Synthetic approach to the AB ring system of ouabain

Article abstract of DOI:10.1021/jo020454+

Several novel hydroxylated cis-decalin derivatives, potential intermediates for the synthesis of the AB ring system of the important cardiotonic steroid ouabain, have been synthesized from commercially available starting materials. The first step in the preparation of these highly functionalized intermediates is a Robinson annulation of the β-keto ester 6 and the 4-silyl-3-buten-2-one 5 to furnish the octalone 4 with good diastereoselectivity in fair yield (due to competition with a novel silicon-to-carbon phenyl migration). Reduction of the epoxy alcohol 3 (derived from 4 in two high-yielding steps) with LiAlH₄ gave a mixture of the desired triol 11 along with the product of an unusual reductive opening at the tertiary carbon, namely the triol 12. A plausible mechanism for this unusual reduction is presented as are possible methods for avoiding it. In particular, reduction of the corrresponding epoxy ketone 15 with aluminum amalgam proceeded in good yield to give the hydroxy ketone 16. Also reduction of the epoxide ester having the inverted stereochemistry at C3 afforded the desired tertiary alcohol 33 in good yield. Another approach using the β,γ-unsaturated ketal 38 permitted the formation of the tertiary alcohol 40. Fleming oxidation of the related, very functionalized silane 39 afforded the desired 1β-alcohol 41 in fair yield. Finally a novel rearrangement was observed when the epoxy alcohol 24 was treated with DIBAL to effect loss of the angular hydroxymethyl group to produce the tetrasubstitued alkene 29 in high yield.

Full text of DOI:10.1021/jo020454+

Syn th etic Ap p r oa ch to th e AB Rin g System of Ou a ba in
Michael E. J ung* and Grazia Piizzi
Department of Chemistry and Biochemistry, University of California, Los Angeles, 405 Hilgard Avenue,
Los Angeles, California 90095-1569
jung@chem.ucla.edu
Received J uly 8, 2002
Several novel hydroxylated cis-decalin derivatives, potential intermediates for the synthesis of the
AB ring system of the important cardiotonic steroid ouabain, have been synthesized from
commercially available starting materials. The first step in the preparation of these highly
functionalized intermediates is a Robinson annulation of the â-keto ester 6 and the 4-silyl-3-buten-
2
-one 5 to furnish the octalone 4 with good diastereoselectivity in fair yield (due to competition
with a novel silicon-to-carbon phenyl migration). Reduction of the epoxy alcohol 3 (derived from 4
in two high-yielding steps) with LiAlH gave a mixture of the desired triol 11 along with the product
4
of an unusual reductive opening at the tertiary carbon, namely the triol 12. A plausible mechanism
for this unusual reduction is presented as are possible methods for avoiding it. In particular,
reduction of the corrresponding epoxy ketone 15 with aluminum amalgam proceeded in good yield
to give the hydroxy ketone 16. Also reduction of the epoxide ester having the inverted stereochem-
istry at C3 afforded the desired tertiary alcohol 33 in good yield. Another approach using the â,γ-
unsaturated ketal 38 permitted the formation of the tertiary alcohol 40. Fleming oxidation of the
related, very functionalized silane 39 afforded the desired 1â-alcohol 41 in fair yield. Finally a
novel rearrangement was observed when the epoxy alcohol 24 was treated with DIBAL to effect
loss of the angular hydroxymethyl group to produce the tetrasubstitued alkene 29 in high yield.
In tr od u ction
The cardenolide ouabain (1), a species of the genus
Aconcanthera isolated from the bark and roots of the
1
ouabaio tree by Arnaud, has been used for more than
two centuries in the clinical treatment of congestive heart
2
failure (CHF). Like all cardiac glycosides, ouabain, which
belongs to the family of cardiotonic steroids (digitalis
+
+
glycosides), works by inhibiting Na ,K -ATPase to in-
crease the force of cardiac musculature contraction
3
(
positive inotropic effect). Structurally, ouabain differs
from most common steroids in that its AB and CD rings
are cis rather than trans fused, it has a tertiary hydroxyl
group at C-14, the 19-methyl group is hydroxylated, and
the substituent at C-17 is a butenolide. The compound’s
novel stereochemistry and high degree of oxygenation
have made it, not surprisingly, a challenging synthetic
target. However, although a number of syntheses using
steroidal starting materials4 and degradation studies5
have been reported, to date no total synthesis of ouabain
has appeared in the literature.
In our proposed total synthesis of ouabain, we envi-
sioned that the AB ring system (2) could be derived from
the epoxy alcohol 3 (Scheme 1), a key intermediate
already possessing three oxygenated carbons and a
phenyldimethylsilyl group as a masked hydroxyl with the
desired all-cis stereochemistry. The decalone derivative
4, which is produced from a Robinson annulation between
the dimethyl(phenyl)silyl enone 5 and the â-keto ester
6, seemed a suitable precursor for the epoxy alcohol 3.
We report herein our approach to the synthesis of the
tetraol 2.
(
(
(
(
1) Arnaud, M. Compt. Rend. Acad. 1888, 107, 1011.
2) Bigelow, N. M.; J acobs, W. A. J . Biol. Chem. 1932, 96, 647.
3) Ruegg, U. T. Experientia 1992, 48, 1102.
Resu lts a n d Discu ssion
4) (a) Hanson, J . R. Nat. Prod. Rep. 1993, 10, 313. (b) Hanson, J .
R. Nat. Prod. Rep. 1998, 15, 298.
5) (a) Overman, L. E.; Rucker, P. V. Tetrahedron Lett. 1998, 39,
643. (b) Overman, L. E.; Rucker, P. V. Heterocycles 2000, 52, 1297.
c) Hynes, J ., J r.; Overman, L. E.; Nasser, T.; Rucker, P. V. Tetrahedron
Lett. 1998, 39, 4647. (d) Deng, W.; J ensen, M. S.; Overman, L. E.;
Rucker, P. V.; Vionnet, J . P. J . Org. Chem. 1996, 61, 6760. (e) Laschat,
S.; Narjes, F.; Overman, L. E. Tetrahedron 1994, 50, 347.
A. Robin son An n u la tion . The Michael acceptor 5
(
4
(
was synthesized in five steps in high yield (Scheme 2)
6
starting from commercially available 3-butyn-2-ol. The
(6) J ung, M. E.; Piizzi, G. J . Org. Chem. 2002, 67, 3911.
1
0.1021/jo020454+ CCC: $25.00 © 2003 American Chemical Society
Published on Web 03/04/2003
2
572
J . Org. Chem. 2003, 68, 2572-2582

Synthetic Approach to the AB Ring System of Ouabain
SCHEME 1
SCHEME 2
TABLE 1. Robin son An n u la tion
product(s) ratio 4:9
yield of
entry
[5] (M)
[6] (M)/equiv
[NaOMe] (M)/equiv
temp (°C)/time
(cis:trans)
4 (%)
1
2
3
4
5
0.09
2.54
1.77
2.17
1.30
0.09/1-2
2.54/1
3.2/1.8
6.5/3
0.09/0.4-1
1.78/0.7
1.77/1
2.17/1
1.3/1
0-80/1-2 d
25-80/2.5 d
25-80/42 h
80/18 h
0.01:1 (6:1)
1.4:1 (6:1)
2.4:1 (6:1)
6.2:1 (1:1)
>7:1 (1.2:1)
<5
21
40
31
29
5/3.8
25-80/38 h
SCHEME 3
results of the Robinson annulation between the enone
3.6 The initial formation of a pentacovalent silicon
intermediate would trigger a 1,2-phenyl migration to the
adjacent electrophilic carbon.8 A new charged silicon
species could then leave behind a stabilized benzyl anion
which, upon protonation, would yield the observed prod-
uct 9.
