Synthesis of beta;-1-homonojirimycin and beta;-1-homomannojirimycin using the enzyme aldolase

摘要

J. CHEM. SOC. PERKIN TRANS. 1 1994 231 Synthesis of p-1 -Homonojirimycin and p-I -Homomannojirimycin using the Enzyme Aldolase Karen E. HoW Finian J. Leeper*.B and Sheetal Handab a University Chemical Laboratory, Lensfield Road, Cambridge CB2 I EW, UK Department of Chemistry, King's College London, Strand, London WC2R 2LS, UK The four stereoisomers of the four-carbon azido sugar 11 have been stereoselectively synthesised by a route involving Sharpless epoxidation and all are found to be substrates for rabbit muscle fructose 1.6- bisphosphate aldolase, giving (after treatment with phosphatase) 6-azido-6-deoxyheptuloses 14; hydrogenation of 14a and 14b gave p-1-homomannojirimycin 15a and p-1-homonojirimycin 15b with high selectivity. Deoxynojirimycin (DNJ) 1 is a potent inhibitor of exogluco- sidases due to its close similarity to glucose and its positive charge when protonated, which resembles the positively charged transition state for glycoside hydrolysis.DNJ, and more particularly its N-butyl derivative, show good activity against HIV as a result of inhibition of one of the glucosidases responsible for processing the glycoproteins that form the outer coat of the virus.* The epimer at C-2, deoxymannojirimycin (DMJ) 2, is also a natural product and a potent inhibitor of manno~idases.~However, because DNJ and DMJ lack any substituents at C-1, they show little selectivity for inhibition of a-us. 0-glucosidases. Furthermore, they show little inhibition towards endoglycosidases, which generally only bind com- pounds with at least two sugar ~nits.~.~ We, therefore, set out to synthesise the a-and p-1-homonojirimycins1 and 19,as well as their C-2 epimers, a-and P-1-homomannojirimycins 15d and 1511.It was reasoned that the extra carbon substituent at C-1 would lead to selective inhibition of a-or P-glycosidases, as appropriate, and that the hydroxy group could provide a point of attachment for further binding groups, such as another sugar unit.Our approach to the synthesis of compounds 15a-d envisaged the use of fructose 1,6-bisphosphate (FBP) aldolase to make the linear C7ketones 14a-d,having an amino function masked as an azide group. Hydrogenation of 14would unmask the amine, which would condense with the ketone and the resulting imine would be further hydrogenated to give the saturated piperidine ring of 15.The reaction catalysed in vivo by FBP aldolase is the reversible aldol reaction between dihy- droxyacetone phosphate (DHAP) 4 and glyceraldehyde 3-phosphate (G3P) 3 to give FBP 5 (which exists predominantly HlwHO X 1 DNJ, X=OH,Y=H 2 DMJ, X=H,Y=OH HO OH 0 0~0~~-5 FBP 4 DHAP scheme 1 in the cyclic furanose form) (Scheme 1).However, many studies have shown that the commercially available FBP aldolase from rabbit muscle (RAMA), although specific for DHAP, can accept a wide range of aldehydes in place of G3P.6 In the past 5 years, several publications have appeared describing similar approaches using RAMA to synthesise imino sugars but only C5 and C, sugar analogues have been made.'