HETEROCYCLES

Communication | Special issue | Vol. 79, No. 1, 2009, pp. 347-352
Received, 1st September, 2008, Accepted, 9th January, 2009, Published online, 9th January, 2009.
DOI: 10.3987/COM-08-S(D)15

■ Regioselective Ring Expansion of 3,3-Dimethylaziridin-2-carboxylate and a Photochemical Entry to the Penem Nucleus

James D. White* and Takuya Furuta

Department of Chemistry, Oregon State University, Gilbert Hall 153, Corvallis, Oregon 97331, U.S.A.

Abstract

3,3-Dimethylaziridin-2-carboxylates undergo ring expansion with thiocyanates to give a 4,4-dimethylthiazolidin-5-carboxylate as the major product. Irradiation of a N-cysteinyl-3,3-dimethylaziridin-2-carboxylate was found to give a penem in low yield, presumably via a transient thioaldehyde which added across the aziridine N-C(3) bond.

The aziridine ring is a valuable synthon for the preparation of larger heterocycles via ring expansion.2 For an unsymmetrically substituted aziridine, ring enlargement can be programmed in a manner that results in scission of either of the C–N bonds (paths a or b), or in the special case of azomethine ylides, at the C–C bond (path c) (Figure 1). Our interest in aziridine ring expansion evolved from the proposition that they can lead to novel thiazolidine derivatives and can be used to construct a portion of the penem nucleus.3 Herein, we report that 3,3-dimethylaziridin-2-carboxylate (1) undergoes regioselective ring expansion via either path a or path b with thiocyanates, a process which affords substituted thiazolidines. We have also found that a photochemical application of aziridine ring expansion provides access to the bicyclic nucleus of penems.

Ethyl 3,3-dimethylaziridin-2-carboxylate (1) was prepared by reaction of ethyl 2,3-dibromo-3-methylbutanoate (2)4 with ammonia-saturated ethanol5 (Scheme 1). Exposure of 1 to potassium thiocyanate in hot formic acid led cleanly to crystalline thiazoline 3 (mp 103-104 °C), indicating that aziridine cleavage in this case followed path a. The same cleavage mode was observed when 1 was reacted with carbon disulfide and resulted in crystalline thiazolidinethione 4 (mp 99 °C).6 On the other hand, treatment of 1 with benzyl isothiocyanate in benzene at reflux gave the hydantoin derivative 5. The latter is believed to be formed via transient N-substituted aziridine 6 which undergoes internal acylation by the carboxylic ester. Aziridine fragmentation at bond b then generates the isopropylidene substituent.

We also examined the reactivity of N-substituted aziridine 7 which was prepared by treatment of dibromo ester 2 with benzylamine in benzene at reflux7 (Scheme 2). In contrast to 1, which gave a single product 3 with potassium thiocyanate, the analogous reaction of 7 produced a mixture of regioisomeric thiazolidines 8 and 9, the ratio of which depended upon polarity of the solvent. In a neutral solvent such as THF, the major product was 8, but as the acidity of the reaction medium was increased the proportion of 9 also increased. This suggests that opening of the protonated aziridine from 7 increasingly favors path b of Figure 1, as would be expected on the basis of relative carbocation stabilization at C2 vs C3.

Alkylation of 2-thionothiazolidines has been investigated,8 and in concordance with previous results 4 was found to react with methyl iodide exclusively at the exo sulfur atom to yield 10 (Scheme 3). Thiazoline 10 provided an opportunity to test a strategy for appending functionality at C2 which could be used to build the isopenem nucleus, and it was found that condensation of 10 with the anion of ethyl α-nitroacetate led efficiently to 11.9 However, neither catalytic hydrogenation nor other reduction methods were successful in saturating the double bond of 11.

The Arnstein postulate10 as extended by Baldwin,11 that the penicillin nucleus is formed in Nature via oxidation of a α-aminoadipoylcysteinylvaline tripeptide to a thioaldehyde has gained experimental support,12 and in an attempt to simulate this pathway using aziridine 1 the latter was acylated with S-protected N-Boc cysteine derivative 12.13 Our hope was that photolysis of N-acylaziridine 13 would lead to cleavage of the acetamidomethyl appendage affording thioaldehyde 14 which would then add to the aziridine and yield a penem. However, the only product that could be isolated from this reaction was 15 resulting from Norrish type II cleavage of the acylaziridine. Further photolysis of 15 or treatment of 15 with mercuric acetate at pH 4 gave disulfide 16, presumably via homolytic fracture of the C–S bond adjacent to the acetamide function followed by homocoupling of the sulfide radical.