In an attempt to minimize intramolecular quenching
of the enone 5 and to favor the desired intermolecular
annulation, several experimental conditions were tested
(Table 1). A significant increase in the concentration of
reactants resulted in a remarkable improvement in the
ratio of the desired product 4 and the rearranged product
9 with a consequent increase in the yield of the former
(entries 2-5). When an excess (1.8 equiv) of the â-keto
ester 6 was used to compensate for its loss during the
retro-Dieckmann process, the yield of the desired Rob-
inson annulation product reached 40% (entry 3). This
result is noteworthy considering that a similar Robinson
6
7
5
and the â-keto ester 6 are summarized in Table 1.
Initial attempts to build the desired decalone system
were carried out under mild conditions (entry 1) to avoid
the retro-Dieckmann condensation of the â-keto ester and
polymerization of the Michael acceptor. However, at low
concentration (<0.1 M) and temperature (<80 °C), none
of the desired product was observed. When the mixture
was refluxed for a prolonged period, only a trace of the
Robinson annulation product 4 was detected, the starting
material being consumed by unproductive pathways. A
significant amount of the retro-Dieckmann fragmentation
product of the â-keto ester was recovered along with
4
-phenyl-2-butanone 9. The formation of the major di-
astereomer (cis-4) is probably due to the greater stability
of the latter in which the bulky dimethyl(phenyl)silyl
group is placed in the equatorial position. A plausible
mechanism for the formation of 9 is shown in Scheme
(
7) Ruest, L.; Blouin, G.; Deslongchamps, P. Synth. Commun. 1976,
(8) Fleming, I.; Newton, T. W.; Sabin, V.; Zammattio, F. Tetrahedron
1992, 48, 7793.
6
, 169.
J . Org. Chem, Vol. 68, No. 7, 2003 2573

J ung and Piizzi
SCHEME 4
SCHEME 5
SCHEME 6
annulation using the trimethylsilyl enone, for which no
such migration can occur, afforded a mixture of diastereo-
meric products in only 45% yield. Subsequent attempts
The allylic alcohol 10 was protected as its TBS ether
13 under standard conditions in 65% yield (Scheme 5).
9
However, attempted oxymercuration of 13 with Hg(TFA)
or Hg(OAc) was unsuccessful. Epoxidation of 13 with
2
to further tune the reaction conditions failed. While the
rearrangement was clearly disfavored (product ratios for
entries 3-5), the yield of the desired enone 4 did not
improve. Therefore we used the conditions of entry 3 to
obtain consistently a 40% isolated yield of the desired
enone.
B. Tow a r d th e AB Rin g System of Ou a ba in . For
the synthesis of the target AB ring system, the Robinson
annulation product cis-4 was subjected to Luche reduc-
tion10 and afforded the desired â-allylic alcohol 10
exclusively in nearly quantitative yield, as shown in
Scheme 4. Subsequent epoxidation occurred only on the
â-face as expected to give the desired epoxy alcohol 3.
Reduction of the ester and opening of the epoxide with
2
mCPBA was very slow and ultimately required 6 days
to go to completion to give only the undesired R-epoxide
14 in excellent yield. Nevertheless, the epoxy alcohol 3
could be oxidized in 72% yield to the corresponding
ketone 15, which upon treatment with aluminum amal-
gam gave exclusively the tertiary alcohol 16 in good
yield.11 Unfortunately, selective reduction of the ketone
16 with NaBH in methanol, lithium in liquid ammonia,
4
or Ni-Raney in ethanol only resulted in complex mix-
tures of products.
We then returned to the triol 11, and attempted to
oxidize the silyl group12 of this advanced intermediate,
the last step required for the synthesis of the target AB
ring system (Scheme 6). Treatment of the silane 11 with
mercury(II) trifluoroacetate in a mixture of acetic and
trifluoroacetic acid followed by addition of peracetic acid
4
LiAlH afforded the all-cis triol 11 in 52% yield. The
regioisomeric triol 12, derived from opening of the epoxide
at the tertiary center, was also isolated in 41% yield.
Even though this represented a very rapid approach to
the desired all-cis triol 11, the low selectivity in the
reduction prompted us to investigate other routes for the
introduction of the tertiary hydroxyl group with improved
selectivity.
(9) Hwu, J . R.; Furth, P. S. J . Am. Chem. Soc. 1989, 111, 8834.
(10) Gemal, A. L.; Luche, J .-L. J . Am. Chem. Soc. 1981, 103, 5454.
(11) Corey, E. J .; Ensley, H. E. J . Org. Chem. 1973, 38, 3187.
(12) For an excellent review see: J ones, G. R.; Landais, Y. Tetra-
hedron 1996, 52, 7599.
2
574 J . Org. Chem., Vol. 68, No. 7, 2003

Synthetic Approach to the AB Ring System of Ouabain
SCHEME 7
SCHEME 8
resulted in a complex mixture of products.13 A similar
result was obtained during alkoxy-mediated intra-
molecular displacement of the phenyl group on silicon
Protection of the C-3 hydroxyl of the alcohol 3 with
TBSOTf gave in 96% yield the epoxide 23, which was
then treated with an excess of lithium aluminum hydride
to give 24 without competitive opening of the epoxide.
The bis-TBS-protected epoxide 27 was prepared from the
allylic alcohol 10 in three steps under the usual condi-
tions. We next subjected these new epoxides to several
hydride-opening conditions (Table 2). When the parent
epoxy-ester 3 was treated with 1 equiv of lithium
aluminum hydride, no reaction took place even after
refluxing at 80 °C over a long period (entry 1). Even using
3 equiv of reducing agent proved insufficient for opening
the epoxide and only succeeded in reducing the ester
(entry 2). Interestingly, 4 equiv of the hydride reagent
was found to be the minimum amount required for the
two-step process to go to completion (entry 4). However,
under these conditions no selectivity was achieved and
the opening occurred at the less hindered secondary
carbon as well as at the tertiary center. Moreover, when
the parent epoxide 3 was subjected to a large excess of
lithium aluminum hydride, a similar result was obtained,
namely, a 1:1 mixture of the regioisomeric triols 11 and
12 (entry 4). To exclude the possibility of a Li cation-
assisted epoxide opening, the epoxide 3 was treated with
sodium aluminum hydride under identical conditions as
(
NaH, THF, 0-22 °C).14
When the oxidation of the silyl triol 11 was carried out
under Fleming conditions,15 although some peaks of the
desired tetraol 2 could be detected in a crude mixture,
this was difficult to purify even after peracetylation.
Acetylation of the two most accessible hydroxyl groups
of 11 was accomplished under standard conditions.
During the oxidation of the resulting diacetate, using the
mild protocol introduced by Fleming et al.,16 the desired
diacetoxy diol 17 was detected as the major product
(
∼50%) in the crude mixture. However, purification by
conventional chromatographic techniques (flash chroma-
tography, preparative TLC, reverse-phase HPLC) was
unsuccessful. Even the bis(tert-butyldimethylsilyl)-pro-
tected silane 18 was found to be resistant to the same
mild oxidation protocol.16
To circumvent the isolation problems, the triol 11 was
treated with 2,2-dimethoxypropane and p-toluenesulfonic
acid to give exclusively the acetonide 20 in high yield
(Scheme 7). The C-3 hydroxyl was also protected as its
acetate to afford the less hydrophilic intermediate 21.
Unfortunately, probably due to the low reactivity of the
more sterically encumbered silyl group, this new silane
2
1 did not undergo oxidation.
C. Un exp ected Ep oxid e Op en in g. The competitive
opening of the epoxide 3 at the tertiary and secondary
carbons might be predicted based on the similar acces-
sibility of these two centers as shown in Figure 1
(geometry optimized by using Spartan PM3 calculation).
This hypothesis prompted us to prepare alternative
substrates on which to test the reductive epoxide opening
(Scheme 8).
(
13) Kolb, C. H.; Ley, S. V.; Slawin, A. M. Z.; Williams, D. J . J . Chem.
