13b having four contiguous stereocenters was obtained as a
unique stereoisomer when irradiation was performed at -50
or -78 °C (entries 5 and 6).10 Interestingly, cyclizations of
8 and 12 were complete within 0.5 h at -50 °C, leading to
high yield of the cyclized products, in contrast with literature
reports on similar cyclizations, where long reaction times
are usually observed (vide infra).
els.15 Applied to our case, these models effectively provide
a good basis for understanding the sense of the stereoinduc-
tion (Figure 1). The major isomer is likely to be formed
We then extended our study to simple acyclic precursors
15 and 17,11 to test the generalization of the strategy to other
allylic systems (Scheme 4). Surprisingly, 15 led, after 14 to
Scheme 4. Cyclization of Precursors 15 and 17
Figure 1. Beckwith-Houk transition state models for 5-exo-trig
cyclization of 8, 12, and 17.
through a chairlike transition state A, in which the bulky
silicon group prefers the pseudoequatorial position. The
minor isomer could then be formed through a boatlike TS
B, which is higher in energy due to higher torsional strain.16
Interestingly, the electron-rich C-Si bond is nearly aligned
with the incipient C-C bond and is thus able to stabilize
the partial positive charge developing at the â-center. This
is reminiscent of the well-known â-silicon effect17 and may
explain both the unusual rate and the level of stereocontrol
of the process. Steric and electronic effects would thus
reinforce one another in cyclizations of models 8 and 12
having an allylic silicon group. Electronic effects are
probably operating in the case of alcohol 17. Major cyclo-
pentane 18a is likely to be formed through a TS similar to
A, in which the OH group at C2 would prefer the pseudoaxial
position to avoid destabilizing interactions between the
C2-O bond and the partial positive charge developing at
C1.18
24 h, to a mixture of four stereoisomers in good yield but
with low stereocontrol.12 When the reaction was performed
at lower temperature (-78 °C), only traces of the cyclized
product 16 were observed after 8 h of irradiation. On the
opposite side, allylic alcohol 17 afforded, after 2 h at -50
°C, a mixture of cyclopentanes 18a,b in reasonable yield
but again with modest diastereocontrol. Interestingly, the all-
cis isomer 18a was obtained as the major isomer, thus
providing an access to the complementary stereochemistry
to that of the silylated analogues.13
We thus noticed that under analogous conditions, cycliza-
tion of allylsilanes 8 and 12 was faster than that of analogues
having a methyl, a hydroxy, or simply no allylic substituent.
This may be mainly attributed to the known reactivity of
allylsilanes toward electrophiles.14 p-TolSO2 radical is elec-
trophilic in nature and is thus likely to react efficiently with
electron-rich allylsilanes.
Finally, cyclopentanes 13 can be elaborated further in a
simple way, using the pendant olefinic moiety and the
sulfonyl function. This is illustrated below with the efficient
preparation of the tricyclic ketone 19 (Scheme 5). Alkylation
Stereocontrol in these 5-exo-trig cyclizations is usually
rationalized by using Beckwith-Houk transition state mod-
(10) Best results were obtained at -50 °C, using 0.25 equiv of p-TolSO2-
SePh (entries 4 and 5). As shown in entry 6, use of 0.5 equiv of p-TolSO2-
SePh at -78 °C slightly increased the diastereoselectivity, but at the expense
of the yield.
(11) 15 and 17 were prepared in 6 and 4 steps from (R)-(-)-citronellene
and hexa-1,5-diene, respectively (see Supporting Information).
(12) Cyclization of the achiral analogue of 15, lacking the allylic Me
group, provided a 3:1 cis/trans ratio at best, in good agreement with earlier
reports.2
(13) A silicon group may be used as a masked hydroxy group, see: Jones,
G. R.; Landais, Y. Tetrahedron 1996, 52, 7599-7662.
(14) Fleming, I.; Dunogue`s, J.; Smithers, R. Org. React. 1989, 37, 57-
575.
(15) (a) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron Lett. 1985,
26, 373-376. (b) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985,
41, 3925-3941. (c) Spellmeyer, D. C.; Houk, K. N. J. Org. Chem. 1987,
52, 959-974. (d) Beckwith, A. L. J.; Zimmermann, J. J. Org. Chem. 1991,
56, 5791-5796.
(16) Boatlike conformation B with a SiR3 group in the pseudoequatorial
position has been preferred over a chairlike conformation with the bulky
silicon group in the pseudoaxial position.
(17) Lambert, J. B. Tetrahedron 1990, 46, 2677-2689 and references
therein.
(18) In TS A, a pseudoequatorial σC2-H bond, which is a better electron
donor than the σC-O bond would be aligned with the incipient C-C bond
and thus stabilize more efficiently the developing positive charge at C1.
Org. Lett., Vol. 6, No. 3, 2004
327