765
alkylation product 1 was obtained, albeit in low yield, along with a large amount of cyclopentanone.
An important observation was made in this reaction: immediately after the addition of TiCl4 the reaction
mixture turned dark-red — the characteristic colour of Ti-enolates.6 This was surprising, as the formation
of Ti-enolates from silyl enol ethers and TiCl4 was reported to take place only at rt, where so-formed
intermediates undergo rapid decomposition.7 Attempts to pre-form the Ti-enolate by allowing the TMS-
enol ether to react with TiCl4 prior to the addition of oxirane resulted in exclusive formation of a
cyclopentanone self-condensation product. However, in situ formation of cyclopentanone Ti-enolate by
the addition of excess TiCl4 (3 equiv.) to a solution of the silyl enol ether (1 equiv.) and ethylene oxide (2
equiv.), allowing the reaction mixture to reach −50°C, and quenching the reaction after the red coloration
fades, afforded 2-(2-hydroxyethyl)-cyclopentanone 1 in 46% isolated yield. Encouraged by this result,
and in order to test the generality of this new reaction, we prepared a number of structurally different
TMS-enol ethers and submitted them to the above reaction conditions. In all cases the desired products
were obtained in moderate to good yields; the results of these experiments are summarized in Table
1. Acyclic, cyclic, aromatic ketones and enones all reacted cleanly giving the alkylated products and
varying amounts of the starting ketone as the only constituents of the reaction mixtures after work-up.8
In the case of the steroid 4 two products were isolated, namely 5 and its acetal derivative 6. The latter was
subsequently hydrolyzed into 5 (H2O, HCl, THF, 93%), to afford the desired compound 5 in 70% total
yield. Remarkably, this reaction proceeded with complete stereoselectivity, a single isomer was obtained,
and was tentatively assigned as 5α.
Scheme 1.
We believe that the reaction proceeds through the intermediacy of a titanium enolate, as outlined in
Scheme 1. This hypothesis is supported by several observations: (a) the reaction mixture is intensively red
coloured, and the coloration vanishes immediately upon aqueous quenching; (b) yields are substantially
reduced, and the amount of unreacted parent ketone increased, when quenching is performed before
the colour fades out; (c) under the same, or even more forcing conditions, the reaction of TMS-enol
ethers with oxirane fails to produce detectable amounts of products when TiCl4 is substituted by other
Mukaiyama-type activators, such as TMSI, TMSOTf, BF3·Et2O or SnCl4, indicating that formation of
the Ti-enolate is essential for the success of the reaction.
To summarize, TiCl4-promoted reactions of enoxysilanes with ethylene oxide offer a new method of
alkylation of carbonyl compounds. Work is in progress in our laboratory in order to establish the scope
and limitations of this reaction, and to provide further mechanistic clarification.
References
1. Reviews on the reactions of epoxides with nucleophiles: (a) with non-stabilized carbanions: Klunder, J. M.; Posner, G. H. In
Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 3, pp. 223–226; (b)
with vinyl carbanions: Knight, D. W. In Comprehensive Organic Synthesis; Trost, B. M.; Fleming, I., Eds.; Pergamon Press:
Oxford, 1991; Vol. 3, pp. 262–266; (c) with alkynyl carbanions: Garratt, P. J. In Comprehensive Organic Synthesis; Trost, B.
M.; Fleming, I., Eds.; Pergamon Press: Oxford, 1991; Vol. 3, pp. 277–280
2. (a) Carboxylic acid dianions: Danishefsky, S.; Kitahara, T.; Schuda, P. F.; Etheredge, S. J. J. Am. Chem. Soc. 1976, 98,
3028–3030; Grieco, P. A.; Ohfune, Y.; Majetich, G. J. Org. Chem. 1979, 44, 3092; Iwai, K.; Kosugi, H.; Uda, H.; Kawai, M.