Stereocontrolled synthesis of (-)-5,11-dideoxytetrodotoxin

Article abstract of DOI:10.1002/(SICI)1521-3773(19991018)38:20<3081::AID-ANIE3081>3.0.CO;2-6

New derivatives of an intriguing marine natural product are now accessible. The first asymmetric synthesis of the simple tetrodotoxin analogue, 5,11-dideoxytetrodotoxin (3), was achieved. Hydroxylation at position C8 of the key intermediate 1 relied on th

Full text of DOI:10.1002/(SICI)1521-3773(19991018)38:20<3081::AID-ANIE3081>3.0.CO;2-6

COMMUNICATIONS
at its total synthesis. In spite of many efforts,[8] no total
synthesis of tetrodotoxin has been reported since the total
synthesis of the racemate by Kishi, Goto, and co-workers in
1972.[9] During the course of our synthetic studies on
tetrodotoxin, we have developed a stereocontrolled synthesis
of polyhydroxylated cyclohexane,[10] the introduction of a
nitrogen functionality by the Overman rearrangement,[11, 12]
and a novel guanidine synthesis.[13] Here we describe the
stereocontrolled synthesis of enantiomerically pure 5,11-
dideoxytetrodotoxin (2), a tetrodotoxin analogue not yet
found in natural sources.
[10] For the nomenclature used to describe hexagonal plates, see Figure 2
and Reference [1].
[11] The thickness of the plates was controlled by cutting the oxidized
hexagonal rod with a razor blade. The thickness is an average value
since the hexagons vary slightly in their thickness (Æ0.1 mm).
1
[12] We periodically changed the rate of agitation to 1.8 s to accelerate
collisions, to break up weak structures, and to stimulate annealing.
[13] We repeated the experiment six times. The yield was calculated by
counting the number of times structure 2 ´ 33 was formed, relative to
the incomplete structures 2 ´ 32 and 2 ´ 31. For example, in the assembly
of 1.4-mm-thick objects of 3 and 0.9-mm-thick 2 we counted 21 arrays
of 2 ´ 33, 3 arrays of 2 ´ 32, and 6 arrays of 2 ´ 31. The yield of 2 ´ 33 was 21/
30 or 70%.
[14] The structure 2 ´ 33 was not stable at w ˆ 1.5 s 1, but was stable at w ˆ
1.2 s 1. We agitated the objects at w ˆ 1.2 s 1 and periodically changed
1
O
the frequency of agitation to 1.8 s
.
HO
[15] We repeated the experiment five times with six objects of 2 and
eighteen of 1 to get statistics for this self-assembly. We agitated the
plates for one hour. The yields of the assembly of 2 ´ 13 were 67%,
100%, 83%, 67%, and 67%. The micrograph shown in Figure 3d is
the most complete assembly (the highest yield) that we observed.
[16] S. A. Jenekhe, X. L. Chen, Science 1999, 283, 372 ± 375.
[17] D. E. Ingber, Sci. Am. 1998, 278(1), 48 ± 57.
