carbinol derivatives. We also show that the level of diaste-
reocontrol is highly dependent on the nature of the protective
groups.
Table 1. Effect of the Nature of the Reagent
We envisioned that the introduction of a bulky protective
group may actually prevent preassociation of the zinc reagent
with the substrate and thus lead to the formation of the anti
diastereomer through a transition structure that minimizes
steric interactions between the reagent and the substituents
(Figure 1, A is favored over B). The ground-state conforma-
entry
reagent
IZnCH2Id
conversion (%)b
anti:sync
1
2
3
4
5
6
7
8
8
48
44
92
88
60
52
>99
84:16
92:8
90:10
94:6
97:3
89:11
93:7
EtZnCH2I
Zn(CH2I)2
e
EtZnCH2Cl
e
Zn(CH2Cl)2
2,4,6,-Cl3C6H2OZnCH2I
2,4,6,-F3C6H2OZnCH2I
CF3CO2ZnCH2I
>99:1
a Reaction conditions: substrate is added to 2.0 equiv of preformed
reagent, CH2Cl2, 0 °C to room temperature. b Determined by 1H NMR using
an internal standard. c Determined by GC analysis of the corresponding
acetate derivatives. d Formed by treating 2.0 equiv of I2 with 2.0 equiv of
Et2Zn and 2.0 equiv of CH2I2. e 1.0 equiv of reagent was used.
useful conversions and diastereoselectivities were observed
with the Denmark’s procedure (entries 4-5).8 The aryloxide-
derived carbenoids9 gave lower conversions, probably due
to the bulkiness of the reagent making the approach to the
alkene more difficult (entries 6-7). A highly activated
reagent prepared from trifluoroacetic acid10 (Shi’s reagent)
gave the best conversion and diastereoselectivity, making it
the optimal carbenoid for the following studies.
Figure 1. Postulated transition structures for directed vs non-
directed cyclopropanation reactions.
tion in which the silyloxy group is synclinal to the alkene
has been proposed by Gung to be the most stable for small
R4 substituents.6 Nonassisted attack of the carbenoid on this
favored conformer (which should also be the most reactive
on the basis of stereoelectronic arguments) should lead to
the anti isomer. This approach contrasts with the cyclopro-
panation of unprotected allylic alcohols in which minimiza-
tion of the A-1,3 strain is the main controlling element for
good diastereocontrol (Figure 1, D is favored over C). One
potential problem with this approach is the lack of reactivity
of the classical zinc carbenoid reagents, which sometimes
are not sufficiently reactive to cyclopropanate alkenes that
do not contain basic proximal groups. However, several
improved versions of the reagent have recently been reported
and should prove to be sufficiently reactive to provide the
cyclopropane product in good yields with this class of
substrates.
The reaction was initially performed by reacting the chiral
racemic triisopropylsilyl ether derived from (E)-4-phenylbut-
3-en-2-ol to different zinc carbenoids (Table 1). As expected,
the classical Simmons-Smith reagent7 gave only poor
conversion, though the selectivity was in favor of the desired
anti diastereomer (entry 1). The selectivities increased with
Furukawa-type carbenoids (entries 2-5), and synthetically
The next step was to study the effect of the protective
group on the selectivity of the reaction (Table 2). We
compared the level of diastereoselection as a function of the
size of the silyl ether protective group. Gratifyingly, all but
two substrates gave the anti diastereomer as a major product
with an excellent ratio (anti:syn ratio superior to 94:6) when
treated with Shi’s reagent. It is also apparent that the
diastereoselectivities usually increased by increasing the
steric bulk of the silyl ether, the triisopropylsilyl being the
optimal protective group for maximizing the anti:syn ratio.11
As expected, the nature of the allylic substituent (R4) also
influences the ratio considerably (entries 6-9). While the
small methyl and ethyl allylic substituents allowed excellent
diastereoselectivities, the level was lower in the case of
isopropyl (entries 2 and 7 vs 8). Conversely, the presence
of a methyl substituent at the R3 position did not seem to
affect the level of diastereoselection, yielding the corre-
sponding trisubstituted cyclopropylcarbinyl silyl ethers in
good yields and excellent diastereomeric ratios (entries 10-
11).
It is well established that (Z)-allylic secondary alcohols
give high levels of syn selectivity when treated under the
(8) Denmark, S. E.; Edwards, J. P. J. Org. Chem. 1991, 56, 6974.
(9) Charette, A. B.; Francoeur, S.; Martel, J.; Wilb, N. Angew. Chem.,
Int. Ed. 2000, 39, 4539.
(10) (a) Yang. Z.; Lorenz, J. C.; Shi, Y. Tetrahedron Lett. 1998, 39,
8621. (b) Charette, A. B.; Beauchemin, A.; Francoeur, S. J. Am. Chem.
Soc. 2001, 123, 8139.
(11) Reaction of a bulkier allylic tert-butyldiphenylsilyl ether led mostly
to decomposition of the starting material, which in this case probably occurs
faster than the cyclopropanation under these Lewis-acidic conditions.
(5) (a) Furukawa, J.; Kawabata, N.; Nishimura, J. Tetrahedron Lett. 1966,
3353; (b) Tetrahedron 1968, 24, 53.
(6) (a) Gung, B. W.; Melnick, J. P.; Wolf, M. A.; King, A. J. Org. Chem.
1995, 60, 1947. (b) Khan, S. D.; Pau, C. F.; Chamberlin, A. R.; Hehre, W.
J. J. Am. Chem. Soc. 1987, 109, 650.
(7) Simmons, H. E.; Cairns, T. L.; Vladuchick, S. A.; Hoiness, H. M.
Org, React. 1973, 20, 1.
3352
Org. Lett., Vol. 4, No. 20, 2002