Chemical synthesis of the GHIJKLMNO ring system of maitotoxin

Article abstract of DOI:10.1021/ja801139f

As the largest secondary metabolite to be discovered as of yet, the polyether marine neurotoxin maitotoxin constitutes a major structural and synthetic challenge. After its originally proposed structure (1) had been questioned on the basis of biosynthetic considerations, we provided computational and experimental support for structure 1. In an effort to provide stronger experimental evidence: of the molecular architecture of maitotoxin, its GHIJKLMNO ring system 3 was synthesized. The ¹³C NMR chemical shifts of synthetic 3 matched closely those corresponding to the same domain of the natural product providing strong evidence for the correctness of the originally proposed structure of maitotoxin (1).

Full text of DOI:10.1021/ja801139f

Published on Web 05/16/2008
Chemical Synthesis of the GHIJKLMNO Ring System of
Maitotoxin
K. C. Nicolaou,* Michael O. Frederick, Antonio C. B. Burtoloso, Ross M. Denton,
Fatima Rivas, Kevin P. Cole, Robert J. Aversa, Romelo Gibe, Taiki Umezawa, and
Takahiro Suzuki
Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research
Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and Department of
Chemistry and Biochemistry, UniVersity of California, San Diego, 9500 Gilman DriVe,
La Jolla, California 92093
Received February 14, 2008; E-mail: kcn@scripps.edu
Abstract: As the largest secondary metabolite to be discovered as of yet, the polyether marine neurotoxin
maitotoxin constitutes a major structural and synthetic challenge. After its originally proposed structure (1)
had been questioned on the basis of biosynthetic considerations, we provided computational and
experimental support for structure 1. In an effort to provide stronger experimental evidence of the molecular
architecture of maitotoxin, its GHIJKLMNO ring system 3 was synthesized. The 13C NMR chemical shifts
of synthetic 3 matched closely those corresponding to the same domain of the natural product providing
strong evidence for the correctness of the originally proposed structure of maitotoxin (1).
toxicus.1d However, it would not be until 1988 that the substance
was actually isolated from a broth of G. toxicus by Yasumoto
Introduction
Maitotoxin is the largest secondary metabolite isolated1 as
yet from any living creature. Its legendary toxicity surpasses
that of any known molecule, other than a few proteomic
substances. As such, this impressive natural product elicited
considerable attention from the scientific community.1–5 Mai-
totoxin was first detected in the gut of the surgeonfish Ctenocha-
etus striatus,1b,c and later in the dinoflagellate Gambierdiscus
and co-workers.1e The gross structure of maitotoxin was
proposed by Yasumoto and co-workers in 1993.2b Its relative
stereochemistry was defined by Kishi et al. in 1996,3a with its
absolute stereochemistry assigned by Tachibana et al. in the
same year (1, Figure 1).4e In 2006, the assigned structure of
maitotoxin came under close scrutiny by Gallimore and Spencer
on the basis of biosynthetic considerations, which suggested
the opposite configuration at the two stereocenters of the JK
junction (structure 1, C-51, C-52, Figure 1).6 Following this
challenge, and in order to test the Gallimore-Spencer hypoth-
esis, we resorted to computational chemistry, which provided
support for the originally proposed (1), rather than a revised,
structure.7 In a subsequent study, we synthesized the GHIJK
ring system 2 (Figure 1) of maitotoxin and compared its 13C
chemical shifts with the corresponding 13C chemical shifts of
the natural product (1), an exercise that provided experimental
evidence in support of the originally proposed structure (1) of
maitotoxin.8 In this article, we describe the chemical synthesis
of the entire GHIJKLMNO ring domain 3 (Figure 1) of structure
1, and the comparison of the 13C chemical shifts of this fragment
to those reported for the same domain of the natural product,
which provided further support for the originally proposed
structure (1) of maitotoxin.
(1) (a) Murata, M.; Yasumoto, T. Nat. Prod. Rep. 2000, 17, 293. (b)
Yasumoto, T.; Bagnins, R.; Randal, J. E.; Banner, A. H. Bull. Jpn.
Soc. Sci. Fish. 1976, 37, 724. (c) Yasumoto, T.; Bagnins, R.; Vernoux,
J. P. Bull. Jpn. Soc. Sci. Fish. 1976, 42, 359. (d) Yasumoto, T.;
Nakajima, I.; Bagnis, R.; Adachi, R. Bull. Jpn. Soc. Sci. Fish. 1977,
43, 1021. (e) Yokoyama, A.; Murata, M.; Oshima, Y.; Iwashita, T.;
Yasumoto, T. J. Biochem. 1988, 104, 184.
(2) (a) Murata, M.; Iwashita, T.; Yokoyama, A.; Sasaki, M.; Yasumoto,
T. J. Am. Chem. Soc. 1992, 114, 6594. (b) Murata, M.; Naoki, H.;
Iwashita, T.; Matsunaga, S.; Sasaki, M.; Yokoyama, A.; Yasumoto,
T J. Am. Chem. Soc. 1993, 115, 2060. (c) Murata, M.; Naoki, H.;
Matsunaga, S.; Satake, M.; Yasumoto, T. J. Am. Chem. Soc. 1994,
116, 7098. (d) Satake, M.; Ishida, S.; Yasumoto, T. J. Am. Chem.
Soc. 1995, 117, 7019.
(3) (a) Zheng, W.; DeMattei, J. A.; Wu, J.-P.; Duan, J. J.-W.; Cook, L. R.;
Oinuma, H.; Kishi, Y. J. Am. Chem. Soc. 1996, 118, 7946. (b) Cook,
L. R.; Oinuma, H.; Semones, M. A.; Kishi, Y. J. Am. Chem. Soc.
1997, 119, 7928. (c) Kishi, Y. Pure Appl. Chem. 1998, 70, 339.
(4) (a) Sasaki, M.; Nonomura, T.; Murata, M.; Tachibana, K. Tetrahedron
Lett. 1995, 36, 9007. (b) Sasaki, M.; Nomomura, T.; Murata, M.;
Tachibana, K.; Yasumoto, T. Tetrahedron Lett. 1995, 36, 9011. (c)
Sasaki, M.; Nonomura, T.; Murata, M.; Tachibana, K. Tetrahedron
Lett. 1994, 35, 5023. (d) Sasaki, M.; Matsumori, N.; Muruyama, T.;
Nonomura, T.; Murata, M.; Tachibana, K.; Yasumoto, T. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1672. (e) Nonomura, T.; Sasaki, M.;
Matsumori, N.; Murata, M.; Tachibana, K.; Yasumoto, T. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1675.
Results and Discussion
1. Retrosynthetic Analysis. Having defined our target as
structure 3, we proceeded to consider a strategy by which to
(6) Gallimore, A. R.; Spencer, J. B. Angew. Chem., Int. Ed. 2006, 45,
(5) (a) Nicolaou, K. C.; Postema, M. H. D.; Yue, E. W.; Nadin, A. J. Am.
Chem. Soc. 1996, 118, 10335. (b) Nakata, T.; Nomura, S.; Matsukura,
H. Chem. Pharm. Bull. 1996, 44, 627. (c) Nagasawa, K.; Hori, N.;
Shiba, R.; Nakata, T. Heterocycles 1997, 44, 105. (d) Sakamoto, Y.;
Matsuo, G.; Matsukura, H.; Nakata, T. Org. Lett. 2001, 3, 2749.
4406.
