Asymmetric total syntheses of (+)-mycoepoxydiene and related natural product (-)-1893A: Application of one-pot ring-opening/cross/ring-closing metathesis to construct their 9-oxabicyclo[4.2.1]nona-2,4-diene skeleton

Article abstract of DOI:10.1021/jo048566j

The total syntheses of (+)-mycoepoxydiene and (-)-1893A have been completed. The present synthetic strategy features the use of one-pot ring-opening/cross metathesis (ROM/CM) followed by a ring-closing metathesis (RCM) reaction, allowing for the concise construction of the 9-oxabicyclo-[4.2.1]nona-2,4-diene framework from a 7-oxabicyclo[2.2.1]hept-2- ene derivative and 1,3-butadiene. The sequential metathesis product was converted into (+)-mycoepoxydiene through the oxidative rearrangement of a furfuryl alcohol to a pyranone, thereby establishing its absolute stereochemistry. From the common intermediate, a structurally related natural product (-)-1893A was also synthesized via the vinylogous aldol reaction.

Full text of DOI:10.1021/jo048566j

Asymmetric Total Syntheses of (+)-Mycoepoxydiene and Related
Natural Product (-)-1893A: Application of One-Pot Ring-Opening/
Cross/Ring-Closing Metathesis to Construct Their
9-Oxabicyclo[4.2.1]nona-2,4-diene Skeleton
Ken-ichi Takao, Hiroyuki Yasui, Shun Yamamoto, Daisuke Sasaki, Soujiro Kawasaki,
Gohshi Watanabe, and Kin-ichi Tadano*
Department of Applied Chemistry, Keio University, Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan
tadano@applc.keio.ac.jp
Received August 17, 2004
The total syntheses of (+)-mycoepoxydiene and (-)-1893A have been completed. The present
synthetic strategy features the use of one-pot ring-opening/cross metathesis (ROM/CM) followed
by a ring-closing metathesis (RCM) reaction, allowing for the concise construction of the 9-oxabicyclo-
[4.2.1]nona-2,4-diene framework from a 7-oxabicyclo[2.2.1]hept-2-ene derivative and 1,3-butadiene.
The sequential metathesis product was converted into (+)-mycoepoxydiene through the oxidative
rearrangement of a furfuryl alcohol to a pyranone, thereby establishing its absolute stereochemistry.
From the common intermediate, a structurally related natural product (-)-1893A was also
synthesized via the vinylogous aldol reaction.
Introduction
(+)-Mycoepoxydiene (1) was isolated from the solid-
state fermentation of a rare fungus designated as OS-
F66617 in 1999 (Figure 1).1 The structure of 1 was
elucidated by extensive spectroscopic studies and finally
determined by X-ray crystallographic analysis, although
its absolute stereochemistry remained unclear. There had
been no publication regarding its biological properties.
This fungal metabolite is the first natural product
reported so far that contains a 9-oxabicyclo[4.2.1]nona-
2,4-diene skeleton.2 This unprecedented structure is
thought to be of polyketide origin.1 On the other hand,
1893A (2) and 1893B (3) were isolated recently from the
fermentation broth of a marine endophytic fungus (No.
1893), showing cytotoxic and insecticidal activities.3 The
two new metabolic products, 2 and 3, were found to
possess the same oxygen-bridged cyclooctadiene core
skeleton. Being interested in their unique structures and
the biological activities of 2 and 3, we became involved
in synthetic studies of mycoepoxydiene (1) as well as
1893A (2) and 1893B (3).4 In a program directed toward
the concise synthesis of the core skeleton of these related
natural products, we have developed an attractive meth-
FIGURE 1. Structures of (+)-mycoepoxydiene, (-)-1893A, and
1893B.
odology for the construction of the 9-oxabicyclo[4.2.1]-
nona-2,4-diene structure5 using one-pot ring-opening/
cross metathesis (ROM/CM) followed by a ring-closing
metathesis (RCM) reaction. Herein, we describe the full
details of the asymmetric total syntheses of (+)-1 and
(-)-2 by the ROM/CM/RCM approach.
Olefin metathesis is established as a remarkably
valuable synthetic tool in current organic chemistry.6
Recently, several new methods using sequential metath-
esis for the formation of medium- and large-sized ring
cycloalkadienes have appeared in the literature. For
example, Grubbs and co-workers have explored a ring-
expansion reaction via a series of three types of olefin
metathesis (ring-opening, cross, and ring-closing metath-
esis) to yield a variety of 18- to 39-membered macro-
cycles.7 Mori and co-workers demonstrated that the ring-
* To whom correspondence should be addressed.
(1) Cai, P.; McPhail, A. T.; Krainer, E.; Katz, B.; Pearce, C.; Boros,
C.; Caceres, B.; Smith, D.; Houck, D. R. Tetrahedron Lett. 1999, 40,
1479-1482.
(2) Oxygen-bridged dibenzocyclooctadiene lignans such as kadsulig-
nans are known. See: Liu, J.-S.; Li, L. Phytochemistry 1995, 38, 241-
245.
(3) Chen, G.; Lin, Y.; Wen, L.; Vrijmoed, L. L. P.; Jones, E. B. G.
