Diastereoselectivity in the Carroll rearrangement of β-keto esters of tertiary allylic alcohols

Article abstract of DOI:10.1016/j.tetlet.2003.10.102

Carroll rearrangement of β-keto esters derived from tertiary allylic alcohols, for example, 7, under basic conditions followed by decarboxylation of the resulting β-keto acids yielded the expected γ, δ-unsaturated methyl ketones 8 with a range of olefin geometries from 100:0 to 1:1.8 E/Z, depending on the relative steric requirements of the two groups at the allylic center.

Full text of DOI:10.1016/j.tetlet.2003.10.102

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 107–109
Diastereoselectivity in the Carroll rearrangement of b-keto esters
of tertiary allylic alcohols
q
*
Michael E. Jung and Brian A. Duclos
Department of Chemistry and Biochemistry, University of California, Los Angeles, 405 Hilgard Ave, Los Angeles,
CA 90095-1569, USA
Received 26 September 2003; revised 13 October 2003; accepted 20 October 2003
Abstract—Carroll rearrangement of b-keto esters derived from tertiary allylic alcohols, for example, 7, under basic conditions
followed by decarboxylation of the resulting b-keto acids yielded the expected c,d-unsaturated methyl ketones 8 with a range of
olefin geometries from 100:0 to 1:1.8 E=Z, depending on the relative steric requirements of the two groups at the allylic center.
Ó 2003Elsevier Ltd. All rights reserved.
For a proposed synthesis of several biologically active
polycyclic triterpene natural products, we investigated
the use of the Carroll rearrangement in order to achieve
the desired olefin geometry in the 7-substituted
hoped for a similar selectivity in the rearrangement of
2 to give 1.
1
We report herein the results of the Carroll rearrange-
ment of b-keto esters of a series of tertiary allylic alco-
hols with alkyl groups of differing sizes. Although the
yields of the rearrangement process are low, we have
demonstrated the ability to access both geometries with
varying functionality on the olefins.
6
-methyl-5-hepten-7-one subunit (Scheme 1). In particu-
lar, we hoped that the steric difference between a silyl-
oxymethyl and a methyl group in an intermediate such
as 4 would afford predominantly the E-olefin geometry
in the Claisen rearrangement product 3a. Likewise, we
Preparation of the substrates 7a–e for the Carroll rear-
2
rangement was straightforward (Scheme 2). The two
silyloxy derivatives 5ab were obtained in good yield
from reaction of either TBSOTf or TIPSOTf with
commercially available acetol and triethylamine or
pyridine. Subsequent treatment with vinylmagnesium
Me
Me
O
Me
O
Me
Me
Me
TBSO
TBSO
1
2
Me
Me
O
MgBr
or
MgBr
O
Me
O
OH
R
O
O
O
Me
Me
R
O
Me
TBSO
Me
DMAP
Et O/23 ˚C
5
a-e
TBSO
TBSO
5a
THF
2
3a
4
6
a-e
O
R'
Scheme 1.
Me
LDA, THF
78 ˚C to 23 ˚C;
O
R
Me
-
O
CCl /²
4
Me
Me
Keywords: Carroll rearrangement; [3,3]-sigmatropic rearrangement;
steric effects; electronic effects.
R
q
Supplementary data associated with this article can be found at
doi:10.1016/j.tetlet.2003.10.102
7a-e
8
3
a-e R' = CO H
a-e R' = H
2
*
Corresponding author. Tel.: +1-3108257954; fax: +1-3102063722;
e-mail: jung@chem.ucla.edu
Scheme 2.
0
040-4039/$ - see front matter Ó 2003Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.10.102