Soc., Perkin Trans. 1 1992, 2735.
14) Harada, T.; Imanaka, S.; Ohyama, Y.; Matsuda, Y.; Oku, A.
Tetrahedron Lett. 1992, 33, 5807.
15) Fleming, I.; Newton, T. W.; Sabin, V.; Zammattio, F. Tetrahe-
dron 1992, 48, 7793.
16) Fleming, I.; Sanderson, P. E. J . Tetrahedron Lett. 1987, 28,
229.
(
(
(
4
F IGURE 1. Epoxide 3.
J . Org. Chem, Vol. 68, No. 7, 2003 2575

J ung and Piizzi
TABLE 2. Ep oxid e Op en in ga
entry
compd
R1
R2
reductant
equiv
time
product(s)b
1
2
3
4
5
6
7
8
9
0
1
2
3
4
3
3
3
3
H
CO2CH3
CO2CH3
CO2CH3
CO2CH3
CO2CH3
LiAlH4
LiAlH4
LiAlH4
LiAlH4
NaAlH4
LiAlH4
LiAlH4
LiAlH4
LiAlH4
DIBALf
DIBALf
Al, NiCl2
LiAlH4
LiAlH4
1
3
4
10
4
2
1
4
15
6
9
3 dc
2 d
2 d
7 h
2 d
6 d
2 d
2 d
2 d
4 d
4 d
2 d
2 d
2 d
SM
26
H
H
H
H
d
11 (52%), 12 (41%)
11, 12 (∼1:1)
11, 12 (∼1:1), SM
SM
24, SM (1:6)
24, SM (1:1)
3
3
26
23
23
23
23
23
23
24
27
R ) H
TBS
TBS
TBS
TBS
TBS
TBS
CO2CH3
e
CO2CH3
CO2CH3
CO2CH3
CO2CH3
CO2CH3
12
1
1
1
1
1
SM
SM
SM
SM
12
3
3
15
3
R ) TBS
TBS
TBS
a
All reactions were carried out in THF at 22 °C unless otherwise stated. Determined from 1H NMR analysis of the crude mixture
b
unless otherwise stated. The reaction was run at 80 °C. d Isolated yields. e Treatment of the crude mixture with an additional 6 equiv
of LiAlH4 and reflux in THF for 2 days yielded 24 as the only isolated product (61%). The reaction was run in CH2Cl2.
c
f
SCHEME 9
before but no selectivity was observed (entry 5). When
the epoxy-diol 26 was treated with 2 equiv of LiAlH , only
unreacted starting material was recovered, suggesting
that the reagent is trapped by the two hydroxyl groups
and is therefore unable to deliver the hydride to the
epoxide (entry 6). Several attempts to reduce compound
Surprisingly, when compound 24 was treated with an
excess of DIBAL a new product was isolated in good yield
(Scheme 9). Preliminary spectroscopic analysis revealed
that the unexpected compound had lost the epoxide and
gained a tetrasubstituted alkene. To rationalize these
findings, we first postulated a DIBAL-catalyzed opening
of the epoxide that could lead to the spiro compound 28
where the angular carbon is still intact (Scheme 9, path
a). Alternatively, in the presence of DIBAL the epoxide
24 could undergo a syn-elimination with complete loss
of the angular hydroxymethyl to yield the alkene 29
(Scheme 9, path b). The intermediate for such an
elimination mechanism, shown in Scheme 9, resembles
the cyclic transition state proposed by Rickborn et al. for
the lithium diethylamide-induced syn-elimination of ep-
4
2
3, in which the C-3 hydroxyl is protected as the TBS
ether, were undertaken (entries 7-12). Using a limited
excess of lithium aluminum hydride gave a mixture of
monoreduced product (24) and unreacted starting mate-
rial (entries 7 and 8). Conversely, a large excess of
reducing agent resulted in reduction of the ester (entry
9
) and opening of the epoxide at the tertiary center,
exclusively, with loss of the protecting group (entries 9
and 14). Other reducing conditions, such as an excess of
diisobutylaluminum hydride (DIBAL) or a slurry of
aluminum and nickel(II) chloride17 (entries 10-12), failed
to give any reaction. The same result was obtained when
20
oxides to the corresponding allylic alcohols. Interest-
ingly, extensive spectroscopic analyses (GC/MS, COSY,
HMQC, and homodecoupling experiments) were consis-
tent with compound 29.
the epoxide 23 was refluxed with Na[PhSeB(OEt)
3
] and
acetic acid in ethanol18 or with PhSeNa and Ti(OiPr)
,
19
4
Intrigued by this finding, we found a literature prece-
dent showing that an analogous substrate (30) underwent
a similar rearrangement in the presence of a strong Lewis
thereby confirming the difficulty of opening these ep-
oxides with nucleophilic reagents.
2
1
(
(
17) Sarmah, B. K.; Barua, N. C. Tetrahedron 1991, 47, 8587.
18) Miyashita, M.; Suzuki, T.; Yoshikoshi, A. J . Am. Chem. Soc.
acid to give the spiro compound 31 (Scheme 10).
1
1
989, 111, 3728.
19) Blay, G.; Cardona, L.; Garc ´ı a, B.; Pedro, J . R. J . Org. Chem.
993, 58, 7204.
(
(20) Thummel, R. P.; Rickborn, B. J . Am. Chem. Soc. 1970, 92, 2064.
(21) Hwu, J . R.; Wetzel, J . M. J . Org. Chem. 1992, 57, 922.
2
576 J . Org. Chem., Vol. 68, No. 7, 2003

Synthetic Approach to the AB Ring System of Ouabain
SCHEME 10
SCHEME 11
The difference in behavior of the two epoxy-decalins
30 and 24) can be rationalized by comparing the stability
2) to yield either a 1,3- or 1,4-diol, respectively. Con-
versely, the more Lewis acidic DIBAL is capable of
coordinating the epoxide oxygen before delivering the
hydride intermolecularly. This opening occurs predomi-
nantly at the center that allows for the formation of a
five-membered-ring chelate and thus produces a 1,2-diol.
(
of the putative carbocations (Scheme 11). In the case of
the epoxide 30, the carbocation intermediate I is stabi-
lized by the silyl group via carbon-silicon σ-bond hyper-
conjugation. Consequently, the optimal overlap allows for
the elimination of the silyl group to give the observed
alkene 31.21 On the other hand, inversion at the stereo-
center R to the silyl group tends to disrupt this stabiliza-
tion thus preventing the formation of the carbocation II.
This is confirmed by the fact that neither the product of
a proton loss (28) nor the product of silyl group elimina-
tion (as in 31) was observed in the reaction of 24. More
likely, the tertiary C-O bond of the epoxide 24 breaks
to give a tertiary carbocation, which is converted into the
alkene 29 with loss of formaldehyde via a syn-elimination
mechanism (Scheme 9).
