18 -0.81 1.76 3.31 Pr" 0.08 0.93 CHzPri 0.15 1.11 -2.00 2.00 CHzCMe3 I .04 1.54 0.09 -1.28 0.00 CHzOMe 1.38 * 2.18 0.71 Pr' 1.43 0.58 -0.80 0.45 CH(Me)Et 1.60 -0.72 c-C~HI 1 1.59 0.79 CH(Et)z 2.00 CH( Me) Pr 1.92 adamantyl 1.61 Values are from ref.3 unless otherwise specified. The entering ligand is P(OMe)3. * Starred data are from ref. i of Table 3. Ref. d of Table 3. Ref. 31. Unpublished results from Emory University. Ref.f of Table 3. R. Dreos-Garlatti, G. Tauzher, and G. Costa, Inorg. Chim. Acra, 1984, 82, 197, the entering ligand is thiourea in HzO solvent. For py logk = -4.8 and for NHzPh logk = -2.31, ref. 35. For py logk = -4.34 and for NHzPh logk = -1.73, ref.35. E. Co-C Bond Dissociation Energy.-Co-C BDEs have usually been estimated by kinetic radical trapping procedures. However, for cobaloximes, two methods have been developed which do not require as many assumptions. This contrasts with the situation in other model systems. In 1979, Halpern l3 introduced the first non-kinetic method to determine cobalt-alkyl bond dissociation energies which was based on the measurement of the rates and activation enthalpies of reactions of the kind shown in equation 4. Randaccio, Pahor, Zangrando, and Marzilli LCo1"(DH)2CHMePh-LCO"(DH)~+ PhCH=CH2 + 4H2 (4) Values of BDE, derived for cobaloximes with different P- and N-donor neutral ligands and R = CHMePh and CH2Ph are given in Table 5.These data show that in cobaloximes containing N-donors the Co-C BDE increases systematically with basicity of the neutral ligand. A linear relationship is observed between BDE and pKa of 4-Xpy having similar bulk. On '+e contrary, for cobaloximes containing P-donor ligands, 5r both CHMePh and CH2Ph derivatives, the BDE decreases with the increasing size of the phosphine and linear relationships with the Tolman cone angle are found.42 In the latter case the magnitude of this dependence is such as to mask the apparently much smaller influence of electronic effects. In fact, for the octaethylporphyrin (OEP) series, LCo(OEP)CH2Ph, data reported in Table 5 reveal a linear dependence of BDE on the pKa of the phosphine, similar to that found for 4-Xpy, but not on the variation in bulk of the phosphine.More recently, Toscano and co-workers 43 have exploited the 12 cleavage of Co-R bonds to give CoI + RI to measure the BDE by a calorimetric method. This has allowed the first determination of a Co-Me BDE, specifically in pyCo(DH)2Me. This Co-C has a BDE of 34.6 1.4 kcal mol-', in contrast to the Pr' analogue which has a BDE of 21.3 f2.0 kcal mol-'. The BDE trends agree with the bond length results for cobaloximes. In fact, the increasing bulk of the neutral ligand trans to R weakens and lengthens the Co-C bond (steric trans-influence). However, no structural evidence of the influence of the neutral ligand basicity on the Co-C distance has been detec- ted. This influence should be too small to be revealed by distance measure- ments, but large enough to be detected from BDE measurements.Further-more, the bulkier the R group, e.g. Me us. Pr', the weaker and longer the Co-C bond. Comparison of BDE data for (DH)2 and OEP series suggests that the porphyrin ligand is not sufficiently flexible to respond to the steric pressure of bulky neutral L ligands. Therefore, it cannot bend towards the alkyl group to weaken the Co-C bond, as occurs in cobaloximes. This effect (see above) is apparent in Co-C bond lengths only when large variation in the bulk of L occur, i.e. differences are detected when HzO or py complexes are compared with PR3 analogues. The lack of flexibility of the OEP equatorial ligand should represent a 'barrier', which almost cancels the steric trans-influence as measured from BDE.42 The flexibility of the Co(DH)2 system is supported by the range of a and d values reported in Table 3. The solid-state thermolysis of a series of 24 H~OCO(DH)~Rand 19 pyCo(DH)2R complexes has been studied by differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) technique^.^^ In the most favour- 42 M. K.Geno and J. Halpern, J. Am. Chem. Soc., 1987,109, 1238. 43 P. J. Toscano, A. L. Seligson, M. T. Curran, A. T. Skrobutt, and D. C. Sonnenberger, Inorg. Chem., 1989, 28, 166. 44 K. L. Brown, G. W. Young, R. Segal, and K. Rajeshwar, Inorg. Chim. Acta, 1987,128, 197. 241 Structural Properties of Organocobalt Coenzyme B12 Models Table 5 Co-C BDE (kcal mol-') for LCo(DH)zR with different L ligands and R = CHMePh and CHZPh.Datu relative to LCo(OEP)CH2Ph are also given La LCo(DH)zCHMePh Lb LCo(DH)zCHMePh 4-NHzpy 21.2 PMe2Ph 24 4-Mepy 20.1 P(n-Bu) 21 PY 19.5 P(CH2CHZCN)3 20 4-CNpy 17.9 PEtPh2 19 Imidazole 20.8 PPh3 17 L' LCo(DH)ZCH2Ph LCo(OEP)CH2Ph PMe2Ph 30.4 27.1 P(Bu")3 28.9 29.3 PEtPh, 26.8 26.1 PPh3 25.8 23.8 P(c-C~H~1)3 22.8 29.6 a Ref. 11. * Ref. 12. Ref. 42. able cases, three transitions ascribable to the loss of L, R, and decomposition of the Co(DH)z unit, are observed. The temperature of the first transition for aquoderivatives correlates well with: (i) the Taft constant o*;(ii) logk for the substitution reaction of H20 by pyridine, as well as for the reverse reaction; (iii) the Co-N(axia1) distances of pyridine analogues. 4 Cationic Organometallic Complexes having a Mixed Oxime-Schiff Base System as Equatorial Ligand In 1969, Costa et al.45 prepared new coenzyme B12 models, namely the organocobalt complexes having the tetradentate ligand N2,N2'-propanediyl-bis(2,3-butanedione-2-imine-3-oxime),[(DO)(DOH)pn], as equatorial ligand.These 'Costa' models differ from cobaloximes by having substituted one of the two oxime bridges with a propylene bridge (Figure 7). Thus, they are a mixed Schiff base-oxime system, with the equatorial ligand having the same -1 charge as the B12 corrin; they are typically ionic compounds with complex cations and PF, or C104 anions.Until recently, only one relatively imprecise structure had been reported, namely { H~O[CO(DO)(DOH)~~]M~)C~O~.~~A neutral di-organometallic complex Me[Co(DO)(DOH)pn] Me was reported 47 a few years later. More recently several { L[Co(DO)(DOH)pn]R}PF6 complexes have been characterized by X-ray analysis in our laboratorie~.~ 1,48-53 We have also prepared and characterized several complexes containing a 2,2'-dimethyl-propylene bridge. 54 A. Comparison with the Structure of Coba1oximes.-As already found in cobaloximes, the bond lengths and angles in the equatorial moiety of the 45 G. Costa, G. Mestroni, and E. de Savorgnani, Inorg. Chim. Acla, 1969,3, 323. 46 S. Bruckner, M. Calligaris, G. Nardin, and L. Randaccio, Inorg. Chim.Acta, 1969,3,278.47 M. Calligaris, J.Chem. Soc., Dalton Trans., 1974, 1628. Randaccio, Pahor, Zangrando, and Marzilli -L Me Me R 1 Figure 7 complex cations are not significantly influenced by the axial ligands. In contrast to cobaloximes, where the Co-N equatorial distances are essentially all equal [mean 1.8901(9) the mean distance of 1.912(8) 8, involving the iminic nitrogen donor, i.e. on the propylene bridge side, is significantly longer than that of 1.880(9) 8, involving the oxime N donor. Correspondingly, the N(l)-Co-N(2) angle becomes slightly narrower (mean 97.6’) and the N(3)-Co-N(4) angle slightly larger (mean 99.2’) than that found in cobaloximes (mean value 98.57(8)’). The other N-Co-N angles average to a value of 81.5(4)”, which does not differ significantly from that of 81.38(5)’ reported for cobalo~imes.~ This small variation in the coordination parameters is in agreement with a slight, but significant, decrease of the 0 0 distance in Costa models [mean 2.45(1) A] as compared with the mean value for cobaloximes [2.487(2) A].The difference in -0 distance should parallel the ease of the bridge deprotonation in the two types of c~mplexes.~ This suggestion is supported by the slower exchange of the proton in the oxime bridge of [(DO)(DOH)pn] complexes as compared to that of (DH)Z complexes with similar axial ligand~.~.” The mean values of the other bond lengths and angles agree with a structure having essentially double C=N and single (Me)C-C(Me) bonds (Figure 7), as are also found in cobalo~imes.~ The six-membered chelate ring, involving the propylene chain, has the 48 L.G. Marzilli, N. Bresciani Pahor, L. Randaccio, E. Zangrando, R. G. Finke, and S. A. Mayers, Inorg. Chim. Acta, 1985, 107, 139. 49 E. Zangrando, W. 0. Parker, N. Bresciani Pahor, L. B. Thomas, L. G. Marzilli, and L. Randaccio, Gazz. Chim. Ital., 1987, 117, 307. 5” W. 0. Parker, N. Bresciani Pahor, E. Zangrando, L. Randaccio, and L. G. Marzilli, Inorg. Chem., 1985,24,3908. 51 W. 0.Parker, E. Zangrando, N. Bresciani Pahor, L. Randaccio, and L. Marzilli, Inorg. Chem., 1986, 25, 3489. 52 E. Zangrando, W. 0.Parker, and A. Mezzetti, Acta Crystallogr., Sect. C, 1987,43,2211. 53 W. 0. Parker, N. Bresciani Pahor, E.Zangrando, L. Randaccio, and L. G. Marzilli, Inorg. Chem., 1986,25, 1303. 54 P. G. Yohannes, N. Bresciani Pahor, L. Randaccio, E. Zangrando, and L. G. Marzilli, Inorg. Chem., 1988,27,4738.’’R. J. Guschl and T. L. Brown, Inorg. Chetn., 1974, 13,959. Structural Properties of Organocobalt Coenzyme BIZ Models expected conformation with the central C atom out of the chelate plane, normally on the side of the less bulky axial ligand. The deviation from planarity of the [Co(DO)(DOH)pn] unit may be described by the bending angle, a, between the two chemically equivalent moieties. The values reported in Table 6 appear related to the bulk of the axial ligands and are larger than those observed in cobaloximes3 (Table 3) having the same axial ligands (see below).The axial Co-C and Co-L bond lengths follow trends similar to those found in cobaloximes. Acomparison is given in Table 6 for some L ligands. For example, in aqua-derivatives, the Co-C bond lengthens with the increasing bulk of the alkyl group. A small increase of the Co-C distance is also observed going from L = OH2 to L = PR3, i.e. approximately in the order of increasing bulk of L. Correspondingly, for alkyl groups with similar bulk, the a values range from negative values (bending towards L) in H20 to large positive values (bending towards R) in MeJBzm derivatives. As already observed in cobaloximesS6 a lengthening of the Co-C bond is accompanied by increasing bending and cobalt displacement towards the L group, induced by the bulk of the trans neutral ligand (steric trans-influence).Asin cobaloximes, for the same L the Co-L bond lengths (Table 6) reflect the o-donor power of the alkyl group. For planar N-ligands, the orientation of their plane with respect to the equatorial moiety in all the structures is close to the ideal one sketched in Figure 5b, i.e. the planar ligand almost crosses the two equatorial five-membered chelate rings, As we described above, the orientation of the planar L in cobaloximes, with the exception of two 1-MeIm derivatives, is that of Figure 5a. Comparison between cobaloximes and Costa models suggests that the orientation of the planar ligand is mainly dictated by the interaction of the axial ligand with the equatorial one.The orientation (a) in Figure 5 is the one most favoured in cobaloximes, but the substitution of one oxime bridge with a CH2-CH2-CH2 chain increases the interaction with the axial ligand oriented as in (a), so that it is forced into orientation (b). This reorientation provokes a significant increase of both Co-N(axia1) bond lengths and 0: bending angles in Costa models when compared with those in cobaloximes having the same axial ligands. Similar differences in Co-OH2 bond lengths are observed in the comparison of H20 derivatives, but no significant difference is detectable in PR3 complexes. As for cobaloximes, the geometry of the CH2X group bonded to Co is significantly modified, with an opening of the Co-Ca-C angle up to 130" in CH2CMe3 and a shortening of the Ca-C bond length down to ~1.408, in Et complexes.These two effects are more pronounced when L is bulky.32 B. Kinetics.-In non-coordinating solvents, the L exchange in { L-Co[(DO)(DOH)pn]R}X complexes follows a first-order rate law, consistent with LIM mechanism, as for cobalo~imes.~ an SN~ Results for complexes with L = py 56 N. Bresciani Pahor, M. Calligaris, G. Nardin, and L. Randaccio, J. Chem. SOL..,Dalton Trans., 1982, 2549. 