Tandem use of cobalt-mediated reactions to synthesize (+)- epoxydictymene, a diterpene containing a trans-fused 5-5 ring system

Article abstract of DOI:10.1021/ja970022u

The diterpene (+)-epoxydictymene has been synthesized in 20 steps using the asymmetry of (R)-pulegone and several substrate-controlled diastereoselective reactions to prepare the natural product in its natural configuration. Three of the four rings were assembled with two consecutive intramolecular reactions involving dicobalt hexacarbonyl complexes of alkynes: a Lewis acid-promoted Nicholas reaction and a Pauson-Khand reaction. The construction of the strained trans-3-oxabicyclo[3.3.0]octane ring system of the natural product presented a significant challenge. To this end, several radical and anionic cyclizations were studied, the latter leading to (+)-epoxydictymene.

Full text of DOI:10.1021/ja970022u

J. Am. Chem. Soc. 1997, 119, 4353-4363
4353
Tandem Use of Cobalt-Mediated Reactions to Synthesize
(+)-Epoxydictymene, a Diterpene Containing a Trans-Fused
5-5 Ring System
Timothy F. Jamison, Soroosh Shambayati, William E. Crowe,† and
Stuart L. Schreiber*
Contribution from the Howard Hughes Medical Institute, Department of Chemistry and Chemical
Biology, HarVard UniVersity, Cambridge, Massachusetts 02138
ReceiVed January 2, 1997X
Abstract: The diterpene (+)-epoxydictymene has been synthesized in 20 steps using the asymmetry of (R)-pulegone
and several substrate-controlled diastereoselective reactions to prepare the natural product in its natural configuration.
Three of the four rings were assembled with two consecutive intramolecular reactions involving dicobalt hexacarbonyl
complexes of alkynes: a Lewis acid-promoted Nicholas reaction and a Pauson-Khand reaction. The construction
of the strained trans-3-oxabicyclo[3.3.0]octane ring system of the natural product presented a significant challenge.
To this end, several radical and anionic cyclizations were studied, the latter leading to (+)-epoxydictymene.
Isolated and fully characterized in 1983 (Scheme 1), the
diterpene epoxydictymene1 (1) contains an 8-membered ring2
with five asymmetric centers as its central structural element
and is one of four terpene natural products3-5 containing the
strained trans-fused 5-5 ring system6,7 (Figure 1). This
structural and stereochemical complexity, along with epoxy-
dictymene’s similarity to carbocyclic systems previously pre-
pared using two reactions of dicobalthexacarbonyl clusters of
alkynes,8-10 led us to explore this cobalt chemistry in the
Figure 1. Three sesquiterpene natural products containing the trans-
bicyclo[3.3.0]octane ring system (bold lines).
preparation of the natural product. These studies, reported
herein, led to the first total synthesis of a natural product
Scheme 1
containing a trans-5-5 moiety.11
Our synthetic strategy is outlined in Scheme 2. Epoxydic-
tymene possesses few functional groups, and therefore a
protecting group strategy was not a chief concern. On the
contrary, because of this dearth of functionality, we reasoned
that synthesis of the natural product might involve temporary
functional group introduction to facilitate carbon-carbon bond
formation. Accordingly, tetracyclic enone 2 became an ap-
† Current address: Department of Chemistry, Emory University, Atlanta,
pealing target since a reductive methylation might install the
GA 30322.
X Abstract published in AdVance ACS Abstracts, May 1, 1997.
(1) Enoki, N.; Furusaki, A.; Suehiro, K.; Ishida, R.; Matsumoto, T.
Tetrahedron Lett. 1983, 24, 4341.
(2) For a review concerning the synthesis of carbocyclic eight-membered
rings, see: Petasis, N. A.; Patane, M. A. Tetrahedron 1992, 48, 5757.
(3) R-Funebrene: Kirtany, J. K.; Paknikar, S. K. J. Indian Chem. 1973,
11, 508.
(4) Paradalianchol: Bohlmann, F.; Abraham, W. F. Phytochemistry 1979,
18, 668.
(5) R-8-Hydroxypresilphiperfolene: Bohlmann, W. F.; Zdero, C.; Jak-
upovic, J.; Robinson, H.; King, R. M. Phytochemistry 1981, 20, 2239.
(6) Chang, S.; McNally, D.; Shary-Tehrany, S.; Hickey, M. J.; Boyd, R.
H. J. Am. Chem. Soc. 1970, 92, 3109.
(7) Allinger, N. L.; Tribble, M. T.; Miller, M. A.; Wertz, D. H. J. Am.
Chem. Soc. 1971, 93, 1637.
(8) Schreiber, S. L.; Sammakia, T.; Crowe, W. E. J. Am. Chem. Soc.
1986, 108, 3128.
(9) Schreiber, S. L.; Klimas, M. T.; Sammakia, T. J. Am. Chem. Soc.
1987, 109, 5749.
(10) Shambayati, S.; Crowe, W. E.; Schreiber, S. L. Tetrahedron Lett.
1990, 31, 5289.
quaternary methyl group at C1. Whether this transformation
would construct the trans-5-5 ring system, however, was
uncertain. In contrast to 6-6 fused systems, the trans-5-5 is
much less stable than the cis (approximately 6 kcal/mol)6 and,
consequently, possesses a degree of strain and conformational
rigidity not usually associated with 5-membered rings. Further,
the difference in the heats of formation of the parent cis- and
trans-3-oxabicyclo[3.3.0]octanes, the 5-5 ring system present
in epoxydictymene, has been calculated to be over 10 kcal/mol.12
Nevertheless, as we and others had prepared similar fused 5-5
enones via intramolecular Pauson-Khand reactions,10,13,14 cobalt
cluster 3 was a logical precursor to 2. Our experience with the
Lewis acid-promoted Nicholas reaction8,9 suggested that 4,
fashioned with a displacement of a triflate derived from 5 by
an acetylide derived from propargylic acetal 6, should be the
initial focus of our synthetic efforts.
(11) (a) For a preliminary account of this work: Jamison, T. F.;
Shambayati, S.; Crowe, W. E.; Schreiber, S. L. J. Am. Chem. Soc. 1994,
116, 5505. (b) An alternative total synthesis of epoxydictymene appeared
after submission of this work: Paquette, L. A.; Sun, L.-Q.; Friedrich, D.;
Savage, P. B. Tetrahedron Lett. 1997, 38, 195.
(12) Schoening, A.; Friedrichsen, W. Z. Naturforsch. 1989, 44b, 975.
(13) For a review on the Pauson-Khand reaction see: Pauson, P. L.
Tetrahedron 1985, 41, 5855.
(14) Schore, N. E. Org. React. 1991, 40, 1.
S0002-7863(97)00022-X CCC: $14.00 © 1997 American Chemical Society

4354 J. Am. Chem. Soc., Vol. 119, No. 19, 1997
Jamison et al.
Scheme 2
Scheme 4
Scheme 5
Scheme 3
Unfortunately, all attempts to prepare propargylic acetal 6
by protic acid-catalyzed transacetalization of propiolaldehyde
diethyl acetal and 2-methyl-3-buten-2-ol were unsuccessful.
Thus, a similar compound, mixed acetal 10, was prepared by
bromodimethylborane treatment17 of propiolaldehyde diethyl
acetal and subsequent trapping of the bromomethyl propargyl
ether in situ with 2-methyl-3-buten-2-ol, giving 10 in 54% yield
(Scheme 4). Since it was unclear whether this mixed acetal
would exhibit the necessary site selectivity in the 8-membered
ring formation, 1,3-dioxane 11 (racemic) was also prepared.
Similar 1,3-dioxanes had been reported to exhibit complete site
selectivity in the desired sense in Lewis acid-catalyzed addition
reactions.18 In contrast to attempts to prepare 6 under protic
acid catalysis, p-toluenesulfonic acid successfully promoted the
reaction between propiolaldehyde diethyl acetal and 3-methyl-
1,3-butanediol, giving 11 in 66% yield. The intramolecularity
of the second transacetalization step in this case, as well as
reduced steric interaction in the product acetal (as compared to
that in 6), are likely contributors to the success of the reaction.
Coupling of 5 and cyclic acetal 11 was achieved under
carefully controlled conditions. Thus, activation of the alcohol
of allylsilane 5 with trifluoromethanesulfonic anhydride in the
presence of a hindered pyridine base afforded a triflate that was
treated immediately with a lithium anion derived from 11
(Scheme 5). A 79% yield of the coupled product 12 was
obtained as a 1:1 diastereomeric mixture at the acetal carbon.
Treatment of 12 in ether with dicobalt octacarbonyl incorporated
the alkyne into a dicobalt hexacarbonyl complex in greater than
90% yield. This compound is typical of such clusters, being
deep red in color, moderately air-sensitive, but nevertheless
stable enough to be purified using silica gel chromatography.
Synthesis and Coupling of the Two Fragments. All of the
asymmetry in the synthesis is derived from (R)-pulegone, and
its conversion to allylsilane 5 is depicted in Scheme 3.
Wolinsky and co-workers demonstrated15 that Favorskii ring
contraction of pulegone affords ester 7 stereospecifically,
favoring the diastereomer in which the methyl and carboxy-
methyl groups are oriented trans to each other. Saponification
of the ester and subsequent acid-catalyzed cyclization to the
fused cis-5-5 system established the configuration of C7
(epoxydictymene numbering) in 56% overall yield for the four
steps. Reduction of the lactone to the diol and acetylation of
the primary alcohol allowed for selective dehydration of the
tertiary alcohol, effected in boiling acetic anhydride. Saponi-
fication of the acetate yielded a 1:2 mixture of primary alcohols
9a and 9b (desired) that were separated chromatographically
after selective epoxidation (MCPBA) of the tetrasubstituted
olefin of the former. Subjecting 9b to 300 mol % of Schlosser’s
base16 and quenching the resulting dianion with chlorotrimeth-
ylsilane afforded, after hydrolysis of the silyl ether, a 50% yield
of allylsilane 5, along with a 36% yield of recovered 9b.