D. Ep oxid e Op en in g: Ra tion a le a n d Solu tion . The
intriguing data shown in Table 2 regarding the unpre-
dicted and undesired opening of a variety of epoxides
prompted us to carry out further investigations. Studies
on the selectivity of the opening of acyclic epoxy-alcohols
can be found in the pioneering work of Kishi et al.22 They
investigated the behavior of epoxy alcohols toward the
different hydride reagents, REDAL and DIBAL (Scheme
While a few examples of unusual epoxide openings on
23
cyclic substrates can be found in the literature, we were
unaware of a systematic approach toward overcoming
this problem. An intramolecular hydride delivery was
clearly not feasible owing to conformational restrains,
and it seemed unlikely that LiAlH could act as a Lewis
4
acid in a manner similar to DIBAL. However, this being
the case, a new chelation model could be postulated
(Scheme 13). After initial deprotonation of the hydroxyl
groups and reduction of the ester in 3, two equilibrating
chelates III and IV can be formed. The opening of this
unreactive epoxide is probably assisted by the lithium
or sodium cation (M) that plays a crucial role in this
process. In the chelate III, the C-3 hydroxyl can form a
five-membered chelate with the epoxide. According to
Kishi’s model, intermolecular hydride delivery should
then occur to give the five-membered-ring diol chelate
preferentially and, as a result, the undesired triol 12 is
produced. When the angular hydroxyl is involved in the
chelation (IV), external delivery of hydride should occur
at the secondary center to favor the formation of a six-
membered-ring chelate and thus the desired triol 11
should be produced. The lack of selectivity observed
during the openings (see Table 2) confirmed that both
chelates are involved in the process. A system for testing
this model and for selectively accessing the desired triol
was therefore needed. The primary consideration here
was to protect the hydroxyl at C-3 with a suitable
protecting group (P ), as shown in A (Scheme 14), a group
1
2). When REDAL is the reducing agent, an intramo-
SCHEME 12
lecular hydride delivery takes place predominantly
through a five- or six-membered-ring chelate (n ) 1 or
(
23) (a) Liu, W.; Rosazza, J . P. N. Synth. Commun. 1996, 26, 2731.
b) Magar, S. S.; Desai, R. C.; Fuchs, P. L. J . Org. Chem. 1992, 57,
360. (c) Pe n˜ a, W.; L o´ pez, J . T.; Cort e´ s, M. Synth. Commun. 1989, 19,
2841.
(
5
(22) Finan, J . M.; Kishi, Y. Tetrahedron Lett. 1982, 23, 2719.
J . Org. Chem, Vol. 68, No. 7, 2003 2577

J ung and Piizzi
SCHEME 13
SCHEME 14
that could withstand the reaction conditions (large excess
more efficient than benzoic acid for inversion of sterically
hindered alcohols.24 The benzoate 34 was treated under
the usual opening conditions to give the desired epimeric
triol 33 in good yield. It is noteworthy that the process
accomplishes three steps: reduction of two esters and
opening of the epoxide exclusively at the secondary
center. However, the isolation of 33 was made difficult
by the presence of a byproduct in which the nitro group
had also been reduced. Therefore, we decided to use
benzoic acid during the Mitsunobu reaction. The result-
ing inverted benzoate 35 was formed in good yield with
of LiAlH
chelation by the C-3 oxygen without preventing the
desirable) chelation by the angular hydroxyl group. With
4
, 2 days at 22 °C) and reduce the possibility of
(
this optimal protecting group, the proposed model pre-
dicted that the opening would occur selectively at the
secondary carbon to afford 32 as the major product.
Unfortunately, finding such a protecting group proved
problematic. Consequently, we changed our strategy.
Chelate B would require inversion of the stereochemistry
at C-3 but it has several advantages over A. The
undesired chelation by the C-3 oxygen would be pre-
vented, without the need of a protecting group. Second,
the opening in B would take advantage not only of the
assistance of the angular hydroxyl but also of an in-
tramolecular hydride delivery by the C-3 alkoxy at the
desired secondary center to give the triol 33 as the major
or exclusive product. Our approach to solving the selec-
tivity problem while giving further support to our che-
lation model is shown in Scheme 15.
4
an appreciably cleaner reduction using LiAlH .
After confirming the validity of our proposed model,
we moved to the inversion of C-3 to the â-stereochemistry
required for the AB ring system of ouabain (Scheme 16).
SCHEME 16
SCHEME 15
After protecting the primary and tertiary alcohols as the
acetonide 36, we attempted the Mitsunobu reaction using
different conditions. However, owing to the steric conges-
tion of the R-alcohol, the inverted ester 37 was not
formed. Even activation with chloromethylsulfonyl chlo-
ride was unsuccessful. Consequently, although this strat-
egy proved successful in the selective opening of the
The epoxy ester 3 was subjected to standard Mitsunobu
conditions with use of p-nitrobenzoic acid, known to be
(24) Martin, S. F.; Dodge, J . A. Tetrahedron Lett. 1991, 26, 3017.
2
578 J . Org. Chem., Vol. 68, No. 7, 2003

Synthetic Approach to the AB Ring System of Ouabain
SCHEME 17
epoxide, the epimeric triol 33 was not suitable for further
manipulation. We therefore undertook a new pathway
to the AB ring system (Scheme 17). This approach starts
with the migration of the double bond of the enone cis-4
to the B ring and concomitant protection of the carbonyl
at C-3 to give alkene 38 in 51% (nonoptimized) yield.25
The stereoselective epoxidation of the double bond is
carried out in high yield by using Yang’s protocol.26 In
the absence of the chelating hydroxyl at C-3, the epoxide
proposed a plausible mechanism and solution for its
unprecedented DIBAL-promoted rearrangement and un-
usual LiAlH
4
opening. Two successful approaches to
improve the selectivity of the opening have been pre-
sented; however, further manipulation of these advanced
intermediates was problematic. Last, it was found that
highly functionalized and sterically congested systems
such as ours are not suitable candidates for the Tamao-
Fleming oxidation sequence. Consequently, we are cur-
rently investigating the possibility of carrying out this
process at an earlier stage in the synthesis.
3
9 opened selectively at the secondary center to afford
the desired diol 40, exclusively. Encouraged by these
results, we then decided to carry out the Tamao-Fleming
oxidation of the silyl group. The epoxide 39 was treated
under the standard Fleming protocol15 but a complex
mixture of products resulted, which was also the case
Exp er im en ta l Section
Gen er a l. 1H NMR spectra were obtained at 400.132 or
and AcOOH.13 Finally, the epoxide 39
13
using Hg(TFA)
was converted to the desired alcohol 41 using the mild
2
500.132 MHz as indicated. C NMR spectra were recorded at
1
13
100.625 or 125.773 MHz as indicated. H NMR and C NMR
data are reported in parts per million (δ) downfield from
tetramethylsilane. Resonance patterns are reported with the
following notations: br (broad), s (singlet), d (doublet), t
1
6
protocol introduced by Fleming et al., which is known
to tolerate even highly substituted epoxides.27 This
advanced intermediate would then be reduced with
(triplet), q (quartet), and m (multiplet). Second-order spectra
LiAlH
4
, as for the preparation of 40. Then, the desired
in which coupling cannot be obtained by inspection are
reported as multiplets, with the center of the signal indicated
by the δ value given. Infrared spectra were recorded as neat
liquid films, and only the most significant absorption bands
tetraol 2 would be accessed after ketal hydrolysis followed
3
by ketone reduction using iBu Al, which has been shown
to give selectively the most stable, equatorial alcohol.2
Unfortunately, the yield of the isolated product 40 was
low, confirming the difficulty of carrying out such an
oxidation on a highly functionalized substrate.
8
-
1
are reported (in cm ). Thin-layer chromatography (TLC) was
carried out by the use of silica gel 60 F254 0.2 mm aluminum-
backed plates. Visual detection was performed with ultraviolet
light or with phosphomolybdic acid or permanganate stain.
Flash column chromatography was performed with silica gel
Su m m a r y
6
0 (230-400 mesh) and compressed air. All solvents/reagents
were purified by using literature procedures. All reactions were
performed under an atmosphere of argon unless otherwise
noted.
While our recent results toward the synthesis of the
AB ring system of ouabain were not entirely successful,
they are nonetheless not without a certain intellectual
interest and instructive value for future attempts at the
synthesis of this compound. Despite the apparent simi-
larities with other known decalins, this all-cis tetraol
system has been shown to be unexpectedly challenging.