(7) Nicolaou, K. C.; Frederick, M. O. Angew. Chem., Int. Ed. 2007, 46,
5278.
(8) Nicolaou, K. C.; Cole, K. P.; Frederick, M. O.; Aversa, R. J.; Denton,
R. M. Angew. Chem., Int. Ed. 2007, 46, 8875.
9
7466 J. AM. CHEM. SOC. 2008, 130, 7466–7476
10.1021/ja801139f CCC: $40.75
2008 American Chemical Society

GHIJKLMNO Ring System of Maitotoxin
A R T I C L E S
Figure 1. Structure of maitotoxin (1), previously synthesized GHIJK ring system 2, and targeted GHIJKLMNO ring system 3.
reach it. An appealing retrosynthetic analysis is shown in Figure
2 in which the molecule (3) is dissected approximately in the
middle by rupturing the indicated C-C and C-O bonds. This
dissection led to advanced intermediates aldehyde 4 and alkyne
5, whose union/elaboration was expected to forge the missing
ring (K). Aldehyde 4 was then dismantled at the indicated bonds
(two C-C and two C-O bonds) furnishing ring systems 6 (G)
and 7 (J) as potential key building blocks, whereas alkyne 5
was envisioned to arise from two molecules of a single key
building block (8) by virtue of the pseudo symmetry of the
molecule. Finally, all three key building blocks (6-8) were
traced back to furan (9) and its derivatives. A strategy based
on this analysis would have the advantage of high convergency
and the aesthetic appeal of deriving a most complex polyether
system from the simplest of oxygen heterocycles, furan itself.
2. Carbohydrate Approach to the LM, NO, and LMNO
Fragments of Maitotoxin. As a prelude to devising a synthetic
strategy to the targeted GHIJKLMNO maitotoxin domain, we
needed to charter a viable route to its LMNO region (we already
had enough intelligence regarding the GHIJK region gathered
from our previous work).8 Toward this end, we targeted the
needed LM/NO common key building block 8 for which we
developed the synthetic route depicted in Schemes 1 and 2.
Thus, starting with diol 10 (conveniently derived from methyl-
D-glucopyranoside),9 the bis-benzyl ether 11 was prepared (NaH,
BnBr, 97% yield) and converted to the new diol 12 by the action
of CSA cat. in the presence of EtSH (91% yield). The primary
alcohol 13 was then prepared from 12 by a two-step procedure
involving bis-silylation (TBSOTf, 2,6-lut.) followed by selective
monodesilylation (CSA cat., MeOH, 80% overall yield for the
two steps). Swern oxidation of the latter compound [(COCl)2,
DMSO, -78 °C; Et3N, 0 °C]10 followed by Baeyer-Villiger
(9) Sasaki, M.; Ishikawa, M.; Fuwa, H.; Tachibana, K. Tetrahedron 2002,
Figure 2. First retrosynthetic analysis of the GHIJKLMNO ring system 3
of maitotoxin.
58, 1889.
(10) Omura, K.; Sharma, A. K.; Swern, D. J. Org. Chem. 1976, 41, 957.
9
J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7467

A R T I C L E S
Nicolaou et al.
Scheme 1. Construction of the L/N Common Building Block 16a
Scheme 2. Synthesis of the LMNO Enone 25a
a Reagents and conditions: (a) BnBr (4.0 equiv), NaH (60% in mineral
oil, 3.5 equiv), DMF, 0 f 25 °C, 3 h, 97%; (b) CSA (0.2 equiv), EtSH:
CH2Cl2 (1:3), 25 °C, 16 h, 91%; (c) TBSOTf (3.0 equiv), 2,6-lut. (5.0 equiv),
CH2Cl2, 0 °C, 2 h; (d) CSA (0.3 equiv), CH2Cl2:MeOH (1:1), 0 °C, 2 h,
80% over the two steps; (e) (COCl)2 (3.0 equiv), DMSO (5.0 equiv), -78
°C, 1 h; Et3N (10 equiv), 0 °C, 30 min; (f) m-CPBA (70%, 4.0 equiv),
CH2Cl2, 25 °C, 3 h, 84% over the two steps; (g) allylTMS (6.0 equiv),
TBSOTf (3.0 equiv), CH3CN, -40 f0 °C, 1 h, 80%; (h) methyl acrylate
(25 equiv), Grubbs II cat. (0.005 equiv), benzene, 85 °C, 2 h, 85%.
Abbreviations: TBDPS ) tert-butyldiphenylsilyl; DMF ) N,N-dimethyl-
formamide; CSA ) (()-camphor-10-sulfonic acid; TBS ) tert-butyldi-
methylsilyl; Tf ) trifluoromethanesulfonyl; lut. ) lutidine; DMSO )
dimethyl sulfoxide; m-CPBA ) meta-chloroperbenzoic acid; TMS )
trimethylsilyl.
a Reagents and conditions: (a) DIBAL-H (1.0 M in CH2Cl2, 2.5 equiv),
CH2Cl2, -78 °C, 1 h, 89%; (b) (-)-DET (1.2 equiv), Ti(i-PrO)4 (1.0 equiv),
t-BuOOH (5.0 M in decane, 4.0 equiv), 4 Å MS, CH2Cl2, -20 °C, 24 h,
86% (6:1 mixture of epoxides); (c) SO3•py. (3.5 equiv), Et3N (10 equiv),
CH2Cl2:DMSO (5:1), 0 °C, 2 h; (d) CH3PPh3Br (2.0 equiv), KHMDS (0.5
M in PhMe, 1.9 equiv), THF, -78 f 0 °C, 1 h, 95% over the two steps;
(e) TBAF (1.0 M in THF, 3.0 equiv), THF, 25 °C, 1 h; (f) CSA (0.2 equiv),
CH2Cl2, 0 f 25 °C, 2 h, 65% over the two steps; (g) BnBr (3.0 equiv),
TBAI (0.2 equiv), NaH (60% in mineral oil, 2.0 equiv), THF, 25 °C, 3 h,
72% yield; (h) O3, CH2Cl2, -78 °C, 10 min; Ph3P (1.5 equiv), -78 f 0
°C, 30 min, 100%; (i) 5% Pd/C (20% w/w), H2, EtOAc, 25 °C, 1 h; (j)
TBAF (1.0 M in THF, 3.0 equiv), THF, 25 °C, 16 h, 96% over the two
steps; (k) SO3•py. (3.5 equiv), Et3N (10 equiv), CH2Cl2:DMSO (5:1), 0
°C, 2 h; (l) n-BuLi (1.6 M in hexanes, 5.0 equiv), MeP(O)(OMe)2 (5.0
equiv), THF, -78 °C, 1 h; then aldehyde from 23, THF, -78 °C, 1 h; (m)
DMP (2.0 equiv), CH2Cl2, 25 °C, 30 min, 75% over the three steps; (n) 22
(1.0 equiv), LiCl (2.0 equiv), i-Pr2NEt (1.0 equiv), CH3CN, 25 °C, 75%.
Abbreviations: DIBAL-H ) diisobutylaluminum hydride; DET ) diethyl
tartrate; MS ) molecular sieves; KHMDS ) potassium bis(trimethylsilyl)-
amide; THF ) tetrahydrofuran; TBAF ) tetra-n-butylammonium fluoride;
TBAI ) tetra-n-butylammonium iodide; DMP ) Dess-Martin periodinane.
oxidation of the resulting aldehyde (m-CPBA)11 afforded
formate ester 14 in 84% overall yield for the two steps, setting
the stage for the required chain extension. Indeed, treatment of
14 with allylTMS in the presence of TBSOTf gave C-glycoside
15 (ꢀ-anomer, 80% yield), which was subjected to cross-
metathesis with methyl acrylate (Grubbs II cat.)12 to afford the
desired R,ꢀ-unsaturated ester 16 in 85% yield.