** For the synthesis of the required C, intermediates 14we need C, aldehydes such as 10(and its stereoisomers).Our synthesis of 10 started with cis-but-2-ene- 1 ,Cdiol 6, which was mono-protected with tert-butyldiphenylsilyl chloride. This large protecting group gave good selectivity for mono-us. di-protection. Sharpless epoxidation g of the resulting allylic alcohol, using L-( +)-diethy1 tartrate, then produced the (2S,3R)-epoxide7.The enantiomeric excess of 7was determined from the I9FNMR spectrum of its Mosher's ester to be ca. 89. Oxidation of the epoxy alcohol 7 to the corresponding aldehyde proceeded well with the mild oxidising agent tetrapropyl- ammonium perruthenate (TPAP) and the aldehyde was protected as its dimethyl acetal 8 using Amberlyst-15 in trimethyl orthoformate. Nucleophilic opening of the epoxide ring of 8with sodium azide proceeded, as expected,' via attack at the end further from the acetal group to give almost exclusively the 3-azido acetal 9.Finally, deprotection in aqueous trifluoroacetic acid gave our target aldehyde 10. i,ii iii,iv HO nOH ROROAoH 6 7 OMe IV tdo;rH: xoMeN3 vii = vi RO RO OMe 1la 10 9 12 13 llb R = Bu'Ph2Si Scheme 2 Reagents: i, BuLi, BDPSiC1; ii, DET, Ti(OPr'),, Bu'OOH; iii, TPAP, NMMO;iv, Amberlyst-15, CH(OMe),; v, NaN,, NH,CI; vi, CF,CO,H, CHCI,, H,O; vii, TBAF; viii, PCC; ix, DiBAl-H Table 1 Inhibition constants (Kimmol dm-') for the inhibition of four glycosidases by P-homomannojirimycin 15a and P-homonojirimycin 15b a-Glucosidasea N.1.' 0.90 fbGlucosidase 3.0 0.43 a-Mannosidase' N.I.' 26 P-Mannosidase 0.08 3.3 From yeast.From almonds. From jack beans. From snail. N.1.- No detectable inhibition at 1 mmol, i.e. Ki ca. 25 mmol dm-3. We had hoped to use the aldehyde 10,having the 4-hydroxy protected, in the enzymic aldol reaction but preliminary tests showed that, though it was a substrate, it reacted very slowly (compared with a good substrate such as propionaldehyde) and the reaction never proceeded beyond about 25 completion, possibly because of the poor solubility of 10.Therefore, the silyl protecting group was removed using tetrabutylammonium fluoride (TBAF).The resulting hydroxy aldehyde cyclised spontaneously to give the lactol lla as a mixture of epimers at the hemiacetal centre. We wondered whether this lactol formation might prevent the enzymic aldol reaction, which presumably requires the open-chain aldehyde; in fact, it is reported that the four-carbon sugar D-erythrose is not a substrate.6 However, when assayed with RAMA and DHAP lla was found to be a good substrate, reacting with an initial rate about one fifth as fast as propionaldehyde, and the reaction proceeded to completion. In order to synthesise the diastereoisomer llb, the trans- allylic alcohol 13 was required. As trans-but-Zene-l,4-diol is not commercially available, 13 was made by monoprotecting the cis-diol6, as before, and then oxidising to the aldehyde using pyridinium chlorochromate (PCC).These (acidic) conditions also caused isomerisation of the cz3-to the trans-aldehyde 12 and subsequent reduction with DiBAL yielded the desired trans-alcohol 13.Sharpless epoxidation of this allylic alcohol, using D-( -)-diethy1 tartrate, gave an enantiomeric excess of 98 (only one set of signals could be seen in the 'H and 19F NMR spectra of its Mosher's ester) and the other reactions proceeded as before to give the lactol llb. Finally, repetition of the same two syntheses using the opposite enantiomers of diethyl tartrate in the epoxidation reactions yielded the enantiomeric lactols llc and lld. In assays with RAMA and DHAP, all four lactols lla-d reacted at similar rates and the reactions all proceeded to completion.Preparative runs typically used about 600 mg (4mmol) of lactol and 2.7 mmol of DHAP with 11 mg (1 50 units) of RAMA. The resulting product was immediately dephosphorylated at pH 4.8 with acid phosphatase to give, in yields of