We have nevertheless demonstrated that during the
Robinson annulation, a peculiar silicon-to-carbon phenyl
migration takes place and destroys the enone 5 intramo-
lecularly. Moreover, we have increased our understand-
ing of the unique reactivity of the epoxide 3 and have
(
()-(4R,4aR) Meth yl 4-(Dim eth ylph en ylsilyl)-2-oxo-2,3,4,
4
a ,5,6,7,8-octa h yd r on a p h th a len e-4a -ca r boxyla te (cis-4)
a n d (()-(4S,4a R) Meth yl 4-(Dim eth ylp h en ylsilyl)-2-oxo-
2,3,4,4a ,5,6,7,8-oct a h yd r on a p h t h a len e-4a -ca r b oxyla t e
(
4
tr a n s-4). Sodium metal was dissolved in methanol (MeOH,
mL) at ambient temperature. After complete dissolution, the
â-ketoester 6 (2.1 mL, 14.7 mmol) was added slowly via
syringe. After 15 min, the enone 5 (1.63 g, 7.99 mmol),
dissolved in MeOH (0.5 mL), was added over 7 h via syringe
pump. At the end of addition, the deep-orange solution was
stirred at ambient temperature for 18 h. Then the mixture
was refluxed in a preheated bath at 80-85 °C for 24 h before
being cooled to ambient temperature. The reaction mixture
was then poured into a separatory funnel containing saturated
(25) Nickisch, K.; Bittler, D.; Cleve, G.; Eckle, E.; Laurent, H. Liebigs
Ann. Chem. 1988, 579.
(
26) Yang, D.; J iao, G.-S. Chem. Eur. J . 2000, 6, 3517.
ammonium chloride (NH
ether (Et O). The combined organic layers were washed with
brine and dried over MgSO . The solvent was removed under
4
Cl, 50 mL) and extracted with diethyl
(27) Singleton, D. A.; Redman, A. M. Tetrahedron Lett. 1994, 35,
2
5
4
3
2
8, 627. (b) Overman, L. E.; McCready, R. J . Tetrahedron Lett. 1982,
3, 2355. (c) Overman, L. E.; Rucker, P. V. Heterocycles 2000, 52, 1297.
vacuum to give a crude orange oil. Purification by flash column
chromatography (5:1 hexanes/ethyl acetate) afforded the enones
J . Org. Chem, Vol. 68, No. 7, 2003 2579

J ung and Piizzi
cis-4 and trans-4 as a 6:1 mixture of diastereomers (1.03 g,
0%). A second column chromatography (8:1 hexanes/ethyl
acetate) yielded an analytical sample of each diastereomer.
cis-4: 1H NMR (CDCl
, 500 MHz) δ 7.48 (m, 2H), 7.34 (m,
10% aqueous sodium hydroxide (1 mL). After 20 min, the
mixture was extracted with ethyl acetate and the combined
organic layers were dried over Na SO . Solvent removal
2 4
afforded a crude colorless oil that was purified by preparative
thin-layer chromatography (2:1 ethyl acetate/hexanes, 1%
triethylamine) to yield the pure triols 11 (18.5 mg, 52%) and
12 (15 mg, 41%) as colorless oils.
4
3
3
2
1
1
H), 5.89 (s, 1H), 3.49 (s, 3H), 2.61 (br d, J ) 13.4 Hz, 1H),
.50 (dd, J ) 16.2, 16.2 Hz, 1H), 2.37 (m, 1H), 2.29 (dd, J )
6.2, 3.1 Hz, 1H), 2.04 (m, 1H), 1.81 (m, 1H), 1.64 (m, 2H),
.51 (app qt, J ) 13.5, 3.5 Hz, 1H), 1.40 (app qt, J ) 13.0, 3.7
1
11: H NMR (CDCl , 500 MHz) δ 7.49 (m, 2H), 7.33 (m, 3H),
3
Hz, 1H), 1.14 (app td, J ) 13.3, 3.8 Hz, 1H), 0.39 (s, 3H), 0.35
4.31 (m, 1H), 3.67 (m, 2H), 3.43 (m, 1H), 3.21 (br s, 2H), 2.06
1
3
(s, 3H). C NMR (CDCl
3
, 125.8 MHz) δ 199.3, 172.5, 164.4,
(m, 1H), 1.88 (m, 2H), 1.77 (m, 2H), 1.63 (m, 2H), 1.50 (m,
1
3
36.9, 134.0, 129.2, 127.6, 125.7, 51.8, 50.9, 39.0, 35.7, 35.1,
13
2
H), 1.38 (m, 2H), 1.25 (m, 2H), 0.38 (s, 3H), 0.36 (s, 3H).
C
-
1
3.6, 26.7, 23.3, -1.7, -4.0. IR (neat) 2937, 1728, 1666 cm
Si 342.1651.
, 500 MHz) δ 7.44 (m, 2H), 7.33
m, 3H), 5.89 (s, 1H), 3.61 (s, 3H), 2.32 (m, 4H), 2.13 (m, 2H),
.
NMR (CDCl
3
, 125.8 MHz) δ 139.4, 133.5, 128.9, 127.7, 68.8,
HRMS (m/e) found 342.1661, calcd for C20
26 3
H O
6
7.6, 46.2, 42.7, 34.8, 32.6, 29.7, 27.2, 25.5, 21.2, 20.4, -0.3,
tr a n s-4: 1H NMR (CDCl
-1
3
-2.0. IR (neat) 3262, 2930, 2862, 1661 cm
.
(
1
1
2: H NMR (CDCl
3
, 500 MHz) δ 7.61 (m, 2H), 7.40 (m, 3H),
1
3
1
2
.86 (m, 1H), 1.61 (m, 3H), 1.28 (m, 1H), 0.35 (s, 3H), 0.33 (s,
4
1
.08 (d, J ) 12.4 Hz, 1H), 3.66 (m, 1H), 3.48 (d, J ) 12.1 Hz,
1
3
H). C NMR (CDCl
33.7, 129.1, 127.7, 125.2, 52.8, 52.2, 35.7, 35.02, 34.1, 30.7,
3
, 125.8 MHz) δ 198.6, 174.2, 164.5, 137.3,
H), 3.45 (m, 1H), 1.91 (ddd, J ) 13.0, 13.0, 13.0 Hz, 1H), 1.78
(m, 2H), 1.67 (m, 2H), 1.30 (m, 6H), 0.91 (m, 1H), 0.83 (br d,
-1
8.01, 22.6, -2.0, -2.6. IR (neat) 2947, 1736, 1726, 1666 cm
.
13
3
J ) 12.3 Hz, 1H), 0.41 (s, 3H), 0.40 (s, 3H). C NMR (CDCl ,
125.8 MHz) δ 139.7, 133.7, 128.8, 127.8, 73.7, 72.7, 64.3, 50.4,
(
()-(2S,4R,4a R) Meth yl 2-Hyd r oxy-4-(d im eth ylp h en yl-
silyl)-2,3,4,4a ,5,6,7,8-octa h yd r on a p h th a len e-4a -ca r boxy-
la te (10). The enone cis-4 (352 mg, 1.03 mmol) was dissolved
in MeOH (10 mL) and the solution was cooled to 0 °C. Then
cesium(III) chloride heptahydrate (382 mg, 1.03 mmol) was
added followed by sodium borohydride (78 mg, 2.04 mmol).
After 5 min, the solution was warmed to ambient temperature
and stirred for 4 h. Then the cloudy mixture was poured into
4
3
+
0.5, 39.5, 36.8, 29.5, 26.6, 26.3, 22.1, -0.6, -2.1. IR (neat)
-1
323, 2922, 2856, 1728, 1714 cm . HRMS (m/e) found for (M
+
H) 335.2049, calcd for C19
H
31
O
3
Si 335.2042.
(
()-(1R,4R,4a R,8a R) Meth yl 2-Oxo-4-(d im eth ylp h en yl-
silyl)decah ydr on aph th [1,8a-b]oxir en e-4a-car boxylate (15).
