With these promising results in hand, we continued
construction of the isotwistane skeleton. Treatment of 19 with
triphenyltin hydride under the various reaction conditions
afforded the desired cyclization product 21 (Scheme 6), the
To our delight, treatment of bromoacetal 24 with tributyltin
hydride and triethylborane delivered 27 as a mixture of
diastereomers in excellent yield for the two steps. Mecha-
nistically, this transformation likely proceeds via initial
formation of primary radical 25 followed by 5-exo-trig
cyclization into the adjacent exocyclic olefin. A second
cyclization event into the maleate followed by trapping of
hydrogen from the less hindered face yields 27. The acetal
diastereomers (27a and 27b) were separated, and the
stereochemistry was confirmed via single-crystal X-ray
analysis of 27a.12
Scheme 6
Curious as to functional group tolerance in the cyclization
cascade, we explored preformation of the maleic anhydride
unit found in the natural product (Scheme 8). Thus, hydroly-
Scheme 8
highest yield being achieved with triethylborane in toluene.10
Stannane 21 was obtained in all cases as a single isomer,
the structure of which was confirmed by X-ray crystal-
lographic analysis.
To complete the isotwistane core, we turned to Stork’s
bromoacetal method in hope of advancing tertiary alcohol
21 via the intramolecular tandem/radical cyclization sequence
outlined in Scheme 7.11 To this end, stannane 21 was
protodestannylated with dry hydrochloric acid to furnish 22
which was efficiently converted to bromoacetal 24 by
exposure to dibromoethyl ether 23 and N,N-dimethylaniline.
Scheme 7
sis and dehydration of maleate 22 with acetic anhydride
provided maleic anhydride 28 in good yield. As before,
derivatization as the corresponding bromoacetal was followed
by exposure to Bu3SnH and Et3B. We were pleased to find
that the radical cascade sequence again delivered a cycliza-
tion product (29) in excellent yield.
Having established an extremely efficient method for
constructing the isotwistane core, we turned our attention to
the acid oxidation state at C27 (Scheme 1).13 Since we had
already successfully incorporated an acetate moiety into the
phenolic oxidation cascade by using Pb(OAc)4 (i.e., 15 f
20), our efforts focused on advancing the latter. In what
proved to be a remarkable one-pot procedure, we discovered
that simply exposing 20 to LiTMP and Eschenmoser’s salt
effects conversion to 30, thus setting the stage for the tandem
radical cascade.14 As before, bromoacetal formation furnished
(8) Anderson, D. R.; Koch, T. H. J. Org. Chem. 1978, 43, 2726.
(9) Fujisawa, T.; Kurita, Y.; Kawashima, M.; Sato, T. Chem. Lett. 1982,
1641.
(10) Nishida, A.; Takahashi, H.; Takeda, H.; Takada, N.; Yonemitsu,
O. J. Am. Chem. Soc. 1990, 112, 902.
(11) Stork, G.; Biller, S. A.; Rychnovsky, S. D.J. Am. Chem. Soc. 1983,
105, 3741.
(12) 27a and 27b were subjected separately to Jones conditions, in which
they both converged to the same lactone.
(13) Phomoidride numbering system as reported in the following:
Dabrah, T. T.; Kaneko, T.; Massefski, W., Jr.; Whipple, E. B. J. Am. Chem.
Soc. 1997, 119, 1594.
(14) Lansbury, P. T.; Zhi, B.-X. Tetrahedron Lett. 1988, 29, 5735.
Org. Lett., Vol. 3, No. 16, 2001
2437