(15) Wolinsky, J.; Gibson, T.; Chan, D.; Wolf, H. Tetrahedron 1965,
21, 1247.
(16) Schlosser, M. Angew. Chem., Int. Ed. Engl. 1974, 13, 701.
(17) Morton, H. E.; Guindon, Y. J. Org. Chem 1985, 50, 5379.
(18) Ishihara, K.; Yamamoto, H.; Heathcock, C. H. Tetrahedron Lett.
1989, 30, 1825.

Co-Mediated Reactions to Synthesize (+)-Epoxydictymene
J. Am. Chem. Soc., Vol. 119, No. 19, 1997 4355
Table 1. Effect of Lewis Acid and Solvent on the Yield and Selectivity of the Cyclization of the Cobalt Complex Derived from 14 (Scheme 7)
entry
Lewis acid
solvent
temp (°C)/time
yield (%)
15:3 (site selectivity)
1
2
3
4
5
6
7
8
9
Et2AlCl
Et2AlCl
TiCl4
EtAlCl2
TMSOTf
TMSOTf
TMSOTf
TBDMSOTf
TESOTf
HOTf
PhMe/CH2Cl2
PhMe/CH2Cl2
CH2Cl2
PhMe/CH2Cl2
CH2Cl2
THF
Et2O
CH2Cl2
CH2Cl2
CH2Cl2
-78/1 h
0
50
73
50-78
87
63
91
85
87
73
na
80:20
na
78:22
70:30
0:100
91:9
65:35
63:37
50:50
-78 to 0/1 h
-78/30 min
-78/30 min
-78/15 min
-78/15 min
-78/15 min
-78/15 min
-78/15 min
-78/15 min
10
a All reactions were performed at 0.005 M in the solvent indicated, and 110 mol % of Lewis acid was used in each case.
Scheme 6
Scheme 7
Lewis Acid-Mediated Nicholas Cyclizations and Pauson-
Khand Reactions of Alkyne-Cobalt Complexes. The prepa-
ration of 8-membered rings by direct cyclization of acyclic
precursors is often a low-yielding process, primarily as a result
of unfavorable torsional strain and transannular interactions
engendered in the cyclization process.2 However, low-temper-
ature treatment of the cobalt complex derived from 12 with a
stoichiometric amount of Et2AlCl in dichloromethane afforded
a 91% yield of the fused 5-8 ring system of epoxydictymene
as cobalt complex 13. Although the cyclization substrate is a
involving allylsilane 5 and 10 was similar to that observed with
cyclic acetal 11 (Scheme 5). As before, the coupled product
14 was isolated as a 1:1 mixture of diastereomers at C10.
However, the cobalt complex derived from this alkyne was more
resistant to Lewis acids than that derived from 12, and several
conditions were screened until the optimum cyclization protocol
was found. In further contrast to the cyclization of the substrate
containing the cyclic acetal, the cyclization involving the acyclic
mixed acetal was not entirely site selective. That is, while the
major product obtained in this reaction was identical to terminal
olefin 3, the minor product, ethyl ether 15, was consistent with
cyclization after loss of the tertiary allylic alcohol of the acetal.
1
1:1 mixture of diastereomers, H NMR analysis detected only
a single diastereomer of 13 in the crude product mixture. Since
the yield of the reaction was greater than 50%, equilibration at
the acetal carbon leading to C10 was necessarily faster than
cyclization.9 In addition to a high level of diastereoselection,
complete site selectivity was observed.
Mild conditions were required for the conversion of the
primary alcohol to terminal olefin 3 in the presence of the
substitutionally and oxidatively labile dicobalthexacarbonyl
cluster (Scheme 6). The method of Grieco and Nicolaou proved
to be most compatible with this system.19,20 Accordingly,
o-nitrophenyl selenocyanate, when combined with tributylphos-
phine in dichloromethane, displaced the hydroxyl group in good
yield. A trace of a second compound, consistent with replace-
ment of one of the liganded CO molecules of the cobalt complex
with tributylphosphine, was detected in the product mixture.
Oxidation and elimination of the selenoether was effected with
the phenyloxaziridine of Davis under biphasic conditions, giving
olefin 3 in 61% overall yield for the two steps. A series of
NOE DIFF experiments was consistent with the configuration
of C10 shown, the same as that found at the corresponding
carbon in epoxydictymene.
Having demonstrated the feasibility of the Lewis acid-
mediated cyclization approach to the synthesis of the fused 5-8
system of epoxydictymene, we next explored a similar procedure
with racemic acyclic acetal 10 with the hope of providing a
more direct route to Pauson-Khand educt 3. The key question
in this case was whether the acyclic acetal would exhibit the
same high level of site and diastereoselectivity in the cyclization
reaction. As depicted in Scheme 7, the yield of coupled product
The Lewis acids that were tested are listed in Table 1, along
with the yield and site selectivity of each cyclization. Et2AlCl
(entry 1), which efficiently induced cyclization in the cyclic
acetal case, failed to afford any of the cyclized material under
the same conditions with the alkyne-cobalt cluster derived from
14. At higher temperatures, this Lewis acid did induce
cyclization with a 4:1 level of selectivity, but in a diminished
yield of 50%. The more potent titanium tetrachloride afforded
an uncyclized product that was consistent with hydrolysis of
the acetal to the aldehyde (1H NMR analysis). EtAlCl2,
intermediate in Lewis acidity, did effect fairly effective cy-
clization in several instances (entry 4), but with inconsistent
yields and varying amounts of the aldehyde side product. The
most consistent and efficient results for this reaction were
obtained with silyl triflates, with a striking solvent dependence
observed for TMSOTf. In dichloromethane (entry 5), a 2.3:1
ratio of desired 3 to undesired cyclization product 15 was
obtained, while the experiment conducted in THF afforded none
of the desired product, a low yield of 15, and the aldehyde as
the major product. With ether as solvent, the highest yield
(91%) and selectivity (10:1) were observed, and no aldehyde
side product was formed (entry 7). The steric bulk of the alkyl
groups on the silyl triflate had minimal effects on the selectivity
(19) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41,
1485.
(20) Grieco, P. A.; Jaw, J. Y.; Claremon, J. A.; Nicolaou, K. C. J. Org.
Chem. 1981, 46, 1215.

4356 J. Am. Chem. Soc., Vol. 119, No. 19, 1997
Jamison et al.
Scheme 8
Figure 2. Proposed transition structure in the Lewis acid-catalyzed
conversions of 12 to 13 (Scheme 5) and 15 to 3 (Scheme 7). The bonds
depicted with untapered bold lines are some of those involved in the
minimized g+/g- interactions. The orientation of the hydrogen atom
of C10 toward the incipient 8-membered ring both minimizes steric
interaction between the larger alkoxy substituent and the rest of the
molecule and affords the desired configuration of C10 in the product.
Of the methods for initiating intramolecular Pauson-Khand
reactions, thermal,13 oxidative,10 and ultrasonic27,28 conditions
proved to be effective in converting cobalt complex 3 into
tetracyclic enone 16 (Scheme 8). Earlier investigations in our
laboratory10 led to the development of a high yielding, room
temperature procedure that employs N-methylmorpholine N-
oxide in the oxidative initiation of cyclization, probably by
oxidative dissociation of one of the CO ligands of the cobalt
complex to CO2, vacating a coordination site for the olefin, and
ultimately leading to enone. Under these conditions, cobalt
complex 3 gave a 70% yield of the desired enone 16 as an 11:1
mixture of diastereomers at C12. NOE DIFF experiments
performed on the major enone product provided two insights
into stereochemistry. While the C12 configuration opposite to
that desired was favored in the Pauson-Khand reaction, these
NMR experiments supported the earlier assignment of config-
uration at C10 following Lewis acid cyclization. X-ray crystal-
lographic analysis of subsequent derivatives confirmed this
stereochemical assignment. Under an atmosphere of air, the
thermally induced Pauson-Khand cyclization of 3 was higher
yielding but less diastereoselective (5:1 at C12), again with the
undesired diastereomer prevailing. All attempts at epimerization
of C12 were unsuccessful.
Reductive and Deconjugative Methylation Attempts. With
a tetracyclic precursor of epoxydictymene in hand, there
remained four tasks for completion of the synthesis: face-
selective delivery of a hydrogen to C11, stereoselective instal-
lation of the angular methyl group at C1, inversion of the C12
configuration, and reduction of the C14 carbonyl (Scheme 8).
The first strategy explored attempted to solve the first two of
these issues in a single step (Scheme 9). Dissolving metal
reduction of the enone followed by alkylation of the enolate
with iodomethane afforded tetracyclic ketone 17 in high yield.
A cis fusion of the 5-5 system of 17 was formed exclusively,
opposite to results usually observed in dissolving metal reduc-
tions of homologous 6-6 enones, which favor the trans-fused
product.29-32 While the fusion of these two rings to the
8-membered ring may also exert a reinforcing effect on the facial
selectivity of this reduction, formation of the cis fusion is
generally observed in dissolving metal reductions of similar 5-5
enones.33,34 Of greater importance to the synthesis of the natural
product was that the methyl group attached to C1 had been
installed in the incorrect configuration.
and yield in reactions performed in dichloromethane (entries 8
and 9) suggesting that triflic acid, generated in the reaction by
trace amounts of water, might be effecting catalysis in the
conditions screened. Triflic acid did indeed promote cyclization,
but not in the same yield or selectivity, affording a 1:1:1 mixture
of desired product, undesired cyclization product, and aldehyde
under otherwise identical reaction conditions (entry 10). As it
was the highest yielding and most site-selective combination,
TMSOTf in ether was used in all subsequent cyclizations.
