Total synthesis of (+)-asteltoxin

Article abstract of DOI:10.1021/ja034332q

A convergent synthesis of (+)-asteltoxin (1) has been achieved by the Horner-Emmons olefination of bis(tetrahydrofuran) aldehyde 53 and α-pyrone phosphonate 5. A key step features the stereoselective construction of a sterically congested quaternary center embedded in the densely functionalized bis(tetrahydrofuran) subunit by a Lewis acid-catalyzed, pinacol-type rearrangement of an epoxy silyl ether. This pivotal rearrangement methodology parallels the proposed biosynthetic pathway of 1 and is ripe for applications to the stereocontrolled synthesis of structurally complex natural products.

Full text of DOI:10.1021/ja034332q

Published on Web 04/09/2003
Total Synthesis of (+)-Asteltoxin
Khee Dong Eom, J. Venkat Raman, Heejin Kim, and Jin Kun Cha*
Contribution from the Department of Chemistry, UniVersity of Alabama,
Tuscaloosa, Alabama 35487
Received January 24, 2003; E-mail: jcha@chem.wayne.edu
Abstract: A convergent synthesis of (+)-asteltoxin (1) has been achieved by the Horner-Emmons
olefination of bis(tetrahydrofuran) aldehyde 53 and R-pyrone phosphonate 5. A key step features the
stereoselective construction of a sterically congested quaternary center embedded in the densely
functionalized bis(tetrahydrofuran) subunit by a Lewis acid-catalyzed, pinacol-type rearrangement of an
epoxy silyl ether. This pivotal rearrangement methodology parallels the proposed biosynthetic pathway of
1 and is ripe for applications to the stereocontrolled synthesis of structurally complex natural products.
Scheme 1. Vleggar’s Biosynthetic Proposal
Introduction
Asteltoxin (1) was isolated by Steyn, Vleggaar, and co-
workers from toxic maize cultures of Aspergillus stellatus
Curzi.1a Its structure, including relative stereochemistry, was
determined by spectroscopic methods and single-crystal X-ray
analysis,1 and the absolute configuration was subsequently
established by a partial synthesis starting with (R)-isopropylidene
glyceraldehyde.2c This mycotoxin belongs to a group of
structurally related trienic R-pyrones, such as citreoviridin,
verrucosidin, and the aurovertins, which are known to function
as inhibitors of oxidative phosphorylation.3,4 Asteltoxin was later
shown to possess similar inhibitory activity of E. coli BF1-
ATPase and to provide a fluorescent probe of mitochondrial
F1- and bacterial BF1-ATPase.4b Unlike other R-pyrone myco-
toxins of the same family, 1 is characterized by the presence of
a unique, highly functionalized 2,8-dioxabicyclo[3.3.0]octane
containing a quaternary carbon embedded in an array of six
stereogenic centers. On the basis of extensive 13C and 18O
labeling experiments, Vleggaar advanced the biosynthesis of 1
involving polyepoxidation of a linear polyene precursor; his
intriguing postulate for formation of the bis(tetrahydrofuran)
moiety featured an epoxide-mediated 1,2-alkyl shift of a
polyketide chain to generate a branched aldehyde (Scheme
1).1b-d A related pinacol-like rearrangement was also implicated
in the oxidative rearrangement of (+)-averufin to versiconal
acetate in the biosynthetic pathway of aflatoxins B1 and B2, in
which the branched aldehyde of versiconal acetate was derived
from the straight side chain of (+)-averufin.5,6 The amalgam-
ation of the fascinating architecture, unusual biogenesis, and
interesting biological activity of 1 has attracted considerable
* To whom correspondence should be addressed at the Department of
Chemistry, Wayne State University, 5101 Cass Ave., Detroit, MI 48202.
(1) (a) Kruger, G. J.; Steyn, P. S.; Vleggaar, R.; Rabie, C. J. J. Chem. Soc.,
Chem. Commun. 1979, 441. (b) Steyn, P. S.; Vleggarr, R. J. Chem. Soc.,
Chem. Commun. 1984, 977. (c) de Jesus, A. E.; Steyn, P. S.; Vleggarr, R.
J. Chem. Soc., Chem. Commun. 1985, 1633. (d) Vleggaar, R. Pure Appl.
Chem. 1986, 58, 239.
(2) (a) Schreiber, S. L.; Satake, K. J. Am. Chem. Soc. 1983, 105, 6723. (b)
Schreiber, S. L.; Satake, K. J. Am. Chem. Soc. 1984, 106, 4186. (c)
Schreiber, S. L.; Satake, K. Tetrahedron Lett. 1986, 27, 2575.