108
M. E. Jung, B. A. Duclos / Tetrahedron Letters 45 (2004) 107–109
Table 1. Formation and Carroll rearrangement of b-keto esters 7a–e
R
Yield (%)
Yield (%)
Yield (%)
E=Z ratio
CH
CH
2
OTBS
OTIPS
6a
6b
6c
6d
6e
70
100
99
7a
7b
7c
7d
7e
88
67
3a
3b
36
531.4:1
86
2.1:1
2
a
CH(OMe)
2
75
3c
3d
3e
100:0
3:1
b
c
CH@CMe
CBCH
2
100
93
100
68
38
131:1.8
a
b
c
Acidification of the aqueous layer to retrieve acid 8c, even to a pH of 5, caused the dimethyl acetal to be completely converted to the corresponding
aldehyde.
The reported yield is for the inseparable mixture of products where both the dimethylvinyl and vinyl substituents participated in the rearrangement.
See text.
See text for discussion.
bromide in THF provided 6ab in 70% and 100% yields,
respectively. The dimethyl acetal derivative 6c was
obtained from reaction of vinylmagnesium bromide and
pyruvaldehyde dimethoxy acetal 5c in THF in 99%
yield. The dimethylvinyl derivative 6d was obtained
from reaction of vinylmagnesium bromide with mesityl
oxide 5d in THF in quantitative crude yield. Finally, the
enyne substrate 6e was obtained from reaction of methyl
vinyl ketone with ethynylmagnesium bromide in THF in
good yield.
H
H
O
O
O
O
O
Me
O
Me
Me
Me
Me
Me
Me
A
B
Me
Me Me
Me
O
Me Me O
All of the tertiary alcohols 6a–e were treated with
diketene and catalytic DMAP in ether to afford the
acetoacetates 7a–e in good to excellent yields (Table 1).
Treatment of the substrates 7a–e with LDA in THF at
Me
Me
Me
Me
3dE (5.7)
Me
3d'E (3.0)
)
78 °C with subsequent warming to 23 °C furnished the
Me
Carroll rearrangement products, the b-keto acids 8a–e,
in good crude yields. These were generally not purified
O
Me Me O
Me
3dZ (1.9)
Me Me
3d'Z (1.0)
4
but immediately refluxed in CCl to effect decarboxyl-
ation and give the ketones 3a–e in poor to good yield.
The dimethyl acetal of 8c was hydrolyzed during the
purification to give the aldehyde 3c, which was some-
what unstable and therefore was reduced to the primary
alcohol 9c for further analysis. The acid sensitivity of
Scheme 3.
3
c–e made column chromatography partially responsi-
transition state, for example, C to give E, in order to
avoid the pseudo 1,3-diaxial interactions that would
result from an axial arrangement as in D, which would
give F (Scheme 4). We see the size of the steric differ-
ences of the diaxial interactions reflected in the olefin
ratios, with the most sterically demanding R group, the
dimethoxymethyl group 7c, having the highest E=Z
selectivity (essentially only one isomer corresponding to
ble for the low yields, even on silica gel columns de-
activated with triethylamine. The olefin geometries of
3a–e were determined by assignment of the cross peaks
in the NOESY spectra.
The rearrangement of the bis-vinyl substrate 7d pro-
duced intriguing results, with a mixture of four com-
pounds being isolated (Scheme 3). Careful proton NMR
analysis with significant use of difference NOE allowed
us to assign the mixture as shown below, namely:
0
0
3
dE:3d E:3dZ:3d Z 5.7:3.0:1.9:1.0. Thus, the rearrange-
H
H
ment takes place predominantly on the unsubstituted
vinyl group via A, rather than via B, presumably due to
steric hindrance, although the selectivity (roughly 2:1) is
less than might have been anticipated. The E=Z ratio
was 3:1 in each case.
H
H
O
Me
O
Me
R
Me
vs
O
O
O
O
O
Me
R
Li
Li
C
D
The stereochemistry resulting from the cyclizations was
in accordance with the anticipated steric differences in
all but one case. For the substrates 7a–c, an E alkene
was expected due to the size difference between a methyl
substituent and the R group. Bulky R groups would
exhibit a preference for the equatorial position in the
Me
R
O
R
Me
Me
Me
E
F
Scheme 4.