The cyclizations of cobalt complexes derived from 12 and
14 illustrate the interplay of two competing factors that
determine which alkoxy group is preferentially removed by the
Lewis acid. In the acyclic acetal case, the ethoxy oxygen is
likely to be more accessible for complexation, but the C10-O
bond of the tertiary alkoxy substituent is probably longer and
therefore weaker than the C10-OEt bond.21 In the cyclizations
of the cobalt complex derived from 12 (Scheme 5), where
complete site selectivity was observed, the cyclic nature of the
acetal suggests the appealing explanation that oxygen acces-
sibility is the more important factor by far. However, several
groups have studied addition reactions of acyclic and cyclic
acetals and found that the mechanism of the reaction is
dependent on Lewis acid, nucleophile, the structure of the acetal,
and solvent.22-26
An analysis of the high level of diastereoselectivity observed
in these cyclizations is depicted in Figure 2. Two critical factors
appear to dominate this aspect of the reaction. Minimization
of steric interactions in the transition state can be achieved with
the conformation shown. Also of importance is the fluxional
nature of such cationic cobalt complexes. In intermolecular
Lewis acid-mediated Nicholas reactions, equilibration occurs
at the propargylic center (C10 in this case) via what is thought
to be a fluxional cationic intermediate, and thus any stereo-
chemical information at this site in the substrate is lost upon
treatment with a Lewis acid.9 Assuming that this equilibration
is operative in this intramolecular process and that it is fast
relative to cyclization, then the stereochemical induction
observed in the cyclization may be representative of the ratio
of the activation energies of the two diastereomeric cyclization
pathways (Curtin-Hammett principle). Positioning of the
hydrogen atom at C10 toward the interior of the cyclizing system
and the larger alkoxy substituent away from this forming ring
leads to the desired configuration at C10.
(27) Billington, D. C.; Helps, I. M.; Pauson, P. L.; Thomson, W.;
Willison, D. J. Organomet. Chem. 1988, 354, 233.
(28) Bladon, P.; Brunner, P. L.; Eder, R. J. Organomet. Chem. 1988,
358, 845.
(29) Stork, G.; Darling, S. D. J. Am. Chem. Soc. 1960, 82, 1512.
(30) Stork, G.; Tsuji, J. J. Am. Chem. Soc. 1961, 83, 2783.
(31) Stork, G.; Rosen, P.; Goldman, N. L. J. Am. Chem. Soc. 1961, 83,
2965.
(21) Schreiber, S. L.; Wang, Z.; Schulte, G. Tetrahedron Lett. 1988, 29,
4085.
(22) Denmark, S. E.; Willson, T. M.; Almstead, N. G. J. Am. Chem.
Soc. 1989, 111, 9258.
(23) Mori, I.; Ishihara, K.; Flippin, L. A.; Nozaki, K.; Yamamoto, H.;
Bartlett, P. A.; Heathcock, C. H. J. Org. Chem. 1990, 55, 6107.
(24) Denmark, S. E.; Almstead, N. G. J. Am. Chem. Soc. 1991, 113,
8089.
(32) Review: Caine, D. Org. React. 1976, 23, 1.
(33) Markgraf, J. H.; Staley, S. W.; Allen, T. R. Synth. Commun. 1989,
19, 1471.
(25) Sammakia, T.; Smith, R. S. J. Am. Chem. Soc. 1992, 114, 10998.
(26) Sammakia, T.; Smith, R. S. J. Am. Chem. Soc. 1994, 116, 7915.
(34) Agnel, G.; Owczarczyk, Z.; Negishi, E. Tetrahedron Lett. 1992,
33, 1543.

Co-Mediated Reactions to Synthesize (+)-Epoxydictymene
J. Am. Chem. Soc., Vol. 119, No. 19, 1997 4357
Scheme 9
Figure 3. The predicted spatial relationship of C1 and C12 in
epoxydictymene’s ABD ring system (abbreviated for clarity), and two
modes (syn and anti) of connecting them.
Scheme 10
prepared by carboxymethylation of an enolate derived from the
enone, following the procedure of Mander.35 The aim was to
generate a dianion derived from 19, with the most reactive
position being C1. Again, deconjugative methylation was
elusive.
It was difficult to imagine an expedient route to epoxydic-
tymene involving reductively alkylated ketone 17, but this
undesired product did serve a purpose. Reduction of the
carbonyl to methylene by a three-step process provided a
tetracyclic product epimeric to epoxydictymene at C1 and C12.
Di-epi-epoxydictymene 18 was useful in gaining familiarity with
the physical properties and chromatographic behavior of these
systems. Since an authentic sample of the natural material was
not available, the technical knowledge gleaned from 18 facili-
tated later isolation and purification of epoxydictymene.
The results of the reductive and deconjugative methylation
studies suggested that alkylation at C1 with the desired facial
preference could be achieved only after inversion of the
configuration at C12. This inference was supported by molec-
ular modeling. As summarized in Figure 3, connecting C1 and
C12 in a syn manner with two carbon atoms results in a ring
system that is considerably less strained than that obtained by
bridging these two positions in an anti manner. Despite the
strained trans-5-5 system that one of the syn connection modes
yields, plastic models of the “in-out” 36,37 8-5-5 system
suggest that the anti connection is much more tenuous.
Ring-Opening/Reclosure Strategy: Radical and Anionic
Cyclizations. Since all attempts at inverting the configuration
of C12 with all rings of the tetracyclic skeleton intact were
unsuccessful, we next considered an approach that opened the
C ring and interconverted functional groups in such a way to
allow reclosure of this ring with the correct configurations at
all sites. This series of experiments began with Pauson-Khand
product 16 (Scheme 10). Although reductive methylation of
this enone installed the C1 methyl with the incorrect configu-
ration (Scheme 9), this process did afford the correct configu-
ration at C11, the 8-5-5 fusion carbon. Thus, Li/NH3
reduction of this enone, followed by an isoprene/NH4Cl quench,
gave ketone 21 in good yield. Omission of the isoprene step
Another strategy seeking simultaneous solution to two
remaining tasks was next explored (Scheme 9). The general
plan was to methylate at C1 while preserving the overall
oxidation state of the system. While the enolate formed in
dissolving metal reduction of enone 16 was alkylated exclusively
from the top face (as drawn), model building and molecular
modeling were consistent with the notion that the dienolate
derived from enone 16 (by γ-deprotonation at C12) would
exhibit the opposite facial selectivity, from the bottom face of
the molecule. Following this deconjugative alkylation, isomer-
ization of the olefin into conjugation with ketone would lead
to enone 20, with the configuration of C11 predicted to favor
the desired stereochemistry based on examination of molecular
models. Further speculation proposed that reduction of 20
would favor construction of the trans-fused 5-5, in contrast to
the reduction of enone 16. Unfortunately, all conditions
explored (e.g., KH, THF; MeI) for the first transformation
(deconjugative methylation) were unsuccessful, affording either
products of alkylation at C13 or decomposition, and thus this
expedient solution could not be investigated further. In order
to promote alkylation at C1 rather than C14, compound 19 was
(35) Mander, L. N.; Sethi, S. P. Tetrahedron Lett. 1983, 24, 5425.
(36) Isolation and structure of ingenol, which contains an “inside-
outside” bridged ring system: Zechmeister, K.; Brandl, F.; Hoppe, W.;
Hecker, E.; Opferkuch, H. J.; Adolf, W. Tetrahedron Lett. 1970, 11, 4075.
(37) For a review on the “inside-outside” bridged ring system: Alder,
R. W. Acc. Chem. Res. 1983, 16, 321.

4358 J. Am. Chem. Soc., Vol. 119, No. 19, 1997
Jamison et al.
Scheme 11
Figure 4. Some of the cross peaks observed in a phase-sensitive
NOESY experiment performed on 24, enabling assignment of stereo-
chemistry at C11 and C12. The identity of each resonance was first
established with a DQF COSY experiment.
produced significant amounts of overreduction to the secondary
alcohol. Separation of C12 diastereomers produced in the
Pauson-Khand reaction was more efficient after the reduction
of these two enones to the ketones (minor diastereomer not
shown).
Opening of this ring and epimerization at C12 were effected
with a three-step sequence that began with double R-hydroxy-
lation and reduction of this keto diol with NaHB(OAc)3 to give
triol 22. The stereochemical arrangement shown is based on
hydroxylation from the less-hindered (top) face of the ketone
followed by directed reduction of the carbonyl.38 Oxidative
excision of C14 with lead tetraacetate gave a compound whose
C12 configuration could be inverted when subjected to DBU
in dichloromethane, giving keto aldehyde 22 as a 3:1 mixture
of diastereomers (minor diastereomer not shown). One recy-
cling of this minor diastereomer with DBU gave a 61% overall
yield of 23. Exclusion of oxygen (three freeze-pump-thaw
cycles) was required in this reaction for maximal yield.
Support for the stereochemical assignment of the major
diastereomer of 23 was provided by a combination of DQF
COSY and phase-sensitive NOESY NMR experiments of a
closely related derivative (Figure 4). Keto diene 24, prepared
by treatment of 23 with Tebbe’s methylenation reagent (Scheme
10), exhibited strong cross peaks in its NOESY spectrum
between the C3 and C11 protons, consistent with the assignment
shown for C11. Further, there was also a strong cross peak
between the C11 proton and those attached to only one of the
diastereotopic methyl groups (C19). The proton at C10 showed
strong cross peaks to the proton at C12 at those attached to the
other diastereotopic methyl (C20). Taken together, these data
are most consistent with the stereochemical assignments made
at C11 and C12 for compound 24 and therefore its precursor,
keto aldehyde 23. Noteworthy was that the Tebbe reagent
reacted only with the aldehyde carbonyl in 23. This difference
in reactivity of the two carbonyl groups of 23 was exploited in
subsequent strategies and was critical to the completion of the
synthesis of the natural product.
Scheme 12
the trans. Reductive debenzylation of this mixture afforded an
aldehyde which was immediately reduced with NaBH4 to the
primary alcohol. The desired iodide 26 was obtained from this
alcohol in a single step.
