Lewis acid mediated cascade reactions of silyl-substituted methylenecyclopropyl ketones

Article abstract of DOI:10.1016/S0040-4039(99)02342-4

The Lewis acid mediated cyclisation of various silyl-substituted methylenecyclopropyl ketones has been investigated. The presence of the silyl-substituent enhances the reactivity of the methylene cyclopropane in comparison to our earlier study on non-sily

Full text of DOI:10.1016/S0040-4039(99)02342-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 1615–1618
Lewis acid mediated cascade reactions of silyl-substituted
methylenecyclopropyl ketones
Guillaume L. N. Peron,a John Kitteringham b and Jeremy D. Kilburn a,∗
aDepartment of Chemistry, University of Southampton, Southampton, SO17 1BJ, UK
bSynthetic Chemistry Department, SmithKline Beecham Pharmaceuticals, New Frontiers Science Park, Third Avenue, Harlow,
Essex CM19 5AW, UK
Received 3 November 1999; accepted 16 December 1999
Abstract
The Lewis acid mediated cyclisation of various silyl-substituted methylenecyclopropyl ketones has been investi-
gated. The presence of the silyl-substituent enhances the reactivity of the methylene cyclopropane in comparison to
our earlier study on non-silyl-substituted methylenecyclopropyl ketones, allowing milder Lewis acids (BF3·Et2O
or BF3·2AcOH) to be used for the cyclisation reaction. The mild conditions used allow the allyl cation, formed as
an intermediate in the cyclisation, to be trapped in further carbon–carbon bond-forming reactions. © 2000 Elsevier
Science Ltd. All rights reserved.
Keywords: methylenecyclopropane; Lewis acid; cascade; cyclisation.
We recently reported that treatment of methylenecyclopropyl ketones or aldehydes such as 1 (R=H)
with suitable Lewis acids (TiCl4 or SnCl4) provides a new route to cycloalkanols.1 The mechanism for
this reaction is presumed to proceed by nucleophilic addition of the alkene to the activated carbonyl,
leading to a cyclopropyl cation intermediate 2, which rearranges to an allyl cation 3, which, in turn, is
trapped by chloride anion derived from the Lewis acid (Scheme 1).2
Scheme 1.
Incorporation of a silyl-substituent on the methylenecyclopropane precursor (e.g. 1, R=SiR3) should
serve to encourage the initial alkene addition to the carbonyl, since the silyl group should stabilise the
intermediate cyclopropyl cation3 — indeed the silyl-substituted methylenecyclopropane 1 is, of course,
∗
Corresponding author. E-mail: jdk1@soton.ac.uk (J. D. Kilburn)
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(99)02342-4
tetl 16270

1616
a strained allylsilane.4 In this paper, we describe our studies with such silyl-substituted methylenecyclo-
propyl ketones, which have led to the development of a novel, Lewis acid mediated, cascade process.
Preparation of suitable silylated methylenecyclopropyl ketones was readily accomplished by sequen-
tial deprotonation of methylenecyclopropane, silylation, deprotonation and alkylation with an iodoketal,
carried out in a one-pot procedure, 5 followed by ketal deprotection (Scheme 2).
Scheme 2.
Cyclisation of ketone 8 using TiCl4 or SnCl4 gave cyclohexene 11,6 formed by trapping the interme-
diate allyl cation 10 with chloride ion, in good yield (Scheme 3).
Scheme 3.
In contrast, the analogous non-silylated methylenecyclopropyl ketone (1, R=H) required higher re-
action temperatures and gave lower yields of cyclised products when treated with TiCl4 or SnCl4.1
Furthermore, whereas the non-silylated methylenecyclopropyl ketones had failed to react with milder
Lewis acids such as boron trifluoride,1 when treated with either BF3·Et2O or BF3·2AcOH ketone 8 gave
the unusual bicyclic ether 12 in good yield. Here, in the absence of a sufficiently nucleophilic anion
derived from the Lewis acid, the intermediate allyl cation 10 is trapped intramolecularly by the alkoxide
nucleophile.7
Cyclisation of the homologous silylated ketone 9 with BF3·2AcOH similarly gave the bicyclic ether
13 in reasonable isolated yield (Scheme 4).
Scheme 4.
Treatment of 9 with TiCl4, however, led to a complex mixture from which four compounds 14–17, all
presumably derived from an allyl cation intermediate, were isolated, with vinyl chloride 17 presumably

