Biomimetic syntheses from squalene-like precursors: Synthesis of ent- abudinol B and reassessment of the structure of muzitone

Article abstract of DOI:10.1021/ja1006806

We achieved the stereoselective syntheses of two different structural patterns corresponding to the enantiomers of the marine natural products abudinol B and muzitone, by developing two-directional tandem biomimetic cyclizations of polyepoxides of squalene analogues in which one alkene was functionalized as an enolsilane. In the course of this work, we demonstrated that the structure of muzitone was misassigned.

Full text of DOI:10.1021/ja1006806

Published on Web 03/24/2010
Biomimetic Syntheses from Squalene-Like Precursors:
Synthesis of ent-Abudinol B and Reassessment of the
Structure of Muzitone
Matthew A. Boone, Rongbiao Tong, Frank E. McDonald,* Sheri Lense,† Rui Cao,†
and Kenneth I. Hardcastle†
Department of Chemistry, Emory UniVersity, Atlanta, Georgia 30322
Received February 2, 2010; E-mail: fmcdona@emory.edu
Abstract: We achieved the stereoselective syntheses of two different structural patterns corresponding to
the enantiomers of the marine natural products abudinol B and muzitone, by developing two-directional
tandem biomimetic cyclizations of polyepoxides of squalene analogues in which one alkene was
functionalized as an enolsilane. In the course of this work, we demonstrated that the structure of muzitone
was misassigned.
Introduction
Polycyclic ether terpenoid marine natural products are
secondary metabolites from algae and sponges, which present
considerable diversity in their structural frameworks with
different ring sizes and ring fusion patterns as well as an
interesting dependence of biological activities on the ring
systems and substituents.1 A variety of terpenoid compounds
have been isolated from a sea sponge (Ptilocaulis spiculifer of
the Axinellidae family) native to the Red Sea waters of the
Dahlak archipelago of Eritrea. Representative compounds arising
from this isolation include abudinols A and B (1 and 2, Figure
1), the corresponding oxidative degradation products nakorone
(3) and durgamone (4), and a macrocyclic diketone muzitone
(5).2 The structural assignment for abudinol A (1) was secured
by crystallography,1g and ozonolysis of abudinol A produced
nakorone (3) and durgamone (4), whereas the structures of
abudinol B (2) and muzitone (5) were assigned solely from
spectroscopic analysis.
Figure 1. Representative marine terpenoids from Ptilocaulis spiculifer.
arise from tandem oxa- and carbacyclizations, by sequential
Markovnikov-mode nucleophilic additions terminated by elimi-
nation of an allylic hydrogen to provide 7. A similar mode of
cyclization in the opposite direction would provide the penta-
cyclic structure of abudinol B (2) after elimination of the
remaining allylic hydrogen (path a, Figure 2), possibly with
the intermediacy of the tertiary cation 8. The biosynthesis of
muzitone (5) has been proposed to involve oxidative cleavage
of the tetrasubstituted alkene in the pentacyclic structure “pre-
muzitone” 11 (isomeric to abudinol B). Protonation of abudinol
B (2) at C14 could form the tertiary cation 8, which might also
arise directly from bicyclization of 6. Ring expansion of 8 to
the spirocyclic 9 with a secondary cation at C14 might be
followed by a second ring expansion to structure 10, with
subsequent elimination of the C14-hydrogen to afford “pre-
muzitone” (11).
Abudinols A and B as well as muzitone have been proposed
to arise biogenetically from polycyclizations of squalene tet-
raepoxide(6,Figure2),involvingtwo-stagecascadecyclizations.3,4
For instance, the all-fused tricyclic sector of abudinol B might
† Emory University X-ray Crystallography Center.
(1) (a) Carmely, S.; Kashman, Y. J. Org. Chem. 1983, 48, 3517. (b)
Carmely, S.; Kashman, Y. J. Org. Chem. 1986, 51, 784. (c) Rudi, A.;
Goldberg, I.; Stein, Z.; Benayahu, Y.; Schleyer, M.; Kashman, Y.
Tetrahedron Lett. 1993, 34, 3943. (d) Rudi, A.; Kashman, Y.;
Benayahu, Y.; Schleyer, M. J. Nat. Prod. 1994, 57, 1416. (e) Rudi,
A.; Goldberg, I.; Stein, Z.; Kashman, Y.; Benayahu, Y.; Schleyer,
M.; Gravalos, G. J. Nat. Prod. 1995, 58, 1702. (f) Rudi, A.; Aknin,
M.; Gaydou, E. M.; Kashman, Y. J. Nat. Prod. 1997, 60, 700. (g)
Rudi, A.; Stein, Z.; Goldberg, I.; Yosief, T.; Kashman, Y.; Schleyer,
M. Tetrahedron Lett. 1998, 39, 1445.
Although neither abudinol B nor muzitone exhibited interest-
ing biological activities, their novel structures made them
compelling synthetic target compounds. Given our longstanding
interests in exploring higher-order polycyclization cascades in
the context of the biomimetic synthesis of oxacyclic compounds,5,6
we initiated a synthetic program directed toward the enantiomers
(2) (a) Rudi, A.; Yosief, T.; Schleyer, M.; Kashman, Y. Tetrahedron 1999,
55, 5555. (b) The stereochemistry of the C23 alcohol of abudinol B
and corresponding C19 alcohol of muzitone is incorrectly shown as
the R isomer in Figure 2 of ref 2a but correctly depicted with S
stereochemistry in Scheme 1 (communication with Prof. Yoel Kash-
man).
