Complete relative stereochemistry of maitotoxin

Article abstract of DOI:10.1021/ja961230+

By addressing the relative stereochemistry of the four acyclic portions via organic synthesis, the complete relative stereochemistry of maitotoxin (MTX) has been established as 1B. The relative stereochemistry of the C.1-C.15 portion was elucidated via a two-phase approach: (1) the synthesis of the eight diastereomers possible for model C, representing the C.1-C.11 portion, and the eight diastereomers possible for model D, representing the C.11-C.15 portion, and the comparison of their proton and carbon NMR characteristics with those of MTX, concluding that 9 and 35 represent the relative stereochemistry of the corresponding portions of MTX; (2) the synthesis of the two remote diastereomers 51 and 52, and comparison of their proton and carbon NMR characteristics with those of MTX, concluding that 51 represents the relative stereochemistry of the C.1-C.15 portion of MTX. The relative stereochemistry of the C.35-C.39, C.63-C.68, and C.134-C.142 acyclic portions was established via (1) the synthesis of the 8, 8, and 16 diastereomers possible for models E, F, and G, respectively, and (2) the comparison of their proton and carbon NMR characteristics with those of MTX, concluding that 81, 117, and 187, respectively, represent the relative stereochemistry of the corresponding portions of MTX. Some biogenetic considerations have been given to speculate on the absolute configuration of MTX. The vicinal proton coupling constants observed for models 51, 81, 117, and 187 were used to elucidate their preferred solution conformation. Assembling the preferred solution conformations found for the four acyclic portions allows one to suggest that the approximate global conformation of MTX is represented by the shape of a hook, with the C.35-C.39 portion being its curvature. MTX appears to be conformationally relatively rigid, except for conformational flexibility around the C.7-C.9 and C.12-C.14 portions. On the basis of the experimental results gained in the current work, coupled with those in the AAL-toxin/fumonisin area, it has been pointed out that the structural properties of 51, 81, 117, 187 and their diastereomers are inherent to the specific stereochemical arrangement of the small substituents on the carbon backbone and are independent from the rest of the molecule. Thus, it has been suggested that each of these diastereomers has the capacity to install a unique structural characteristic through a specific stereochemical arrangement of substituents on the carbon backbone, and that fatty acids and related classes of compounds may be able to carry specific information and serve as functional materials in addition to structural materials.

Full text of DOI:10.1021/ja961230+

7946
J. Am. Chem. Soc. 1996, 118, 7946-7968
Complete Relative Stereochemistry of Maitotoxin
Wanjun Zheng, John A. DeMattei, Jiang-Ping Wu, James J.-W. Duan,
Laura R. Cook, Hitoshi Oinuma, and Yoshito Kishi*
Contribution from the Department of Chemistry, HarVard UniVersity,
Cambridge, Massachusetts 02138
ReceiVed April 15, 1996X
Abstract: By addressing the relative stereochemistry of the four acyclic portions via organic synthesis, the complete
relative stereochemistry of maitotoxin (MTX) has been established as 1B. The relative stereochemistry of the C.1-
C.15 portion was elucidated via a two-phase approach: (1) the synthesis of the eight diastereomers possible for
model C, representing the C.1-C.11 portion, and the eight diastereomers possible for model D, representing the
C.11-C.15 portion, and the comparison of their proton and carbon NMR characteristics with those of MTX, concluding
that 9 and 35 represent the relative stereochemistry of the corresponding portions of MTX; (2) the synthesis of the
two remote diastereomers 51 and 52, and comparison of their proton and carbon NMR characteristics with those of
MTX, concluding that 51 represents the relative stereochemistry of the C.1-C.15 portion of MTX. The relative
stereochemistry of the C.35-C.39, C.63-C.68, and C.134-C.142 acyclic portions was established via (1) the synthesis
of the 8, 8, and 16 diastereomers possible for models E, F, and G, respectively, and (2) the comparison of their
proton and carbon NMR characteristics with those of MTX, concluding that 81, 117, and 187, respectively, represent
the relative stereochemistry of the corresponding portions of MTX. Some biogenetic considerations have been given
to speculate on the absolute configuration of MTX. The vicinal proton coupling constants observed for models 51,
81, 117, and 187 were used to elucidate their preferred solution conformation. Assembling the preferred solution
conformations found for the four acyclic portions allows one to suggest that the approximate global conformation of
MTX is represented by the shape of a hook, with the C.35-C.39 portion being its curvature. MTX appears to be
conformationally relatively rigid, except for conformational flexibility around the C.7-C.9 and C.12-C.14 portions.
On the basis of the experimental results gained in the current work, coupled with those in the AAL-toxin/fumonisin
area, it has been pointed out that the structural properties of 51, 81, 117, 187 and their diastereomers are inherent to
the specific stereochemical arrangement of the small substituents on the carbon backbone and are independent from
the rest of the molecule. Thus, it has been suggested that each of these diastereomers has the capacity to install a
unique structural characteristic through a specific stereochemical arrangement of substituents on the carbon backbone,
and that fatty acids and related classes of compounds may be able to carry specific information and serve as functional
materials in addition to structural materials.
I. Introduction
variety of biological events; most notably, it has been shown
to act on the voltage-sensitive Ca2+ channels.5
Ciguatera is a type of poisoning induced by ingestion of coral
reef fish, which annually affects an estimated 20 000 people
worldwide.1 The toxification mechanism of these species was
not known until 1976, when Yasumoto and co-workers showed
that the epiphytic dinoflagellate Gambierdiscus toxicus is the
causative organism whose toxins are transferred to the coral
biota through the food chain, ultimately residing in carnivorous
fish.2 Several extraordinarily complex natural products, includ-
ing ciguatoxin, which was isolated by Scheuer in 19673 and
the structure of which was elucidated by Yasumoto in 1989,4
were identified as the toxic principles of G. toxicus.1a Among
these, maitotoxin (MTX; 1A, Figure 1) is unique in terms of
its molecular size and structural complexity as well as its
extremely potent bioactivity. MTX is known to trigger a wide
Yasumoto has played the seminal role in the advancement
of the chemistry of MTX, including its isolation and structure
elucidation.6 The gross structure of MTX was elucidated in
1993. The relative stereochemistry of the fused rings as well
as the directly connected ether rings was reported in 1994.
However, the relative stereochemistry of the chiral centers
embedded in the four acyclic portions, i.e., at the C.1-C.15,
C.35-C.39, C.63-C.68, and C.134-C.142 portions,7 remained
unknown. In this paper, we present our efforts to establish the
relative stereochemistry of these four acyclic portions, and
consequently the complete relative stereochemistry of MTX,
via organic synthesis. On the basis of the vicinal proton
coupling constants observed for the models of these four acyclic
portions, we also suggest the preferred solution conformation
of MTX.8-10
X Abstract published in AdVance ACS Abstracts, August 1, 1996.
(1) For recent reviews, see: (a) Yasumoto, T.; Murata, M. Chem. ReV.
1993, 93, 1897-1909. (b) Faulkner, D. J. Nat. Prod. Rep. 1995, 12, 223-
269. (c) Strichartz, G.; Castle, N. Pharmacology of marine toxins. Effects
on membrane channels; ACS Symposium Series 418; American Chemical
Society: Washington, DC, 1990; pp 2-20.
(5) (a) Takahashi, M.; Ohizumi, Y.; Yasumoto, T. J. Biol. Chem. 1982,
257, 7287-7289. (b) Kobayashi, M.; Ochi, R.; Ohizumi, Y. Br. J.
Pharmacol. 1987, 92, 665-671. (c) Soergel, D. G.; Yasumoto, T.; Daly, J.
W.; Gusovsky, F. Mol. Pharmacol. 1992, 41, 487-493.
(6) (a) Murata, M.; Iwashita, T.; Yokoyama, A.; Sasaki, M.; Yasumoto,
T. J. Am. Chem. Soc. 1992, 114, 6594-6596. (b) Murata, M.; Naoki, H.;
Iwashita, T.; Matsunaga, S.; Sasaki, M.; Yokoyama, A.; Yasumoto, T. J.
Am. Chem. Soc. 1993, 115, 2060-2062. (c) Murata, M.; Naoki, H.;
Matsunaga, S.; Satake, M.; Yasumoto, T. J. Am. Chem. Soc. 1994, 116,
7098-7107. (d) Satake, M.; Ishida, S.; Yasumoto, T. J. Am. Chem. Soc.
1995, 117, 7019-7020.
(2) Yasumoto, T.; Bagnis, R.; Vernoux, J. P. Bull. Jpn. Soc. Sci. Fish.
1976, 42, 359-365.
(3) (a) Scheuer, P. J.; Takahashi, W.; Tsutsumi, J.; Yoshida, T. Science
1967, 155, 1267-1268. (b) Nukina, M.; Koyanagi, L. M.; Scheuer, P. J.
Toxicon 1984, 22, 169-176.
(4) Murata, M.; Legrand, A. M.; Ishibashi, Y.; Yasumoto, T. J. Am.
Chem. Soc. 1989, 111, 8929-8931.
(7) The numbering used throughout this paper represents the position of
the corresponding carbon atom in MTX.
S0002-7863(96)01230-9 CCC: $12.00 © 1996 American Chemical Society

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7947
Figure 1. 1A: Maitotoxin.
Scheme 1
II. Relative Stereochemistry
1. C.1-C.15 Acyclic Portion. This portion of MTX
contains seven unassigned chiral centers on the acyclic back-
bone. In order to establish the complete relative stereochemistry
of MTX, the configuration of these chiral centers relative to
the C.15 position must also be addressed. Therefore, 128
possible diastereomers exist to represent the relative stereo-
chemistry of this portion of MTX. A new concept was recently
advanced to address this type of problem, and its validity and
usefulness were demonstrated by the stereochemical assignment
of the AAL-toxin/fumonisin family of natural products.11
This concept was first illustrated using AAL toxin TA. The
backbone of AAL toxin TA had seven unknown asymmetric
centers, and therefore there were 64 possible enantiomeric pairs
of diastereomers to represent its stereochemistry. Its stereo-
chemistry could be established via organic synthesis; namely,
all of the possible diastereomers could be synthesized and
compared with the aminopentol backbone derived from the
natural product. However, there were two concerns about this
standard approach. First, the synthesis of 64 possible enantio-
meric pairs of diastereomers represented a substantial effort.
Second, and more critically, even if all of the possible
diastereomers were prepared, there was no assurance that all of
them could be differentiated by currently available spectroscopic
and/or chromatographic techniques.
The new approach consisted of three phases. The amino-
pentol backbone derived from AAL toxin TA was composed of
two distinct halves with the asymmetric centers separated by
five methylene units (Scheme 1). This separation was assumed
to be great enough for each half to have chemical properties
independent from its opposite half. Thus, models A and B
should represent the relative stereochemical characteristics of
the left and right halves of the AAL toxin TA backbone,
respectively. With this assumption, the first phase of this
approach involved (1) the choice of models A and B properly
representing the structural characteristics of the left and right
halves of AAL toxin TA, respectively; (2) the synthesis of all
the diastereomers possible for A and B; and (3) the determi-
nation of which stereoisomers of A and B represent the left
and right halves of the AAL toxin TA aminopentol backbone,
respectively. There were eight diastereomers possible for the
left half and four diastereomers possible for the right half.
Therefore, this operation required the synthesis of only 12
diastereomers instead of 64 and would allow us to determine
the relative stereochemistry of the left and right halves of the
aminopentol backbone of AAL toxin TA.
(8) The complete relative stereochemistry of MTX was given as a part
of the presentation at the Glaxo symposium (November 1995), the Prelog
lecture (November 1995), and the 211th National Meeting of the American
Chemical Society (March 1996) by one (Y.K.) of us.
(9) Since we began our work, Tachibana and co-workers reported the
relative stereochemistry of the C.35-C.39 and C.63-C.68 portions of
MTX: (a) Sasaki, M.; Matsumori, N.; Murata, M.; Tachibana, K.;
Yasumoto, T. Tetrahedron Lett. 1995, 36, 9011-9014. (b) Sasaki, M.;
Nonomura, T.; Murata, M.; Tachibana, K. Tetrahedron Lett. 1995, 36,
9007-9010.
(10) We thank Professor Murata for a preprint, discussing the C.5-C.9
relative stereochemistry assignment by long-range carbon-proton coupling
constants: Matsumori, N.; Nonomura, T.; Sasaki, M.; Murata, M.; Ta-
chibana, K.; Satake, M.; Yasumoto, T. Tetrahedron Lett. 1996, 37, 1269-
1272.
(11) (a) Boyle, C. D.; Harmange, J.-C.; Kishi, Y. J. Am. Chem. Soc.
1994, 116, 4995-4996. (b) Harmange, J.-C.; Boyle, C. D.; Kishi, Y.
Tetrahedron Lett. 1994, 35, 6819-6822. (c) Boyle, C. D.; Kishi, Y.
Tetrahedron Lett. 1995, 36, 5695-5698 and references cited therein.
The second phase was to determine the relative stereochem-
istry between the left and right halves of the AAL toxin TA
backbone. The left half is represented by an oval, and since
only the relative stereochemistry of the left half would be known