244 Table 6 Compurr\oti of Co-C, Co-L, d und x hetiteen L[Co(DO)(DOH)pn]R und LCO(DH)~Rcomplexey. The firyt number in each entrv referv lo (DO)(DOH)pnderri utrrcy. rhr \e(ond to tobuloximer For the Iutior, dutu ure from ref: 3 und from Table 3 L R co-c (A) co-L (A) d (A) ("1 Re6X H 20 Me 1.977(4) 1.990(5) 2.103(3) 2.058(3) 0.01 0.00 2.0 -4.0 a -bEt 2.020(3) -2.109(2) --0.03 --9.3 -h~~CHzPh 2.052(3) 2.099(1) --0.02 --5.8 -bPr' 2.073(2) -2.133(3) --0.05 --6.6 1 -MeIm Me 2.001 (3) 2.009(7) 2.042(2) 2.058(5) 0.05 0.05 4.8 4.4 C PY Me 2.003(3) 1.998(5) 2.106(3) 2.068(3) 0.07 0.04 6.9 3.2 d CH2Me3 2.083(4) 2.060(6) 2.121(3) 2.081(4) 0.03 0.00 14.3 -5.2 d Me3Bzm Me 2.0 1 1(3) 1.989( 2) 2.100( 3) 2.060( 2) 0.09 0.06 13.8 4.7 e -Et 2.041(4) -2.105(3) -0.10 --16.7 e 1,2-Me21m Me 2.003(5) 2.001(2) 2.1OO(4) 2.086( 1 ) 0.10 0.06 19.7 3.4 c "a 3-NH2Ph Me 1.991(4) 1.992(2) 2.147(3) 2.129( I) 0.00 0.04 -11.3 3.5 .f Et 2.030(4) 2.030(3) 2.I74(3) 2.147(2) 0.01 0.01 -7.1 2.8 g PPh3 Me 2.018(5) 2.026(6) 2.405( 1) 2.418( 1) 0.10 0.11 8.2 14.0 h P(OMe)J Me 2.02 l(5) 2.01 (1) 2.265( 1) 2.256(4) 0.05 0.10 3.3 10.0 h R 9 " Ref.4X. Ref. 49.' Unpublished results. Ref. 50.' Ref. 32. Ref. 51. Ref. 52. Ref. 53 h) PVI Structural Properties of Organocobalt Coenzyme B1 ModeIs 00 R’ Co (ac acen) -CH2CH2- R2 = H Co(salen) -CH(Me)CH2- R2 = H Co(salpn) -CHzCHz- R2 = F Co(3F -sal en1 R2 = H Co(saloph) n, Co (acsalen) Figure 8 and Me3Bzm and different R groups show that the dissociation rate increases by lo5 in both model complexes going from the weak electron-donor CH2C02Me to the good electron-donating group Pr’. Comparison of logk values with those of the analogous series of cobaloximes containing 4-CNpy and Me3Bzm shows that there is a very good linear relationship in both cases, with slope close to 1.Furthermore, the reactivities of the Costa-type complexes are 10 times greater, on average, than those of the analogous cobalo~imes.~~ This result is consistent with longer Co-N(axia1) distances and larger a bending angles in the former complexes than those found in cobaloximes (see previous paragraph). 5 Schiff-base Complexes The structural properties of cobalt complexes prepared with bis(salicyla1-dehyd0)ethylendiimine (salenH2) and bis(acety1acetone)ethylendiimine (aca-cenH2) were reviewed in 1972.57Since then, the structure of the free ligands has been reported as well as many other structures of Co complexes having the tetradentate dianionic salen equatorial ligand or its derivatives (Figure 8).58 57 M.Calligaris and G. Nardin, and L. Randaccio, Coorcl. Chem. Rev., 1972,7. 385. 58 M. Calligaris and L. Randaccio in ‘Comprehensive Coordination Chemistry’, ed. G. Wilkinson, R. Gillard, and J. McCleverty, Pergamon Press, Oxford, 1987, Vol. 2. p. 7 15. Randaccio, Pahor, Zangrando, and Marzilli Table 7 Axial Co-L and Co-C bond lengths (A) in octahedral cobalt(nl) complexes of tetradentate Schiff bases co-L co-c Ref: [MeCo(saloph)] 2 2.435(4) 1.963(7) a [EtCo(salen)]2 2.342(3) 1.990(7) b PYCCo(salpn)lCH(CN)~ 2.060(9) 2.02( 1) c py[Co(saloph)]CH2CN 2.098(4) 2.000(5) d py [Co(saloph)] C H C F 2.126(9) 1.99( 1) a py[Co(salen)]CH=CHz 2.12(1) 1.93(2) b py[Co(saloph)] Et 2.2 14(4) 2.042( 6) d H20[Co(3-F-salen)]CHzCOMe 2.13(1) 2.01(1) e H2O[Co(acsalen)] E t 2.2 19(4) 1.996(6) f MeOH[Co(salen)]CH2COMe 2.202(9) 2.02( 1) b py[Co(acacen)] Me 2.16( 1) 1.99( 1) b H20[Co(acacen)]CH=CHz 2.221(7) 1.89(1) b * Ref.59. Ref. 57. N. A. Bailey, B. M. Higson, and E. D. Mckenzie, J. Chem. SOC.,Dalton Trans., 1975, 1105. Ref. 63. W. P. Schaefer, R. Waltzman, and B. T. Huie, J. Am. Chem. Soc.,1978,100,5063. M. H. Darbien, F. Dahan, J. P. Costes, J. P. Laurent, and G. Cros, J. Chem. SOC.,Dalton Trans., 1988, 129. A. Structural Properties.-Organocobalt Schiff base complexes have been found as penta-and hexa-coordinated species. It should be noted that sometimes potentially pentacoordinate species XCo(che1) dimerize in the crystal as [XCo- (chel)l2 reaching six-coordination through a long bond from the cobalt of one Co(che1) unit to the oxygen of the other unit.57 In Co"' pentacoordinate complexes the Co-C bond of 1.96(1) A in the Me derivative is significantly shorter than that of 2.031(8) 8, in the Pr' analogue.This difference is strongly indicative of the greater bulk of the Pr' group compared to that of the Me, a relationship already observed in hexacoordinate cobaloximes. Comparison with the corresponding hexacoordinated complexes reported in Table 7 indicates a slight but significant increase of the Co-C bond lengths in the latter.59 This result appears to be in agreement with the lower BDE found in the base-on than in the base-off forms of BIZ coenzyme, but the coordination number of the base-off species is not known.Since the homolysis leads to Co" species the structural features of Co" compounds are of some interest. Although data for pentacoordinate Co" complexes are limited, it appears that high-spin species, for which a significant increase of the basal coordination distances is found6' with respect to the low- spin complexes, have shorter axial bonds. This feature may have implication in the mechanism of the Co-C bond cleavage in the coenzyme. In hexacoordinate Co"' complexes (Table 7), more data are available for axial distances and the following relationships may be derived. The Co-C bond length (i) is not influenced significantly by the nature of the equatorial Schiff base, (ii) J. Am. Chem. SOC.,1985,107,6880. 59 L.G. Marzilli, M. F. Summers, N. Bresciani Pahor, E. Zangrando, J. P. Charland, and L. Randaccio, 6o M. Calligaris, G. Nardin, and L. Randaccio, J. Chem. SOC.,Dalton Trans., 1974, 1903; B. J. Kennedy, G. D. Fallon, B. M. K. C. Gatehouse, and K. Murray, Inorg. Chem., 1984,23,580. Structural Properties of Organocobalt Coenzyme B12 Models increases with the increasing bulk of the R group, but (iii) depends strongly upon the hybridization of the carbon atom bonded to cobalt, as already observed in cobaloximes (see above). The Co-L distances are strongly dependent upon the nature of the trans-ligand. In fact, the Co-py bond length increases from 2.060(9) to 2.215(4) 8,in the order of the trans ligands CH(CN)2 < CHzCN < CH~CFJ2 CH=CH2 < Me < Et irrespective of the Schiff base ligand.This trend follows the increasing o-donor power of the R group, with the exception of the pyCo(salen)CH=CHz derivative. This exception could indicate that the nature of the equatorial Schiff base may influence Co-L distances, but this contrasts with the observation that the Co-OH2 bond lengths in HzOCo(acsa1en)Et C2.219(4) A] and HzOCo(aca- cen)CH=CH2 [2.221(7) A] are very close, as expected on the basis of the 0-donor power of Et and CH=CH2 groups. Nevertheless, it may be concluded that both electronic trans-influence and steric cis-influence, originated by the steric interac- tion of the axial ligand with the equatorial moiety, are reflected in bond length trends of Schiff base complexes in the same way as in those of cobaloximes.However, the Co-L distances in Schiff base complexes are significantly longer than those in other models having the same R group, so that the order of increasing Co-L distances is: (DH)2 < (DO)(DOH)pn < Schiff base < corrin B. Solution Properties.-Since the general review of Costa et aL61 and the recent review on synthesis and properties of similar B12 models,62 relatively little work has been reported on organocobalt Schiff base complexes. Recently, Halpern et a1.14' applied the kinetic method to estimate the BDE of the Co-C bond in the pyCo(sa1oph)R series. Values of 25, 22, 20, 18 kcal mol-' were obtained for R = Pro, CH2Ph, Pr', CH2CMe3, respectively. These data confirm that steric factors may play an important role in promoting the Co-C bond homolysis and show that the BDEs follow a trend similar to that found for Co-C bond lengths.Dynamic NMR measurements of L dissociation rates of 3,5-lutidineCo-(sa1oph)R compounds reveal rates 10'' times larger than those in the cobaloxime analogues. This enormous cis-effect is partly a ground-state effect, which follows the same trend of the cis-influence and partly arises from the higher stability of saloph pentacoordinate intermediate compared to that of cobaloxime~.~~The latter point is illustrated in the unusual example of both dimeric and monomeric molecules in the same crystal of the MeCo(sa1oph) compound. Furthermore, the rate constant increases in the order CH2CN (2.25 x lo4 s-') < CH~CFJ 61 A.Bigotto, G. Costa. G. Mestroni, G. Pellizer, E. Reisenhofer, L. Stefan;, and G. Tauzher, Inorg. Chim. Acra, Ret.., 1970,4,41. 62 P. J. Toscano and L. G. Marzilli, Prop. Inorg. Chrw., 1984,31, 105. 63 M. F. Summers. L. G. Marzilli, N. Bresciani Pahor. and L. Randaccio, J. Am Chem. Soc., 1984, 106, 4478. 248 Randaccio, Pahor, Zangrando, and Marzilli (2.76 x lo5 s-’) < Me (2.75 x lo8 s-I). This dependence on electron donation by R is greater than in other classes of B12 compounds (see above).59 The quantitatively very different solution properties of organometallic Schiff base and cobaloxime complexes are in good agreement with trends expected on the basis of structural properties. The Schiff base complexes appear to have properties closer to Cbls than do cobaloximes, but unfortunately they are more difficult complexes to prepare and study.Probably, for both electronic (i.e. they are ‘electron rich’ systems closer to Cbls than ‘electron poor’ cobaloximes) and steric reasons [i.e. the equatorial ligand is more flexible than the (DH)2 moiety] stable pentacoordinated organocobalt Schiff-base species can be characterized. Such species have been identified only as intermediates in cobaloximes. In base-off Cbls and in Cbi’s, five-coordination is difficult to assess since these species are dissolved in H20,which may occupy the trans-axial position. 6 Conclusions In this review, we have focused on a comparison of the most extensively studied model system for B12 coenzymes, namely cobaloximes.The structural properties of cobaloximes in the solid state have been very useful in interpreting solution data. The trend in structural data can be compared with Cbls on the one hand and Costa type and Schiff-base type models on the other hand. In general, these properties suggest the following trend: CO(DH)~z Co(DO)(DOH)pn 6 Cbls d Co(Schiff-base) The Co-C bond energies decrease for closely analogous species following this general trend, although the differences are not large. Indeed, relatively little difference is found in Co-C bond lengths for a given type of R-group. The more important structural factor is the length of the Co-C bond that reflects roughly the relative Co-C cleavage rates given in Section 2B above.A second important aspect of our findings is that distortions induced by L also lengthen and weaken the Co-C bond. These findings support the concept espoused by many groups (see reference 4) that the protein may distort the coenzyme and thereby labilize the Co-C bond. Interestingly, this suggestion was first made nearly two decades ago.64 On the other hand, models have shown that the distortion of the Co-C-C angle does not affect substantially the stability of the organometallic complex, since neopentylcobaloximes, where this angle is found to be 1 30°, are particularly stable. In contrast to the similar Co-R bond lengths for a given R with different equatorial ligands, Co-N bond lengths are quite different and these as well as L dissociation rates indicate that Co-N bonding is weakest in the Schiff-base compounds.Equatorial ligand flexibility is more difficult to assess and does not necessarily strictly follow the trends in bond energies and bond lengths noted above. In this regard, the porphyrin analogues deserve greater scrutiny. The 64 H. A. 0.Hill, J. M. Pratt, and R.J. P. Williams, Ckem. Brit., 1969, 5, 156. Structural Properties of Organocobalt Coenzyme B12 Models solid-state information has laid a foundation of parameters that are likely to be important for interpreting NM R data and for developing methods of calculating conformations of Cbls. Indeed, new two-dimensional NMR methods for gaining insight into solution structure are usually most powerful when combined with molecular mechanics calculations. Therefore, despite their remoteness from Cbls in terms of physical properties, we expect cobaloximes to continue to play a central role in the development of methods for understanding B12 properties, especially the factors influencing the Co-C bond. Acknowledgement. We thank the MPI (Rome), CNR (Rome), and the NIH (GM 29225) for support.

