Asymmetric total synthesis of 11-deoxytetrodotoxin,a naturally occurring congener

Article abstract of DOI:10.1021/ja0265153

Tetrodotoxin, a toxic principle of puffer fish poisoning, is a specific blocker of sodium channel. Despite many synthetic efforts since the structure elucidation in 1964, the only total synthesis of the racemic tetrodotoxin has been reported by Kishi and

Full text of DOI:10.1021/ja0265153

Published on Web 06/07/2002
Asymmetric Total Synthesis of 11-Deoxytetrodotoxin,
a Naturally Occurring Congener
Toshio Nishikawa, Masanori Asai, and Minoru Isobe*
Contribution from the Graduate School of Bioagricultural Sciences, Nagoya UniVersity,
Chikusa, Nagoya 464-8601, Japan
Received April 12, 2002
Abstract: Tetrodotoxin, a toxic principle of puffer fish poisoning, is a specific blocker of sodium channel.
Despite many synthetic efforts since the structure elucidation in 1964, the only total synthesis of the racemic
tetrodotoxin has been reported by Kishi and co-workers. In the course of our studies directed toward the
total synthesis to analyze biologically interesting issues associated with tetrodotoxin, we accomplished a
highly stereocontrolled synthesis of (-)-5,11-dideoxytetrodotoxin in 1999. Based on the synthesis, we
describe herein the first total synthesis of 11-deoxytetrodotoxin, a naturally occurring analogue. The synthesis
started from an allylic alcohol, the same intermediate for the synthesis of 5,11-dideoxytetrodotoxin.
Epoxidation of the allylic alcohol was followed by isomerization with Ti(i-PrO)4 to give an R-hydroxy allylic
alcohol, in which the configurations of the two hydroxyl groups were inverted by oxidation and then a 2-step
reduction. Further epoxidation of the allylic alcohol and ozonolysis of the remaining vinyl group gave an
aldehyde, which reacted with magnesium acetylide to give a propargyl alcohol in a stereoselective manner.
Oxidative cleavage of the acetylenic moiety with RuO4 afforded a fully functionalized lactone for
11-deoxytetrodotoxin. Crucial guanidinylation was achieved from trichloroacetamide according to our own
method to give acetyldibenzylguanidine. Finally, deprotection of benzyl groups, acetates, and acetal furnished
11-deoxytetrodotoxin.
Introduction
Tetrodotoxin 1 (Figure 1), originally isolated from puffer fish,
is one of the best-known marine natural products because of
its novel chemical architecture and its potent biological activity,
which can lead to fatal poisoning.1 The structure of this
compound was revealed by three groups, in 1964, including
Hirata-Goto, Tsuda, and Woodward, to be an unprecedented
dioxa-adamantane skeleton functionalized by hydroxy groups,
an ortho ester showing acidity (pKa ) 8.7), and cyclic guanidine
with hemiaminal.2 The toxicity is attributed to a specific
blockage of voltage-dependent sodium channels, a finding that
has led to its use as an indispensable biochemical tool in the
field of neurophysiology.3 However, the detailed bound structure
of tetrodotoxin to sodium-channel protein has not been solved
despite many efforts.4 Tetrodotoxin and its deoxy analogues such
as 2, 3, and 45 have been isolated not only from puffer fish, but
also from many other animals such as newt, frog, octopus, and
crab. These analogues are regarded as biosynthetic precursors
or metabolites of tetrodotoxin, although the biosynthetic pathway
has not been elucidated.6 Cultured puffer fish were not toxic,7
Figure 1. Structures of tetrodotoxin and its analogues.