O
H2N
H
N
O
OH
N
R
HO
H
OH
R = CH2OH
R = CH3
Tetrodotoxin 1
11-Deoxytetrodotoxin
O
HO
H2N
H
N
O
Stereocontrolled Synthesis of
( )-5,11-Dideoxytetrodotoxin**
5
OH
11
N
1
CH3
R2
HO
6
Toshio Nishikawa, Masanori Asai, Norio Ohyabu,
Noboru Yamamoto, and Minoru Isobe*
R1
R1 = H
R2 = H
5,6,11-Trideoxytetrodotoxin
1-Hydroxy- 5,11-Dideoxytetrodotoxin
5,11-Dideoxytetrodotoxin 2
R1 = OH R2 = OH
R1 = OH R2 = H
Tetrodotoxin (1)[1] is a well-known marine natural product
that is the toxic principle of puffer fish poisoning. Due to its
highly selective inhibition of voltage-dependent sodium
channels, tetrodotoxin has served widely as an important
biochemical tool in neurophysiological studies.[2] Recently,
tetrodotoxin analogues have been isolated from many kinds
of animals.[3] These studies brought to light a number of new
interesting problems associated with tetrodotoxin,[4] for
example its biosynthesis, mechanisms of accumulation, detox-
ification, and binding to the sodium channel protein,[5] and its
actual biological function.[6] In order to examine such new
problems on a molecular level, suitably labeled tetrodotoxins
are desirable. Such labeled compounds, however, are difficult
to prepare from naturally occurring tetrodotoxin.[7] The
complex structure, which has a variety of functional groups
and unique chemical properties, has thwarted recent attempts
Our retrosynthetic plan for 5,11-dideoxytetrodotoxin (2) is
shown in Scheme 1. The guanidine part of 2 is introduced at
the latest stage of the synthesis because of the instability and
low solubility of the intermediates. Lactone intermediate 3
O
O
H
N
HO
NH2
NH
O
O
O
COCCl3
COCCl3
NH
NH
OH
OAc
HO
BnO
8
Me
HO
Me
HO
Me
OH
O
O
O
3
2
4
[*] Prof. M. Isobe, Dr. T. Nishikawa,
M. Asai, N. Ohyabu, N. Yamamoto
Laboratory of Organic Chemistry
School of Bioagricultural Sciences
Nagoya University
O
O
O
O
OH
O
O
COCCl3
NH
+
O
X
Chikusa, Nagoya 464 ± 8601 (Japan)
Fax: (‡81)52-789-4111
Me
Me
Me
7
6
5
E-mail: isobem@agr.nagoya-u.ac.jp
Scheme 1. Retrosynthetic plan for 5,11-dideoxytetrodotoxin (2).
[**] We are grateful to Prof. T. Yasumoto and Dr. M. Yotsu-Yamashita
(Tohoku University) for fruitful discussions. We also acknowledge the
valuable contributions of Y. Fukuda and Dr. S. Pikul to the early
stages of this work. 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, and Iyakushigen
Zaidan (the Fujisawa foundation).
was envisioned to be synthesized from trichloroacetamide
diene 5 via diene alcohol 4 in a stereoselective manner.
Compound 5 has already been synthesized from levoglucose-
none 7 (X ˆ H) as a chiral starting material[14] by means of a
Angew. Chem. Int. Ed. 1999, 38, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3820-3081 $ 17.50+.50/0
3081

COMMUNICATIONS
nine-step sequence that included the Diels ± Alder reaction of
bromolevoglucosenone 7 (X ˆ Br) with isoprene and a highly
stereoselective introduction of the amino group by the
Overman rearrangement of the exo-allylic alcohol 6[12, 15] as
key steps.
Stereoselective a-hydroxylation at the C8 position of the
key intermediate 5 was critical because the hydroxy group has
this configuration in all of the naturally occurring tetrodo-
toxin analogues. Bromination of 5 with pyridinium bromide
perbromide gave dibromide 8, which was treated with DBU in
DMF to give oxazoline 9 in a high yield (Scheme 2;
abbreviations are defined in the legend). This reaction
involves regioselective dehydrobromination of the dibromide
8 and subsequent SN2' reaction of the resulting allylic bromide
i with the neighboring trichloroacetamide group in a stereo-
specific manner. The product 9 was hydrolyzed with pTsOH
in aqueous pyridine[16] to give trichloroacetamide 10 having an
8S hydroxy group without further hydrolysis. The hydroxy
group of 10 was inverted by oxidation with PCC and highly
stereoselective reduction with NaBH4/CeCl3 to afford largely
the desired 8R alcohol 4 (20:1).[17] This high stereoselectivity is
attributed to the steric hindrance of the axially oriented vinyl
group of the cyclohexene ring, whose conformation is fixed by
the bulky equatorial trichloroacetamide group.