ISSN:0306-0012    

DOI:10.1039/CS9891800225

出版商:RSC    

年代:1989

数据来源: RSC

摘要:

Chem. Soc. Rrc., 1989, 18, 25 1-283 Oxidative Coupling of Methane using Oxide Catalysts By G. J. Hutchings* LEVERHULME CENTRE FOR INNOVATIVE CATALYSIS. D E PART M E NT 0F CH E M I STR Y. U N I V E R SIT Y 0F LIVER PO 0 L, PO BOX 147. LIVERPOOL L693BX M. S. Scurrell DEPARTMENT OF ENERGY TECHNOLOGY. CSIR, PO BOX 395. PRETORIA 0001, SOUTH AFRICA J. R. Woodhouse CATALYSIS RESEARCH PROGRAMME. DEPARTMENT OF CHEMISTRY. UNlVFRSlTY OF THE WITWATERSRAND, PO WITS. JOHANNESBURG 2050, SOUTH AFRICA 1 Introduction In the past fifteen years there has been a resurgence in research into the production of chemicals and liquid transportation fuels from fossil fuel sources other than petroleum. This was largely spurred by the increased cost of petroleum and the majority of this effort was focused not only on improving the existing pathways based on the hydrogenation of carbon monoxide 192 but also led to the development of a new commercial process based on methanol conversion.1.3,4However, in recent years petroleum has again become inexpensive and this has significantly decreased the attention given to such research.In its place the research emphasis has now been directed at the utili~ation of natural gas as a source of chemicals and fuels. The reasons for this are twofold: first, the current reserves of natural gas are considered as a large underutilized energy resource, but a large proportion of these reserves are remotely located from high densities of consumers and chemical conversion of the methane is required to facilitate transportation.Second, vast quantities of natural gas are currently flared, particularly at locations where the gas is associated with crude oil, and consequently chemical conversion could be both economically and environ- mentally beneficial. The conversion of natural gas ria the process of partial oxidation to obtain more valuable chemical products, c.g. methanol and ethane, is not a new research topic. A large amount of pioneering research was completed in the 1920s and 1930s in which the partial oxidation reaction was investigated as a high pressure gas phase reaction, and this work has recently been reviewed by Gesser (it dSA * To u horn correspondence should be addressed.' G. J. Hutchings, S.A/,-.J. C'/wni..1986. 39. 65.'P. Bilocn and W. H. Sachtler. ,4dr. C'u/d., 1981, 31. 165.'G. J. Hutchings. C%rm. &//.. 1987. 23. 767. C. D Chang. CU/C//.Rcr..-Su. Eir,q..26. 373. ' H. D. Gesser. N.R. Hunter. and C. R. Prakash. C/wm. Rw.. 1985.85 235. 25 1 Oxidative Coupling qf'Metlianr using 0.uide Catalysts small proportion of recent research effort has been focused on the production of methan~l,~.'but to a large extent the identification of catalytic materials for this reaction has been unsuccessful. By far the majority of research has been carried out into the oxidative coupling of methane using oxide catalysts to form higher hydrocarbon products, mainly ethane and ethene. In 1988 three major scientific meetings8 devoted over 60 papers solely to this reaction, which emphasizes the importance given to this topic at the present time.The aim of this review is to highlight the salient features of the recent research literature on the oxidative coupling of methane, particularly that published since 1986. In particular this review seeks to discuss the mechanism of methane oxidation over oxide catalysts and to consider experimental approaches that could enable the design of an improved catalyst. 2 Evaluation of Oxide Catalysts A. Molecular Oxygen as Oxidant.-Initial work by Keller and Bhasin 9.10 considered that non-selective methane oxidation, either gas phase or catalysed, could be dominant when oxygen and methane were co-fed over a catalyst. To minimize such an effect they operated the reaction cyclically, i.e.methane and air were fed individually over the catalyst interspaced by a short purge of an inert gas. In this way methane was reacted with the catalyst which was stoich-eiometrically reduced by the reaction, once all the readily reacted oxygen was depleted the reaction ceased. A range of oxides were identified which became active above 600 "C and produced ethene and ethane as major products. Of the oxides identified, supported manganese oxide was found to give the best results. Later studies '*" demonstrated that cyclic operation was not necessary to obtain high selectivities, and in the presence of suitable catalysts co-fed methane and oxygen also gave high yields of CZhydrocarbons. Subsequently Lunsford identified that lithium-promoted magnesium oxide could give a much improved catalytic performance.These early studies demonstrated that the oxidative coupling of methane to form CZhydrocarbons was a high temperature reaction and temperatures in excess of 600 "C were required to observe selective products (Figure 1). At lower temperatures only COz and CO were observed as carbon- containing products. The early studies failed to observe that hydrogen was also a major reaction product, but subsequent studies noted that significant levels of hydrogen can be present." In general, the products observed in the high N. R. Foster, Appl. Cutnl., 1985. 19. I. R. Pitchai and K. Klier, Ctrttrl.Rar.-Sci. Eng., 1986,28, 13. " First Workshop on the Catalytic Methane Conversion, May 1988, Bochum, Cuiul.7oticrj.. 1989. 4, 271 500: Proc. Ytli /fir. Cong. Cattrl., Calgary, ed. M. J. Phillips and M. Ternan, Chemical Institute of Canada, 1988.2.883 997; Symposium on the Direct Conversion of Methane to Higher Hydrocarbons, Div. Pet. Chem., A.C.S. Symp. Ser., Los Angeles. September 1988.'G. E. Keller and M. M. Bhasin, J. C~titrl..1982. 73. 9. '(I M. M. Bhasin. Srzrti. Sw/. Sc,i.Ct//d.31988. 36. 343.'' W. Hinsen and M. Baerns. Chc,r~.Zcg.. 19x3. 107. 223. '' W. Hinsen. W. Bytyn, and M. Baerns. Proc. #//I1/11.Gong. Cu~td.,Verlag Chemie, Weinheim, 1984, 3, 581. l3 T. Ito and J. H. Lunsford. Na/irro (Loritloir).1985, 314, 721. Is G. J. Hutchings. M. S. Scurrell, and J. R. Woodhouse, AppL Critd..1988.38. 157. 70 60 40 30 600 650 700 750 800 TemperatureR 1 Figure I Ef7i~r.tof' rcucrion tcniperutuw on the purtiul osidutioii of twrhutie mer Lii MgO;0 .sclcctiritj. to C2 cmipoutidv; A sc~lwliritj~to CO und C02: CH4 concersion (Reproduced by permission from Nature, 1985,314. 721) temperature reaction are: C2H6, C2H4, CO, C02, H2, HlO and to a lesser extent higher hydrocarbons (C3 hydrocarbons are typically found at an order of magnitude lower than the Cz hydrocarbons). Oxygenated products, e.g. CH30H and CH20 are found in only trace amounts, which is in contrast to earlier gas phase st~dies.~.'~ Based on these early studies two catalysts are now being evaluated at pilot plant scale. Sofranko rt ri1.'6,'7 have described a cyclically fed process based on a supported manganese oxide catalyst.Under conditions of cyclic feeding they have shown that selectivity and conversion are related by a 'lOOo/, rule' (methane conversion and Cz selectivity roughly sum to loo%, see Figure 2). Co-feeding of methane and oxygen results in a loss of at least 10% in C2 selectivity. It is apparent that methane conversion decreases significantly in only a short reaction time, and hence rapid cycling between methane and oxygen is required for this process to be viable. Edwards and Tyler l8.I9 have described a co-feeding process based on lithium-promoted magnesium oxide operated in a fluidized bed reactor, ''I E. H. Boomer and V. Thomas, C'rr/i. J. Rcc.. Swi. B. 1937, 15, 401. "I C.PI. Jones. J. J. Leonard. and J. A. Sofranko. J. C'trrtd., 1987. 103, 31 1. J. A. Sofranko. J. J. Leonard. C. .4 Jones. A. M. Gaffneq, and H. P. Withers. Cuid Tork/j,,1988. 3, 127. J. H. Edwards and R. J. Tyler. Siud Sw/. S<,i.Cr//rr/..1988, 36. 395. "J. H. Edwards and R. J. Tyler, in Proceedings 1st Workshop on Catalytic Methane Conversion, Bochum 1988, C'r/ic//. Tot/tr,v. 1989. 4. 345. 1 2 3 4 5 Reaction Time (Min.) Figure 2 Mefhcirieomlution it1 flie cihwitc of t o-fd o\jyyvi crf 850 "C cirirl 860 ml CH4 ml cat./h over 15":, Mn, 5",, Na4P202SiOz (Reproduced by permission from J Card, 19x7. 103, 3 1 1) which can also give C2 selectivities in excess of go";,, at low conversions (Figure 3), and so it is clear that cyclic feed operation is not necessary for the attainment of high product selectivity. An important question remains concerning the comparison of methane oxidation using cyclic and co-feed experimental regimes; it has to be considered whether the methane activation process is the same for the two processes. Recent results2' have shown that, in the absence of gas phase oxygen, lattice oxygen in MgO or LiiMgO is inactive for methane oxidation, i.~.the stoicheiometric reaction is not observed, and hence methane oxidation in the co-fed regime must involve reaction with an adsorbed oxygen species. With supported manganese oxide the stoicheiometric oxidation is observed implicating the involvement of lattice oxide ions O2 , possibly at low co-ordinate defect sites as found for ammoxidation catalysts,21 as active centres.Hence, superficially it could be concluded that the methane activation step could be different for the two experimental regimes. However, in the presence of gas phase oxygen an adsorbed oxygen species could also become important for the manganese oxide catalysts, and consequently further experimental work is required to address this question fully. 2 0 G. J. Hutchings, M. S. Scurrell, and J. R. Woodhouse, J. C/icm. Soc,.,Formk/j,Trtrm. I, in press 21 K. Aykam. 0.Halvorson. A. W. Sleight. and D. B. Rodgers. J. Cu/d.,1975. 35. 401. 254 7;,, Li/MgO 1.6 2990 720 42.8 -45.4 19.4 0.57 u,h 7y0 Li/MgO 2.0 2990 720 37.8 -50.3 19.0 0.62 u,h 7% Li/MgO 3. I 2990 720 29.1 -58.1 16.9 0.66 u,h Be0 9.3 31.1* 74G750 9 99 22 2.0 c.~ __ L'MgO 9.3 0.2* 74@-750 12 93 47 5.6 ._ cCaO 9.3 0.1 * 74&750 11 94 55 6.I ~SrO 9.3 23.1* 74G750 8 88 72 5.8 c. BaO 9.3 14.5* 74&750 14 98 70 9.8 -L' 37; Li/CaO 9.3 0.4* 740-750 14 97 72 10.1 c ~3% Na/CaO 9.3 0.3* 74&750 16 95 76 12.2 c, 32, K/CaO 9.3 0.4* 740 --750 14 83 72 10.1 -c, MgO 200 12* 800 -~ -__ 0.1 0.56 rl 20% BiiMgO 100 12* 800 --__ 0.3 3.00 d 15% I00 12* 800 ---2.3 0.6 I Cl $ 1 ~3.5 667 550 3.0 -0 0 c,f;g,h >MgO -.MgO 3.5 667 720 5.2 I00 58.1 3.0 I .83 0,f;gh 5 5% Li/MgO 3.3 648 550 2.0 -22.5 0.45 3.50 P,f;gh ?5% Li/MgO 3.3 648 720 2.7 37.0 65.4 1.76 1.47 Ll,f;g,h c,-Q -i37; LiiBeO 3.0 0.61* 640 19.8 69.6 51.0 10.1 -2 3":, LiiBeO 3.0 0.61* 680 27.5 96.6 52.0 14.3 -i $ c\3?