tetrodotoxin-producing bacteria were isolated,8 and tetrodotoxin
was detected in many marine animals with a diet of puffer fish.9
From these facts, it has been accepted that tetrodotoxin is
(4) For example, (a) Nakayama, H.; Hatanaka, Y.; Takai, M.; Yoshida, E.;
Kanaoka, Y. Ann. N.Y. Acad. Sci. 1993, 707, 349. (b) Noda, M.; Suzuki,
H.; Numa. S.; Stuehmer, W. FEBS Lett. 1990, 259, 213. (c) Terlau, H.;
Heinemann, S. H.; Stuehner, W.; Pusch, M.; Conti, F.; Imoto, K.; Numa,
S. FEBS Lett. 1991, 293, 93. (d) Lipkind, G. M.; Fozzard, H. A. Biophys.
J. 1994, 66, 1. (e) Recently, the gross 3D structure at 19 Å was determined.
See: Sato, C.; Ueno, Y.; Asai, K.; Takahashi, K.; Sato, M.; Engel, A.;
Fujiyoshi, Y. Nature 2001, 409, 1047.
* Corresponding author. E-mail: isobem@agr.nagoya-u.ac.jp.
(1) Review: (a) Kao, C. Y., Lovinson, S., Eds. Tetrodotoxin, Saxitoxin, and
the Molecular Biology of the Sodium Channel. Ann. N.Y. Acad. Sci. 1986,
479. (b) Kobayashi, J.; Ishibashi, M. In ComprehensiVe Natural Products
Chemistry; Pergamon: New York, 1999; Vol. 8, p 480.
(5) For a review, see: Yotsu-Yamashita, M. J. Toxicol. Toxin ReV. 2001, 20,
51.
(6) Shimizu, Y.; Kobayashi, M. Chem. Pharm. Bull. 1983, 31, 3625.
(7) Shimizu, U.; Matsui, H.; Sato, H.; Yamamori, K. Mar. Sci. Mon. 1984,
16, 560 (in Japanese).
(2) (a) Goto, T.; Kishi, Y.; Takahashi, S.; Hirata, Y. Tetrahedron 1965, 21,
2059. (b) Tsuda, K.; Ikuma, S.; Kawamura, M.; Tachikawa, R.; Sakai, K.;
Tamura, C.; Amakasu, O. Chem. Pharm. Bull. 1964, 12, 1357. (c)
Woodward, R. B. Pure. Appl. Chem. 1964, 9, 49.
(8) (a) Yasumoto, T.; Yamamura, D.; Yotsu, M.; Michishita, T.; Endo, A.;
Kotaki, Y. Agric. Biol. Chem. 1986, 50, 793. (b) Noguchi, T.; Jeon, J.-K.;
Arakawa, O.; Sugita, H.; Deguchi, Y.; Shida, Y.; Hashimoto, K. J. Biochem.
1986, 99, 311. (c) Shimizu, Y. Ann. N.Y. Acad. Sci. 1986, 479, 24.
(9) Yasumoto, T.; Nagai, H.; Yasumura, D.; Michishita, T.; Endo, A.; Yotsu,
M.; Kotaki, Y. Ann. N.Y. Acad. Sci. 1986, 479, 44.
(3) (a) Narahashi, T. Physiol. ReV. 1974, 54, 813. (b) Hucho, F. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 39.
9
10.1021/ja0265153 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 7847-7852
7847

A R T I C L E S
Nishikawa et al.