Tetrahedron 2003, 59, 4907-4909. The whole stereochemistry of 1893B
(3) has not been determined.
(4) The first total synthesis of (()-1 was published by us in a
preliminary communication. See: Takao, K.; Watanabe, G.; Yasui, H.;
Tadano, K. Org. Lett. 2002, 4, 2941-2943.
(5) Recently, an elegant route to enantiopure 9-oxabicyclo[4.2.1]non-
3-ene was reported. See: Paquette, L. A.; Kim, I. H.; Cunie`re, N. Org.
Lett. 2003, 5, 221-223.
(6) For recent reviews on olefin metathesis, see: (a) Schrock, R. R.;
Hoveyda, A. H. Angew. Chem., Int. Ed. 2003, 42, 4592-4633. (b)
Blechert, S.; Connon, S. J. Angew. Chem., Int. Ed. 2003, 42, 1900-
1923. (c) Arjona, O.; Csa´ky¨, A. G.; Plumet, J. Eur. J. Org. Chem. 2003,
611-622. (d) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34,
18-29. (e) Fu¨rstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012-3043.
(7) Lee, C. W.; Choi, T.-L.; Grubbs, R. H. J. Am. Chem. Soc. 2002,
124, 3224-3225.
10.1021/jo048566j CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/09/2004
J. Org. Chem. 2004, 69, 8789-8795
8789

Takao et al.
SCHEME 1
SCHEME 2a
opening and successive ring-closing metathesis of cyclo-
alkene bearing an alkyne side chain proceeded to provide
bicyclic compounds and/or dimeric compounds.8 Kulkarni
and Diver reported cycloheptadiene synthesis by tandem
intermolecular enyne metathesis between 1-alkynes and
cyclopentene.9 In the present report, we describe the
synthesis of the core skeleton of 1 and 2 through two
synthetic routes by the use of an intramolecular or
sequential inter- and then intramolecular olefin metath-
esis strategy. The first route relies on the RCM of a
doubly 2-propen-1-yl-substituted tetrahydrofuran deriva-
tive at C2 and C5. The resulting RCM product, a
9-oxabicyclo[4.2.1]non-3-ene, was transformed into a
9-oxabicyclo[4.2.1]nona-2,4-diene skeleton. Our second
route involves one-pot ROM/CM followed by RCM to
construct directly the 9-oxabicyclo[4.2.1]nona-2,4-diene
skeleton from a functionalized 7-oxabicyclo[2.2.1]hept-
2-ene derivative.
a Reagents and conditions: (a) lipase PS (Amano), vinyl acetate,
83%; (b) TBDPSCl, imidazole, DMF; (c) NaOMe, MeOH, 92% for
2 steps; (d) CH2dCHOEt, PPTS, CH2Cl2, 97%; (e) NaOMe, MeOH,
quant.; (f) TBDPSCl, imidazole, DMF, quant.; (g) 60% aq AcOH,
THF, 97%; (h) LiAlH4, THF; (i) Ac2O, pyridine, 67% for 2 steps;
(j) lipase PS (Amano), THF-H2O, 88%.
acetate (+)-10 in an enantiomeric excess (ee) of 96%.13
Silylation of (+)-10 followed by deacetylation provided
tert-butyldiphenylsilyl (TBDPS) ether (+)-7. In addition,
the enantiomer (-)-7 was prepared from (+)-10 via a
protection-deprotection sequence, i.e., (1) ethoxyethyl
(EE) etherification of the hydroxy group in (+)-10, (2)
deacetylation, (3) silylation with TBDPSCl, and then (4)
deprotection of the EE group under mild acidic condi-
tions. An alternative and more concise synthesis of (-)-7
was performed from meso-diacetate 14, readily prepared
from 8. By treatment of 14 with lipase PS in THF-H2O,
the enantioselective hydrolysis of one acetyl ester oc-
curred effectively to provide monoacetate (-)-10 in 95%
ee,13 the antipode of the monoacetate prepared from 9.
By using the same procedure for the preparation of (+)-7
from (+)-10, (-)-10 was converted into silyl ether (-)-7.
As shown in Scheme 3, the functionalized 7-oxabicyclo-
[2.2.1]hept-2-ene derivative (-)-7 was converted into the
2,5-dipropenylated tetrahydrofuran (+)-6. For the radi-
cal-mediated removal of the hydroxy group, (-)-7 was
converted into the xanthate ester (-)-15. Dihydroxylation
of the double bond in (-)-15 with osmium tetroxide and
trimethylamine N-oxide (TMNO)14 stereoselectively pro-
vided diol (-)-16,15 which was treated with tri-n-butyltin
hydride to provide the deoxygenated product (+)-17.16
Oxidative cleavage of the cis-diol in (+)-17 with lead
tetraacetate resulted in the formation of a dialdehyde
hydrate, the 1H NMR spectrum of which showed no
signals attributable to two aldehyde groups. The oxida-
tive cleavage product was reduced with sodium boro-
Results and Discussion
The First-Generation Approach to the Function-
alized Oxygen-Bridged Cyclooctadiene (+)-4. Our
first-generation retrosynthetic analysis of 1 and 2 is
shown in Scheme 1. We envisioned synthesizing the
target compounds 1 and 2 from a common synthetic
intermediate, 7,8-dialkylated 9-oxabicyclo[4.2.1]nona-2,4-
diene 4. As a precursor for the oxygen-bridged cyclooc-
tadiene 4, the oxygen-bridged cyclooctene derivative 5
seemed to be a promising intermediate. The formation
of 5 would be achieved by the RCM of 2,5-dipropenylated
tetrahydrofuran 6. The substrate for this RCM reaction
was expected to be prepared from a 7-oxabicyclo[2.2.1]-
hept-2-ene derivative such as 7, which in turn could be
prepared in an optically active form through asymmetric
desymmetrization of a corresponding meso-compound
such as the known 8,10 the Diels-Alder exo-adduct of
furan and maleic anhydride. The chirality incorporated
into the ring junction of 8 was anticipated to be preserved
throughout the remaining process.