M. E. Jung, B. A. Duclos / Tetrahedron Letters 45 (2004) 107–109
109
H
H
the pentadienyl cations being partially formed in the
transition state outweighs the very small difference in the
steric size of the two groups and determines the stereo-
chemistry of the product.
H
H
O
Me
Me
O
Me
Me
Me
vs
O
O
O
O
Me
Me
Me
Me
Me
Li
H
Li
H
G
H
Me
In summary, we have shown that the substrates 7a–e can
be tuned to provide varying olefin geometries according
_
_
H
H
5
O
O
Me
to the steric demand of the substituents. By changing
Me
+
Me
+
the size of the group at the olefinic carbon, one can
change the selectivity from modestly Z-selective (1.8:1)
to completely E-selective (100:1).
vs
O
O
O
O
Me
Li
Li
I
J
W
Me
S
Scheme 5.
Acknowledgements
E was formed). The substrate with the smallest R group,
the ethynyl substrate 7e, afforded the expected product
We thank the NIH and the University of California
Universitywide AIDS Research Program (UARP-ID02-
LA-007) for generous financial support.
3
eZ, namely the Z-alkene corresponding to F, since the
methyl substituent is now the larger one and prefers the
equatorial position, for example D. The only somewhat
surprising case is the olefin 7d where both vinylic units
prefer the equatorial position vs. the sterically slightly
References and Notes
larger methyl group (A values: methyl, 1.74; vinyl, 1.49–
3
1
.68). The small difference in A values does not corre-
1. (a) Carroll, M. F. J. Chem. Soc. 1940, 704–706; (b) Carroll,
M. F. J. Chem. Soc. 1941, 507–511; (c) Kimel, W.; Cope,
A. C. J. Am. Chem. Soc. 1943, 65, 1992–1998.
spond well with the 3:1 ratio of E to Z products
observed, which is in the opposite direction for a purely
steric influence. Therefore we believe that an electronic
effect must be responsible and suggest the following. If
the bond breaking is fairly advanced in the transition
state (which may be reasonable for a trimethyl-substi-
tuted pentadienyl acetoacetate), then the relative
stability of the conformations of the nascent pentadienyl
cations would play a role. Thus for the transition states
leading to the products 3d, G and H, breaking the C–O
bond to generate the enolate and the pentadienyl cation
would be more favorable for G generating I since that
forms the pentadienyl cation in its more stable W con-
formation whereas breaking the same bond in H to
generate J would be less favorable since now the pen-
tadienyl cation is formed in its less stable S conforma-
tion (Scheme 5). Streitweiser has calculated that the
difference in energy favors the W conformation over the
S conformation by about 3.7 kcal/mol with the U con-
2
.
Wilson, S. R.; Augelli, C. E. Org. Synth. 1990, 68, 210–
19.
2
3
.
(a) Eliel, E. L.; Wilen, S. H. Stereochemistry of Organic
Compounds; Wiley: New York, 1994, pp. 695–698; (b)
Eliel, E. L.; Manoharan, M. J. Org. Chem. 1981, 46, 1959–
1
962.
Pratt, L. M.; Streitweiser, A. J. Org. Chem. 2000, 65, 290–
94.
4
.
2
5. The methyl ester derivative of 7 (R ¼ CO Me) was
2
also prepared, but when the b-keto ester derived from
it was subjected to treatment with LDA the corre-
sponding tetronic acid was formed instead of the
desired a; b-unsaturated ester. For a similar result, see
6
Takaiwa et al. Similarly, the dithiane derivative of 7
] did not provide the desired b-keto
CH ], perhaps due to complica-
[
R ¼ CH(SCH
2
)
2
CH
2
acid, 8 [R ¼ CH(SCH
2
)
2
2
tions due to the presence of the relatively acidic dithiane
proton.
6
.
(a) Takaiwa, A.; Yamashita, K. Agric. Biol. Chem. 1982,
46, 1721–1722; (b) Takaiwa, A.; Yamashita, K. Agric. Biol.
Chem. 1984, 48, 2061–2065.
4
formation being quite high in energy. Thus we suggest
that the energy difference of the two conformations of