Although the Bailey protocol for anionic cyclization of
5-hexenyl iodides was successful in our hands on simpler
systems, exposure of iodide 26 to these conditions resulted in
the exclusive formation of protodehalogenation product 27. No
cyclized material could be detected in the reaction mixture
(Scheme 12). Since another method commonly used in the
synthesis of five-membered rings is the 5-exo-trig cyclization
of the hexenyl radical,42 iodide 26 was also subjected to such
conditions. In contrast to the metal-halogen exchange condi-
tions, cyclization did occur upon addition of a solution of tris-
(trimethylsilyl)silane43,44 and AIBN to a refluxing benzene
solution of the iodide. However, NMR analysis (1H, 13C, DEPT
90, DEPT 135) of the reaction product revealed that the product
did not exhibit the appearance of the methyl singlet characteristic
of epoxydictymene and was most consistent with the product
of 6-endo cyclization, 5-8-6-5 tetracycle 28. In retrospect,
this result is not surprising since 6-endo products are often
favored over the 5-exo products in radical cyclizations of
5-alkyl-5-hexenyl iodides.42
On the basis of results of studies of the cyclization of
5-hexenyllithium to (cyclopentyl)methyllithium,39 Bailey and
co-workers reported an elegant method for the synthesis of trans-
bicyclo[3.3.0]octanes.40,41 A suitable substrate for this trans-
formation in the context of the synthesis of epoxydictymene
was prepared from keto aldehyde 23 in five steps by the
sequence shown in Scheme 11. Selective Wittig olefination of
the aldehyde with the phosphorane derived from benzyl bro-
momethyl ether provided a compound whose ketone, while
resistant to methylenation by the Tebbe reagent, was smoothly
olefinated with the reagent system of Lombardo (Zn, CH2Br2,
TiCl4) to give 25 as a 6:1 mixture of olefin isomers, favoring
The 6-endo cyclization of iodide 26 prompted examination
of another substrate in which cyclization could occur only at
the desired C1. Iodide 30 was thus prepared by means of a
six-step sequence whose key transformation was lithium di-
(38) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.
1988, 110, 3560.
(39) Bailey, W. F.; Patricia, J. J.; DelGobbo, V. C.; Jarret, R. M.; Okarma,
P. J. J. Org. Chem. 1985, 50, 2000.
(40) Bailey, W. F.; Khanolkar, A. D. Tetrahedron Lett. 1990, 31, 5993.
(41) Bailey, W. F.; Khanolkar, A. D.; Gavaskar, K. V. J. Am. Chem.
Soc. 1992, 114, 8053.
(42) Review: Curran, D. P. Radical Cyclization and Sequential Radical
Reactions; Curran, D. P., Ed.; Pergamon Press: New York, 1991; Vol. 4.2,
pp 779-831.
(43) Chatgilialoglu, C.; Griller, D.; Lesage, M. J. Org. Chem. 1988, 53,
3642.
(44) Kulicke, K. J.; Giese, B. Synlett 1990, 91.

Co-Mediated Reactions to Synthesize (+)-Epoxydictymene
J. Am. Chem. Soc., Vol. 119, No. 19, 1997 4359
Scheme 13
Figure 5. NOE DIFF experiments of the major isomer of 32: (A)
Irradiation of the C11 proton resulted in enhancement of the C3, C13,
and C19 protons. (B) Irradiation of the C10 resonance enhanced the
C12 and C20 protons, and one of the C17 protons. These data are most
consistent with stereochemical assignment shown.
Scheme 14
Scheme 15
methylcuprate displacement of a vinyltriflate45 at C1 (Scheme
13). Modeling of this compound predicted that C2 would be
much more distant than C1 and therefore inaccessible to an
approaching carbon centered radical at C14. Indeed, generation
of this radical under the conditions used in the previous case
(Scheme 12) confirmed that C2 was not the preferred site of
attack. However, cyclization was not detected in this experiment
either. A major product (31) appeared to be the result of
generation of a radical at C14, a 1,6 transfer of one of the
hydrogens at C15 to C14, and trapping of the allylic radical by
a species generated in the fragmentation of the radical initiator
AIBN. Thus, C15 was again the most accessible site to the
approaching C14 radical. In fact, the second-lowest energy
conformation (within 0.3 kcal of the global mininum, Macro-
model, Monte Carlo, MM2) of 32, seen in trace amounts in the
crude reaction mixture, predicted that the the distance between
C14 and a proton attached to C15 was approximately equal to
that between C14 and C1.
Adjustment of the electronic characteristics of the C1-C15
olefin in 26 in order to bias cyclization in favor of a 5-exo mode
rather than a 6-endo mode was next investigated. Attachment
of a cyano group to C15 was attractive, for it provided the
desired electronic bias for both radical46 and anionic cyclizations.
Further, the cyano group could be removed following cyclization
in a single reductive step.47 Preparation of the requisite
acrylonitrile was accomplished in 5 steps from keto aldehyde
23 (Scheme 14).
case was that derived from chloromethyl methyl ether, as
selective reduction of a benzyl enol ether in the presence of the
acrylonitrile and the exocyclic olefin was unlikely. Wads-
worth-Emmons olefination of the C1 ketone with diethyl
(cyanomethyl)phosphonate under forcing conditions (toluene,
reflux, 24 h), afforded 33 in 74% overall yield as an 8.5:1
mixture of olefin isomers (enol ether), favoring the trans. Only
one configuration of the acrylonitrile could be detected.
Due to the harshness of the Wadsworth-Emmons conditions,
the major isomer of 33 was examined with NOE DIFF
experiments to confirm that epimerization at C11 had not
occurred. These NMR experiments, summarized in Figure 5,
exhibited NOE behavior analogous to that seen previously on
a similar system (24, Figure 4). Acid-catalyzed hydrolysis of
the enol ether liberated an aldehyde that was converted with
two reactions into the cyclization substrate, iodide 34.
Wittig olefination of the aldehyde was again the first step
(cf. Schemes 11 and 13), but the phosphorane of choice in this
(45) Scott, W. J.; McMurry, J. E. Acc. Chem. Res. 1988, 21, 47.
(46) Giese, B. Angew. Chem., Int. Ed. Engl. 1983, 22, 753.
(47) Ohsawa, T.; Kobayashi, T.; Mizuguchi, Y.; Saitoh, T.; Oishi, T.
Tetrahedron Lett. 1985, 26, 6103.
Scheme 15 illustrates the two cyclizations that were ac-
complished with 34. Despite the cyano group, radical cycliza-
tion conditions still effected a 6-endo cyclization, in the same

4360 J. Am. Chem. Soc., Vol. 119, No. 19, 1997
Jamison et al.
resulting mixture was allowed to warm slowly to 0 °C over 1 h, during
which time a white precipitate had formed and the mixture had paled
to light yellow. Quenched with saturated NaHCO3 (30 mL) and diluted
with ether (300 mL), the mixture showed no residual color. The
organics were then separated from the heavily salted aqueous layer
and washed with water (25 mL) and brine (3 × 20 mL). The ethereal
solution was dried (MgSO4), filtered, and carefully concentrated in
Vacuo.
maner as that observed in the case lacking the cyano group
(Scheme 12). Reductive decyanation47 of cyclization product
35 produced 5-8-6-5 system 36. Surprisingly, this compound
was not identical to 28. 1H, 13C, DEPT 90, and DEPT 135
NMR experiments supported the connectivity shown, and since
epimerization at any site is unlikely in the three steps (33 to
34, Scheme 14) following stereochemical assignment, 36 is most
likely epimeric to 28 at C1. The factors governing this divergent
stereoinduction are unclear.
The resulting yellow oil was dissolved in MeOH (24.0 mL), and
treated with K2CO3 (2.76 g, 20.0 mmol). After 15 min TLC analysis
indicated complete hydrolysis of the silyl ether. The solution was
diluted with ether (200 mL), washed with pH 7.00 buffer (10 mL).
The aqueous layer was extracted with ether (50 mL). The combined
organics solutions were dried (MgSO4), filtered and concentrated in
Vacuo. Semipreparative HPLC (4:1 hexanes/methyl tert-butyl ether)
produced 2.15 g (50%) of 5 as a colorless oil, along with 1.10 g (36%)
of the starting alcohol 9b (Rf ) 0.4 in 4:1 hexane/EtOAc): [R]23D -73.5°
While the electronic influence of the cyano group was not
sufficient to favor the strained trans-5-5 in a radical cyclization,
it proved sufficient in an anionic cyclization. Metal-halogen
exchange of iodide 34 effected a cyclization in 74% yield. As
1
the H and 13C NMR spectra of this compound were nearly
identical to those reported for epoxydictymene (except for the
differences expected for substitution of CN for H), the structure
shown for 37 having the complete 5-8-5-5 skeleton of
epoxydictymene was assigned to this material.
1
(c 0.0264, CHCl3); IR (film) 3327, 3082, 2951, 2868 cm-1; H NMR
(400 MHz, CDCl3) δ 4.70 (s, 1H), 4.64 (s,1H), 3.48 (br m, 1H), 3.36
(br m, 1H), 2.43 (overlapping dd, J1 ) J2 ) 7.8, J3 ) 10.3, 1H), 1.92-
1.50 (m, 7H), 1.07 (m, 1H), 1.01 (d, J ) 6.4, 3H), 0.00 (s, 9H); 13C
NMR (100 MHz, CDCl3) δ 149.2, 106.8, 64.1, 50.6, 48.5, 36.1, 33.5,
29.5, 28.3, 21.7, -1.4; HRMS (EI) m/z calcd for M+ 226.1573, found
226.1752.
Final confirmation that this cyclization had occurred in the
desired 5-exo mode was realized by completion of the synthesis
of epoxydictymene. Reductive decyanation47 of 37 provided
synthetic (+)-epoxydictymene (1), whose spectral and physical
properties were identical in every respect with those reported
for the natural material (1H and 13C NMR, HRMS, IR, [R]D).1
The Lewis acid-promoted Nicholas reaction and subsequent
intramolecular Pauson-Khand reaction successfully assembled
three of the four rings of epoxydictymene, and an anionic
conjugate addition provided a means to construct the fourth.
These efficient transformations of dicobalthexacarbonyl clusters
with a bridging alkyne ligand await further use in the synthesis
of complex carbocyclic ring systems.