The epoxide 3 (25 mg, 0.069 mmol) was dissolved in CH Cl
3 mL) and treated with pyridinium chlorochromate (PCC, 18
2
2
(
a separatory funnel containing saturated NH
extracted with ethyl acetate. The combined organic layers were
dried over MgSO . Solvent removal under vacuum resulted in
a clear crude oil. Purification by flash column chromatography
3:1 hexanes/ethyl acetate, 1% triethylamine) yielded the
4
Cl (20 mL) and
mg, 0.076 mmol). After being stirred for 6 h at ambient
temperature, the mixture was filtered through a pad of Celite
and silica gel, washing with CH Cl and ethyl acetate. After
2 2
4
solvent removal under vacuum, the dark crude oil was purified
by flash column chromatography (4:1 hexanes/ethyl acetate)
(
1
3
alcohol 10 as a colorless oil (348 mg, 98%). H NMR (CDCl ,
1
to give the epoxy ketone 15 (18 mg, 72%) as a colorless oil. H
5
1
2
1
3
1
2
00 MHz) δ 7.48 (m, 2H), 7.31 (m, 3H), 5.51 (s, 1H), 4.02 (m,
NMR (CDCl
3
, 500 MHz) δ 7.47 (m, 2H), 7.31 (m, 3H), 3.61 (s,
H), 3.47 (s, 3H), 2.48 (m, 1H), 2.12 (m, 1H), 1.91-1.79 (m,
H), 1.68 (m, 1H), 1.55 (m, 2H), 1.37 (m, 1H), 1.24 (m, 1H),
.09 (d, J ) 13.4 Hz, 1H), 1.00 (m, 1H), 0.37 (s, 3H), 0.29 (s,
3
2
1
1
H), 2.94 (s, 1H), 2.77 (dd, J ) 16.4, 7.9 Hz, 1H), 2.41 (m, 1H),
.23 (m, 2H), 1.83 (m, 1H), 1.65 (m, 1H), 1.58 (app td, J )
2.7, 3.6 Hz, 1H), 1.53 (dd, J ) 7.8, 5.9 Hz, 1H), 1.38 (m, 2H),
1
3
H). C NMR (CD CO, 125.8 MHz) δ 174.4, 140.1, 138.8,
6
.17 (m, 1H), 0.33 (s, 3H), 0.30 (s, 3H). 13C NMR (CDCl
, 125.8
3
33.8, 128.6, 128.1, 127.4, 67.4, 50.8, 49.9, 39.0, 34.4, 32.6, 30.5,
7.5, 23.7, -1.5, -5.2. IR (neat) 3515, 2932, 1728, 1699 cm-1
MHz) δ 206.1, 173.6, 138.4, 133.9, 129.0, 127.6, 69.4, 61.1, 51.6,
.
4
2
9.4, 37.3, 35.3, 32.7, 32.1, 24.8, 22.3, -1.9, -3.2. IR (neat)
(
()-(1S,2S,4R,4a R,8a R) Meth yl 2-Hyd r oxy-4-(d im eth -
-1
943, 1732, 1703 cm
()-(4R,4a R,8a S) Met h yl 2-Oxo-4-(d im et h ylp h en yl-
silyl)-8-h yd r oxyd eca h yd r on a p h th a len e-4a -ca r boxyla te
.
ylph en ylsilyl)decah ydr on aph th [1,8-b]oxir en e-4a-car boxy-
la te (3). The allylic alcohol 10 (290 mg, 0.843 mmol) was
(
2 2
dissolved in dichloromethane (CH Cl , 12 mL) and treated with
(
(
16). The epoxide 15 (54 mg, 0.15 mmol) was dissolved in THF
6 mL), H O (3 mL), ethanol (EtOH, 3 mL), and saturated
(0.6 mL) and the mixture was cooled to 0 °C. Then
m-chloroperbenzoic acid (mCPBA, 217 mg, 1.26 mmol). After
the solution was stirred for 3 d at ambient temperature, the
reaction was quenched by addition of saturated sodium sulfite
2
NaHCO
3
aluminum foil (600 mg) was cut in small pieces and immersed
in 2% aqueous mercury(II) chloride (10 mL) for 20 s. The
aluminum pieces were then washed with EtOH (50 mL) and
(
8 mL). The aqueous layer was extracted with CH
combined organic layers were washed with saturated sodium
bicarbonate (NaHCO ) and dried over MgSO . Solvent removal
2 2
Cl . The
3
4
Et
C, the aluminum pieces were removed by filtration through
filter paper washing with ethyl acetate. The aqueous layer was
washed with ethyl acetate and dried over MgSO . After solvent
2
O (60 mL) and added to the reaction flask. After 1 h at 0
under vacuum resulted in a colorless oil that was purified by
flash column chromatography (4:1 to 2:1 hexanes/ethyl acetate,
°
1
% triethylamine) to afford the epoxy alcohol 3 (300 mg, 99%)
1
4
as a white solid: mp 115-117 °C. H NMR (CDCl
δ 7.48 (m, 2H), 7.31 (m, 3H), 3.74 (m, 1H), 3.63 (s, 3H), 3.07
3
, 500 MHz)
removal, the crude oil was purified by flash column chroma-
tography (2:1 hexanes/ethyl acetate, 1% triethylamine) to give
(
(
s, 1H), 2.36 (m, 2H), 2.10 (app td, J ) 13.8, 4.8 Hz, 1H), 1.73
m, 2H), 1.55 (m, 2H), 1.37-1.22 (m, 2H), 1.15 (br d, J ) 13.7
1
the ketol 16 (39 mg, 72%) as a colorless oil. H NMR (CDCl
3
,
5
00 MHz) δ 7.48 (m, 2H), 7.35 (m, 3H), 3.65 (s, 3H), 3.31 (br
Hz, 1H), 0.83 (dd, J ) 13.3, 2.1 Hz, 1H), 0.34 (s, 3H), 0.31 (s,
1
3
s, 1H), 3.05 (br d, J ) 13.7 Hz, 1H), 2.53 (dd, J ) 14.1, 14.1
Hz, 1H), 2.34 (m, 2H), 2.06 (m, 1H), 1.86 (m, 2H), 1.54 (m,
4H), 1.36 (m, 2H), 0.38 (s, 3H), 0.34 (s, 3H). 13C NMR (CDCl3,
125.8 MHz) δ 208.9, 176.4, 138.2, 133.6, 129.1, 127.8, 75.5,
3
1
2
H). C NMR (CDCl
3
, 125.8 MHz) δ 174.4, 139.7, 133.7, 128.6,
27.5, 70.0, 66.0, 64.8, 51.5, 47.0, 39.1, 34.8, 33.7, 28.0, 24.1,
-
1
2.1, -1.1, -3.7. IR (neat) 3435, 2943, 1732 cm . HRMS (m/
Si 360.1756.
()-(2R,4R,4a S,8a S) 4-(Dim et h ylp h en ylsilyl)-4a -h y-
d r oxym et h yld eca h yd r on a p h t h a len e-2,8a -d iol (11) a n d
e) found 360.1766, calcd for C20
28 4
H O
5
3.5, 52.5, 51.7, 40.1, 34.6, 29.1, 26.5, 20.9, 20.3, -1.9, -3.0.
(
-1
IR (neat) 3433, 2924, 2864, 1720, 1713, 1693 cm
.
(
()-(1R,2S,4R,4a R,8a R) 4-(Dim et h ylp h en ylsilyl)-4a -h y-
(()-(1S,2S,4R,4aR,8aR) Meth yl 2-((1,1-Dim eth yl)eth yldi-
m e t h y ls ily lo x y )-4-(d im e t h y lp h e n y ls ily l)d e c a h y d r o -
n a p h th [1,8a -b]oxir en e-4a -ca r boxyla te (23). The alcohol 3
d r oxym eth yld eca h yd r on a p h th a len e-1,2-d iol (12). The ep-
oxide 3 (40 mg, 0.11 mmol) was dissolved in tetrahydrofuran
(
0
4 mL) and treated with lithium aluminum hydride (0.44 mL,
.44 mmol). After being stirred for 2 d at ambient temperature,
the cloudy mixture was quenched at 0 °C by addition of
saturated sodium sulfate (Na SO , 3 mL). After 5 min the
mixture was stirred at ambient temperature and treated with
2 2
(92 mg, 0.26 mmol) was dissolved in CH Cl and treated with
triethylamine (0.1 mL, 0.76 mmol). The resulting mixture was
cooled to -78 °C and TBSOTf (95 µL, 0.4 mmol) was added.
