Organic Letters
Letter
product, ester 6, was obtained in 65% yield over two steps.
Hydroboration of 6 from the convex face, followed by
treatment with H2O2 afforded diol 7 in 38% yield (58%
based on the recovery of starting material). Bicycle 7 was first
hydrolyzed to the corresponding carboxylate, which formed
the tricyclic lactone upon exposure to 5 M HCl in
tetrahydrofuran (THF) at room temperature.5 Subsequent
i
acetylation delivered Pr-chomodorolide B (8) in 63% yield
over two steps, the structure of which was confirmed by XRD.
Encouraged by the successful model study, we carried out
further synthetic studies toward chromodorolide B (Scheme
2). The trans-hydrindanone 9 was first prepared via the
procedures reported by the Overman group5 and underwent a
Van Leusen reaction14 followed by a DIBAL-H reduction to
afford aldehyde 10 with a d.r. ratio of 5:1 in 60% yield over two
steps. The Corey−Fuchs reaction converted aldehyde 10 to
the corresponding dibromo compound,15 and the subsequent
n
addition of BuLi generated lithium acetylide 12 that was
trapped in situ by Weinreb amide 11 to afford alkynone 13 in
70% yield over two steps.16 The Paterno−Bu
̈
chi reaction of
̀
Figure 2. Synthesis of 3-oxabicyclo[3.3.0]octanes by the gold-
catalyzed alkoxycyclization.
alkynone 13 and furan afforded inseparable diastereomeric
oxetanes 14a and 14b (1:1) in 62% yield (91% brsm). The
poor diastereoselectivity was expected given the distance
between the chiral hydrindane and the ketone carbonyl group
(reaction center). Gratifyingly, by screening a variety of
reaction conditions (Table S1), we found that 5 mol %
tBu3PAuCl/AgSbF6 catalyzed the alkoxycyclization of 14a/b to
afford a pair of separable diastereomers 15a and 15b in 37 and
15% yield, respectively. The structure of 15b was unambigu-
ously determined by XRD, suggesting that the major product,
15a, was the desired stereoisomer for the total synthesis of 1.
The epimerization of the C12 stereogenic center was
accomplished by a similar two-step sequence shown in our
model study, but we found the α-hydroxylation step worked
better with O2 as the oxidant, leading to 16 in 54% yield over
two steps.
(25% isolation yield of the desired product), presumably due
to the increased steric hindrance on the alkenyl nucleophile.
Interestingly, with substrate 4g (R1 = −Me, R2 = −H, R3 =
−H), product 5g was obtained in 55% yield instead of the
direct cascade product, the stereochemistry of which was
determined by a nuclear Overhauser effect spectroscopy
(NOESY) experiment. (See the SI.) This observation could
be rationalized by the facile gold-catalyzed allylic ether
formation from the allylic alcohol.11 Replacing the methanol
with benzyl alcohol also smoothly provided the corresponding
3-oxabicyclo[3.3.0]octane in a similar yield (5h).
On the basis of our retrosynthetic analysis of chromodor-
olide B (Figure S3), we first carried out a model study using 3-
oxabicyclo[3.3.0]octane 5a in hand (Scheme 1). The inversion
of the C12 stereogenic center was achieved by (1) using excess
SmI2/pivalic acid in THF−HMPA to reductively remove the
hydroxyl group12 and (2) deprotonation of the resulting ester,
followed by treatment with Davis oxaziridine.13 The desired
However, we encountered an insurmountable hurdle in the
hydroboration of C8−C17 alkene. A variety of conditions
screened did not afford compound 17 from 16, and the
decomposition of starting material (16) to unidentifiable
products was observed. The difficulty of this hydroboration
could be ascribed to two reasons: (1) The hydrindane motif
created a stifling congestion on the convex face, preventing the
rhombic transition state of hydroboration (TS A); (2) the
transformations on the convex face of cis-bicyclo[3.3.0]octenes
were torsionally unfavored.17 Consequently, other side
reactions, including the hydroxyl-directed reduction of ester18
and the decomposition of ketals due to the Lewis acidity of
borane reagents, predominated. Consistent with this explan-
ation, even the hydrogenation of C8−C17 alkene turned out to
be extremely sluggish, which was completed after 6 days at 35
°C in the presence of PtO2. Compound 18 was isolated as a
single diastereomer in 80% yield and subjected to a hydrolysis/
acidification/acetylation sequence to afford 17-deacetoxyl
chomodorolide B (19) in 66% yield over two steps, the
structure of which was confirmed by XRD.
a
Scheme 1. Model Study of Chromodorolide B
In summary, we have developed a novel strategy to
synthesize highly substituted 3-oxabicyclo[3.3.0]octanes by
a
Reagents and conditions: (a) SmI2 (5.0 equiv), pivalic acid (4.0
equiv), HMPA, THF, rt; (b) KHMDS (1.1 equiv), (1S)-(+)-(10-
camphorylsulfonyl oxaziridine (1.2 equiv), HMPA, THF, −78 °C,
65% (two steps); (c) BMS (2.5 equiv), DCE, 0 to 35 °C, then
NaHCO3, H2O2, THF, rt, 38% (58% brsm); (d) NaOH, THF, rt,
then HCl, THF, rt; (e) DMAP (0.8 equiv), pyridine (17.0 equiv),
Ac2O (12.0 equiv), rt, 63% (two steps).
̀
combining the Paterno−Buchi reaction of furan and a gold-
̈
catalyzed cascade reaction. Even though we were not able to
accomplish chromodorolide B, the efficiency of the current
approach to construct the skeleton of these rearranged
spongian diterpenoids was noteworthy. The caveat of our
B
Org. Lett. XXXX, XXX, XXX−XXX