up to SO, the 6-azido-6-deoxyheptuloses14a-d,which existed predominantly in their furanose forms. Hydrogenation of the azido heptuloses 14a-dis, as explained above, expected to proceed with cyclisation and further reduction of the imine formed. This reduction of the imine introduces the fifth asymmetric centre and the hydrogen could, in principle, be added from below to give imino sugars of D-configuration 15or from above to give the L-configuration, e.g.16. Hydrogenation of similar compounds to 14 (lacking the terminal hydroxymethyl group) has been reported by Wong's group and others7*8,13*'4 to give only the compounds of D-configuration (DNJ and DMJ) but in our case the extra substituent (CH,OH) on the carbon adjacent to the nitrogen atom would be expected to have some directing influence. This is indeed what has been observed in a published synthesis of p-L-homofuconojirimycin." In our case also, the influence of the CH,OH group was marked: the hydrogenations of 14aand 14b J. CHEM.soc. PERKIN TRANS. 1 1994 N3 OH ,lla'SOH llbG , OH i, li i, iiI1 H? ?HHOWOH N3 OH 0 N, OH 0 14a 14b J 1iii iii HO 15a 15b 'SOH i, ii lid I H O G O H N, OH 0 14d ,:iii /lii IHO (15c OH 15d 'OH Ho 16 Scheme 3 Reagents: i, RAMA, DHAP; ii, acid phosphatase; iii, 10 Pd/C, H2, 40psi.were both stereospecific giving p-1-homomannojirimycin 15s and p-1 -homonojirimycin 15b,respectively, in 90 yield. The latter product is symmetrical and so has very much simplified 'H and "C NMR spectra. In both these cases the CH20H group at C-1 of the product is up and in the equatorial position thus presenting no hindrance to addition of hydrogen to the imine from below, which results in the CHzOH group at C-5 also being in the more favourable equatorial position. Hydrogenation of 14, however, was not stereospecific and yielded a mixture of 15d and 16in a 1 :1 ratio and 73combined yield.Clearly when the CH,OH group at C-1 is pointing down (i.e. a) it does interfere with the hydrogenation process. Hydrogenation of the fourth isomer 14c has not, as yet, yielded any well characterised product. It is not certain whether this is because the reduction does not proceed as expected or the product(s) are unstable; further investigation of this reaction is needed. The two a-isomers 15c and 15d are both known compounds: 15c is a natural product 16*" and both isomers have been synthesised by lengthy synthetic procedures involving much use of protecting The two p-isomers 15a and 15b are, to our knowledge, both new compounds. The relatively short and efficient syntheses described here illustrate well the advantages of using enzymes for key reactions in synthesis.The two p-imino sugars synthesised here, p-1-homo-mannojirimycin 15a and p-1-homonojirimycin 19, have been tested as inhibitors of four representative glycosidases, a-and J. CHEM. soc. PERKIN TRANS. 1 1994 p-glucosidases and a-and p-mannosidases. Compound 15a, which has the p-mannose stereochemistry, showed excellent selectivity, inhibiting P-mannosidase strongly but the other three enzymes much less, see Table 1. Compound 15b, with the p-glucose stereochemistry, does inhibit P-glucosidase more than the other three enzymes but the difference is less marked than for 15a. In comparison, it is reported that a-homonojirimycin 1 has a much greater effect on a-than p-glucosidase activity16 and a-mannosidases are inhibited by 15d but p-mannosidases are not.,' The specificity of inhibition of glyco- sidases shown by the l -homonojirimycin and l-homomanno- jirimycin