(3) Synthesis of (-)-citreoviridin: (a) Nishiyama, S.; Shizuri, Y.; Yamamura,
S. Tetrahedron Lett. 1985, 26, 231. (b) Williams, D. R.; White, F. H. J.
Org. Chem. 1987, 52, 5067. (c) Suh, H.; Wilcox, C. S. J. Am. Chem. Soc.
1988, 110, 470. (d) Bowden, M. C.; Patel, P.; Pattenden, G. J. Chem. Soc.,
Perkin Trans. 1 1991, 1947. (e) Whang, K.; Venkataraman, H.; Kim, Y.
G.; Cha, J. K. J. Org. Chem. 1991, 56, 7174. Synthesis of (-)-aurovertin
B: (f) Nishiyama, S.; Toshima, H.; Kanai, H.; Yamamura, S. Tetrahedron
Lett. 1986, 27, 3643. Synthesis of (+)-verrucosidin: (g) Hatakeyama, S.;
Sakurai, K.; Numata, H.; Ochi, N.; Takano, S. J. Am. Chem. Soc. 1988,
110, 5201. (h) Whang, K.; Cooke, R. J.; Okay, G.; Cha, J. K. J. Am. Chem.
Soc. 1990, 112, 8985. Synthesis of (+)-citreoviral: (i) Hanaki, N.; Link,
J. T.; MacMillan, D. W. C.; Overman, L. E.; Trankle, W. G.; Wurster, J.
A. Org. Lett. 2000, 2, 223. (j) Peng, Z.-H.; Woerpel, K. A. Org. Lett. 2002,
4, 2945 and references therein for previous syntheses.
(4) (a) Linnett, P. E.; Beechey, R. B. Methods Enzymol. 1979, 55, 472. (b)
Satre, M. Biochem. Biophys. Res. Commun. 1981, 100, 267.
9
10.1021/ja034332q CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 5415-5421
5415

A R T I C L E S
Eom et al.
Scheme 2. Retrosynthetic Analysis
synthetic interest that has culminated in two total syntheses:
Schreiber’s first synthesis utilized an innovative application of
the [2 + 2] furan-carbonyl photocycloaddition.2 Takano and
co-workers employed a D-glucose-based chiron approach in the
second synthesis.7 Stereoselective syntheses of the bis(tetrahy-
drofuran) centerpiece have been achieved by two other groups8,9
and also in our laboratory.10a,b We herein report the details of
our synthetic studies leading to an enantioselective synthesis
of (+)-1.10c
Results and Discussion
Retrosynthetic Analysis. Our initial task focused on the
enantio- and diastereoselective preparation of the unusual bis-
(tetrahydrofuran) core 2 or 3 for eventual coupling with
phosphonate 4 or 5 for a convergent synthesis of (+)-1. Inspired
for preparing enantiomerically pure or enriched 2,3-epoxy
alcohols.14 The requisite substrate 7 for the Tsuchihashi-Suzuki
rearrangement (7 f 6) was expected to be readily available by
employing a straightforward sequence of well-precedented
transformations involving 8. In comparison, the Yamamoto
rearrangement (10 f 9) required the preparation of threo-
epoxide 10, which was deemed to be more challenging and
could possibly entail a mismatched case of the Sharpless
asymmetric epoxidation, depending on the choice of a side
chain. These considerations thus prompted us to investigate the
epoxy silyl rearrangement by the method of Tsuchihashi and
Suzuki. During the course of our synthetic investigations, this
rearrangement has received renewed attention by other labora-
tories, and further progress in the methodology development
for preparing various â-hydroxy carbonyls and 1,3-diols was
reported in the literature.15-17 In passing, we also note that Jung
by Vleggaar’s biosynthetic postulate, we were attracted to an
epoxide-mediated pinacol-type rearrangement approach. Par-
ticularly alluring was the preparative power of the underlying
methodology for the convenient, enantioselective construction