O
Prior to the transformation of the terminal olefinic bond
O
DBU
O
ˆ
into a hydroxycarboxylic acid function, the trisubstituted C C
O
COCCl3
NH
H
bond of 4 was epoxidized with MCPBA to give a b-epoxide as
a sole product,[18] whose hydroxy group was protected as
benzyl ether 11 in a high yield. In accord with our model
studies,[19] the vinyl group of 11 was strongly shielded by the
bulky trichloroacetamide group and therefore was extremely
a
b
5
N
O
CCl3
Me
O
Me
O
Br
Br
i
Br
8
ˆ
inert toward a variety of reagents. This C C bond was,
O
however, cleaved by ozonolysis to afford the aldehyde 12,
which was transformed into the hydroxycarboxylic acid by the
addition of a carboxylic acid equivalent. Vinylmagnesium
bromide, which we had employed previously,[19] did not react
with the aldehyde 12. On the other hand, a smaller sp
nucleophile, an alkynylmagnesium bromide, added in a highly
stereoselective manner to give a propargylic alcohol as a
single stereoisomer. The resulting unstable product was
immediately transformed into 13 by successive desilylation
and acetylation in high yield. The acetylenic moiety was
cleaved with ruthenium oxide[20] to give a carboxylic acid,
which spontaneously opened the epoxide to furnish six-
membered lactone 3[21] in moderate yield.
With the key intermediate 3 in our hands, we were ready for
the crucial construction of the guanidine moiety. The trichloro-
acetamide group of 3 was difficult to deprotect to the amine,
because the hydroxy group at the C9 position was easily
epimerized even under mild basic conditions such as K2CO3 in
MeOH. Consequently, we could not synthesize the guanidine
unit in the usual way, from a nonprotected amine with one of a
variety of guanylation reagents. To overcome this difficulty,
we developed a novel guanidine synthesis from the trichloro-
acetamide group proceeding through a dibenzylguanidine.[13]
This new route did not proceed through the corresponding
free amine as an intermediate. First, we applied this route to
the intermediate 3, which was transformed into benzylurea
14 with benzylamine and Na2CO3 in DMF,[13, 22] and the
resulting urea was dehydrated with Ph3P and CBr4 to give
benzylcarbodiimide 15 (Scheme 2). However, extensive ex-
amination under a variety of conditions revealed that the
addition of an amine to carbodiimide 15 to obtain the
corresponding dibenzylguanidine was very difficult because
of the labile C9 acetoxy group.[23] Second, we planned an
alternative route going through a stable tricyclic intermediate
17, which could survive the guanidine synthesis. The synthesis
of 17 is shown in Scheme 3. Oxidative cleavage of the ketal[24]
in 3 was followed by protection of the resulting aldehyde and
tertiary alcohol to give the dimethylacetal 16, which was
O
8
COCCl3
NH
c
d, e
N
CCl3
Me
O
Me
Me
Me
OH
10
9
O
O
O
O
O
COCCl3
NH
COCCl3
NH
f, g
h
Me
OH
OBn
4
11
O
O
O
O
COCCl3
NH
COCCl3
NH
i, j, k
l
CHO
OBn
BnO
Me
OAc
O
O
O
13
12
O
O
R
OAc
R = NHCOCCl3
R = NHCONHBn
R = N=C=NBn
BnO
OH
3
14
15
m
n
9
Me
O
Scheme 2. Synthesis of lactone 15. a) Pyridinium bromide perbromide,
K2CO3, CH2Cl2, 108C, 3 h, 87%; b) DBU, DMF, rt, 6 h; c) pTsOH,
pyridine ± H2O, 708C, 1.5 h, 71% over two steps; d) PCC, MS 4 Š, CH2Cl2,
rt, 11 h; e) NaBH4, CeCl3(H2O)7, EtOH ± CH2Cl2, 08C, 2 h, !rt, 30 min,
75% over two steps; f) MCPBA, CH2Cl2, rt, 24 h; g) NaH, BnBr, THF-
DMF, rt, 22 h, 86% over two steps; h) O3, MeOH, 788C, 70 min, Me2S,
ꢀ
96%; i) Me3SiC CMgBr, THF, rt, 2 h; j) nBu4NF, THF, rt, 2.5 h; k) Ac2O,
pyridine, rt, 5.5 h, 86% over three steps; l) RuO2(H2O)n, NaIO4,
CCl4 CH3CN H2O, rt, 4.5 h, 75%; m) BnNH2, Na2CO3, DMF, 1258C,
30 min; n) Ph3P, CBr4, Et3N, CH2Cl2, rt, 7 h, 87% over two steps. DBU ˆ
1,8-diazabicyclo[5.4.0]undec-7-ene, DMF ˆ dimethylformamide, rt ˆ room
temperature, pTsOH ˆ 4-toluenesulfonic acid, PCC ˆ pyridinium chloro-
chromate, MCPBA ˆ meta-chloroperbenzoic acid.