Ester 16 was then reduced (Scheme 2) with DIBAL-H (89%
yield), and the resulting allylic alcohol (17) was subjected to
Sharpless asymmetric epoxidation [(-)-diethyl tartrate, Ti(i-
PrO)4, t-BuOOH]13 to afford epoxide 18 in ca. 6:1 diastereo-
meric ratio and 86% combined yield. Subsequent oxidation of
the latter with SO3•py. and Et3N furnished the corresponding
aldehyde, which was subjected to methylenation (CH3PPh3Br,
KHMDS) to give olefin 20 in 95% overall yield for the last
two steps. The next ring was forged by selectively removing
the TBS group from 20 (TBAF) and then inducing the expected
regioselective cyclization of the resulting hydroxy epoxide by
exposure to catalytic amounts of CSA to form bicycle 21 in
65% overall yield for the two steps.14 Benzylation (NaH, BnBr)
of the latter compound then gave the coveted common inter-
mediate 8 (72% yield), whose divergence into key building
blocks 22 (LM ring system) and 24 (NO ring system) was
anticipated as part of the plan to reach the targeted LMNO ring
system. Thus, the desired LM aldehyde fragment 22 was
generated through ozonolysis (O3; Ph3P) of olefin 8 in quantita-
tive yield, while the other required fragment NO ketophospho-
(11) Renz, M.; Meunier, B. Eur. J. Org. Chem. 1999, 737.
(12) (a) Schwab, P.; France, M. B.; Ziller, J. W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 107, 2179. (b) Scholl, M.; Ding, S.; Lee,
C. W.; Grubbs, R. H. Org. Lett. 1999, 1, 953. For a review on the use
of Grubbs’ catalysts in total synthesis see: (c) Nicolaou, K. C.; Bulger,
P. G.; Sarlah, D. Angew. Chem., Int. Ed. 2005, 44, 4490.
(14) (a) Nicolaou, K. C.; Duggan, M. E.; Hwang, C.-K.; Somers, P. K.
J. Chem. Soc. Chem. Commun 1985, 1359. (b) Nicolaou, K. C.; Prasad,
C. V. C.; Somers, P. K.; Hwang, C.-K. J. Am. Chem. Soc. 1989, 111,
5330.
(13) (a) Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974.
(b) Katsuki, T.; Martin, V. S. Org. React. 1996, 48, 1.
9
7468 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

GHIJKLMNO Ring System of Maitotoxin
A R T I C L E S
Scheme 3. Synthesis of LMNO Model Compound 31a
Scheme 4. Synthesis of L/N Fragment 46a
a Reagents and conditions: (a) (S)-CBS (1.0 M in PhMe, 1.2 equiv),
BH3•THF (1.0 M in THF, 1.2 equiv), THF, -20 °C, 1 h, 70%; (b) m-CPBA
(3.0 equiv), CH2Cl2, 25 °C, 8 h, 60% + 21% of the other isomer; (c) DMP
(1.5 equiv), CH2Cl2, 25 °C, 1 h, 100%; (d) SmI2 (0.1 M in THF, 4.0 equiv),
MeOH (10 equiv), THF, 0 °C, 10 min, 96%; (e) Me4NHB(OAc)3 (2.0
equiv), CH3CN:AcOH (2:1), -20 °C, 5 h, 88%; (f) TBAF (1.0 M in THF,
2.0 equiv), THF, 25 °C, 12 h; (g) 20% Pd(OH)2/C (30% w/w), H2, MeOH:
EtOAc (1:1), 25 °C, 12 h, 75% over the two steps. Abbreviations: CBS )
2-methyl-Corey-Bakshi-Shibata-oxazaborilidine.
a Reagents and conditions: (a) n-BuLi (1.6 M in hexanes, 1.5 equiv), 9
(1.5 equiv), THF, 0 °C, 30 min; then 32 (1.0 equiv), -78 °C, 1 h; AcOH:
H2O (7:1), 50 °C, 12 h; (b) PivCl (1.4 equiv), Et3N (2.5 equiv), CH2Cl2, 0
f 25 °C, 4 h, 69% yield over the three steps; (c) 34 (0.005 equiv), TBAB
(0.3 equiv), HCO2Na (10 equiv), CH2Cl2:H2O (1:1), 25 °C, 36 h, 98%(g
95% ee); (d) NBS (1.0 equiv), NaOAc (1.0 equiv), NaHCO3 (2.0 equiv),
THF:H2O (2.5:1), 0 °C, 1 h; (e) PivCl (1.0 equiv), Et3N (1.25 equiv),
4-DMAP (0.05 equiv), CH2Cl2, -78 °C, 10 min, 65% yield over the two
steps (+20% of the other anomer); (f) NaBH4 (1.0 equiv), CeCl3•7H2O
(0.5 equiv), CH2Cl2:MeOH (1:1), -78 °C, 1 h, 98%; (g) BnBr (5.0 equiv),
TBAI (0.5 equiv), 25% NaOH (aq.):PhMe (1:1), 25 °C, 18 h, 85%; (h)
OsO4 (1.0% in H2O, 0.02 equiv), NMO (2.0 equiv), acetone:H2O (5:1), 25
°C, 18 h, 80%; (i) n-Bu2SnO (1.0 equiv), BnBr (1.4 equiv), TBAI (1.0
equiv), C6H6, reflux, 18 h, 89%; (j) Ac2O (3.0 equiv), py. (6.0 equiv), 4-
DMAP (0.05 equiv), CH2Cl2, 25 °C, 1 h, 99%; (k) allylTMS (3.0 equiv),
BF3•Et2O (2.0 equiv), CH3CN, 50 °C, 5 h, 94%; (l) K2CO3 (0.4 equiv),
MeOH, 25 °C, 4 h, 79%; (m) 4-NO2C6H6CO2H (2.0 equiv), DIAD (2.25
equiv), Ph3P (2.25 equiv), PhMe, 70 °C, 5 h, 75%; (n) K2CO3 (0.4 equiv),
MeOH, 25 °C, 0.5 h, 99%; (o) TBSCl (2.0 equiv), imid. (4.0 equiv), DMF,
70 °C, 18 h, 92%; (p) LiAlH4 (1.0 M in THF, 2.0 equiv), Et2O, -78 °C,
0.5 h, 91%; (q) NAPBr (1.2 equiv), TBAI (0.5 equiv), NaH (2.0 equiv),
THF:DMF (1:1), 25 °C, 1 h, 92%. Abbreviations: Piv ) trimethylacetyl;
TBAB ) tetra-n-butylammonium bromide; NBS ) N-bromosuccinimide;
DMAP ) dimethylaminopyridine; NMO ) 4-methylmorpholine N-oxide;
py. ) pyridine; DIAD ) diisopropyl azodicarboxylate; imid. ) imidazole;
NAP ) naphthyl.
nate (24), was reached through the five-step sequence depicted
in Scheme 2. Thus, hydrogenation of the olefinic bond of 8
(H2, 5% Pd/C) followed by cleavage of the TBDPS protecting
group (TBAF) resulted in the formation of primary alcohol 23,
whose conversion to 24 required: (i) oxidation with SO3•py. in
the presence of Et3N; (ii) reaction of the resulting aldehyde with
the lithio derivative of dimethyl methylphosphonate [Me-
P(O)(OMe)2, n-BuLi]; and (iii) oxidation of the resulting 1:1
mixture of diastereomeric alcohols with Dess-Martin reagent,15
a sequence that proceeded in 75% overall yield for the three
steps. Coupling of fragments 22 and 24 was then accomplished
through a Horner-Wadsworth-Emmons olefination (LiCl,
i-Pr2NEt),16 providing enone 25 in 75% yield.