of quaternary carbons starting with readily available, enantio-
merically pure epoxides.11 Among the known repertoire of
stereoselective 1,2-rearrangement reactions of epoxides and their
derivatives, the Tsuchihashi-Suzuki12 and Yamamoto13 pro-
cedures seemed particularly well suited for an enantioselective
synthesis of 2 or 3 in close parallel with the proposed biogenesis
(Scheme 2). At the inception of our synthetic studies, the
Sharpless asymmetric epoxidation of allylic alcohols was
demonstrated to be one of the most general and reliable methods
(12) (a) Maruoka, K.; Hasegawa, M.; Yamamoto, H.; Suzuki, K.; Shimazaki,
M.; Tsuchihashi, G.-i. J. Am. Chem. Soc. 1986, 108, 3827. (b) Suzuki, K.;
Shimazaki, M.; Tsuchihashi, G.-i. Tetrahedron Lett. 1986, 27, 6237. (c)
Suzuki, K.; Matsumoto, T.; Tomooka, K.; Matsumoto, K.; Tsuchihashi,
G.-i. Chem. Lett. 1987, 113. (d) Suzuki, K.; Miyazawa, M.; Tsuchihashi,
G.-i. Tetrahedron Lett. 1987, 28, 3515. (e) Shimazaki, M.; Hara, H.; Suzuki,
K.; Tsuchihashi, G.-i. Tetrahedron Lett. 1987, 28, 5891. (f) Suzuki, K.;
Miyazawa, M.; Shimazaki, M.; Tsuchihashi, G.-i. Tetrahedron 1988, 44,
4061. (g) Suzuki, K. J. Synth. Org. Chem. Jpn. 1988, 46, 49. (h) Shimazaki,
M.; Hara, H.; Suzuki, K. Tetrahedron Lett. 1989, 30, 5443. (i) Shimazaki,
M.; Morimoto, M.; Suzuki, K. Tetrahedron Lett. 1990, 31, 3335. (j)
Nagasawa, T.; Taya, K.; Kitamura, M.; Suzuki, K. J. Am. Chem. Soc. 1996,
118, 8949. Cf. (k) Cheer, C. J.; Johnson, C. R. J. Am. Chem. Soc. 1968,
90, 178.
(13) (a) Maruoka. K.; Ooi, T.; Yamamoto, H. J. Am. Chem. Soc. 1989, 111,
6431. (b) Maruoka. K.; Nagahara, S.; Ooi, T.; Yamamoto, H. Tetrahedron
Lett. 1989, 30, 5607. (c) Maruoka. K.; Bureau, R.; Yamamoto, H. Synlett
1991, 363. (d) Maruoka. K.; Ooi, T.; Nagahara, S.; Yamamoto, H.
Tetrahedron 1991, 47, 6983. (e) Maruoka. K.; Sato, J.; Yamamoto, H. J.
Am. Chem. Soc. 1991, 113, 5449. (f) Maruoka. K.; Ooi, T.; Yamamoto, H.
Tetrahedron 1992, 48, 3303. (g) Maruoka. K.; Sato, J.; Yamamoto, H.
Tetrahedron 1992, 48, 3749. (h) Ooi, K.; Maruoka. K.; Yamamoto, H.
Org. Synth. 1995, 72, 95. See also: (i) Nemoto, H.; Ishibashi, H.;
Nagamochi, M.; Fukumoto, K. J. Org. Chem. 1992, 57, 1707. (j) Nemoto,
H.; Nagamochi, M.; Ishibashi, H.; Fukumoto, K. J. Org. Chem. 1994, 59,
74.
(5) (a) Sankawa, U.; Shimada, H.; Kobayashi, T.; Ebizuka, Y.; Yamamoto,
Y.; Noguchi, H.; Seto, H. Heterocycles 1982, 19, 1053. (b) Townsend, C.
A.; Christensen, S. B. J. Am. Chem. Soc. 1985, 107, 270. For an excellent
review on the biosynthesis of aflatoxins, see: (c) Minto, R. E.; Townsend,
C. A. Chem. ReV. 1997, 97, 2537.
(6) No information is presented regarding the biosynthetic details of the key
oxidative rearrangement, but insightful in vitro model studies have been
reported: (a) Townsend, C. A.; Davis, S. G.; Koreeda, M.; Hulin, B. J.
Org. Chem. 1985, 50, 5428. (b) Townsend, C. A.; Isomura, Y.; Davis, S.
G.; Hodge, J. A. Tetrahedron 1989, 45, 2263.
(7) (a) Tadano, K.-i.; Yamada, H.; Idogaki, Y.; Ogawa, S.; Suami, T.
Tetrahedron Lett. 1988, 29, 655. (b) Tadano, K.-i.; Yamada, H.; Idogaki,
Y.; Ogawa, S.; Suami, T. Tetrahedron 1990, 46, 2353.
(8) Nishiyama, S.; Kanai, H.; Yamamura, S. Bull. Chem. Soc. Jpn. 1990, 63,
1322.
(9) Mulzer, J.; Mohr, J.-T. J. Org. Chem. 1994, 59, 1160.