(5) (a) McDonald, F. E.; Bravo, F.; Wang, X.; Wei, X.; Toganoh, M.;
Rodr´ıguez, J. R.; Do, B.; Neiwert, W. A.; Hardcastle, K. I. J. Org.
Chem. 2002, 67, 2515. (b) Valentine, J. C.; McDonald, F. E. Synlett
2006, 1816. (c) McDonald, F. E.; Tong, R.; Valentine, J. C.; Bravo,
F. Pure Appl. Chem. 2007, 79, 281.
(3) Ferna´ndez, J.; Souto, M.; Norte, M. Nat. Prod. Rep. 2000, 17, 215.
(4) Kashman, Y.; Rudi, A. Phytochem. ReV. 2004, 3, 309.
9
5300 J. AM. CHEM. SOC. 2010, 132, 5300–5308
10.1021/ja1006806  2010 American Chemical Society

Biomimetic Syntheses of Oxacyclic Triterpenes
A R T I C L E S
Figure 2. Proposed biosyntheses of abudinol B (2) and muzitone (5) from squalene tetraepoxide (3).
Figure 3. Retrosynthetic analysis for ent-abudinol (ent-2) and ent-“pre-muzitone” (ent-11).
of abudinol B (ent-2) and “pre-muzitone” (ent-11).7,8 Our
synthesis of each target structure involved a hybrid of polyene9
and polyepoxide cyclizations5,6 as the key steps. Specifically,
we envisioned that the enantiomer of abudinol B (ent-2) would
directly form by the bicyclization of the proposed biosynthetic
intermediate (ent-7), which in turn could arise from squalene-
like diepoxide 12 bearing an enolsilane as a terminating
nucleophile10 at C15 rather than the methyl group normally
present in squalene (Figure 3). For the hypothetical alkene of
“pre-muzitone” (ent-11), our synthesis would not be biomimetic
but rather inspired by the mechanistic similarity of forming two
fused seven-membered rings of ent-11 by bicyclization of
diepoxide (13), arising from a C14 enolsilane diepoxide 14.
(6) (a) Fujiwara, K.; Hayashi, N.; Tokiwano, T.; Murai, A. Heterocycles
1999, 50, 561. (b) Tokiwano, T.; Fujiwara, K.; Murai, A. Synlett 2000,
335. (c) Xiong, Z. M.; Corey, E. J. J. Am. Chem. Soc. 2000, 122,
4831. (d) Xiong, Z. M.; Corey, E. J. J. Am. Chem. Soc. 2000, 122,
9328. (e) Heffron, T. P.; Jamison, T. F. Org. Lett. 2003, 5, 2339. (f)
Zakarian, A.; Batch, A.; Holton, R. A. J. Am. Chem. Soc. 2003, 125,
7822. (g) Simpson, G. L.; Heffron, T. P.; Merino, E.; Jamison, T. F.
J. Am. Chem. Soc. 2006, 128, 1056. (h) Vilotijevic, I.; Jamison, T. F.
Science 2007, 317, 1189. (i) Morimoto, Y.; Yata, H.; Nishikawa, Y.
Angew. Chem., Int. Ed. 2007, 46, 6481. (j) Wan, S.; Gunaydin, H.;
Houk, K. N.; Floreancig, P. E. J. Am. Chem. Soc. 2007, 129, 7915.
(k) Morten, C. J.; Jamison, T. F. J. Am. Chem. Soc. 2009, 131, 6678.
(l) Tanuwidjaja, J.; Ng, S. S.; Jamison, T. F. J. Am. Chem. Soc. 2009,
131, 12084. (m) Vilotijevic, I.; Jamison, T. F. Angew. Chem., Int. Ed.
2009, 48, 5250. (n) Morten, C. J.; Byers, J. A.; van Dyke, A. R.;
Vilotijevic, I.; Jamison, T. F. Chem. Soc. ReV. 2009, 38, 3175.
(7) Preliminary communications: (a) Tong, R.; Valentine, J. C.; McDonald,
F. E.; Cao, R.; Fang, X.; Hardcastle, K. I. J. Am. Chem. Soc. 2007,
119, 1050. (b) Tong, R.; McDonald, F. E. Angew. Chem., Int. Ed.
2008, 47, 4377.
