Versatile diastereoselectivity in formal [3,3]-sigmatropic shifts of substituted 1-alkenyl-3-alkylidenecyclobutanols and their silyl ethers

Article abstract of DOI:10.1021/ja051663p

A new method for the preparation of highly substituted cyclohexenones is reported. [2 + 2] Cycloaddition of 2-silyloxydienes with allenecarboxylate affords the 1-alkenyl-3-alkylidenecyclobutanol silyl ethers. Thermolysis of these compounds affords the met

Full text of DOI:10.1021/ja051663p

Published on Web 07/21/2005
Versatile Diastereoselectivity in Formal [3,3]-Sigmatropic Shifts of Substituted
1-Alkenyl-3-alkylidenecyclobutanols and Their Silyl Ethers
Michael E. Jung,* Nobuko Nishimura, and Aaron R. Novack
Department of Chemistry and Biochemistry, UniVersity of California, Los Angeles, California 90095-1569
Received March 15, 2005; E-mail: jung@chem.ucla.edu
Scheme 1
We recently reported the formal [3,3]-sigmatropic shift of
1-alkenyl-3-alkylidenecyclobutyl silyl ethers to give various
2-methylenecyclohex-4-enecarboxylates and the application of this
methodology to the total synthesis of the antitumor agent dysidiolide
in its correct enantiomeric form.1 We proposed a mechanism to
account for the unusual exo selectivity of this thermal process. We
now report the rearrangement of more substituted systems to include
unsymmetric 2,2-dialkylethenyl derivatives, which have caused us
to revise this original mechanism. In addition, we report herein the
base-catalyzed rearrangement of the corresponding cyclobutanols,
which proceeds at very low temperatures and affords the opposite
diastereoselectivity. This work now permits the formation of very
hindered substituted cyclohexene systems.
Heating the cyclobutyl silyl ether 3, prepared by the [2 + 2]
cycloaddition of the silyloxybutadiene 1 and the 2-methylallene-
carboxylate 2, gave only the exo cyclohexene 4x (Scheme 1). We
believe that this rearrangement occurs via a diradical mechanism,
as proposed by Dolbier,2 in which bond a of the cyclobutane is
Scheme 2
selectively cleaved over bond b since the electron-deficient meth-
ylene radical in I would be expected to be more stable cis to the
electron-donating methyl group rather than cis to the electron-
withdrawing ester. This seems to be due to two factors: a repulsive
steric interaction between the methyl and the sp3 carbon in II and
a slight attractive interaction of the carbonyl oxygen and protons
on the sp3 methylene group in I. Rotation and electronic reorganiza-
tion, as shown in III, would give the observed exo product 4x.3
Because it is often extremely difficult to prepare very hindered
cyclohexene systems by cycloadditions, for example, compounds
with contiguous quaternary centers, we studied the rearrangement
Scheme 3
of the 2,2-dialkylethenyl systems. Thus, cycloaddition of the
silyloxydiene 5 prepared from mesityl oxide and the allenecar-
boxylate 2 gave the cyclobutane 6, which could be cleanly
rearranged to the hindered cyclohexene 7 on heating in a sealed
tube at 135 °C for 48 days or by flash vacuum pyrolysis4 or
microwave heating5 (Scheme 2). We also prepared the unsym-
metrical 2,2-disubstituted alkenyl systems, namely, the three aryl
alkenes, 9a-c, via cycloaddition of the analogous E-arylbutadienes
8 and the allenecarboxylate 2. Heating of these cyclobutanes gave
the desired cyclohexenes but with poor diastereoselectivity (roughly
2:1) (Scheme 3).
Attempts to promote the rearrangement using TBAF to prepare
the alkoxide in situ failed due to elimination of the tertiary allylic
silyl ether by the basic fluoride ion.6 To try to improve the
diastereoselectivity of this process, we decided to prepare the
trimethylsilyl ether rather than the corresponding tert-butyldimeth-
ylsilyl ether. Thus, the TMS enol ether 11 was prepared from
mesityl oxide and reacted with the allenecarboxylate 2 to give the
desired silyl ether 12 in 58% yield (Scheme 4). Hydrolysis of the
silyl ether in acidic ethanol furnished the cyclobutanol 13 in 45%
yield. After several failed attempts to rearrange the alcohol with
sodium bases (with and without metal additives), we found that
treatment of 13 with methyllithium at -78 °C gave the desired
rearrangement product 14 in 33% yield along with 10% of recovered
13 and some of the tertiary alcohol from addition of methyllithium
to the ester. Thus, as hoped, the anion of the alcohol rearranged at
low temperatures to give the enone ester.
Now that we had developed a method for the rearrangement of
these systems at low temperature, we decided to reexamine the
unsymmetrical 2,2-disubstituted alkenyl systems. The diene 15
(prepared from the E-enone) reacted with 2 to give the silyl ether
16, which was hydrolyzed to give the alcohol 17 in 70% yield.
Treatment of this alcohol with lithium bases afforded the cyclo-
9
11206
J. AM. CHEM. SOC. 2005, 127, 11206-11207
10.1021/ja051663p CCC: $30.25 © 2005 American Chemical Society