isomers, and especially by compound 15a described here, may be of considerable value in studying the pathways of oligosaccharide biosynthesis. Experimental Experimental details are given here for the synthesis of p-1-homomannojirimycin 15a.The other compounds described in the text were synthesised in a similar way. Asymmetric epoxidation of the monosilylated butenediol and oxidation of the resulting epoxy alcohol 7using TPAP 24 followed published general procedures. (2R, 3R)-4-( tert-Butyldiphenylsilyloxy)-2,3-epoxybutanal Dimethyl Acetal S.-Amberlyst-15 (2.2 g) was added portion- wise to a solution of (2R,3R)-4-(tert-butyldiphenylsilyloxy)-2,3-epoxybutanal (13.0 g, 38.2 mmol) in anhydrous trimethyl orthoformate (25.1 cm3, 230 mmol) at 0deg;C to 4deg;C.'' The mixture was stirred at this temperature for 2 h and then filtered to remove the Amberlyst-15 resin.Excess of trimethyl orthoformate was removed under reduced pressure and the residue was purified by flash chromatography on a column of SO, (first treated with a solution of 1 Et,N in methanol, then washed with dichloromethane), eluting with dichloromethane, to yield the epoxy acetal 8 (13.1 g, 89) as a colourless oil (Found: M + NH,', 404.2303. C,,H,,O,Si requires M + NH,, 404.2257); RF (CH,Cl,) 0.37; v,,,(CHCl,)/cm-' 3065 (m, Ph-H), 3040 (w, Ph-H), 1255 (m, CH-O-CH), 1108 (s, C-0), 914 (m, CH-O-CH) and 855 (m, CH-0-CH); 6,(300 MHz, CDCl,) 1.08 (9 H, s, Bu'), 3.1 1 (1 H, dd, J4.2 and 6.0,2-CH), 3.24 (1 H, ddd, J3.9,4.2 and 6.3, 3-CH), 3.31 and 3.38 (each 3 H, s, OCH,),3.78(1 H,dd, J6.3and 11.7,4-CHAHB),3.93(1 H,dd, J 3.9 and 11.7, ~-CHAHB), 4.14 (1 H, d, J 6.0, 1-CH), 7.37-7.48 (6 H, m, Ph-H) and 7.70-7.73 (4 H, m, Ph-H); 6,(75 MHz, CDCl,) 19.1 (me,), 26.7 (CMe,), 53.6, 53.76 (2 x OCH,), 55.5 and 56.0 (2 and 3-CH), 62.2 (4-CH,), 101.6 (I-CH) and 127.6, 129.7, 133.0, 133.2 and 135.5 (Ph); m/z (CI, NH,) 404 (loo, M + NH4+)and 340 (7, M + NH, -2 x CH,OH); a;' +2.5 O (c 6 in CHCl,).(2R,3S)-3-Azido-4-(tert-but~l~~henylsilyloxy)-2-hydroxy-butanalDimethylAcetal9.--Sodium azide (10.95 g, 168.5 mmol) and ammonium chloride (3.96 g, 74.0 mmol) were added to a solution of the epoxy acetal 8 (13.0 g, 33.7 mmol) in MeOH- H,O 8 : 1 (300 cm3) and the mixture was heated at reflux with stirring for 44 h.' The mixture was cooled, diluted with water (100 cm3) and extracted with diethyl ether (3 x 200 cm3). The combined organic phases were dried (Na,SO,), filtered and evaporated under reduced pressure. The residue was chroma- tographed on a column of silica (first treated with 3 Et,N in methanol and then washed with 10 EtOAc in hexane), eluting with 10 to 20 ethyl acetate in hexane, to give the starting epoxy acetal8 (2.8 g) and the azido alcohol 9 (8.0 g, 71 based on unrecovered starting material) as an oil (Found: C, 61.6; H, 7.2; N, 9.8 M + NH4+, 447.2499. C2,H3,N,0,Si requires C, 61.5; H, 7.25; N, 9.8; M + NH,, 447.2428); R, (20 EtOAc in hexane) 0.24; v,,,,,(CHCl,)/cm-' 3600-3260 (m, br, OH), 3065 233 (m, Ph-H), 3040 (w, Ph-H), 2100 (s, N,) and 1110 (s, CH-0-CHj);amp;(300MHz,CDCl,) 1.10(9H, S, B~'),2.43(1 H,d, J3.6, OH),3.44and 3.45(each3 H,s,0CH3),3.73-3.65(2H,m,2,3-CH), 3.90(1 H,dd,J5.1 and 10.5,4-CH,HB),3.97(1 H,dd, J7.5 and 10.5, 4-CHAHB), 4.42 (1 H, d, J6.9, 1-CH), 7.38-7.48 (6 H, m, Ph-H) and 7.72-7.68 (4 H, m, Ph-H); 6x75 MHz, CDCl,) 19.0 (me,), 26.6 (CMe,), 54.7 and 55.5 (2 x OCH,), 62.7 (3-CH), 64.4 (4-CH,), 70.4 (2-CH), 104.4 (1-CH) and 127.7, 129.8, 132.8 and 135.5 (Ph); m/z (CI, NH,) 447 (loo, M + NH4+) and 383 (15, M + NH, -2 x CH,OH); a?:+ 15.8" (c 1.6 in CHCl,).