The vitamin C route to the ciguatoxins: Enantioselective synthesis of a ring F building block

Article abstract of DOI:10.1039/a910163m

A ten-step enantioselective synthesis of an F ring lactone for CTX antibody assay development, employing vitamin C as a starting material, is described.

Full text of DOI:10.1039/a910163m

The vitamin C route to the ciguatoxins: enantioselective synthesis of a ring F
building block
Silas Bond and Patrick Perlmutter*
Department of Chemistry, Monash University, Clayton, Victoria 3168, Australia.
E-mail: Patrick.Perlmutter@sci.monash.edu.au
Received (in Cambridge, UK) 21st December 1999, Accepted 2nd February 2000
Published on the Web 17th March 2000
A ten-step enantioselective synthesis of an F ring lactone for
CTX antibody assay development, employing vitamin C as a
starting material, is described.
to 3, has recently been reported by Hirama’s group to be a very
useful intermediate in CTX synthesis.11
The synthesis of 4 is shown in the Schemes and begins with
the cyclohexylidene acetal 5 of vitamin C (Scheme 1). (The
corresponding acetonide12 was employed initially, however it
proved to be insufficiently robust in subsequent transforma-
tions). Selective O-allylation of the C4 hydroxyl of 5 in
preference of the hydroxyl at C3, using potassium carbonate/
THF–DMSO/allyl bromide, has been reported for vitamin C
acetonide.12 Unfortunately under these conditions we were not
able to obtain 6 sufficiently pure for further reactions. Selective
C4 O-allylation of 5 was ultimately achieved by treatment with
allyl bromide in aqueous THF maintaining neutral to slightly
alkaline pH. The corresponding C-allyl derivative 7 was
obtained as a 4+1 mixture of diastereomers by heating a toluene
solution of 6 at reflux for 6 h. Next it was necessary to generate
a diene from 7 which could be closed under ring closing
metathesis conditions. Addition of a variety of allyl nucleo-
philes to 7 gave mixtures of diastereomers which proved very
difficult to purify. After considerable experimentation it was
found that addition of allylzinc bromide to the O-trimethylsilyl
derivative 8 (obtainable as a single diastereomer, after purifica-
tion, by silylation of crude 7) gave pure diene 9 in excellent
yield and complete diastereoselectivity. The preference for
attack from the Si-face at C4 is probably due to a combination
of steric hindrance from the C3 allyl group and coordination to
the oxygens at C3 and/or those contained in the side chain
acetal. Although the stereochemistry at C3 and C4 is lost later it
Ciguatera, a disease caused by consumption of ciguateric fish
(i.e. those fish which have accumulated ciguatoxins (CTXs) in
their flesh as a result of ingesting Gambierdiscus toxicus, a
benthic dinoflagellate which synthesizes these toxins) is
endemic throughout the tropics.1 Consequently there is a need
to develop a simple assay for the presence of these toxins in fish.
Because of the very low concentrations ( < 0.1 ppb) of
ciguatoxins in fish such an assay will need to be highly
sensitive, such as an immunoassay.2 Hence as part of a
program3 directed towards the synthesis of immunogenic
domains of the CTXs, we report here, the enantioselective
synthesis of a ring F building block.4
From a synthetic point of view, the ciguatoxins constitute one
of the most synthetically challenging classes of naturally
occurring marine toxins.5 Each of these toxins consists of
thirteen contiguous, fused cyclic ethers ranging in size from five
to nine members. The structures of two potential targets, the
Pacific ciguatoxins 2,3-dihydroxy-P-CTX-3C 1a6 and P-CTX-
3C 1b7 are shown in Fig. 1.
Fig.
1 The structures of two Pacific ciguatoxins: P-CTX-3C and
2,3-dihydroxy-P-CTX-3C.
We envisaged the ring F target to be of general structure 4
(Fig. 2). This nine-membered lactone has the correct relative
and absolute stereochemistry in place as well as a reactive
lactone carbonyl. Related oxoninones have been shown to be
valuable intermediates in natural product synthesis.8–10 In
principle 4 should be accessible from vitamin C 2 via bicyclic
lactone 3 (Fig. 2). Indeed an intermediate bicyclic ether, similar
Scheme 1 Reagents and conditions: i, K2CO3, allyl bromide, DMSO–THF,
r.t., 6 h or allyl bromide, 1 M NaOH–THF pH 7–8, r.t., 24 h; ii, toluene,
reflux, 6 h; iii, TMSCl, imidazole, THF, r.t., 2 h, 20% as a single isomer
from 5; iv, allylzinc bromide, THF, r.t., 30 min, 94%; v, RuCHPh(P-
Cy3)2Cl2, toluene, 60–70 °C, 24 h, 69%.
Fig. 2 The structures of vitamin C 2 and two key synthetic targets.
DOI: 10.1039/a910163m
Chem. Commun., 2000, 567–568
This journal is © The Royal Society of Chemistry 2000
567