The overall plan of converting LMNO enone 25 to the
targeted LMNO ring system 31 through epoxide formation and
opening was implemented as shown in Scheme 3. Thus, having
failed to realize a diastereomerically useful direct epoxidation
of enone 25, we adopted a sequence that involved first selective
reduction of the enone moiety of 25 [BH3•THF, (S)-CBS cat.,
70% yield],17 then epoxidation (m-CPBA, 60% yield, plus 21%
yield of its diastereomeric epoxide) of the resulting allylic
alcohol (26), and, finally, DMP oxidation (100% yield) of epoxy
alcohol 27 to afford the desired epoxy ketone 28. It should be
noted that other epoxidation methods, such as Sharpless
asymmetric epoxidation,13 failed to provide any epoxide product,
presumably due to steric hindrance at the reaction site. The
epoxide within the latter compound was then regioselectively
opened with the aid of SmI2 in the presence of MeOH to give
ꢀ-hydroxy ketone 29 in 96% yield.18 Finally, stereoselective
(15) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
(16) Blanchette, M. A.; Choy, W.; Davis, J. T.; Essenfeld, A. P.; Masamune,
S.; Roush, W. R.; Sakai, T. Tetrahedron Lett. 1984, 25, 2183.
(17) Corey, E. J.; Helal, C. J. Angew. Chem., Int. Ed. 1998, 37, 1987.
(18) (a) Molander, G. A.; Harris, C. R. Chem. ReV. 1996, 96, 307. (b)
Kagan, H. B. Tetrahedron 2003, 59, 10351.
9
J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7469

A R T I C L E S
Nicolaou et al.
Scheme 5. Completion of the Furan-Based Synthesis of LM and
NO Building Blocks 55 and 57a
Scheme 6. Construction of the GIJ Ring System 62a
a Reagents and conditions: (a) O3, CH2Cl2:MeOH (5:1), -78 °C, 10 min;
then NaBH4 (1.0 equiv), -78 f 25 °C, 30 min; (b) TBSCl (1.5 equiv),
imid. (2.0 equiv), DMF, 25 °C, 2 h, 92% over the two steps; (c) DIBAL-H
(1.0 M in CH2Cl2, 5.0 equiv), CH2Cl2, -78 °C, 30 min, 98%; (d) TEMPO
(0.2 equiv), PhI(OAc)2 (5.0 equiv), CH2Cl2, 25 °C, 18 h, 80%; (e) PhNTf2
(5.0 equiv), KHMDS (0.5 M in PhMe, 5.0 equiv), THF, -78 °C, 30 min,
89%; (f) 6 (1.0 equiv), 9-BBN (0.5 M in THF, 4.0 equiv), THF, 60 °C,
3 h; then KHCO3 (1.0 M in H2O, 20 equiv), 25 °C, 30 min; then SPHOS
(0.2 equiv), Pd(OAc)2 (0.1 equiv), 56 (1.0 equiv), 25 °C, 48 h, 78%.
Abbreviations: TEMPO ) 2,2,6,6-tetramethyl-1-piperidinyloxy; 9-BBN )
9-borabicyclo[3.3.1]nonane; SPHOS ) 2-dicyclohexylphosphino-2′,6′-
dimethoxybiphenyl; TES ) triethylsilyl; PMB ) para-methoxybenzyl.
this juncture we had the option of producing the initially targeted
advanced LMNO alkyne fragment 5 as originally planned, or
to seek a more practical route to prepare a more advanced, and
therefore more relevant, maitotoxin fragment through our furan-
based technology8 that proved highly efficient in the synthesis
of the GHIJK ring system of maitotoxin. We chose the latter
option; but before abandoning this study, we decided to carry
out two additional manipulations on our advanced intermediate
30 in order to reach LMNO nonaol 31, an intermediate whose
enantiomer was previously synthesized by Tachibana et al.4b
To this end, diol 30 was sequentially treated with TBAF (to
remove the TBDPS group) and H2 in the presence of 20%
Pd(OH)2/C (to remove the benzyl groups) affording, in 75%
overall yield for the last two steps, compound 31. Indeed, the
1H and 13C NMR spectroscopic data of our synthetic intermedi-
ate 31 matched those reported for the enantiomer of 31,4b thus
confirming our stereochemical assignments along the described
route.
a Reagents and conditions: (a) methyl acrylate (20 equiv), Grubbs II cat.
(0.005 equiv), benzene, 80 °C, 1 h, 91%; (b) DIBAL-H (1.0 M in CH2Cl2, 2.5
equiv), CH2Cl2, -78 °C, 1 h, 89%; (c) (-)-DET (1.2 equiv), Ti(i-PrO)4 (1.0
equiv), t-BuOOH (5.0 M in decane, 4.0 equiv), 4 Å MS, CH2Cl2, -20 °C,
24 h, 84%; (d) SO3•py. (3.5 equiv), Et3N (10 equiv), CH2Cl2:DMSO (5:1), 0
°C, 2 h; (e) CH3PPh3Br (2.0 equiv), KHMDS (0.5 M in THF, 1.9 equiv), THF,
-78 °C, 1 h; then 49 (1.0 equiv), THF, -78 f 0 °C, 1 h, 93% over the two
steps; (f) TBAF (1.0 M in THF, 3.0 equiv), THF, 25 °C, 1 h; (g) CSA (0.2
equiv), CH2Cl2, 0 f 25 °C, 3 h, 65% yield over the two steps; (h) BnBr (3.0
equiv), TBAI (0.2 equiv), NaH (60% in mineral oil, 2.0 equiv), THF, 25 °C,
3 h, 91%; (i) O3, CH2Cl2:MeOH (5:1), -78 °C, 15 min; NaBH4 (1.0 equiv),
-78 f 25 °C, 30 min, 75%; (j) TBDPSCl (1.5 equiv), imid. (2.0 equiv),
CH2Cl2, 25 °C, 3 h, 89%; (k) DDQ (1.5 equiv), CH2Cl2:H2O (5:1), 0 °C, 2 h,
80%; (l) DMP (1.5 equiv), CH2Cl2, 25 °C, 1 h; (m) 54 (4.0 equiv), K2CO3
(2.0 equiv), MeOH, 25 °C, 16 h, 86% over the two steps; (n) 5% Pd/C (20%
w/w), H2, EtOAc, 25 °C, 1 h; (o) DDQ (1.5 equiv), CH2Cl2:H2O (5:1), 0 °C,
2 h, 82% yield over the two steps; (p) DMP (1.5 equiv), CH2Cl2, 25 °C, 1 h;
(q) (MeO)CH2PPh3Cl (5.0 equiv), LHMDS (1.0 M in hexanes, 4.5 equiv), THF,
0 °C, 30 min; then aldehyde from 56 (1.0 equiv), THF, -78 °C, 2 h; Hg(OAc)2
(5.0 equiv), THF:H2O (2:1), 25 °C, 15 min; HI (8% aq.), 25 °C, 30 min, 76%
yieldoverthetwosteps.Abbreviations:DDQ)2,3-dichloro-5,6-dicyano-p-benzo-
quinone.