7948 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
at this stage, it could exist in either an R or S configuration.
The same holds true for the right half, which is depicted as a
rectangle. Coupling of the R left half with the S right half would
provide an R-S aminopentol backbone. Likewise, coupling of
the R left half with the R right half would yield an R-R
aminopentol backbone. In the case of diastereomers bearing
chiral centers in close proximity, the pair could be distinguished
from each other. However, as the diastereomers in question
had chiral centers at remote positions (we refer to such cases
as remote diastereomers), the degree of chemical communication
between the two remote moieties should be negligibly small, if
any exists. Indeed, the proposed first phase of this approach
relied on this very assumption. We planned to apply the concept
of molecular recognition to distinguish these remote diastere-
omers from each other. Interactions of these diastereomers with
an achiral foreign molecule might differ significantly enough
to be detected by currently available spectroscopic and/or
chromatographic techniques.
Scheme 2a
The final phase of this approach was to distinguish the
enantiomers of the AAL toxin TA aminopentol backbone, i.e.,
R-S vs S-R or R-R vs S-S. This problem could be solved
by several methods, including the possibility of using a chiral,
instead of achiral, foreign molecule for molecular recognition.
The validity and usefulness of this approach was demonstrated
first for determination of the relative and absolute stereochem-
istry of the aminopentol backbone of AAL toxin TA and then
for the aminotetrol backbone of fumonisin B2.11
Returning to the C.1-C.15 moiety of MTX, we were curious
to test whether two structural moieties spaced by only two
methylene groups could be treated independently, in order to
determine its relative stereochemistry by synthesizing only 18
diastereomers instead of 128, i.e., eight diastereomers possible
for model C, eight diastereomers possible for model D, and
two remote diastereomers. At the same time, however, we still
hoped that some residual chemical communication between the
left and right halves of this portion of MTX could be detected
by currently available spectroscopic means. In this context, it
should be noted that even a trace amount of the natural product
was not available for our work, and we solely depended on the
published NMR data to conclude the complete relative stereo-
chemical assignment.12
a (a) (E)-(R,R)-Crotylboronate, 4 Å molecular sieves, PhMe, -78
°C, 79%; (b) (1) TBSOTf, pyr, CH2Cl2, 23 °C; (2) catecholborane,
RhCl(PPh3)3, THF, 23 °C;64 10% NaOH, H2O2, 23 °C; (3) (COCl)2,
DMSO, Et3N, CH2Cl2, -78 °C; (c) vinyllithium, THF, -78 °C, 3 h, 0
°C, 10 min, 79% over four steps; (d) (1) TBSOTf, pyr, CH2Cl2, 23 °C;
(2) OsO4, NMO, EtOH-THF-H2O (1:2:1), 23 °C, 16 h; NaIO4, 23
°C, 24 h, 76% over two steps; (3) (E)-4-(tert-butyldimethylsiloxy)-2-
buten-2-yl)lithium,65 THF, -78 °C, 30 min; (4) (COCl)2, DMSO, Et3N,
CH2Cl2, -78 °C, 88% over two steps; (5) DDQ, CH2Cl2-pH 7 buffer
(10:1), 23 °C, 1 h, 80%; (6) Ph3PdCH2, pentane, 0 °C, 90 s, 67%; (7)
(COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 96%; (e) (1) (4-(1-ethoxy-
ethoxy)-3,3-dimethylbutyl)lithium,66 THF, -78 °C, 15 min, 85%; (2)
SO3‚pyr, pyr, 23 °C, 15 h;67 (3) HF‚pyr,68 MeOH, 23 °C, 48 h, 89%
over two steps.
chemistry at C.7 and C.8 was set by using the Roush boronate
chemistry.13 The reaction of the (E)-(R,R)-crotylboronate with
the aldehyde 214 was predicted to yield 3 preferentially. Indeed,
this pair of reagent and substrate led almost exclusively to a
single diastereomer. The olefin 3 was converted to the aldehyde
4, which was then coupled with vinyllithium, to yield a 1:1.2
mixture of the two expected allylic alcohols 5 and 6, which
In order to test this experimentally, models C and D were
chosen to represent the C.1-C.11 and C.11-C.15 portions,
respectively. Scheme 2 summarizes the synthesis of all dia-
stereomers possible for C. The relative and absolute stereo-
(13) (a) Roush, W. R.; Walts, A. E.; Hoong, L. K. J. Am. Chem. Soc.
1985, 107, 8186-8190. (b) Roush, W. R.; Palkowitz, A. D.; Ando, K. J.
Am. Chem. Soc. 1990, 112, 6348-6359 and references cited therein.
(14) England, P.; Chun, K. H.; Moran, E. J.; Armstrong, R. W.
Tetrahedron Lett. 1990, 31, 2669-2672.
(12) Yasumoto and co-workers completed an astonishing work to make
the proton and carbon chemical shift assignment for each atom with 0.01
and 0.1 decimal place accuracy, respectively.6

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7949
were chromatographically separated. The relative and absolute
configuration of 5 and 6, consequently of 3, was established
via NMR analyses of their cyclic derivatives, and the optical
purity of 3 was estimated to be approximately 70-80%.15
The diastereomers 5 and 6 were subjected to the seven-step
transformation, i.e., (1) protection of the C.5 alcohol, (2)
oxidative cleavage of the vinyl group to generate the C.4
aldehyde, (3) coupling with the lithium reagent to form the C.3-
C.4 bond, (4) oxidation of the resultant allylic alcohol to form
the enone, (5) deprotection of the C.9 alcohol, (6) olefination
to form the diene, and (7) oxidation of the C.9 alcohol to form
the C.9 aldehydes. Each of the aldehydes 7 and 8 was then
subjected to the next carbon-carbon bond forming reaction to
yield the two possible C.9 diastereomers, which were separated
by chromatography. Nuclear Overhauser effect (NOE) experi-
ments with the C.8 and C.9 acetonides, prepared from both
diastereomers, established their C.9 stereochemistry.16 Each
diastereomer was then subjected to sulfation at C.9, followed
by deprotection, to furnish the four diastereomers 9-12.
Following the same sequence of reactions on the syn-adduct
13, obtained from the reaction of 2 with (Z)-(R,R)-crotylbor-
onate, the remaining four diastereomers 14-17 were also
obtained (Scheme 2B).
Chart 1. Difference in Proton Chemical Shifts between
MTX and Each of 9-12 and 14-17 (500 MHz, 2:1
CD3CN-D2O)a
1
All of the eight diastereomers were subjected to H and 13C
NMR studies. Charts 1 and 2 show the difference in the
chemical shifts of protons and carbons between MTX12 and each
diastereomer synthesized. This exercise clearly demonstrated
that (1) each diastereomer exhibited distinct spectroscopic
characteristics differing from the others, and (2) only the
diastereomer 9 displayed spectroscopic characteristics that were
virtually identical to those of MTX, therefore establishing that
the relative stereochemistry of the C.1-C.11 portion of MTX
is represented by that of 9.
Model D was chosen for the study of the spectroscopic
characteristics of the C.11-C.15 portion of MTX. Scheme 3
outlines the synthesis of the C.13-C.19 portion from the
aldehyde 18,17 readily available from D-ribose. Addition of the
allyl group was best achieved by the reaction of allylindium
bromide with 18 to yield a 6:1 mixture of the two possible
diastereomers. The major product was tentatively assigned to
a The x- and y-axes represent carbon number and ∆δ (∆δ ) δMTX
- δSyntheticModel in ppm), respectively, for all the charts in this paper.
be the desired diastereomer, which was confirmed later (see
Scheme 4). The major diastereomer was then converted into
the aldehyde 19. Addition of the Roush (Z)-(R,R)-crotylboronate
to 19 furnished the expected product 20 exclusively, whereas
addition of the corresponding (Z)-(S,S)-reagent gave 1.5:1
mixtures of 20 and 21. Similarly, addition of (E)-(R,R)- and
(E)-(S,S)-crotylboronates to 19 yielded a 9:1 and 1:1.2 mixture
of 22 and 23, respectively.
On the basis of the precedents of Roush’s work,13 the
exclusive product formed with (Z)-(R,R)-crotylboronate was
anticipated to be 20, whereas the major product formed with
(E)-(R,R)-crotylboronate was expected to be 22. Thus, 20 and
22 should have the same stereochemistry at C.14 but the
opposite stereochemistry at C.15, which was confirmed from
the following experiments (Scheme 4). Oxidation of the C.15
hydroxy group of 20, deprotection of the C.19 alcohol, and
reductive cyclization furnished exclusively the expected C.15
equatorial product 26; the vicinal proton coupling constants
(J15,16 ) 9.7 Hz and J18,19 ) 9.5 Hz) observed for 26 led to the
conclusion that the C.15 and C.19 substituents were equatorial.
The same sequence of reactions on 22, the major product formed
by the reaction of (E)-(R,R)-crotylboronate with 19, also
furnished the identical equatorial product 26. Since 20 was
formed from (Z)-crotylboronate and 22 was formed from (E)-
crotylboronate, their C.14 and C.15 relative stereochemistry
should be syn and anti, respectively.
(15) The absolute configuration and optical purity of 3 were established
via its transformation into the Mosher ester of 2-methyl-1-butanol and
comparison of it with the corrresponding Mosher ester prepared from (S)-
(-)-2-methyl-1-butanol (Aldrich). The relative stereochemistry of 5 and 6
was established via the NMR study of the cyclic compounds i and ii,
prepared from them in four steps, i.e., (1) BnBr/NaH, (2) TBAF, (3) OsO4/
NaIO4, and (4) Ac2O/pyr. At the early phase of the work, the mixture of 5
and 6 was used for the following transformation. However, using the pure
materials, 5 and 6 were demonstrated to yield 7 and 8, respectively.
(16) The relative configuration of 9 and 10 was established from the
NOE experiments of the acetonides iii and iv, prepared in three steps, i.e.,
(1) pyr/dioxane/120 °C, (2) MeC(OMe)2Me/PPTS, and (3) Ac2O/pyr. The
acetonide iii exhibited clear cross peaks between the C.8/C.9 and C.7/C.10
protons, whereas the acetonide iv did not show the corresponding cross
peaks but did show cross peaks between the C.7/C.9 and C.9/C.145 protons.
This conclusion was further supported by two additional
experiments. First, 20 gave exclusively 26 via the alternative
mode of reductive cyclization, i.e., 20 f 25 f 26; the C.15
stereochemistry of 20 was retained in this transformation,
(17) Tadano, K.; Maeda, H.; Hoshino, M.; Iimura, Y.; Suami, T. J. Org.
Chem. 1987, 52, 1946-1956.