{) Li/BeO 3.0 0.061* 760 28.4 93.3 58.5 16.6 ._ i -r, -3MgO 20 0.75* 750 11.3 39.I 4.4 3.76 i T 5-10;,;, Li/MgO 20 0.75* 750 7.5 .-76.6 5.7 2.73 i $ h)cn '' Ref'. 13. Ref. 38. ' Ref. 28. * K. I. Aika and T. Nishiyama. PI.OC.9rh //I/. Co/ip.Ctrltrl.. Cdgtrr\., ed. M. J. Phillips and M. Ternan. Chemical Institute of Canada. b Q1988, 2, 907. " Ref. 64. f Rcf. 35. 'I Ref. 25. Ref. 36. ' T. Doi, Y. IJtsumi, and I. Matsuura, Pro(,. 9//1//I/. C'o/zgr. Co/tr/.. CdgorJ,.ed. M. J. Phillips and M. Ternan. Chemical Institute of Canada, 1988, 2, 937. j R. Burch, Ci. D. Squire, and S. C. Tsang, AppI. C'utd,. 1988. 43. 105. * denotes W 'F g s ml-I "t I1Nile mole mtio h' "C CH4 02 SeIectii~itj~",, Yield ',, ct I1ctic 4.6 1.O* 650 10 74 8 0.8 1.7 4.6 I .O* 650 9 I00 3 0.3 0.0 4.6 1 .O* 650 II I00 20 2.2 2.3 4.6 I .O* 650 14 99 37 5.2 3.0 4.6 1.O* 650 15 96 37 5.6 2.4 4.6 I .O* 650 13 91 24 3.1 2.7 -10.0 0.22* 740 8.8 40 7.5 10.0 0.22* 780 2.2 -54 6.6 -_ ~10.0 0.22* 7 70 2.I -48 5.8 10.0 0.22* 820 7.8 28 2.2 --10.0 0.22* 793 2.9 -53 6.8 3 7.0 2.4* 725 3.6 88 71.0 2.5 1.5 8.8 2.4* 725 9.4 91 46.6 4.4 1 .o 3.4 2.4* 725 19.7 93 23.7 4.7 0.8 3 .0 900 710 22.7 00 15.8 3.6 2.2 3.3 700 710 10.3 33 49.9 5.1 2.2 3.1 592 710 19.1 00 15.3 2.9 14.3 3.0 857 710 24.5 87 52.9 13.0 0.9 La203 6.0 37 500 750 19.6 100 56.6 11.1 I .2 I Li/La203 6.0 37 500 800 21.6 100 70.2 15.2 1.2 f I", Na/LkizO_i 6.0 37 500 800 20.0 I00 64.6 12.9 1.1 I' loo K/La203 6.0 37 500 800 20.6 100 60.4 12.4 1.0 1' I", Sr/La203 6.0 37 500 750 20.9 100 63.6 13.3 I .2 .fig SmZ03 6.0 37 500 750 12.9 100 53.0 9.5 1.3 .I' Sm203 20.0 0.75* 750 13.3 ~~ 52.3 7.0 2.4 I1 5 -lo',, Li/Sm203 20.0 0.75* 750 9.3 -64.0 6.0 1.5 I1 -iSm203 2.5 0.6* 750 29.0 I 00 40.0 11.6 Li/Sm203 2.5 0.6* 750 37.0 100 57.0 21.1 i l0lA) Na/Sm203 2.5 0.6* 750 37.0 I00 50.0 18.5 -i -iIOO,, K/Sm203 2.5 0.6* 750 37.0 I00 52.0 19.2 IOU, Rb/Sm203 2.5 0.6* 750 37.0 100 46.0 17.0 -i 105, Cs/Sm203 2.5 0.6* 750 37.0 100 45.0 16.6 i ~CeO 2.0 0.56* 750 --0 0 .f' 202, LI/Ce02 2.0 0.56* 750 28.6 -50.0 14.3 -I' 20'2, Na/Ce02 2.0 0.56* 750 33.4 ~ 41.0 13.7 ./' 20':/,K/Ce02 2.0 0.56* 750 56.5 -23.0 13.0 -f' 1 h h/i., 1985, 499.K. Otsuka and T. Nakajima. I~ioq.C/I~I." K. Otsuka, K. Jinno, and A. Morikawa, C/~PIII. A(,/o. 1986. 120. L27. Ref. 68. Ref. 52. " Ref. 35. ' J. M. 2 DeBoy and R. F. Hicks, J. C'hcni. Soc.. C/ICWI.C'ommui., 1988. 982. J. M. DeBoy and R. F. Hicks, J. Cu/d., 1988, 113, 517. R. Burch. G. D. Squire, and S. C. Tsang. 5Appl. C"//d.,1988.43, 105. K. Otsuka, Q. LIu, M. Hatano, and A. Morikawa, C/7~7.Lr,//., 1987, 1835. * denotes W/F g s ml-' c/,5 -3 9 a moa-bdv-==I0III I N +tm N .i -t, u. s NiO 2.0 0.6* 758 29.9 81 0 0 ~ /I LiCI,INi0 2.0 0.6* 750 25.9 57 71.8 18.6 0.27 I1 Li2C03,"i0 2.0 0.6* 750 25.9 70 48.9 12.7 0.85 I? LiNOJiNiO 2.0 0.6* 750 26. I 75 55.8 14.6 1.34 I1 NaCliNiO 2.0 0.6* 750 17.9 44 34.7 6.2 7.20 h 12",, Li/ZnO 2 0.167* 680 16.7 42.5 56.2 9.4 12';,) Li/ZnO 2 0.I67* 720 30.6 81.6 60.5 18.5 122)Li/ZnO 2 0.167* 760 34.6 89.9 59.5 20.6 0.18", Li/ZnO 2 4.8* 650 5. I 11.7 51.1 2.6 3.9 I i &0.IV;/(, Li/ZnO 7 4.8* 700 13.9 25.3 59.0 8.2 1.15 i 0.I87;, LiiZnO 2 4.8* 720 21.4 32.4 52.6 11.3 1.15 i 0.I8",;, Li/ZnO 2 4.8* 750 28.2 55.5 53.3 15.0 0.85 i " Ref. 9. G. S. Lane and E. Wolf, I'm.. 9//i /fir. C'orzg. C'ufd,. C'tdqcir-~.,cd. M. J. Phillips and M. Ternan. Chemical Institute of Canada. 1988. 2. 944. ' Ref. 16. Ref. 17. " K. Otsuka and T. Komatsu, J. C/i~m. C.o/mntti,. 1987, 338. ' Kh. M. Minachev, N. Ya. Usachev. Yu. S. Khodakov. I.. I-. Ko/lov. V. N. Udut. and 0.A.Soc,.,C/i~m. Fomin, Izc. Akorl.Nid. SSSR, Scr-. K/zh., 1987. 504. R. Burch. G. D. Squire, rind S. C. Tsang. Appl. Crc/ctl., 1988. 43, 105. K. Otsuka. Q. Liu. M. Hatano. and A. Morikawa, Inor-g. C'hini. Ac,to, 1986, 118, L24. ' I. Matsuura, Y. Utsumi. M. Nakai. and T. Doi, C/iwi,/.o//..1986. 1981. Ref. 70. * denotes W F g s ml-'. t cyclic operation, no co-fed oxidant 50 I I J 0 10 20 30 METHANE CONVERSION It could be concluded from these early studies that the selectivity to Cz hydrocarbons for any particular catalyst is related to the concentration of gas phase oxygen co-fed with the methane (e.g. see Figure 3). This is not surprising since the initial products resulting from the activation of methane will be considerably more reactive with respect to oxygen and hence excess oxygen will result in consecutive oxidation of the initial products.Since the early studies of Keller and Bhasin and Lunsford a large scale catalyst screening and evaluation exercise has been carried out by a large number of workers. As a result, the majority of the simple metal oxides have been examined as catalysts for the partial oxidation of methane. Representative data from these studies are summarized in Tables 1-5 for different classification of metal oxides. A striking feature of these studies is that the product distribution remains invariant to a greater extent (i.e. C2H6, C2H4, CO, C02, H2, HzO and lesser amounts of C3+ hydrocarbons) and CH30H, CH2O or other oxygenates have been reported for only a very small range of catalyst formulations.22-24 It is possible that at the elevated temperature, >600 OC, required to observe 22 K.Otsuka, T. Komatsu. K. Jinno. Y. Uragami, and A. Morikawa, Proc. 9th Inr. Cong. Cliid.,Calgary. ed. M. J. Phillips and M. Ternan, Chemical Institute ofCanada, 1988.2.905. 23 E. Mac Giolla Coda, E. Mulhall. Rivan Hoek. and B. K. Hodnett, in Proceedings 1st Workshop on Catalytic Methane Conversion, Bochurn 1988. Crrial. Toclq,, 1989. 4. 383. 24 S. Kasztelan and J. B. Moffat, J. Ciimi. Soc.. Clirm. Commun., 1987, 1663. CH4i02 GHSV T ethanc uiolc rcrtio h' "C CH4 02 Selectirity 7;) Yield ";, cthcne ~-f 96 800 6.8 ~ 44 3.0 -t 96 800 8.4 -4.8 0.4 -t 96 800 4.4 ~ 43 2.0 ~t 96 800 4.4 -9. I 0.4 t 600 700 3.2 20.9 0.7 1.90 t 600 800 10.1 -29.3 3.0 0.86 t 600 700 1.2 11.8 0.1 1.11 f 600 800 2.7 -21.0 0.6 1.13 -5 29 800 75 I 14.2 38.2 5.4 10 0.1 I* 740 -99 12.8 -2.88 10 0.22* 740 ._ 95.I 56.2 -2.2 -10 0.01 I * 640 6.6 87 23.9 1.6 ~10 0.011 * 640 4.2 63 7.6 0.3 -10 0.01 I * 640 6.4 85 14.7 0.9 -I 00 2* 750 -0.I 8 710 12.9 100 59 7.6 -" Ref.9. C. A. Jones, J. J. Leonard. and J. A. Sofronko. CIS Prrroit. 4444984 (1984). ' C. A. Jones, J. J. Leonard. and J. A. Sofranko. US Putenr, 4443664 (1984). Ref. 17. " Ref. 12. ' W. Bytyn and H. Baerns. .4ppl. Ccrrd., 1986, 28. 199. 1. T. A. Ernesh and Y. Arnenorniya. J. Phj..c.. Chm., 1986. 90.4785. K.-I. Aika and T. Nishiyama. .I. C'hn. Soc,.. C/run. ('om/mm..1988. 70. ' G. W. Kculks and M.Yu, Rrtrc,f.Kinrr. ('lifd. Lrft.. 1987, 35, 361. * denotes WIF g s rnl-'. t denotes cyclic operation, methane and oxidant fed separately z N 0.N Conwrsion 'i, ICH4 02 GHSV -cfhuno I? 101. 1'11I10 h-' "C CH4 02 Selectiritj. Oi, Yield Oo eflietw ~5.0 2 000 710 25.3 94.7 48.4 12.2 -8.3 33.6 740 I 3 19 0.2 -8.3 55.3 740 8 100 5 0.4 -8.3 16.1 740 10 I00 4 0.4 -8.3 13.6 740 9 87 51 0.5 ~8.3 52.4 740 8 63 63 0.5 -4.8 0.22* 508 13 ~ 25 3.3 2.0 0.72* 750 44.2 ~ 49.3 21.8 1.18 2.0 0.72* 750 52.6 ~ 60.1 31.6 0.66 2.0 0.72* 750 25.1 .-2.9 0.7 0.0 2.0 0.72* 750 31.5 -12.7 4.0 0.11 2.0 0.72* 750 25.3 ~ 62.4 15.8 0.37 2.0 0.72* 750 28.3 -44.5 12.6 0.98 6 1500 697 10.5 99.0 8.4 0.9 0.0 6 1500 69 7 9.1 99.2 9.I 0.8 8.0 6 I 500 697 25.9 46.6 46.7 12.1 1.5 263 Mctii(inc I O4 Sprcifrc Actiritj.Imo1 procluc,t m h ' .J;; CH4;02 T f~wl-ru kIil Q t?I oIc ruf io "C mol gcat ' h ' Hl coz co ClH4 CzH6 Tot~lC2 TOILIICt Ref. 3 3 710 0.06 16* 7.7 16.1 1.3 2.5 8.0 10.5 38.4 N 2 73 710 0.06 I6* 3.7 4.5 1.3 0.8 0.8 1.6 9.0 u 4.z 3 710 0.06 I6* 20 24 5.1 1.8 8.9 10.7 50.5 a # 3 710 0.06 16* 11 17 1.8 4.2 4.9 9.1 37.0 u I'* 3 710 0.06 16* 16 97 3.9 29.0 42.0 71.0 242.9 u 3 710 0.06 16* 19 28 3.5 5.0 6.8 11.8 55.1 (I 2.I 700 0.0148 0.02 0.06 0.08 1.4 h 2.1 700 0.0185 3.0 3.0 6.0 16.5 h 2.1 700 0.0 I 85 0.21 0.43 0.64 4.2 h 2. I 700 0.01 85 0 0.02 0.02 3.0 h 2. I 700 0.01 85 0.05 0.54 0.59 7.9 h 2.I 700 0.0 I85 2.8 3.9 6.7 16.7 h 2.1 700 0.0185 4.9 5.5 10.4 30.8 h 2.1 700 0.0 I85 3.5 4.9 8.4 28.7 h am r-b C? --y! r! -.qoc-00330 II formation of Cz hydrocarbons, the oxygenated products, such as CH30H and CHz0, are unstable and decompose possibly to carbon oxides and hydrogen. However, formation of oxygenates at lower temperatures is not reported, and in studies where low temperature investigations have been made 25 no enhanced level of oxygenated products is observed. While the product type remains invariant, product selectivity is markedly affected by the choice of metal oxide as catalyst, but even for the most selective oxides the C2 yield rarely exceeds 25",, an observation consistent with the 100",,rule previously noted by Sofranko et d17 It is also apparent that the ethene'ethane product ratio varies markedly for all these studies. However, as has been recently noted by van Kasteren et (il.," this may be caused by the design of the reactor, particularly the volume of the heated zone downstream of the catalyst bed, and for most published studies summarized in this review precise details of the reactor are not given.Another general feature from these studies is that catalyst activity is related to and the increase in catalyst basicity increases activity. This is most readily observed from the addition of alkali metal additives to single oxides such as MgO and CaO (Table I) which results in catalysts that are significantly more selective for Cz production. The role of alkali promoters will be considered in Section 3A.The main observation that can be made from the summary of the catalytic studies (Tables 1-5) is that a very wide range of experimental conditions have been used. For example, studies have used CH4/02 mole ratios from 2: I to 200: 1 and in some studies additional inert diluent is utilized. In view of the diversity of the experimental work it is extremely difficult to draw any broad conclusions from these studies concerning catalyst selectivity. It is apparent that metal oxides exhibit a range of catalytic activities but to some extent these variations are due solely to the intrinsic surface area, and hence density of surface sites, of the oxide. The importance of surface area with respect to the yield of C7 hydrocarbons has been noted by Aika 29 and it was concluded that low surface area materials were preferable, this being one of the major reasons for the large scale use of alkali metal compound additives since this results in loss of surfice area due to sintering.It is therefore important to compare different catalysts in terms of their specific activity (moles product m 'h-') and these data are summarized for the most active catalysts in Table 6. On this basis it is now possible to compare the relative activities of these oxides. At 700-710 "C catalysts can be ranked in order of decreasing specific activity for C2 production as follows: Li G. .I.Hutchings. M. S.Scurrell. and J. R. Woodhouse. Srrccl. Sirrf. Sc,i. C~(c/o/.. IYXX. 36. 