catalyzed π-allyl hydrogenolysis3 led to complex product
mixtures (eq 1). Although reaction of allylic acetate 10 did
generate desired cyclohexene 11, byproducts 12 and 13 were
also formed, and 11 could not be isolated by conventional
purification techniques (eq 2).
Table 1. Reductive Isomerization Reactiona
While attempting to solve these problems, we serendipi-
tously discovered a highly selective reductive isomerization
reaction that can be executed under extremely mild and
neutral conditions. Namely, exposure of lactone 9 to
traditional heterogeneous catalytic hydrogenation conditions
(Pd/C, 1 atm H2) afforded key fragment 14 en route to (+)-
dragmacidin F (8) in nearly quantitative yield (eq 3).4 This
smooth conversion of lactone 9 to carboxylic acid 14 stood
in stark contrast to our initial studies using standard π-al-
lylpalladium hydrogenolysis methods to attempt this trans-
formation (10 f 11 + 12 + 13). On the basis of these results,
we set out to examine the unusual reactivity of the
heterogeneous palladium system. Herein, we detail our
exploration of the reductive isomerization reaction, an
expanded substrate scope, and a mechanistic investigation
of its regiochemical fidelity.5
Intrigued by our initial result (9 f 14), we prepared a
number of substrates to assess the generality of this reaction
(Table 1).6 As a starting point, a simple variant of lactone 9
bearing an acetate on the secondary allylic alcohol was
synthesized (i.e., 15). We were pleased to see that 15 could
be converted to carboxylic acid 16 in good yield (entry 1).
The use of allylic acetate 10 as a substrate (entry 2), on the
other hand, led to an unexpected result. We anticipated that
a Standard conditions: H2 (balloon, 1 atm), 10% Pd/C (2 mol % Pd),
MeOH, 0 °C. b Isolated yield. c Yield based on 1H NMR integration.
d Performed with 10% Pd/C (0.5 mol % Pd). e Performed with 10% Pd/C
(1 mol % Pd). f Reaction performed at 23 °C. g Product formed in 7.2%
ee.
methyl ester 11 would be the observed product because
acetate is a superior leaving group to silanolate (cf. eq 2).
However, the compound obtained (i.e., 17) resulted from a
net loss of the OTBS group.7 Notably, none of the byproducts
formed under homogeneous π-allyl protocols were observed
under these heterogeneous conditions (eq 2, 11-13). To
probe this result further, a version of 9 with exchanged
protecting groups on the secondary alcohols was prepared
(18, entry 3). In this case, elimination of acetate occurred.7
Due to the success of the rigid bicyclic lactone framework
in this reaction, we reasoned that the reactivity of 10 and 18
might be improved by restricting them as bicyclic carbonates
(20 and 21). These carbonate-containing substrates were well
tolerated and led to competent production of the correspond-
(3) For a review, see: Tsuji, J.; Mandai, T. Synthesis 1996, 1-24.
(4) 10% Pd/C was purchased from Aldrich (20,569-9). This has been
demonstrated to be a safe and nearly neutral hydrogenation catalyst; see:
(a) Sajiki, H.; Ikawa, T.; Hirota, K. Tetrahedron Lett. 2003, 44, 7407-
7410. (b) Ikawa, T.; Sajiki, H.; Hirota, K. Tetrahedron 2004, 60, 6189-
6195.
(5) Isolated examples of similar reactivity using Pd/C, H2, and a protic
solvent have been reported in the literature; see: (a) Paulson, D. R.; Gilliam,
L. S.; Terry, V. O.; Farr, S. M.; Parker, E. J.; Tang, F. Y. N.; Ullman, R.;
Ribar, G. J. Org. Chem. 1978, 43, 1783-1787. (b) Dauben, W. G.; Hance,
P. D. J. Am. Chem. Soc. 1955, 77, 2451-2453. (c) Dauben, W. G.; Hayes,
W. K.; Schwarz, J. S. P.; McFarland, J. W. J. Am. Chem. Soc. 1960, 82,
2232-2238.
(6) With the exception of 31 (entry 10), all of the compounds in Table
1 are enantiopure.
(7) The major product in this reaction resulted from direct hydrogenation
of the olefin moiety and was formed as a mixture of diastereomers.
2514
Org. Lett., Vol. 7, No. 12, 2005
Caspi, Daniel D.