7950 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Chart 2. Difference in Carbon Chemical Shifts between
MTX and Each of 9-12 and 14-17 (125 MHz, 2:1
CD3CN-D2O)
Scheme 3a
a (a) (1) Allyl bromide, In, DMF, 23 °C, ratio of R:â ) 6:1, 85%;69
(2) H2, Lindlar catalyst, EtOAc, 23 °C, overnight; (3) NaH, MPMCl,
DMF, 23 °C, 12 h; (4) n-Bu4NF, THF, 50 °C, 30 min, 84% over three
steps; (5) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 92%; (b) crotyl-
boronate, 3 Å molecular sieves, PhMe, -78 °C.
Scheme 4a
whereas the C.19 stereochemistry was retained in the previous
mode of reductive cyclization. Second, the major anti-product
23, obtained with the (E)-(S,S)-crotylboronate, furnished ex-
clusively 28 in the three-step sequence of reductive cyclization
(Scheme 4B). The vicinal proton coupling constants (J15,16
)
9.7 Hz and J18,19 ) 9.5 Hz) demonstrated that the C.15 and
C.19 substituents of 28 were equatorial, and consequently 28
was the C.14 diastereomer of 26. These results were consistent
with the conclusion derived from the previous experiments.
The Roush crotylboronate chemistry should allow the addition
of the next propionate unit to 26 and 28. However, it was
difficult to predict the influence of the pre-existing stereocenters
on the overall stereochemical outcome.18 The aldehydes 29 and
38, derived from 26 and 28, respectively, were subjected to (E)-
(R,R), (E)-(S,S)-, (Z)-(R,R)-, and (Z)-(S,S)-crotylboronates, fol-
lowed by hydroboration and deprotection, to yield the expected
products (Scheme 5). All of the possible diastereomers 34-
37 were obtained from the aldehyde 29, whereas only two
diastereomers 43 and 44 were obtained from the aldehyde 38.
While these observations were intriguing, it was necessary
to obtain the remaining two diastereomers for the present work,
which was accomplished as summarized in Scheme 6. The
conjugate reduction of the enone 45, followed by hydride
reduction and deprotection of the TBS and benzyl group, gave
a 1.0:1.9:1.6:1.0 mixture of the four possible diastereomers 43,
44, 46, and 47. The last two products corresponded to the
diastereomers not obtained in the previous experiments.
a (a) (COCl)2, DMSO, Et3N, CH2Cl2 -78 °C; (b) (1) DDQ, CH2Cl2-
H2O (10:1), 23 °C, 1 h; (2) BF3‚OEt2, Et3SiH, CH3CN-CH2Cl2 (10:
1), -20 °C, 15 min; 75% over three steps; (c) (1) TESOTf, 2,6-lutidine,
CH2Cl2; (2) DDQ, CH2Cl2-H2O (10:1), 23 °C, 1 h, 80% over two
steps; (3) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 72%; (d) (1) HF‚pyr,
THF, 23 °C, 3 h, 76%; (2) same as step b2, 82%.
1
All of the eight diastereomers were subjected to H and 13C
NMR studies. Charts 3 and 4 show the difference in the
(18) For a review of double stereodifferentiation, for example, see:
Masamune, S.; Choy, W.; Petersen, J. S.; Sita, L. R. Angew. Chem., Int.
Ed. Engl. 1985, 24, 1-76.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7951
Scheme 5a
Chart 3. Difference in Proton Chemical Shifts between
MTX and Each of 34-37, 43, 44, 46, and 47 (500 MHz, 2:1
CD3CN-D2O)
Chart 4. Difference in Carbon Chemical Shifts between
MTX and Each of 34-37, 43, 44, 46, and 47 (125 MHz, 2:1
CD3CN-D2O)
a (a) Crotylboronate, 4 Å molecular sieves, PhMe, -78 °C; (b) (1)
catecholborane, RhCl(Ph3P)3 (cat.), THF, 0-23 °C, 1 h; 10% NaOH,
30% H2O2, 0-23 °C, 3 h, 80%; (2) H2, Pd(OH)2 on C (cat.), EtOH, 23
°C, 3 h, 99%.
Scheme 6a
a (a) (1) 4-(tert-Butyldimethylsiloxy)-1-buten-2-yl iodide,70 1% NiCl2/
CrCl2, THF-DMF (2:1), 23 °C, 23 h; (2) (COCl)2, DMSO, Et3N,
CH2Cl2, -78 °C, 33% over two steps (48% recovered aldehyde 38);
(b) (1) (CuH‚PPh3)6, wet PhH, 23 °C, 3 h, ratio of R: â at C.12)1:1,
89%;71 (2) DIBAL, CH2Cl2, -78 °C, 0.5 h, then 23 °C, 10 min, 85%;
(3) n-Bu4NF, THF, 23 °C, 1 h, 95%; (4) H2, Pd(OH)2 on C (cat.), EtOH,
23 °C, 3 h, 95%.
chemical shifts for protons and carbons between MTX and each
of the diastereomers synthesized. This exercise once again
demonstrated that (1) each diastereomer exhibited distinct
spectroscopic characteristics differing from the others and (2)
only diastereomer 35 displayed spectroscopic characteristics that
were virtually identical to those of MTX.
The C.12 and C.13 stereochemistry of 35 was then addressed.
As noted earlier (Scheme 5), addition of (E)-(R,R)-crotylbor-
onate to the aldehyde 29 yielded the two expected products in
a 2.2:1 ratio. Therefore, the C.12 and C.13 stereochemistry of
these products should be anti, and they should correspond to
either 30 and 31 or vice versa. The major product 30 was
subjected to a sequence of degradation reactions,19 which
furnished the triol 49 (Scheme 7). Seven well-separated signals
(19) This degradation was carried out according to the protocol used in
the palytoxin case: Ko, S. S.; Finan, J. M.; Yonaga, M.; Kishi, Y.; Uemura,
D.; Hirata, Y. J. Am. Chem. Soc. 1982, 104, 7364-7367.

7952 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Scheme 7a
throughput and, more importantly, unambiguous stereochemical
assignment of the newly introduced chiral center at C.9. The
first carbon-carbon bond formation was accomplished by
coupling of the aldehyde 53 with the dibromo olefin 50, to yield
an easily separable 1.5:1 mixture of the two expected products
54 and 55. The acetylenic bond in 54 and 55 provided a useful
handle to establish the stereochemistry of the newly introduced
chiral center via degradation reactions.24
Formation of the C.3-C.4 bond was achieved by coupling
the aldehyde 56, derived from 54, with the lithium reagent
prepared from (E)-4-(tert-butyldimethylsiloxy)-2-buten-2-yl bro-
mide.25 Oxidation of the resultant allylic alcohol and depro-
tection of the C.9 alcohol provided the enone 57. Wittig
reaction, sulfation of the C.9 alcohol, and deprotection of all
the silyl protecting groups furnished the final product 51, which
a (a) (1) O3, -78 °C, MeOH; NaBH4, 94%; (2) MeC(OMe)2Me,
p-TsOH (cat.), 23 °C, 1 h; (3) H2, Pd(OH)2 on C (cat.), EtOAc, 52%
over two steps; (b) (1) NaIO4, THF-pH 7 buffer (1:1), 23 °C, 0.5 h;
(2) MCPBA, CH2Cl2-pH 7 buffer (2:1), 23 °C, 5 h; (3) LAH, THF,
23 °C, 0.5 h; (4) MeOH, p-TsOH (cat.), 23 °C, 2 h, 46% over four
steps.
was isolated as its sodium salt by silica gel column chroma-
tography. Using the same chemistry but with 59, the antipode
of 53, the remote diastereomer 52 was also synthesized (Scheme
8B).
Figure 2. Crystal structure of 50. The TBS groups in the ORTEP
representation were removed for clarity.
The two remote diastereomers 51 and 52 were subjected to
NMR studies. Charts 5 and 6 summarize the difference in the
chemical shifts observed between MTX and the remote dia-
stereomers. As expected, both the remote diastereomers 51 and
were observed in its 13C NMR spectrum,20 establishing the
relative stereochemistry for 49, and consequently that for 30.
This conclusion was further supported by a single crystal X-ray
analysis of 50 derived from 31 (Figure 2).21 These experimental
results established that the relative stereochemistry of the C.12-
C.15 portion of MTX is represented by that of 50.
Having established the relative stereochemistry for both the
C.1-C.11 and C.11-C.15 portions, we planned to synthesize
the two possible remote diastereomers 51 and 52, and to
compare their spectroscopic characteristics with those of MTX.
Obviously, the olefin 31 (Scheme 7) was an ideal starting
material for this synthesis, but one practical problem had to be
addressed, i.e., 31 was obtained only as the minor product via
the crotylboronate chemistry. Among several reagents and
conditions tested, the addition of the zinc reagent prepared from
crotyl bromide to the aldehyde 29 in THF at -93 °C gave the
most satisfactory result.22 The method to construct the two
remote diastereomers was then studied, and the synthetic route
shown in Scheme 8 met our requirements,23 including material
52 were found to exhibit NMR characteristics, particularly 13
C
NMR, very similar to each other, and very similar to the NMR
characteristics of MTX in that region.26 However, upon close
1
examination of their H NMR spectra, small but significant
differences were detected between 51 and 52 and also between
52 and MTX. It is important to recognize that the differences
(23) In the model C series, the C.9-C.10 bond was formed by coupling
the aldehyde with the alkyllithium reagent. However, in the model 51/52
series, the attempted preparations of the corresponding lithium reagent failed.
The coupling of the acetylene anion prepared from 50 with the complete
C.1-C.9 aldehyde was very efficient, but the selective reduction of the
acetylene over the diene proved problematic.
(24) The relative configuration of 54 and 55 was established from the
vicinal coupling constants observed for the bis-acetonides v and vi, prepared
from them in 5 steps, i.e., (1) H2/Lindlar catalyst, (2) BnOC(dNH)CCl3/
TfOH, (3) O3, followed by NaBH4, (4) HCl (cat.)/aqueous MeOH, and (5)
MeC(OMe)2Me/PPTS.
(20) 13C NMR (125 MHz, CDCl3) δ 75.4 ppm, 66.8, 66.6, 39.5, 38.7,
14.0, 9.7.
(21) We thank Dr. A. Marcus Semones in this group for this X-ray
analysis.
(22) Zn(0), Cr(II), and In(0) were tested in several solvent systems at a
wide range of temperatures. Among them, Zn(0) in THF at -93 °C yielded
a 51:35:9:5 mixture (1H NMR) of the products, with the major product
being the desired 31 and the second major product being 30. The desired
product 31 was isolated in 44% yield on a preparative scale.
(25) Meyer, S. D.; Miwa, T.; Nakatsuka, M.; Schreiber, S. L. J. Org.
Chem. 1992, 57, 5058-5060.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7953
Scheme 8a
Chart 5. Difference in Proton Chemical Shifts between
MTX and Each of 51 and 52 (500 MHz, 1:1
C5D5N-CD3OD)
Chart 6. Difference in Carbon Chemical Shifts between
MTX and Each of 51 and 52 (125 MHz, 1:1
C5D5N-CD3OD)
a (a) (1) TBSOTf, Et3N, CH2Cl2, 98%; (2) O3, CH2Cl2-MeOH-
pyr (10:5:0.3); NaBH4, 78%; (3) TESCl, Et3N, DMAP (cat.), CH2Cl2,
23 °C, 1 h, 96%; (4) H2, Pd(OH)2 on C (cat.), EtOAc, 23 °C, 1 h,
90%; (5) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 79%; (b) (1) Na,
NH3, -33 °C, 1 h, 100%; (2) TBSCl, AgNO3, pyr, DMF, 23 °C, 12 h,
86%; (3) TBSOTf, i-Pr2NEt, CH2Cl2, 23 °C, 42 h, 88%; (4) O3,
CH2Cl2-MeOH (12:1), -78 °C; NaBH4, 23 °C, 1 h, 82%; (5) (COCl)2,
DMSO, Et3N, CH2Cl2, -78 °C, 99%; (6) PhHgCBr3, PPh3, PhH, 80
°C, 1 h, 97%;72 (c) 50, n-BuLi, -78 °C, 1 h, then 0 °C, 6 min, -78
°C; 53, 25 min, 97%; (d) (1) H2, 1% Rh on Al2O3 (cat.), EtOAc, 100%;
(2) p-methoxybenzyl 2,2,2-trichloroacetimidate, TfOH (trace), Et2O,
23 °C, 12 h, 37% (60% recovered starting material); (3) PPTS, MeOH-
Et2O (3:2), 23 °C, 0.5 h, 93%; (4) (COCl)2, DMSO, Et3N, CH2Cl2,
-78 °C, 85%; (e) (1) ((E)-4-tert-butyldimethylsiloxy)-2-buten-2-
yl)lithium, THF, -78 °C, 20 min; (2) (COCl)2, DMSO, Et3N, CH2Cl2,
-78 °C, 76% over two steps; (3) DDQ, CH2Cl2-pH 7 buffer (10:1),
23 °C, 40 min, 90%; (f) (1) CH2dCHOEt, PPTS, CH2Cl2, 23 °C, 6 h,
95%; (2) Ph3PdCH2, pentane, 0 °C, 3 h, 76%; (3) PPTS, Et2O-MeOH
(3:1), 23 °C, 6 h; (4) TBSCl, Et3N, DMAP, CH2Cl2, 23 °C, 2.5 h, 86%
over two steps; (g) (1) SO3‚pyr, pyr, 23 °C, 12 h, 96%; (2) HF-pyr,
MeOH, 23 °C, 20 days, 79%.
were particularly distinct in the chemical shifts of the protons
in the bridging area between the two remote sides.27 This being
(26) We thank Professor Murata for information on the revised assign-
ment of the carbon chemical shift for the C.12 methyl group (δ 18.0 ppm).
The chemical shift for the C.3 carbon of MTX6 appears to have been
misassigned.
(27) We chose a geminal dimethyl group at the C.12 position of model
C, in order to simplify the spectroscopic analysis. Interestingly, we observed
that the geminal dimethyl groups of 9 exhibited a small but distinct
1
diastereotopic property in its H NMR spectrum, which was the first sign
of the presence of some detectable residual communication between the
two halves.