Scheme 1. Synthetic Plan for 11-Deoxytetrodotoxin
Results and Discussion
accumulated through the food chain. Recently tetrodotoxin-
binding proteins were isolated from puffer fish,10 and these
proteins might play some role in the specific accumulation and
the resistance against tetrodotoxin.11 Actual biological functions
of the toxin also have been of significant interest.12
Synthetic Plan.19 Our retrosynthetic plan for 11-deoxytet-
rodotoxin (2) is shown in Scheme 1, based on the synthesis of
(-)-5,11-dideoxytetrodotoxin (5). 11-Deoxytetrodotoxin (2) and
4,9-anhydro-11-deoxytetrodotoxin (6) are interconvertible under
acidic conditions.2 The ortho ester was synthetically equivalent
to δ-hydroxy lactone. As the 4,9-anhydro structure played a
crucial role in the protection of the unstable C-9 hydroxyl group
under gunanidinylation15 in the synthesis of 5, we planned to
use a lactone intermediate bearing the 4,9-anhydro structure for
the synthesis of 2. When an aldehyde was generated at the C-4
position, facile â-elimination of the hydroxyl group at C-5 was
anticipated, because of the diaxial relationship between the
proton at C-4a and the hydroxy group at C-5.20 Consequently,
the lactone 7 at C-5 was chosen from the two possible lactone
candidates at the C-5 and C-7 positions, because the lactone
ring formation in 7 would prevent the â-elimination.21 The
guanidine group could be prepared from the trichloroacetamide
8 according to the guanidine synthesis developed in this
laboratory.22 The anhydro moiety of 8 would be prepared from
the lactone-acetonide 9 in an analogous way described in the
synthesis of 5,11-dideoxytetrodotoxin (5). The lactone 9 was
then retrosynthesized into vinylepoxide 10, which was envisaged
to synthesize from the same intermediate 11 as that of 5. Allylic
alcohol 11 was readily prepared from 12, the common synthetic
intermediate for various tetrodotoxin analogues in our studies.23
To elucidate the above problems on a molecular level, a
supply of suitably labeled tetrodotoxin derivatives has been
highly desired by means of total synthesis, because derivatization
of natural tetrodotoxin is extremely difficult.13 Despite many
synthetic efforts,14 however, the only total synthesis of (()-
tetrodotoxin 1 was achieved by Kishi and co-workers in 1972.15
Our continuous efforts directed toward the total synthesis
culminated in achieving a stereocontrolled synthesis of a chiral
(-)-5,11-dideoxytetrodotoxin (5) in 199916 as the first asym-
metric synthesis among tetrodotoxin analogues. However, the
synthetic 5 showed little biological activity (toxicity toward mice
and affinity to Na channel proteins),17 which prompted us to
synthesize biologically active tetrodotoxin analogues having the
characteristic ortho ester group. Herein, we describe the first
total synthesis of 11-deoxytetrodotoxin (2), a naturally occurring
as well as biologically active tetrodotoxin analogue isolated from
puffer fish and newts by Yasumoto and co-workers.18
(10) (a) Matsui, T.; Yamamori, K.; Furukawa, K.; Kono, M. Toxicon 2000, 38,
463. (b) Yotsu-Yamashita, M.; Sugimoto, A.; Terakawa, T.; Shoji, Y.;
Miyazawa, T.; Yasumoto, T. Eur. J. Biochem. 2001, 268, 5937.
(11) (a) Shahjahan, M.; Yamada, M.; Nagaya, M.; Kawai, M.; Nakazawa, A.
Ann. N.Y. Acad. Sci. 1993, 707, 346. (b) Yotsu-Yamashita, M.; Nishimori,
K.; Nitanai, Y.; Isemura, M.; Sugimoto A.; Yasumoto, T. Biochem. Biophys.
Res. Commun. 2000, 267, 403.
(12) (a) Saito, T.; Noguchi, T.; Harada, T.; Murata, O.; Hashimoto, K. Bull.
Jpn. Soc. Sci. Fish 1985, 51, 1175. (b) Matsumura, K. Nature 1995, 378,
563.
(13) For example, Mosher, H. S. Ann. N.Y. Acad. Sci. 1986, 479, 32.
(14) For leading references from other laboratories, see: (a) Noya, B.; Paredes,
M. D.; Ozores, L.; Alonso, R. J. Org. Chem. 2000, 65, 5960. (b) Burgey,
C. S.; Vollerthun, R.; Fraser-Reid, B. J. Org. Chem. 1996, 61, 1609. (c)
Sato, K.; Kajihara, Y.; Nakamura, Y.; Yoshimura, J. Chem. Lett. 1991,
1559. (d) Nachman, R. J.; Ho¨nel, M.; Williams, T. M.; Halaska, R. C.;
Mosher, H. S. J. Org. Chem. 1986, 51, 4802. (e) Keana, J. F. W.; Bland,
J. S.; Boyle, P. J.; Erion, M.; Hartling, R.; Husman, J. R.; Roman, R. B. J.
Org. Chem. 1983, 48, 3621, 3627. (f) Speslacis, J. Ph.D. Thesis, Harvard
University, 1975.
