[3 + 2] annulation of β-heteroatom-substituted α,β-unsaturated acylsilanes with methyl ketone enolates: Scope and investigation of the reaction course

Article abstract of DOI:10.1021/jo0160219

A new route to (Z)-β-silylacryloylsilanes 10 and the improved conditions for the [3 + 2] annulation using 10 and alkyl methyl ketone enolates are reported. Also, details of investigations defining a reaction course of the [3 + 2] annulation using β-phenylthio- and β-trimethylsilyl-acryloylsilanes 1 (X = SPh, SiMe₃) and alkyl methyl ketone enolates are described.

Full text of DOI:10.1021/jo0160219

1786
J . Org. Chem. 2002, 67, 1786-1794
[3 + 2] An n u la tion of â-Heter oa tom -Su bstitu ted r,â-Un sa tu r a ted
Acylsila n es w ith Meth yl Keton e En ola tes: Scop e a n d In vestiga tion
of th e Rea ction Cou r se
Kei Takeda,*,† Kenji Yamawaki, and Noriaki Hatakeyama
Faculty of Pharmaceutical Sciences, Toyama Medical and Pharmaceutical University,
2630 Sugitani, Toyama 930-0194, J apan
takedak@hiroshima-u.ac.jp.
Received August 9, 2001
A new route to (Z)-â-silylacryloylsilanes 10 and the improved conditions for the [3 + 2] annulation
using 10 and alkyl methyl ketone enolates are reported. Also, details of investigations defining a
reaction course of the [3 + 2] annulation using â-phenylthio- and â-trimethylsilyl-acryloylsilanes
1 (X ) SPh, SiMe3) and alkyl methyl ketone enolates are described.
Sch em e 1
In tr od u ction
The development of synthetic methodologies for the
highly efficient and stereoselective construction of five-
membered rings has been a very active area of research.1
In 1993 we reported a novel [3 + 2] annulation approach
to obtain cyclopentenol derivatives that involves the
reaction of â-heteroatom-substituted acryloylsilanes with
ketone enolates.2 The synthetic utility of the reaction has
been successfully demonstrated in the synthesis of natu-
ral products, including untenone A3 and clavulones
(claviridenones).4
An intriguing feature of the reaction of acryloylsilanes
with ketone enolates is that the products and the product
distributions depend on the â-substituent of the acry-
loylsilanes. Thus, in contrast to the observation with (E)-
and (Z)-â-phenylthio derivatives 1 (X ) SPh) in which
isomeric cyclopentenols 3 (major) and 4 (minor) were
obtained in almost the same ratio irrespective of the
acylsilane geometry, the trimethylsilyl derivative 1 (X )
SiMe3) afforded cyclopentenols 5 and uncyclized enol silyl
ether 6 in a ratio depending on the vinylsilane geometry
(Scheme 1). Thus, cyclopentenol 5 was a major product
from (Z)-acylsilane, whereas (E)-acylsilane afforded un-
cyclized enol silyl ether 6 as a major product. On the
other hand, crotonoylsilane 1 (X ) Me) did not afford
cyclopentenol but cyclopropanol derivative 7.5
Resu lts a n d Discu ssion
[3 + 2] An n u la tion Usin g (â-(Tr im eth ylsilyl)- a n d
â-(Dim et h ylp h en ylsilyl)a cr yloyl)sila n es. Although
the [3 + 2] annulation methodology using (â-(phenylthio)-
acryloylsilanes 1 (X ) SPh) has proved to be very efficient
for the synthesis of highly functionalized five-membered
carbocycles, the process of preparing the acryloylsilanes
produces the unpleasant-smelling byproduct thiophenol.
Consequently, we decided to reinvestigate the [3 + 2]
annulation using (â-(trimethylsilyl)acryloyl)silane, which
we previously reported to afford a cyclopentenol deriva-
tive in lower yields relative to the corresponding â-phe-
nylthio derivative and in a better yield from (Z)-1 (X )
SiMe3) than from (E)-1 (X ) SiMe3). Our first task was
to find a stereoselective route to (Z)-10a , because the
reported procedure provides only (E) or a mixture of (E)
and (Z) derivatives6 and to optimize the conditions for
the annulation process using (Z)-10a leading to 5. Also,
we examined [3 + 2] annulation using (Z)-â-(dimethyl-
(phenyl)silyl)acryloylsilane (Z)-10b, which would signifi-
In this paper, we present additional information on the
reaction pathway depending on the â-substituent as well
as the scope of the [3 + 2] annulation.
† Current address: Institute of Pharmaceutical Sciences, Faculty
of Medicine, Hiroshima University, 1-2-3, Kasumi, Minami-Ku,
Hiroshima 734-8551, J apan.
(1) (a) Ramaiah, M. Synthesis 1984, 529-570. (b) Hudlicky, T.; Rulin
F.; Lovelace, T. C.; Reed, J . W. In Studies in Natural Products
Chemistry; Attaur-Rahman, T., Ed.; Elsevier Science: Essex, U.K.