3. Furan Approach to the LM, NO, and LMNO Fragments.
The furan approach to the LM, NO, and LMNO fragments of
maitotoxin proceeded through enone 36, an intermediate
prepared from furan in a manner similar to those described by
us8 and others.20 Thus, lithiation of furan (9, n-BuLi, Scheme
4) followed by addition of amide 32 led to, after removal of
(20) (a) Achmatowicz, O.; Bielski, R. Carbohydr. Res. 1977, 55, 165. (b)
Guo, H.; O’Doherty, G. A. Angew. Chem., Int. Ed. 2007, 46, 5206.
(c) Zhou, M.; O’Doherty, G. A. J. Org. Chem. 2007, 72, 2485. (d)
Guo, H.; O’Doherty, G. A. Org. Lett. 2006, 8, 1609. (e) Harris, J. M.;
Keranen, M. D.; Nguyen, H.; Young, V. G.; O’Doherty, G. A.
Carbohydr. Res. 2000, 328, 17. (f) Li, M.; Scott, J.; O’Doherty, G. A.
Tetrahedron Lett. 2004, 45, 1005. (g) Henderson, J. A.; Jackson, K. A.;
Phillips, A. L. Org. Lett. 2007, 9, 5299.
reduction of the carbonyl group within ꢀ-hydroxy ketone 29
was achieved through the use of Me4NHB(OAc)3 to afford the
anti diol 30 as a single diastereoisomer, and in 88% yield.19 At
(19) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc.
1988, 110, 3560.
9
7470 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

GHIJKLMNO Ring System of Maitotoxin
A R T I C L E S
Scheme 7. Failed Attempt to Construct the GHIJK Ring System
66a
Figure 3. Second generation retrosynthetic analysis of the GHIJKLMNO
ring system 3. The order 1, 2, and 3 refers to the synthetic direction.
a Reagents and conditions: (a) BH3•THF (1.0 M in THF, 10 equiv), THF,
0 °C, 18 h; then H2O2 (35% aq., 100 equiv), NaOH (1.0 M aq., 200 equiv),
25 °C, 5 h, 78%; (b) DMP (2.0 equiv), CH2Cl2, 25 °C, 3 h, 92%; (c) TsOH
(2.0 equiv), MeOH, 50 °C, 18 h; (d) TMSOTf (3.0 equiv), Et3SiH (5.0
equiv), CH3CN, 0 °C, 30 min, 90% over the two steps; (e) TESCl (30 equiv),
imid. (50 equiv), CH2Cl2, 25 °C, 10 h, 91%; (f) (COCl)2 (50 equiv), DMSO
(75 equiv), CH2Cl2, -78 °C, 1 h; then Et3N (125 equiv), 0 °C, 20 min,
90%. Abbreviations: Ts ) para-toluenesulfonyl.
39 were differentiated by selective protections [(i) n-Bu2SnO,
BnBr, TBAI cat., 40, 89% yield;23 (ii) Ac2O, py., 4-DMAP,
99% yield] to afford the corresponding 2-acetoxy-1-pivaloate.
Designed for its potential to undergo a stereoselective C-
allylation (by virtue of the directing effect of the 2-acetoxy
group), the latter intermediate was reacted with allylTMS in
the presence of BF3•Et2O to afford, indeed, the desired C-
allylation product 41 in 94% yield as a single diastereoisomer.
Having accomplished its mission, the acetate group was then
removed from 41, generating alcohol 42 and setting the stage
for the required inversion, which was carried out through a
Mitsunobu reaction (p-NO2C6H4COOH, DIAD, Ph3P)24 fol-
lowed by ester hydrolysis (K2CO3, MeOH), affording the desired
alcohol 43 in 59% overall yield for the three steps. Anticipated
operations going forward dictated protection of the free hydroxyl
group of 43 as a TBS ether (TBSCl, imid., 44, 92% yield) and
exchange of the pivaloate ester for a naphthyl ether [(i) LiAlH4;
(ii) NAPBr, TBAI cat., NaH, 84% overall yield], leading to the
L/N ring system 46 through intermediate 45.
With all stereocenters on the L/N rings of 46 set, the next
task was to append the M/O rings. To this end, alkene 46
was subjected to olefin cross metathesis with methyl acrylate
in the presence of Grubbs II catalyst12 to afford R,ꢀ-
unsaturated ester 47 in 91% yield as a single geometrical
isomer (Scheme 5). DIBAL-H reduction of 47 provided the
corresponding allylic alcohol (89% yield), whose Sharpless
asymmetric epoxidation [(-)-DET, Ti(i-PrO)4, t-BuOOH]13
resulted in the formation of epoxide 48 (84% yield) as a
the TBS group (AcOH, 50 °C) and protecting the resulting
primary alcohol as a pivaloate ester (PivCl, Et3N), ketone 33
(69% yield over the three steps). Ketone 33 was subjected to a
Noyori reduction (HCO2Na, n-Bu4NBr, 34 cat.)21 to afford
alcohol 35 in 98% yield and g95% ee. The latter compound
was converted to 36 through an Achmatowicz rearrangement20a
(NBS, NaOAc, NaHCO3) followed by pivaloate formation
(PivCl, Et3N, 4-DMAP) at the anomeric position, in 65% overall
yield for the two steps. This highly efficient process also yielded
the ꢀ-anomer of 36, isolated in 20% overall yield from 35. Luche
reduction (NaBH4, CeCl3•7H2O)22 of enone 36 proceeded
stereoselectively (as expected based on steric grounds) and
provided allylic alcohol 37, in 98% yield, which was benzylated
under biphasic conditions (25% aq. NaOH:PhMe 1:1, BnBr,
TBAI cat.) to afford benzyl ether 38 in 85% yield. Dihydroxyl-
ation of the latter compound (NMO, OsO4 cat.) then furnished
diol 39 in 80% yield and with complete stereocontrol. The
conversion of this diol to the next key intermediate required
allylation at the anomeric position and inversion of stereochem-
istry at C-2. To this end, the C-2 and C-3 hydroxyl groups within
(21) (a) Fujii, A.; Hashiguchi, S.; Uematsu, N.; Ikariya, T.; Noyori, R. J. Am.
Chem. Soc. 1996, 118, 2521. (b) Ferrie, L.; Reymond, S.; Capdevielle,
P.; Cossy, J. Org. Lett. 2007, 9, 2461.
(23) David, S.; Hanessian, S. Tetrahedron 1985, 41, 643.
(24) For reviews on the Mitsunobu inversion, see: (a) Mitsunobu, O.
Synthesis 1981, 1. (b) Hughes, D. L. Org. Prep. 1996, 28, 127.
(22) Luche, J. L. J. Am. Chem. Soc. 1978, 100, 2226.
9
J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7471

A R T I C L E S
Nicolaou et al.