415. '"H.M. N. van Kastcrcn. J. W. M. H. Gccrts. and K. van dcr Wiele. rcl. 22. p. 930.''K.-I. Aiku and T. Nishiyama, in Proceedings 1st Workshop on Catalytic Methane Conversion. Bochum 1988. Cu/cc/.7otltr.v. 19x9. 4. 27 I 2n J. A. S. P. Carreiro. (i. Follmer. LA. I.ehmnnn. and M. Haerns. ref. 22. p. S9l.''E. Iwamatsu. T. Mosiqama. N Takasaki. and K. Aika. S/irt/. .Srcr/. Si.C'cr/cr/.. I9XX. 36. 373. Hutc'liings, Scurwll md Woodiouse At the higher temperature of 750"C a recent study by Machida and Enyo 30 indi-cates that SrCe03 may exhibit a higher specific activity than LiiSrnzO3. For this summary it may be concluded that, although catalyst selectivities may be similar, there exists a very wide range of specific activities spanning almost two orders of magnitude.Considerable scope exists to further improve on these specific activities, and as yet no detailed study has been published concerning optimiza- tion of the specific activity by using improved methods of catalyst preparation. B. Nitrous Oxide as Oxidant.--A number of studies have specifically utiliLed nitrous oxide as oxidant because of the specific selectivity effect that can be .~achieved. Liu et ~1 and Zhen ~ et d3'noted that with Mo and V catalysts methane was oxidized to methanol with NIO as oxidant. Subsequently, Solymosi ef ul.33demonstrated that for Bi2O3 as catalyst use of NzO resulted in the forma- tion of high selectivities for formaldehyde at 550 "C, a temperature at which O2is inactive. Addition of SnOz significantly enhanced this effect and it was considered that one role of the SnOz was to facilitate the decomposition of NzO enabling it to participate directly in the oxidation process.However, a number of other studies for Sm203,34.35 Mg0,36 and Li'Mg036 have shown that use of NzO does not significantly enhance the production of oxygenates for these catalysts, and hence this is not a general feature of this oxidant. Indeed, a recent study by Kasztelan and Moffatz4 demonstrated that when 01 was used as oxidant with silica as catalyst significant yields of formaldehyde were observed. in contrast to the results with NzO where complete oxidation was observed. Few comparative studies of the use of N20 and O2 as oxidants have been reported, and these are summarized in Table 7.The main conclusion from these studies is that under comparable reaction conditions, with NzO as oxidant, significantly lower methane conversions are observed with increased Cz selec-tivity, particularly to ethane. The increase in selectivity to ethane is most marked at lower reaction temperatures. The marked increase in Cz selectivity may just be a result of the selectivityiconversion relationships that exist for these catalysts, i.c.. higher Cz selectivities are generally observed at lower methane conversions (eg. see Figure 3). However, recent evidence" has shown this not to be the case, and at comparable conversions NzO is always more selective for Cz formation than 02 as oxidant (Figure 4). The decreased oxidation activity observed with N20 is probably due to two factors. First, on a molar basis NzO can be considered to have only half the potential for stoicheiometric reaction compared to 02,but experiments where this has been taken into account have shown that NzO is much less active than 01.34Second, the decomposition of NzO to an active "' K.I. Machida and M. EnLo .I C/ie,/?i. Soc...C/ic/ji, ('OIII/IIU/I.. 1987. 1639. I' H. F. Liu. R. S.Liu. K. Y. Liew. R. W. Johnson, and J. H. tunsford, J. .-1/?i C'hcvti. So(.., IY84.106.41 17. " K. J. Zhen. M. M. Khan. C. fl. Mak. K. R. Leu& and G. A. Sornorjai. J. Ccr/u/..1985. 94. 501.''F. Solyrnosi. I.Tornbac/. and G. Katson, ./. Chrtii. Sor.. C'/icw. Co/~i~~iir~i..1985. 1455. K. Otsuka and K. Nakajirna../.('hcwi. Sot.. Ftr,oc/crI. Trtr/i\.. 19x7. 83. I3 15. .''G. J. Hutchings. M. S. Scurrcll. and J. R. Woodhouse. in Proceedings 1st Workshop on Catalytic Methane Conversion, Bochurn 1988. Co/(t/.Toc/c/j..19x9, 4. 371. .'"G.J. Hutchinps. M. S. Scurrell. and J. R. Woodhouse. J. C/iwi. Socc. C/i~wi('o/1i/1iir/i.. 1987. 1388. -.Conrersion nzole Product Selectiviry mole 7; .oCH4/osidunl T GHSV Crltcll~~.st Oriricint mole rcitio "C h ' CH4 Ovidunt C02 CO CzH4 CzH6 ToralC2 Ref 59 MgO 02 3.5 550 667 3.0 -66.2 31.8 0.4 1.6 2.0 h,c ~N2O 3.3 550 563 1.8 96.1 2.0 0.3 1.6 I .9 h,c W .a02 3.5 710 667 5.2 100 50.0 3.2 28.4 18.4 46.8 h,c \ N2O 3.3 710 545 3.0 98 88.9 6.8 2.7 1.6 4.3 h,c % %Li: MgO 02 3.3 550 648 2.0 -59.8 17.7 5.0 17.5 22.5 b,c Q N2O 3.4 550 638 0.2 -17.2 0 7.0 75.8 82.8 h,c % 02 3.3 710 648 2.7 37 33.0 1.6 26.5 38.9 65.4 h,c 5 NzO 3.4 710 638 0.3 2.8 51.9 0 9.3 38.9 48.3 h,c s.Sm203 02 3.I 710 592 19.1 100 82.4 2.3 1.0 14.3 15.3 c P NlO 3.0 710 720 14.6 99 64.4 3.7 14.6 17.3 31.8 c g 5",, LitSmz03 02 3.0 710 857 24.5 87 45.8 1.3 27.0 25.0 52.9 2c 4 41N2O 2.4 710 756 3.0 13 29.1 0 11.5 60.3 71.8 C 12 e1 La203 02 4.6 550 3 660 2 25 --0 0 0 d e N2O 4.6 550 3660 tr. 4 -~ 0 57 57 d 02 4.6 700 3 660 10 75 --2 6 8 d N2O 4.6 700 3 660 4 24 --15 70 85 d Srn203 02 4.6 550 3660 11 93 --1 10 11 d NzO 4.6 550 3660 0.5 5 --3 82 85 d 02 4.6 700 3 660 15 98 --10 21 31 d N2O 4.6 700 3 660 5 35 --10 68 78 d " Selectivities calculated on carbon basis. Ref.36. c Ref. 3s. ,I K. Otsuka and T. Nakajima, Jnorg. C'him. Ac~ri,1986, 120. L27; K. Otsuka, J. Jq.Per. Jnst., 1987, 30. 385 Hutchings, Scurrell, and Woodhouse 100 90 80 0 be \ c 2,-70 0w-1 Lf v) 60 0 0 0 50 0 40 0 1 2 3 4 5 CONVERSION/% Figure 4 Comparison qf C2 selectiuity oi'er Li/MgO, 710 OC, CH4/osidunt = 3: 002 us 0-uidant,0NzO as oxidanr, data taken from reJ20 oxidizing species may be slow, and experiments by the present authors have indicated that only partial N20 decomposition occurs under most conditions. N20 has been utilized in these studies as a source of surface 0-as an oxidizing species3' and some evidence for this is further obtained from the observation (Table 7) that with N20, CO is oxidized to C02 which is produced in significantly higher quantities than with 02.Based on a detailed analysis of the reaction selectivities at 550 "C and 710 "C it has been shown 36 that 0 has two ~ oxidizing roles on these catalyst systems, a selective oxidation role which is particularly marked at the lower reaction temperature, and a non-selective role which becomes significant at the higher temperature.The mechanistic significance of the comparison between 02 and N2O as oxidants will be discussed in more detail in Section 3. C. Ozone as Oxidant.-Earlier studies reviewed by Foster have indicated that ozone could, by virtue of its enhanced oxidizing activity, be active for methane oxidation at temperatures lower than those observed with molecular oxygen (i.e.<500 "C). A comparative study of ozone versus oxygen as oxidant 25 has shown "C. Naccache, Chem. Phix Lett., 1971. 11. 323. 0.vidcrtiw Coupling of'Mc. tlimw rrsitig Ox itkc Ctrtci1y.ct.c PRODUCTS 1 103mol.h ' 2.0 m P 1.5 1.0 0.5 0.0 0 100 200 300 400 500 600 700 that use of ozone does not enhance the production of either Cz hydrocarbon or oxygenates at lower temperatures (Figure 5). Ozone was observed to be more reactive than molecular oxygen at <500 "Cbut the only products observed were hydrogen and carbon oxides. D. Methane Oxidation in the Presence of Carbon Dioxide.-Aika and Nishiyama2' have shown that addition of COz to CH4/02reactant mixture can significantly enhance the yield of C2hydrocarbons for a broad range of catalysts, and in particular with PbO-MgO catalysts. It was concluded that the carbon dioxide was involved in these reactions as an oxidant, providing a monatomic oxygen species viu degradation to carbon monoxide.However, no isotopic studies have been reported to substantiate the involvement of oxygen from carbon dioxide. It is more probable that the role of carbon dioxide is to enhance or maintain the formation of surface carbonate species which produces a more selective catalyst for methane activation. It is clear that further surface and isotopic studies are required to rationalize the experimental observations. 3 Studies of the Reaction Mechanism A. The Role of the Alkali Promoter and Nature of the Active Site.-A number of studies have been made concerning the elucidation of the nature of the active catalytic sites particularly for the non-reducible oxide catalysts typified by MgO and Smz03, and particular attention has been given to establishing the role of the alkali promoter.Addition of Li+ to MgO has been shown38.39 to decrease the surface area, increase specific activity, both for methane oxidation and the selective production of C2 hydrocarbons (Table 6), and significantly increase CZ selectivity, particularly for C2H6. Even the addition of very low levels of Li+ have been shown to exhibit these effects eg. 0.4",, by mass.18 Lunsford and co-workers 13.38.40 44 have extensively studied the Li'MgO catalyst using e.p.r.spectroscopy and have shown that in the presence of 02 this catalyst forms 0-as a surface species, which is stabilized by the formation of [Li'O-] sites. Previous studies of Abraham and co-~orkers"~ 49 have demon- strated that [Li'O-] centres can be found in Li+-doped MgO or CaO, either by thermal treatment or by high intensity electron bombardment. Lunsford and co-workers 43 have further shown that gas phase methyl radicals were found when CH4/02 were reacted over Li/MgO at 500°C. Direct correlations were observed between the concentration of CH; and the concentra- tion of [Li+O-] centres, as well as between the degree of methane conversion and the concentration of the [Li'O-] centres. It was therefore concluded that the CH; radical was produced from the interaction of CH4 with the 0-of the [Li+O-] centre.The active [Li'O -1 was subsequently regenerated from the resultant LiOH by reaction of LiOH with molecular oxygen. For MgO, doping with Na' was not found to be as effective as Li+. This has been explained in terms of the ionic radii of Li+, Na', and ME'+. The [Li'O-] centres are formed by substitution of Li' for Mg2+ in the MgO lattice and these ions have similar radii (Li' 68 pm, Mg2+ 66 pm). The larger Na' ion can not be so readily accommodated in the MgO lattice, and hence is less effective at generating the equivalent [Na+O -1 centre. However, Na+ and Ca2+ also have very similar ionic radii (Na' 99 pm, Ca2+ 97 pm) and it is important to note that for CaO the most effective dopant is Na+ ." T.Ito, J.