3082
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3820-3082 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 20

COMMUNICATIONS
The crude product was separated by HPLC with an ion-
exchange resin (Hitachi gel 3013c)[3] to afford 5,11-dideoxy-
tetrodotoxin (2), 4-epi-5,11-dideoxytetrodotoxin (21), and 4,9-
anhydro-epi-5,11-dideoxytetrodotoxin (22)[28] in yields of 29,
16, and 36%, respectively. These structures were confirmed
by full characterization by the NMR techniques COSY,
HMBC, and HSQC, and by FAB mass spectrometry.
We have achieved a highly stereocontrolled synthesis of
( )-5,11-dideoxytetrodotoxin and its isomer. This is the first
asymmetric synthesis of tetrodotoxin analogues and provides
a practical route accessible to the labeled compounds for
biochemical studies. Further studies toward naturally occur-
ring tetrodotoxin (1) and other analogues are in progress in
our laboratory.
MeO OMe
COCCl3
OAc
H
OMe
NH
OAc
Me
H
g, h
a, b, c
d, e, f
N
O
3
BnO
H
OBn
O
O
O
NHBn
Me
OAc O
O
H
16
17
OAc
H
OAc
H
OMe
NBn
OMe
Me
Me
j
i
NBn
N
N
H
H
H
OBn
O
OBn
O
O
O
O
O
N
NHBn
Bn
Ac
H
H
19
18
OH
H
OAc
H
OH
NH
Received: April 15, 1999 [Z13284IE]
German version: Angew. Chem. 1999, 111, 3268 ± 3271
OMe
NAc
Me
Me
k, l
N
N
+
H
H
H
OH
NH2
AcO
OAc
O
O
O
NHAc
Keywords: asymmetric synthesis ´ natural products ´ syn-
thetic methods ´ tetrodotoxins
OH
H
O
O
20
2
[1] For the structure: a) T. Goto, Y. Kishi, S. Takahashi, Y. Hirata,
Tetrahedron 1965, 21, 2059 ± 2088; b) K. Tsuda, S. Ikuma, M.
Kawamura, K. Tachikawa, K. Sakai, C. Tamura, O. Akamatsu, Chem.
Pharm. Bull. 1964, 12, 1357 ± 1374; c) R. B. Woodward, Pure Appl.
Chem. 1964, 9, 49 ± 74.
OH
H
OH
H
4
NH
4
NH
Me
Me
N
N
H
H
+
OH
O
NH2
AcO
H
H
OH
OH
OH
O
NH2
AcO
O
O
O
9
[2] T. Narahashi, Physiol. Rev. 1974, 54, 813 ± 889.
[3] a) 5,6,11-Trideoxytetrodotoxin: M. Yotsu-Yamashita, Y. Yamagishi, T.
Yasumoto, Tetrahedron Lett. 1995, 51, 9329 ± 9332; b) 1-hydroxy-5,11-
dideoxytetrodotoxin: Y. Kotaki, Y. Shimizu, J. Am. Chem. Soc. 1993,
115, 827 ± 830; c) 6-epi- and 11-deoxytetrodotoxin: T. Yasumoto, M.
Yotsu, M. Murata, H. Naoki, J. Am. Chem. Soc. 1988, 110, 2344 ± 2345.
[4] ªTetrodotoxin, Saxitoxin, and the Molecular Biology of the Sodium
Channelº: Ann. N.Y. Acad. Sci. 1986, 479 (review series).
[5] a) H. Nakayama, Y. Hatanaka, M. Takai, E. Yoshida, Y. Kanaoka,
Ann. N.Y. Acad. Sci. 1993, 707, 349; b) G. M. Lipkind, H. A. Fozzard,
Biophys. J. 1994, 66, 1 ± 33.