Results and Discussion
Initial Methodology Studies for Bicyclizations of Terpene-
Derived Diepoxides. In earlier work from our laboratory directed
toward the synthesis of a bicyclic substructure of abudinol B,
we reported that bicyclization of diepoxide-enolsilane 15 was
accomplished with substoichiometric tert-butyldimethylsilyl
trifluoromethanesulfonate (TBSOTf) in the presence of 2,6-di-
(8) Given that D-epoxone (derived from D-fructose) was more easily
prepared than L-epoxone, we intended from the outset to prepare the
enantiomers of the natural products 2 and 5. (a) Zhao, M.-X.; Shi, Y.
J. Org. Chem. 2006, 71, 5377. Review: (b) Wong, O. A.; Shi, Y.
Chem. ReV. 2008, 108, 3958.
9
J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010 5301

A R T I C L E S
Boone et al.
Scheme 1. Lewis Acid Promoted Bicyclizations of Diepoxide
Substratesa
a Reagents and conditions: (a) TBSOTf (20 mol %), DTBMP (20 mol
%), CH2Cl2, -78 °C (60% yield of 16, 7.5% yield of 17); (b) TMSOTf (20
mol %), DTBMP (1 equiv), CH2Cl2, -78 °C (90% yield); (c) TMSOTf
(1.2 equiv), DTBMP (2 equiv), CH2Cl2, -78 °C; (65% yield).
tert-butyl-4-methylpyridine (DTBMP) to provide the bicyclic
ketone 16 (Scheme 1).7a The use of DTBMP was necessary to
ensure reproducibility by trapping any triflic acid generated from
trace amounts of adventitious water. Although the diepoxide
15 was a mixture of diastereomers due to lower stereoselectivity
in the epoxidation of the internal C5-C6 alkene precursor to
15, the ratio of diastereomeric products 16:17 matched the
diastereomer ratio of diepoxide 15. However, the minor product
17 was formed at a much slower rate than 16, which was
observed by carefully monitoring the reaction by TLC. Subse-
quently, we observed that the allylic silane of 18 was a superior
terminating group for bicyclizations,11 which upon reaction with
substoichiometric trimethylsilyl triflate (TMSOTf) and DTBMP
afforded an excellent yield of the bicyclic structure 19 bearing
an exocyclic methylene. Of greater relevance to the biosynthetic
pathway for the formation of abudinol B (2) from alkenyl-
diepoxide 7, the disubstituted alkene of diepoxide 20 also
participated in the tandem oxacyclization transformation with
stoichiometric TMSOTf/DTBMP to provide trisubstituted alkene
21 as the major product. Notably, the isomeric product 19 was
not observed as a byproduct in the bicyclization of 20.
Figure 4. Mechanisms for bicyclizations of diepoxide substrates.
bicyclo[4.1.0]epoxonium ion 23 (Figure 4),5 which then under-
goes addition of nucleophilic C1 with inversion of configuration
at C6 in the formation of cation 24. The catalytic cycle is then
propagated by transfer of trialkylsilicon from 24 to the terminal
epoxide oxygen in substrates 15/18, regenerating activated
epoxide 22 along with the stable products 16/19. For the
bicyclization of disubstituted alkene substrate 20 with stoichio-
metric TMSOTf, product 21 arises through the intermediacy of
a nonclassical carbenium ion 25 or tertiary carbocation 26,
followed by elimination of a C3 hydrogen to provide the
thermodynamically stable trisubstituted alkene. These prelimi-
nary experiments have demonstrated that (1) tandem bicycliza-
tions proceed with endo-selectivity in anti-parallel addition with
a variety of nucleophilic alkenes, (2) trialkylsilyl Lewis acids
(TMSOTf or TBSOTf) in the presence of a bulky non-
nucleophilic base (DTBMP) are optimal reagents for these
transformations, and (3) the stereochemistry of the epoxides
affects the cyclization rate, but not the course, so that the rate
of cyclization is faster with epoxides of the same configuration,
i.e. all-R,R-diepoxides rather than mixed R,R and S,S-diepoxides.
Our working hypothesis for these results is that initial
activation of the terminal epoxide results in the generation of a
Preparation of Squalene Diepoxide Analogues Bearing
Enolsilanes. The squalene-like substrates for the syntheses of
ent-abudinol B and ent-muzitone were each prepared by a
common strategy, involving alkylation of the anion generated
from 1,2-Brook rearrangement,12 either with the diepoxy-allylic
bromide 29 or with geranyl bromide. The diepoxy-allylic
bromide 29 was prepared beginning with regio- and enantiose-
lective epoxidation8 of farnesyl p-nitrobenzoate (27),13 with the
(9) For recent reviews, see: (a) de la Torre, M.; Sierra, M. A. Angew.
Chem., Int. Ed. 2004, 43, 160. (b) Yoder, R. A.; Johnston, J. N. Chem.