2-Ethynyl-4,4-dimethyl-1,3-dioxane (11). A mixture of propiol-
aldehyde diethyl acetal (39 mmol, 5 g, 5.6 mL), 3-methyl-1,3-butanediol
(58.5 mmol, 6.1 g, 5.4 mL), and p-toluenesulfonic acid (1.95 mmol,
371 mg) in benzene (120 mL) was heated at reflux under an atmosphere
of nitrogen for 4 h. The resulting pale yellow solution was allowed to
cool to room temperature and diluted with 120 mL of ether. The
mixture was washed with saturated NaHCO3 (25 mL) and brine (3 ×
20 mL). The aqueous layers were combined and extracted with ether
(3 × 20 mL). The combined organic layers were dried (MgSO4),
filtered, and carefully concentrated in Vacuo (note: 11 is volatile) to
give a yellow, pungent oil. Silica gel chromatography of this residue
(1:2 pentane/CH2Cl2) gave 4.6 g (33 mmoL, 84%) of pure 11 as a
clear oil. Ku¨gelrohr distillation (200 °C; 1 atm) provided analytically
pure 11 (3.6 g, 26 mmoL; 66%; Rf ) 0.24 in 1:2 pentane/CH2Cl2): IR
(thin film/NaCl) 3266, 2131, 1738 cm-1; 1H NMR (400 MHz, CDCl3)
δ 5.41 (d, J ) 1.3 Hz, 1H), 3.95 (ddd, J ) 11.8, 5.5, 1.7 Hz, 1H), 3.85
(apparent dt, J ) 12, 2.8 Hz, 1H), 2.49 (d, J ) 1.3 Hz, 1H), 1.89
(apparent dt, J ) 12, 5.5 Hz, 1H), 1.30 (obscured m, 1H), 1.28 (s,
3H), 1.26 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 84.45, 79.53, 72.90,
72.36, 63.14, 35.35, 31.15, 21.37.
Experimental Section
Unless otherwise noted, all reactions were performed in oven-dried
or flame-dried glassware, sealed with a rubber septum or a ground-
glass stopper, under an atmosphere of dry nitrogen. Nitrogen was
bubbled through concentrated sulfuric acid, then passed over a tower
of KOH/Drierite (CaSO4), and supplied through a glass manifold.
Solvents and reagents were transferred via oven-dried syringes, under
a positive N2 pressure.
Dicobalt octacarbonyl was exclusively purchased from Strem
Chemicals, as other suppliers’ stock of this reagent was consistently
contaminated with oxidation byproducts. Ether, hexanes, and THF were
distilled from a deep blue or purple solution over benzophenone ketyl.
Dichloromethane (CH2Cl2), acetonitrile, and TMSCl were distilled from
CaH2. All other reagents were used as supplied by commercial
manufacturers without any further purification unless specifically noted.
Thin layer chromatography (TLC) was performed on analytical (0.25
mm) silica gel 60 plates coated with a 254 nm fluorescent indicator
from E. Merck. Compounds were visualized using a Mineralight
ultraviolet lamp (254 nm), or by staining with a p-anisaldehyde (in
EtOH) or a cerium ammonium molybdate (aqueous) stock solution.
Semipreparative scale HPLC was performed at 380 psi on a Septec
Novaprep 5000 instrument using an ultraviolet detector, and a gradient
of HPLC-grade hexanes and tert-butyl methyl ether through a silica
100 A column (51 × 250 mm).
1-Ethoxy-1-(2-methyl-3-buten-2-oxy)-2-propyne (10). To a solu-
tion of propioaldehyde diethyl acetal (2.56 g, 2.9 mL, 20.0 mmol) in
CH2Cl2 (200 mL) at -78 °C was added bromodimethylborane (4.83
g, 3.9 mL, 40.0 mmol) dropwise, and the resulting solution was allowed
to warm to -40 °C over the course of 2 h. Diisopropylethylamine
(6.46 g, 8.7 mL, 50.0 mmol) was added, followed by 2-methyl-3-buten-
2-ol (5.17 g, 6.3 mL, 60.0 mL). The solution was allowed to warm to
0 °C over 2 h, stirred at this temperature for an additional 2 h, and
poured into a mixture of THF (100 mL) and saturated aqueous NaHCO3
(100 mL). After addition of 200 mL ether, the mixture separated into
two layers, and the organic layer was washed in succession with 25
mL each of saturated aqueous NaHCO3, water, and saturated aqueous
NaCl. The combined aqueous layers were extracted with 50 mL ether.
The combined organic solutions were dried (MgSO4), filtered, and
concentrated in Vacuo (10 is volatile.). The crude product mixture was
purified by silica gel chromatography (20:1 pentane/ether) to give 10
(1.8 g, 54%, Rf ) 0.52 in 10:1 hexane/ethyl acetate) as a clear oil: IR
(film) 3943, 3299, 2980, 2934, 2890, 2122 cm-1; 1H NMR (400 MHz,
CDCl3) δ 5.87 (dd, J ) 10.7, 17.6, 1H), 5.31 (d, J ) 1.7, 1H), 5.20 (d,
J ) 17.6, 1H), 5.15 (d, J ) 10.7, 1H), 3.69 (overlapping dq, J ) 2.1,
7.1, 2H), 2.49 (d, J ) 1.7, 1H), 1.35 (s, 3H), 1.32 (s, 3H), 1.19 (t, J )
7.1, 3H); 13C NMR (100 MHz, CDCl3) δ 142.8, 114.7, 87.0, 80.5, 77.7,
72.7, 58.9, 27.6, 26.0, 15.1; HRMS (EI) m/z calcd for [M + NH4]+
186.1494, found 186.1489.
1(S)-(Hydroxymethyl)-2(R)-[3-(trimethylsilyl)isopropenyl]-5(R)-
methylcyclopentane (5). A solution of sublimed potassium tert-
butoxide (6.73 g, 60.0 mmol) in dry hexane (30 mL) was cooled under
an argon atmosphere to -20 °C, and neat 9b15 (3.08 g, 20.0 mmol)
was added dropwise. After 5 min of vigorous stirring, a solution of
n-BuLi (33.3 mL, 1.8 M, 60.0 mmol) was added dropwise. The reaction
was allowed to warm to 0 °C over 30 min, during which time the
heterogeneous mixture became pale yellow and then deep orange. After
1.5 h at 0 °C nearly all solids had dissolved, but there was no change
in overall color. The reaction was cooled to -78 °C and treated with
freshly distilled TMSCl (6.52 g, 7.6 mL, 60.0 mmoL), dropwise, in a
manner such that the reaction temperature did not exceed -60 °C. The
1(S)-[4-Ethoxy-4-(2-methyl-3-buten-2-oxy)butyn-1-yl]-2(R)-[3-
(trimethylsilyl)isopropenyl]-5(R)-methylcyclopentane (14). A mix-
ture of allysilane 9b (113 mg, 0.50 mmol) and 2,6-di-tert-butyl-4-
methylpyridine (133 mg, 0.65 mmol) in CH2Cl2 (8 mL) was cooled to

Co-Mediated Reactions to Synthesize (+)-Epoxydictymene
J. Am. Chem. Soc., Vol. 119, No. 19, 1997 4361
-30 °C and treated, dropwise, with trifluoromethanesulfonic anhydride
(155 mg, 0.093 mL, 0.55 mmol). The reaction was warmed to -10
°C and stirred for 10 min. Gradually, a white solid precipitated. The
cloudy suspension was diluted with dry hexanes to triturate any salts
remaining in solution, and the mixture was concentrated in Vacuo with
particular care to avoid bumping. The nonvolatile residue was
resuspended in hexanes and filtered through a glass frit, resulting in a
clear solution of the desired triflate.
subsequent recrystallization: Rf ) 0.37 in 4:1 hexane/EtOAc; [R]23
D
+85° (c 0.020, CHCl3); IR (film) 3490, 2939, 2861, 1707 cm-1; H
1
NMR (400 MHz, CDCl3) δ 5.15 (br s, 1H), 4.94 (br s, 1H), 4.32 (m,
1H), 2.95 (m, 1H), 2.81 (dd, J ) 5.9, 11.3, 1H), 2.58-2.52 (m, 2H),
2.30 (m, 1H), 2.09 (dd, J1 ) J2 ) 11.3, 1H), 2.07-1.92 (m, 3H), 1.75-
1.50 (m, 4H), 1.44-1.36 (m, 2H), 1.42 (s, 3H), 1.01 (obscured m, 1H),
1.00 (d, J ) 6.3, 3H), 0.96 (s, 3H); 13C NMR (100 MHz, CDCl3) δ
208.1, 182.9, 149.1, 135.7, 116.3, 81.6, 76.3, 52.2, 52.1, 48.3, 46.7,
38.4, 37.7, 34.2, 33.1, 27.5, 22.6, 19.9, 18.3; HRMS (FAB) m/z calcd
for M+ 286.1927, found 286.1922.
1,11-Dehydro-1-desmethyl-12-epi-14-oxoepoxydictymene (16)
(NMO-Promoted Pauson-Khand). A solution of cobalt complex
14 (30 mg, 55 mmol) in CH2Cl2 (10 mL) was treated with a single
portion of solid N-methylmorpholine N-oxide (39 mg, 0.33 mmol) at
room temperature. After 12 h a purple precipitate had formed and TLC
analysis indicated consumption of all starting material. The mixture
was passed through a small plug of silica gel and the filtrate was
concentrated in Vacuo. The resulting yellow oil was purified by SGC
(3:1 hexane/ether) to give 11 mg (38 mmol, 70%) of a solid, shown to
be an 11:1 mixture of the diastereomeric enones 16 and 2 by spectral
analysis.
Acetal 10 (168 mg, 0.189 mL, 1.0 mmol) in dry THF (20 mL) was
cooled to -78 °C and treated dropwise with a 1.9 M solution of n-BuLi
(0.79 mL, 1.5 mmol). HMPA (3.58 mmol, 0.35 mL, 2.0 mmol) was
added, and the mixture was stirred 1 h, warmed to -35 °C, and kept
at this temperature for 30 min. In the meantime, the triflate solution
described above was concentrated in Vacuo, dissolved in THF (6 mL),
and added dropwise via cannula to the acetylide solution at -35 °C.