(14) Finn, M. G.; Sharpless, K. B. In Asymmetric SynthesissChiral Catalysis;
Morrison, J. D., Ed.; Academic Press: Orlando, FL, 1985; Vol. 5, Chapter
8.
(10) (a) Raman, J. V.; Lee, H. K.; Vleggaar, R.; Cha, J. K. Tetrahedron Lett.
1995, 36, 3095. Taken in part from the following: (b) Kim, H. Ph.D.
Dissertation, The University of Alabama, 1997. (c) Eom, K. D. Ph.D.
Dissertation, The University of Alabama, December 2000.
(11) For excellent reviews, see: (a) Fuji, K. Chem. ReV. 1993, 93, 2037. (b)
Corey, E. J.; Guzman-Perez, A. Angew. Chem., Int. Ed. 1998, 37, 388. (c)
Christoffers, J.; Mann, A. Angew. Chem., Int. Ed. 2001, 40, 4591. (d)
Martin, S. F. Tetrahedron 1980, 36, 419.
(15) For a recent review of rearrangements of epoxy alcohols, see: (a)
Magnusson, G. Org. Prep. Proced. Int. 1990, 22, 547. (b) Rickborn, B. In
ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Eds.;
Pergamon: Oxford, 1991; Vol. 3, pp 733-775.
(16) (a) Marson, C. M.; Walker, A. J.; Pickering, J.; Hobson, A. D.; Wriggles-
worth, R.; Edge, S. J. J. Org. Chem. 1993, 58, 5944. (b) Marson, C. M.;
Harper, S.; Walker, A. J.; Pickering, J.; Campbell, J.; Wrigglesworth, R.;
Edge, S. J. Tetrahedron 1993, 49, 10339.
9
5416 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Total Synthesis of (
+
)-Asteltoxin
A R T I C L E S
Scheme 3 Diasteroselective Rearrangement of Epoxide 17b
increase (∼5:1; 70% yield) was observed by employing L-(+)-
diisopropyl tartrate. More significantly, it became apparent that
the acetonide group [e.g., 17a] was incompatible with a Lewis
acid required to induce the key Tsuchihashi-Suzuki rearrange-
ment (vide infra). Thus, silyl protecting groups were chosen
because of their anticipated robustness toward Lewis acids.
Removal of the acetonide protecting group of 12 with HCl/
THF afforded diol 13 (85%), which was sequentially protected
by standard methods to give 14 (88%). Subsequent DIBAL
reduction provided alcohol 15b in 90% yield. The Sharpless
asymmetric epoxidation of 15b furnished epoxy alcohol 16b
in 80-88% yield and 94% diastereoselectivity. PCC or TPAP
oxidation of 16b, followed by addition of 2-propenylmagnesium
bromide and silylation with TMSCl, provided the rearrangement
substrate 17b in 70% overall yield. For convenience, com-
mercially available 2-propenylmagnesium bromide was em-
ployed instead of 2(E)-pentenylmagnesium bromide in the first-
generation synthesis. The pivotal rearrangement of the epoxy
silyl ether 17b was accomplished by the action of TiCl4 to
provide aldehyde 18 in 90% yield, which was found to be
surprisingly robust and well behaved. Since both epimers
smoothly underwent the acid-catalyzed pinacol-type rearrange-
ment, no attempt was made to enhance the diastereoselectivity
of addition of the Grignard reagent to the aldehyde. As noted
above, the cognate rearrangement of 17a in the presence of TiCl4
or SnCl4 gave only poor (40-45%) yields of the desired product.
Conversion of the aldehyde 18 to the corresponding acetal
19 was achieved by transacetalization with 2,2-dimethyl-1,3-
dioxolane or conventional acetalization with ethylene glycol,
and subsequent protection of 19 with p-methoxybenzyl chloride
gave 20 in 80% overall yield. No conditions were found for
selective removal of the TBDMS group from 20. Both silyl
protecting groups were thus removed by using n-Bu4NF, and
subsequent treatment with methanolic hydrogen chloride cleanly
gave 21 as a single diastereomer. A fully protected acetal, 22,
was then obtained in 80% overall yield by treatment of 21 with
TIPSCl and imidazole.