1989; pp 3-72. (c) Ho, T.-L. Carbocycle Construction in Terpene
Synthesis; VCH: New York, 1988.
(2) (a) Takeda, K.; Fujisawa, M.; Makino, T.; Yoshii, E.; Yamaguchi,
K. J . Am. Chem. Soc. 1993, 115, 9351-9352. (b) Takeda, K.; Ohtani,
Y.; Ando, E.; Fujimoto, K.; Yoshii, E.; Koizumi, T. Chem. Lett. 1998,
1157-1158.
(3) Takeda, K.; Nakayama, I.; Yoshii, E. Synlett 1994, 178-178.
(4) Takeda, K.; Kitagawa, K.; Nakayama, I.; Yoshii, E. Synlett 1997,
255-256.
(5) Takeda, K.; Nakatani, J .; Nakamura, H.; Sako, K.; Yoshii, E.;
Yamaguchi, K. Synlett 1993, 841-843.
(6) (a) Reich, H. J .; Kelly, M. J .; Olson, R. E.; Holtan, R. C.
Tetrahedron 1983, 39, 949-960. For stereoselective preparation of (E)-
10a , see: (b) Takeda, K.; Nakajima, A.; Takeda, M.; Yoshii, E.; Zhang,
J .; Boeckman, R. K., J r. Org. Synth. 1999, 76, 199-213.
10.1021/jo0160219 CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/26/2002

[3 + 2] Annulation of R,â-Unsaturated Acylsilanes
J . Org. Chem., Vol. 67, No. 6, 2002 1787
Ta ble 1.
Sch em e 2
ability of the â-substituent of the acryloylsilanes. In the
case of an anion-stabilizing substituent such as a phe-
nylthio group, the carbanion 17 (X ) SPh) generated by
Brook rearrangement immediately delocalizes to give an
intermediate 15. Consequently, the olefin geometry of 15
arising from (E)- and (Z)-1 can be the same if conforma-
tional interconversion occurs faster than cyclization. On
the other hand, in the case of a methyl group, which does
not have R-carbanion-stabilizing ability, carbanion 17
(X ) Me) attacks the â-carbonyl group before allylic
delocalization to give 7 via 16. If it is assumed that a
trimethylsilyl group has less ability to stabilize the
R-carbanion than does a phenylthio group, the results
with (â-(trimethylsilyl)acryloyl)silane 1 (X ) SiMe3) could
be explained as being due to the formation of vinylcyclo-
propanolate followed by oxyanion-accelerated vinylcyclo-
propane rearrangement13 to cyclopentenone (16 f 5) in
which (E)- and (Z)-16 can retain their geometries in 1
(X ) SiMe3) and possibly have different reactivities. This
hypothesis is based on the fact that 3, 4, and 5 were
recovered unchanged after treatment with LDA (1 equiv)
in THF at -80 to 0 °C, respectively, and the fact that
the reaction of 1 (X ) SPh) with lithium enolate 18 of
alkenyl methyl ketone produces [3 + 2] annulation
product 19 in sharp contrast with the reaction of 1 (X )
SiMe3) that affords cycloheptenone derivatives, 21, a [3
+ 4] annulation product (Scheme 4). Our previous study14
has established that the latter process proceeds via
anionic oxy-Cope rearrangement of divinylcyclopropano-
late 20.
yield (%)
R1
R2
Et
n-Pr
i-Pr
Et
11
12
13
10a
10a
10a
10b
10b
10b
Me
Me
Me
Ph
Ph
Ph
75
70
76
72
71
80
5
4
9
3
3
4
9
19
10
10
10
5
n-Pr
i-Pr
cantly enhance the synthetic versatility of the process
because of the ability of the silyl group to act as a
hydroxyl surrogate in combination with Fleming’s oxida-
tive desilylation protocol.7 The new route to (Z)-10a ,b
started from O-protected propargyl alcohol 8,8 which was
converted to 9a ,b by a 2-fold silylation followed by
deprotection. Transformation of 9a ,b into acryloylsilanes
10a ,b was carried out with hydroboration9 followed by
Swern oxidation.10
Having developed an efficient route to (Z)-10, we
optimized the reaction conditions for [3 + 2] annulation
using (Z)-10. The best results were obtained when the
reaction mixture was warmed to 0 °C. A trace amount of
diastereomeric cyclopentenol 12 was detected in addition
to 11, in contrast to the results obtained in our previous
study,2a in which 11a was formed exclusively (Table 1).
Also, unlike in our previous study in which (E)- and (Z)-
13 were obtained in various proportions, (E)-13 was
formed exclusively.
Although we also reexamined the reaction using (E)-
10a by elevating the reaction temperature to 0 °C, no
improvement in the yield of 11a and no formation of
cyclopentenol 12a were observed. The difference in the
product distribution of E and Z derivatives will be
discussed later.
A Mech a n istic P r op osa l for [3 + 2] An n u la tion .