21
22
Scheme 3. Synthesis of 5,11-dideoxytetrodotoxin (2). a) H5IO6, AcOEt, rt,
24 h; b) CSA, CH(OMe)3, MeOH, rt, 10 h; c) Ac2O, pyridine, DMAP, rt,
24 h, 78% over three steps; d) BnNH2, Na2CO3, DMF, 1408C, 15 min;
e) KCN, EtOH, rt, 1 h; f) CSA, acetone, rt, 2 h, 70% over three steps;
g) Ph3P, CBr4, Et3N, CH2Cl2, rt, 4 h; h) BnNH2 ´ HCl, pyridine, reflux, 5.5 h;
i) Ac2O, pyridine, Et3N, rt, 4.5 h, 85% over three steps; j) H2 (1 atm),
Pd(OH)2/C, Ac2O, rt, 14 d, 81%; k) aq. NH3, MeOH H2O, rt, 20 h; l) TFA,
H2O, rt, 34 h, 81% over two steps. CSA ˆ 10-camphorsulfonic acid,
DMAP ˆ 4-dimethylaminopyridine, TFA ˆ trifluoroacetic acid.
[6] Tetrodotoxin was identified as a pheromonelike substance among the
puffer fish. K. Matsumura, Nature 1995, 378, 563.
[7] For examples of some labeled tetrodotoxin derivatives, see: a) M.
Balerna, A. Lombet, R. Chicheportiche, G. Romey, M. Lazdunski,
Biochim. Biophys. Acta 1981, 644, 219 ± 225; b) J. Bontemps, R.
Cantineau, C. Grandfils, P. Leprince, G. Dandrifosse, E. Schoffeniels,
Anal. Biochem. 1984, 139, 149 ± 157; c) B. Q. Wu, L. Yang, C. Y. Kao,
S. R. Levinson, M. Yotsu-Yamashita, T. Yasumoto, Toxicon 1996, 34,
407 ± 416.
[8] Recent reports: a) B. Noya, R. Alonso, Tetrahedron Lett. 1997, 38,
2745 ± 2748; b) C. S. Burgey, R. Vollerthun, B. Fraser-Reid, J. Org.
Chem. 1996, 61, 1609 ± 1618; c) K. Sato, Y. Kajihara, Y. Nakamura, J.
Yoshimura, Chem. Lett. 1991, 1559 ± 1562; d) J. F. W. Keana, J. S.
Bland, P. J. Boyle, M. Erion, R. Hartling, J. R. Husman, R. B. Roman,
G. Ferguson, M. Parvez, J. Org. Chem. 1983, 48, 3627 ± 3631; J. F. W.
Keana, P. J. Boyle, M. Erion, R. Hartling, J. R. Husman, J. E.
Richman, R. M. Wah, J. Org. Chem. 1983, 48, 3621 ± 3626.
[9] a) Y. Kishi, M. Aratani, T. Fukuyama, F. Nakatsubo, T. Goto, S. Inoue,
H. Tanino, S. Sugiura, H. Kakoi, J. Am. Chem. Soc. 1972, 94, 9217 ±
9219; b) Y. Kishi, T. Fukuyama, M. Aratani, F. Nakatsubo, T. Goto, S.
Inoue, H. Tanino, S. Sugiura, H. Kakoi, J. Am. Chem. Soc. 1972, 94,
9219 ± 9221; c) Y. Kishi, F. Nakatsubo, M. Aratani, T. Goto, S. Inoue,
H. Kakoi, S. Sugiura, Tetrahedron Lett. 1970, 5127 ± 5128; d) Y. Kishi,
F. Nakatsubo, M. Aratani, T. Goto, S. Inoue, H. Kakoi, Tetrahedron
Lett. 1970, 5129 ± 5132.