Hydroxylation of the Cyclohexane Ring. The synthesis
commenced with epoxidation of allylic alcohol 11 with m-
chloroperbenzoic acid (MCPBA) to give exclusively â-epoxide
13 (Scheme 2). This selectivity is due to steric hindrance of
the axially oriented vinyl group in the cyclohexane ring as well
(18) Yasumoto, T.; Yotsu, M.; Murata, M.; Naoki, M. J. Am. Chem. Soc. 1988,
110, 2344.
(19) The numbering used in this paper corresponds to that of tetrodotoxin.
(20) Kishi and co-workers encountered the serious problem of â-elimination.
See ref 15.
(21) A lactone at the C-5 hydroxy group was synthesized by Kishi and co-
workers to prevent the â-elimination in their synthesis of (()-tetrodotoxin.
See: (a) Kishi, Y. J. Synth. Org. Chem. Jpn. 1974, 32, 855 (in Japanese).
(b) Aratani, M.; Fukuyama, T.; Tanino, H.; Kishi, Y.; Mataura, T.; Kakoi,
H.; Inoue, S. Annual Meeting of Japan Chemical Society, 1974; abstract
paper III, p 1427.
(22) (a) Yamamoto, N.; Isobe, M. Chem. Lett. 1994, 2299. (b) Nishikawa, T.;
Ohyabu, N.; Yamamoto, N.; Isobe, M. Tetrahedron 1999, 55, 4325.
(23) Nishikawa, T.; Asai, M.; Ohyabu, N.; Yamamoto, N.; Fukuda, Y.; Isobe,
M. Tetrahedron 2001, 57, 3875.
(15) (a) Kishi, Y.; Aratani, M.; Fukuyama, T.; Nakatsubo, F.; Goto, T.; Inoue,
S.; Tanino, H.; Sugiura, S.; Kakoi, H. J. Am. Chem. Soc. 1972, 94, 9217.
(b) Kishi, Y.; Fukuyama, T.; Aratani, M.; Nakatsubo, F.; Goto, T.; Inoue,
S.; Tanino, H.; Sugiura, S.; Kakoi, H. J. Am. Chem. Soc. 1972, 94, 9219.
(16) (a) Nishikawa, T.; Asai, M.; Ohyabu, N.; Yamamoto, N.; Isobe, M. Angew.
Chem., Int. Ed. 1999, 38, 3081. (b) Asai, M.; Nishikawa, T.; Ohyabu, N.;
Yamamoto, N.; Isobe, M. Tetrahedron 2001, 57, 4543.
(17) Yotsu-Yamashita, M.; Nishikawa, T.; Asai, M.; Isobe, M. Unpublished
results.
9
7848 J. AM. CHEM. SOC. VOL. 124, NO. 26, 2002

Asymmetric Total Synthesis of 11-Deoxytetrodotoxin
A R T I C L E S
Scheme 2 a
Figure 2. Conformation of 16 and 17.
on the cyclohexane ring.16,22,29 However, stereoselective reduc-
tion of the diketone 15 proved to be quite difficult. Reduction
with a variety of hydrides gave a mixture of all four possible
diol products with poor selectivities. Fortunately, much experi-
mentation led us to find that the two-step reduction involving
LiAlH(t-BuO)3 in the presence of LiBr30 followed by exposure
31
to NaBH4-CeCl3 afforded diol 17 with the desired configu-
ration in 80% isolated yield from 14. The first reduction gave
a mixture of R-hydroxyketone 16 and the desired diol 17 in a
ratio of 3:1.32 Since the subsequent Luche reduction of the crude
products30 gave 17 in good isolated yield, the reduction of 16
should proceed in high diastereoselectivity. In fact, 16 prepared
from oxidation of diol 17 with IBX (1.5 equiv) was treated under
Luche conditions at -78 °C to afford largely diol 17. Interest-
ingly, an excess of the first reducing agent LiAlH(t-BuO)3 did
not complete the reaction, while reduction of diketone 15 under
Luche conditions gave a mixture of diols in a significantly
decreased stereoselectivity. The configurations of the newly
generated asymmetric centers of 16 and 17 were assigned on
the basis of NOESY spectra as shown in Figure 2. Fortunately,
only the desired diol 17 was easily separated by silica gel
column chromatography, while a mixture of the undesired other
diols could be transformed to diketone 15 through re-oxidation
with IBX as described above.