(3R,4S)-4-Azidotetrahydrofuran-2,3-diollla.-To a solution of the azido acetal9 (100 mg, 0.23 mmol) in chloroform (1 cm3) and water (ca. 20 mm3) at 0deg;C was added dropwise trifluoroacetic acid (1 cm3). The mixture was stirred at 0 "C for 2.5 h and then concentrated under reduced pressure. The residue was dissolved in THF (0.5 cm3) and stirred with a solution of tetrabutylammonium fluoride in THF (1.1 mol dm-,; 0.63 cm3, 0.69 mmol) at 20 OC for 0.75 h. The mixture was concentrated under reduced pressure and the residue purified by flash column chromatography on silica gel (first treated with 1 Et,N in methanol, then washed with 40 EtOAc in hexane), eluting with 40 to 80 ethyl acetate in hexane, to yield the unstable hemiacetaz lla (27 mg, 81) as an oil; RF(80EtOAc in hexane) 0.38, (0.5 MeOH in CH,Cl,) 0.25; v,,,(liquid film)/m-' 3500-3160 (s br, OH), 2100 (s, N,) and 1040 (m, C-0); 6,-(100 MHz, CD3OD) 63.2, 63.4 (4-CH, two anomers), 65.3, 65.5 (5-CH,, two anomers), 74.1, 74.2 (3-CH, two anomers) and 99.2, 104.1 (2-CH, two anomers).(3S,4R,5R,6S)-6-Azido-1,3,4,5,7-pentahydroxyheptan-2-0ne 14a.-A solution of rabbit muscle aldolase (0.536 cm3, 202 U) in TRIS-maleate buffer (50 mmol drn-,; pH 6.6; 2.29 cm3) was treated with a solution of the triose phosphate isomerase inhibitor 3-bromo-1 -hydroxyacetone phosphate in the same buffer (190 mg drn-,; 1.01 cm3) for 10 min and then added to a mixture of solutions of the hemiacetal lla (650 mg, 4.5 mmol) in DMSO (4.5 cm3) and DHAP', (0.1 mol drn-,; 28 an3, 2.8 mmol) in triethanolamine buffer (0.2 mol drn-,; pH 7; 8.05 cm3).The solution was stirred slowly at room temperature and 50 mm3 aliquots were assayed for DHAP at various time intervals. A control reaction containing no hemiacetal was run and assayed at the same time. When no significant DHAP remained (24 h), the mixture was adjusted to pH 1.5 and left for 2 min. The pH of the mixture was adjusted to 4.8 using aqueous sodium hydroxide (2 rnol dm-,), after which argon was bubbled through it for 15 min and then acid phosphatase (250 mg, 100 U) added to it. After the mixture had been stirred slowly at 37 "C overnight, the test for inorganic phosphate indicated complete reaction.The mixture was then stirred at 75 "C for 5 min, filtered, its pH adjusted to 7 and water removed from it under reduced pressure. The residue was extracted with hot methanol (2 x 100 cm3) and the extracts were evaporated under reduced pressure. The residue was purified by flash column chromatography on silica gel, eluting with 0 to 10 MeOH in EtOAc, to yield the heptanone 14a (371 mg, 56) as an oil. The compound exists as a mixture of furanose anomers (Found: M + NH,', 253.1148. C,H,,N,O, requires M + NH,, 253.1 148); RF (10 MeOH in EtOAc) 0.23; v,,,(liquid film)/cm-' 3500-3200s, br (OH) and 21 10s (N,); 6,(400 MHz, CD,OD) major anomer: 3.34 (1 H, ddd, J 3.9, 4.2 and 7.8, 6- CH), 3.40-3.51 (3 H, m, 7-cH~H~and 1-CH,), 3.61 (I H, dd, J 3.9 and 7.8,5-CH), 3.63 (1 H, dd, J4.2 and 1 1.6, ~-CHAHB), 3.86 (1 H, d, J8.0, 3-CH) and 3.98 (1 H, dd, J7.8 and 7.9, 4-CH); 6,(100 MHz, CD,OD) major anomer: 62.6 (C-7), 64.2 (C-1), 67.0 (C-6), 77.5 and 76.8 (C-4 and 5), 81.3 (C-3) and 103.3 (C-2); minor anomer: 63.1 (C-7), 65.1 (C-1), 65.7 (C-6), 79.3 and 81.6 (C-4 and 5), 84.4 (C-3) and 105.8 (C-2); m/z (CI, NH,) 253 (50, M + NH4+), 235 (92, M + NH, -HZO), 192 (75, M + NH, -HOCH,CHOH) and 60 loo, (CHOH),;+ 17.4O (c 1.2 in CH,OH).p-1-Homomannojirimycin 15a.