1617
resulting from a protodesilylation of an allylsilane intermediate cf. 14. Using SnCl4, on the other hand,
cleanly gave cis hydroxychloride 14 as the only isolated product in 65% yield.8
The cyclisation of ketone 9 to give seven-membered ring products is in stark contrast to the cyclisation
of the non-silylated analogue 18, which on treatment with TiCl4 or SnCl4 gave cyclopentanol 19, via
nucleophilic attack by the σ-bond of the methylenecyclopropane (Scheme 5).1
Scheme 5.
In our earlier work we sought to trap the allyl cation, formed as an intermediate in these cyclisations,
with a range of nucleophilic groups grafted on to the starting methylenecyclopropyl ketone. However,
most of these attempts were thwarted by rapid trapping of the allyl cation by chloride ion derived from
the Lewis acid TiCl4 or SnCl4, which were required for the cyclisation. The introduction of a silyl-
substituent, however, allows the cyclisation to be carried out in the absence of a strongly nucleophilic
counteranion, so we again sought to exploit the reactivity of the intermediate allyl cation for further
reaction. To this end we studied the cyclisations of allylsilane 20 and phenylsilane 22 — prepared in
analogous fashion to the trimethylsilyl-substituted ketones 8 and 9.
Treatment of allylsilane 20 with TiCl4 or SnCl4 did not lead to any isolable cyclised products.
Treatment with either BF3·Et2O or BF3·2AcOH, however, gave the silyl fluoride 21, presumably
resulting from intramolecular addition of the allylsilane moiety to the allyl cation intermediate, in
moderate yield (Scheme 6).
Scheme 6.
Treatment of phenylsilane 22 with TiCl4 or SnCl4, gave the cyclised allyl chloride 23 in very good yield
(Scheme 7), again as a consequence of rapid trapping of the intermediate allyl cation by chloride, and
treatment of 22 with BF3·Et2O gave the bicyclic ether 24, also in very good yield, analogous to reaction of
the trimethylsilyl-substituted methylenecyclopropyl ketone 8 (Scheme 3). Treatment of phenylsilane 22
with BF3·2AcOH, however, did now lead to silylfluoride 25, as a single diastereoisomer as shown,8 and
in reasonable yield, accompanied by bicyclic ether 24. Clearly under these conditions the phenyl transfer
from silicon to the allyl cation can compete with intramolecular trapping by the alkoxide nucleophile.
In conclusion, silylated methylenecyclopropyl ketones have proved to be more reactive than the non-
silylated analogues, and cyclise under relatively mild conditions to give six- and seven-membered ring
products in good yields. The mild conditions used also allow the allyl cation, formed as an intermediate
in the cyclisation, to be trapped in further carbon–carbon bond-forming reactions, as exemplified by the
reactions of allylsilane 20 and phenylsilane 22.

1618
Scheme 7.
Acknowledgements
We thank SmithKline Beecham for supporting this work and Thomas Gelbrich for help in determining
the stereochemistry of 14 and of 25.
References
1. Peron, G. L. N.; Kitteringham, J.; Kilburn, J. D. Tetrahedron Lett. 1999, 40, 3045.
2. The mechanism was first proposed by Hosomi et al. who described the Lewis acid mediated intermolecular addition of
methylenecyclopropanes to ketones and aldehydes: Miura, K.; Takasumi, M.; Hondo, T.; Saito, H.; Hosomi, A. Tetrahedron
Lett. 1997, 38, 4587.
3. Hosomi et al. reported that 2-trimethylsilylmethylenecyclopropane was considerably more reactive than simple
methylenecyclopropane in their study on intermolecular additions to ketones and aldehydes; see Ref. 2.
4. For the original preparation of 2-trimethylsilylmethylenecyclopropane and further alkylation reactions, see: Sternberg, E.;
Binger, P. Tetrahedron Lett. 1985, 26, 301.
5. For the original description of this one-pot procedure, see Ref. 4. See also: Thomas, E. W. Tetrahedron Lett. 1983, 24, 1467.
1
1
1
6. All new compounds were characterised by IR, MS, H and 13C NMR, with H–1H and H–13C correlation spectra, where
necessary, to aid the assignments, and by HRMS. Full details will be reported in due course.
7. The conversion of 8 to 12 is essentially a [3+2] cycloaddition of a silylated trimethylenemethane (TMM) equivalent to a
ketone. The use of methylenecyclopropanes as TMM equivalents in [3+2] cycloaddition reactions with olefins, using transition
metal catalysis, is well-documented: (a) Binger, P.; Büch, H. M. Top. Curr. Chem. 1987, 135, 98; (b) Lewis, R. T.; Motherwell,
W. B.; Shipman, M.; Slawin, A. M. Z.; Williams, D. J. Tetrahedron 1995, 51, 3289; (c) Lautens, M.; Klute, W.; Tam, W. Chem.
Rev. 1996, 96, 49.
8. The stereochemistry of 14 and 25 was determined by X-ray crystallographical analysis. Full details will be published in due
course.