Further functionalization of diol 17 also proved to be
problematic. We initially planned to protect two hydroxyl groups
at the C-7 and -8 position as benzyl ether, because the benzyl
groups could be concomitantly removed in the deprotection of
dibenzylguanidine at the last stage of the synthesis. However,
benzylation [NaH, benzyl bromide in DMF-THF] of the two
hydroxyl groups gave dibenzyl ether 18 in a very poor yield
(23%) along with monobenzyl ethers as byproducts.33 Further-
more, attempted epoxidation of the resulting dibenzyl ether 18
with MCPBA did not proceed even under forcing conditions at
dichloroethane reflux temperature in the presence of Kishi’s
radical inhibitor34 and with dimethyldioxirane35 in acetone.36
We then examined the trimethylsilyl (TMS) protecting group
a Conditions: (a) MCPBA, Na2HPO4, CH2Cl2. (b) Ti(Oi-Pr)4, (CH2Cl)2,
reflux. (c) IBX (4 equiv), DMSO, 60 °C. (d) LiAlH(Ot-Bu)3, LiBr, THF,
-78 °C. (e) NaBH4, CeCl3(H2O)7, MeOH, -78 °C. (f) IBX (1.5 equiv),
DMSO, room temperature.
as the directed effect of the â-hydroxyl group at C-8. Lewis
acid-mediated isomerization of epoxide 13 to 14 was then
examined, since it was likely that a conventional strong base
such as lithium amide was incompatible with the base-labile
trichloroacetamide group.24 Al(i-PrO)325 was used to isomerize
13 to a 4:1 mixture of the desired endo-allylic alcohol 14 and
the undesired exo-allylic alcohol. In contrast, 13 was treated
with Ti(i-PrO)426 in dichloroethane at the reflux temperature to
afford 14 as a single product in about 90% yield. The high
regioselectivity was attributed to coordination of the titanium
ion to both the epoxide oxygen and hydroxy group followed
by epoxide ring cleavage, as suggested by Sharpless.26 Both
configurations of the resulting diol 14 would be inverted at the
same time by the oxidation-reduction sequence, utilizing the
preferential approach of reagents from the less hindered â-face
of these intermediates. Oxidation of the diol 14 was best
accomplished with IBX (o-iodoxybenzoic acid)27 to give dike-
tone 15, which was successively reduced in the next reduction
without purification.28 We expected that stereoselective reduc-
tion of diketone 15 would be easily realized, because steric
hindrance of the axially oriented vinyl group had been employed
as an important control element in many stereoselective reactions
(28) No enolization of 15 was observed by 1H NMR spectra of the crude product.
Since enolization of the diketone 15 was our major concern, we initially
attempted the reduction of epoxydiketone that was prepared from epoxi-
dation of 14 with MCPBA followed by oxidation of 1,2-diol (Ac2O-DMSO).
However, little selectivity was observed in the reduction under a variety
of conditions.
(29) Yamamoto, N.; Nishikawa, T.; Isobe, M. Synlett 1995, 505.
(30) Mori, Y.; Suzuki, M. Tetrahedron Lett. 1989, 30, 4383.
(31) Luche, J. L.; Gamal, A. L. J. Am. Chem. Soc. 1979, 101, 5848.
(32) A small amount (less than 5%) of the undesired diols was also observed
by 1H NMR; however, the stereochemistries have not been determined.
(33) Other benzylation conditions with benzyl trichloroacetimidate in the
presence of acid failed. Wessel, H.-P.; Iversen, T.; Bundle, D. R. J. Chem.
Soc., Perkin Trans. 1 1989, 54, 465.
(24) Trichloroacetamide produced isocyanate under basic conditions, see:
Atanassova, I. A.; Petrov, J. S.; Mollov, N. M. Synthesis 1987, 734.