The syntheses of 1 and 2 began with the asymmetric
desymmetrization of meso-compounds 9 and 14 by the
well-known enzymatic methods (Scheme 2). According to
Bloch’s precedent,11 the lipase-catalyzed transesterifica-
tion of vinyl acetate with meso-diol 9, the hydride-
reduction product of 8,12 was conducted to afford mono-
(12) Das, J.; Vu, T.; Harris, D. N.; Ogletree, M. L. J. Med. Chem.
1988, 31, 930-935.
(13) The ee values of (+)- and (-)-10 were determined by chiral
HPLC analysis after conversion into the corresponding p-tert-butyl-
benzoates. The absolute stereochemistry of (+)-10 has been unambigu-
ously determined by Bloch’s group.11
(14) Ray, R.; Matteson, D. S. Tetrahedron Lett. 1980, 21, 449-450.
(15) In the 1H NMR spectrum of (-)-16, two characteristic coupling
constants (J1,6 ) 0 Hz and J4,5 ) 0 Hz) were observed. The fact
indicated that the dihydroxylation occurred on the â-face of (-)-15 to
provide exo-diol (-)-16.
(16) An attempt at radical-induced deoxygenation of (-)-15 with
n-Bu3SnH in the presence of AIBN gave a complex mixture.
(8) Mori, M.; Kuzuba, Y.; Kitamura, T.; Sato, Y. Org. Lett. 2002, 4,
3855-3858.
(9) Kulkarni, A. A.; Diver, S. T. Org. Lett. 2003, 5, 3463-3466.
(10) (a) Woodward, R. B.; Baer, H. J. Am. Chem. Soc. 1948, 70,
1161-1166. (b) Chem. Abstr. 1966, 65, 16924e-16925b.
(11) Cinquin, C.; Schaper, I.; Mandville, G.; Bloch, R. Synlett 1995,
339-340.
8790 J. Org. Chem., Vol. 69, No. 25, 2004

Asymmetric Total Synthesis of (+)-Mycoepoxydiene
SCHEME 3a
SCHEME 5
(four-time addition of each 5 mol % equiv of 21 over a
period of 20 h) to provide the oxygen-bridged cyclooctene
(+)-5 in 83% yield.20 Increasing the concentration of (+)-6
in benzene (0.005 M) decreased the yield to 70%, presum-
ably due to the co-occurrence of competing oligomeriza-
tion. The use of the second-generation Grubbs catalyst
2221 enabled the reduction of the total amount of catalyst
to 3 mol % and the reduction of the reaction time to 6 h,
with the formation of (+)-5 in a slightly higher yield of
86%. Then we explored the introduction of the diene unit
in the 9-oxabicyclo[4.2.1]nonene ring. The addition of 1
equiv of bromine to (+)-5 and the subsequent â-elimina-
tion of two-molar hydrogen bromide from the resulting
stereoisomeric mixture of dibromide 23, with potassium
tert-butoxide in hot tert-butyl alcohol, provided the
desired 9-oxabicyclo[4.2.1]nona-2,4-diene (+)-24. The
fluoride-mediated desilylation of (+)-24 provided the
advanced synthetic intermediate (+)-4. The first-genera-
tion synthesis of the enantiomerically enriched oxygen-
bridged cyclooctadiene (+)-4 required 17 steps in 13.6%
overall yield from meso-diol 9 or 15 steps in 14.1% overall
yield from meso-diacetate 14.
a Reagents and conditions: (a) NaH, CS2, MeI, THF, 96%; (b)
OsO4, TMNO, aq acetone, 85%; (c) n-Bu3SnH, AIBN, toluene,
reflux, 89%; (d) Pb(OAc)4, PhH; (e) NaBH4, MeOH, 81% for 2 steps;
(f) n-BuLi, TsCl, THF, 0 °C, 79%; (g) NaI, 2-butanone, reflux, 96%;
(h) vinylMgBr, PhH, 89%.