ReV. 2005, 105, 4730. For selected papers, see: (c) Corey, E. J.; Luo,
G.; Lin, L. S. Angew. Chem., Int. Ed. 1998, 37, 1126. (d) Corey, E. J.;
Luo, G.; Lin, L. S. J. Am. Chem. Soc. 1997, 119, 9927. (e) Zhang, J.;
Corey, E. J. Org. Lett. 2001, 3, 3215. (f) Mi, Y.; Schreiber, J. V.;
Corey, E. J. J. Am. Chem. Soc. 2002, 124, 11290. (g) Johnson, W. S.;
Chenera, B.; Tham, F. S.; Kullnig, R. K. J. Am. Chem. Soc. 1993,
115, 493. (h) Johnson, W. S.; Fletcher, V. R.; Chenera, B.; Bartlett,
W. R.; Tham, F. S.; Kullnig, R. K. J. Am. Chem. Soc. 1993, 115,
497. (i) Johnson, W. S.; Buchanan, R. A.; Bartlett, W. R.; Tham, F. S.;
Kullnig, R. K. J. Am. Chem. Soc. 1993, 115, 504. (j) Johnson, W. S.;
Plummer, M. S.; Reddy, S. P.; Bartlett, W. R. J. Am. Chem. Soc. 1993,
115, 515.
(11) (a) Fish, P. V.; Johnson, W. S. J. Org. Chem. 1994, 59, 2324. (b)
Fish, P. V. Tetrahedron Lett. 1994, 35, 7181.
(12) (a) Brook, A. G. J. Am. Chem. Soc. 1958, 80, 1886. (b) Brook, A. G.
Acc. Chem. Res. 1974, 7, 77. (c) Moser, W. H. Tetrahedron 2001,
57, 2065.
(10) (a) Corey, E. J.; Sodeoka, M. Tetrahedron Lett. 1991, 32, 7005. (b)
Corey, E. J.; Roberts, B. E. Tetrahedron Lett. 1997, 38, 8921. (c)
Corey, E. J.; Luo, G.; Lin, L. S. J. Am. Chem. Soc. 1997, 119, 9927.
(13) Tong, R.; Boone, M. A.; McDonald, F. E. J. Org. Chem. 2009, 74,
8407.
9
5302 J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010

Biomimetic Syntheses of Oxacyclic Triterpenes
A R T I C L E S
Scheme 2. Synthesis of Diepoxy-Enolsilane Substrates for ent-Abudinol B and ent-Muzitone Synthesesa
a Reagents and conditions: (a) D-epoxone (0.5 equiv), Oxone (2.8 equiv), K2CO3 (8.0 equiv), Bu4NHSO4 (0.1 equiv), Na2EDTA, Na2B4O7, DMM/MeCN
(2:1), H2O, 0 °C (49-61% yield); (b) K2CO3, MeOH (88% yield); (c) MsCl, Et3N, THF, -40 °C, then LiBr, THF, 0 °C (96% yield); (d) LDA, THF, -30
to -10 °C, then geranyl bromide, -30 to -10 °C; (e) NaOAc, HOAc, H2O/pentane (84% yield, 2 steps); (f) vinylmagnesium chloride, Et2O/THF, 0 °C
(88% yield); (g) n-BuLi, hexane/THF, -78 °C, then 29, -40 to -20 °C (50% yield); (h) LDA, THF, -30 to -10 °C, then 29, -30 to -10 °C; (i) NaOAc,
HOAc, H2O/pentane (74% yield, 2 steps); (j) vinylmagnesium bromide, Et2O/THF, -40 °C (99% yield); (k) n-BuLi, Et2O/THF, -78 °C, then geranyl
bromide, -20 °C (39% yield).
electron-withdrawing nature of the p-nitrobenzoate ester sup-
pressing epoxidation of the alkene bearing the allylic ester. The
diepoxide product 28 was then converted into the highly reactive
diepoxy-allylic bromide 29 by a standard sequence of ester
saponification, mesylation, and substitution with bromide.13
Acylsilane 31 was prepared by alkylation of geranyl bromide
with the metalloenamine derived from 30,14 followed by imine
hydrolysis to reveal the alpha-silyl ketone in 31. Addition of
vinylmagnesium chloride to the acylsilane 31 was followed by
the Brook rearrangement-in situ alkylation sequence pioneered
by Reich and Kuwajima,15 for incorporation of the diepoxy-
allylic bromide 29 to provide 12 bearing the silyloxy substituent
at C15 and 29 of the 30 carbons required for the synthesis of
abudinol B. The isomeric substrate 14 with a silyloxy substituent
at C14 was prepared by reordering the steps in this protocol,
so that initial alkylation of the metalloenamine derived from
30 with diepoxy-allylic bromide 29 afforded the acylsilane 32,
which was similarly subjected to vinylmagnesium bromide
addition and Brook rearrangement-in situ alkylation with
geranyl bromide to provide isomeric substrate 14.