-X. Wang. C.-H. Lin. and J. H. 1,unsfot-d. J. Aii7. C/wi7. SOL..,1985. 107. 5062. jYG. J. Hutchings, M. S. Scurrell, and J. R. Woodhouse. J. Chon. So(,.,Chcn7. Cor,7n7un., 1987, 1862. "' D. J. Driscoll and J. H. Lunsford. J. PI7j.s. C/ZP~H.,1983,87,301. " D J. Driscoll. W. Martir, J.-X. Wang, and J. H. Lunsford, in 'Adsorption and Catalysis on Oxide Surfaces'. ed. M. Che and G. C. Bond, Elsevier, Amsterdam. 1986.403. D. J. Driscoll and J. H. Lunsford, J. Phj..~..C'i~wi.,1985,89, 4415. "D. J. Driscoll. W. Martir. J.-:Y '.i'LLiig.and J. H. Lunsford, J. .4,17. Chcw7. Sot,.. 1985, 107. 58. I4 J. H. Lunsford. C.-H. LkIi. .I.-X. Wang. and K. D. Campbell, in 'Microstructure and Properties of C~~iiilysts'. Materialsed.M. M. J. Treacy. J. M. Thomas, and J. M. White, iM~it.R6.s. Soc. Sjwp. PIYIL~., i Soc~ely,Pittsburgh, 1988, 3. 305. I5M. M. Abraham. Y. Chen. L. A. Boatner, and R. W. Reynolds, Plij,s. RPI..Le/i., 1986.37, 849. Ih Y. Chen. H. T. Tohver, J. Narayan, and M. M. Abraham, Phj*.s. Rer.. B, 1977, 16. 5535. 47 J. B. Lacy. M. M. Abraham. J. L. Boldu, Y. Chen. J. Narayan. and H. T. Tohver. P/7j..s. Rrr.. B, 1978. 18.4236. "J. 1.. Holdu. M. M. Abraham. and Y. Chen. PhI..s. Her.. S, 1979. 19, 4421. "J. L. Olson. V. M.Orera. Y. Chen. and M. M. Abraham. P/7I..s.Rrr. B, 1980, 20. 1258. 27 1 Oxidutiw Couplirig of Metliunr usirzg Oxide Cutuly.rts (Table 6), although both Lit and Kf also significantly enhance the specific activity of CaO.Confirmation of the role of 0-in the selective oxidation of methane has come from photoluminescence studies 50 which have shown a direct correlation between the concentration of 0-on the surface of MgO and the conversion of methane. In addition, model studies using X-ray photoelectron spectroscopy on oxidized Mg for the analogous activation of N-H bonds in ammonia have also demonstrated the importance of the 0-species." Use of N20 as oxidant,36 a known source of surface 0-,37has also confirmed the role of 0-in the selective oxidation of methane. Hence, there appears to be a general consensus that 0-on the surface of MgO and CaO, stabilized by the addition of Li' on Na', is the species responsible for the initial activation of methane cia hydrogen atom abstraction to form a methyl radical.It must be noted that the presence of alkali cations is not crucial in this respect since undoped MgO and CaO are also active for methane partial oxidation. The nature of the active centre on lanthanide oxide catalysts has been extensively studied. An e.p.r. study by Lunsford and co-workers 52 demonstrated that the rate of methyl radical formation was higher on La203 than on Li/MgO under comparable conditions. They found that superoxide, 02, was the most abundant species and they considered that the broad e.p.r. spectrum of 0; might have masked the presence of 0-. It is interesting to note that doping La203 with Li' does not significantly enhance the specific activity (Table 6) which is in direct contrast to the effect of Li' on MgO and Sm203, and hence this could indicate that 0-may not be the active surface species for La203 oxide.However, it is clear that the methane activation on La203 is analogous to that on MgO and Li/MgO, i.e. ciu the generation of methyl radicals. 34353Otsuka and co-workers have carried out a detailed kinetic study of Sm2O3 and have concluded that the active species for methane activation on this oxide is a diatomic oxygen species. Otsuka et uI.~~subsequently demonstrated that solid peroxides Na202, Ba02, Sr02 were active for the formation of ethane from methane in the absence of molecular oxygen. Based on these studies Otsuka has suggested that 0;-could be the active species on this oxide. More recently, Lee and Oyama 55 have noted that O', -, which is difficult to characterize on an oxide surface, can be regarded as a dimer of 0-,i.~.O--O-, and the formation of 0-from 0;-has been previously considered in the extensive reviews of Che and Ten~h.~~ ~Further evidence that 0 is the active species on Smz03 is provided by the observation that addition of Li' gives a significant increase in specific activity (Table 6) which is directly analogous to the effect of Li' dopijig on MgO.In addition, use of N20 as oxidant in place of molecular oxygen (Table '37) significantly increases the formation of C2Hh decreased methane 5 0 M. Anpo, M. Sunamoto, T. Doi. and I. Matsuura, C/icwi.Lf,i/.. 1988, 701. '' C. T. Au and M. W. Roberts, J. C/icwi.Soc..Frrrtrt/tr~~Trertis. I. 1987, 83. 2047.''C.-H. Lin, K. D. Kampbell, J.-X. Wang, and J. H. Lunsford, J. P/IJ..s.C/iwi.. 1986. 90. 534. 53 K. Otsuka and K. Jinno, Inorg. Cliiin. Ai./u, 1986. 121, 237. s4 K. Otsuka, A. A. Said. K. Jinno. and T. Komatsu, C/iet?i.Lcit., 1987, 77. ''J. S. Lee and S. T. Oyama, Cli[d.Rev. .Si,i.Etig., 1988. 30. 249.'' M. Che and A. J. Tench. Ad,. C'trtcil., 1982, 32, 77; A&. Clttd.,1983, 32. 1. conversion. These effects of NZ0 are similar to those observed with the Li MgO catalyst, and hence by analogy it can be concluded that 0-is also the active oxidizing species on Smz03 and Li, Smz03. Methane activation over reducible oxides, c.g. MnOz, PbO, has not been studied in any particular detail. Keller and Bhasin9 noted the active oxides for stoicheiometric methane activation had two accessible positive oxidation states.M"' and M'" 2, ', and they postulated the involvement of lattice oxide ions 0'-in the generation of activated surface methyl species. Subsequently Jones et it/.'' considered that the product distribution of Cz-C6 hydrocarbon products was consistent with the involvement of gas phase methyl radicals in carbon-carbon bond formation. Driscoll and Lunsford 42 did not observe methyl radical formation in the CHs/02 reaction over PbO at 475 "C whereas methyl radicals were observed over both MgO and Li MgO at this temperature. It is clear that this aspect of methane oxidation requires further detailed study before any definitive conclusions can be drawn.B. Primary Hydrocarbon Product Formation.-(i) /dc~r.rzt$cutiotz of' Priincir!? Hjdroc.uvhoti Products. Until recently only limited attention had been given in the research literature to consideration of the nature of the primary hydrocarbon products. This is surprising since without this detailed knowledge of the reaction mechanism it is unlikely that future improvements in catalyst design will be effected in a scientific manner. In part this is the result of early studies,".'* which concluded that ethane was the only primary hydrocarbon product and that the direct formation of ethene from methane was negligible. Subsequent studies have therefore worked on this basis, which is unfortunate since ethene is considered to be the more desired product on an economic basis and the identification and enhancement of a direct formation pathway would be a significant advance in catalyst design.A recent detailed study by the present a~thors~~.~'.~~ has shown that direct formation of ethene from methane is not negligible for all catalysts, and must therefore be considered in a mechanistic scheme. This investigation involved the standard procedure for determining primary selectivity by varying reactant feed-rate at constant reaction temperature and methanejoxidant ratio, both molecular oxygen and nitrous oxide were used as oxidants, and repre- sentative data are shown for MgO and LiIMgO in Figure 6. From this study it was concluded that for MgO, SmrO3, and La203 with both 0' and NzO, and for LiiSmzO3 with 02, both ethene and ethane are the primary hydrocarbon products.For LiiMgO with 02 and NlO (Figure 6). Li;La203 with 02,35 and Li/Sm203 with N203' as oxidant the addition of Li' had a marked effect enhancing the formation of ethane relative to the unpromoted system. From these catalysts it was possible to conclude that ethane was a primary product, but the situation was less clear for ethene, since although linear extrapolation of the plots (e.g. Figure 6b) could have indicated a positive intercept for both ethane '-C. A. Jones. J. J. Lconard. and J. A. Sofranko. Eiirrgj. (md FM~.1987. 1. 17. jXK. Otsuka. K. Jinno, and A. Morikawa. J. C'circd.. 1986. 100. 