SCHEME 4a
The Second-Generation Access to (+)-4 by Using
the ROM/CM/RCM Approach. To streamline the syn-
thetic route to the 9-oxabicyclo[4.2.1]nona-2,4-diene struc-
ture, we undertook the development of a methodology
based on sequential metathesis. Our second-generation
synthetic approach is summarized in Scheme 5. In this
novel approach, the construction of oxygen-bridged cy-
clooctadiene (+)-24 was expected to be performed by the
RCM of triene 25a or 25b, which in turn would be
synthesized by the ring-opening/cross metathesis (ROM/
CM) of 7-oxabicyclo[2.2.1]hept-2-ene 26 with 1,3-buta-
diene. The intermediate 26 would be prepared from (-)-
7. This sequential ROM/CM/RCM approach seemed
particularly attractive because the 9-oxabicyclo[4.2.1]-
nona-2,4-diene skeleton can be formed from 7-oxabicyclo-
[2.2.1]hept-2-ene in fewer steps.
a Reagents and conditions: (a) Br2, Et2O, -78 °C, 96%; (b)
t-BuOK, t-BuOH, 75 °C, 53%; (c) n-Bu4NF, THF, quant.
hydride to provide tetrahydrofuran 2,5-dimethanol (+)-
18. With use of n-butyllithium as a base, simultaneous
ditosylation of (+)-18 was performed to provide (+)-19,17
which was transformed into diiodide (+)-20 with sodium
iodide in boiling 2-butanone. An attempt to substitute
simultaneously the two iodo groups in (+)-20 with a vinyl
group with use of a variety of vinylcopper reagents gave
no desired product. Eventually, (+)-20 was found to react
smoothly with an excess amount of vinylmagnesium
bromide in the absence of a copper(I) catalyst in benzene
at room temperature, providing the 2,5-diallylated
tetrahydrofuran (+)-6.18
Although 7-oxabicyclo[2.2.1]hept-2-ene derivatives are
known to undergo ring-opening metathesis generating a
variety of functionalized compounds,22 precedents for
successive cross metathesis with use of 1,3-butadiene are
With the diene (+)-6 in hand, the RCM reaction was
investigated (Scheme 4). In a high-dilution (0.003 M)
solution in refluxing benzene, the substrate (+)-6 was
treated with the first-generation Grubbs catalyst 2119
(17) Use of pyridine in place of the n-BuLi/THF system resulted in
the undesired intramolecular cyclization of intermediary monotosylate,
providing a 3,8-dioxabicyclo[3.2.1]octane derivative.
(20) When 20 mol % of 21 was added in one portion, the product
(+)-5 was obtained in 41% yield.
(21) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999,
1, 953-956.
(18) When tetrahydrofuran was used as the solvent, the reaction
proceeded less effectively.
(22) (a) Schneider, M. F.; Lucas, N.; Velder, J.; Blechert, S. Angew.
Chem., Int. Ed. Engl. 1997, 36, 257-259. (b) Arjona, O.; Csa´ky¨, A. G.;
Murcia, M. C.; Plumet, J. J. Org. Chem. 1999, 64, 9739-9741. Also
see refs 6b and 6c.
(19) (a) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2039-2041. (b) Schwab, P.; Grubbs,
R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996, 118, 100-110.
J. Org. Chem, Vol. 69, No. 25, 2004 8791

Takao et al.
TABLE 2. The Ring-Closing Metathesis of 25a-d
SCHEME 6a
entry
conditionsa
yield of 4,b %
1
21, benzene (0.003 M), reflux
22, benzene (0.003 M), reflux
22, CH2Cl2 (0.003 M), reflux
22, toluene (0.003 M), reflux
22, Ti(O-i-Pr)4, benzene (0.003 M), reflux
2
29
28
27
10
2
3
4
5c
a Triene substrates 25a-d were prepared by using the proce-
dure of entry 1 in Table 1. 20 mol % of catalyst 21 or 22 was used.
b Isolated yield of (+)-4 after desilylation. c 1 equiv of Ti(O-i-Pr)4
was used.
hept-2-ene, an excess (4 equiv) of 1,3-butadiene was
employed (entry 5). As expected, the yield was improved
with a minimal decrease in the Z/E-selectivity. As shown
in Table 2, the RCM reaction of the mixture of trienes
25a-d occurred more effectively by use of the catalyst
22 in refluxing benzene (entry 2), affording the 9-oxabi-
cyclo[4.2.1]nona-2,4-diene compound (+)-24 accompanied
by inseparable and unidentified byproducts. By exposure
of the mixture to tetrabutylammonium fluoride, homo-
geneous (+)-4 was obtained in a yield of 29% after
chromatography on silica gel. The first-generation cata-
lyst 21 did not catalyze the cyclization (entry 1). Whereas
the RCM of (+)-6 smoothly produced the 9-oxabicyclo-
[4.2.1]non-3-ene (+)-5 as shown in Scheme 4, that of
25a-d proceeded much more slowly and less effectively.
The solvent did not affect this reaction (entries 2-4).25
Assuming that a lone pair of electrons of the bridge-
oxygen might diminish the catalyst reactivity, we also
attempted RCM in the presence of titanium tetraisopro-
poxide26 (entry 5). Under these conditions, the desired
product (+)-4 was obtained in a lower yield of 10% after
desilylation.
a Reagents and conditions: (a) TsCl, Et3N, DMAP, CH2Cl2,
quant.; (b) NaBH4, DMPU, 90 °C, 76%; (c) n-Bu4NF, THF.