Synthesis of ent-Abudinol B from Diepoxy-enolsilane 12. For
the first stage of biomimetic tricyclization (Scheme 3), we
observed that 1.1 equiv of TMSOTf selectively activated the
terminal epoxide of 12 and effectively promoted regio- and
stereoselective tandem tricyclization to provide the trans-anti-
trans-fused tricyclic ketone 33 as the major product, consistent
with anti-parallel addition and an expected chairlike conforma-
tion 34.7b To our surprise, substrate 12 was unreactive with
TBSOTf, in contrast to previous observations with compound
15 (Scheme 1).16 The best yield of 33 (50%, single diastereomer)
was achieved when the reaction was quenched with 1.1 equiv
of tetrabutylammonium fluoride at -78 °C within 10 min of
TMSOTf addition, whereas longer reaction times or aqueous
quench resulted in the generation of the C14-epimeric byproduct
37, perhaps through silyloxonium ion intermediates 35 and 36.
Bicyclic ether 38 and monocyclic ether 39 were also isolated
as minor products resulting from truncated cyclization and beta-
hydrogen elimination, processes that have been well documented
in polyene carbacyclizations.17
To explore the ultimate bicyclization of the hypothetical
biogenesis of ent-abudinol B from the enantiomer of compound
7 (Figure 2), our synthetic compound 33 underwent methyl-
enation of the C15-ketone to introduce the 30th carbon of the
triterpenoid structure (Scheme 4). After evaluating a variety of
reagents for this transformation,18 we found that the classical
Wittig reagent prepared in situ from a refluxing benzene solution
of methyltriphenylphosphonium bromide (Ph3PCH3Br) and
potassium tert-butoxide (KO-t-Bu)9c gave good yields of the
disubstituted alkene product, albeit with some epimerization at
C14 under the basic reaction conditions of the Wittig methyl-
enation. Due to the steric hindrance of the C10-methyl sub-
stituent, olefination of 33 proceeded more slowly than with
epimer 37, so that a mixture of C14-epimers 40 and 41 was
consistently produced, even when isomerically pure samples of
ketone 33 were used. However, diastereomers 40 and 41 were
separable by careful silica gel chromatography. Regio- and
stereoselective double epoxidations of each triene 40 and 41
provided the corresponding diepoxides ent-7r and ent-7ꢀ, by
careful control of the reaction temperature, concentration,
amount of D-epoxone, and reaction time. To the best of our
knowledge, this was the first example of epoxone-catalyzed
regioselective epoxidation of trisubstituted alkenes in the
presence of a gem-disubstituted alkene.
At the outset of this study, we were concerned that only one
of the diastereomers of ent-7r/ꢀ would have the proper
(17) For selected papers, see: (a) Stork, G.; Burgstahler, A. W. J. Am. Chem.
Soc. 1955, 77, 5068. (b) Eschenmoser, A.; Ruzicka, L.; Jeger, O.;
Arigoni, D. HelV. Chim. Acta 1955, 38, 1890. (c) Gamboni, G.; Schinz,
H.; Eschenmoser, A. HelV. Chim. Acta 1954, 37, 964. (d) van Tamelen,
E. E.; Leiden, T. M. J. Am. Chem. Soc. 1982, 104, 2061. (e) Ishihara,
K.; Ishibashi, H.; Yamamoto, H. J. Am. Chem. Soc. 2001, 123, 1505.
(f) Ishihara, K.; Ishibashi, H.; Yamamoto, H. J. Am. Chem. Soc. 2002,
124, 3647.
(14) (a) Nowick, J. S.; Danheiser, R. L. Tetrahedron 1988, 44, 4113. (b)
Corey, E. J.; Lin, S. J. Am. Chem. Soc. 1996, 118, 8765.
(15) (a) Reich, H. J.; Olson, R. E.; Clark, M. C. J. Am. Chem. Soc. 1980,
102, 1423. (b) Kato, M.; Mori, A.; Oshino, H.; Enda, J.; Kobayashi,
K.; Kuwajima, I. J. Am. Chem. Soc. 1984, 106, 1773. (c) Enda, J.;
Kuwajima, I. J. Am. Chem. Soc. 1985, 107, 5495.
(16) Other Lewis acids, including BF3-OEt2, MeAlCl2, and TBSCl/AgClO4,
resulted in lower yields of products or no reaction.
(18) Other methylenation reagents attempted included the Tebbe reagent,
Petasis reagent, and Julia tetrazole sulfone under various conditions.
9
J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010 5303

A R T I C L E S
Boone et al.
Scheme 3. First-Stage Tricyclization of Diepoxy-enolsilane 12
Scheme 4. Preparation of Diepoxyalkene Substrates ent-7r and ent-7ꢀa
a Reagents and conditions: (a) Ph3PCH3Br, KO-t-Bu, benzene, 80 °C (82% yield; from 33, ratio 40/41 ) 1:1.8; from 37, ratio 40/41 ) 1:2.5); (b)
D-epoxone (0.5 equiv), Oxone (2.9 equiv), K2CO3 (11 equiv), Bu4NHSO4 (0.1 equiv), Na2EDTA, Na2B4O7, DMM/MeCN (2:1), H2O, 0 °C (37% yield; 50%
combined yield after a second cycle with 0.5 equiv of reagents).