This solution was allowed to warm to room temperature over 1 h. The
progress of the coupling reaction was monitored by TLC analysis, as
the spot corresponding to the nonpolar triflate (or a decomposition
product thereof) slowly transformed to a more polar, lower Rf spot that
was distinct from the starting acetal 10. When TLC revealed no further
progress (after 40 min at room temperature), the reaction was quenched
with 1:1 saturated NaHCO3/H2O (5 mL) and diluted with ether (20
mL). The organic layer was washed with brine (3 × 5 mL), and the
combined aqueous layers were extracted with ether (2 × 10 mL). The
combined organic layers were dried (MgSO4), filtered, and concentrated
in Vacuo. SGC of the crude oil (20:1 hexanes/EtOAc) gave 156 mg
(82%) of pure 14 (Rf ) 0.47 in 10:1 hexane/EtOAc) (single diastere-
omer): IR (film) 2953, 2869, 2238 cm-1; 1H NMR (400 MHz, CDCl3)
δ 5.87 (dd, J ) 10.7, 17.6, 1H), 5.30 (s, 1H), 5.18 (d, J ) 17.6, 1H),
5.11 (d, J ) 10.7, 1H), 4.57 (s, 1H), 4.52 (s, 1H), 3.65 (overlapping
dq, J ) 2.0, 7.1, 2H), 2.43 (m, 1H), 2.08-1.77 (m, 5H), 1.61-1.54
(m, 2H), 1.58 (d, J ) 13.4, 1H), 1.39 (d, J ) 13.4, 1H), 1.33 (s, 3H),
1.30 (s, 3H), 1.17 (t, J ) 7.1, 3H), 1.10 (m, 1H), 1.03 (d, J ) 6.7, 3H),
-0.01 (s, 9H); 13C NMR (100 MHz, CDCl3) δ 147.1, 143.2, 114.2,
107.5, 87.6, 85.7, 77.2, 58.5, 48.9, 46.9, 38.1, 32.9, 28.1, 27.7, 27.5,
26.0, 22.6, 20.4, 15.2, -1.3; HRMS (EI) m/z calcd for M+ 376.2798,
found 376.2812.
1-Desmethyl-14-oxo-1,12-di-epi-epoxydictymene (21). Ammonia
(15 mL) was distilled from sodium into a dry three-necked flask
equipped with a dry ice condenser, a three-way stopcock, and a rubber
septum. Lithium (14 mg, 2.0 mmol) was added and the mixture was
stirred at -78 °C until the deep blue color was uniformly dispersed.
The mixture was treated with a THF solution of 16 (110 mg, 0.38 mmol
in 6 mL of THF) dropwise, via cannula. After 15 min at -78 °C, the
reaction was quenched by addition of 1.5 mL of isoprene, at which
point the solution lost its blue color, turning light brown at first, and
then into a clear pink. Saturated ammonium chloride (8 mL) was added
with caution; the mixture was diluted with ether (35 mL) and then
allowed to warm to room temperature. The clear solution was stirred
for 4 h at room temperature, open to air inside a fume hood, until excess
ammonia had evaporated. The remaining biphasic mixture was poured
into a separatory funnel containing 6 mL of 6 N HCl and 30 mL of
ether. The layers were separated, and the organic layer was washed
with saturated NH4Cl (10 mL) and then saturated NaHCO3 (2 × 10
mL). The aqueous layer was washed with two portions of ether (20
mL each). The combined organic layers were dried (MgSO4), filtered,
and concentrated in Vacuo to give a yellow oil. SGC (4:1 hexane/
EtOAc) gave three major fractions. The least polar fraction contained
10 mg (35 mmol, 9%) of the minor diastereomer (not shown in figure);
a highly crystalline solid (Rf ) 0.41 in 4:1 hexane/EtOAc). A second,
more polar fraction gave 65 mg (0.22 mmol, 60%) of the major
diastereomer 21, also as a crystalline solid (Rf ) 0.36 in 4:1 hexane/
EtOAc). Finally, the most polar fraction contained a mixture of 21
and starting materials 18 (12 mg, ∼11%). This fraction could be
purified with a second SGC purification, giving an additional 6% of
[3(S)-(1,1-Dimethylpropenoxy)-5-methylene-9(R)-methyl-(6R,-
10S)-bicyclo[6.3.0]undecyne]dicobalt Hexacarbonyl Complex (3)
(From 14). To a solution of 14 (1.56 g, 2.35 mmol) in ether (235
mL) at -78 °C was added with trimethylsilyl trifluoromethanesulfonate
(580 mg, 0.63 mL, 2.60 mmol. The deep red solution was allowed to
stir at this temperature for 15 min and poured into a mixture of saturated
NaHCO3 (100 mL) and ether (200 mL). The organic layer was washed
with saturated NaCl (100 mL), dried (MgSO4), filtered through silica,
and concentrated in Vacuo giving 1.26 g (98%) of a red oil. 1H NMR
analysis of this crude product revealed a 10:1 mixture of desired 3 to
ethyl ether 15, each as a single diastereomer. This material was used
in subsequent Pauson-Khand reactions without further purification:
IR (film) 2953, 2870, 2087, 2047 (vs), 2022 (vs) cm-1; 1H NMR (300
MHz, CDCl3) δ 5.97 (dd, J ) 8.8, 14.2, 1H), 5.20 (br d, J ) 14.2,
1H), 5.16 (br d, J ) 8.8, 1H), 5.04 (br s, 1H), 4.98 (br s, 1H), 4.62
(dd, J ) 3.6, 9.0, 1H), 2.79 (br dd, J ) 1.7, 9.0, 1H), 2.68 (obscured,
the major diastereomer: IR (thin film/NaCl) 2957, 2930, 1738 cm-1
;
1H NMR (400 MHz, CDCl3) δ 4.99 (s, 1H), 4.70 (s, 1H), 3.59 (apparent
dt, J ) 10.3, 3.9 Hz, 1H), 3.14 (dd, J ) 9.5, 7.9 Hz, 1H), 2.75 (dd, J
) 12, 3.9 Hz, 1H), 2.74 (m, 1H), 2.57 (m, 1H), 2.38 (apparent t, J )
10.4 Hz, 1H), 2.22 (dd, J ) 19.2, 10.4 Hz, 1H), 2.1-1.98 (m, 3H),
1.88 (m, 1H), 1.62-1.54 (m, 2H), 1.32 (s, 3H), 1.22 (s, 3H), 1.20 (m,
1H), 1.06 (d, J ) 6.3 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 218.78,
150.02, 114.93, 81.20, 80.17, 53.02, 52.37, 50.29, 49.44, 49.28, 47.62,
38.54, 38.45, 33.57, 32.79, 30.37, 26.31, 24.86, 18.34; HRMS (FAB/
1H), 2.66 (obscured, 1H), 2.54 (m, 1H), 1.96 (dd, J ) J2 ) 10.1,
1
1H), 1.85 (m, 1H), 1.76 (m, 1H), 1.68-1.50 (m, 5H), 1.36 (s, 3H),
1.34 (s, 3H), 1.06 (d, J ) 6.3, 3H); 13C NMR (100 MHz, CDCl3) δ
200.4, 146.6, 144.3, 114.0, 111.9, 76.3, 75.5, 50.4, 49.6, 45.9, 42.9,
42.8, 33.9, 30.0, 26.3, 26.1, 20.9, 20.7; MS (CI) m/z calcd for [M +
NH4]+ 562, found 562.
20
NaI) m/z calcd for C19H28O2Na 311.1987, found 311.1986; [R]D
+105.5° (c 0.015, CHCl3).
1,11-Dehydro-1-desmethyl-12-epi-14-oxoepoxydictymene (16)
(Thermal Pauson-Khand). A solution of 14 (1.1 g, 2.0 mmol) in
acetonitrile (200 mL) was brought rapidly to reflux under an atmosphere
of air and heated at that temperature for 15 min. The solution was
cooled to room temperature and filtered through layered Celite (top, 5
g) and silica gel (bottom, 5 g), eluting with ethyl acetate. A bluish
solid was retained by the Celite. Evaporation in Vacuo of the clear
solution afforded a yellow oil which was purified by silica gel
chromatography (4:1 hexanes/EtOAc) to give 486 mg (85%) of enone
16 as a 5:1 mixture of diasteromers at C12 (1H NMR analysis) that
were more efficiently separated after subsequent dissolving metal
reduction. A pure sample of the major diastereomer was prepared by
the NMO-promoted Pauson-Khand10 (see Supporting Information) and
1-Desmethyl-1,13,14-trihydroxy-1,12-di-epi-epoxydictymene (22).
A solution of KHMDS in toluene (740 mL of a 0.54 M solution, 0.4
mmol) was diluted with THF (2.5 mL), cooled to -78 °C, and treated
with a solution of ketone 21 (29 mg, 0.1 mmol) in 500 mL of THF
(rapid addition via cannula). HMPA (TOXIC!, 0.5 mL) was added
after 5 min to the pale yellow solution, and the mixture was stirred at
-78 °C for 2 h. A solution of the oxaziridine of Davis (130 mg, 0.5
mmol) in 1 mL of THF was then added via syringe and the solution
was stirred at -78 °C for an additional 20 min. The reaction was
quenched at -78 °C with saturated NaHCO3 (1 mL), then diluted with
EtOAc, and allowed to warm to room temperature. The layers were
separated, and the organic layer was washed with brine (2 × 5 mL)

4362 J. Am. Chem. Soc., Vol. 119, No. 19, 1997
Jamison et al.
and dried over MgSO4. Concentration in Vacuo gave a crude oil
(HMPA contaminated) that was dissolved in THF (3 mL), and treated
with Na(AcO)3BH (64 mg, 0.3 mmol). After 5 h the reaction was
quenched by addition of excess MeOH (2 mL) and a few drops of 1 N
NaOH. The mixture was concentrated in Vacuo, and the resulting thick
oil was passed through a short (5 cm long) plug of silica gel with 9:1
EtOAc/MeOH to remove any salts. A thick gel had formed after
concentration of the filtrate in Vacuo. The residue was purified by
SGC (3% then 5% then 10% MeOH in chloroform) to give 23 mg
(0.071 mmol, 71%) of the desired triol 22 (Rf ) 0.16 in 1:3 hexane/
EtOAc) along with 3 mg (9.3 mmol, 9%) of a minor diastereomer.
