Direction of Ring Opening of Styrene Oxide and Butadiene Monoxide by Ester Carbanions$ Stefan T. Orszulik*,% and Maurice Porter Malaysian Rubber Producers' Research Association, Tun Abdul Razak Laboratory, Brickendonbury, Hertford, Herts SG13 8NL, UK Attack takes place at both primary and secondary epoxy carbons of styrene oxide and butadiene monoxide on reaction with ester carbanions, providing evidence for the conjugative effect in these systems. Styrene oxide 1 and butadiene monoxide (3,4-epoxybut-1- ene) 2 have been used extensively in the investigation of the in�Puence of conjugative e€ects on the direction of ring opening of epoxides by nucleophiles1,2 but the results have been confused by structural misassignments and failure to identify some signiRcant products. Early workers,3,4 who did not have available the beneRts of modern analytical methods, claimed that diethyl sodiomalonate 3 attacks 1 exclusively at the terminal carbon to give eventually the 4-phenyl lactone 10h via the sequence shown (Scheme 1).This Rnding was widely quoted1,5�}7 but later shown to be incorrect, evidence being obtained that both 10h and 11h are formed in signiRcant amounts,8�}11 though one worker12 obtained only 11h via the isolated acid 9g. Reaction of 3 with the vinyl epoxide 2 was claimed also to take place exclusively at the primary epoxide carbon atom to give 10i3,4 and this assertion has also received wide acceptance;1,6,7 similarly the reaction of sodiocyanoacetate with styrene oxide 113 and with butadiene monoxide 213 was claimed to proceed by attack on the terminal carbon to give the products 10h and 10i respectively. In these cases the reactions do not appear to have been studied thoroughly since: there is one report11 in which styrene oxide 1 has been shown to give both 3- and 4-phenyl lactones on reaction with cyanoacetate 5, though there is some con- fusion regarding the melting points given in this report (see below). The position of attack on epoxides such as 1 and 2 by nucleophiles like 3, 4 and 5 has been thought to be deter- mined by a balance between the conjugative e€ect exerted by the phenyl or vinyl group, which should facilitate attack at the neighbouring secondary carbon atom to give the ion 7, and steric hindrance, which will predispose attack at the primary carbon to give ion 6.As can be seen from Table 1, the experimental observations prior to this paper have pro- duced a confusing picture.The degree of steric hindrance in the attacking carbanion falls in the order: 3>4w5, whilst that of the epoxide is 1>2. At Rrst sight, the results for 4 and 5 are anomalous since these should produce much larger proportions of the isomers 7. However, reference to the original literature shows that the identiRcation of 6c and 6f is doubtful. IdentiRcation of 6e is almost certainly wrong, the product isolated having been assigned13 the structure 8e on the basis of its hydrolysis to 8g and comparison with a supposedly authentic sample of 8g,4 the source of which is doubtful.9 Comparison of melting points of the acids 8g10 and 9g9�}12 given in the literature suggests that the product said to be 8e is almost certainly 9e, derived from the ion 7e.Since we required a pure sample of lactone 10i for another purpose, we attempted its preparation