geometry for the formation of abudinol B, and thus the
epimerization of the C14 chiral center in the Wittig reaction of
ketone 33 was not entirely unwelcome. TMSOTf-promoted
reaction of ent-7r did provide ent-abudinol B (ent-2, Scheme
5), albeit in modest yield, and was accompanied by several
byproducts including the trisubstituted alkene isomer 43 as well
as the truncated monocyclization structure 44. The spectroscopic
properties of our synthetic product ent-2 matched the reported
literature data as well as direct comparison with another sample
of ent-abudinol B generated in our first generation synthesis,
which was confirmed by X-ray crystallography of the bis-silyl
ether of ent-abudinol B.7a The byproduct 44 presumably resulted
9
5304 J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010

Biomimetic Syntheses of Oxacyclic Triterpenes
A R T I C L E S
Scheme 5. Bicyclization of ent-7r to ent-Abudinol B (ent-2)
Scheme 6. Bicyclization of ent-7ꢀa
a Reagents and conditions: (a) TMSOTf (1.5 equiv), DTBMP (1.5 equiv), CH2Cl2, -78 °C; (b) Bu4NF, THF; (c) p-BrBzCl, DMAP, CH2Cl2.
from beta-hydrogen elimination from the C15-methyl substituent
after diepoxide cyclization. The mechanism for the tandem
bicyclization process to ent-1 and the trisubstituted alkene isomer
43 is consistent with the intermediacy of a C15 carbenium ion
46 or ion pair, perhaps due to the relatively poor nucleophilicity
of the gem-disubstituted C15-alkene of ent-7r.
Interestingly, TMSOTf-promoted bicyclization of the C14-
diastereomer ent-7ꢀ provided a mixture of trisubstituted alkenes
47 and 48 (Scheme 6), but no trace of ent-abudinol B or the
corresponding Z-tetrasubstituted alkene isomer of abudinol B
was found in the crude reaction mixture or in any of the isolated
byproducts. The structure of 48 was unambiguously confirmed
by X-ray crystallography of the bis-p-bromobenzoate derivative
49, which also clarified the structural assignments for the key
precursor compounds 40, 41, and ent-7r/ꢀ and provided a lead
structure for confidently assigning the structures of analogous
compounds such as 43 and 47.
the first chemical evidence for the viability of the biosynthetic
pathway proposed for abudinol B (2) from precursor compound
7.
Synthesis of the Structure Proposed for ent-Muzitone
from Diepoxy-enolsilane 14. We subsequently explored the
analogous tandem cyclizations beginning with the regioisomeric
substrate 14 bearing the silyloxy substituent on C14, thus making
C25 (muzitone numbering) the more nucleophilic terminus of
the diepoxide substrate. We perceived that this all-endo-
regioselective cyclization would be more challenging, due to
formation of a six-membered ring by closure of the C10-C25
bond, but would be a critical component of a synthesis of the
tetrasubstituted alkene precursor to muzitone. The first-stage
tricyclization of 14 was initiated with TMSOTf in the presence
of DTBMP at -78 °C (Scheme 7). This transformation was
somewhat slower than the corresponding transformation of 12,
as expected for six-membered ring formation, but the desired
product 50 was formed as an inseparable 1:1 mixture of C25
diastereomers, along with the monocyclization byproduct 51.19
Although it was not certain if this diastereomeric mixture
resulted from nonstereoselective formation of intermediate 54
or occurred in the course of the slow conversion of possible
bicyclic epoxonium ion 53 to tricyclic 54, we could convert
the mixture of C25 diastereomers by reaction with KOH in
refluxing methanol20 to provide a single C25 stereoisomer of
The dependence of the regioselectivity of alkene formation
on C14 stereochemistry was unexpected, especially since
examination of models had suggested that the C14-H14 bond
was better aligned with the C18-O bond of epoxide ent-7ꢀ vs
the diastereomer ent-7r. Moreover, if a C15-tertiary carbenium
ion 49 was indeed an advanced mechanistic intermediate, one
would have expected that some of the product distribution would
have included the tetrasubstituted alkene of abudinol B arising
from both substrate diastereomers ent-7r/ꢀ. Nonetheless, the
successful bicyclization of ent-7r to ent-abudinol B has provided
(19) A bicyclic compound bearing exo-methylene at C10 was also observed
as a minor byproduct. However, this byproduct could not be purified
and thus was not unambiguously characterized.
9
J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010 5305

A R T I C L E S
Boone et al.
Scheme 7. Tricyclization of Diepoxy-enolsilane 14
Scheme 8. Preparation of Diepoxyalkene 57a
a Reagents and conditions: (a) K2CO3, MeOH/THF (4:1), (89% yield); (b) Ph3PCH3Br, KO-t-Bu, THF, benzene, 70 °C (81% yield); (c) D-Epoxone (0.5
equiv), Oxone (2.6 equiv), K2CO3 (8 equiv), Bu4NHSO4 (0.1 equiv), Na2B4O7, DMM/MeCN (2:1), H2O, -5 °C (four cycles total); (d) TMSCl (2 equiv),
imidazole (4 equiv), CH2Cl2 (39% yield from 56).
