hexanol without moving the exomethylene unit into the
more stable endocyclic position and installation of the
3-furyl ketone. We used the same technique that had
worked in our earlier synthesis of hedychilactone B;3
namely, reduction of the ester of 10x with DIBAL afforded
in 88% yield the alcohol 11, which was oxidized under
Dess-Martin periodinane conditions to afford the aldehyde
7 in 85% yield (Scheme 3). Formation of the ylide 13
Scheme 1
Scheme 3
ane-1,3-dione 6 via the known vinyl iodide 7 which has
been used by Danishefsky in a synthesis of taxol deriva-
tives5 (Scheme 2). The vinyllithium prepared from 7 was
Scheme 2
from methoxymethyltriphenylphosphonium iodide,7 pre-
pared in two steps from methylal via reaction with
trimethylsilyl iodide (TMSI) and triphenylphosphine, by
reaction with n-butyllithium and addition of the aldehyde
gave a mixture of stereoisomers of the enol methyl ether
14 in 51% yield. The ketal functionality survived all of
these transformations without any problems.
The key step of hydrolysis of the silyl enol ether without
concomitant conjugation of the exocyclic methylene unit
was effected by the use of conditions shown to work in
our earlier system. Thus careful treatment of 14 with HF-
pyridine in acetonitrile at low temperature afforded a 92%
yield of the desired trans-decalone 15 in which the
exocyclic methylene remained untouched (Scheme 4). As
expected, reduction of the decalone with DIBAL occurred
from the less-hindered equatorial direction to give the
desired axial alcohol 16 in quantitative yield. Very mild
acidic hydrolysis of the methyl enol ether (aq. HCl in
THF) afforded the keto aldehyde 17 in 84% yield.
Attachment of the 3-furyl ketone unit to the side chain
was all that was required to complete the synthesis. The
3-furyllithium species,8 prepared from commercially
available 3-bromofuran, was added to the keto aldehyde
17 to afford chemoselectively the aldehyde adduct in fair
yield. However, attempted selective oxidation of this
benzylic alcohol using manganese dioxide proceeded
rather poorly, and only very small amounts of keller-
manoldione 1 could be obtained. Therefore we decided
to examine a different sequence for the conversion of the
key aldehyde 12 into the desired product 1.
trapped with acetaldehyde and oxidized, and the silyl enol
ether was prepared by the standard route to give the
desired diene 8. Heating a neat mixture of the diene 8
with the allene carboxylate 3 (prepared in 64% yield by
reaction of acetyl chloride with ethoxycarbonylmethyl-
enephosphorane)6 at 110 °C for 14 days gave a separable
mixture of three products, the [2 + 2] cycloadduct 9, the
desired exo [4 + 2] cycloadduct 10x, and the endo adduct
10n in 7.2%, 26.5%, and 13.7% yield, respectively, with
32.4% of the recovered starting diene 8. The cyclobutane
9 could be converted into the same 2:1 ratio of 10x and
10n, resulting in a 39.2% overall yield of 10x based on
recovered starting material.
The conversion of the adduct 10x into kellermanoldione
1 required two key steps, formation of the axial cyclo-
(7) Jung, M. E.; Mazurek, M. A.; Lim, R. M. Synthesis 1978, 588–90.
(8) (a) Collins, M. A.; Tanis, S. P. “3-Lithiofuran,” e-EROS Encyclo-
pedia of Reagents for Organic Synthesis ( 2001). (b) Chen, C.-L.; Sparks,
S. M.; Martin, S. F. J. Am. Chem. Soc. 2006, 128, 13696.
(5) Di Grandi, M. J.; Jung, D. K.; Krol, W. J.; Danishefsky, S. J. J.
Org. Chem. 1993, 58, 4989–92.
(6) Paik, Y. H.; Dowd, P. J. Org. Chem. 1986, 51, 2910.
Org. Lett., Vol. 11, No. 17, 2009
3883