TABLE 1. The Ring-Opening/Cross Metathesis of (+)-26
with 1,3-Butadiene
entry
conditionsa
yield,b %
Z/Ec
1
21, benzene (0.01 M), rt
22, benzene (0.01 M), rt
21, CH2Cl2 (0.01 M), rt
22, CH2Cl2 (0.01 M), rt
21, benzene (0.01 M), rt
60
40
59
50
75
1.8:1
1:1.2
1.5:1
1:2.3
1.5:1
2
3
4
5d
a Unless otherwise noted, 2 equiv of 1,3-butadiene and 5 mol %
of catalyst 21 or 22 were used. b Combined yield for 25a-d.
c Determined by 1H NMR (270 MHz). d 4 equiv of 1,3-butadiene
was used.
To construct the 9-oxabicyclo[4.2.1]nona-2,4-diene struc-
ture from the 7-oxabicyclo[2.2.1]hept-2-ene derivative
more concisely,27 we next performed the sequential ROM/
CM/RCM reaction in a one-pot process using the sub-
strates (+)-26, (+)-28, and (-)-13 (Scheme 7). The
treatment of (+)-26 with 1,3-butadiene (2 equiv) in the
presence of 21 (2 mol %) in benzene (0.006 M) at room
temperature initially produced trienes 25a-d. After (+)-
26 was consumed, the reaction mixture was heated to
reflux and bubbled in a stream of argon to remove excess
1,3-butadiene. Then each 2 mol % of 22 was added in
five portions over a period of 4 days. Desilylation of the
reaction mixture provided the key intermediate (+)-4 in
a yield of 23% from (+)-26. When 4 equiv of 1,3-butadiene
was used in this one-pot reaction, the yield was de-
creased, contrary to our expectation. The ROM/CM/RCM
reaction of substrate (+)-28, prepared from (+)-26 by
desilylation, proceeded less effectively. On the other
hand, the di-O-protected diol derivative (-)-13 underwent
one-pot ROM/CM/RCM to afford (+)-30, after removal of
the ethoxyethyl group, in almost the same yield as that
scarce.23 To confirm the feasibility of the designed ap-
proach, we first examined the two reactions in a stepwise
manner. The substrate (+)-26 was prepared from (-)-7
through tosylation followed by the reductive removal of
the sulfonate in the resultant (-)-27 with sodium boro-
hydride in N,N′-dimethylpropyleneurea (DMPU) (Scheme
6). The results for the first ROM/CM of (+)-26 with 1,3-
butadiene are summarized in Table 1. With use of the
first-generation Grubbs catalyst 21 in benzene (entry 1),
the reaction proceeded smoothly at room temperature to
give a mixture of trienes with a favorable formation of a
mixture of the Z-isomers 25a and 25b, which were
indispensable for the next ring-closing step.24 In contrast,
the second-generation Grubbs catalyst 22 produced pref-
erentially the undesired E-isomers (25c and 25d) in a
diminished yield of 40% (entry 2). Lower Z-selectivity was
also observed when the reaction was conducted in dichlo-
romethane (entries 3 and 4). To prevent a competing self-
metathesis (polymerization) of the 7-oxabicyclo[2.2.1]-
(23) (a) Randall, M. L.; Tallarico, J. A.; Snapper, M. L. J. Am. Chem.
Soc. 1995, 117, 9610-9611. (b) Michaut, M.; Parrain, J.-L.; Santelli,
M. Chem. Commun. 1998, 2567-2568.
(24) The Grubbs group reported that treatment of 21 with a 10-fold
excess of 1,3-butadiene resulted in the high-yield formation of a
vinylalkylidene carbene RuCl2(dCHCHdCH2)(PCy3)2 and that no
formation of a methylidene carbene RuCl2(dCH2)(PCy3)2 occurred even
after a prolonged reaction time.19b We presume that the vinylalkylidene
ruthenium complex, which behaves as an active catalytic species, is
initially formed. For the reaction mechanism of ROM/CM, see ref 23.
(25) For discussions regarding the effect of solvent and reaction
temperatures in metathesis, see: (a) Fu¨rstner, A.; Thiel, O. R.;
Ackermann, L.; Schanz, H.-J.; Nolan, S. P. J. Org. Chem. 2000, 65,
2204-2207. (b) Yamamoto, K.; Biswas, K.; Gaul, C.; Danishefsky, S.
J. Tetrahedron Lett. 2003, 44, 3297-3299.
(26) Fu¨rstner, A.; Langemann, K. J. Am. Chem. Soc. 1997, 119,
9130-9136.
(27) Under the ROM/CM conditions described in Table 1, no
formation of the 9-oxabicyclo[4.2.1]nona-2,4-diene skeleton was ob-
served.