7954 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
recognized, it became evident that the remote diastereomer 51
represents the stereochemistry of the corresponding moiety of
MTX.28
The experimental results given in this section provided the
answers to the two questions raised earlier. Although bridged
by only two methylene groups, the left-side and right-side
moieties could still be treated independently, yet small but
distinct communications between these two moieties were
detected by currently available spectroscopic means, which
allowed the assignment of the relative stereochemistry of the
C.1-C.15 portion of MTX.
2. C.35-C.39 Acyclic Portion. The C.35-C.39 portion
of MTX contained two unassigned chiral centers on the acyclic
backbone. It is important to note that, in order to establish the
complete relative stereochemistry of MTX, their configuration
relative not only to C.35 but also to C.39 must be determined.
By this means, the relative stereochemistry at the remote
positions, for example C.35 and C.39, can be established. We
decided to address these questions via (1) the selection of a
model properly representing this portion of MTX; (2) the
synthesis of all the possible diastereomers with respect to the
C.35, C.36, C.37, and C.39 chiral centers; (3) the comparison
of the NMR characteristics of each diastereomer with those of
MTX; and (4) the identification of which diastereomer represents
the spectroscopic properties of the corresponding part of MTX.
Unlike the C.1-C.15 moiety, the chiral centers present in this
region are in close proximity to each other so that each
diastereomer should exhibit spectroscopic characteristics dif-
fering from those of the others. We chose model E, which
contained the four chiral centers in question. This strategy called
for the synthesis of eight possible diastereomers.29
or 66, followed by dihydroxylation; (2) the acetylene 64 could
stereoselectively be synthesized by coupling the lactone 62 with
the acetylene anion generated from the dibromo olefin 63,
followed by silane reduction; (3) both 62 and 63 could be
prepared from the common synthetic intermediate 61; (4) the
Scheme 9
The synthetic plan of all of the eight diastereomers is outlined
in Scheme 9. Several comments are in order: (1) the four
diastereomers 72-75 could be prepared via the acetylene 64,
i.e., reduction of 64 selectively into the cis- or trans-olefin 65
(28) In order to eliminate the possibility that any electronic and/or steric
factor might give an unforeseen effect on their proton and carbon NMR
spectra, we synthesized two additional models, vii and viii, both of which
exhibited NMR characteristics very similar to those of 51, and vii is even
closer to those of MTX than 51. Below is depicted the difference in proton
(500 MHz, 1:1 C5D5N-CD3OD) and carbon (125 MHz, 1:1 C5D5N-CD3-
OD) chemical shifts between MTX and each of vii and viii. The x- and
y-axes represent carbon number and ∆δ (∆δ ) δMTX - δSyntheticModel in
ppm), respectively.
(29) Tachibana and co-workers recently reported the relative stereo-
chemistry of this portion of MTX.9 At the onset, they relied on the NMR
data to select one diastereomer, corresponding to 81, over the remaining
seven diastereomers, and they synthesized that diastereomer to conclude
the relative stereochemistry of this portion of MTX. It is interesting to note
that 74 and 75 are the C.39 and C.35 diastereomers of 78, respectively.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7955
Scheme 11a
remaining four diastereomers 76-79 should be available via
the same synthetic sequences but with the lactone 68, the
antipode of 62.
Scheme 10 outlines the synthesis of the lactone 86 from 82.30
The silane reduction of the ketol, formed via addition of
butenylmagnesium bromide to the lactone 82, was expected to
give the equatorial product 83;31 indeed, this reduction was
highly stereoselective, yielding the desired equatorial product
with a stereoselectivity greater than 20:1. The cyclization to
form the second ring was best achieved by sodium hydride
treatment of the pentol monomesylate 84. The C.33 deoxy-
genation was effected by the Barton procedure,32 and removal
of the hydroxymethyl group to form the lactone was carried
out in three steps.
a (a) (1) TBSCl, DMAP, Et3N, CH2Cl2, 23 °C, 3 days, 77%; (2)
1,1′-thiocarbonyldiimidazole, PhMe, 110 °C, 4 h, 95%; (3) 1,3-
dimethyl-2-phenyl-1,3,2-diazaphospholidine, 23 °C, 3 h, 81%; (4) 1%
HCl in 95% EtOH, 23 °C, 5 h, 96%; (5) TBDPSCl, imidazole, pyr, 23
°C, 11 h; (b) OsO4, NMO, t-BuOH-acetone-H2O (1:1:1), 23 °C, 48
h, 67% over two steps; (c) (1) BnBr, Ag2O, Et2O, 35 °C, 12 h; (2)
BnBr, n-Bu4NI (cat.), NaH, THF, 23 °C, 5 h; (3) 3% HCl in MeOH,
23 °C, 12 h, 72% over three steps; (d) (1) MsCl, Et3N, CH2Cl2, 0 °C,
15 min, 96%; (2) NaCN, DMSO, 90 °C, 1.5 h, 98%; (3) DIBAL, PhMe,
-78 °C, 10 min, 94%; (4) PhHgCBr3, PhH, 80 °C, 1.5 h, 87%.
Scheme 10a
Scheme 12a
a (a) (1) 3-Buten-1-ylmagnesium bromide, Et2O, -78 °C, 15 min,
91%; (2) Et3SiH, BF3‚OEt2, CH2Cl2-CH3CN (1:1), -20 °C, 15 min,
98%; (b) (1) O3, MeOH-CH2Cl2 (2:1); NaBH4, 88%; (2) MsCl, Et3N,
CH2Cl2, 0 °C, 15 min, 98%; (3) H2, Pd(OH)2 on C (cat.), MeOH, 23
°C, 4 h, 100%; (c) NaH, DMF, 23 °C, 12 h, 74%; (d) (1) p-
MeOPhCH(OMe)2, p-TsOH, DMF, 23 °C, vacuum, 12 h, 97%; (2)
NaH, CS2, MeI, THF, 23 °C, 30 min, 95%; (3) n-Bu3SnH, AIBN, PhMe,
110 °C, 1 h, 75%; (4) DIBAL, CH2Cl2, -78 °C, 30 min, then 0 °C, 1
h, 98%; (e) (1) Ph3P, imidazole, I2, PhH, 70 °C, 3 h, 98%; (2) DBU,
1,2-dichlorobenzene, 200 °C, 15 min;73 (3) O3, MeOH, -78 °C; DMS,
23 °C, 1 h, 32% over two steps.
a (a) (1) EtO2CCHdPPh3, PhH, 23 °C, 2 h, 71%; (2) H2, Lindlar
catalyst, EtOAc, 23 °C, 16 h, 100%; (3) DIBAL, CH2Cl2, 0 °C, 15
min, 96%; (4) MsCl, Et3N, CH2Cl2, 0 °C, 15 min, 99%; (5) H2, Pd(OH)2
on C (cat.), MeOH, 23 °C, 12 h, 100%; (b) NaH, DMF, 23 °C, 4 h,
92%; (c) p-MeOPhCH(OMe)2, HBF4, DMF, 23 °C, 12 h, 88%; (d) (1)
DIBAL, CH2Cl2, 0 °C, 1 h, ratio of products ) 1:1.3, 96%; (2) for the
minor isomer, i.e., C.33-OMPM, NaH, n-Bu4NI (cat.), BnBr, THF-
DMF (3:1), 23 °C, 2.5 h, 93%; CAN, CH3CN-H2O (9:1), 23 °C, 40
min, 80%; NaH, imidazole (cat.), CS2, MeI, THF, 90%; Bu3SnH, AIBN,
PhMe, 110 °C, 0.5 h, 63%; for the major isomer, i.e., C.34-OMPM:
NaH, imidazole (cat.), CS2, MeI, THF, 15 min, 93%; Bu3SnH, AIBN,
PhMe, 110 °C, 0.5 h, 61%; DDQ, CH2Cl2-H2O (10:1), 23 °C, 40 min,
83%; BnBr, NaH, n-Bu4NI (cat.), THF-DMF (3:1), 23 °C, 2.5 h, 76%;
(e) (1) THF-CF3CO2H-H2O (1:1:1), 65 °C, 26 h, 85%; (2) (COCl)2,
DMSO, Et3N, CH2Cl2, -78 °C, 76%.
We planned to use the same triol 61, an intermediate in the
synthesis of the lactone 86, for the synthesis of the dibromo
olefin 90 as well (Scheme 11). For this purpose, it was
necessary to invert the stereochemistry of both C.40 and C.41
alcoholic groups of 61, which was effectively achieved via
selective protection of the primary alcohol, olefin formation from
the remaining cis-diol,33 and then osmylation, to yield the desired
diol 88 as the major product.
The synthesis of the lactone 96, the antipode of 86, from
9134 is summarized in Scheme 12. Once again, the cyclization
to form the second ring was achieved with the tetraol monomes-
ylate 92. Although reductive cleavage of the anisylidene group35
in 94 derived from the diol 93 was not selective, both
regioisomers were usable for the synthesis of 96 by deoxygen-
ation at C.33 under the Barton conditions, hydrolysis of the
methyl glycoside, and oxidation of the resultant lactol to the
lactone.
Coupling of the lactone 96 with the acetylene anion generated
in situ from the dibromide 90,36 followed by silane reduction,
gave once again the desired, equatorial acetylene 97 exclusively
(Scheme 13). Partial reduction of the acetylene 97 (Lindlar
catalyst) to the cis-olefin 98, followed by osmylation, gave a
17:1 mixture of the two possible diols 99 and 100. On the basis
of the empirical rule,37 their stereochemistry was tentatively
assigned as indicated. The trans-olefin 101 was stereospecifi-
cally obtained from the acetylene 97 in three steps38 and then
subjected to osmylation, to yield a 6:1 mixture of the two
possible diols 102 and 103. Once again on the basis of the
(30) Dondoni, A.; Scherrmann, M.-C. J. Org. Chem. 1994, 59, 6404-
6412.
(31) Lewis, M. D.; Cha, J. K.; Kishi, Y. J. Am. Chem. Soc. 1982, 104,
4976-4978.
(32) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans.
1 1975, 1574-1585. For a review on this subject, for example, see: Hartwig,
W. Tetrahedron 1983, 39, 2609-2645.
(33) Corey, E. J.; Hopkins, P. B. Tetrahedron Lett. 1982, 23, 1979-
1982.
(36) Corey, E. J.; Fuchs, P. L. Tetrahedron Lett. 1972, 3769-3772.
(37) (a) Cha, J. K.; Christ, W. J.; Kishi, Y. Tetrahedron Lett. 1983, 24,
3943-3946. (b) Christ, W. J.; Cha, J. K.; Kishi, Y. Tetrahedron Lett. 1983,
24, 3947-3950.
(38) Attempted photochemically- or radically-induced direct cis-to-trans
isomerization gave no promising results.
(34) Barnes, J. C.; Brimacombe, J. S.; Kabir, A. K. M. S.; Weakley, T.
J. R. J. Chem. Soc., Perkin Trans. 1 1988, 3391-3397.
(35) Takano, S.; Akiyama, M.; Sato, S.; Ogasawara, K. Chem. Lett. 1983,
1593-1596.