In an earlier stage of this research, as a mechanism to
explain the reaction outcome outlined in Scheme 1, we
proposed the reaction mechanism illustrated in Scheme
3, which involves two competing pathways.11
It seemed that verification of the proposed mechanism
would require (1) comparison of relative R-carbanion-
stabilizing abilities of trimethylsilyl and phenylthio
groups, (2) verification of the stereochemical integrity of
the starting acryloylsilanes under the reaction conditions,
and (3) examination of whether the vinylcyclopropane-
cyclopentene rearrangement (16 f 5) proceeds at tem-
peratures below -30 °C.
(12) For reviews on the Brook rearrangement, see: (a) Brook, M.
A. Silicon in Organic, Organometallic, and Polymer Chemistry; J ohn
Wiley & Sons: 2000. (b) Brook, A. G.; Bassindale, A. R. In Rearrange-
ments in Ground and Excited States; de Mayo, P., Ed.; Academic
Press: New York, 1980; pp 149-221. (c) Brook, A. G. Acc. Chem. Res.
1974, 7, 77-84. For the use of the Brook rearrangement in tandem
bond formation strategies, see: (d) Moser, W. H. Tetrahedron 2001,
57, 2065-2084. Also, see: (e) Ricci, A.; Degl’Innocenti, A. Synthesis
1989, 647-660. (f) Bulman Page, P. C.; Klair, S. S.; Rosenthal, S.
Chem. Soc. Rev. 1990, 19, 147-195. (g) Qi, H.; Curran, D. P. In
Comprehensive Organic Functional Group Transformations, Katritzky,
A. R.; Meth-Cohn, O.; Rees, C. W.; Moody, C. J ., Eds.; Pergamon:
Oxford, 1995; pp 409-431. (h) Cirillo, P. F.; Panek, J . S. Org. Prep.
Proc. Int. 1992, 24, 553-582. (i) Patrocinio, A. F.; Moran, P. J . S. J .
Braz. Chem. Soc. 2001, 12, 7-31.
The first one, path a, involves delocalized allylic
carbanion intermediate 15 that is derived from the 1,2-
adduct 14 by way of Brook rearrangement.12 Path b
involves the cyclopropanolate intermediate 16 that is
generated by internal attack on the carbonyl group by
the carbanion derived by Brook rearrangement. We
considered that the reaction proceeds through either path
a or path b, depending on the R-carbanion-stabilizing
(13) (a) Denheiser, R. L.; Martinez-Davila, C.; Morin, J . M., J r. J .
Org. Chem. 1980, 45, 1340-1341. (b) Danheiser, R. L.; Martinez-
Davila, C.; Auchus, R. J .; Kadonaga, J . T. J . Am. Chem. Soc. 1981,
103, 2443-2446. For a carbanion-accelerated version, see: Danheiser,
R. L.; Bronson, J . J .; Okano, K. J . Am. Chem. Soc. 1985, 107, 4579-
4581.
(7) Fleming, I.; Henning, R. Parker, D. C.; Plaut, H. E.; Sanderson,
P. E. J . Chem. Soc., Perkin Trans. 1 1995, 317-337.
(8) Hoff, S.; Brandsma, L.; Arens, J . F. Recl. Trav. Chim. Pays-Bas
1968, 87, 916. Also, see ref 6b.
(9) Denmark, S. E.; Herbert, B. J . Org. Chem. 2000, 65, 2887-2896.
(10) Mancuso, A. J .; Swern, D. Synthesis 1981, 165-185.
(11) Takeda, K. J . Synth. Org. Chem. J pn. 1997, 55, 774-784.
(14) Takeda, K.; Nakajima, A.; Takeda, M.; Okamoto, Y.; Sato, T.;
Yoshii, E.; Koizumi, T.; Shiro, M.J . Am. Chem. Soc. 1998, 120, 4947-
4959.

1788 J . Org. Chem., Vol. 67, No. 6, 2002
Takeda et al.
Sch em e 3
Ta ble 2.
Sch em e 4
yield (%)
30 (E:Z)
27
conditions
27 (E:Z)
29
31
32
E
E
E
E
E
Z
Z
Z
Z
Z
-80 °C, 10 min
-80° to -70 °C
-80° to -60 °C
-80° to -50 °C
-80° to -30 °C
-80 °C, 10 min
-80° to -70 °C
-80° to -60 °C
-80° to -50 °C
-80° to -30 °C
23 (1:1.4)
20 (1:1.2)
12 (1:1.2)
-
14
-
-
-
-
-
-
-
-
12 (1:1.3)
17 (1:1.2)
24 (1:1)
-
33
53
48
63
68
8
21
40
42
73
-
-
Sch em e 5
1
1
2
2
2
2
2
5
-
-
73 (1:1.9)
66 (1:1.2)
38 (1:1.2)
24 (1:1.3)
-
3 (1:3.3)
9 (1:2.4)
12 (1:1.9)
8 (1:1.2)
-
-
sila n es a n d Lith iu m En ola te of Aceton e. To verify
whether the stereochemical integrity of the acryloyl-
silanes 27 was preserved under the reaction conditions,
low-temperature quenching experiments were carried
out, and the results are shown in Table 2. Loss of the
olefin geometry of 27 was observed in all cases in which
the starting material was recovered. The possibility that
isomerization occurred during the workup/purification
process was ruled out by subjecting 27 without 28 to the
same workup conditions and purification process.