These could be combined for use in the next reaction: IR (thin film/
NaCl) 3372, 2951, 2868 cm-1; 1H NMR (400 MHz, CDCl3) δ 5.03 (s,
1H), 4.74 (s, 1H), 4.10 (dd, J ) 7.3, 4.5 Hz, 1H), 3.86 (s, 1H), 3.83
(apparent dt, J ) 9.7, 4.5 Hz, 1H), 2.76-2.68 (overlapping m, 3H),
2.43 (d, J ) 11.2 Hz, 1H), 2.05-1.95 (m, 1H), 2.02 (s, 1H), 1.96-
1.90 (m, 1H), 1.89 (s, 1H), 1.85 (d, J ) 3.8 Hz, 1H), 1.7-1.64
(overlapping m, 4H), 1.33 (overlapping s, 6H), 1.23 (m, 1H), 0.96 (d,
J ) 6.0 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 150.00, 114.48, 89.67,
82.36, 81.05, 80.98, 79.36, 62.33, 60.41, 49.24, 47.30, 45.50, 38.65,
33.40, 32.93, 32.59, 31.67, 24.91, 18.43; HRMS (FAB/NaI) m/z calcd
Acrylonitrile 33. To a solution of (methoxymethyl)triphenylphos-
phonium chloride (360 mg, 1.05 mmol) in THF (8.6 mL) at -78 °C
was added a solution of KHMDS in THF (0.6 M, 1.09 mL, 0.655
mmol). The resulting yellow solution was stirred at this temperature
for 30 min, and a solution of keto aldehyde 23 (38 mg, 0.131 mmol)
in THF (4.5 mL) was added. This mixture was stirred at this
temperature for 30 min, warmed to room temperature, and stirred at
room temperature 1.5 h. Hexanes (15 mL) were added, and the mixture
was filtered through a plug of Celite. The yellow oil afforded upon
evaporation of the filtrate was purified by silica gel chromatography
(4:1 hexanes/EtOAc), giving the desired methyl enol ether (35 mg,
83%) as an 8.5:1 mixture of olefin isomers. This material (25 mg,
0.0786 mmol) was added to a premixed solution of NaH (32 mg, 60%
suspension, 0.786 mmol) and diethyl (cyanomethyl)phosphonate (278
mg, 1.57 mmol, 0.254 mL) in toluene (16 mL). This solution was
heated at reflux for 48 h, cooled to room temperature and diluted with
water (10 mL) and EtOAc (10 mL). The organic phase was washed
with water (3 × 5 mL) and saturated NaCl (5 mL), dried (MgSO4),
filtered, and concentrated in Vacuo. Purification by SGC (4:1 hexanes/
EtOAc) afforded 33 (24.8 mg, 93%) as a clear oil as an 8.5:1 mixture
of olefin isomers (see above). Data for major olefin isomer: IR (film)
20
1
for C19H30O4Na 345.2042, found 345.2039; [R]D -4.5° (c 0.009,
CHCl3).
2953, 2868, 2215 cm-1; H NMR (400 MHz, CDCl3) δ 6.21 (d, J )
12.6, 1H), 5.05 (s, 1H), 4.88 (s, 1H), 4.87 (s, 1H), 4.44 (dd, J ) 9.8,
12.6, 1H), 3.62 (ddd, J1 ) 4.4, J2 ) J3 ) 10.3, 1H), 2.81 (dd, J ) 4.4,
12.3, 1H), 2.63 (dd, J ) 10.7, 11.2, 1H), 2.63 (obscured m, 1H), 2.53
(dd, J ) 12.3, 16.0, 1H), 2.36-2.32 (m, 2H), 2.02-1.90 (m, 4H), 1.85-
1.67 (m, 4H), 1.24 (s, 3H), 1.16 (m, 2H), 1.11 (s, 3H), 1.05 (d, J )
6.8, 3H); 13C NMR (100 MHz, CDCl3) δ 165.4, 149.4, 145.3, 117.1,
114.8, 99.4, 94.9, 85.0, 82.2, 56.4, 54.5, 53.3, 48.6, 46.1, 43.9, 40.0,
37.7, 32.1, 28.8, 27.9, 25.5, 21.4; HRMS (FAB) m/z calcd for [M +
Na]+ 364.2253, found 364.2269.
Keto Aldehyde 23. Triol 22 (72 mg, 0.226 mmol) was dissolved
in dry EtOAc (11 mL) and treated with solid Pb(OAc)4 (201 mg, 0.453
mmol) in one portion, at room temperature. After 5 min the canary
yellow mixture was poured over a plug of silica gel (5 cm long), and
eluted with EtOAc. The filtrate was concentrated in Vacuo to give a
clear, colorless oil. This was dissolved in CH2Cl2 (11 mL), and after
three freeze-pump-thaw cycles (argon), treated with DBU (0.100 mL,
0.68 mmol). The mixture was stirred at room temperature, under an
atmosphere of argon, for 24 h, diluted with EtOAc (25 mL), washed
with HCl (1 N, 5 mL), water (5 mL), and saturated NaCl (5 mL). The
organic solution was dried (Na2SO4), filtered, concentrated in Vacuo,
and purified by SGC (4:1 hexane/EtOAc) gave 32.1 mg (49%) of pure
23, along with 4 mg (12.9 mg, 20%) of a minor diastereomer (not
shown) (Rf ) 0.29 in 4:1 hexane/EtOAc). This minor component was
resubjected to the above conditions for further equilibration toward 23.
Following this step, a total of 40 mg (61%) of pure 23 (colorless oil)
and 4 mg (6%) of the minor diastereomer were obtained: [R]23D +15°
Iodoacrylonitrile 34. To a solution of 33 (23.0 mg, 0.0674 mmol)
in THF (6.9 mL) was added aqueous 1 N HCl (6.9 mL), and this
solution was stirred at room temperature for 36 h. The pale yellow
solution was diluted with water (10 mL) and EtOAc (10 mL), and the
organic phase was washed with saturated NaCl (3 mL). After extraction
of the aqueous phase with EtOAc (10 mL), the combined organics were
dried (MgSO4), filtered, and concentrated in Vacuo, giving 25 mg
(100%) of an aldehyde that was immediately dissolved in MeOH/EtOH
(1.7 mL, 5.1 mL), cooled to -78 °C, and treated with sodium
borohydride (26 mg, 0.674 mmol). After this mixture was stirred at
this temperature for 1 h, water (5 mL) and EtOAc (15 mL) were added,
and the aqueous layer was extracted with EtOAc (10 mL). The
combined organic solutions were dried (Na2SO4), filtered, concentrated
in Vacuo, and purified by silica gel chromatography (1:2 hexanes/
EtOAc) to give the desired primary alcohol (21 mg, 95%) as a clear
oil: [R]23D +250° (c 0.0043, CHCl3); IR (film) 3447, 2947, 2215, 1613,
1
(c 0.003, CHCl3); IR (film) 2951, 2868, 1720 cm-1; H NMR (500
MHz, CDCl3) δ 9.62 (br s, 1H), 4.96 (br s, 1H), 4.94 (br s, 1H), 3.66
(ddd, J1 ) 4.3, J2 ) J3 ) 9.7, 1H), 3.61 (dd, J1 ) J2 ) 9.7, 1H), 3.52
(d, J ) 9.7, 1H), 2.85 (dd, J ) 4.0,12.1, 1H), 2.75 (m, 1H), 2.41 (m,
1H), 2.36 (dd, J1 ) J2 ) 11.7, 1H), 2.15 (dd, J ) 3.5, 12, 1H), 2.0 (p
obs dd, J1 ) J2 ) 12, 1H), 2.0-1.9 (m, 1H), 1.8-1.64 (m, 4H), 1.54
(s, 3H), 1.21 (s, 3H), 1.17 (m, 1H), 1.10 (d, J ) 7.0, 3H); 13C NMR
(100 MHz, CDCl3) δ 207.0, 198.8, 144.0, 116.1, 82.4, 80.2, 64.3, 53.6,
49.2, 46.6, 43.7, 42.4, 40.9, 32.4, 29.7, 28.9, 26.5, 21.8; HRMS (FAB)
m/z calcd for [M + Na]+ 313.1780, found 313.1796.
Keto Diene 24. A solution of the keto aldehyde 23 (4 mg, 0.014
mmol) in THF (1 mL) was treated with the Tebbe reagent48 (0.050 mL
of a 0.4 M solution in toluene, 0.020 mmol) at -40 °C. After a total
of 1 h of stirring at -40 °C, the reaction was warmed to -25 °C and
quenched with one drop of 15% aqueous NaOH. The mixture was
diluted with ether and allowed to warm to room temperature. After
an additional hour of stirring, the mixture was pale yellow in color.
This was passed through a plug of silica gel, concentrated in Vacuo,
and purified by SGC (8:1 hexane/EtOAc), affording 3 mg (74%) of
24: IR (thin film/NaCl) 2928, 2866, 1701 cm-1; 1H NMR (400 MHz,
CDCl3) δ 5.58 (m, 1H), 5.01 (s, 1H), 4.99 (d, J ) 4.9 Hz, 1H), 4.92
(overlapping s, 2H), 3.64 (apparent dt, J ) 10.6, 4.4 Hz, 1H), 3.22
(apparent t, J ) 10.8 Hz, 1H), 3.13 (dd, J ) 10.8, 8.3 Hz, 1H), 2.85
(dd, J ) 12.3, 4.4 Hz, 1H), 2.69 (m, 1H), 2.37 (overlapping m, 2H),
2.08 (d, J ) 14.8 Hz, 1H), 1.94 (overlapping m, 2H), 1.74-1.64
(overlapping m, 3H), 1.30 (s, 3H), 1.17-1.10 (m, 1H), 1.10 (s, 3H),
1.09 (d, J ) 6.5 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 208.83,
143.67, 135.09, 117.53, 115.99, 83.71, 82.65, 58.93, 56.54, 49.54, 48.00,
43.54, 42.58, 40.94, 32.37, 28.69, 28.33, 26.16, 22.26; HRMS (FAB/
NaI) m/z calcd for C19H28O2Na 311.1987, found 311.1981.