subsequently developed a useful variant of the Yamamoto
rearrangement involving vinyl epoxides and that 10 f 9 could
be considered an example of the Jung rearrangement.18 Recent
impressive advances in enantioselective epoxidation reactions
of (Z)- and (E)-olefins, namely, the Jacobsen and Shi epoxida-
tions,19,20 would certainly allow several variants of these
stereoselective 1,2-epoxide rearrangements to be synthetically
viable.21
Our attention was next directed to the stereocontrolled
attachment of the left-hand tetrahydrofuran moiety. Osmylation
of 22 took place with complete stereoselectivity to give diol 23
as an anomeric mixture in 85% yield (Scheme 4). Extensive
scrambling at the anomeric center was observed under the
dihydroxylation conditions, but the other possible stereoisomers
were not found in the crude reaction mixture. The stereochemical
assignment of the dihydroxylation products was initially made
on the basis of difference NOE measurements of the cyclization
product 25 (and also 26, vide infra) and was unequivocally
confirmed by its ultimate conversion to 2. The origin of the
observed diastereoselectivity in osmylation is discussed in detail
later. Swern oxidation of 23 and subsequent chelation-controlled
addition of EtMgBr to the resulting aldehyde allowed an
efficient, stereoselective introduction of the ethyl side chain to
provide 24 as a single epimer at the newly generated stereo-
center, but as a 1.3:1 anomeric mixture. Acid-catalyzed cy-
clization of 24 with p-TsOH or CSA in CH2Cl2 then afforded
bis(tetrahydrofuran) 25 in 70% overall yield (from 23). As noted
by Mulzer,9 a large difference in the rate of cyclization was
found between these anomers when a small (10 mol %) amount
of an acid catalyst was employed. This difference in rate could
First-Generation Synthesis of 2. Our synthesis commenced
with the known and readily available allylic alcohol 15a
(Scheme 3). However, acetonide 15a had previously been shown
to be one of the very rare E-allylic alcohols among poor
substrates for the Sharpless asymmetric epoxidation: use of (+)-
DET had been reported to give the desired diastereomer 16a in
4:1 diastereoselectivity (75% yield),14 and only a modest
(17) (a) Tu, Y. Q.; Sun, L. D.; Wang, P. Z. J. Org. Chem. 1999, 64, 629. (b)
Fan, C.-A.; Wang, B.-M.; Tu, Y.-Q.; Song, Z.-L. Angew. Chem., Int. Ed.
2001, 40, 3877.
(18) (a) Jung, M. E.; D’Amico, D. C. J. Am. Chem. Soc. 1995, 117, 7379. (b)
Jung, M. E.; Marquez, R. Org. Lett. 2000, 2, 1669. See also: (c) Kita, Y.;
Kitagaki, S.; Yoshida, Y.; Mihara, S.; Fang, D.-F.; Kondo, M.; Okamoto,
S.; Imai, R.; Akai, S.; Fujioka, H. J. Org. Chem. 1997, 62, 4991.
(19) (a) Allain, E. J.; Hager, L. P.; Deng, L.; Jacobsen, E. N. J. Am. Chem. Soc.
1993, 115, 4415. (b) Jacobsen, E. N. In Catalytic Asymmetric Synthesis;
Ojima, I., Ed.; VCH: Weinheim, Germany, 1993; Chapter 4.2. (c) Annis,
D. A.; Jacobsen, E. N. J. Am. Chem. Soc. 1999, 121, 4147.
(20) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J.-R.; Shi, Y. J. Am. Chem. Soc.
1997, 119, 11224.
(21) For recent synthetic applications of the Jacobsen and Shi epoxidations in
the synthesis of structurally complex natural products, see: (a) Xiong, Z.;
Corey, E. J. J. Am. Chem. Soc. 2000, 122, 9328. (b) Burova, S. A.;
McDonald, F. E. J. Am. Chem. Soc. 2002, 124, 8188. (c) 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.
9
J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5417

A R T I C L E S
Eom et al.
Scheme 4. Stereoselective Synthesis of Bis(tetrahydrofuran) 2
Scheme 5. Stereoselective Synthesis of 36 and 37
be attributed to the stereoelectronic control, where the R-anomer
could more readily adopt the requisite conformation with an
electron pair of the ring oxygen antiperiplanar to the leaving
group.22 By utilizing 1 equiv of p-TsOH, both anomers could
be cyclized to 25 conveniently within 2 h, but the reaction had
to be carefully monitored to avoid the unwanted cleavage of
the PMB group. Finally, deprotection [(1) TBAF; (2) H2, Pd/
C)] gave bis(tetrahydrofuran) 2 (R1 ) R2 ) H), the spectral
data of which were in excellent agreement with literature
values.2a,8,9 For additional characterization, 26 was also con-
verted into diene ester 28.2a,7 Thus, Swern oxidation of 26 and
subsequent olefination with trimethyl 4-phosphonocrotonate or
the corresponding ethyl ester23 furnished diene ester 27, along
with a small amount of its C-8 (asteltoxin numbering) epimer.