8792 J. Org. Chem., Vol. 69, No. 25, 2004

Asymmetric Total Synthesis of (+)-Mycoepoxydiene
SCHEME 7a
SCHEME 8a
a Reagents and conditions: (a) Dess-Martin periodinane, CH2Cl2;
(b) furan, n-BuLi, THF, 0 °C, 43% for 2 steps; (c) MnO2, CH2Cl2,
72%; (d) L-Selectride, THF, -78 °C, 85%; (e) NaClO2, NH2SO3H,
Na2HPO4, t-BuOH-H2O; (f) Me(MeO)NH‚HCl, EDC‚HCl, HOBt,
Et3N, DMAP, CH2Cl2; (g) furan, n-BuLi, THF, 0 °C, 43% from (+)-
4; (h) VO(acac)2, t-BuOOH, CH2Cl2, 0 °C, 90%; (i) TBSCl, Et3N,
DMAP, CH2Cl2, 46% for (+)-37 and 19% for (+)-38.
a Reagents and conditions: (a) n-Bu4NF, THF; (b) PPTS, MeOH;
(c) TsCl, Et3N, DMAP, CH2Cl2, quant.; (d) NaBH4, DMPU, 90 °C,
76%; (e) n-Bu4NF, THF, quant.
SCHEME 9a
obtained in the case of (+)-26.28 The product (+)-30 was
deoxygenated to (+)-4 via tosylate (-)-31. Although the
yield is still moderate, the present one-pot ROM/CM/
RCM approach realizes access to (+)-4 in 10 steps with
an overall yield of 13.7% from meso-diol 9 or in 8 steps
from meso-diacetate 14 with a 14.2% overall yield.
Completion of the Synthesis of (+)-Mycoepoxy-
diene (1). Having established a practical synthetic route
to the 9-oxabicyclo[4.2.1]nona-2,4-diene intermediates,
we next focused our efforts on the assembly of the
δ-lactone moiety attached to the main scaffold. The key
step in achieving this purpose was the oxidative rear-
rangement of a furfuryl alcohol, incorporated into the
main scaffold, to a hydroxylated pyranone.29 The total
synthesis of (+)-1 is illustrated in Schemes 8 and 9.
Oxidation of (+)-4 with a Dess-Martin reagent30 pro-
vided the corresponding aldehyde, which was reacted
with 2-furyllithium derived from furan and n-butyl-
lithium, providing an inseparable diastereomeric mixture
(ca. 3:2) of 32 and 33. For the synthesis of 1, the
diastereomer 32 possesses the correct stereochemistry.
Thus, more efficient preparation of 32 was explored and
accomplished by an oxidation-reduction procedure. The
mixture of furfuryl alcohols 32 and 33 was oxidized to
a Reagents and conditions: (a) NaBH4, CeCl3‚7H2O, MeOH-
CH2Cl2, -78 °C, 86%; (b) Ac2O, DMAP, pyridine, 89%; (c) aq HCl,
THF, quant.; (d) NaBH4, CeCl3‚7H2O, MeOH, 0 °C, 63% for (+)-
42 and 30% for (+)-43; (e) Ac2O, DMAP, pyridine; (f) aq HCl, THF,
80% for 2 steps; (g) MnO2, CH2Cl2, 76%; (h) Ac2O, DMAP, pyridine;
(i) aq HCl, THF; (j) MnO2, CH2Cl2, 65% for 3 steps.
(28) It was confirmed that the ROM/CM of (-)-13 with use of 21 in
benzene produced preferentially the Z-isomers (Z/E ) 1.6:1).
(29) For a review on furan as a building block in organic synthesis,
see: Maier, M. In Organic Synthesis Highlights II; Waldmann, H., Ed.;
VCH: Weinheim, Germany, 1995; pp 231-242.
(30) (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-
4156. (b) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277-
7287. (c) Ireland, R. E.; Liu, L. J. Org. Chem. 1993, 58, 2899.
(-)-34 with manganese dioxide. Reduction of the carbonyl
group in (-)-34 with L-Selectride provided the desired
J. Org. Chem, Vol. 69, No. 25, 2004 8793

Takao et al.
SCHEME 10a
furfuryl alcohol (+)-32 as a sole product. This reduction
proceeded possibly through the Li-chelation-assisted
transition state occurring between the carbonyl group
and the bridge oxygen. A more efficient route to ketone
(-)-34 was achieved as follows. The primary alcohol in
(+)-4 was oxidized to the carboxylic acid by two-step
oxidation, which was coupled with N,O-dimethylhydroxy-
lamine by using water-soluble carbodiimide to provide
the Weinreb amide 35. The addition of 2-furyllithium to
35 afforded the ketone (-)-34 in an improved yield of 43%
from (+)-4. Among the known methods for the rearrange-
ment of furfuryl alcohols to pyranones,29 the procedure
with vanadium-mediated oxidation conditions31 seemed
to be well-suited to (+)-32 because these conditions were
expected to prevent the oxidation of the diene in the
eight-membered ring. In fact, the treatment of (+)-32
with tert-butyl hydroperoxide in the presence of a cata-
lytic amount of vanadyl acetylacetonate (VO(acac)2) ef-
fectively provided the rearranged product (+)-36 as a 2:1
hemiacetal mixture through the rearrangement depicted
in brackets.32 Protection of the hemiacetal hydroxy group
in (+)-36 as the tert-butyldimethylsilyl (TBS) ether gave
a 2.4:1 mixture of R-isomer (+)-37 and â-isomer (+)-38,
which were readily separated.
a Reagents and conditions: (a) Dess-Martin periodinane, CH2Cl2;
(b) 45, TESOTf, CH2Cl2, -78 °C, 60% for 2 steps; (c) MsCl,
pyridine, rt to 80 °C for 2 days, 63% for (-)-2 and 13% for (-)-47;
(d) MsCl, pyridine, 90 °C, 63% for (-)-2 and 25% for recovered
(-)-47.