After being stirred for 2 h, the mixture was poured into a
2
4
2
separatory funnel containing water and extracted with Et O.
2
580 J . Org. Chem., Vol. 68, No. 7, 2003

Synthetic Approach to the AB Ring System of Ouabain
The combined organic layers were dried over MgSO and
solvent removal yielded a crude white solid. Purification by
flash column chromatography (3:1 hexanes/ethyl acetate, 1%
4
8.6 Hz, 2H), 7.98 (d, J ) 8.6 Hz, 2H), 7.47 (m, 2H), 7.25 (m,
3H), 5.38 (m, 1H), 3.69 (s, 3H), 3.09 (s, 1H), 2.41 (m, 1H), 2.27
(ddd, J ) 13.7, 13.7, 4.8 Hz, 1H), 2.09 (m, 1H), 1.81 (m, 1H),
triethylamine) afforded the silyl ether 23 (116 mg, 96%) as a
1.68 (m, 1H), 1.57 (m, 3H), 1.39 (m, 1H), 1.31 (m, 1H), 1.18
1
13
white solid: mp 97-99 °C. H NMR (CDCl
3
, 500 MHz) δ 7.48
(m, 1H), 0.37 (s, 3H), 0.36 (s, 3H). C NMR (CDCl
3
, 125.8 MHz)
(
m, 2H), 7.31 (m, 3H), 3.77 (ddd, J ) 10.2, 4.9, 0.8 Hz, 1H),
.65 (s, 3H), 2.92 (s, 1H), 2.36 (br d, J ) 14 Hz, 1H), 2.10 (dd,
δ 174.4, 163.6, 150.4, 140.0, 135.2, 133.5, 130.5, 128.4, 127.5,
3
123.4, 70.2, 62.6, 59.5, 51.6, 48.1, 38.6, 33.1, 28.3, 24.5, 24.2,
-
1
J ) 9.0, 9.0 Hz, 1H), 1.81 (m, 2H), 1.53 (m, 2H), 1.32 (m, 3H),
.16 (m, 1H), 0.84 (s, 9H), 0.78 (dd, J ) 13.5, 2.1 Hz, 1H), 0.346
s, 3H), 0.342 (s, 3H), -0.007 (s, 3H), -0.01 (s, 3H). 13C NMR
CDCl , 125.8 MHz) δ 174.6, 140.0, 133.6, 128.5, 127.5, 71.4,
5.0, 64.7, 51.4, 47.3, 39.5, 35.1, 33.8, 27.8, 25.7, 24.3, 22.2,
22.0, -0.9, -3.7. IR (neat) 2949, 1732, 1728, 1608 cm .
1
(()-(1S,2R,4R,4aR,8aR) Meth yl 2-Ben zoyloxy-4-(dim eth -
ylp h en ylsilyl)d eca h yd r on a p h t h [1,8a -b]oxir en e-4a -ca r -
boxyla te (35). The alcohol 3 (140 mg, 0.39 mmol) was
dissolved in THF (6 mL) and treated with triphenylphosphine
(204 mg, 0.77 mmol) and benzoic acid (95 mg, 0.77 mmol) at
ambient temperature. After 5 min, the white suspension was
treated with diethyl azodicarboxylate (130 µL, 0.77 mmol). The
solution turned orange immediately and clear within 3 min.
Then the mixture was refluxed for 18 h and stirred at ambient
temperature for an additional 18 h. The solvent was then
removed in vacuo and the residue was purified by preparative
(
(
3
6
1
-
1
8.0, -1.2, -3.99, -4.5, -4.8. IR (neat) 2951, 1732 cm
Si 474.2621.
()-(1S,2S,4R,4a S,8a R) 2-((1,1-Dim eth yl)eth yld im eth -
ylsilyloxy)-4-(d im et h ylp h en ylsilyl)d eca h yd r on a p h t h -
1,8a -b]oxir en e-4a -m eth a n ol (24). The ester 23 (35 mg, 0.07
mmol) was dissolved in THF (2 mL) and treated with LiAlH
0.15 mL, 0.14 mmol) at ambient temperature. After being
.
HRMS (m/e) found 474.2615, calcd for C26
42 4
H O
(
[
4
(
thin-layer chromatography (6:1 hexanes/EtOAc, 2 runs) to give
stirred for 44 h, the mixture was poured into a separatory
funnel containing water and extracted with ethyl acetate. The
combined organic layers were dried over Na SO and solvent
2 4
removal under vacuum afforded a crude colorless oil. Purifica-
tion by flash column chromatography (5:1 hexanes/ethyl
acetate, 1% triethylamine) yielded the alcohol 24 (28 mg, 88%)
1
the ester 35 (130 mg, 72%) as a white solid. H NMR (CDCl
3
,
5
3
00 MHz) δ 7.9 (m, 2H), 7.58 (m, 1H), 7.46 (m, 4H), 7.24 (m,
H), 5.38 (m, 1H), 3.65 (s, 3H), 3.11 (d, J ) 2.2 Hz, 1H), 2.39
(
1
m, 1H), 2.24 (ddd, J ) 13.7, 13.7, 4.9 Hz, 1H), 2.09 (m, 1H),
.80 (m, 1H), 1.68 (m, 1H), 1.57 (m, 3H), 1.42 (m, 2H), 1.17
1
3
1
(m, 1H), 0.36 (s, 3H), 0.35 (s, 3H). C NMR (CDCl
3
, 125.8 MHz)
as a white solid. H NMR (CDCl
7
1
)
(
3
, 500 MHz) δ 7.50 (m, 2H),
δ 174.5, 165.5, 139.8, 133.6, 132.9, 129.9, 129.4, 128.5, 128.2,
.32 (m, 3H), 4.05 (dd, J ) 11.4, 1.8 Hz, 1H), 3.77 (dd, J )
0.3, 5.0 Hz, 1H), 3.63 (dd, J ) 11.3, 8.5 Hz, 1H), 2.57 (dd, J
8.6, 3.0 Hz, 1H), 2.12 (ddd, J ) 13.6, 13.6, 4.9 Hz, 1H), 1.76
m, 3H), 1.4-1.12 (m, 7H), 0.88 (m, 1H), 0.82 (s, 9H), 0.74 (br
d, J ) 12.7 Hz, 1H), 0.41 (s, 3H), 0.38 (s, 3H), -0.02 (s, 6H).
, 125.8 MHz) δ 140.0, 133.6, 128.6, 127.6, 71.6,
7.7, 67.4, 65.1, 39.3, 37.8, 35.2, 33.1, 28.6, 25.7, 24.4, 21.5,
8.1, -0.9, -3.3, -4.6, -4.9. IR (neat) 3437, 2930 cm . HR-
MALDI (m/e) found for (M + Na) 469.2564, calcd for C25
Si Na 469.2586.
()-(1R,2S,4R) 2-((1,1-Dim eth yl)eth yldim eth ylsilyloxy)-
-(d im et h ylp h en ylsilyl)-1,2,3,4,5,6,7,8-oct a h yd r on a p h -
th a len -1-ol (29). The alcohol 24 (8 mg, 0.02 mmol) was
dissolved in CH Cl (1 mL) and treated with diisobutylalumi-
1
2
4
4
27.5, 68.9, 62.4, 59.7, 51.5, 48.0, 38.6, 33.2, 27.9, 24.6, 24.3,
-1
2.1, -0.9, -3.4. IR (neat) 2949, 1724 cm . HRMS (m/e) found
+
64.2016, calcd for C27
65.2093, calcd 465 2097.