50, which was fully purified and characterized as the p-
nitrobenzoate ester 55.
though the placement of the alkene relative to the proximal
epoxide required the formation of two seven-membered rings,
this transformation proceeded in relatively high yield, especially
when compared to our results in the abudinol synthesis.
Surprisingly, we found no trace of products corresponding to
the desired C14-C15 tetrasubstituted alkene (“ent-pre-muzi-
tone”, ent-11, Figures 2, 3), nor could we modify the reaction
conditions to provide this regioisomer, such as changes of
temperature (i.e., -40 °C) or reactant stoichiometry (i.e.,
equivalents of DTBMP). As we anticipated that a C14 carboca-
tionic intermediate would have given at worst a mixture of
alkene products including the desired tetrasubstituted isomer,
we suspected that the formation of trisubstituted alkene 58 was
the result of a concerted cyclization process, at least with regard
to formation of the C14-C19 bond and the C13-C14 alkene.
Thus, we focused on the isomerization of the trisubstituted
alkene of 58 to generate the desired alkene regioisomer. As we
anticipated that the alcohol and tertiary-substituted oxepane
functional groups would be sensitive to the strongly acidic
conditions for alkene migration, we protected the alcohols as
the diacetate 61 (Scheme 10), with the expectation that the
electron-withdrawing esters would inductively deactivate the
After saponification of the ester from compound 55, instal-
lation of the 30th carbon was accomplished by Wittig reaction
as described in our abudinol B synthesis, to provide triene 56
in good yield (Scheme 8). However, the regio- and enantiose-
lective epoxidation of triene 56 was a considerable challenge
and required careful monitoring by TLC to minimize overoxi-
dation of the gem-disubstituted alkene at C14. The epoxidation
reaction was stopped upon evidence of significant triepoxide
formation, and after isolation of the triene, monoepoxide, and
desired diepoxide via silica gel chromatography, the recovered
monoepoxide was resubjected to the Shi epoxidation conditions
following the same procedure, but modifying the stoichiometry
of K2CO3 and Oxone. Ultimately, this required four cycles of
the Shi epoxidation. Although this process was cumbersome,
the desired diepoxide 57 was cleanly obtained upon protection
of the alcohol as the trimethylsilyl ether derivative.
The second-stage bicyclization was conducted on diepoxy-
alkene 57, to provide pentacyclic structure 58 (Scheme 9). Even
(20) Zhang, J.; Corey, E. J. Org. Lett. 2001, 3, 3215.
9
5306 J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010

Biomimetic Syntheses of Oxacyclic Triterpenes
A R T I C L E S
Scheme 9. Bicyclization of Diepoxyalkene 57
Scheme 10. Isomerization of Trisubstituted Alkene 61 to Tetrasubstituted Alkene 62a
a Reagents and conditions: (a) Ac2O, DMAP, CH2Cl2 (91% yield); (b) HI, benzene, 70 °C, 1 h (95% yield); (c) K2CO3, MeOH, rt to reflux (100% yield).
nearby oxepanes from protonation. Compound 61 was also
crystalline, which allowed for crystallographic confirmation of
the bond connectivity and stereochemistry achieved in the
previous cyclization steps. Reaction of diacetate 61 with catalytic
HI in hot benzene21 resulted in the predominant formation of
the tetrasubstituted alkene isomer 62 (62:61 ) 20:1), and basic
methanolysis of the acetate esters afforded the enantiomer of
“pre-muzitone” (ent-11). With the spectra of ent-11 in hand,
we firmly demonstrated that this isomer was not produced in
detectable quantities in the course of the bicyclization of
diepoxyalkene 57. The structure of ent-11 was also confirmed
by crystallography.
The synthesis then concluded with oxidative cleavage of the
tetrasubstituted alkene of ent-11. Preliminary experiments
indicated that the diketone arising from oxidative cleavage of
diacetate ester 62 was susceptible to transannular aldol reaction22
under the basic conditions required for acetate removal, and
thus we recognized that C14-C25 bond cleavage to the diketone
would be the final step of our synthesis. Several different
approaches from diol 63 afforded the desired diketone ent-5,
albeit in low yields, including ozonolysis with reductive workup,
as well as the Johnson-Lemieux method for ruthenium-
catalyzed oxidative cleavage (Scheme 11).23 Ultimately, the best
yields were obtained by conducting the oxidative cleavage
process stepwise, beginning with initial osmylation of the
tetrasubstituted alkene of diacetate 62. This alkene reacted
sluggishly with osmium tetraoxide, and regardless of the
additives used (N-methylmorpholine-N-oxide or methanesulfona-
mide), this transformation required an excess of osmium
tetraoxide. Furthermore, the cyclic osmate ester was stable to
typical reduction reagents such as sodium bisulfite or sodium
thiosulfate and was subjected to silica gel chromatography
without degradation. However, the crude osmate ester from 62
was rapidly reduced with sodium borohydride in methanol,24
thus liberating the diacetate diol 63, which was deacetylated to
the corresponding tetraol 64 by basic methanolysis. Although
the oxidative cleavage of tetraol 64 with sodium periodate was
extremely sluggish (only a trace of ent-5 was produced after
48 h), lead tetraacetate promoted the rapid and clean oxidative
cleavage of 64 to afford diketone ent-5, which matched the
material obtained by ozone-promoted or ruthenium-catalyzed
oxidative cleavages of ent-11.
