11 can be conducted with high stereocontrol.11 Treatment
of triol 14 with trifluoroacetic acid in CH2Cl2 served to
accomplish deprotection and concomitant cyclization to
provide spirocyclic lactone 15.
Scheme 3a
At this stage, it appeared that model system 6 was easily
within our grasp by (1) oxidation of the secondary alcohol
in 15 to the ketone level (cf. 16), (2) removal of the benzyl
ether with spontaneous cyclization to form a hemiacetal, and
(3) acid-catalyzed methanolysis to provide 6. Unfortunately,
all attempts to oxidize diol 15 to ketone 16 were fruitless.
Compound 15 was amazingly unreactive to many protocols,
including Dess-Martin, Jones, PCC, TPAP-NMO, TEMPO,
and chromyl chloride. For instance, heating 15 at 150 °C in
DMSO for 24 h in the presence of 5 equiv of IBX resulted
in quantitative recovery of starting material. In contrast,
Swern oxidation conditions led to destruction of the ring
system. Prior protection of the tertiary alcohol in 15 also
led to a substrate that was completely resistant to oxidation.
Similarly, the product arising from deprotection of the benzyl
ether followed by protection of the 1,2-diol as an acetonide
(cf. 17) was totally resistant to oxidation as well.
a (a) OsO4, NMO, acetone/H2O (71%); (b) 90% TFA/H2O,
CH2Cl2 (83%); (c) BBr3, CH2Cl2, -78 °C (85%); (d) AlCl3, Et2O/
acetone, 0 °C (79%). Abbreviations: NMO N-methylmorpholine
N-oxide; TFA ) trifluoroacetic acid.
found to be a single diastereomer, subsequently shown to
be 14 (Vide infra).
While 17 was a noncompetent substrate for oxidation to
18, it is a crystalline compound, mp ) 228 °C. This
crystallinity was used to advantage in that an X-ray structure
of 17 served to corroborate its structure, and thus the
structures of 14 and 15 (Figure 2).12 Examination of the
crystallographically derived structure of 17 reveals that the
crucial methine proton (see asterisk in 17), whose abstraction
is required for oxidation, is in a hindered environment.
Efforts to reach the desired ketone level of oxidation at the
stage of triol 14 were also unsuccessful.13
The high stereoselectivity of the osmylation, providing the
triol with a syn arrangement between the “directing” second-
ary alcohol moiety and the newly installed 1,2-diol unit,
deserves some comment. Although it is generally believed
that osmylations can be directed by certain proximal
functional groups,6 seminal work by Kishi in acyclic systems
demonstrated that OsO4 is not “directed” per se by alcohols
but in fact approaches from the face of the alkene opposite
to that of the hydroxyl group.7,8 Although there appears to
be scant literature on the dihydroxylation of tetrasubstituted
allylic alcohols, the examples we identified all proceed with
high stereochemical fidelity on the face opposite to that of
the hydroxyl group.9
(6) For reports on the directing abilities of specific functional groups,
see: (a) Sulfoxide: Hauser, F. M.; Ellenberger, S. R.; Clardy, J. C.; Bass,
L. S. J. Am. Chem. Soc. 1984, 106, 2458. (b) Sulfoximine: Johnson, C. R.;
Barbachyn, M. R. J. Am. Chem. Soc. 1984, 106, 2459. (c) Nitro: Trost, B.
M.; Kuo, G.-H.; Benneche, T. J. Am. Chem. Soc. 1988, 110, 621.
(7) (a) Cha, J. K.; Christ, W. J.; Kishi, Y. Tetrahedron Lett. 1983, 3943.
(b) Christ, W. J.; Cha, J. K.; Kishi, Y. Tetrahedron Lett. 1983, 3947. (c)
Cha, J. K.; Christ, W. J.; Kishi, Y. Tetrahedron 1984, 40, 2247.
(8) A recent report has convincingly shown that an OsO4‚TMEDA
complex in nonpolar solvents can be effeciently directed by hydrogen
bonding to allylic alcohols, see: Donohoe, T. J.; Moore, P. R.; Waring, M.
J.; Newcombe, N. J. Tetrahedron Lett. 1997, 5027.
We hypothesize that minimization of the steric interactions
felt by the ortho side chains of the CdC double bond favors
a reactive conformer in which entry of OsO4 anti to the -OH
group provides the observed product (Figure 2).10 We note
(9) (a) Shibaski, M.; Mase, T.; Ikegami, S. Chem Lett. 1983, 1737. (b)
Honda, T.; Tomitsuka, K.; Tsubuki, M. J. Org. Chem. 1993, 58, 4274. (c)
Kita, Y.; Yoshida, Y.; Mihara, S.; Furukawa, A.; Higuchi, K.; Fang, D.-F.;
Fujioka, H. Tetrahedron 1998, 54, 14689.
(10) Also deserving of consideration is a conformer featuring a planar
hydrogen bond between the secondary alcohol and the adjacent quinone
carbonyl group. In such a reacting ensemble, the phenyl group of the benzyl
ether side chain can fold over and engage in a π-stacking interaction with
the quinone, successfully explaining the observed stereochemistry. However,
two additional considerations do not support this explanation: (1) the
proposed hydrogen bond is expected to deactivate the CdC bond toward
dihydroxylation and (2) when the hydroxyl group in 13a is protected as a
TBS ether, 14a is again obtained with very high stereochemical fidelity
(between 8:1 and >20:1, depending on conditions). A full discussion of
the stereochemistry observed in this dihydroxylation will be provided in a
forthcoming full paper on the subject.
(11) We have recently shown that a titanium-carbohydrate complex, first
introduced by Duthaler,12a is capable of introducing >95% ee for the aldol
reaction on an equivalent of 11 that is projected to be useful for our total
synthesis of lactonamycinone. (a) Duthaler, R. O.; Herold, P.; Lottenbach,
W.; Oertle, K.; Riediker, M. Angew. Chem., Int. Ed. Engl. 1989, 28, 495.
(12) Crystallographic data for 17 have been deposited with the Cambridge
Crystallographic Data Center as supplementary publication number CCDC-
165580.
Figure 2. Proposed reactive conformer of 13a and Chem-3D
representation of the X-ray structure of 17.
that this highly stereoselective dihydroxylation reaction
provides a practical solution to the enantioselective construc-
tion of lactonamycinone, provided that an aldol reaction of
Org. Lett., Vol. 3, No. 18, 2001
2901