Total synthesis of gambierol by using oxiranyl anions

Article abstract of DOI:10.1002/chem.201000497

Gambierol was isolated as a neurotoxin from the cultured cells of the ciguatera causative dinoflagellate Gambierdiscus toxicus and classified as a member of the polycyclic ether family of marine toxins. The structure consists of a ladder-shaped frans-fuse

Full text of DOI:10.1002/chem.201000497

DOI: 10.1002/chem.201000497
Total Synthesis of Gambierol by Using Oxiranyl Anions
Hiroki Furuta, Yuki Hasegawa, Mariko Hase, and Yuji Mori*[a]
Abstract: Gambierol was isolated as a
methyl-substituted tetrahydropyranyl
rings, and a partially conjugated triene
side chain. The total synthesis of gam-
bierol has been achieved by utilizing
an oxiranyl anion strategy in an itera-
neurotoxin from the cultured cells of
the ciguatera causative dinoflagellate
Gambierdiscus toxicus and classified as
tive manner. Synthetic highlights of
this route include direct carbon–carbon
formation on epoxides, sulfonyl-assist-
ed 6-endo cyclization, and an expansion
reaction of the tetrahydropyranyl rings
to oxepanes to forge the polycyclic ar-
chitecture of the target molecule.
a
member of the polycyclic ether
family of marine toxins. The structure
consists of a ladder-shaped trans-fused
octacyclic ring system that includes 18
stereogenic centers, two 1,3-diaxial di-
Keywords: cyclization · gambierol ·
oxiranyl anions · polycyclic ethers ·
ring expansion
Introduction
ing of dihydrobrevetoxin B to voltage-sensitive sodium
channels[5] has also attracted attention, leading to structure–
activity relationship (SAR) studies.[6] Further evaluation of
its molecular target on the ion channels has revealed that
gambierol inhibits the voltage-gated potassium channels in
mouse taste cells,[7] which might be associated with the taste
alteration caused by ciguatera intoxication, and modulates
ion fluxes by acting as a partial agonist of sodium channels
in human neuroblastoma cells.[8] Moreover, gambierol has
been reported to act as a functional antagonist of neurotoxin
site 5 on voltage-gated sodium channels and to be a subtype
selective voltage-gated potassium channel antagonist that
binds to a novel binding site in cerebellar granule neurons.[9]
Gambierol consists of a ladder-shaped trans-fused octacy-
clic ring system that includes 18 stereogenic centers and a
partially conjugated triene side chain, including a conjugated
Dinoflagellate toxins constitute a large family of com-
pounds, many of which exhibit a wide range of physiological
activities. Accordingly, intense activity has been devoted to
the isolation and structure determination of toxins. Cigua-
tera poisoning is a toxicological syndrome resulting from the
ingestion of seafood contaminated by certain toxic polycy-
ACHTUNGTRENNUNGclic ethers produced by the epiphytic dinoflagellate Gam-
bierdiscus toxicus. The toxin suite comprises ciguatoxins,
maitotoxin, and gambierol.[1] The large complex architecture
of these compounds and their potent neurotoxicity have at-
tracted the attention of chemists and have made them the
focus of numerous synthetic efforts.[2]
Gambierol (1) was isolated as a neurotoxin from the cul-
tured cells of G. toxicus in 1993.[3] The absolute stereochem-
istry was later determined by NMR spectroscopic analysis of
the (S)- and (R)-phenylmethoxy(trifluoromethyl)acetate
(MTPA) derivatives.[4] The toxin exhibits potent toxicity
against mice at LD50 50 mgkgꢀ1 (ip), and its symptoms occur-
ring in mice resemble those shown for ciguatoxins, which in-
dicates that gambierol is also responsible for ciguatera sea-
food poisoning. The ability of gambierol to inhibit the bind-
[a] Dr. H. Furuta, Y. Hasegawa, M. Hase, Prof. Y. Mori
Faculty of Pharmacy, Meijo University
150 Yagotoyama, Tempaku-ku
Nagoya 468-8503 (Japan)
(Z,Z)-diene system. The complex architecture and the need
for biologically active analogues for SAR studies continue
to interest organic chemists. To date, three total syntheses
have been reported. The Sasaki group reported the first
total synthesis based on the B-alkyl Suzuki–Miyaura cross-
coupling strategy.[10] Shortly thereafter, Kadota and co-work-
Fax : (+81)52-834-8090
E-mail: mori@meijo-u.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000497.
7586
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7586 – 7595