Deprotection of the PMB group with DDQ then furnished ester
28, which exhibited spectral characteristics identical to literature
values.2a,7b,9
Second-Generation Synthesis of 2. Use of 2(E)-pentenyl-
magnesium bromide in place of 2-propenylmagnesium bromide
in the preparation of the requisite rearrangement substrate (i.e.,
16b f 17b) should streamline the above-mentioned synthesis
of 2 by eliminating several transformations which were neces-
sary to introduce the ethyl side chain. Additionally, the
diastereoselectivity of osmylation of the resulting E-trisubstituted
olefin was anticipated to be comparable to that of the corre-
sponding isopropenyl moiety (e.g., 22). Toward this end, we
undertook the second-generation synthesis of 2 and planned to
further reduce attendant protection/deprotection steps as well.
2(E)-Pentenyllithium is known to be readily available from
transmetalation of 2(E)-pentenyl(tributyl)stannane (29a), which
was in turn prepared starting with the trisylhydrazone of
acetone;24 in our hands, the reported sequence of transformations
was capricious, and more importantly, it proved to be very
difficult to obtain pure 29a free from impurities. We thus
developed a convenient, preparative-scale route to (E)-2-bromo-
2-pentene (29b) by relying on reiteration of the bromination-
decarboxylative elimination sequence on (E)-2-methyl-2-
pentenoic acid.25 Since it was unnecessary to differentiate two
silyl protecting groups (R4 and R5), our second-generation
synthesis began with bis(TIPS) ether 16c (Scheme 5). Sequential
treatment of 29b with 2.0 equiv of tert-BuLi and 1.0 equiv of
MgBr2, followed by addition of the aldehyde derived from 16c,
cleanly afforded 30, silylation (with TMSCl) of which provided
the rearrangement substrate 31 in 85% overall yield. In parallel
with 16b f 17b, treatment of the epoxy silyl ether 31 with
TiCl4 gave aldehyde 32 in 96% yield. By adaptation of the
previously developed procedure for 17b, the aldehyde 32 was
then converted to the fully protected tetrahydrofurans 36 and
37 in good overall yield; surprisingly, protection of 32 by means
of transacetalization with 2,2-dimethyl-1,3-dioxolane was con-
siderably slower than that of 18 at room temperature. Direct
acetal formation with ethylene glycol was instead achieved under
standard conditions (a catalytic amount of p-TsOH, benzene,
reflux) without competing desilylation in excellent (95%) yield.
Upon exposure to methanolic HCl, 32 gave the unprotected
tetrahydrofuran 38 as a 2:1 mixture of the two anomers in 50%
(unoptimized) yield.
Osmylation of the trisubstituted olefin 36 was drastically
slower than that of the respective disubstituted olefin 22 (Scheme
6). For example, in marked contrast to facile dihydroxylation
(4 h, room temperature; 85% yield) of 22, osmylation of 36
took 2 weeks at room temperature (0.15 equiv of OsO4 and
2-3 equiv of NMO in 2:1 acetone-water) or 2 days at 50 °C
(in 2:1 acetonitrile-water). Osmylation of the free alcohol 35
proceeded more slowly than that of 36 or 37. More surprisingly,
diastereoselectivity of these osmylation reactions was found to
(22) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry; Perga-
mon: Oxford, 1983.
(23) Cf. (a) Lythgoe, B.; Moran, T. A.; Nambudiry, M. E. N.; Ruston, S. J.
Chem. Soc., Perkin Trans. 1 1976, 2386. (b) Moune´, S.; Niel, G.; Busquet,
M.; Eggleston, I.; Jouin, P. J. Org. Chem. 1997, 62, 3332.
(24) Cooke, M. P. Jr. J. Org. Chem. 1982, 47, 4963.
(25) Kim, H.; Lee, S.-K.; Lee, D.; Cha, J. K. Synth. Commun. 1998, 28, 729.
9
5418 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Total Synthesis of (
+
)-Asteltoxin
A R T I C L E S
Scheme 6. Osmylation Reactions of Tetrahydrofurans 35-37
Scheme 7. Second-Generation Synthesis of 26
double bond. Lack of diastereocontrol in the osmylation
reactions of 38 and Takano’s example 41 is consistent with the
allylic strain-based rationalization. These results suggested that
the presence of both groups was essential for high diastereo-
selectivitity in osmylation. On the other hand, the unusually
sluggish dihydroxylation of the trisubstituted olefins 35-37 and
the observed low diastereoselectivities were incongruent with
the Mulzer model; they were unexpected and perplexing,
especially because these olefins are tantalizingly similar to
Mulzer’s substrate 40; the origin for the striking divergence
between disubstituted olefin 22 and related trisubstituted olefins
35-37 in osmylation is unclear at present, whereas difference
NOE measurements of 22 and 36 indicated their similar
conformational preference. Subtle, yet unidentified, factors seem
to play an important role in determining the stereochemical
outcome of these dihydroxylation reactions. However, these
factors were not thoroughly examined because of the successful
outcome of an alternate approach.