C O M M U N I C A T I O N S
Scheme 4
Scheme 7
Scheme 5
diastereomer 22n predominated. Thus, either the endo or the exo
diastereomer can be prepared as the major product of the rear-
rangement of 20. It is interesting to note that when HMPA is added
to the solution to help dissociate the lithium ion, the selectivity
decreases to only 1.3:1. Therefore, a tight metal alkoxide bond is
required for high stereoselectivity, presumably because it is
necessary for the selective cleavage of bond b in Scheme 6.
In summary, we have developed a route to very hindered
cyclohexene systems via a thermal [3,3]-sigmatropic rearrangement
of cyclobutyl silyl ethers and the base-promoted cyclobutanol
openingsintramolecular Michael cyclization to give either diaste-
reomer of the desired product. Further work on the application of
these rearrangements in synthesis is underway and will be reported
in due course.
Scheme 6
Acknowledgment. We thank the National Science Foundation
(CHE0314591) for generous support of this work.
Supporting Information Available: Experimental procedures and
proton and carbon NMR data for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
hexenone products with a great preference for the endo isomer,
for example, LiHMDS afforded a 67% yield of 18n with only a
trace of 18x (23:1).7 Thus, we have developed a method for the
low temperature conversion of such cyclobutanols into cyclohex-
enones but with a complete inversion of the stereochemical
preference.
(1) (a) Jung, M. E.; Nishimura, N. J. Am. Chem. Soc. 1999, 121, 3529. (b)
Jung, M. E.; Nishimura, N. Org. Lett. 2001, 3, 2113.
(2) (a) Dolbier, W. R., Jr.; Mancini, G. J. Tetrahedron Lett. 1975, 2141. (b)
Dolbier, W. R., Jr.; Piedrahita, C. A.; Al-Sader, B. H. Tetrahedron Lett.
1979, 2957. (c) Spitzner, D. Angew. Chem. 1982, 94, 639. (d) Kozhushkov,
S. I.; Kuznetsova, T. S.; Zefirov, N. S. Dokl. Akad. Nauk SSSR 1988,
299, 1395-1399. (e) Spitzner, D.; Klein, I. Liebigs Ann. Chem. 1990,
63.
(3) These conclusions are drawn from an extensive theoretical study of the
mechanism of this reaction, which is currently underway. Houk, K. N.;
Zhao, Y. L. Unpublished results. The mechanism proposed here is different
than the one proposed earlier by us for this process.1a
(4) For a review of flash vacuum pyrolysis, see: Trahanovsky, W. S.; Lee,
S. K. Synthesis 1996, 1085-1086 and references therein.
(5) The use of various solvents for this process showed that moderately polar
solvents, such as chlorobenzene, 1,2-dichlorobenzene, THF, etc., acceler-
ated the rearrangement by a factor of about 4-5, whereas very polar
solvents, such as DMF and DMSO, gave the two triene products
(approximately 1:1 mixture) via solvolysis of the tertiary silyl ether.
(6) This produced the same approximately 1:1 mixture of triene esters as did
heating in DMSO or DMF.
Presumably, the selective formation of the endo isomer 18n from
17 proceeds via the mechanism outlined in Scheme 6. The metal
alkoxide IV formed by deprotonation of the alcohol 17 would open
bond b cis to the ester group in great preference to a since Weiler
and Harris have shown that the cis anion is formed in complete
preference to the trans in similar systems.8 Rotation in V followed
by an intramolecular Michael addition as shown would then afford
18n. Thus, the reaction proceeds via two distinct mechanisms
depending on conditions.
(7) Although hexa-1,5-dien-3-olates are well-known to rearrange to enones
(e.g., anionic oxy-Cope), there are no examples of this process with an
3-methylene-1-alkenylcyclobutan-1-ol system such as these.
(8) (a) Harris, F. L.; Weiler, L. Tetrahedron Lett. 1985, 26, 1939. (b) Harris,
F. L.; Weiler, L. Tetrahedron Lett. 1984, 25, 1333. Deprotonation of
stereospecifically deuterated 3-methylcrotonate was carried out with LDA
in the presence of HMPA, so that it is not clear whether an association of
the lithium base with the carbonyl was the determining factor in the
stereochemical course of the reaction (syn deprotonation).
(9) The yield of the [2 + 2] product 20 is somewhat low because the aryl
group in monoaryl substituted systems accelerates the rearrangement to
give the exo products (e.g., 21x) during the heating required to prepare
the cyclobutane. This is not the case for monoalkyl substituted systems
(e.g., 3), where the yield is generally 80% or so.
The consequences of this mechanistic dichotomy imply that either
diastereomer can be made from the same substrate by choice of
conditions. We have now shown that to be the case (Scheme 7). [2
+ 2] Cycloaddition of the diene 19 with 2 gave the cyclobutyl
silyl ether 20 in 35% yield (along with some of the [4 + 2] product
21x).9 Thermolysis of 20 in toluene afforded only the exo product
21x in 90% yield, which was desilylated to give the exo cyclo-
hexenone 22x as the sole product. To access the opposite diaste-
reomer, the silyl ether 20 was hydrolyzed in acid to the alcohol 23
in quantitative yield. Treatment of 23 with LHMDS in THF at -78
°C afforded a 73% yield of a 5:1 mixture in which the endo
JA051663P
9
J. AM. CHEM. SOC. VOL. 127, NO. 32, 2005 11207