Com p a r ison of r-Ca r ba n ion -Sta bilizin g Abilities
of th e P h en ylth io a n d Tr im eth ylsilyl Gr ou p s. We
compared the R-carbanion-stabilizing abilities of the two
groups based on the relative rates of the base-catalyzed
Brook rearrangement of the â-heteroatom-substituted
R-silylallyl alcohol 22 to 25 and 26.15 This stems from
the fact that the Brook rearrangement is facilitated by
electron-withdrawing substituents.12 Thus, the rates of
formation of 25 and 26 would increase as the R-carban-
ion-stabilizing ability of the substituent X increases. The
result that the half-life of 22a is much shorter than that
of 22b suggests that the phenylthio group stabilizes
R-carbanion more strongly than does the trimethylsilyl
group, at least in this system (Scheme 5), thus, providing
good support for the above-postulated mechanism.
Low -Tem p er a tu r e Qu en ch in g of th e Rea ction of
â-(P h en ylt h io)- a n d (â-(Tr im et h ylsilyl)a cr yloyl)-
The fact that the stereochemical outcome was almost
the same irrespective of the olefin geometry in the [3 +
2] annulation of 27 can be partly attributed to the loss
of the stereochemical integrity of the starting acylsilanes.
To obtain information on the origin of the E-Z isomer-
ization, reaction of (E)-29 with LDA at -80 °C for 10 min
was carried out, and the results are shown in Scheme 6.
The nearly stereorandom formation of acryloylsilanes
27 suggests that a retro-aldol process (29 f 27 + 28)
exists in the [3 + 2] annulation in which isomerization
of the double bond can occur; however, the mechanism
is not clear.
On the other hand, the corresponding reaction of
â-trimethylsilyl derivative 10a with 28 was found to be
different from that of â-phenylthio derivative 27 (Table
3). In the reaction of (E)-10a , 10a was not recovered;
instead, 33, a reduction product by a slight excess of LDA,
(15) Takeda, K., Ubayama, H.; Sano, A.; Yoshii, E.; Koizumi, T.
Tetrahedron Lett. 1998, 39, 5243-5246.

[3 + 2] Annulation of R,â-Unsaturated Acylsilanes
J . Org. Chem., Vol. 67, No. 6, 2002 1789
Sch em e 6
cyclopropanolate intermediates 38 or of allyllithium
derivatives 39 and 42, and an intermolecular protonation
of 41 (Scheme 8). The allyllithium derivatives 39 and 41
may also arise from ring-opening of 38 and/or directly
from 14 (X ) SiMe3). Protonation of 38 followed by a ring-
opening would provide both (E)- and (Z)-35. On the other
hand, while protonation of 39 should afford some amount
of 1,3-silyl-migration product 4017 in addition to (Z)-35,
(E)-35 would become a major product upon protonation
of 42. In the reaction of both (E)- and (Z)-10a , the greater
formation of (E)-35 with increase in reaction temperature
and with progress of the reaction and no formation of
detectable amounts of 40 suggests that 35 could come
from both unreacted 38 and dienolate 42 at lower
temperatures and that 42 becomes a major precusor for
35 with increasing temperature.
Ta ble 3.
Low -Tem p er a tu r e Oxya n ion -Acceler a ted Vin yl-
cyclop r op a n e-Cyclop en ten e Rea r r a n gem en t. Al-
though the vinylcyclopropane-cyclopentene rearrange-
ment only proceeds at high temperatures, usually at
tempertures over 250 °C,18 Danheiser found that an
oxyanion substituent on the cyclopropane ring dramati-
cally accelerated the rearrangement.13a,b,19 However, even
in these cases, a temperature in excess of 25 °C is
required for the reaction. Consequently, in our case (16
f 11, 12), it was questionable whether the rearrange-
ment would proceed at a temperature below -30 °C. We
decided to prepare four diastereomeric vinylcyclopro-
panolate intermediates 16 (X ) SiMe3, R ) Me) by an
independent route and to examine the reactivity to the
cyclopentene derivatives and the stereochemistry of the
reaction. We sought a synthetic route that would permit
the rapid generation of cyclopropanolate even at -80 °C,
and we found that the reaction of 2 equiv of MeLi with
the corresponding cyclopropyl acetates 45 was suitable
for this purpose. The acetates 45 were prepared by the
reaction of dienol silyl ethers 4314 with acetoxy carbene
complex 44 (Chart 1).20,21 The dienol silyl ethers, in turn,
were prepared from the corresponding acylsilanes and
lithiomethyl phenyl sulfone utilizing the Brook rear-
rangement according to Reich’s method.22
yield (%)
10a
conditions
10a (33)
34
35 (E:Z)
36 37
E
E
E
E
E
E
Z
Z
Z
Z
Z
Z
-80 °C, 10 min
-80° to -70 °C
-80 to -60 °C
-80 to -50 °C
-80 to -30 °C
-80 to 0 °C
-80 °C, 10 min
-80 to -70 °C
-80 to -60 °C
-80 to -50 °C
-80 to -30 °C
-80 to 0 °C
11a
9a
31c 30 (1:1.8)
18c 35 (1:1.5)
8c 35 (1.7:1)
6
9
-
-
6a
12
15
17
19
6
-
-
-
-
4a
-
-
35 (3.5:1)
34 (5.6:1)
34 (6.2:1)
2 (1:1.9)
3 (1:1.8)
7 (1.3:1)
6 (2.0:1)
-
-
65b
61b
56b
57b
32b
-
12d
14d
-
-
7
-
12
16
-
3d
-
3
5
6
7 (only E) 42
20 (only E) 60
-
Yield of 33. Yield of (Z)-10a . c Yield of (E)-34. Yield of (Z)-
a
b
d
34.