This disappointing result of the key osmylation reaction
prompted us to perform the Sharpless asymmetric dihydroxy-
lation (AD)27 prior to formation of the tetrahydrofuran ring
(Scheme 7). The Sharpless AD reaction of 34 with AD-mix-â
was not encouraging, and the desired diol was isolated in only
poor yield from a complex reaction mixture. However, when
both silyl protecting groups were first removed, the resulting
diol 42 was found to be an excellent substrate for the Sharpless
AD reaction, which proceeded with a 10:1 diastereoselectivity.
The tetrol 43 was isolated in 74% yield after purification by
column chromatography. Interestingly, the Sharpless AD reac-
tion of 42 with AD-mix-R was less stereoselective (a 1:6
diastereoselectivity to give 26 and 39a, respectively, after
cyclization) and less clean (43%). Subsequent treatment of 43
with methanolic HCl gave the desired bis(tetrahydrofuran) 26
in nearly quantitative yield. The Sharpless AD reactions of 35
and 36 were not diastereoselective and offered no particular
advantage over the above-mentioned catalytic osmylation reac-
tions. Overall, the second generation of 26 was thus achieved
efficiently by utilizing the Sharpless AD reaction in nine steps
from 16c (in 45% overall yield).
be not only modest, but also dependent on R3 (e.g., 35-37)
and reaction temperature; since all four possible diastereomers
were produced in each case, the crude reaction mixtures were
subjected, after partial purification (instead of separation and
individual characterization of each isomer), to acid-mediated
cyclization with CSA in CH2Cl2, to afford two diastereomeric
products (i.e., 25/26 and 39a-c) in low diastereoselectivity, thus
negating any potential advantage of directly introducing the
2(E)-pentenyl group. Osmylation of diol 38 proved to be
nonstereoselective and also suffered from poor yields.
This stereorandom dihydroxylation of 38 might be suggestive
of the stereodirecting effect of the alkoxy substituent at C-7
(asteltoxin numbering), presumably as a consequence of allylic
strain.26 Other laboratories reported osmylation of the related
compounds 40 and 41, as summarized in Scheme 6. The Mulzer
group put forward an attractive rationalization on the basis of
difference NOE measurements: the reactive conformer was
believed to be A, where osmium tetroxide was expected to
approach the double bond away from the aryloxy group. The
minor conformer B would suffer from nonbonding interactions
between the aryloxy group and the methyl substituent on the
Total Synthesis of (+)-1. Primarily because of the ready
availability of phosphonates 4a and 4b, which had been prepared
in our laboratory during our total synthesis of (-)-citreoviridin,3e
we first examined the final coupling of 4b and the aldehyde
(26) (a) Johnson, F. Chem. ReV. 1968, 68, 375. (b) Hoffmann, R. W. Chem.
ReV. 1989, 89, 1841.
(27) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV. 1994,
94, 2483.
9
J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5419

A R T I C L E S
Eom et al.
Scheme 8. Studies on Final Coupling
Scheme 9. Total Synthesis of (+)-1
47 in 79% yield, along with less than 5% of its easily separable
epimer. On the other hand, use of carbethoxy(triphenyl)-
phosphorane gave a 1:1 mixture of 47 and its C-8 epimer.
Aldehyde 50 was then prepared by standard methods in three
straightforward steps and was found to be configurationally
stable. The final union of the two segments was achieved by
treatment of phosphonate 53e with LiHMDS, followed by
addition of aldehyde 50 at -78 °C; the product yield was
1
estimated to be 80% on the basis of H NMR analysis of the
crude reaction mixture, but we were disappointed that removal
of an excess amount of the R-pyrone partner 5 by chromatog-
raphy was an acutely onerous task and that pure (+)-1 was
isolated in only 35% yield.
44, obtained by Swern oxidation of 26, to provide the PMB-
protected asteltoxin derivative 45 admixed with a small amount
of its epimer (structure not shown) at C-8 (Scheme 8).