(25) (a) Eschinasi, E. H. J. Org. Chem. 1970, 35, 1598. (b) Terao, S.; Kato, K.
Synthesis 1987, 467.
(26) Morgans, D. J., Jr.; Sharpless, K. B.; Traynpor, S. G. J. Am. Chem. Soc.
1981, 103, 462.
(27) Frigerio, N.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
(34) Kishi, Y.; Aratani, M.; Tanino, H.; Fukuyama, T.; Goto, T.; Inoue, S.;
Sugiura, S.; Kakoi, H. J. Chem. Soc., Chem. Commun. 1972, 64.
(35) (a) Murray, R. W. Chem. ReV. 1989, 89, 1187. (b) Adam, W.; Edwards, J.
O. Acc. Chem. Res. 1989, 22, 205.
9
J. AM. CHEM. SOC. VOL. 124, NO. 26, 2002 7849

A R T I C L E S
Nishikawa et al.
Scheme 3 a
Scheme 4 a
a Conditions: (a) TMSOTf, Py, MeCN. (b) MCPBA, Na2HPO4, CH2Cl2,
room temperature.
Figure 3.
that also could be removed at the last stage of the synthesis.
Although the conventional conditions (TMSCl, imidazole, DMF
or TMSOTf, 2,6-lutidine, CH2Cl2) for silylation were not
applicable to 17, we fortunately found that the conditions
(TMSOTf, pyridine, CH3CN) described by Heathcock37 gave
the desired bis-TMS protected diol 19 in high yield. In sharp
contrast to dibenzyl ether 18, 19 was smoothly epoxidized with
MCPBA at room temperature to give 22 as a single product in
high yield.38,39
a Conditions: (a) O3, MeOH, -78 °C; Me2S. (b) TMS-CtC-MgBr,
THF, 0 °C to room temperature. (c) Ac2O, Py. (d) n-Bu4NF, MeOH, THF.
(e) HF-Py, THF. (f) Ac2O, Py. (g) RuO2(H2O)x, NaIO4, CCl4, MeCN, H2O.
the configuration could not be determined at this stage. The
propargyl alcohol 24a was transformed to propargyl acetate 25
in two steps involving acetylation followed by selective desi-
lylation of silylacetylene with n-Bu4NF in the presence of MeOH
in THF.42 Surprisingly, the acetylenic moiety of 25 was com-
pletely inert to RuO4 oxidation,43 probably due to steric hin-
drance of the two TMS ethers and trichloroacetamide. Thus, the
two TMS ethers of 25 were converted to acetate’s smaller pro-
tective group in two steps: (i) HF/Py, and (ii) Ac2O/Py.44 In
this case, the acetylenic group of 26 was successfully cleaved
with RuO4,43 and the resulting carboxylic acid intermediate
spontaneously opened the epoxide under these conditions to
furnish lactone 27 in good overall yield.45 The structure of 27
was confirmed by extensive 1H NMR analysis (Figure 3). The
configuration of the C-9 position that was generated by the ad-
dition of acetylide to the aldehyde 23 was confirmed by ob-
serving the W-shaped long-range coupling (J ) 1 Hz) between
H-4a and H-9. Long-range coupling (J ) 1.5 Hz) observed
between H-5 and H-7 and NOESY correlation between H-4a
and H-8 confirmed the configurations of the two hydroxyl
groups as shown in Figure 3. Thus lactone 27 constitutes a fully
Synthesis of Lactone Intermediate. With multigram quanti-
ties of 22 in hand, the vinyl group was transformed into
R-hydroxycarboxylic acid according to the methods for (-)-
5,11-dideoxytetrodotoxin. The aldehyde 23 obtained by ozo-
nolysis of 22 was treated with magnesium acetylide in THF to
give a 5.5:1 separable diastereomeric mixture in favor of the
desired isomer 24a (Scheme 4),40 while the reaction with lithium
acetylide in THF gave a lower selectivity (1.7:1).41 On the basis
of the proposed mechanism of stereoselective addition of the
acetylide to the aldehyde in the synthesis of 5, we assumed that
the configuration of the C-9 position of the major product 24a
would be the same as that of 11-deoxytetrodotoxin, although
(36) Epoxidation of 17 with MCPBA also did not proceed at room temperature,
while the epoxidation in the presence of Kishi’s radical inhibitor at elevated
temperature (dichloroethane reflux) gave 20 in only 52% yield.