was obtained.16 In contrast, in the reaction of (Z)-10a , a
significant amount of 10a was recovered without E/Z
isomerization. Moreover, since we previously showed that
(Z)-10a cannot be reduced by LDA, in contrast to the fact
that reduction of (E)-10a by LDA is a rapid process even
at -80 °C, the fact that 33 was not obseved in the
reaction of (Z)-10a suggests that a detectable amount of
(E)-10a was not formed and, hence, that E/Z isomeriza-
tion did not occur.
The reaction of 45 with MeLi (2.2 equiv) was carried
out in both THF and toluene at -80 °C for 30 min and
followed by elevation of the temperature (from -80 ° to
-50 °C and -30 °C) over a 30-min period and then
(17) Nakahira, H.; Ryu, I.; Ogawa, A.; Kambe, N.; Sonoda, N. Bull.
Chem. Soc. J pn. 1990, 63, 3361-3363.
(18) For reviews on vinylcyclopropane-cyclopentene rearrangement,
see: Hudlicky, T.; Reed, J . W. In Comprehensive Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 5, pp
899-970. Hudlicky, T.; Kutchan, T. M.; Naqvi, S. M. Org. React. 1985,
33, 247-335. Also, see: Wong, H. N. C.; Hon, M.-Y.; Tse, C.-W.; Yip,
Y.-C.; Tanko, J .; Hudlicky, T. Chem. Rev. 1989, 89, 165-198.
Also, the reaction of (E)- and (Z)-34 with LDA afforded
33 and (Z)-10a , respectively, without E/Z isomerization
(Scheme 7).
(19) For
a review of charge-accelerated arrangement, see: (a)
The slower reaction of (Z)-10a with 28 relative to (E)-
10a is attributed in part to an unfavorable equilibrium
of 1,2-adduct (Z)-34 toward the starting materials, which
can be ascribed to the relatively severe steric repulsions
in the (Z)-1,2-adduct. A similar trend was previously
observed for reactions of (E,Z)-10a with lithium enolate
of alkenyl methyl ketone.14
Bronson, J . J .; Danheiser, R. L. In Comprehensive Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, 1991; Vol. 5, pp
999-1036. (b) Wilson, S. R. Org. React. 1993, 43, 93-250.
(20) (a) Murray, C. K.; Yang, D. C.; Wulff, W. D. J . Am. Chem. Soc.
1990, 112, 5660-5662. (b) Takeda, K.; Okamoto, Y.; Nakajima, A.;
Yoshii, E.; Koizumi, T. Synlett 1997, 1181-1183. For reviews on the
synthetic applications of Fischer carbene complexes, see: (c) Wulff,
W. D. In Comprehensive Organic Synthesis; Trost, B. M., Fleming, I.,
Eds.; Pergamon: Oxford, 1991; Vol. 5, pp 1065-1113. (d) Wulff, W. D.
In Advances in Metal-Organic Chemistry; Libeskind, L. S., Ed.; J AI
Press Inc.; Greenwich, CT, 1989; Vol. 1, pp 209-393. (e) Do¨tz, K. H.
Angew. Chem., Int. Ed. Engl. 1984, 23, 587-608.
There are several possibilities concerning the origin of
35 in Table 3, including protonation during workup of
(21) The relative stereochemistry was assigned on the basis of
results of NOESY experiments.
(16) (a) Takeda, K.; Ohnishi, Y. Org. Lett. 1999, 1, 237-239. For a
review of reduction with lithium dialkylamides including LDA, see:
(b) Majewski, M.; Gleave, D. M. J . Organomet. Chem. 1994, 470, 1-16.
(22) Reich, H. J .; Holtan, R. C.; Bolm, C. J . Am. Chem. Soc. 1990,
112, 5609-5617.

1790 J . Org. Chem., Vol. 67, No. 6, 2002
Takeda et al.