Scheme 8. Synthesis of JKLM Ring System 72a
Scheme 9. Synthesis of IJKLM Vinyl Triflate Coupling Partner 76a
a Reagents and conditions: (a) BH3•THF (1.0 M in THF, 10 equiv), THF,
0 °C, 18 h; then H2O2 (35% aq., 100 equiv), NaOH (1.0 M aq., 200 equiv),
25 °C, 5 h, 79%; (b) TEMPO (0.2 equiv), PhI(OAc)2 (5.0 equiv), CH2Cl2,
25 °C, 18 h, 80%; (c) TESCl (20 equiv), imid. (30 equiv), CH2Cl2, 25 °C,
4 h, 92%; (d) PhNTf2 (5.0 equiv), KHMDS (0.5 M in PhMe, 5.0 equiv),
THF, -78 °C, 15 min, 89%.
a Reagents and conditions: (a) n-BuLi (1.6 M in hexanes, 2.0 equiv), 55
(2.0 equiv), THF, -78 °C, 1 h; then 67 (1.0 equiv), -78 °C, 30 min, 78%;
(b) DMP (2.0 equiv), CH2Cl2, 25 °C, 2 h, 96%; (c) HF (48% aq.):CH3CN
(1:3), 25 °C, 16 h, 89%; (d) AgOTf (0.5 equiv), CH2Cl2, 35 °C, 48 h, 95%;
(e) TBDPSCl (2.0 equiv), imid. (3.0 equiv), CH2Cl2, 25 °C, 3 h, 89%; (f)
NaBH4 (1.1 equiv), CeCl3•7H2O (0.3 equiv), CH2Cl2:MeOH (1:1), 0 °C,
30 min; (g) TESCl (2.0 equiv), imid. (3.0 equiv), CH2Cl2, 25 °C, 2 h; (h)
DIBAL-H (1.0 M in CH2Cl2, 10 equiv), CH2Cl2, -78 °C, 30 min, 98%
over the three steps.
aldehyde was reacted with Ohira-Bestmann reagent (54)25 to
afford the targeted key building block (55, 86% yield for the two
steps). The NO ring system ketophosphonate 24 was obtained from
the same intermediate 52 via primary alcohol 56 [(i) H2, 5% Pd/
C; (ii) DDQ, 82% yield overall for the two steps] and aldehyde 57
[(i) DMP; (ii) (MeO)CH2PPh3Cl-LHMDS; Hg(OAc)2,26 76%
overall yield for the two steps]. The elaboration of the latter
compound to ketophosphonate 24 has already been described above
(Scheme 2).
4. Initial Attempt to Construct the GHIJKLMNO Ring
System. With ample supplies of the LM and NO coupling
partners 55 and 57, we then turned our attention to the next
phase of the drive toward the target molecule, namely the
construction of the GHIJ advanced intermediate 4, as required
by the strategy outlined in Figure 2. As shown in Scheme 6,
this synthesis began with our furan-derived J ring system 588
and proceeded through coupling of vinyl triflate 61 and the
borane obtained from vinyl ether G ring system 6.8 Thus,
ozonolysis of 58 followed by NaBH4 reduction of the resulting
single isomer. Oxidation of the latter compound with SO3•py.
led to aldehyde 49, which was subjected to methylenation
(CH3PPh3Br, KHMDS, 93% yield over the two steps) to
afford olefinic epoxide 50. Removal of the TBS group
(TBAF) from the latter intermediate, followed by exposure
of the resulting hydroxy epoxide to CSA, led to bicyclic
system 51 in 65% yield over the two steps (plus 12% of the
chromatographically separable 5-membered ring).14 Benzyl-
ation of the remaining hydroxyl group within the latter
compound (NaH, BnBr, TBAI cat. 91% yield) then furnished
fully protected common intermediate 52.
From intermediate 52, the synthetic route diverged toward LM
alkyne 55 and NO ring system ketophosphonate 24 as shown in
Scheme 5. For the synthesis of alkyne 55, intermediate 52 was
first converted to primary alcohol 53 by a three-step sequence
involving: (i) ozonolysis/NaBH4 reduction (75% yield); (ii) silyl-
ation with TBDPSCl-imid. of the resulting primary alcohol (89%
yield); and (iii) removal of the NAP group with DDQ (80% yield).
The latter compound was then oxidized (DMP) and the resulting
(25) (a) Ohira, S. Synth. Commun. 1989, 19, 561. (b) Muller, S.; Liepold,
B.; Roth, G.; Bestmann, H. J. Synlett 1996, 521. (c) Roth, G.; Liepold,
B.; Muller, S.; Bestmann, H. J. Synthesis 2004, 59.
(26) Su, Q.; Dakin, L. A.; Panek, J. S. J. Org. Chem. 2007, 72, 2.
9
7472 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

GHIJKLMNO Ring System of Maitotoxin
A R T I C L E S
Scheme 10. Construction of the GHIJKLM ring system 80a
a Reagents and conditions: (a) 6 (3.0 equiv), 9-BBN (0.5 M in THF, 8.0 equiv), THF, 60 °C, 3 h; then KHCO3 (1.0 M in H2O, 20 equiv), 25 °C,
30 min; then SPHOS (0.2 equiv), Pd(OAc)2 (0.1 equiv), 76 (1.0 equiv), 25 °C, 48 h, 72%; (b) BH3•THF (1.0 M in THF, 10 equiv), THF, 0 °C, 18 h;
then H2O2 (35% aq., 100 equiv), NaOH (1.0 M aq., 200 equiv), 25 °C, 5 h, 85%; (c) DMP (2.0 equiv), CH2Cl2, 25 °C, 2 h, 93%; (d) TsOH (2.0
equiv), MeOH, 50 °C, 18 h; (e) TMSOTf (5.0 equiv), Et3SiH (10 equiv), CH3CN, 0 °C, 30 min, 95% over the two steps; (f) TESCl (20 equiv), imid.
(25 equiv), CH2Cl2, 25 °C, 3 h, 96%.
ozonide led to the corresponding primary alcohol, which was
silylated (TBSCl, imid.) to yield intermediate 7 in 92% overall
yield for the two steps. Removal of both pivaloate groups from
7 (DIBAL-H, 98% yield) followed by selective oxidation of
the primary alcohol of the resulting diol [59, TEMPO, PhI(O-
Ac)2]27 then furnished lactone 60, through the intermediate
lactol, in 80% yield. Lactone 60 was converted to vinyl triflate
61 by the action of PhNTf2 and KHMDS (89% yield). The latter
intermediate was then subjected to a Suzuki coupling with the
borane derived from vinyl ether 6 and 9-BBN,28 carried out in
the presence of KHCO3 and catalytic amounts of SPHOS29 and
Pd(OAc)2, to afford the desired coupling product 62 in 78%
yield.
The elaboration of the GIJ intermediate 62 to the tetracyclic
GHIJ ring system 4 required the casting of the missing ring
(H), an objective that was achieved through the sequence
depicted in Scheme 7. Thus, regioselective hydroboration/
oxidation of 62 (BH3•THF; H2O2, NaOH) followed by DMP
oxidation of the resulting secondary alcohol furnished ketone
63 (72% overall yield). Exposure of 63 to TsOH in MeOH at
50 °C for 18 h then caused cleavage of the TES, TBS, and
PMB groups, forming the corresponding H ring methyl acetal
(ꢀ anomer), which was reduced stereoselectively with Et3SiH
in the presence of TMSOTf to afford tetracyclic triol 64 (90%
overall yield for the two steps).30 The latter compound was then
loaded with three TES groups by exhaustive silylation (TESCl,
imid., 91% yield), and the resulting tri-TES ether (65) was
subjected to Swern oxidation conditions [(COCl)2, DMSO;
Et3N]31 to afford, selectively and as expected, aldehyde 4 with
the two secondary TES groups remaining intact. Unfortunately,
all attempts to forge the remaining ring (K) on this intermediate
proved unsuccessful, bringing this route to the targeted penta-
cyclic intermediate 66 to a hold. The inability to generate the
fifth ring of the desired system in this instance was attributed
to the diaxial arrangement of the tail-end substituent on ring J
(determined by 1H NMR spectroscopic analysis) and the rigidity
of the J ring caused by the GH and I rings, facts that keep the
two functionalities too far apart for productive engagement. A
new plan clearly had to be devised in order to overcome this
hurdle.