transformed into the acetal urea under the above-mentioned
conditions. Mild deacetylation with potassium cyanide in
ethanol,[25] and partial hydrolysis of the acetal gave the urea 17
(as a 5:1 diastereomeric mixture at the acetal carbon) as the
precursor to the guanidine installation. The major isomer 17
was separated and dehydrated with Ph3P and CBr4 to afford
the carbodiimide, which was further treated with benzylamine
hydrochloride in pyridine[26] to furnish the dibenzylguanidine
18 in a high yield. Our model studies revealed that the benzyl
group of benzylguanidinium salts was difficult to remove
under hydrogenolytic conditions, while the benzyl group of
acetylbenzylguanidine could be easily removed under the
same conditions.[12b] In the event, the corresponding acetyl-
benzylguanidine 19 was hydrogenolyzed in acetic anhydride
under one atmosphere of hydrogen to give diacetylguanidine
20 in a good yield.[27] Under the reaction conditions, the benzyl
ether in 19 was also converted to an acetyl group. At this
stage, all the protective groups were unified to acetyl groups
except for the acetal.
[10] M. Isobe, T. Nishikawa, S. Pikul, T. Goto, Tetrahedron Lett. 1987, 28,
6485 ± 6488.
[11] Reviews: a) L. E. Overman, Acc. Chem. Res. 1980, 13, 218 ± 224; b) K.
Ritter in Stereoselective Synthesis (Houben-Weyl) Vol. 21/9 E21, Vol. 9
(Eds.: G. Helmchen, R. W. Hoffmann, J. Mulzer, E. Schaumann),
Thieme, Stuttgart, 1996, pp. 5677 ± 5699.
To complete the synthesis of 5,11-dideoxytetrodotoxin (2),
two steps of deprotection were carried out. Thus, hydrolysis of
the acetyl groups with ammonium hydroxide and then acid
hydrolysis of the acetal with aqueous TFA was performed.
Angew. Chem. Int. Ed. 1999, 38, No. 20
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3820-3083 $ 17.50+.50/0
3083

COMMUNICATIONS
[12] M. Isobe, F. Fukuda, T. Nishikawa, P. Chabert, T. Kawai, T. Goto,
Tetrahedron Lett. 1990, 31, 3327 ± 3330.
[13] a) N. Yamamoto, M. Isobe, Chem. Lett. 1994, 2299 ± 2302; b) T.
Nishikawa, N. Ohyabu, N. Yamamoto, M. Isobe, Tetrahedron 1999, 55,
4325 ± 4340.
[14] M. Isobe, N. Yamamoto, T. Nishikawa in Levoglucosenone and
Levoglucosans, Chemistry and Applications (Ed.: Z. J. Witczak), ATL
Press, 1994, pp. 99 ± 118.
[15] T. Nishikawa, M. Asai, N. Ohyabu, M. Isobe, J. Org. Chem. 1998, 63,
188 ± 192.
Regio- and Diastereoselective Aldol Products
through Three-Component Coupling Reactions
of Difluoroboroxy Fischer Carbene
Molybdenum Complexes**
Â
Â
Jose Barluenga,* Felix Rodríguez,
Â
Ä
Francisco J. Fananas, and Eduardo Rubio
The formation of a large number of carbon ± carbon bonds
with high efficiency in the minimal number of chemical steps
is the requirement of modern synthetic design.[1] In this
context, radical and carbanion chemistry has become a
valuable tool in organic synthesis. However, both reactive
intermediates can be used in a complementary manner and
sequential reactions have been devised to prepare complex
molecules.[2, 3] On the other hand, despite the importance of
Fischer carbene complexes for the development of novel
organic transformations[4] there is only a very limited number
of examples in which these complexes have been used as
precursors of radicals. Thus, tetramethylammonium acyl
chromate complexes are adequate precursors for alkyl and
acyl radicals, which are formed by oxidation with manga-
nese(iii)[5] and copper(ii)[6] salts.
In 1997 we reported a simple method for generating acyl
radicals from difluoroboroxy Fischer carbene molybdenum
complexes in very mild reaction conditions and in the absence
of an oxidant.[7] Herein we describe a sequential radical-
anionic three-component coupling reaction of difluoroboroxy
Fischer carbene molybdenum complexes, vinyl ketones, and
aldehydes to give syn-b-hydroxyketones diastereoselectively.
This process is amenable for synthesizing enantiomerically
pure compounds.
[16] H. W. Pauls, B. Fraser-Reid, Carbohydr. Res. 1986, 150, 111 ± 119.