using the reaction conditions of Russell and VanderWerf.3 In fact, this procedure gave us a 30:70 mixture of 10i and 11i with an overall yield of 54%.The two isomers were cleanly separated by preparative HPLC and were identiRed by their 1H NMR spectra.14 In our hands reaction of ethyl cyanoacetate 5 with styrene oxide 1 using established conditions13 gave a mixture of 8e and 9e, separated by fractional crystallisation and identiRed by their 1H NMR spectra.15 Zuidema et al.13 claim to have obtained the diastereoisomers of 8e, the melting points of which correspond to those of 8e and 9e of the present work.17 The melting point of the acid 8g given by Zuidema J.Chem. Research (S), 1998, 258�}259$ Scheme 1 Table 1 Ions 6a�}f and 7a�}f formed by attack of ester carbanions 3�}5 on epoxides 1 and 2 deduced from the literature Ion(s) formed from Ester carbanion Styrene oxide 1 Butadiene monoxide 2 Malonate3 6a3,4 6b3,4 7a12 6ba7ba (30:70) 6aa7a8�}11 Acetoacetate4 6c16 6da7d16 Cyanoacetate5 6e13b 6f13 6ea7ea (46:54) 6fa7fa (32:68) aThis paper.bAlmost certainly 7e, see text. $This is a Short Paper as deRned in the Instructions for Authors, Section 5.0 [see J. Chem. Research (S), 1998, Issue 1]; there is there- fore no corresponding material in J. Chem. Research (M). %Present address: 6 The Kestrels, Grove, Wantage, Oxfordshire OX12 0QA, UK. *To receive any correspondence (e-mail: stefan@orszulik.demon. co.uk). 258 J. CHEM. RESEARCH (S), 1998et al.13 corresponds to that of 9g given elsewhere.9¡¾12 Hashem and Shaban11 claim to have obtained 8e and 9e, but their melting points do not accord with those of the present work or elsewhere.17 Furthermore, they appear to have interchanged the melting points of 10h and 11h17 compared with their authentic sample.8 Similarly, reaction of 5 with butadiene monoxide 2, using standard conditions,13 proceeds via 6f and 7f, the isolated products being identi¢çed as 8f and 9f by 1H NMR spec- troscopy18 of the materials obtained after preparative HPLC.Zuidema et al.13 claim to have obtained the diaster- eoisomers of 8f which gave, on removal of the cyano group, 10i. Comparison of their IR data with those of the present work suggests that Zuidema's samples of 8f and 10i are all mixtures of the 3- and 4-vinyl lactones. Both 8e and 9e would be expected to be a mixture of dia- stereoisomers. Careful examination of the 1H NMR data19 for 8e, and comparison with published values for other sub- stituted g-lactones, indicates that the major diastereoisomer is cis-2-cyano-4-phenyl-g-butyrolactone 12, the ratio of 12 to 13 being 70:30 by HPLC.Isomers 12 and 13 were separated by HPLC, but subsequent examination by HPLC and NMR spectroscopy showed the samples to have reverted back to the equilibrium mixture. In the case of 9e, the ratio of dia- stereoisomers is 85:15 by HPLC, thought to be the 14 and 15 isomers respectively. In both these cases (8e and 9e) the CHCN hydrogen is labile as observed in D2O experiments using NMR spectroscopy, and thus the ratios of diastereoisomers probably represent their equilibrium mixtures. Taking the results of the present work into account, there is ¢çrm evidence for attack by malonate ion 3 at both epoxide carbon atoms in 1 and 2.Similarly, cyanoacetate ion 5 reacts at both epoxy