(37) Heathcock, C. H.; Young, S. D.; Hagen, J. P.; Pilli, R.; Badertscher, U. J.
Org. Chem. 1985, 50, 2095.
(38) One plausible explanation for the different reactivity might be the difference
in the conformation of these substrates 18 and 19. However, we could not
find a difference in the coupling constants between protons at C-7 and C-8
of 18 and 19 (both 3.5 Hz).
(39) The stereochemistry of the epoxide could not be determined at this stage,
although the structure was assumed to be â from the steric hindrance of
the allylic siloxy group. The structure was finally determined by NMR
experiments of lactone 27 derived from 22 (vide infra).
(42) Nishikawa, T.; Ino, A.; Isobe, M. Tetrahedron 1994, 50, 1449.
(43) Carisen, P. H.; Katsuki, T.; Martin, V. S.; Sharpless, K. B. J. Org. Chem.
1981, 46, 3936.
(44) Attempted desilylation of all the TMS groups of 24a in one operation with
n-Bu4NF in THF in the absence of MeOH gave not triol but a cyclic
carbamate that was formed by attack of alkoxide at the C-8 of trichloro-
acetamide.
(45) Spontaneous lactonization may be due to the special nature of such a caged
molecule. For other examples from our laboratory, see ref 29.
(40) The slightly better diastereomeric ratio (6:1) was obtained by titanium
triisopropoxy acetylide in THF, but the reaction was sluggish (for 3 days
at 5 °C). For titanium acetylide, see: (a) Tabusa, F.; Yamada, T.; Suzuki,
K.; Mukaiyama T. Chem. Lett. 1984, 405. (b) Krause, N.; Seebach, D.
Chem. Ber. 1987, 120, 1845.
(41) Interestingly, the same reaction in diethyl ether as a solvent gave the reverse
selectivity (24a/24b ) 1:3.5 by 1H NMR).
9
7850 J. AM. CHEM. SOC. VOL. 124, NO. 26, 2002

Asymmetric Total Synthesis of 11-Deoxytetrodotoxin
A R T I C L E S
Scheme 5 a
same conditions gave the corresponding guanidine in poor
yield, probably because of steric hindrance of the methoxy
group at the acetal position. Thus, 28a could be guanidinyl-
ated through the minor isomer 28b obtainable by equilibration
of 28a under the acidic conditions (p-TsOH, CH(OMe)3,
MeOH).
The synthesis was completed by the successive removal of
three types of protecting group as follows. (i) The benzyl groups
of guanidine (31) were hydrogenolyzed in acetic anhydride to
give diacetylguanidine (32); (ii) all the acetates were removed
with aqueous ammonia; and (iii) the acetal was hydrolyzed with
aqueous TFA. The crude products were separated by HPLC on
an ion-exchange resin to give 11-deoxytetrodotoxin (2) and
anhydro-11-deoxytetrodotoxin (6) in 29 and 55% yields,
respectively.18 These products 2 and 6 are interconvertible in
acidic media. For example, a solution of anhydro-11-deoxytet-
rodotoxin (6) dissolved in 1% TFA-d in D2O was kept at room
temperature for 7 days to attain the equilibrium composed of
2, 6, and 4-epi-11-deoxytetrodotoxin in an ca. 5:1:1 ratio (from
a Conditions: (a) KCN, EtOH. (b) HIO4(H2O)2, AcOMe. (c) p-TsOH,
CH(OMe)3, MeOH.
functionalized cyclohexane derivative for the synthesis of 11-
deoxytetrodotoxin.