FULL PAPER
ers achieved the second synthesis by using the intramolecu-
lar allylation of a-chloroacetoxy ether followed by ring-clos-
ing metathesis.[11] Most recently, the Rainier group complet-
ed the third synthesis, with iterative C-glycoside/enol ether–
olefin metathesis and olefin metathesis/carbonyl olefination
reactions being employed.[12] Moreover, interesting ap-
proaches to the construction of the partial ring system of
gambierol have been designed.[13]
We describe herein a full account of our total synthesis of
gambierol based on our oxiranyl anion coupling method-
ACHTUNGTRENNUNG
ology.[14]
Results and Discussion
Retrosynthetic analysis: Our laboratory has been interested
in the synthesis of bioactive polycyclic ether natural prod-
ucts through use of oxiranyl anions in an iterative manner.
Along these lines, we have developed a unique method for
the formation of tetrahydropyran III by the reaction of trif-
ACHTUNGTRENNUNGlate I with an anion derived from an epoxide followed by a
sulfonyl-assisted 6-endo cyclization of II (Scheme 1).[15] By
Scheme 2. Retrosynthetic analysis of gambierol (1).
Scheme 1. Synthesis of six- and seven-membered rings by an oxiranyl
anion strategy.
active epoxy sulfones 5 and 6 as building blocks. The ad-
vanced BCD-ring fragment 7 was then retrosynthetically
broken by considering the hypothetical polyepoxide 8 to
afford epoxy sulfones, 10 and 11, and triflate 12 as potential
starting materials. To facilitate the gram-scale preparation of
the starting epoxy sulfones, we decided to employ racemic
10 and 11, which could readily be prepared as a mixture of
cis and trans-isomers according to the reported method.[17]
using this approach, we circumvent direct formation of oxe-
pane IV by employing a ring-expansion reaction of III.[16]
Such structural units found in gambierol make it an attrac-
tive target for possible applications of our methodology.
Our synthetic strategy is outlined in Scheme 2. The re-
ported total syntheses have successfully employed the Stille
coupling reaction of vinyl iodide 2 or the corresponding
vinyl bromide with a (Z)-dienyl stannane at the final
step.[10–12] This conversion to gambierol is efficient and relia-
ble from the perspective of constructing the stereochemical-
ly labile conjugated (Z,Z)-diene system. As a result, our
synthetic plan has focused on the construction of vinyl
iodide 2. We envisioned that two seven-membered E and H
rings in 2 would be constructed by a ring-expansion reaction
of tetrahydropyranyl rings at suitable stages of synthesis, so
we tentatively regarded 3 as a hypothetical polyepoxide pre-
cursor for retrosynthetic disassembly. Disconnection of 3 at
the indicated bonds with the aid of sulfonyl groups allowed
for generation of the ABCD-ring diol 4 with the optically
Synthesis of the CD rings: Our synthesis of gambierol com-
menced with the D-ring diol 13, which was prepared from
epoxy sulfone 11 and triflate 12 according to the previously
reported procedure[18] (Scheme 3). Selective triflation of the
primary alcohol of diol 13 with triflic anhydride at ꢀ808C
followed by triethylsilylation of the secondary alcohol with
TESOTf in one pot gave triflate 14 in high yield. Treatment
of a mixture of epoxy sulfone 10 and triflate 14 with nBuLi
in THF/HMPA at ꢀ1008C afforded, after removal of the
TES group, hydroxy epoxy sulfone 15 in 89% overall yield.
As the latter compound is a mixture of four diastereoiso-
mers around the stereochemistry of the epoxide functionali-
Chem. Eur. J. 2010, 16, 7586 – 7595
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7587