The stereoselective 1,2-reduction of the major product
(+)-37 was conducted under Luche conditions33 to afford
allylic alcohol (+)-39 exclusively, which was converted
into acetate (+)-40 (Scheme 9). The stereochemistry of
(+)-40 was confirmed by 1H NMR analysis, including
NOE experiments.34 Consequently, the configurations of
stereogenic centers introduced by the L-Selectride reduc-
tion of (-)-34 (C5, mycoepoxydiene numbering) and by
the Luche reduction of (+)-37 (C4) were determined as
depicted. By hydrolysis with dilute hydrochloric acid, the
TBS group in (+)-40 was removed to provide lactol (+)-
41. The minor diastereomer (+)-38 was also converted
into (+)-41 by using the same procedure as that described
for the transformation of (+)-37 to (+)-41. The less
stereoselective result was observed in the Luche reduc-
tion of (+)-38. The stereoselectivity of the reduction of
(+)-38 was affected by the reaction temperature. When
the reaction was carried out at -78 °C, the ratio of (+)-
42 to (+)-43 was 1:2. At 0 °C, the desired (+)-42 was
obtained favorably (dr ) 2:1). Finally, the lactol (+)-41
was oxidized to (+)-mycoepoxydiene (1) with manganese
dioxide. The spectroscopic data (1H and 13C NMR) of
synthetic (+)-1 were well matched with those reported
for natural 1.1,35 Comparison of the specific rotation of
synthesized from (+)-7 by using exactly the same syn-
thetic steps as those used for the total synthesis of (+)-
1. Furthermore, the minor reduction product (+)-43 was
converted into the 4-epimer of (+)-mycoepoxydiene (-)-
44, which apparently showed different spectra (1H and
13C NMR) from those of 1.
Total Synthesis of (-)-1893A (2). The total synthesis
of (-)-1893A (2) was achieved from the advanced inter-
mediate (+)-4 as shown in Scheme 10. As a butenolide
anion equivalent, 2-(trimethylsilyloxy)furan (45) was
chosen due to its high γ-regioselectivity.36 Thus, oxidation
of (+)-4 and the subsequent vinylogous aldol reaction of
the resulting aldehyde with 45 in the presence of trieth-
ylsilyl triflate (TESOTf)37 provided γ-adduct 46 as a
mixture of stereoisomers (ca. 8:1).38,39 The sulfonylation
of 46 with methanesulfonyl chloride (MsCl) in pyridine,
followed by heating at 80 °C for 30 min,40 gave (-)-1893A
(2) and its E-isomer (-)-47 in 45% and 38% yields,
respectively. Gratifyingly, a prolonged reaction time (2
days) for converting 46 into the mixture of (-)-2 and (-)-
47 increased the Z:E ratio to ca. 5:1, and a 63% yield of
(-)-2 was isolated. In addition, the minor product (-)-
47 was subjected to the same conditions (MsCl, pyridine)
to afford additional (-)-2. This result is reasonably
attributable to a reversible conjugate addition of a
nucleophilic species such as a chloride ion to δ-position
synthetic (+)-1 ([R]20 +227; c 0.072, MeOH) with that
D
of the natural product ([R]D +210; c 0.106, MeOH)
established the absolute stereochemistry as depicted in
Scheme 9. Analogously, the unnatural enantiomer (-)-
mycoepoxydiene (1) ([R]19 -223; c 0.093, MeOH) was
D
(31) Ho, T.-L.; Sapp, S. G. Synth. Commun. 1983, 13, 207-211.
(32) Reproducible results were obtained by a nonaqueous workup.
For an example of water-unstable hemiacetal, see: Trost, B. M.; King,
S. A.; Schmidt, T. J. Am. Chem. Soc. 1989, 111, 5902-5915.
(33) Gemal, A. L.; Luche, J.-L. J. Am. Chem. Soc. 1981, 103, 5454-
5459.
(34) When H5 was irradiated, significant signal enhancements of
H1 (14%), H4 (7.6%), and H7 (6.9%) were observed. The lack of signal
enhancement of H6 as well as a large coupling constant (J5,6 ) 10.8
Hz) indicated that H5 and H6 are in an antiperiplanar relationship.
NOEs between H4/H7 (9.6%) and H5/CH3-13 (9.3%) were also observed.
(35) The optical purity of synthetic (+)-1 was confirmed by chiral
HPLC analysis to be 94%.
(36) For a review regarding vinylogous aldol reactions, see: Casi-
raghi, G.; Zanardi, F.; Appendino, G.; Rassu, G. Chem. Rev. 2000, 100,
1929-1972.