5. SecondGenerationRetrosyntheticAnalysisoftheGHIJKLM-
NO Ring System of Maitotoxin. Having failed to reach the
target molecule (3) through the originally planned strategy,
we then devised a new plan based on the modified retrosyn-
thetic analysis shown in Figure 3. Although the same
disconnections were adopted in the second generation ret-
rosynthetic analysis of the GHIJKLMNO ring system 3, the
order of the main coupling reactions was projected differently.
Thus, it was envisioned that the four key building blocks (6,
24, 55, and 67) defined by this new analysis, would be
processed in the synthetic direction as follows: 1 alkyne
addition, followed, after appropriate elaboration, by cycliza-
tion to form the JKLM ring system; 2 Suzuki coupling of a
borane derived from G ring system 6 and the appropriate
IJKLM vinyl triflate followed by cyclization to cast the
GHIJKLMringsystem;and3aHorner-Wadsworth-Emmons
olefination between ketophosphonate 24 and the appropriate
GHIJKLM aldehyde followed by final elaborations. The
formation of the K ring that proved problematic in our first
approach (vide supra) was hoped to be more facile by the
new approach now involving the more flexible ring J alone,
and without the restraints associated with the polycyclic
system employed previously.
6. Construction of the Maitotoxin GHIJKLMNO Ring
System. The construction of the targeted maitotoxin GHIJKL-
MNO ring system 3 began with the coupling of the LM
acetylenic fragment 55 with J ring aldehyde 678 and the
elaboration of the product to JKLM ring system 72 as shown
in Scheme 8. Thus, addition of the lithium acetylide generated
from acetylene 55 and n-BuLi (-78 °C) to aldehyde 67 at
-78 °C followed by DMP oxidation of the resulting alcohol
furnished ynone 68 in 75% overall yield. In preparation for
the pending ring closure to form ring K, the TBS group was
(27) Hansen, T. M.; Florence, G. J.; Lugo-Mas, P.; Chen, J.; Abrams, J. N.;
Forsyth, C. J. Tetrahedron Lett. 2003, 44, 57.
(28) Chemler, S. R.; Trauner, D.; Danishefsky, S. J. Angew. Chem., Int.
Ed. 2001, 40, 4544.
(29) Barder, T. E.; Walker, S. D.; Martinelli, J. R.; Buchwald, S. L. J. Am.
Chem. Soc. 2005, 127, 4685.
(30) Nicolaou, K. C.; Hwang, C.-K.; Nugiel, D. A. J. Am. Chem. Soc. 1989,
111, 4136.
(31) Tolstikov, G. A.; Miftakhov, M. S.; Adler, M. E.; Komissarova, N. G.;
Kuznetsov, O. M.; Vostrikov, N. S. Synthesis 1989, 940.
9
J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7473

A R T I C L E S
Nicolaou et al.
Scheme 11. Completion of the Synthesis of the Maitotoxin GHIJKLMNO Ring System 3a
a Reagents and conditions: (a) (COCl)2 (50 equiv), DMSO (75 equiv), CH2Cl2, -78 °C, 2 h; then Et3N (125 equiv), 0 °C, 20 min; (b) 24 (1.5 equiv),
i-Pr2NEt (3.0 equiv), LiCl (3.0 equiv), CH3CN, 25 °C, 5 h, 74% over the two steps; (c) (S)-CBS (1.0 M in PhMe, 2.0 equiv), BH3•THF (1.0 M in THF, 2.0
equiv), THF, -20 °C, 2 h, 79%; (d) m-CPBA (4.0 equiv), CH2Cl2, 18 h, 55% + 20% of the other diastereoisomer; (e) DMP (2.0 equiv), CH2Cl2, 25 °C,
2 h, 99%; (f) SmI2 (0.1 M in THF, 4.0 equiv), MeOH (10 equiv), THF, 0 °C, 30 min, 100%; (g) Me4NHB(OAc)3 (2.0 equiv), CH3CN:AcOH (2:1), -20 °C,
6 h, 86%; (h) TBAF (1.0 M in THF, 10 equiv), THF, 25 °C, 5 h, 76%; (i) 20% Pd(OH)2/C (50% w/w), H2, EtOH, 100 h, 71%.
removed from 68 by the action of aq. HF in acetonitrile in
a process that also cleaved the TBDPS group from the
primary hydroxyl moiety, leading to diol 69 in 89% yield.
Upon considerable experimentation, it was found that gentle
heating of dihydroxy ynone 69 in the presence of AgOTf
induced the desired cyclization to form the JKLM enone 70
in 95% yield.32 Protection of the remaining hydroxyl group
within 70 as a TBDPS ether (TBDPSCl, imid., 89% yield)
then led to 71, which underwent smooth and stereoselective
Luche reduction (NaBH4, CeCl3•7H2O) to afford, after TES
protection (TESCl, imid.) and removal of the pivaloate groups
(DIBAL-H), diol 72 in 98% overall yield for the three steps.
Tetracycle 72 was then elaborated to the desired pentacyclic
vinyl triflate 76 by the short sequence depicted in Scheme 9.
Thus, hydroboration/oxidation (BH3•THF; aq. H2O2, NaOH)5a
of the enol ether in 72 afforded triol 73 as a single stereoisomer
in 79% yield. Subsequent oxidation of the latter compound with
PhI(OAc)2 in the presence of catalytic amounts of TEMPO27
proceeded smoothly and selectively to afford hydroxy lactone
74 (80% yield), whose silylation (TESCl, imid.) furnished the
fully protected pentacyclic lactone 75 in 92% yield. Finally,
the IJKLM vinyl triflate 76 was generated from lactone 75
through the sequential action of PhNTf2 and KHMDS in 89%
yield.
The next task was fusion of the GH ring system onto the
growing molecule (i.e., 76) by attaching ring G fragment 6 and
appropriately elaborating the product. This operation required
another short sequence as shown in Scheme 10. Thus, the Suzuki
coupling28 of the IJKLM vinyl triflate 76 with the alkyl borane
generated in situ from alkene 6 and 9-BBN, as facilitated by
the catalytic action of Pd(OAc)2 cat. and SPHOS cat.29 in the
presence of KHCO3, afforded coupling product 77 in 72% yield.
Regio- and stereoselective hydroboration/oxidation of 77
(BH3•THF; aq. H2O2, NaOH, 85% yield)5a followed by DMP
oxidation (93% yield) of the resulting alcohol then led to ketone
78. Exposure of this hexacyclic compound (78) to TsOH in
(32) Wang, C.; Forsyth, C. J. Org. Lett. 2006, 8, 2997.
9
7474 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008

GHIJKLMNO Ring System of Maitotoxin
A R T I C L E S
Figure 4. Graphically depicted 13C chemical shift differences (∆δ, ppm) for each carbon between C-39 to C-75 for maitotoxin (1) and GHIJKLMNO ring
system 3. See Table 1.