The formation of 16 upon extended photolysis of 15 and the failure of 13 to furnish a substance with the penem skeleton suggested that a more labile photocleavable group than the acetamidomethyl substituent was needed. Accordingly, 1 was acylated with S-benzoylmethyl-N-Boc cysteine derivative 1714 in the expectation that photolysis of sulfide 18 would result in activation via 19 to produce thioaldehyde 20. In the event, irradiation15 of a solution of 18 in benzene afforded known penicillanic ester 2116 as a single epimer. The major product was again disulfide 16 from homolytic cleavage of 18 and coupling, accompanied by Norrish type II cleavage of the acylaziridine.

Although 18, in which an aziridine serves as surrogate for dehydrovaline gave 21 in low yield, the formation of this penem lends circumstantial support to the Arnstein proposal for penicillin biosynthesis involving a thioaldehyde. Ring strain in 20 and the propensity for the aziridine to undergo expansion is probably responsible for the difference between our result and that of Scott.14

ACKNOWLEDGEMENTS
We are grateful to Professor Sigeru Torii, Okayama University, for helpful discussions and to the U.S. National Science Foundation for financial support.

References

1. Dedicated to the memory of Dr. John Daly, a distinguished alumnus of Oregon State University and a pioneer in the discovery and structural elucidation of heterocyclic natural products.
2. O. C. Dermer and G. E. Ham, ‘Ethylenimine and Other Aziridines’, Academic Press, Inc., New York, 1969, pp. 205-303.
3. T. Kametani and T. Honda, ‘Advances in Heterocyclic Chemistry: Application to the Synthesis of Natural Products,’ Vol. 39, ed. by A. R. Katritzky, Academic Press, Inc., London, 1986, pp. 181-236.
4. M. Prostenik, N. P. Salzman, and H. E. Carter, J. Am. Chem. Soc., 1955, 77, 1856. CrossRef
5. N. H. Cromwell and G. D. Mercer, J. Am. Chem. Soc., 1957, 79, 3819. CrossRef
6. L. B. Clapp and J. W. Watjen, J. Am. Chem. Soc., 1953, 75, 1490. CrossRef
7. P. L. Southwick and D. R. Christman, J. Am. Chem. Soc., 1952, 74, 1886. CrossRef
8. A. G. Goddin and N. E. Searle, U.S. Patent 2,516,313 (1950); (Chem. Abstr., 1950, 45, 810a).
9. K. Hirai, H. Matsuda, and Y. Kishida, Chem. Pharm. Bull., 1971, 20, 97.
10. S. W. Queener and N. Neuss, ‘Chemistry and Biology of β-Lactam Antibiotics,’ Vol. 3, ed. by R. B. Morin and M. Gorman, Academic Press, Inc., New York, 1982, pp. 2-81.
11. M. E. Byford, J. E. Baldwin, C. Y. Shiau, and C. J. Schofield, Chem. Rev., 1997, 97, 2631. CrossRef
12. Y. Kita, N. Shibata, N. Kawano, T. Tohjo, C. Fujimori, and H. Ohishi, J. Am. Chem. Soc., 1994, 116, 5116. CrossRef
13. D. F. Veber, J. D. Milkowski, S. L. Varga, R. G. Denkewalter, and R. Hirschmann, J. Am Chem. Soc., 1972, 94, 5456. CrossRef
14. J. Cheney, C. J. Moores, J. A. Raleigh, A. I. Scott, and D. W. Young, Chem. Commun., 1974, 47. CrossRef
15. Irradiation of a 0.01M solution of substrate in benzene at 6 °C was carried out for 2 h through Pyrex glass (> 300m) with a 450W medium pressure mercury lamp.
16. (a) H. R. Pfaendler and H. Hoppe, Heterocycles, 1985, 23, 265; CrossRef (b) Y. Asano, Y. Kato, A. Yamada, and K. Kondo, Biochemistry, 1992, 31, 2316. CrossRef