Unfortunately, we observed some discrepancies in our
characterization data for ent-5 when compared to those reported
for muzitone by the Kashman laboratory. Most obvious among
these differences was the report that the natural product was an
oil, whereas our synthetic material was a white crystalline solid
with a melting point of 210-213 °C. There were also differences
in the chemical shifts and coupling constants for some of the
1H NMR resonances. After some effort, we were able to solve
the crystal structure for the derived bis-p-nitrobenzoate deriva-
tive 65, confirming that we had indeed prepared the intended
structure ent-5, corresponding to the enantiomer of the structure
proposed for muzitone. Thus the Kashman laboratory’s assign-
ment for muzitone was incorrect.
Upon further consideration of key characterization details,
1
we focused on the discrepancies in the H NMR data. Specif-
ically, the four most deshielded hydrogens were the obvious
choice for comparisons. In particular, we were intrigued by the
coupling constant values of the hydrogen assigned at C19 for
the synthetic and natural materials. Our synthetic material
showed a chemical shift for the C19 hydrogen at 4.24 ppm (600
MHz, C6D6) and was a doublet of doublets with coupling
constants of 2.4 and 6.0 Hz; for instructive comparison, the shift
of this hydrogen in the pentacyclic precursor alkene ent-11 was
3.49 ppm (400 MHz, CDCl3) with coupling constants of 5.2
and 11.6 Hz. In contrast, the Kashman laboratory reported a
chemical shift of 3.88 ppm (500 MHz, C6D6) and coupling
(21) Marcos, I. S.; Pedrero, A. B.; Sexmero, M. J.; Diez, D.; Garc´ıa, N.;
Escola, M. A.; Basabe, P.; Conde, A.; Moro, R. F.; Urones, J. G.
Synthesis 2005, 3301.
(22) Herr, M. E.; Fonken, G. S. J. Org. Chem. 1967, 32, 4065.
(23) Pappo, R.; Allen, D. S.; Lemieux, R. U.; Johnson, W. S. J. Org. Chem.
1956, 21, 478.
(24) Herranz, E.; Sharpless, K. B. Org. Synth., Coll. Vol. VII 1990, 375.
9
J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010 5307

A R T I C L E S
Boone et al.
Scheme 11. Completion of the Synthesis of the Structure ent-5 Assigned to ent-Muzitonea
a Reagents and conditions: (a) OsO4 (2 equiv), THF/H2O (3:1), then NaBH4, MeOH (64% yield of 63 + 15% recovered 62); (b) K2CO3, MeOH/THF
(2.5:1), reflux (100% yield); (c) Pb(OAc)4, benzene (71% yield); (d) O3, CH2Cl2, -78 °C; then Me2S (21% yield); (e) RuCl3 (2 mol %), NaIO4, CCl4/
MeCN/H2O (2:2:3), 15 min (24% yield); (f) p-NO2BzCl, DMAP, CH2Cl2 (83% yield).
constants of 5.3 and 11.3 Hz. Considering the conformational
freedom achieved upon oxidative cleavage of the tetrasubstituted
alkene, our data were more consistent with the expected
conformational change, whereas Kashman’s coupling constant
data were more consistent with a rigid system, such as ent-11.
Nonetheless, the Kashman laboratory clearly isolated a diketone,
as they reported δ 218.0, 207.0 in C6D6 (we observed δ 216.3,
211.6 in C6D6 for compound ent-5). As an authentic sample of
muzitone is no longer available, it is not possible to determine
the correct structure at this time.
that the structure of muzitone has been misassigned, and thus
it is pointless to comment further on the biogenesis of muzitone
until its structure has been correctly determined. Nonetheless,
the synthetic approach has demonstrated that six- and even
seven-membered carbocyclic rings can be generated by tandem
bicyclizations of the appropriate substrates, thus providing an
important expansion of the types of structures that can be
prepared.
Acknowledgment. Initial stages of this work were funded by
the National Science Foundation (CHE-0516793). We thank Prof.
Yoel Kashman (Tel Aviv University) for correspondence and
spectroscopic data regarding abudinol B and muzitone.
Conclusion
In the course of achieving the total synthesis of the enantiomer
of the natural product abudinol B, we have demonstrated the
chemical viability of a proposed biogenetic pathway for this
structurally complex triterpenoid natural product, particularly
the final bicyclization stage which exactly matches the structural
elements that would be present in a natural product intermediate.
Unfortunately, our efforts to extend this concept to prepare a
hypothetical biosynthetic intermediate to muzitone have revealed
Supporting Information Available: Experimental details and
characterization of new compounds, including 1H and 13C NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA1006806
9
5308 J. AM. CHEM. SOC. VOL. 132, NO. 14, 2010