-A solution of the heptanone 14a (330 mg, 1.40 mmol) in methanol (100 cm3)was degassed with nitrogen for 20 min and then 10 palladium-on-carbon (150 mg) was added to it. The mixture was hydrogenated at 40 psi overnight and then filtered through Celite.The filtrate was evaporated under reduced pressure and the residue purified using a column of Dowex-SOX8 (100 mesh) H+ resin, eluting with water then aqueous ammonia (0.1 mol drn-,), to yield p-1-homomannojirimycin 15a (240 mg, 89) as a solid (Found: MH+, 194.1028; C7H,,N0, requires MH, 194.1028); R, (EtOAc-MeOH-H,O, 12:6 :3) 0.58; v,,,(Nujol)/cm-' 3540-3440s, br (OH) and 3420-32OOs, br (NH); 6,(400 MHz, CD,OD)2.46(1 H,ddd, J2.6,5.6and9.5,5-CH),2.72(1H,dd, J6.2 and 6.5, 1-CH), 3.29 (1 H, dd, J2.6 and 9.5, 3-CH), 3.48 (1 H, t, J9.5 4-CH), 3.58-3.63 (3 H, m, 5'-CH,HB and 1'-CH,), 3.78 (1 H, dd, J2.6 and 11.0, 5'-CHAH,) and 3.81 (1 H, d, J2.6, 2-CH); 6,(100 MHz, CD,OD) 60.5 and 62.6 (C-1 and 5), 63.5 and 62.9 (C-1 'and 5') and 70.5,70.8 and 77.4 (C-2,3 and 4); m/z (CI, NH,) 194 (loo, MH'); Calk0 -4.3 (c 1.3 in MeOH).Acknowledgements We thank the SERC and Shell Research for a CASE studentship (for K. E. H.). References 1 M. L. Sinnott, Chem. Rev., 1990,90, 1 171. 2 G. W. J. Fleet, A. Karpas, R. A. Dwek, L. E. Fellows, A. S. Tyms, S. Petursson, S. K. Namgoong, N. G. Ramsden, P. W. Smith, J. C. Son, F. Wilson, D. R. Witty, G. S. JacobandT. W. Rademacher, FEBS Lett., 1988,237, 128. J. CHEM. SOC. PERKIN TRANS. 1 1994 3 B. Winchester, S. Aldaher, N. C. Carpenter, I. C. Dibello, S. S. Choi, A. J. Fairbanks and G. W. J. Fleet, Biochem. J., 1993,290,743. 4 S. Hiraizumi, U. Spohr and R. G. Spiro, J.Biol. Chem., 1993,268,9927.5 L. J. Liotta, R. C. Bernotas, D. B. Wilson and B. Ganem, J. Am. Chem. SOC.,1989,111,783. 6 M. D. Bednarski, E. S. Simon, N. Bischofberger, W.-D. Fessner, M. J. Kim, W. Lees, T. Saito, H. Waldmann and G. M. Whitesides, J. Am. Chem. SOC.,1989,111,627. 7 K. K. C. Liu, R. L. Pederson and C. H. Wong, J. Chem. SOC.,Perkin Trans. I, 199 1,2669. 8 G. C. Look, C. H. Fotsch and C. H. Wong, Ace. Chem. Res., 1993,26, 182. 9 K. B. Sharpless, Y. Gao, R. M. Hanson, J. M. Klunder, S. Y. KO and H. Masamune, J. Am. Chem. SOC.,1987,109,5765. 10 S. A. Patwardhan and S. Dev, Synthesis, 1974, 348. 11 C. H. Behrens and K. B. Sharpless, J. Org. Chem., 1985,50,5696. 12 F. Effenberger and A. Straub, Tetrahedron Lett., 1987,223, 1641. 13 A. Straub,F. EffenbergerandP. Fischer, J. Org. Chem., 1990,55,3926. 14 T. Kajimoto, L. R. Chen, K. K. C. LiuandC. H. Wong, J.Am. Chem. SOC.,199 1, 113, 6678. 15 G. W. J. Fleet, S. K. Namgoong, C. Barker, S. Baines, G. S. Jacoband B. Winchester, Tetrahedron Lett., 1989,30,4439. 16 G. C. Kite, L. E. Fellows, G. W. J. Fleet, P. S. Liu, A. M. Scofieldand N. G. Smith, Tetrahedron Lett., 1988,29,6483. 17 G. C. Kite, J. M. Horn, J. T. Romeo, L. E. Fellows, D. C. Lees, A. M. Scofield and N. G. Smith, Phytochem., 1990,29, 103. 18 P. S. Liu, J. Org. Chem., 1987,52,4717. 19 P. B. Anzeveno, L. J. Creemer, J. K. Daniel, C.-H. R. King and P. S. Liu, J. Org. Chem., 1989,54,2539. 20 I. Bruce, G. W. J. Fleet, I. C. Dibello and B. Winchester, Tetrahedron Lett., 1989,30, 7257. 21 I. Bruce, G. W. J. Fleet, I. C. Dibello and B. Winchester, Tetrahedron, 1992,48, 10191. 22 S. Aoyagi, S. Fujimaki and C. Kibayashi, J. Chem. SOC.,Chem. Commun., 1990, 1457. 23 G. Legler, Pure Appl. Chem., 1987,59, 1457. 24 W. P. Griffith and S. V. Ley, Aldrichim. Acta, 1990,23, 13. Paper 3/07019K Received 25th November 1993 Accepted 25th November 1993