7956 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Scheme 13a
Scheme 14a
a (a) (1) TBSOTf, Et3N, CH2Cl2, 23 °C, 68%; (2) (COCl)2, DMSO,
Et3N, CH2Cl2, -78 °C, 99%; (3) H2, Pd(OH)2 on C (cat.), EtOAc, 23
°C, 0.5 h, 98%.
Scheme 15a
a (a) (1) n-BuLi, 90, THF, -78 °C, 1 h, then 0 °C, 10 min; 96, -78
°C, 0.5 h, then 23 °C, 95%; (2) Et3SiH, BF3‚OEt2, CH2Cl2-CH3CN
(1:1), -25 °C, 0.5 h, 83%; (b) H2, Lindlar catalyst, EtOAc, 50 min,
61%; (c) OsO4, NMO, t-BuOH-acetone-H2O (1:1:1), 23 °C, 12 h,
91%; (d) (1) Ph3SnH, Et3B, PhMe, 23 °C, 3 h, 88%;74 (2) NIS, THF,
23 °C, 1 h, 89%;75 (3) n-BuLi, THF, -78 °C, 10 min, 63%;76 (e) OsO4,
NMO, t-BuOH-acetone-H2O (1:1:1), 23 °C, 4 days, 96%; (f) H2,
Pd(OH)2 on C (cat.), MeOH, 23 °C, 12 h, 100%.
empirical rule, their stereochemistry was tentatively assigned
as indicated.
Although the empirical rule is very reliable for predicting
the stereochemical outcome of the osmylation of allylic alcohols
and their derivatives,37 it was desirable to have further experi-
mental support. Thus, the two diastereomers 102 and 103 were
converted into the six-membered ketols 104 and 105, respec-
tively (Scheme 14). The vicinal coupling constants observed
for the C.37 and C.38 protons indeed verified the stereochemical
assignment based on the empirical rule.
Following exactly the same sequence of reactions summarized
in Scheme 13 but with the enantiomeric lactone 86, the
remaining four diastereomers 108, 109, 111, and 112 were
obtained, and their stereochemistry was assigned as indicated
(Scheme 15).
The C.40 alcohol of MTX exists as a sulfate, which must
affect the NMR characteristics of the protons and carbons in
this region of the molecule. However, the degree of the effects,
particularly on the protons and carbons at C.35 through C.38,
may be so small that some meaningful information is expected
to be gained in comparison of the NMR characteristics of the
eight diastereomeric polyols 72-79 derived from these diaster-
eomers with those of MTX. Thus, all of the eight diastereomers
a Follow steps a-f of Scheme 13.
were diastereomers with respect to the C.35 and C.39 positions.29
On the basis of these experiments, we were confident in
concluding that the stereochemistry of 78 represents the
stereochemistry of the corresponding part of MTX. Neverthe-
less, further experimental support to verify this conclusion would
be desirable, and we decided to examine the spectroscopic
properties of the C.40 sulfates corresponding to the 78 and 75
diastereomers.
Scheme 16 summarizes the synthesis of the C.40 sulfates 80
and 81 from 112 and 102, respectively. The key feature in this
transformation was the regiospecific reductive cleavage of the
C.40-C.41 benzylidene group with DIBAL,35 to yield exclu-
1
were deprotected and then subjected to the H and 13C NMR
studies. The NMR characteristics observed for the C.34 through
C.39 protons and carbons of each diastereomer were compared
with those of the corresponding protons and carbons of MTX
(Charts 7 and 8). This exercise indeed demonstrated that (1)
the diastereomer 78 was the best match with the relevant portion
of the natural product and (2) the diastereomer 75 was not as
good as 78 but close. It is interesting to note that 78 and 75

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7957
Scheme 16a
Chart 7. Difference in Proton Chemical Shifts between
MTX and Each of 72-79 (500 MHz, 1:1 C5D5N-CD3OD)
a (a) (1) Ac2O, pyr, 65 °C, 2 h, 92%; (2) H2, Pd(OH)2 on C (cat.),
EtOAc, 23 °C, 5 h, 100%; (3) DDQ, p-MeOPhCH2OMe, CH2Cl2, 23
°C, 2 h, 93%; (b) (1) MeONa, MeOH, 23 °C, 1.5 h, 95%; (2) NaH,
BnBr, THF-DMF (3:1), 23 °C, 3 h, 95%; (3) DIBAL, CH2Cl2, -78
°C, 5 min, then 0 °C, 20 min, 85%; (c) (1) SO3‚pyr, pyr, 23 °C, 12 h,
84%; (2) H2, Pd(OH)2 on C, MeOH, 23 °C, 12 h, 100%.
Chart 8. Difference in Carbon Chemical Shifts between
MTX and Each of 72-79 (125 MHz, 1:1 C5D5N-CD3OD)
Chart 9. Difference in Proton (500 MHz, 1:1
C5D5N-CD3OD) and Carbon (125 MHz, 1:1
C5D5N-CD3OD) Chemical Shifts between MTX and Each
of 80 and 81
same approach as the one used for the C.35-C.39 portion. We
chose model F, which possessed the relevant chiral centers at
C.63, C.64, C.66, and C.68. This strategy required the synthesis
of all eight possible diastereomers and the comparison of their
NMR characteristics with those of the natural product.39
sively the C.40 alcohol 114. The sulfates 80 and 81 were
isolated as their sodium salts and then subjected to NMR studies.
1
Chart 9 shows the H and 13C NMR characteristics in the
C.34-C.40 region of these diastereomers in comparison with
MTX, clearly supporting the previous conclusion that the C.35
through C.39 stereochemistry of 81, corresponding to 78,
represents the relative stereochemistry of the C.35-C.39 portion
of MTX.
3. C.63-C.68 Acyclic Portion. The relative stereochemical
assignment of the C.63-C.68 portion was addressed with the
(39) Tachibana and co-workers recently reported the relative stereo-
chemistry of the C.63-C.68 portion of MTX.9 They opted to synthesize
the four diastereomers sharing the same relative stereochemistry at the C.63
and C.64 positions. They relied on the NMR experiments in order to exclude
the diastereomers bearing the opposite relative stereochemistry at the C.63
and C.64 positions. It is interesting to note that 140 and 141 were the C.68
and C.66 diastereomers of 117, respectively.

7958 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Scheme 18a
Scheme 17 outlines a synthetic plan for all eight diastereo-
mers. The 1,3-diol group in 115-118 may be introduced via
reduction of â-hydroxy ketones 119. The â-hydroxy ketones
119 could be obtained via a 1,3-dipolar cycloaddition process,
i.e., 121 + 122 f 120 f 119, and the proposed 1,3-dipolar
cycloaddition was expected to yield the two possible C.64
diastereomers. Both the olefin and the nitrile oxide required
for the cycloaddition reaction could be obtained from a common
intermediate 123. The four diastereomers 115-118 should be
secured by these experiments, and the remaining four diaster-
eomers 138-141 (Scheme 21) should be available by repeating
the same sequence of synthesis with the antipode of 121.
Interestingly, we noticed that the common intermediate 123
shares its structural feature with that found in the C.28-C.37
portion of halichondrins, suggesting that its ring system could
a (a) (1) K2CO3, MeOH, 23 °C, 12 h; (2) PhCH(OCH3)2, CSA, DMF,
50 °C, 2 h; 81% over two steps; (3) TBSCl, AgNO3, pyr, DMF, 23 °C,
4 h, 90%; (4) OsO4, NMO, t-BuOH-THF-H2O (3:1:1); (5) NaIO4,
MeOH; 80% over two steps; (b) (E)-ICHdCHCO2CH3, 1% NiCl2/
CrCl2, THF-DMSO (trace), 23 °C, ratio of 126a:126b ) 2.5:1, 80%;
(c) (1) p-nitrobenzoic acid, DEAD, Ph3P, Et2O-PhMe (2:1), 23 °C;
(2) K2CO3, MeOH, 86% over two steps; (d) (1) Ac2O, pyr, 91%; (2)
n-Bu4NF-imidazole (4:1), THF; (3) DBU, CH2Cl2; 77% over two steps;
(e) (1) TBSOTf, i-Pr2NEt, CH2Cl2, 83%; (2) DIBAL, Et2O, -78 °C,
10 min, 23 °C, 1 h, 90%; (3) DMTrCl, pyr, DMAP, DMF, 0 °C, 10
min, 23 °C, 3 h; TBSCl, AgNO3, pyr, DMF; PPTS, MeOH, CH2Cl2;
88%, three steps; (4) CH3I, NaH, THF, 23 °C, 89%; (5) n-Bu4NF, THF,
23 °C, 12 h; (6) BnBr, n-Bu4NI (cat.), NaH, DMF, 23 °C, 24 h; 82%
over two steps; (f) (1) DIBAL, CH2Cl2, -78 °C, 30 min, 23 °C, 12 h,
83%; (2) I2, Ph3P, imidazole, PhH, 23 °C, 1 h; (3) n-Bu4NCN, DMF,
55 °C, 4 h; 82% over two steps; (4) DIBAL, PhMe, -78 °C, 2 h, then
1 N HCl, 23 °C, 2 h; (5) H2NOH‚HCl, NaOAc, EtOH, 23 °C, 1 h;
76% over two steps.
Scheme 17
be constructed by adopting the method developed for the
synthesis of halichondrins.40 Indeed, this was effectively carried
out (Scheme 18). The axial C-allylglucose tetraacetate 12441
was used as the starting material, and the C.74-C.75 bond was
formed via the Ni(II)/Cr(II)-mediated coupling42 of the aldehyde
125 with methyl (E)-â-iodoacrylate, to yield a 5:2 mixture of
two possible diastereomers. These diastereomers were inter-
convertible via the Mitsunobu reaction.43 On the basis of the
previous examples,40 the stereochemistry of the major product
was tentatively assigned. The bicyclic ring system was ste-
reospecifically constructed via the Michael reaction.44 The 1H
NMR analysis of the Michael adduct acetates 127a,b, derived
(40) (a) Aicher, T. D.; Buszek, K. R.; Fang, F. G.; Forsyth, C. J.; Jung,
S. H.; Kishi, Y.; Scola, P. M. Tetrahedron Lett. 1992, 33, 1549-1552. (b)
Aicher, T. D.; Buszek, K. R.; Fang, F. G.; Forsyth, C. J.; Jung, S. H.; Kishi,
Y.; Matelich, M. C.; Scola, P. M.; Spero, D. M.; Yoon, S. K. J. Am. Chem.
Soc. 1992, 114, 3162-3164 and references cited therein.
(41) Giannis, A.; Sandhoff, K. Tetrahedron Lett. 1985, 26, 1479-1482.
(42) (a) Jin, H.; Uenishi, J.-i.; Christ, W. J.; Kishi, Y. J. Am. Chem.
Soc. 1986, 108, 5644-5646. (b) Takai, K.; Kimura, K.; Kuroda, T.; Hiyama,
T.; Nozaki, H. Tetrahedron Lett. 1983, 24, 5281-5284.
(43) Mitsunobu, O. Synthesis 1981, 1-28.
(44) Protection of the allylic hydroxy group was necessary to have a
clean and efficient cyclization.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7959
Scheme 20a
Scheme 19a
a (a) (1) Benzylation and silylallylation, see ref 31; (2) MeI, NaH,
THF, 23 °C; (3) Na, NH3, THF, -78 °C; (4) TBSOTf, i-Pr2NEt,
CH2Cl2, 23 °C; (b) (1) O3, CH2Cl2-MeOH (3:1), -78 °C; Ph3P, 98%;
(2) (E)-ICHdCHCO2CH3, 1% NiCl2/CrCl2, THF, 23 °C, ratio of 132a:
132b ) 1:1, 90%; (c) (1) p-nitrobenzoic acid, DEAD, Ph3P, PhMe, 23
°C, 4 h; (2) K2CO3, MeOH, 99%; (d) (1) Ac2O, pyr; (2) HF‚pyr,
CH3CN, 0 °C, 4 h, 67%; (3) DBU, CH2Cl2, 20 h, 81%; (4) TBSCl,
AgNO3, imidazole, DMF, 89%; (e) (1) DIBAL, Et2O, -78 °C, 10 min,
23 °C, 1 h, 95%; (2) PvCl, pyr, 0 °C, 92%; TBSCl, imidazole, DMF,
23 °C, 2 h; DIBAL, Et2O, -78 °C, then 10 min, 23 °C, 1 h, 100%
over two steps; (3) I2, Ph3P, imidazole, PhH, 95%; (PhSe)2, NaBH4,
EtOH, 23 °C; NaIO4;77 i-Pr2NH, 81% over two steps; (4) n-Bu4NF,
THF; BnBr, n-Bu4NI (cat.), NaH, DMF, 23 °C, 24 h, 91% over two
steps.
a (a) (1) NCS, CHCl3, pyr (trace), 23 °C, 1 h; (2) 121, i-Pr2NEt,
CHCl3, 0 °C, 15 min, then 60 °C, 24 h; 50-60%; (b) Mo(CO)6,
CH3CN-H2O (8:1), 80 °C, 70-80%; (c) NaBH(OAc)3, AcOH-
CH3CN (1:1), 23 °C, 10 h, 80-90%; (d) Pd(OH)2 on C (cat.), H2 (1
atm), MeOH-EtOAc (1:1), 90%.
from the major and minor diastereomers 126a,b obtained in the
Ni(II)/Cr(II)-mediated coupling reaction, confirmed the tentative
stereochemical assignment; most critically, the coupling constant
between the C.74-H and C.75-H was found to be 10 Hz for
127a, whereas it was approximately 2 Hz for 127b. With this
evidence for the stereochemical assignment, the major Michael
product 127a was then transformed into the aldoxime 129.
As pointed out, the olefin partner 121 for the 1,3-dipolar
cycloaddition could be prepared from some intermediates in
Scheme 18. For technical reasons, however, it was decided that
the C.55 alcohol should be converted to its methyl ether at the
very beginning of the synthesis. Thus, using the same chemistry
as the one shown in Scheme 18, the axial C-allylglucoside
derivative 131, readily available from 1,6-anhydro-D-glucose
(130),31,45 was converted into the olefin 121 (Scheme 19A).
Similarly, its antipode 135 was obtained from 1,6-anhydro-L-
glucose (134; Scheme 19B).46
subjected to the three-step sequence of transformation, i.e., (1)
reductive cleavage of the N-O bond with molybdenum hexa-
carbonyl,49 to yield â-hydroxy ketones 119a,b; (2) reduction
of the resultant â-hydroxy ketone 119a with sodium triacetoxy-
borohydride,50 to give a ca. 4:1 ratio of the two possible diols,
which were separated by silica gel column chromatography; and
(3) deprotection of the benzyl groups, to furnish the polyols
115 and 116. From the second cycloaddition product 119b,
the third and fourth polyols 117 and 118 were similarly obtained
(Scheme 20). Exactly the same chemistry was applied for the
olefin 135 derived from 1,6-anhydro-L-glucose (134), to furnish
the remaining four diastereomers 138-141 (Scheme 21).
All of the eight diastereomers were then subjected to the 1H
and 13C NMR study, and the chemical shifts observed for the
C.62 through C.69 protons and carbons of each diastereomer
were compared with those corresponding to MTX (Charts 10
(47) With TBS groups as the hydroxy protecting groups, the 1,3-dipolar
cycloaddition failed to give acceptable yields. Change of the TBS groups
into benzyl groups resulted in improved and reproducible yields.
(48) For a review on the nitrile oxide coupling reaction, for example,
see: Torssell, K. B. G. Nitrile Oxides, Nitrones, and Nitronates in Organic
Synthesis; VCH: New York; 1987; Chapter 2.
The 1,3-dipolar cycloaddition between 121 and 129 took place
smoothly at 60 °C, to yield a 3:2 mixture of the expected adducts
120a,b in 50-60% combined yield.47,48 The two adducts were
separated by silica gel column chromatography and then
(45) Coleman, G. H. In Methods in Carbohydrate Chemistry; Whistler,
R. L., Wolfrom, M. L., Eds.; Academic Press: New York and London,
1963; Vol. II, pp 397-399.
(49) Baraldi, P. G.; Barco, A.; Benetti, S.; Manfredini, S.; Simoni, D.
Synthesis 1987, 276-278.
(50) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.
1988, 110, 3560-3578.
(46) 134 was prepared from commercially available L-glucose.45