sites of 1 and 2. In all the cases studied here the major products have been the result of attack on the more substituted carbon of the epoxide, giving predominantly the 3-phenyl and 3-vinyl products.Since simple alkyl mono-substituted epoxides undergo nucleophilic attack under these conditions exclusively at the primary carbon atom, this is evidence for the partici- pation of the conjugative e€ect in the ring opening of these epoxides. Experimental Reaction of butadiene monoxide with malonate was carried out in the manner described in ref. 3; reactions of styrene oxide and butadiene monoxide with cyanoacetate were carried out according to ref. 13. We are grateful to the board of the Malaysian Rubber Producers' Research Association for permission to publish this work, and to M. J. R. Loadman and A. D. Edwards for NMR and HPLC analyses. Received, 10th November 1997; Accepted, 22nd January 1998 Paper E/7/08062J References and notes 1 R. E. Parker and N. S. Isaacs, Chem. Rev., 1959, 59, 737. 2 J. G. Buchanan and H. Z. Sable, in Selective Organic Transformations, ed. B. S. Thyagarajan, Interscience, New York, 1972, vol. 2, p. 1. 3 R. R. Russell and C. A. VanderWerf, J. Am. Chem. Soc., 1947, 69, 11. 4 G. Van Zyl and E. E. van Tamelen, J. Am. Chem. Soc., 1950, 72, 1357. 5 C. K. Ingold, in Structure and Mechanism in Organic Chemistry, Cornell University Press, 1953, p. 342. 6 A. Rosowsky, in Heterocyclic Compounds with Three- and Four- membered Rings, Part 1, ed. A. Weissberger, Interscience, New York, 1964, el, in Steric E€ects in Organic Chemistry, ed. M.S. Newman, Wiley, New York, 1956, pp. 106¡¾114. 8 C. H. DePuy, F. W. Breitbeil and K. L. Eilers, J. Org. Chem., 1964, 29, 2810. 9 P. M. G. Bavin, D. P. Hansell and R. G. W. Spickett, J. Chem. Soc., 1964, 4535. 10 S. A. M. Tayyeb Hussain, W. D. Ollis, C. Smith and J. F. Stoddart, J. Chem. Soc., Perkin Trans. 1, 1975, 1480. 11 A. I. Hashem and M. E. Shaban, Indian J. Chem., Sect. B, 1981, 20, 807. 12 T. Jakobiec, Diss. Pharm. Pharmacol., 1971, 23(4), 401. 13 G. D. Zuidema, P.L. Cook and G. Van Zyl, J. Am. Chem. Soc., 1953, 75, 294. 14 10i: 2.4 (2 H, m, H2); 2.4¡¾1.9 (2 H, m, H3); 4.82 (1 H, m, H4); 5.1¡¾5.5 (2 H, m, vinyl CH2); 5.7¡¾6.1 (1 H, m, vinyl CH). 11i: 2.4 (2 H, m, H2); 3.2 (1 H, sextuplet, H3); 3.92¡¾4.36 (2 H, dq, H4); 5.1¡¾5.3 (2 H, m, vinyl CH2); 5.6¡¾6.1 (1 H, dq, vinyl CH). 15 8e: 7.3 (5 H, C6H5): 5.85 (0.22 H, PhCH0O); 5.6 (0.55 H); 4.35 (0.9 H); 2.2¡¾3.2 (2, 5 H). 9e: 7.3 (5 H, C6H5); 4¡¾5 (3 H, m). 16 R. M. Adams and C. A. VanderWerf, J. Am. Chem. Soc., 1950, 72, 4368. 17 8e: mp 131.4¡¾132.5 8C (113¡¾115,11 diastereoisomers 93¡¾94 and 133¡¾134 8C13). 9e: mp 93.4¡¾94.0 8C (93¡¾95,12 126¡¾128 8C11). 10h: mp (45¡¾46,11 37¡¾38,8 36¡¾37 8C9). 11h: mp (37¡¾38,11 45¡¾46,8 49¡¾50,12 47¡¾48 8C9). 18 8f: 5.6¡¾6.0 (CH1CH2); 5.2¡¾5.5 (CH1CH2); 4.8¡¾5.0 (CH0O); 3.5¡¾4.0 (CHCN); 2.0¡¾3.0 (CH2). 9f: 5.0¡¾6.0 (CH1CH2); 3.5¡¾4.5 (CHCN and CH20O); 3.0¡¾4.0 (CHCH2). 19 12: 5.5¡¾6.0 (dd, J3 ¢§4,3 ¢§4, 5.4, 10.9 PhCH); 4.2¡¾4.6 (dd, J3 ¢§2,3 ¢§2, 8.3, 12.0, CNCH, D2O exchangeable); 3.1¡¾3.5 (m, H3 ); 2.5¡¾3.0 (m, H3 ). 13: 5.8¡¾6.0 (dd, J3 ¢§4,3 ¢§4, 8.0, 7.0, PhCH); 4.2¡¾4.4 (dd, J3 ¢§2,3 ¢§2, 9.0, 10.0, CNCH). Assignments con¢çrmed by decoupling experiments. J. CHEM. RESEARCH (S), 1998 259