Sch em e 7
Sch em e 8
even at -80 °C. This result indicates that oxyanion-
accelerated vinylcyclopropane-cyclopentene rearrange-
ment in the case of substrates having the substitution
pattern shown in structure 16 (X ) SiMe3) (Scheme 3)
can proceed at temperatures below -30 °C25 and conse-
quently provides support to our assumption of interme-
diacy of vinylcyclopropanolate derivative 16 in the [3 +
2] annulation. Furthermore, the same trend in product
distribution and difference in reactivity depending on the
vinylsilane geometry as that observed in the [3 + 2]
annulation was observed. The greater formation of (E)-
35 with increases in reaction temperature and with
progress of the reaction is also consistent with the results
of [3 + 2] annulation. Although these results seem to
support the proposed pathway involving cyclopropanolate
(16 (X ) SiMe3), Scheme 3), there remains the question
of why cyclopentenol 36 was obtained as a major product
irrespective of the stereochemistry of the vinylcyclopro-
panolates and the question of why (Z)-45 gives more
cyclopentenol 36 than does (E)-45 under comparable
conditions. To explain the stereochemistry of the process
(45 f 36, 37), we propose a mechanism in which
cyclopentenol 36 is produced via a [1,3]-sigmatropic shift
of the internally oxygen-silicon-coordinated anti-(Z)-46
(Scheme 9).26 The stereochemical course from anti-(Z)-
46 to the cyclopentenol 36 is in agreement with that
predicted by orbital symmetry considerations (supra-
facial-inversion). Anti-(Z)-46 may be generated from three
other diastereomers by a ring-opening, a geometric
isomerization, and a ring-closure sequence via 47 and/
or 48. The intermediate should undergo more facile
Ch a r t 1
quenched with acetic acid (2.2 equiv). The results are
shown in Table 4.23
When the reaction was performed in THF, in most
cases, cyclopentenols 36 and 37, which are exactly the
same as those obtained in the [3 + 2] annulation, and a
ring-opening product 35 were obtained.24 The fact that
the minor diasetereomer 37 was obtained only from (Z)-
45 is also consistent with the results of the [3 + 2]
annulation. Particularly noteworthy is the substantial
formation of cyclopentenol 36 from the (Z)-derivative
(25) For the effect of a substituent on the rate of thermal vinyl-
cyclopropane-cyclopentene rearrangement, see: Trost, B. M.; Scudder,
P. H. J . Org. Chem. 1981, 46, 506-509.
(26) For reviews of hypervalent silicon compounds, see: (a) ref 12a,
chapter 4. (b) Kost, D.; Kalikhman, I. Hypervalent Silicon Compounds.
In The Chemistry of Organic Silicon Compounds, Rappoport, Z.,
Apeloig, Y., Eds.; Wiley: Chichester, UK, 1998; Vol. 2, Chapter 23. (c)
Chuit, C.; Corriu, R. J . P.; Reye, C.; Young, J . C. Chem. Rev. 1993, 93,
1371-1448.
(23) Takeda, K.; Sakurama, K.; Yoshi, E. Tetrahedron Lett. 1997,
38, 3257-3260.
(24) Although we previously reported in a preliminary communica-
tion (ref 23) that cyclopentenol 36 was the only cyclized product in
the reaction of (Z)-45, reexamination of the original spectral data
revealed the formation of a trace amount of 37.

[3 + 2] Annulation of R,â-Unsaturated Acylsilanes
J . Org. Chem., Vol. 67, No. 6, 2002 1791
Ta ble 4.
yield (%)
in THF
in toluene
45
conditions
36
37
35 (E:Z)
36
35 (E:Z)
45
syn
syn
syn
anti
anti
anti
syn
syn
syn
-80 °C, 30 min
-80 to -50 °C
-80 to -30 °C
-80 °C, 30 min
-80 to -50 °C
-80 to -30 °C
-80 °C, 30 min
-80 to -50 °C
-80 to -30 °C
-80 °C, 30 min
-80 to -50 °C
-80 to -30 °C
0
45
54
0
45
63
52
61
76
59
72
76
0
0
0
0
0
0
6
7
8
6
9
9
81 (1.1:1)
40 (2.9:1)
31 (17:1)
89 (1.1:1)
39 (3.3:1)
34 (only E)
20 (1.7:1)
5 (3.9:1)
0
-
0
0
-
0
0
-
7
0
-
0
-
96
-
83
93
-
53
96
-
35
95
-
E
Z
13 (1.5:1)
0
-
44 (1.3:1)
0
-
54 (3.9:1)
1
-
16 (only E)
10 (1.7:1)
6 (2.8:1)
anti
anti
anti
14 (5.7:1)
30
25 (1.8:1)
45
Sch em e 9
Ch a r t 2
atom in Scheme 9, we carried out reactions of dimeth-
ylphenylsilyl derivatives 49 and 50 with MeLi, anticipat-
ing that, in both reactions, the syn isomers would afford
less 36 analogue relative to the corresponding anti
isomers because of internal chelation such as 51 and 52,
which can be more stabilized by the phenyl group than
the corresponding trimethylsilyl and tert-butyldimethyl-
silyl derivatives, respectively (Chart 2).27 Thus, if the
assumption that the formation of 36 from anti-(Z)-46 is
faster than that from other isomers is correct, the rate
from anti-(Z)-51 and from syn-52 would increase and
decrease, respectively, owing to increased concentrations
of anti-(Z)-51 and syn-52. Consequently, in both cases,
more 36 would be obtained from anti-(Z) derivatives
relative to that from syn-(Z) derivatives.