7960 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Chart 11. Difference in Carbon Chemical Shifts between
Scheme 21a
MTX and Each of 115-118 and 138-141 (125 MHz, 1:1
C5D5N-CD3OD)
a Follow steps a-d of Scheme 20.
Chart 10. Difference in Proton Chemical Shifts between
MTX and Each of 115-118 and 138-141 (500 MHz, 1:1
C5D5N-CD3OD)
Scheme 22a
and 11). This exercise clearly demonstrated that (1) each
diastereomer exhibited distinct spectroscopic characteristics
differing from the others and (2) only the diastereomer 117
displayed NMR characteristics that were virtually identical to
those of this portion of MTX.
Having demonstrated that the diastereomer 117 represents
the relative stereochemistry of this portion of MTX, we then
established its stereochemistry from the following experiments.
Since the diastereomer 117 originated from the cycloaddition
product between 121 and 129, the stereochemistry at the C.63
and C.68 positions was firmly known, but the stereochemistry
at the C.64 and C.66 positions was not established at this stage.
In order to establish the C.64 stereochemistry, the diastereomers
119a,b were transformed into the cyclic hemiketal peracetates
142a,b, respectively (Scheme 22). The C.64 proton signal was
observed as a doublet of doublets of doublets (J ) 11.3, 9.5,
a (a) H2, Pd(OH)2 on C (cat.), EtOAc-MeOH (1:1); (b) CHCl3-
MeOH (5:1), PPTS, 23 °C; (c) Ac2O, pyr.
and 5.3 Hz) at 5.56 ppm for the isomer 142b, derived from
119b, whereas a broad doublet of doublets was observed (J )
5.5 and 2.6 Hz) at 5.39 ppm for the isomer 142a, derived from
119a. Coupled with the fact that the newly formed six-
membered ring clearly adopted a chair conformation (for
example, J62,63 ) 9.8 Hz for 142a and J62,63 ) 9.6 Hz for 142b),
the values of these vicinal coupling constants established the
C.64 stereochemistry of 117 as depicted in Scheme 23.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7961
Scheme 23a
Scheme 24
a (a) MeC(OMe)2Me, PPTS, CH2Cl2, 23 °C.
The relative stereochemistry between C.64 and C.66 was then
determined from the following experiments. Both the diols
137a,b, obtained from the reduction of the â-hydroxy ketone
119b with sodium triacetoxyborohydride, were converted into
the acetonides 143a,b (Scheme 23). The vicinal coupling
constants J64,65(â) ) 9.8 Hz and J65(â),66 ) 6.1 Hz were observed
for the acetonide 143a derived from 137a, whereas the constants
J64,65(â) ) J65(â),66 ) 12.0 Hz were observed for the acetonide
143b derived from 137b. These vicinal proton coupling
constants confirmed the stereochemistry at C.64 and C.66,
thereby establishing that 117 (Scheme 23) represents the relative
stereochemistry of this acyclic portion of MTX.
4. C.134-C.142 Acyclic Portion. The C.134-C.142 por-
tion of MTX contained five chiral centers in question, giving
rise to 16 possible diastereomers. We decided to adopt the same
approach as the one used for the C.35-C.39 and the C.63-
C.68 portions, and we chose G as the model for this purpose.
The angular methyl substituents were substituted by hydrogens
to simplify the synthesis with the belief that this substitution
would not greatly alter the structural characteristics of the acyclic
chain.
of the acetylene anion generated from the dibromo olefin 144
with the lactone 145, followed by silane reduction, is expected
to yield stereoselectively the equatorial acetylene 146, which
could, in turn, be reduced to either the trans- or cis-olefin 147
or 148. Osmylation of these olefins should give the four diols
149-152. Application of this synthesis to the remaining three
acetylene equivalents 153-155 should result in the remaining
12 diastereomeric diols.
The synthesis of the lactone 145 began with methyl D-
glucopyranoside derivative 15651 (Scheme 25). The C.129
alcohol was removed under the Barton conditions, and the
equatorial C-glycosidation was stereoselectively accomplished
by addition of allylmagnesium bromide to the lactone 157,
followed by silane reduction.
Scheme 26 summarizes the synthesis of anti- and syn-
dimethyl dibromo olefins 161 and 164. The relative stereo-
chemistry of the C.138 and C.139 methyl groups was set by
the Diels-Alder reaction. It should be noted that these
syntheses were purposely carried out in a racemic form so that
all of the four required diastereomers of dibromo olefins were
obtained in only two syntheses.
The anion, generated from the racemic anti-dimethyl dibromo
olefin 161, was coupled with the lactone 145, followed by
triethylsilane reduction in the presence of BF3‚Et2O. The NMR
spectra of the crude reaction mixture showed no detectable
amount of the axial products, but the two equatorial anti-
Scheme 24 outlines a synthetic plan for all 16 diastereomers.
The proposed synthesis is convergent, in which the problem of
preparing these diastereomers was first reduced to the prepara-
tion of four acetylene intermediates. For example, the coupling
(51) Koto, S.; Takebe, Y.; Zen, S. Bull. Chem. Soc. Jpn. 1972, 45, 291-
293.

7962 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Scheme 25a
Scheme 28a
a (a) (1) (i) NaH, CS2, MeI, THF, 23 °C; (ii) Bu3SnH, AIBN, PhMe,
110 °C; 90% over two steps; (2) 1N HCl, AcOH, 60 °C, 70%; (3)
(COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 60%; (b) (1) allylmagnesium
bromide, THF, -78 °C; (2) Et3SiH, BF3‚OEt2, CH2Cl2, -20 °C; 55%
over two steps; (c) (1) (i) catecholborane, (PPh3)3RhCl, THF, 23 °C;
(ii) 30% H2O2, 3 N NaOH, 23 °C, 90% over two steps; (2) Jones
oxidation, 80%; (3) H2, Pd(OH)2 on C, MeOH, 23 °C, 85%; (4)
PhCH(OMe)2, p-TsOH, DMF, in vacuum at 50 °C, 90%; (5) CSA,
PhH, 80 °C, 95%.
Scheme 26a
a (a) (1) LAH, THF, 23 °C, 85%; (2) p-TsCl, pyr, 23 °C, 90%;78 (3)
(i) LAH, THF, 65 °C; (ii) O3, CH2Cl2, -78 °C; (iii) NaBH4, THF-
CH2Cl2-MeOH, 23 °C, 30% over three steps; (b) (1) TBDPSCl, pyr,
DMF, 23 °C, 33%; (2) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 80%;
(3) PPh3, CBr4, 0 °C, 95%.
a (a) (1) Ph3SnH, BEt3, PhMe, 23 °C; (2) NIS, THF, 23 °C; (3)
n-BuLi, THF, -78 °C, 75% over three steps; (b) H2, Lindlar catalyst,
EtOAc, 23 °C, 95%; (c) OsO4, NMO, H2O-acetone-t-BuOH (1:1:1),
23 °C, 2:1.
under Lindlar conditions to the cis-olefins 172a,b, which were
subjected to osmylation, followed by chromatographic separa-
tion, to furnish the four possible erythro diols 173-176 (Scheme
28).
Scheme 27a
Using the same methods, the four threo diols 178-181 and
the four erythro diols 183-186 were obtained from the two
equatorial syn-dimethyl acetylenes 166a,b (Scheme 29).
These diastereomers were subjected to a sequence of reactions
to furnish all of the 16 models 187-202 (Scheme 30). As
1
before, their H and 13C NMR characteristics were examined
in reference to those of the natural product (Charts 12 and 13).
This exercise demonstrated that (1) each diastereomer once again
exhibited distinct spectroscopic characteristics differing from
those of the others; (2) only one diastereomer, 187, exhibited
spectroscopic characteristics that were virtually identical to those
of the C.134-C.142 portion of MTX; and (3) this matching
diastereomer was found to be one of the four possible diaster-
eomers belonging to the syn-diol/anti-dimethyl series, i.e., the
diastereomers originating from 168-171.
The relative stereochemistry of the matching diastereomer
was deduced from the experiments shown in Scheme 31. The
methyl group at C.3 of (S)-citronellal (203) was transformed
into the C.164 methyl group of 167a, establishing the relative
stereochemistry at C.134, C.138, and C.139. Osmylation of
167a yielded a 2:1 mixture of the two possible threo diols, with
the matching diastereomer being the minor product 187. On
the basis of the empirical rule,37 the stereochemistry of the
minor, matching diol was assigned to be 187. This assignment
a (a) (1) n-BuLi, THF, -78 °C; (2) Et3SiH, BF3‚OEt2, CH2Cl2, -25
°C; 70% over two steps.
dimethyl acetylenes 165a,b were chromatographically insepa-
rable at this stage. Similarly, the combination of 145 and 164
furnished exclusively the two equatorial syn-dimethyl acetylenes
166a,b as an inseparable mixture (Scheme 27).
The anti-dimethyl acetylene mixture 165a,b was converted
into the trans-olefins 167a,b in three steps, which were then
subjected to osmylation, to furnish the four possible threo diols
168-171. Fortunately, these four diastereomers were easily
separable by silica gel column chromatography at this stage.
The anti-dimethyl acetylene mixture 165a,b was also reduced