Treatment of 13 with L-selectride {LiB[CHMe(Et)]3H} at
278 °C gave 15 as a 6+1 mixture of diastereomers. Lowering
the reaction temperature to ca. 2130 °C improved the
selectivity to 14+1 in favour of the desired isomer. Reduction of
11 with 2 equivalents of L-selectride at ca. 2130 °C resulted in
a mixture of products consisting of 60% 16 and a 1+1+1 mixture
of the other three possible diastereomers. The relative ster-
eochemistry of 15 was established by examination of the
coupling between H8 and H9 (J8,9 8.7 Hz). For closely related
nine-membered cyclic ethers, typical coupling constants are
J
trans 8.5 Hz, Jcis 2.2 Hz.11 The reasons for the remarkably high
diastereoselectivity in the first reduction remain unclear as
initial molecular mechanics modelling failed to reveal a low-
energy conformation likely to lead to the observed ster-
eochemistry at C3. However, the source of the selectivity in the
second reduction appears to be a folded conformation for 13
similar to that shown in Fig. 3 for crystalline 14. If this
conformation is maintained in solution then attack at the Si face
of the ketone is clearly favoured, generating the observed
sterochemistry at C8.
In conclusion, we have demonstrated the usefulness of
vitamin C as an enantiomerically pure starting material for the
synthesis of oxonins. Key steps included (i) a ring closing
metathesis/oxidative cleavage sequence to form the nine-
membered ring and (ii) a sequence of two highly diaster-
eoselective reductions. Compounds such as oxonin 15 represent
valuable intermediates for the synthesis of F-ring containing
CTX domains as well as other naturally occurring oxonins such
as obtusenyne and other nine-membered cyclic ethers of marine
origin.16
We are grateful to Dr Gary D. Fallon for determining the X-
ray structures of 10 and 15. P. P. is grateful to the Australian
Research Council for financial support. S. B. is grateful to the
Australian Government for an Australian Postgraduate Award
scholarship.
Scheme 2 Reagents and conditions: i, HgO, I2, benzene, hn, reflux, 8 h; ii,
NaCNBH3, AcOH, ca. 13 °C, 20 min, 43% from 10; iii, TMSCl, imidazole,
THF, r.t., 12 h, 99%; iv, (PhCO)2O, Et3N, DMAP, THF, r.t., 2 h, 99%; v,
1 M LiB[CHMe(Et)]3H, THF, ca. 2130 °C, 30 min, 73%.
proved operationally more practicable to handle a single
stereoisomer through the next sequence of reactions. Thus ring
closing metathesis13 of 9 gave bicyclic lactone 10 in excellent
yield. This crystalline product was subjected to single crystal X-
ray analysis and the relative stereochemistry of the new
stereogenic centres at C3 and C8 was confirmed to be as
shown.
Oxidative ring opening of 10, following adaptations of the
conditions of O’Dell et al.14 and Ito et al.15 yielded oxonintrione
11 in good yield (Scheme 2). We assumed that an intermediate
such as 10a would be generated under these conditions. Such
intermediates have been proposed before for the oxidative
cleavage of silyl ethers.15 With the nine-membered lactone in
hand it remained to reduce the two ketone carbonyls. We found
that this was best achieved in a stepwise manner. Thus treatment
with sodium cyanoborohydride regio- and stereo-selectively
reduced 11 to alcohol 12. The (R)-stereochemistry of the new
centre was confirmed by single crystal X-ray analysis of the
corresponding benzoate (14, R = benzoyl in Scheme 2 and Fig.
3).
Notes and references
1 R. J. Lewis and M. J. Holmes, Comp. Biochem. Physiol. C, Comp.
Pharmacol., 1993, 106, 615.
2 R. J. Lewis, Mem. Queensland Museum, 1993, 34, 541.
3 L. Eriksson, S. T. Guy, P. Perlmutter and R. J. Lewis, J. Org. Chem.,
1999, 64, 8396.
4 M. Inoue, M. Sasaki and K. Tachibana, Tetrahedron, 1999, 55,
10 949.
5 E. Alvarez, M.-L. Candenas, R. Perez, J.-L. Ravelo and J. D. Martin,
Chem. Rev., 1995, 95, 1953.
6 M. Satake, M. Fukui, A.-M. Legrand, P. Cruchet and T. Yasumoto,
Tetrahedron Lett., 1998, 39, 1197.
7 M. Satake, M. Murata and T. Yasumoto, Tetrahedron Lett., 1993, 34,
1975.
8 J. G. Bendall, A. N. Payne, T. E. O. Screen and A. B. Holmes, Chem.
Commun., 1997, 1067.
9 K. Fujiwara, M. Tsunashima, D. Awakura and A. Murai, Tetrahedron
Lett., 1995, 36, 8263.
10 K. Fujiwara, D. Awakura, M. Tsunashima, A. Nakamura, T. Honma and
A. Murai, J. Org. Chem., 1999, 64, 2616.
11 T. Oishi, M. Maruyama, M. Shoji, K. Maeda, N. Kumahara, S. Tanaka
and M. Hirama, Tetrahedron, 1999, 55, 7471.
12 K. Wimalasena and M. P. D. Mahindarate, J. Org. Chem., 1994, 59,
3427.
Fig. 3 Single crystals of C23H26O7 14 were recrystallised from diethyl
ether–hexanes, mounted in inert oil and transferred to the cold gas stream of
the diffractometer. Crystal data: C23H26O7, M = 414.44, monoclinic, space
group P21 (no. 4), a = 10.1336(2), b = 9.5207(2), c = 11.1957(2) Å, b =
104.750(1)°, U = 1044.55(4) Å3, T = 123 K, Z = 2, m(Mo-K) = 0.097
mm21, 14784 reflections measured, 4923 unique (Rint = 0.023) which were
used in all calculations. The final wR(F2) was 0.074 (all data), and flack
parameter c = 0.1(6). CCDC 182/1556. See http://www.rsc.org/suppdata/
cc/a9/a910163m/ for crystallograpic files in .cif format.
13 M. Schuster and S. Blechert, Angew. Chem., Int. Ed. Engl., 1997, 36,
2036.
14 D. E. O’Dell, J. T. Loper and T. L. Macdonald, J. Org. Chem., 1988, 53,
5225.
15 Y. Ito, S. Fujii and T. Saegusa, J. Org. Chem., 1976, 41, 2076.
16 D. J. Faulkner, Nat. Prod. Rep., 1998, 15, 113 and earlier reports in this
series.
Communication a910163m
568
Chem. Commun., 2000, 567–568