[17] The original conditions described by Luche and Gamal using MeOH
as a solvent gave a low yield of 4 and recovered starting material; J. L.
Luche, A. L. Gamal, J. Am. Chem. Soc. 1979, 101, 5848. On the other
hand, the reduction with the solvent system (EtOH CH2Cl2) reported
by Danishefsky et al. gave our desired product 4 in better yield and
with higher selectivity; S. J. Danishefsky, H. G. Selnick, R. E. Zelle,
M. P. DeNinno, J. Am. Chem. Soc. 1988, 110, 4368 ± 4378.
[18] The stereochemistry was controlled by a steric factor rather than an
orientation of the neighboring hydroxy group; see ref. [9d].
[19] N. Yamamoto, T. Nishikawa, M. Isobe, Synlett 1995, 505 ± 506.
[20] P. H. Carisen, T. Katsuki, V. S. Martin, K. B. Sharpless, J. Org. Chem.
1981, 46, 3936 ± 3938.
[21] Determination of the configuration at the C9 position is included in
the Supporting Information.
[22] J. Petrov, I. Atanassova, A. Balabanova, N. Mollov, Izv. Khim. 1990,
23, 53 ± 57.
[23] Detailed analysis of the crude product revealed that one of the
products lacked H9, indicating that an oxidation occurred at the C9
position. Kishi et al. reported on the lability of the acetoxy group at
the C9 position in their total synthesis of tetrodotoxin. See: a) ref. [9b];
b) H. Tanino, S. Inoue, M. Aratani, Y. Kishi, Tetrahedron Lett. 1974,
335 ± 338.
[24] M. Xie, D. A. Berges, M. J. Robins, J. Org. Chem. 1996, 61, 5178 ± 5179.
[25] K. Mori, M. Tominaga, T. Takigawa, M. Matsui, Synthesis 1973, 790 ±
791.
[26] Pyridine was indispensable in this addition, while we used DMF in our
model studies.[13b]
[27] In this specific case, acetic anhydride was an indispensable solvent.
This reaction was much slower or unreproducible when a mixture of
acetic anhydride and triethylamine was employed as in our previous
(successful) model studies.[13b]
[28] 4-epi- and 4,9-Anhydrotetrodotoxins have been found as minor
constituents with tetrodotoxins from natural sources.[1, 3a, 3c] For
example, 5,6,11-trideoxytetrodotoxin was also isolated with 4-epi-
trideoxytetrodotoxin in the ratio (ca. 1:1.2).[3a] Compound 22 is the
first example of 4,9-anhydrotetrodotoxin among a series of 5-deoxy-
tetrodotoxin analogues.
The treatment of pentacarbonyl(difluoroboroxycarbene)-
molybdenum complexes 1 (formed by reaction of lithium
pentacarbonyl acyl molybdate and boron trifluoride diethyl
ether complex)[7] with vinyl ketones 2 and aldehydes 3 in
diethyl ether at temperatures from 60 to 208C led to b-
hydroxyketones 4. According to the NMR studies the syn
products were formed exclusively (Scheme 1 and Table 1).[3a, 8]
OH
O
R3CHO
3
R2
R3
OBF2
R1
(CO)5Mo
1. –60°C
2. H2O
RT
R1
1
+
4
R4
CHO
OH
O
R2
R4
R2
O
5
X
2
X
1. –60
2. H2O
–20°C
R1
6
Scheme 1. Reaction of difluoroboroxycarbenemolybdenum complexes
with vinyl ketones and aldehydesÐFormation of syn-b-hydroxyketones 4
or 6.
[*] Prof. Dr. J. Barluenga, Dr. F. Rodríguez,
Ä
Â
Dr. F. J. Fananas, Dr. E. Rubio
Instituto Universitario de Química
Â
Organometalica ªEnrique Molesº
Unidad Asociada al CSIC, Universidad de Oviedo
Julian Clavería 8, E-33071-Oviedo (Spain)
Fax: (‡349)8-510-34-50.
E-mail: barluenga@sauron.quimica.uniovi.es
[**] This work was supported by DGICYT (Grant PB97 ± 1271).
3084
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3820-3084 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 20