1
1369, 1053, 897 cm-1; H NMR (400 MHz, CDCl3) δ 5.15 (d, J )
2.4, 1H), 4.89 (s, 1H), 4.81 (br s, 1H), 3.55 (obscured m, 1H), 3.54 (t,
J ) 6.9, 2H), 2.80 (dd, J ) 4.7,12.3, 1H), 2.76 (dd, J ) 12.5, 16.4,
1H), 2.56 (m, 1H), 2.48 (dd, J ) 9.6, 11.1, 1H), 2.30-2.00 (m, 5H),
1.85 (ddd, J1 ) 1.0, J2 ) J3 ) 12.3, 1H), 1.85-1.70 (m, 3H), 1.63
(dddd, J1 ) 6.7, J2 ) J3 ) 6.9, J4 ) 7.0, 1H), 1.53 (dddd, J1 ) J2 )
6.9, J3 ) J4 ) 7.2, 1H),1.33 (s, 3H), 1.30-1.15 (m, 2H), 1.15 (s, 3H),
1.07 (d, J ) 7.0, 3H); 13C NMR (100 MHz, CDCl3) δ 166.1, 144.1,
117.1, 114.5, 94.5, 87.0, 81.8, 60.9, 54.0, 50.5, 48.8, 43.8, 41.9, 40.1,
38.7, 32.5, 31.6, 29.4, 27.8, 25.2, 22.6; HRMS (FAB) m/z calcd for
[M + Na]+ 352.2252, found 352.2259.
To a solution of this alcohol (18.6 mg, 0.0565 mmol) in ether (3.3
mL) and acetonitrile (1.1 mL) at room temperature were added in the
order listed triphenylphosphine (44.5 mg, 0.170 mmol), imidazole (11.5
mg, 0.170 mmol), and iodine (43.0 mg, 0.170 mmol). The yellow
solution exothermed slightly, and a white precipitate developed over 5
min. The mixture was diluted with ether (15 mL), filtered through a
short (5 cm) plug of celite, washed with saturated sodium hydrogen
carbonate (5 mL) and saturated NaCl (5 mL), dried (MgSO4), filtered,
and concentrated in Vacuo. The resulting yellow oil was purified by
silica gel chromatography (10:1 hexanes/EtOAc) to give 34 (24.1 mg,
94%) as a clear oil.
15-Cyanoepoxydictymene (37). A frozen ether (0.9 mL) solution
of 34 (6.2 mg, 0.014 mmol) was degassed under high vacuum and
then cooled to -78 °C under argon. Stirring of the ethereal solution
(48) Tebbe, F. N.; Parshall, G. W.; Reddy, G. S. J. Am. Chem. Soc.
1978, 100, 3611.

Co-Mediated Reactions to Synthesize (+)-Epoxydictymene
J. Am. Chem. Soc., Vol. 119, No. 19, 1997 4363
was initiated with a magnetic stirring bar, and solution of t-BuLi in
pentane (0.031 mmol, 1.5 M, 0.021 mL) was then added. The solution
turned yellow immediately. After the mixture was stirred at this
temperature for 15 min, the stirring bar was stopped, and the solution
was warmed to room temperature and held at this temperature for 30
min. Water (0.1 mL) was added, and the mixture was diluted with
ether (5 mL). The organic phase was washed with saturated NaCl,
dried (MgSO4), filtered, and concentrated in Vacuo to give a yellow
oil that was purified with silica gel chromatography (10:1 to 4:1
11.5, 1H), 2.45 (ddd, J ) 7.8, 11.2, 13.6, 1H), 2.00 (dd, J ) 10.3,
13.7, 1H), 1.94 (ddd, J1 ) 4.0, J2 ) J3 ) 12.5, 1H), 1.75-1.60 (m,
4H), 1.60-1.40 (m, 3H), 1.40-1.16 (m, 4H), 1.24 (s, 3H), 1.15 (s,
3H), 1.02 (p obscured, 1H), 0.98 (d, J ) 6.5, 3H), 0.87 (s, 3H); 13C
NMR (100 MHz, CDCl3) δ 146.3, 114.5, 78.0, 74.8, 59.6, 57.5, 51.0,
48.4, 44.2, 43.9, 43.3, 41.5, 36.7, 34.7, 34.2, 29.3, 24.4, 24.1, 20.3,
19.2; HRMS (EI) m/z calcd for M+ 288.2455, found 288.2453.
Reported for natural product:1 [R]23 +72.1° (c 0.9, hexane); IR
D
(film) 1630, 1140, 999, 893 cm-1; 1H NMR (400 MHz, CDCl3) δ 4.98
(s, 1H), 4.77 (s, 1H), 3.57 (dt, J ) 7, 10, 1H), 2.82 (dd, J ) 7, 12,
1H), 2.64 (ddd, J ) 7, 10, 12, 1H), 2.48 (ddd, J ) 7, 11, 14, 1H), 2.03
(dd, J ) 10, 14, 1H), 1.26 (s, 3H), 1.17 (s, 3H), 1.01 (d, J ) 7, 3H),
0.89 (s, 3H); 13C NMR (25 MHz, CDCl3) δ 145.8, 114.1, 77.7, 74.6,
59.5, 57.4, 50.8, 48.3, 44.1, 43.8, 43.2, 41.4, 36.6, 34.6, 34.1, 29.2,
24.3, 24.0, 20.2, 19.1; HRMS (EI) m/z calcd for M+ 288.2455, found
288.2495.
hexanes/EtOAc) to give 37 (3.1 mg, 74%) as a clear oil: [R]23 +68°
D
1
(c 0.0020, CHCl3); IR (film) 2948, 2867, 2243 cm-1; H NMR (400
MHz, CDCl3) δ 5.00 (s, 1H), 4.87 (s, 1H), 3.52 (ddd, J1 ) 6.8, J2 )
J3 ) 9.8, 1H), 2.84 (dd, J ) 6.7,11.8, 1H), 2.65 (ddd, J ) 6.6, 9.8,
12.1, 1H), 2.45 (d, J ) 16.3, 1H), 2.33 (ddd, J ) 7.5, 11.1, 13.7, 1H),
2.25 (d, J ) 16.3, 1H), 2.15 (m, 1H), 2.02 (ddd, J1 ) 3.1, J2 ) J3 )
12.4, 1H), 1.79 (m, 1H), 1.70 (dd, J ) 10.3, 11.1, 1H), 1.60-1.20 (m,
9H), 1.25 (s, 3H), 1.17 (s, 3H), 1.07 (m, 1H), 1.01 (d, J ) 6.5, 3H);
13C NMR (100 MHz, CDCl3) δ 145.9, 119.0, 115.4, 78.0, 74.1, 59.3,
57.6, 50.9, 48.0, 43.9, 43.3, 41.8, 39.2, 39.0, 34.6, 34.0, 29.2, 26.7,
24.1, 20.3, 19.3; HRMS (EI) m/z calcd for M+ 313.2406, found
313.2407.
Acknowledgment. We thank the National Institute of
General Medical Sciences for support of this research and the
NSF for a predoctoral fellowship to T.F.J. Mass spectral data
were obtained at the Harvard University Mass Spectrometry
Facility. We acknowledge the NIH BRS Shared Instrumentation
Grant Program (1 S10 RR01748-01A1) and NSF (CHE88-
14019) for providing NMR facilities. Additional spectral and
physical data for epoxydictymene were kindly provided by Dr.
Fuyuhiko Matsuda (Hokkaido University, Sapporo, Japan) and
Dr. Mikiko Sodeoka (University of Tokyo, Hongo, Tokyo,
Japan). S.L.S. is an Investigator at the Howard Hughes Medical
Institute.
(+)-Epoxydictymene 1. To a suspension of potassium metal (2.6
mg, 0.064 mmol) in toluene (0.4 mL) at room temperature was added
a solution of 18-crown-6 (6.8 mg, 0.026 mmol) in toluene (0.1 mL).
To this mixture was added a solution of 35 (4.0 mg, 0.013 mmol) in
toluene (0.5 mL), and the heterogenous mixture was stirred at room
temperature for 2 h. During this time a fine white powder appeared,
and the liquid phase became pale yellow. The mixture was quenched
with i-PrOH (0.1 mL) and transferred into a mixture of water and EtOAc
(1 mL of each). The organic layer was washed with saturated NaCl,
dried (MgSO4), and filtered through a short (2 cm × 0.5 cm) plug of
silica gel. The silica gel was washed with 1:1 hexanes/EtOAc.
Evaporation of the combined filtrates yielded an oil that was purified
by silica gel chromatography (20:1 hexanes/EtOAc), giving 1 (2.5 mg,
73%) as an oil. Synthetic 1: [R]23D +70° (c 0.0015, hexane); IR (film)
Supporting Information Available: Experimental proce-
dures and complete physical and spectral data for compounds
3, 12, 13, 17-19, 25-31, and 35-36 (14 pages total). See
any current masthead page for ordering and Internet access
instructions.
1
2948, 2869, 1634, 1456, 1138, 999, 893 cm-1; H NMR (400 MHz,
CDCl3) δ 4.95 (s, 1H), 4.79 (s, 1H), 3.54 (ddd, J1 ) 6.7, J2 ) J3 )
9.9, 1H), 2.82 (dd, J ) 6.6, 11.6, 1H), 2.63 (ddd, J1 ) 6.7, J2 ) J3 )
JA970022U