Unfortunately, the PMB protecting group could not be removed
from 45, but instead extensive decomposition took place on
exposure to DDQ. This result was not surprising in view of the
presumably facile oxidation of the trienic R-pyrone moiety,
coupled with the location of the PMB group in the hindered,
concave face of the bis(tetrahydrofuran) subunit. Of some
concern was the observation that the aldehyde 44 was somewhat
prone to epimerization, especially under basic conditions (see
also 27 in Scheme 4). In view of the lability of 44, it seemed
prudent to eschew the coupling A approach (i.e., 2 + 4), but to
adopt coupling B (i.e., 3 + 5). It is noteworthy that both of the
previous two syntheses of 1 relied on a third tactic of utilizing
the bis(tetrahydrofuran) subunit containing a diene unit.2b,7b In
both syntheses, moreover, the introduction of the diene func-
tionality was carried out prior to the construction of the bis-
(tetrahydrofuran) moiety, thus bypassing potential complication
due to epimerization.
To preclude laborious chromatographic purification, the use
of only a stoichiometric amount of 5 was necessary, which in
turn prompted us to prepare a fully protected derivative of 50.
The presence of a free alcohol had previously been found to be
detrimental to a similar coupling reaction in the synthesis of
citreoviridin.3b,c,e A straightforward sequence of functional group
manipulation gave aldehyde 53 in good overall yield starting
with 48 (Scheme 9). The final union of 53 and 5 in the presence
of LiHMDS furnished a bis(trimethylsilyl)-protected asteltoxin
derivative, 54, free from the epimerization product, in 88% yield.
Finally, both alcohols were unmasked by TBAF to afford, in
87-95% yield, (+)-asteltoxin (1), the spectral data and chro-
matographic properties of which were identical to the literature
values.28
Future Studies
This work illustrates the synthetic utility of the stereoselective
1,2-rearrangement of readily available 2,3-epoxy alcohols or
Toward this end, 3 (e.g., 47-49) was next prepared virtually
free from epimerization under carefully controlled conditions:
Swern oxidation of 26 and subsequent Horner-Emmons ole-
fination of 44 with phosphonate 46 provided the desired ester
(28) The 1H NMR spectrum of natural (+)-asteltoxin was kindly provided by
Professor Vleggaar, but unfortunately an authentic sample is no longer
available due to extensive decomposition during storage.
9
5420 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Total Synthesis of (
+
)-Asteltoxin
A R T I C L E S
silyl ethers in an efficient construction of a new quaternary
center embedded in a challenging array of multiple stereocenters
in an easily predictable and well-defined configuration. The
mechanistic details of the underlying pinacol rearrangement and
its several variants have been investigated through extensive
and imaginative studies in the laboratories of Tsuchihashi-
Suzuki, Yamamoto, Jung, Fukumoto-Nemoto, and others.12-18
Compared to innovation in the methodology development,
applications of these powerful rearrangement reactions of
epoxides in natural product synthesis lag behind. A few
spectacular examples notwithstanding,29 a paucity of these
synthetic applications is surprising, especially in view of recent
dazzling advances in enantioselective epoxidation reactions.
(tetrahydrofuran) aldehyde 53 and R-pyrone phosphonate 5. A
sterically congested quaternary center embedded in the densely
functionalized bis(tetrahydrofuran) subunit was stereoselectively
assembled by a Lewis acid-catalyzed, pinacol-type rearrange-
ment of epoxy silyl ethers. This pivotal rearrangement strategy
is analogous to the proposed biosynthetic pathway of 1. Further
applications of the reliable 1,2-rearrangement reactions of
enantiomerically pure epoxides to the total synthesis of structur-
ally complex natural products are currently in progress.
Acknowledgment. This work is dedicated to Professor Drury
S. Caine. We thank the National Institutes of Health (Grant
GM35956) for generous financial support.
Conclusion
In summary, we have achieved a convergent synthesis of (+)-
asteltoxin (1) by the Horner-Emmons olefination of bis-
Supporting Information Available: Experimental procedures
and spectral data for new intermediates (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
(29) Two recent notable examples are Suzuki’s synthesis of furaquinocins and
Harran’s synthesis of nominal diazonamide A featuring an epoxy alcohol
rearrangement and a pinacol rearrangement, respectively: (a) Saito, T.;
Suzuki, T.; Morimoto, M.; Akiyama, C.; Ochiai, T.; Takeuchi, K.;
Matsumoto, T.; Suzuki, K. J. Am. Chem. Soc. 1998, 120, 11633. (b) Li, J.;
Jeong, S.; Esser, L.; Harran, P. G. Angew. Chem., Int. Ed. 2001, 40, 4765.
JA034332Q
9
J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5421