(37) Jefford, C. W.; Jaggi, D.; Boukouvalas, J. Tetrahedron Lett.
1987, 28, 4037-4040.
(38) The configurations of the two introduced stereogenic centers
in each diastereomer of 46 could not be determined.
(39) By using the mixture 46, the total synthesis of 1893B (3) was
also investigated. But we have not succeeded in the synthesis of 3 from
46 yet. Synthetic studies of 3 are now in progress.
(40) Pohmakotr, M.; Tuchinda, P.; Premkaisorn, P.; Reutrakul, V.
Tetrahedron 1998, 54, 11297-11304.
8794 J. Org. Chem., Vol. 69, No. 25, 2004

Asymmetric Total Synthesis of (+)-Mycoepoxydiene
(C5), followed by â-elimination, leading to (-)-2. On the
basis of the NOE experiment as shown in Scheme 10,
the stereochemistries including geometrical structures of
(-)-2 and (-)-47 were established.41 Furthermore, the
relative configuration of natural 2 was originally con-
firmed by X-ray crystallographic analysis.3 The spectro-
antibacterial activities against a variety of Gram-positive
and Gram-negative bacteria.
Conclusion
We have accomplished the asymmetric total syntheses
of (+)-mycoepoxydiene (1) and (-)-1893A (2). For the
construction of the core skeleton, a new methodology
based on an ROM/CM/RCM approach has been devel-
oped. Thus, ROM/CM of 7-oxabicyclo[2.2.1]hept-2-ene
(+)-26 with 1,3-butadiene, followed by RCM of the
resulting trienes 25a-d, produced 9-oxabicyclo[4.2.1]-
nona-2,4-diene (+)-4. For the conversion of (+)-4 into the
desired natural product (+)-mycoepoxydiene (1), we
applied the rearrangement of a furfuryl alcohol append-
age to a pyranone form. Through the present completion
of the total synthesis of (+)-1, the absolute stereochem-
istry of natural 1 was established. The total synthesis of
novel natural product (-)-1893A (2) was also achieved
by use of the vinylogous aldol reaction with 2-(trimeth-
ylsilyloxy)furan (45) as the introduction of the side chain.
Through the two total syntheses, we have verified that
the ROM/CM/RCM strategy is a valuable approach to
synthesize oxygen-bridged cyclooctadiene-type natural
products.
1
scopic data (IR, H and 13C NMR, HRMS) of synthetic
(-)-2 matched well those for natural 2:42 therefore it is
sure that naturally occurring 1893A (2) and our synthetic
(-)-2 possess the same relative stereochemistry. On the
other hand, the discrepancy in the absolute value of
specific rotation between synthetic and natural 2 was
observed. While our synthetic sample had [R]23D -285 (c
0.064, acetone) or [R]22 -293 (c 0.32, acetone), the
D
specific rotation of natural 2 was reported to be [R]20
D
+4.0 (c 0.07, acetone) in the original isolation paper.3
Later, the authors revised the value of [R]D to +148.43
As the amount of the existing natural sample was so
small, we could not examine a direct comparison.
Biological Activities of Mycoepoxydiene. The es-
tablishment of the total synthesis of mycoepoxydiene (1)
provided an opportunity to investigate the unknown
biological activity of this natural product. As a result of
biological studies with synthetic samples, natural (+)-1
and unnatural (-)-1 were both found to show cytotoxicity
toward human tumor cells in vitro with IC50s (µg/mL) of
2.3 for (+)-1 and 1.8 for (-)-1 against human chronic
myelogenous leukemia, K562, and of 3.1 for (+)-1 and
2.2 for (-)-1 against human hepatocellular carcinoma,
HepG2. Compounds (+)-1 and (-)-1 did not show any
Acknowledgment. We thank Professors Yongcheng
Lin (Zhongshan University) and Guangying Chen (City
University of Hong Kong) for providing spectral copies
of natural 2 and information on the optical rotation for
2. We also thank Infectious Disease Research Labs.,
Meiji Seika Kaisha, Ltd., for performing the biological
assays for 1, and Amano Enzyme, Inc., for donating
lipase PS. Professor Takeshi Sugai (Keio University)
provided us helpful advice regarding the enzymatic
experiment, to whom our thanks are due.
(41) In the case of (-)-2, a signal enhancement of H5 (4.5%) was
observed when H3 was irradiated, whereas NOE was observed between
H3 and H6 (11.1%) in (-)-47. When CH3-13 in (-)-2 was irradiated, a
signal enhancement at H5 (4.5%) was also observed. It was confirmed
that no epimerization at C6 occurred under Dess-Martin oxidation
and/or the vinylogous aldol reaction conditions.
Supporting Information Available: Experimental pro-
cedures, full characterization data, and 1H and 13C NMR
spectra for new compounds. This material is available free of
charge via the Internet at http://pubs.acs.org.
(42) For the spectroscopic comparison of natural and synthetic 2 (1H
and 13C NMR), see the Supporting Information.
(43) Chen, G. Personal communication. They explained that their
miscalculation caused this change.
JO048566J
J. Org. Chem, Vol. 69, No. 25, 2004 8795