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7963
Scheme 29a
Chart 12. Difference in Proton Chemical Shifts between
MTX and Each of 187-202 (500 MHz, 1:1
C5D5N-CD3OD)
a The structures of 177a,b and 182a,b correspond to those of 167a,b
and 172a,b in Scheme 28, respectively, except that the C.138
configuration is opposite in this series. Follow steps a-c of Scheme
28.
Scheme 30a
a (a) (1) MeC(OMe)2Me, PPTS, DMF, 23 °C, 95%; (2) n-Bu4NF,
THF, 23 °C, 95%; (3) (COCl)2, DMSO, Et3N, CH2Cl2; (4) n-BuLi,
Ph3PCH3Br, THF, 0 °C, 77% over two steps; (5) aqueous AcOH, 50
°C, 99%.
was further supported from the results on asymmetric osmylation
using both AD-mix R and AD-mix â reagents;52 the diol 187
was formed as the major product with AD-mix R, whereas the
diol 188 was the exclusive product with AD-mix â. These
results were consistent with the principle of double stereo-
differentiation,18 i.e., the precedents known for asymmetric
induction with these AD-mix reagents52 coupled with the
stereochemical outcome predicted by the empirical rule.37 Thus,
the relative stereochemistry of this diastereomer, and conse-
quently this portion of MTX, was established as 187.
5. Complete Relative Structure. MTX contains 32 rings,
all of which exist as fused-ring systems. Using primarily the
NMR methods, Yasumoto and co-workers have successfully
determined the relative stereochemistry of these fused-ring
portions. At three places, the fused-polycyclic systems are
(52) For a recent review, see: Kolb, H. C.; VanNieuwenhze, M. S.;
Sharpless, K. B. Chem. ReV. 1994, 94, 2483-2547.
connected directly via a single carbon-carbon bond, i.e., the

7964 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Chart 13. Difference in Carbon Chemical Shifts between
MTX and Each of 187-202 (125 MHz, 1:1
C5D5N-CD3OD)
Scheme 31a
a (a) (1) NaBH4, MeOH, 23 °C; (2) TBDPSCl, pyr, DMF, 23 °C;
(3) O3, NaOH, MeOH-CH2Cl2, -78 °C, 50% over three steps; (4) (i)
LDA, PhSeBr, THF, -78 °C; (ii) 30% H2O2, CH2Cl2, 0 °C, 59% over
two steps; (b) (1) MeLi, CuI, THF, BF3‚OEt2, -78 °C to 0 °C (3:2
anti/syn);79 (2) LAH, THF, 23 °C, 51% over two steps; (3) (R)-(-)-
1-(1-naphthyl)ethyl isocyanate, Et3N, PhMe, 60 °C, 82%; (4) LAH,
THF, 50 °C, 51%; (5) (COCl)2, DMSO, Et3N, CH2Cl2, -78 °C, 79%;
(6) PPh3, CBr4, 0 °C, 95%; (c) (1) n-BuLi, 145, -78 °C; (2) Et3SiH,
BF3‚OEt2, CH2Cl2, -25 °C; 70% over two steps; (3) Ph3SnH, BEt3,
PhMe, 23 °C; (4) NIS, THF, 23 °C; (5) n-BuLi, THF, -78 °C, 70%
over three steps; (d) (1) AD-mix R, H2O-t-BuOMe-t-BuOH (1:1:1),
23 °C, 53%; (2) MeC(OMe)2Me, PPTS, DMF, 23 °C, 90%; (3)
n-Bu4NF, THF, 23 °C, 95%; (4) (COCl)2, DMSO, Et3N, CH2Cl2, -78
°C; (5) n-BuLi, Ph3PCH3Br, THF, 0 °C, 77% over two steps; (6)
aqueous TFA-MeOH-H2O (1:1:1), 50 °C, 99%.
these ring junctures and have proposed the structure 1A as the
partial relative stereostructure of MTX.53 Coupled with this
information, the present work allows the assignment of the
structure 1B (Figure 3) as the complete relative stereochemistry
of MTX.
(53) The relative stereochemistry of the K/L- and O/P-ring junctures has
3
been determined on the basis of the NOE and JH,H data with the aid of
molecular mechanics calculations (MM2). However, the presence of the
3J
C.155 methyl group precludes the use of the
data between the ring-
H,H
juncture protons for elucidating the relative stereochemistry at C.99-C.100.
Thus, their conclusion on the relative stereochemistry of the V/W-ring
juncture was less sound than that on the O/P- and V/W-ring junctures. The
experimental data given in ref 6c and its supplementary material do not
necessarily exclude the alternative relative stereochemistry, cf. the partial
structure x. In principle, the two possible stereostructures ix and x should
be distinguishable from the patterns of the NOE connectivity among the
protons at C.98(R), C.98(â), C.99, C.101, and C.155. However, because of
many overlapping peaks present in the spectrum given in the supplementary
material, this distinction was not obvious. It should be noted that the relative
stereochemistry of the alternative possibility x coincides with that of the
K/L- and O/P-ring junctures. If the alternative possibility is the case, the
stereochemistry of MTX is represented by the structure bearing the opposite
relative stereochemistry at C.5 through C.99 of 1B and 1C.
K/L-, O/P-, and V/W-ring junctures. Relying on the NOE and
3JH,H data with the aid of molecular mechanics calculations
(MM2), they have also studied the relative stereochemistry of

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7965
Figure 3. 1B: Complete relative structure of maitotoxin.
Figure 4. 1C.
III. Absolute Stereochemistry
and ciguatoxin.56,57 That suggested absolute configuration of
gambiertoxin 4B/ciguatoxin matches that of brevetoxin B,
assuming that the C.1-ring A moiety of gambiertoxin 4B/
ciguatoxin and the C.42-ring K moiety of brevetoxin B
represent the head (or tail) in a biogenetic sense. The biogenetic
structural correlation of MTX with gambiertoxin 4B/ciguatoxin
is less obscure than that with brevetoxin B; focusing on a vicinal
dimethyl group present in both natural products, the C.142-
ring F′ moiety of MTX appears to correspond to the ring M/L
moiety of gambiertoxin 4B/ciguatoxin. On the basis of these
considerations, the more likely absolute structure of MTX then
seems to be 1C (Figure 4), the antipode of 1B.
Obviously, MTX is structurally similar to brevetoxins,
ciguatoxins, and related marine natural products.1a It is likely
that these natural products share the same absolute configura-
tion.54 On this assumption, the absolute stereochemistry of
MTX may be suggested, but there are two obstacles to consider.
First, among these natural products, brevetoxin B (205) is the
only member whose absolute stereochemistry has been defini-
tively established.55 Second, the structural correlation of
brevetoxin B and MTX is not trivial, particularly assigning, in
a biogenetic sense, the head and tail to these natural products.
However, focusing on the structural resemblance between the
ring E′-Y portion of MTX to the ring K-D portion of
brevetoxin B, one may suggest that the absolute configuration
of MTX is more likely 1B. However, the work by Hirama,
Yasumoto, and co-workers should also be noted. On the basis
of the CD spectrum of the synthetic AB ring model, they
suggested the absolute configuration of gambiertoxin 4B (206)
IV. Preferred Solution Conformation
As noted, MTX contains 32 rings, all of which exist as fused-
ring systems. Except for two, they are trans-fused, and they
are expected to be relatively rigid conformationally, which is
indeed demonstrated through the extensive NMR studies by
Yasumoto and co-workers.6 At three places, two of these fused-
polycyclic portions are connected directly via a single carbon-
carbon bond, i.e., the K/L-, O/P-, and V/W-ring junctures.
Yasumoto and co-workers have also studied the preferred
(54) It should be noted that ciguatoxins and MTX are known to originate
from Gambierdiscus toxicus, whereas brevetoxins are from Gymnodinium
breVe.
(55) Lin, Y.-Y.; Risk, M.; Ray, S. M.; Van Engen, D.; Clardy, J.; Golik,
J.; James, J. C.; Nakanishi, K. J. Am. Chem. Soc. 1981, 103, 6773-6775.
The absolute configuration of brevetoxin B was determined by using the
dibenzoate chirality method on a di-p-bromobenzoate derivative. The recent
total synthesis of brevetoxin B by Nicolaou (Nicolaou, K. C.; Rutjes, F. P.
J. T.; Theodorakis, E. A.; Tiebes, J.; Sato, M.; Untersteller, E. J. Am. Chem.
Soc. 1995, 117, 10252-10263) confirmed the assignment.
(56) Suzuki, T.; Sato, O.; Hirama, M.; Yamamoto, Y.; Murata, M.;
Yasumoto, T.; Harada, N. Tetrahedron Lett. 1991, 32, 4505-4508.
(57) The original suggestion has recently been supported by further
experiments: a private communication from Professor Masahiro Hirama
at Tohoku University (March 1996).

7966 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
Figure 6. Preferred solution conformation of 117 (top), based on
selected vicinal proton coupling constants in Hz (bottom: 500 MHz;
1:1 C5D5N-CD3OD).
conformation, which is fully consistent with numerous examples
studied in this laboratory.59 Second, as seen from the confor-
mation drawn on a diamond lattice,59 the C.135 hydroxyl, C.138
methyl, and C.139 methyl do not cause steric destabilization
due to 1,3-diaxial-like interactions,59,60 whereas the C.136
hydroxyl group has a 1,3-diaxial-like relationship with the C.134
C-O bond. However, it should be noted that 1,3-diaxial-like
interactions involving two C-O bonds are much less significant
than ones involving one or two C-C bonds. In addition, in
this conformation, the C.136 hydroxy group and the ring oxygen
on the F′ ring are placed well to form a hydrogen bond, which
may provide some stabilization. Third, the backbone of this
acyclic portion predominantly adopts an extended conformation.
Fourth, as noted above, the NMR characteristics observed for
model 187 were virtually superimposable on those of the
relevant protons and carbons of the natural product, and the
preferred conformation of this portion of MTX should ad-
equately be represented by the preferred conformation of 187.
Analysis of the vicinal proton coupling constants observed
for model 117 (Figure 6) clearly demonstrates that the C.63-
C.68 portion of 117 and, consequently, of MTX preferentially
adopts a well-defined, extended conformation, 117A. The L/M-
and N/O-rings are dioxa-cis-decalins, for which two conforma-
tions bearing both rings in a chair form were possible. However,
the vicinal coupling constants of the ring protons, for example
J62,63 ) 10.5 Hz and J68,69 ) 9.4 Hz, demonstrate the preferred
conformation of these ring systems to be the one depicted, in
which the C.63 and C.68 substituents are equatorial.
conformation for these moieties by NMR methods, coupled with
force-field computation. It is worth noting that the K/L- and
O/P-ring systems are structurally similar to the case studied by
Still,58 and their conformational characteristics match each other
well.
In order to estimate the overall conformational properties of
MTX, the information regarding the remaining acyclic portions,
i.e., C.1-C.15, C.35-C.39, C.63-C.68, and C.134-C.142,
must be addressed. The vicinal proton coupling constants
observed for models 51, 81, 117, and 187 provide indispensable
conformational information. It should be restated that, as the
NMR characteristics observed for these models were virtually
superimposable on those of the relevant protons and carbons
of MTX itself, the preferred conformations found for these
models should adequately represent the preferred solution
conformation of MTX as well.
Figure 5 summarizes the vicinal proton spin coupling
constants observed for model 187, suggesting the preferred
conformation of this portion of MTX to be represented by 187A.
Several characteristics are worth noting. First, the C.134-C.135
C-glycosidic bond preferentially adopts the exo-anomeric
Similarly, the preferred solution conformation of the C.35-
C.39 portion of the model 81 and, consequently, of MTX is a
well-defined, extended conformation, 81A (Figure 7). Once
again, both C-glycosidic bonds, i.e., the C.35-C.36 and C.38-
C.39 bonds, preferentially adopt the exo-anomeric conforma-
tion,59 and there is no obvious steric destabilization due to 1,3-
diaxial-like interactions in this acyclic region. Intriguingly, this
preferred conformation makes the backbone of MTX turn in
the shape of the letter U, cf. the arrows indicating the direction
of the carbon backbone.
The C.1-C.15 portion of 51 and, consequently, of MTX is
conformationally less well-defined than the other acyclic
portions (Figure 8). Assuming an extended conformation for
the carbon backbone of 51 and placing the established relative
(59) (a) Wu, T. C.; Goekjian, P. G.; Kishi, Y. J. Org. Chem. 1987, 52,
4819-4823. (b) Wei, A.; Haudrechy, A.; Audin, C.; Jun, H.-S.; Haudrechy-
Bretel, N.; Kishi, Y. J. Org. Chem. 1995, 60, 2160-2169 and references
cited therein.
(60) The destabilization energy for 1,3-diaxial Me/Me, Me/OH, and OH/
OH groups on a cyclohexane ring have been estimated to 3.7, 2.7-1.9,
and 1.9 kcal/mol, respectively; see: Corey, E. J.; Feiner, N. F. J. Org. Chem.
1980, 45, 765-780 and references cited therein.
Figure 5. Preferred solution conformation of 187 (top), based on
selected vicinal proton coupling constants in Hz (bottom: 500 MHz;
1:1 C5D5N-CD3OD).
(58) Li, G.; Still, W. C. J. Am. Chem. Soc. 1993, 115, 3804-3805 and
references cited therein.