MeOH at 50 °C resulted in the removal of the TES, PMB, and
TBDPS groups and caused concomitant ring closure to form
the corresponding heptacyclic methyl acetal (as a single
ꢀ-isomer), which was reduced stereoselectively with Et3SiH in
the presence of TMSOTf30 to afford the heptacyclic tetraol 79
in 95% yield over the two steps. The four hydroxyl groups of
79 were then silylated (TESCl, imid.) to provide the desired
GHIJKLM ring system 80 in 96% yield.
Table 1. C-39 to C-75 Chemical Shifts (δ) for Maitotoxin (MTX, 1)
and GHIJKLMNO Ring System 3 and Their Differences (∆δ,
ppm)a
carbon
δ for MTX (1) (ppm)
δ for 3 (ppm)
difference (ppm)
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
72.3
78.9
68.5
80.6
69.6
36.3
77.7
77.4
37.4
67.8
85.8
70.1
74.9
72.1
79.4
69.8
78.2
71.4
70.6
69.6
76.5
71.7
33.9
66.1
77.2
67.7
43.2
66.0
42.5
71.4
76.7
69.2
76.8
71.2
34.2
65.1
72.4
74.4
73.0
71.7
81.0
69.6
36.3
77.8
77.4
37.6
67.8
85.8
70.1
74.9
72.0
79.4
69.8
78.1
71.5
70.2
69.7
76.2
71.8
33.9
66.1
77.3
67.5
43.2
66.0
42.4
71.4
76.3
69.3
76.7
71.1
34.2
64.2
64.6
-2.1
5.9
-3.2
-0.4
0.0
0.0
-0.1
0.0
-0.2
0.0
0.0
0.0
0.0
0.1
0.0
0.0
0.1
-0.1
0.4
-0.1
0.3
-0.1
0.0
0.0
-0.1
0.2
0.0
0.0
0.1
0.0
0.4
The final drive toward the desired GHIJKLMNO ring
system 3 required the attachment of the NO ring system to
the growing molecular assembly (i.e., 80) and further
elaboration of the product. This task was accomplished as
summarized in Scheme 11. Thus, upon exposure to Swern
oxidation conditions [(COCl)2, DMSO, -78 °C; Et3N, 0 °C]31
the tetra-TES protected heptacycle 80 underwent selective
deprotection/oxidation at the primary TES ether site, furnish-
ing the corresponding aldehyde, which entered smoothly into
a Horner-Wadsworth-Emmons olefination reaction with
ketophosphonate 24 (i-Pr2NEt, LiCl)16 to afford enone 81 in
74% overall yield for the two steps. This enone was then
subjected to a CBS reduction (BH3•THF, (S)-CBS cat. 79%
yield)17 to afford, stereoselectively, a major allylic alcohol
(82, ca. g 10:1 ratio), which was presumed to possess the
desired R-stereochemistry on the basis of the chirality of the
catalyst used and by analogy to our previous results with
the LMNO system 25 (vide supra, Scheme 3). Allylic alcohol
82 was then epoxidized with m-CPBA, furnishing the desired
R-epoxide (83) as the major product (55% yield) together
with the undesired ꢀ-epoxide (20% yield), the two isomers
being chromatographically separated. Again, the stereochem-
istry of the major isomer (83) was presumed to be the desired
one on the basis of the expected directing effect of the nearby
hydroxyl group and by analogy to our previous studies with
the LMNO model system 26 (vide supra, Scheme 3).
Oxidation of alcohol 83 then gave epoxy ketone 84, which
underwent regioselective opening with SmI2 to afford hy-
droxy ketone 85 in 99% overall yield for the two steps.
-0.1
0.1
0.1
0.0
0.9
7.8
a 150 MHz, 1:1 methanol-d4/pyridine-d5.
9
J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008 7475

A R T I C L E S
Nicolaou et al.
Stereoselective reduction of the ꢀ-hydroxy ketone moiety
within 85 with Me4NHB(OAc)3 provided 1,3-anti-diol 86
proposed structure of this natural product. Relying on furan as
starting material and a Noyori asymmetric reduction as a source
of chirality, the developed synthetic strategy for the maitotoxin-
like structures differs significantly from the previously used
methods33 toward polyether marine toxins, which relied heavily
on carbohydrate-based starting materials. The efficiency and
flexibility of the furan approach to polyether marine natural
products through asymmetric catalysis bodes well for its scope
and future applications beyond the content described within this
article.
19
as a single stereoisomer, and in 86% yield. Finally, desilyl-
ation of 86 with TBAF afforded pentaol 87 (76% yield), and
subsequent debenzylation of 87 with 20% Pd(OH)2/C and
H2 led to the coveted maitotoxin GHIJKLMNO ring system
3 in 71% yield.34
7. Comparisonofthe13CNMRChemicalShiftsofGHIJKLM-
NO Ring System 3 with Those Corresponding to the Same
Region of Maitotoxin. The C-39 to C-75 13C NMR chemical
shifts (δ) exhibited by the GHIJKLMNO ring system 3 and those
reported for the corresponding carbons of maitotoxin,2d together
with their differences, are listed in Table 1. The observed
chemical shift (∆δ) differences for these carbons are graphically
displayed in Figure 4. As seen in both depictions, there is
compelling agreement between the two sets of 13C chemical
shifts, although the two compounds exhibit significant differ-
ences at the tail-ends of the GHIJKLMNO domain. This is to
be expected, of course, due to the rather drastic changes of
substituents at those ends. Thus, the average difference (∆δ)
between the two sets of δ values for carbons C-42 to C-73 is
0.09 ppm, and the maximum difference (∆δ) between these
values is 0.4 ppm. These striking experimental results render
compelling support for the correctness of the Yasumoto-Kishi-
Tachibana structure 1 originally proposed for maitotoxin.
Acknowledgment. We thank Drs. D.-H. Huang and G. Siuzdak
for spectroscopic and mass spectrometric assistance, respectively.
Financial support for this work was provided by the National
Institutes of Health (U.S.A.) and Skaggs Institute of Chemical
Biology, along with fellowships from the National Science Founda-
tion (to M.O.F.), CNPq (Brazil) (to A.C.B.B.), UCSD/SDSU
IRACDA (to F.R.), National Institutes of Health (U.S.A.) (to
K.P.C.), A*STAR Singapore (to R.G. and T.S.), and Uehara
Memorial Fund, Japan (to T.U.).
Supporting Information Available: Experimental procedures
and compound characterization. This material is available free
of charge via the Internet at http://pubs.acs.org.
JA801139F
Conclusion
(33) For selected reviews on the synthesis of fused polyether natural
products, see: (a) Nicolaou, K. C. Angew. Chem., Int. Ed. Engl. 1996,
35, 588. (b) Nakata, T. Chem. ReV. 2005, 105, 4314. (c) Inoue, M.
Chem. ReV. 2005, 105, 4379. (d) Sasaki, M. Bull. Chem. Soc. Jpn.
2007, 80, 856. For a directing group-free synthesis of cyclic polyether
systems based on oligoepoxide openings in water, see: (e) Vilotijevic,
I.; Jamison, T. F. Science 2007, 317, 1189.
The described chemistry rendered a substantial portion (3)
of the maitotoxin molecule (1) readily available for further
chemical and biological studies. The synthesized GHIJKLMNO
ring system 3 demonstrated strong correlation with the corre-
sponding region of maitotoxin in terms of its 13C NMR chemical
shifts, thus lending compelling support for the originally
(34) The stereochemistry within 3 and its precursors were based on COSY,
ROESY, and HMQC NMR spectroscopic experiments.
9
7476 J. AM. CHEM. SOC. VOL. 130, NO. 23, 2008