Installation of Guanidine and Completion of the Total
Synthesis. In the synthesis of 5,11-dideoxytetrodotoxin, instal-
lation of the guanidine functionality was the most difficult task
due to the hydroxy group instability at the C-9 position and the
severe steric hindrance around the nitrogen-containing func-
tionality.16 As such, the hydroxyl group at the C-9 position was
protected as an intramolecular mixed acetal with the aldehyde
at the C-4 position. This protection proved to be critical not
only during the installation of the guanidine, but also for secure
deprotection at the final stages of the synthesis. Thus, the lactone
27 was converted to a 1.5:1 diastereomeric mixture of acetal
28a and 28b46 in three steps involving selective deacetylation
at the C-9 position,47 oxidative cleavage of the acetonide,48 and
acetalization (Scheme 5). After separation with silica gel
chromatography, the minor isomer 28b was acetylated, and then
subjected to the guanidinylation developed in this laboratory.23
Thus, the trichloroacetamide (28b) was transformed to benzyl-
urea (29), which was dehydrated under the conditions reported
by Appel49 to give a benzylcarbodiimide 30. The addition of
benzylamine was best accomplished with benzylamine hydro-
chloride under pyridine reflux conditions16,22 to afford the
dibenzylguanidinium salt, which was subsequently acetylated
to give acetyldibenzylguanidine (31) in good overall yield. On
the otherhand, the major isomer 28a could be transformed to
the corresponding carbodiimide in the same way as shown in
Scheme 6. However, the addition of benzylamine to the
carbodiimide derived from the major isomer 28a under the
1
1H NMR). The H and 13C NMR spectra of both products 2
and 6 were identical with those of the natural products.18
Conclusion
In summary, we have achieved total synthesis of naturally
occurring 11-deoxytetrodotoxin in an enantiomerically pure
form. This work opens up the possibility of preparing suitable
probes to elucidate related biological issues. Further synthetic
studies toward tetrodotoxin and unnatural tetrodotoxin analogues
are currently under investigation in our laboratory.
Acknowledgment. We are grateful to Professors M. Yotsu-
Yamashita (Tohoku University) and T. Yasumoto (Japan Food
Research Laboratories, Tama laboratory) for providing the NMR
data of the natural products. Elemental analyses and measure-
ments of HR-MS and 2D-NMR spectra were performed by Mr.
S. Kitamura and Mr. K. Koga (analytical laboratory in thiss-
chool), whom we thank. This work was financially supported
by JSPS-RFTF, a Grant-in-Aid for Scientific Research from the
Ministry of Education, Science, Sports, and Culture of Japan,
Pfizer’s Award in Synthetic Organic Chemistry, and Mitsubishi
Chemical Corporation Fund.
Scheme 6 a
a Conditions: (a) Ac2O, DMAP, Py. (b) BnNH2, Na2CO3, DMF, reflux. (c) Ph3P, CBr4, Et3N, CH2Cl2. (d) BnNH2‚HCl, Py, reflux. (e) Ac2O, Et3N, Py.
(f) H2, Pd(OH)2-C, Ac2O. (g) NH4OH, MeOH, H2O. (h) TFA, H2O.
9
J. AM. CHEM. SOC. VOL. 124, NO. 26, 2002 7851

A R T I C L E S
Nishikawa et al.
Supporting Information Available: Full experimental details
and spectroscopic data including NMR spectra (PDF). This
material is available free of charge via the Internet at
http://pubs.acs.org.
(46) The configurations of the acetal carbons (C-4) of 28a and 28b were
determined from coupling constants between H-4 and H-4a (0 and 4 Hz,
respectively). The dihedral angle between H-4 and H-4a of 28a is almost
90°. The observed coupling constants are also in good agreement with
calculated values (0.9 and 6.3 Hz) by MacroModel (MM2*).
JA0265153
(47) Mori, K.; Tominaga, M.; Takigawa, T.; Matsui, M. Synthesis 1973, 790.
(48) Xie, M.; Berges, D. A.; Robins, M. J. J. Org. Chem. 1996, 61, 5178.
(49) Appel, R.; Kleinstu¨ck, R.; Ziehn, K. D. Chem. Ber. 1971, 104, 1335.
9
7852 J. AM. CHEM. SOC. VOL. 124, NO. 26, 2002