Complete RelatiVe Stereochemistry of Maitotoxin
J. Am. Chem. Soc., Vol. 118, No. 34, 1996 7967
Figure 7. Preferred solution conformation of 81 (top), based on
selected vicinal proton coupling constants in Hz (bottom: 500 MHz;
1:1 C5D5N-CD3OD).
stereochemistry on this extended backbone provide a clear
picture for conformational analysis, cf. the conformer 51A; the
-
C.7 Me/C.9 OSO3 groups and the C.12 Me/C.14 Me groups
are in 1,3-diaxial-like relationship and cause severe steric
destabilization.59,60 The vicinal proton coupling constants
observed for the C.7-C.9 and C.12-C.14 regions indeed
indicate that the carbon backbone of 51 does not adopt
preferentially the ideal extended conformation.
With respect to the C.7-C.9 region, there are two conformers,
51B and 51C, which are free from the steric destabilization due
to the 1,3-diaxial-like interactions, 51B resulting from rotation
of the C.7-C.8 bond of 51A, and 51C resulting from rotation
of the C.8-C.9 bond of 51A (Figure 8). The observed vicinal
proton coupling constants, J7,8 ) 5.7 Hz and J8,9 ) 5.1 Hz,61
may suggest the conformational properties of this region to be
described primarily by rapid interconversion between these two
conformers. Indeed, the NOESY experiments62 showed clearly
the cross peak between the C.6 and C.9 protons, which is
obvious in the conformer 51B but not in the conformer 51C,
as well as between the C.9 and C.7-methyl protons and the C.7
and C.10 protons, which are contrarily obvious in the conformer
51C but not in the conformer 51B.
The conformational analysis of the C.12-C.14 region can
be treated in a similar manner. The extended conformer 51A
suffers from steric destabilization due to the 1,3-diaxial-like
interactions, whereas the conformer 51D is free from such steric
destabilization and the conformer 51E is less sterically desta-
bilized than the extended conformer 51A.59,60 The vicinal proton
spin coupling constants, J12,13 ) 5.4 Hz and J13,14 ) 6.3 Hz,61
may indicate the conformational properties of this region to be
described primarily by rapid interconversion between these two
conformers, with more weight on 51D. Once again, the NOESY
experiments61 showed clearly the cross peaks between the C.12
and C.14-methyl protons and the C.11 and C.14 protons, cf.
the conformer 51D, as well as between the C.12 and C.15
protons and the C.14 and C.12-methyl protons, cf. the conformer
51E. It is interesting to note that the conformation found in
the X-ray analysis of 50 (Figure 2) corresponds to the conformer
51D.
Figure 8. Preferred solution conformation of 51: 51A, a hypothetical,
extended conformer; 51B,C: two preferred conformers with respect
to the C.7-C.9 portion and selected vicinal proton coupling constants
in hertz (500 MHz; 1:1 C5D5N-CD3OD); 51D,E, two preferred
conformers with respect to the C.12-C.14 portion and selected vicinal
proton coupling constants in hertz (500 MHz; 1:1 C5D5N-CD3OD).
Assembling the preferred solution conformations found for
the four acyclic portions allows us to suggest that the ap-
proximate global conformation of MTX is represented by the
shape of a hook, with the C.35-C.39 portion being the
curvature. As pointed out, MTX is conformationally rather
rigid, except for the C.7-C.9 and C.12-C.14 portions, which
are expected to provide conformational flexibility.
(61) The very similar vicinal proton coupling constants were recorded
on the smaller models 9 (J7,8 ) 6.1 Hz and J8,9 ) 5.4 Hz) and 35 (J12,13
)
V. Closing Comments
5.5 Hz and J13,14 ) 6.8).
(62) These NOESY experiments were conducted on 51, viii (footnote
28), 9, and 35. Interestingly, no definitive cross peaks which might be
expected for conformer 51A were detected.
It is remarkable and intriguing that the four small acyclic
models 51, 81, 117, and 187, even much smaller models 9 and

7968 J. Am. Chem. Soc., Vol. 118, No. 34, 1996
Zheng et al.
35, beautifully represent the structural characteristics of the
corresponding portions of MTX itself. In connection with this,
we should point out once again the fact that all eight diaster-
eomers of the left half of the AAL toxin backbone as well as
all four diastereomers of the right half exhibited differing and
distinct spectroscopic behavior from each other, and this trend
was also recognized on all the acyclic portions of MTX. This
implies that the structural properties of these substances are
inherent to the specific stereochemical arrangement of the small
substituents on the carbon backbone, and independent from the
rest of the molecules. In other words, each of these diastere-
omers has the capacity to install a unique structural characteristic
through a specific stereochemical arrangement of small sub-
stituents on the carbon backbone. Then, it is tempting to suggest
the possibility that fatty acids and related classes of compounds
may be able to carry specific information and serve as functional
materials in addition to structural materials.
(54 pages). See any current masthead page for ordering and
Internet access instructions.
Note Added in Proof. (a) On exchanging manuscripts with
the Tachibana/Yasumoto group (June 27, 1996), we learned that
they have reached the same conclusion by a different approach.
(b) According to a fax letter from Dr. Murata (July 1, 1996),
the C.3 carbon chemical shift was incorrectly reported (see ref
26); the correct chemical shift is 136.0 ppm, which matches
perfectly with the values observed for the remote diastereomers
51 (135.8 ppm) and 52 (135.8 ppm) and also for two additional
models vii (136.0 ppm) and viii (135.9 ppm) mentioned in ref
27.
(c) We have recently completed the synthesis and NMR
studies of the models xi and xii, representing the two possibilities
for the stereochemistry of the V/W-ring juncture (see ref 53).
As seen from the charts below (∆δ(ppm) ) δMTX - δSyntheticModel
1
in H (500 MHz) and 13C (125 MHz) chemical shifts in 1:1
C5D5N-CD3OD between MTX and xi and xii, respectively), it
is now evident that the original assignment, cf. ix in ref 53, of
the ring juncture is correct: Oinuma, H.; Cook, L. R.; Kishi,
Y. Unpublished work.
VI. Experimental Section
In order to save the journal space, the experimental details, including
synthetic procedures and spectroscopic data, are included in the
supporting information.
Acknowledgment. Financial support from the National
Institutes of Health (NS 12108), the National Science Foundation
(CHE 94-08247), and Eisai Pharmaceutical Company is grate-
fully acknowledged. Dr. A. Marcus Semones in this group is
gratefully acknowledged for the X-ray analysis of 50.
Supporting Information Available: The experimental
details, including the synthetic procedures and spectroscopic data
(63) Related to this, we should mention a recent example from this
laboratory, which demonstrated that the binding affinity of the H-type II
human blood group determinant toward the lectin I was strongly affected
by changes in their preferred solution conformation at the ground state (Wei,
A.; Boy, K. M.; Kishi, Y. J. Am. Chem. Soc. 1995, 117, 9432-9436).
(64) Evans, D. A.; Fu, G. C.; Hoveyda, A. H. J. Am. Chem. Soc. 1988,
110, 6917-6918; 1992, 114, 6671-6679.
(65) Compound ((E)-4-(tert-butyldimethylsiloxy)-2-buten-2-yl)lithium
was prepared by lithiation (t-BuLi, THF, -78 °C, 30 min) of (E)-4-(tert-
butyldimethylsiloxy)-2-buten-2-yl bromide.25
(66) (4-(1-Ethoxyethoxy)-3,3-dimethylbutyl)lithium was prepared by
lithiation (t-BuLi, THF, -78 °C, 30 min) of 4-(1-ethoxyethoxy)-3,3-
dimethylbutyl bromide, which was in turn prepared from 3,3-dimethylox-
etane in four steps: (1) vinyllithium, BF3‚OEt2, -78 °C; (2) ethyl vinyl
ether, p-TsOH; (3) O3, MeOH then NaBH4; (4) Ph3P, imidazole, Br2.
(67) Bocker, T.; Lindhorst, T. K.; Thiem, J.; Vill, V. Carbohydr. Res.
1992, 230, 245-256.
(68) HF‚pyr was prepared by mixing 30.0 g of aqueous HF (49%) and
14.9 mL of pyridine.
(69) (a) Araki, S.; Ito, H.; Butsugan, Y. J. Org. Chem. 1988, 53, 1831-
1833. (b) Kim, E.; Gordon, D. M.; Schmid, W.; Whitesides, G. M. J. Org.
Chem. 1993, 58, 5500-5507 and references cited therein.
(70) Compound 4-(tert-butyldimethylsiloxy)-1-buten-2-yl iodide was
prepared from 3-butyn-1-ol in three steps: (1) n-Bu3SnH, CuCN, n-BuLi;
(2) TBSCl, imidazole, DMF; (3) I2.
(71) (a) Mahoney, W. S.; Brestensky, D. M.; Stryker, J. M. J. Am. Chem.
Soc. 1988, 110, 291-293. (b) Koenig, T. M.; Daeuble, J. F.; Brestensky,
D. M.; Stryker, J. M. Tetrahedron Lett. 1990, 31, 3237-3240 and references
cited therein.
(72) (a) Seyferth, D.; Simmons, H. D., Jr.; Singh, G. J. Organomet. Chem.
1965, 3, 337-339. (b) Seyferth, D.; Heeren, J. K.; Singh, G.; Grim, S. O.;
Hughes, W. B. J. Organomet. Chem. 1966, 5, 267-274.
(73) Williams, D. R.; White, F. H. J. Org. Chem. 1987, 52, 5067-5079.
(74) (a) Nozaki, K.; Oshima, K.; Utimoto, K. J. Am. Chem. Soc. 1987,
109, 2547-2549. (b) Nozaki, K.; Oshima, K.; Utimoto, K. Tetrahedron
1989, 45, 923-933.
(75) Hofmeister, H.; Laurent, H.; Schulze, P.-E.; Wiechert, R. Tetrahe-
dron 1986, 42, 3575-3578.
(76) Mahler, H.; Braun, M. Tetrahedron Lett. 1987, 28, 5145-5148.
(77) Grieco, P. A.; Miyashita, M. Tetrahedron Lett. 1974, 1869-1871.
(78) Mori, K.; Ueda, H. Tetrahedron 1982, 38, 1227-1233.
(79) Yamamoto, Y.; Nishii, S.; Ibuka, T. J. Chem. Soc., Chem. Commun.
1987, 1572-1573.
(d) We learned that Tachibana and co-workers have recently
concluded the absolute configuration of MTX to be represented
by 1C via (1) degradation of MTX (NaIO4 oxidation, followed
by NaBH4 reduction) to the C.136-C1.42 primary alcohol and
(2) chiral gas chromatographic analysis of the resultant primary
alcohol. We have also developed a method to determine its
absolute configuration, which involves (1) degradation of MTX
to the C.136-C.142 primary alcohol via the known procedure,6c
(2) its esterification with a â-naphthyl analog of Mosher acid,
and (3) HPLC analysis. The naphthyl analog of Mosher acid
was used for improvement of the sensitivity of detection; it
allowed an analysis of the absolute configuration of MTX with
much less than 0.01 mg. The absolute configuration of MTX
from two different sources, Wako Chemicals USA (Catalog No.
131-10731, Lot YLE9504) and Calbiochem-Novabiochem
(Catalog No. 442620, Lot B 11288), has recently been examined
by this method: Kim, H.; Kishi, Y. Unpublished results.
JA961230+