353 0 Id 15 14 0-4 I---3t O 1-.1 I'I 4Oo i i 5 6; 8 9 ibi'1h CH, CONVERSION % Figure 7 Ethrtw sekectiritj.IIS u futictioti of CH4 conrorsion; Li:MgO; 710 "C; 0 CH4/02 = 3 mole rcttio; 0CH4!02 = 5; CH4/02 = 10 and ethene, it was possible that the shortest residence times used in these initial experiments may still have been too long to conclude that ethene was a minor primary product with these catalysts/oxidants. Further detailed studies 2033 were then carried out with Li/MgO using a small catalyst bed volume (0.5 ml) with a minimized post reactor heated zone to limit the oxidative dehydrogenation of ethane downstream of the catalyst.26 It was observed that for CH4/02 = 3 a significant selectivity to ethene of cci. 4"/,, was obtained on extrapolation to zero methane conversion (Figure 7), indicating that ethene selectivity and methane conversion is not linear in the conversion range of &lo%, and in particular that extrapolation of data obtained for 25%) conversion could lead to a misleading conclusion concerning ethene as a primary product.This may have been the case in the earlier studies concerning primary product It was also shown 2o that decreasing the oxygen concentration did not significantly affect the ethene selectivity at low conversion while it did enhance the ethane selectivity, which indicated that the ethene primary selectivity was reasonably independent of the CH4,02 ratio. It was observed that this finding strongly supported the conclusion that ethene was a primary reaction product, since if ethene were to be formed solely from oxidative dehydrogenation of ethane, as postulated in earlier studies,' .5 5.s then variation in [02]would have been expected to affect the selectivity to ethene, and this was not observed.Based on this detailed study it was concluded that a small but significant primary selectivity to ethene could be observed for these catalysts. Hence for MgO, Sm203, and La203, with and without Li' and for both 02 and N2O as oxidants, a reaction pathway exists for the direct formation of ethene from methane and independent of C2Hh. (11) Mechanism of' Prinicrrjy Hjdroccirhon Product Forniritioti. Based on the literature review in Section 3B(i) it is clear that both ethane and ethene must be considered as the primary hydrocarbon products, although for most catalysts ethane is the dominant primary product.There is general36.43.44.50.51 .59 that surface 0~,derived from NzO decomposition or stabiliLed by Li + doping, is the selective oxidizing species for ethane formation. Interaction of 0-with methane, either gas phase or surface associated (since recent studies 60-61 have indicated that significant amounts of methane are associated with some oxide surfaces even at 700 "C),leads to the formation of gas phase methyl radical vicc hydrogen atom abstraction. Cant or ri1.62 have shown the existence of a kinetic isotope effect indicating that this hydrogen atom abstraction is the rate determining step in the formation of ethane. The radical nature of this reaction has been further confirmed in experiments using nitric oxide as a radical ~cavenger.'~ It is generally agreed that dimerization of methyl radicals leads to the formation of the ethane primary product, and this process occurs in the gas phase.It has been noted20.64 that the presence of Li' as dopant, which aids the stabilization of 0-, has significantly different effects on ethene and ethane selectivities, and from this observation it has been concluded that the surface 0-species cannot be the oxidizing species responsible for the formation of the ethene primary product. In addition it is probable that the primary formation of ethene must involve a surface-catalysed process since the diverse range of oxide catalysts give a wide range of primary ethene selectivitie~.~~~~~~~~ Recent studies by Martin and Mirodatos "have indicated that methylene carbene may be involved as a reactive intermediate in methane oxidation.They have postulated that the formation of the primary ethane product results from insertion of methylene carbene into a carbon-hydrogen bond of methane. Subsequent studies 66 using isotopic labelling have shown that this reaction does not occur. Oxidation of CD4 and CH4 mixtures resulted in the formation of ethane with isotopic distribu- tions of CH3-CH3, CD3-CD3, and CH3-CD3; no ethane product containing CD2 or CH2 units were observed indicating that the methylene carbene species were not involved in the production of ethane. However, this species could be '')S.Mehandru, A. B. Anderson, and J. F. Bradzil, J. Am. Chrni. Soc.. 19x8. 110, 1715. 6o A. Ekstrom and J. A. Lapszewicz, J. Clirm. Soc., C'Iiiwi. C'oniniuri.. 1988. 797. 61 A. Ekstrom and J. A. Lapszewicz. Prcyr.-Ani. C%riii.Soc., Dir. Pel. C/ioni.,1988. 33, 430. 62 N. W. Cant. C. A. Lukey, P. F. Nelson. and R. J. Tyler, J. Chrr?i.Sot,., Chrm. C'onmun.. 1988. 766. 63 G. J. Hutchings. M. S. Scurrell. and J. R. Woodhouse, J. C'/iot?i. SOC.,Ch~wi.Cwiiuim, 1989, 765. 64 G. J. Hutchings, M. S. Scurrell and J. R. Woodhouse. Proc. 9th /I!/. C'ongr. C(t/d..Calgary, ed. M. J. Phillips and M. Ternan. Chemical Institute of Canada, 1988.2.923.''G. A. Martin and C. Mirodatos, J. Chwii. Soc,.,Clicwz. Coniniwi., 19x7, 1393; C. Mirodatos and G.A. Martin. Pro(,.9l/i /ti/. C.otigr. C'CII~.,Calgary, ed. M. J. Phillips and M. Ternan, Chemical Institute of Canada, 1988,2.899. 66 P. F. Nelson. C. A. Lukey and N. W. Cant. J. P/IJXCT/ieni.\ 19XX. 92, 6176. Hutchings, Scur.rt.11,and Woodhouse involved in the primary formation of ethene uia the following reaction pathways: fH2 -C2H4 (1) It has been concluded,20 on the basis of use of N2O as oxidant or Li+ addition to MgO, that methylene carbene could not be formed in significant amounts from an interaction of CH\(,, with surface 0-, i.c. a second hydrogen abstraction. Hence it has been concluded that the oxidizing species responsible for methylene carbene formation, and consequently the primary ethene product, must be able to abstract two hydrogens from methane in a single interaction of methane with the surface.Detailed reviews by Che and Tench 56 have shown that oxide surfaces such as MgO are capable of stabilizing a range of oxygen species including superoxide Or and peroxide 0s-in addition to 0~.From analysis of the experimental data 2o it was concluded that a dioxygen species, probably 0: -, is the most likely oxygen species responsible for the formation of the methylene carbene species. It is interesting to note that 0;-has been proposed by Otsukas3.s4 as an active oxidizing species for Sm203, a catalyst which gives significantly higher ethene selectivities than other oxides (Table 6). C. Secondary Conversion and Non-selective Oxidation Reactions.-(i) Secondur.j* Cotiiwsioiz of Ethinc to Etlimc~.Ethene can be formed from the primary ethane product tiin two reaction pathways: For MgO, La2O3.and Sm203 it has been noted35.64 that the selectivities of ethene and ethane are comparable at most reaction conditions. Hence the ethene,ethane ratio observed experimentally is close to that expected for thermodynamic equilibrium according to equation 3.67 Hydrogen is a major reaction product for methane oxidation over undoped oxides which is consistent with this reaction. Subsequent studies 2o demonstrated that this non-oxidative pathway occurred only to a limited extent in the presence of an oxidant, and hence reaction 4 is the major pathway for secondary ethene formation under conditions of non-total oxidant conversion.Only recently 20.68.6y have investigations been addressed at the relative rates of ''-R. Stull. E. I;. Westrum. Jr.. and G. C. Sinke. 'The Chemical Thermodynamics of Organic Compounds'. John Wile). Nev. York. 1969. '" J. A. Rooa. S. J. Korf. R. J. H. Veehof. J. G. van Omrnen. and J. R. H. Ross, in Proceedings 1st Workshop on Catalytic Methane Conversion. Bochurn 1988. Curd. TohI.. 1989. 4. 441. '"'J. W. M. H. Geerts. J. N. M. van Kasteren. and K. van der Wiele. in Proceedings Iht Workshop on Catalytic Methane Concersion. Bochum 1 YXX. Cu/d ToduI.. 19x9. 4. 453. oxidation of methane, ethane, and ethene. In general, as expected, ethane and ethene are oxidized far more rapidly than methane, a feature which has been neglected in some kinetic modelling studies.28 Comparison of N2O with O2as an oxidant for ethane has indicated2" that N2O is always more selective for ethene formation than 02,and hence a monatomic oxygen species is the selective oxidizing species for this reaction.These studies 20.68.h9 have all provided evidence that reaction 4 is a gas phase reaction and the presence of a catalyst only leads to enhance the parallel total oxidation pathway. However, these studies also indicate that reaction 4 is facile under the conditions required for methane partial oxidation and at high conversions, >5°/o, it must be considered to be the major route for ethene production. (ii) Secondurj. Totul O.\-icirrtion of Ethmc~crnd E[heric. The mechanism of total oxidation of the C2 primary hydrocarbon products has not been well studied.Lunsford and co-workers'" have proposed that ethane is oxidized to carbon oxides viti ethene and a direct pathway from ethane is not significant. However. no firm experimental evidence was provided in support of this hypothesis, and other studies have indicated that direct oxidation of ethane to carbon oxides occurs for Li'Mg04' and Sm203.58 From a detailed comparison of ethane oxidation by either 02 or NzO as oxidant 2o it has been concluded that the non- selective oxidation of ethane involves mainly a surface-catalysed reaction in- volving a dioxygen species, which was not identified and was designated as Ol-. A similar study2' for ethene total oxidation also indicated the involvement of a dioxygen species, but in this case the reaction involved both surface catalysed as well as gas phase reactions.(iii) Non-st.lectiiv> Metliernr 0\-idtition trnd the Formition of Iljdrogm. Studies 36 have indicated that the surface 0 species has two distinct roles: (a) it is the selective oxidizing species responsible for the formation of methyl radicals, the precursor to ethane, and (b) it participates in non-selective methane oxidation- recent studies2' have indicated this is i'iii a parallel pathway to the selective methane activation route. Both of these pathways involve the surface species 0-. Studies have also shown2'.lS that N20 and 03/02can react in the gas phase and in this case it is considered that a monatomic oxygen species is the reactive gas phase species.Recent studiesh3 involving reaction in the presence of the known radical scavenger nitric oxide have indicated that a second oxidation species or site, separate from the surface species 0-,is also important in the total oxidation of methane. However this oxidation species,'site has not yet been iden- tified. It is apparent that under most reaction conditions hydrogen is a major reaction prod~ct.'~.'' It is particularly marked for MgO as a catalyst (Table 6). but the selectivity to Hr relative to that of C2 hydrocarbons decreases on doping with Li+. It can be considered that the increased surfdce concentration of 0--11 IH.-S. Zhan. J.-X. Wang. D. J. Driscoll. and J. 11. Lunsford. J. CU/C//..IYXX. 112.366.-'J. A. Roos, A. G. Bahcr. H. Bosch, J. G. can Ommcn. and J. R. H. Ross. ('o/d.lothi.. 