Although 4921 could be prepared from the correspond-
ing dienol silyl ethers in a manner similar to that
described for the preparation of 45, in the case of 50, the
cyclopropanation step failed. We prepared 5021 via de-
silylation of 45 followed by dimethyl(phenyl)silylation.
The results of the reactions of 49 and 50 with MeLi are
shown in Tables 5 and 6, respectively. Replacement of
the methyl group in the trimethylsilyl group with a
rearrangement to cyclopentenol than the other isomers
presumably because of its fixed conformation that is
suitable for the overlap of the orbitals required for the
rearrangement. The observations that 35 is produced to
a larger extent from the (E)-derivatives than from the
(Z)-derivatives and that the products ratio is unaffected
by the C-1 stereochemistry can be explained by assuming
that the syn/anti isomerization proceeds much faster
than does the E/Z isomerization. The assumption made
above that the origin of (E)-35 is 38 and/or 41 seems to
rationalize the result that (E)-46 produces more 35 than
(Z)-46 does.
If interconversion between the four diastereomeric 46
via ionic intermediates 47 and/or 48 exists, as we
assumed, the use of a less-polar solvent would cause a
decrease in the rate of isomerization and would conse-
quently result in decrease in the formation of 36 from
syn-(Z)-46 and (E)-46 relative to anti-(Z)-46. In fact, when
the reaction of 45 with MeLi was conducted in toluene,
syn-(Z)-45 afforded less cyclopentenol 36 than did anti-
(Z)-45, and reaction of the (E)-derivatives resulted in
recovery of the starting material (Table 4).
(27) Brook and coworkers reported the rate of C-to-O migration of
silyl group in the reaction of diphenyl(trimethylsilyl)carbinol with
diethylamine. Replacement of methyl on trimethylsilyl group by phenyl
accelerates the 1,2-migration (the Brook rearrangement), which is
believed to proceed via a pentavalent silicon intermediate, by a factor
of about 6. Brook, A. G.; LeGrow, G. E.; MacRae, D. M. Can. J . Chem.
1967, 45, 239-253.
Next, to try to obtain further evidence supporting the
assumption of a role of internal chelation of the silicon

1792 J . Org. Chem., Vol. 67, No. 6, 2002
Takeda et al.
Ta ble 5.
Sch em e 10
yield (%)
49
conditions
53
54
55 (E:Z)
syn
syn
syn
anti
anti
anti
syn
syn
syn
-80 °C, 5 min
-80 °C, 30 min
-80° to -30 °C
-80 °C, 5 min
-80 °C, 30 min
-80° to -30 °C
-80 °C, 5 min
-80 °C, 30 min
-80° to 30 °C
-80 °C, 5 min
-80 °C, 30 min
-80° to -30 °C
16
20
80
18
21
83
64
70
81
66
72
83
0
0
0
0
0
0
7
9
8
8
8
8
80 (1.1:1)
75 (1.6:1)
12 (only E)
77 (0.96:1)
74 (0.92:1)
8 (only E)
21 (1.4:1)
19 (2.2:1)
6 (only E)
16 (1.6:1)
18 (1.7:1)
4 (only E)
E
Z
anti
anti
anti
Ta ble 6.
tant role of the stabilization by chelation involving
pentacoordinate silicon species, as shown in Scheme 9.
Furthermore, our previous observation on the reaction
of 1,2-divinylcyclopropanolates 59 where the formation
of seven-membered ring via [3 + 4] annulation can
compete with the formation of five-membered ring via
[3 + 2] annulation provides additional support for the
chelation and the rapid interconversion between the syn
and anti isomers (Scheme 10). Thus, the five-membered
carbocycle 63 was obtained as a major product from anti-
(Z)-59, which can involve a five-coordinated silicon atom,
in contrast to the fact that no formation of detectable
amounts of 63 was observed in the reaction of anti-(E)-
59 where chelation cannot take place.
Next, we turned our attention to [3 + 2] annulation
using (â-(phenylthio)acryloyl)silanes 27. Although the
results of the low-temperature quenching experiments
described above showed that the independence of the
stereochemical outcome of [3 + 2] annulation using 27
to its olefin geometry is partly due to loss of the
stereochemical integrity of 27 during the reaction, this
does not tell us whether the annulation occurs via a
cyclopropanolate intermediate. We decided to examine
the reaction of the corresponding phenylthio derivatives
64. Cyclopropyl acetates 6421 were prepared in a manner
similar to 45. Unfortunately, syn- and anti-(E)-64 could
not be separated. The results of the reaction with MeLi
are shown in Table 7.
yield (%)
entry
50
conditions
56
57 (E:Z)
58
1
2
3
4
5
6
7
8
syn-(E)
syn-(E)
anti-(E)
anti-(E)
syn-(Z)
syn-(Z)
anti-(Z)
anti-(Z)
-80 °C, 30 min
-80 to -30 °C
-80 °C, 30 min
-80 to -30 °C
-80 °C, 30 min
-80 to -30 °C
-80 °C, 30 min
-80 to -30 °C
0
4
0
11
8
28
20
28
92 (1:5.8)
90 (1:2.5)
96 (1:1.6)
83 (1:1.5)
40 (2.1:1)
65 (5.3:1)
35 (2.6:1)
48 (only-E)
0
0
0
0
43
0
21
14
phenyl group resulted in a significant enhancement of
the reaction rate in both (E)- and (Z)-49. As expected,
there was little difference in the ratios of 53/55 between
syn and anti derivatives. This is thought to be due to the
delicate balance between enhanced reactivity and in-
creased stabilization of the internal chelation induced by
introduction of a phenyl group.