32 5
H O Si 464.2019; (M + H) found
1
3
(()-(2S,4R,4a S,8a S) 4-(Dim et h ylp h en ylsilyl)-4a -h y-
d r oxym eth yld eca h yd r on a p h th a len e-2,8a -d iol (33). The
epoxide 35 (23 mg, 0.05 mmol) was dissolved in THF (3 mL)
C NMR (CDCl
3
6
1
-
1
+
4
and treated with LiAlH (0.25 mL, 0.24 mmol) at ambient
42 3
H O -
temperature. The resulting mixture was stirred for 2 d and
then was quenched by addition of aqueous 10% NaOH (1 mL)
at 0 °C. After being stirred for 15 min at 0 °C and 20 min at
ambient temperature, the mixture was diluted with ethyl
acetate and poured into a separatory funnel and the layers
were separated. The aqueous layer was extracted with ethyl
acetate and the combined organic layers were dried over
NaSO . After solvent removal under vacuum, the crude residue
4
was purified by preparative thin-layer chromatography (1:2
hexanes/ethyl acetate, 1% triethylamine) to give the triol 33
2
(
4
2
2
num hydride (DIBAL, 0.1 mL, 0.1 mmol) at ambient temper-
ature. After being stirred for 24 h, the reaction was quenched
by addition of saturated sodium, potassium tartrate (2 mL).
After being stirred for 30 min, the mixture was extracted with
ethyl acetate and the combined organic layers were dried over
1
(
14 mg, 88%) as a white solid. H NMR (CD
3
OD, 500 MHz) δ
2 4
Na SO . Solvent removal under vacuum gave a crude colorless
oil. Purification by preparative thin-layer chromatography (1:1
7
1
3
.54 (m, 2H), 7.29 (m, 3H), 4.03 (m, 2H), 3.46 (m, 1H), 2.0-
.9 (m, 3H), 1.77 (ddd, J ) 13.8, 13.8, 3.5 Hz, 1H), 1.6 (m,
hexanes/ethyl acetate) afforded the alcohol 29 (6.2 mg, 82%)
13
13
3
H), 1.5-1.2 13 (m, 6H), 0.35 (s, 6H). C NMR (CD OD, 125.8
1
as a white solid. H NMR (CDCl
7
1
3
, 500 MHz) δ 7.53 (m, 2H),
MHz) δ 140.2, 133.4, 128.1, 127.2, 74.8, 66.5, 65.8, 43.2, 42.9,
.32 (m, 3H), 3.71 (ddd, J ) 11.4, 4.0, 4.0 Hz, 1H), 3.68 (m,
H), 2.47 (m, 1H), 2.38 (m, 1H), 1.84 (m, 2H), 1.76 (t, J ) 12
3
6.0, 30.0, 28.2, 21.3, 20.5, -1.4, -2.3 (one high-field carbon
-1
unresolved). IR (neat) 3306, 2912 cm
()-(4′R,4′a R,8′R,8′a S) Meth yl 4′-Hyd r oxyl-1′,3′,4′,5′,6′,
′-8′-d eca h yd r osp ir o[1,3]d ioxola n e-2,2′-n a p h th [8′,8′a -b]-
.
Hz, 1H), 1.68 (m, 1H), 1.58 (m, 4H), 1.41 (m, 1H), 1.34 (m,
H), 0.87 (s, 9H), 0.34 (s, 3H), 0.28 (s, 3H), 0.05 (s, 3H), 0.03
(
1
7
1
3
(s, 3H). C NMR (CDCl
3
, 125.8 MHz) δ 139.4, 133.9, 133.4,
28.6, 127.5, 127.2, 72.1, 70.4, 31.3, 30.0, 29.3, 27.4, 25.7, 22.9,
2.5, 18.0, -2.8, -3.7, -4.5, -4.8. IR (neat) 3540, 2928 cm-1
oxir en e-4′a -ca r boxyla te (41). The epoxide 39 (25 mg, 0.06
mmol) was dissolved in glacial acetic acid (3 mL) and was
treated with potassium bromide (15 mg, 0.93 mmol) and
sodium acetate (76 mg, 0.93 mmol). After 5 min the reaction
was cooled to 0 °C and peracetic acid (32 wt % in AcOH, 0.2
mL, 0.9 mmol) was added. After 15 min, another aliquot of
peracetic acid (0.2 mL, 0.9 mmol) was added. After being
stirred for 10 min at 0 °C, the reaction was warmed to ambient
temperature and stirred for 28 h. Then the reaction was
quenched by addition of saturated Na S O . After being stirred
1
2
.
For HMBC, HMQC, COSY, and homodecoupling experiments
see the Supporting Information.
(
()-(1S,2R,4R,4a R,8a R) Meth yl 2-[4-Nitr oben zoyloxy]-
4
4
-(dim eth ylph en ylsilyl)decah ydr on aph th [1,8a-b]oxir en e-
-ca r boxyla te (34). The alcohol 3 (30 mg, 0.08 mmol) was
dissolved in benzene (2 mL) and treated with triphenylphos-
phine (110 mg, 0.42 mmol) and p-nitrobenzoic acid (70 mg,
2
2
3
0
.42 mmol) at ambient temperature. After being stirred for 5
for 20 min, the mixture was poured into a separatory funnel
containing saturated NaHCO and CH Cl was added. Solid
min, the white suspension was treated with diethyl azodicar-
boxylate (65 µL, 0.42 mmol). The solution turned orange
immediately and clear within 3 min. After the mixture was
stirred for 18 h, the solvent was removed in vacuo and the
residue was purified by flash column chromatography (5:1
3
2
2
NaHCO was added slowly until bubbling stopped. Then the
3
aqueous layer was extracted with CH Cl and the combined
2
2
2 4
organic layers were dried over Na SO . Solvent removal under
vacuum gave a yellowish residue that was purified by flash
hexanes/ethyl acetate) to afford the ester 34 (31 mg, 78%) as
column chromatography (1:3 hexanes/ethyl acetate) to afford
1
1
a pale yellow solid. H NMR (CDCl
3
, 500 MHz) δ 8.25 (d, J )
the alcohol 41 (3.2 mg, 19%) as a clear oil. H NMR (CDCl
3
,
J . Org. Chem, Vol. 68, No. 7, 2003 2581

J ung and Piizzi
5
1
00 MHz) δ 3.96 (m, 5H), 3.91 (s, 3H), 3.68 (d, J ) 11.2 Hz,
H), 2.93 (d, J ) 4.6 Hz, 1H), 2.35 (d, J ) 13.3 Hz, 1H), 2.16
Ack n ow led gm en t. We thank the National Insti-
tutes of Health for generous support.
(
1
ddd, J ) 13.0, 4.0, 2.9 Hz, 1H), 2.03 (m, 2H), 1.88 (dd, J )
2.6, 12.6 Hz, 2H), 1.72 (ddd, J ) 13.2, 13.2, 3.7 Hz, 1H), 1.52
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails for compounds 3, 4, 10, 11, 12, 15, 16, 23, 24, 29, 33, 34,
5, and 41; proton and carbon NMR spectra for all new
compounds reported. This material is available free of charge
via the Internet at http://pubs.acs.org.
13
(
(
5
1
m, 2H), 1.45 (m, 1H), 1.27 (dd, J ) 13.4, 2.6 Hz, 1H). C NMR
CDCl , 125.8 MHz) δ 176.4, 106.5, 67.9, 64.5, 64.2, 58.5, 58.1,
2.2, 51.6, 41.9, 41.1, 24.9, 22.4, 14.6. IR (neat) 3474, 2924,
3
3
-
1
+
714 cm . HRMS (m/e) found for M 284.1259, calcd for
+
C
C
14
H
H
20
O
O
6
284.1259; found for (M + H) 285.1342, calcd for
14
21
6
285.1338.
J O020454+
2
582 J . Org. Chem., Vol. 68, No. 7, 2003