1987, I. 133 278 Hutchings, Scurrell, and Woodhouse from doping with Li' results in oxidation of hydrogen to water. Roos et al.?l considered that the hydrogen could result from a number of different pathways including: dehydrogenation of ethane, the water gas shift reaction, the decomposi- tion of formaldehyde, and steam reforming reactions. In addition, hydrogen is well known as a reaction product from previous gas phase methane oxidation studies,' and Kimble and Kolts 72 account for hydrogen formation uia decomposi-tion of ethyl radicals to ethylene and hydrogen atoms, i.e. reaction 2. At the current level of definition it is possible to discount the significant involvement of ethane dehydrogenation, reaction 3, when unconverted oxidant is present.This has been shown by direct studies with ethane,20 and in addition it is apparent that considerably more hydrogen than ethene is produced in the overall reaction (Table 6). It has also been observed" that the ratio of hydrogen and carbon oxides remains constant at variable contact time for a specific catalyst at constant temperature. The ratio H,/CO, was found to be dependent on the volume of the catalyst bed. It was concluded that a major pathway for hydrogen formation was from a reaction which is linked directly to the formation of carbon oxides. It was considered that a possible route was from the gas phase decomposition of oxygenated products, r.g.CH30H, CH20, formed from the reaction of CH; or CH2 with O2 in the gas phase, cia known reactions.' However, it is clear that a definitive study is still required to fully unravel the origin of hydrogen in this high temperature oxidation reaction. D. Importance of Gas Phase Reactions.-A number of the individual reaction pathways considered in the preceding discussion are cited as involving gas phase reactions, rather than being wholly surface-catalysed. In particular, it appears that the prime function of the catalyst surface is to activate methane uia interaction with a surface 0 species, which generates methyl radicals. These ~ radicals are released into the gas phase and subsequently dimerize to form ethane or could react with oxygen to form the methyl peroxy radical which is a precursor for the formation of carbon oxides.Previous studies by Boomer et ~il." have demonstrated that at high reaction pressures methane can be activated uiu a radical pathway in the absence of a catalyst acting as a radical initiator. It was therefore clear that the effect of increased reaction pressure in the catalysed methane oxidative coupling reaction should be evaluated. An initial study by the present authors 73 demonstrated that increasing reaction pressure from 85 kPa to 585 kPa negated the requirement for a catalyst, since at this pressure significant gas phase radicals can be thermally generated. It was noted that the presence of a catalyst could be deleterious since it tended to catalyse the non-selective pathways and hence decrease the yield of the desired products. Subsequent -'J.B. Kimble and J. H. Kolts. ('hruitcdi. 198X. 501. -'G. J. Hutchings. M. S.Scurrell. and J. R. Woodhouse. J. Chrni.So(...C'hcrii. Commrz.. 1988. 253. G. S.Lane and E. E. Wolf, J. Ct//d..19x8. 113. 144. -.' D. Y. C. Yaws and N.E. Zlotln. J. Ctrtd.. 19XX. 11I, 3 17. -"O.-T. Onsager, R. Lodeng, P. Soraker, A. Anundskaas. and B. Helleborg, in Proceedings 1st Workshop on Catalytic Methane Conversion, Bochum 198X. <'t//td.Todtri.. 19x9. 4. 355. studies 7L76 have confirmed these initial findings, and have demonstrated that gas phase reactions must be taken into account in the oxidative coupling of methane.77 This is of particular importance with respect to the design of improved catalysts.Since if the overall activation of methane should be via the oxidative coupling reaction to ethane, with consecutive oxidative dehydrogena- tion to ethene, then based on current evidence it is probable that use of enhanced reaction pressure may yield the best results. It is interesting to note that the enhancement in the selectivity to Cz hydrocarbons observed on addition of LiCl to NiO, in comparison to the effect of addition of Li2C03 or LiN03 (Table 31, may also be due to the involvement of gas phase reactions. It is possible that LiCl acts as a source of gas phase chlorine atoms that can initiate methane activation ilicr hydrogen atom abstraction to form methyl radicals.The observation that catalyst lifetime is proportional to [LiCl] further confirms the direct involvement of the chloride in a non-catalytic mode. It is clear that further experimental work is required to elucidate whether chloride additives to catalysts (Tables 3 and 5) do have a true surface catalytic function. Recent studies by Burch ct have started to address this area and preliminary results indicate that surface reactions are involved in the production of ethene with chloride modified catalysts. E. Overall Reaction Mechanism.--Most studies, summarized recently by Lee and Oyama” have concluded that a simplified reaction network can be utilized to describe the oxidative coupling of methane (Figure 8) and this has been used in various forms in a number of kinetic studies.’ ‘*28.58 However, the present review of the current literature indicates that this level of simplification omits essential mechanistic details.Lunsford ” has considered the reaction in more detail (Figure 9) and has postulated that reaction of a gas phase methyl radical with a surface oxide anion can lead to the formation of a surface methoxy intermediate which is involved in non-selective oxidation. However, the mechanistic proposal does not account for the direct oxidation of ethane to carbon oxides, nor does it consider the direct formation of ethene from methane. It is apparent that for the range of non-reducible oxide catalysts investigated to date (Tables 1 --5) the -_K. Seshm. Appl. C’(iid. 19XX. 44. 275.-x R. Burch. E. M. Crubb, <;. D. Squire, and S. C. Tsang. C’o/trl. Lo//.. 19x9. 2. 249. 280 Hutchings, Scurrell, and Woodhouse Figure 9 Reuction niechanisni .for the osiclutice coupling of'metlianc. (Reproduced by permission from J. CutuI., 1988,111,302) 0 00 CH3 Og 0-(4 products observed in all cases are broadly similar and comprise C2H4, CzH6, COZ, CO, Hz, higher hydrocarbons together with H20 and traces of oxygenates. Against the background of the reviewed literature and on the basis of a detailed study 20 the present authors have proposed a more detailed reaction mechanism of the oxidative coupling of methane (Figure 10). It is apparent that 28 1 Oxidative Coupling of' Motliunr using O.ridc C~rto/~~st.s further study is required to fully characterize all the active oxygen species, but it is possible that the knowledge gained could be used as the basis for an approach to the design of an improved catalyst.For the oxides investigated to date the direct formation of ethene from methane occurs only as a minor pathway; it is clear, however, that Sm203 and Li/Sm203 demonstrate the highest primary selectivity to ethene.3s In addition Li/Sm203 has been identified as having a specific activity for ethene formation that is over an order of magnitude higher than Li/MgO (Table 6). However, a detailed study of the surface characterization of Smz03 and Li/Sms03 has yet to be carried out. Such a study used in combination with similar studies for LiiMgO can be expected to be of value in the preparation of Sm103-based catalysts that exhibit improved ethene selectivity and activity.4 Concluding Remarks The aim of this review has been to highlight the most important aspects of the recent extensive literature concerning the oxidative coupling of methane to higher hydrocarbon products. In this respect a general mechanism for the methane oxidation over non-reducible catalysts has been described and discussed. A major feature of these oxides is that they only became active for carbon- carbon bond formation reactions at high temperatures (>700 "C)and at such temperatures it is clear that the gas phase nature of the overall reaction is dominant. Indeed, the use of increased reaction pressure can be more significant in effect than the use of the oxide catalysts described to date.However, it is also apparent that alkali-doped oxide catalysts exhibit a wide range of catalytic activities-a feature that has yet to be optimized. This, therefore, represents a major aspect identified in this review that is worthy of further detailed study. It is anticipated that an investigation of the solid state and surface chemistry of these doped oxides, together with optimization of the surface area, should provide the basis for the identification of improved activity catalysts containing an enhanced concentration of active surface sites. While the current studies indicate that use of increased pressure may be more effective, the use of a suitable catalyst will be of paramount importance in the control of the reaction selectivity.It is further suggested that significant efforts should now be given to the identification of oxide catalysts that are active for selective methane oxidation at lower tempera- tures (<500°C) when surface reactions, in preference to gas phase radical chemistry, can be used to control selectivity to the desired product. A feature which is emphasized in this review is that specific oxygen species are responsible for both selective and nonselective oxidation. In particular, studies have indicated the existence of an oxidizing species or site that is involved specifically with total oxidation, but at the present it has not been identified. It can be expected that identification and control of this species or surface site could be of benefit in the preparation of improved catalysts.The most striking feature of the wide range of catalyst evaluation studies reviewed in this article is that few general conclusions can be drawn from the vast body of data. This is primarily a result of the diverse reaction conditions Hutchings, Scurrell, and Woodliouse employed by the different research groups involved. There is therefore a need to standardize the reaction conditions so that catalysts can be directly compared, and any advance in catalyst design can then be readily assessed. In particular, this comparison should be on the basis of specific act vity data quoted for a range of operating temperatures. Acknoii-ledgement. We wish to thank Richard Joyner, Norman Parkyns, and Justin Hargreaves for useful discussions.

ISSN:0306-0012    

DOI:10.1039/CS9891800251

出版商:RSC    

年代:1989

数据来源: RSC