In contrast to the enhanced reactivity by introduction
of a phenyl group in the vinylsilyl group, the introduction
of a dimethyl(phenyl)siloxy group in place of the tert-
butyldimethylsiloxy group resulted in a decrease in the
rate of the reaction. Particularly noteworthy is that a
more predominant formation of cyclopentenol 56 from
anti-50 than from syn-50 was observed at -80 °C as
expected (Table 6, entries 2, 4, 5, and 7). Moreover, syn-
cyclopropandiol derivative 58 was obtained from both syn
and anti derivatives 50, suggesting facile interconversion
between syn- and anti-cyclopropanolates and an impor-
The remarkable difference between the [3 + 2] annu-
lation using 27 and the reaction of 64 with MeLi was
the product distribution depending on the olefin geometry
of the starting materials, in addition to the enhanced
reactivity in the latter case. Thus, the reaction of (Z)-64
provided two diastereomers 31 and 32 in about a 5:4
ratio, while the reaction of (E)-64 almost exclusively
provided 31. Also, since the change of the vinyl substitu-
ent in the vinylpropanolates from trimethylsilyl to di-

[3 + 2] Annulation of R,â-Unsaturated Acylsilanes
J . Org. Chem., Vol. 67, No. 6, 2002 1793
Sch em e 11
Ta ble 7.
Con clu sion s
The results described above seem to indicate a need
for a slight modification of the initially proposed mecha-
nistic pathway for the [3 + 2] annulation shown in
Scheme 3. (E)- and (Z)-27 were found in equilibrium
under the reaction conditions (Scheme 6 and Table 2).
Although the possibility that either of (E)- or (Z)-27 is
the only precursor leading to cyclopentenols 3 and 4
cannot be entirely ruled out, the existence of the equi-
librium does not necessarily mean that the delocalized
allylic carbanion 15 is not involved as a common inter-
mediate. The intermediacy of 15 is consistent with the
failure in [3 + 4] annulation of 27, which requires the
formation of a cyclopropanolate intermediate. Conse-
quently, we believe that delocalized allylic anion 15 is
an intermediate in the major reaction pathway in the [3
+ 2] annulation of 27 (Scheme 11). The reason for the
preferential formation of 3, which seems to be thermo-
dynamically less stable than 4 due to steric repulsion
between the phenylthio group and alkyl groups, is
unclear at present because of potential complications
associated with the structures of sulfur-stabilized allylic
carbanions, their relative reactivities, and the relative
conformation of the carbonyl group.
In the case of the trimethylsilyl derivatives 10a , we
propose a reaction pathway for the formation of 11 that
involves the initial formation of vinylcyclopropanolate
intermediates syn-(E)- and syn-(Z)-67 followed by the
isomerization to anti-(Z)-67 and rearrangement to 11 via
a 1,3-sigmatropic shift, although the mechanistic origin
of the minor product 12 is still not clear.
Although a full understanding of the actual mechanism
of the [3 + 2] annulation must await further detailed
mechanistic experimentation, our current data have
provided significant insights into the reaction sequence
of the annulation.
yield (%)
64
conditions
31
32
anti-E
-80 °C, 30 min
-80° to -30 °C
-80 °C, 30 min
-80 to -30 °C
-80 °C, 30 min
-80 to -30 °C
-80 °C, 30 min
-80 to -30 °C
84
85
87
87
50
48
48
48
2
3
1
anti-E:syn-E (1:1.1)
anti-Z
2
39
39
35
40
syn-Z
methyl(phenyl)silyl and to phenylthio groups increased
the reactivity of the rearrangement, it appears to be a
general trend that the rate of the rearrangement in-
creases, when the R-carbanion-stabilizing character of the
substituent on the vinyl group increases. The enhanced
reactivity is attributed to acceleration of the ring-opening
of cyclopropane caused by the generation of a carbanion
character at the R-position of the â-heteroatom substitu-
ent.
Although the origin of the observed stereoselectivity
is not clear at present, the results rule out the possibility
of the intermediacy of cyclopropanolate 16 (X ) SPh) in
the major reaction pathway of [3 + 2] annulation of 27.

1794 J . Org. Chem., Vol. 67, No. 6, 2002
Takeda et al.
described in the text. This material is available free of charge
via the Internet at http://pubs.acs.org.
Ack n ow led gm en t. This research was partly sup-
ported by a grant from the Takeda Science Foundation.
Su p p or tin g In for m a tion Ava ila ble: Full experimental
detail and characterization data for all new compounds
J O0160219