Y. Mori et al.
steps from triflate 14. We therefore decided to explore an al-
ternative route requiring the use of epoxy sulfone 25, an ad-
vanced building block with an allylic alcohol moiety prein-
stalled into epoxy sulfone 10, at the second oxiranyl anion
coupling step (Scheme 4). The epoxy sulfone 25 was pre-
pared from the known alcohol 21.[21] Protection of the pri-
mary alcohol as the PMB ether was followed by removal of
the acetonide, and oxidative cleavage of the resulting diol
with NaIO4 provided aldehyde 22. Addition of the anion de-
rived from chloromethyl phenyl sulfoxide and LiHMDS pro-
ceeded in good yield to give chlorohydrin 23 as a diastereo-
isomeric mixture. Cyclization of the chlorohydrin to epoxide
24 by using tBuOK and careful chemoselective oxidation of
the sulfoxide in the presence of a trisubstituted olefin with
MMPP[22] in MeOH at 08C afforded epoxy sulfone 25 as a
mixture of racemic cis and trans-isomers in 57% yield for
the three steps.
Scheme 3. Synthesis of the CD rings—the first approach: a) Tf2O, 2,6-lu-
tidine, CH2Cl2, ꢀ808C, then TESOTf, 94%; b) 10, nBuLi, THF, HMPA,
ꢀ1008C, 90%; c) pTsOH·H2O, CH2Cl2, MeOH, RT, 99%;
d) MgBr2·OEt2, LiBr, CH2Cl2, ꢀ15 to 08C, 99%; e) DBU, CH2Cl2, 08C,
90%; f) AlMe3, CH2Cl2, ꢀ50 to ꢀ208C, 94%; g) TMSOTf, 2,6-lutidine,
CH2Cl2, 08C, 95%; h) H2, Pd(OH)2/C, EtOAc, 94%; i) Dess–Martin peri-
odinane, py, CH2Cl2, RT, 91%; j) 9, toluene, RT, 96%; k) DIBALH,
CH2Cl2,
DIBALH=diisobutylaluminum hydride, HMPA=hexamethylphosphor-
amide, py=pyridine, TES=triethylsilyl, Tf2O=trifluoromethanesulfonic
ꢀ808C,
93%.
DBU=1,8-diazabicycloACTHNGUETRNNU[G 5.4.0]undec-7-ene,
ACHTUNGTRENNUNG
anhydride, TMS=trimethylsilyl.
ty, the C ring was constructed in a stepwise manner accord-
ing to the procedure developed in this laboratory.[19]
Thus, exposure of 15 to MgBr2·OEt2 in the presence of
LiBr provided bromoketone 16 as a 1:1 mixture of diaste-
reoisomers, which was then subjected to intramolecular
etherification with DBU to afford the desired ketone 17 in
90% yield with 95:5 diastereoselectivity. This cyclization has
the advantage that neither the stereochemistry of bromoke-
tone 16, in turn, nor that of epoxy sulfone 15 is relevant, be-
cause the initial cyclization products undergo facile base-cat-
alyzed equilibration to afford a thermodynamically more
stable isomer possessing an equatorial side chain. Stereose-
lective introduction of the b-axial methyl group was accom-
plished by using an excess of AlMe3 (3 equiv) to provide a
tertiary alcohol with 95:5 selectivity,[20] and the alcohol was
protected as a TMS ether to afford 18 in 89% yield for the
Scheme 4. Synthesis of the CD rings—the second approach: a) PMBCl,
NaH, DMF, RT, 91%; b) pTsOH·H2O, MeOH, RT, 93%; c) NaIO4,
MeOH, H2O, RT, 97%; d) chloromethyl phenyl sufoxide, LiHMDS, THF,
ꢀ808C, 87%; e) tBuOK, tBuOH, CH2Cl2, 08C to RT, 88%; f) MMPP,
MeOH, 08C, 75%; g) nBuLi, THF, HMPA, ꢀ1008C, 89%;
h) pTsOH·H2O, CH2Cl2, MeOH, RT, 99%; i) MgBr2·OEt2, CH2Cl2, 4 ꢁ
MS, 08C, 60%; j) DBU, CH2Cl2, 08C, 90%; k) AlMe3, CH2Cl2, ꢀ50 to
ꢀ208C, 90%; l) DDQ, CH2Cl2, H2O, 0 8C, then NaBH4, CeCl3, MeOH,
92%. DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone, LiHMDS=lith-
ium hexamethyldisilazide, PMBCl=p-methoxybenzyl chloride, MMPP=
monoperoxyphtalic acid magnesium salt.
two steps. Removal of the benzyl group with H2/PdACTHNUTRGENUG(N OAc)2
followed by oxidation of the resulting alcohol with Dess–
Martin periodinane yielded aldehyde 19 in 86% overall
yield. The aldehyde was then subjected to the Wittig reac-
tion with 9 to afford an unsaturated ester with high E selec-
tivity (94:6), which was reduced with DIBALH to give a
CD-ring fragment 20 containing an allylic alcohol side chain
in 89% yield for the two steps.
The route described above was effective for reaching the
CD rings but was laborious, especially for the stepwise elab-
oration of the allylic alcohol side chain, which required 10
With the requisite epoxy sulfone 25 in hand, we turned
our attention to the oxiranyl anion coupling reaction with
triflate 14.[23] This was accomplished by using the same con-
ditions employed for 10 to provide 26 in 89% yield. After
removal of the TES group, transformation of 26 to bromo-
7588
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7586 – 7595

Total Synthesis of Gambierol by Using Oxiranyl Anions
FULL PAPER
ketone 27 under the standard conditions (MgBr2·OEt2, LiBr,
CH2Cl2, ꢀ158C) proved to be less effective than the preced-
ing case, yielding 27 in only 47% yield. The major side reac-
tion observed was a 1,2-shift of the sulfonyl group, leading
to the corresponding a-sulfonyl ketone. Efforts to optimize
the yield by changing the reaction conditions to
MgBr2·OEt2/4 ꢁ MS/CH2Cl2/08C improved the yield up to
60%. The subsequent cyclization with DBU (diastereomeric
ratio d.r. 94:6) to ketone 28 and methylation with AlMe3
(d.r. 94:6) proceeded without trouble to afford tertiary alco-
hol 29 in 80% yield for the two steps. Removal of the PMB
group with DDQ gave a 3:2 mixture of an allylic alcohol
and the corresponding unsaturated aldehyde. To avoid
ACHTUNGTRENNUNGhandling the labile aldehyde, the reaction mixture was fur-
ther treated with NaBH4 in one pot, yielding allylic alcohol
30 in 92% yield. Utilization of the advanced epoxy sulfone
25 enabled us to diminish the number of reaction steps from
17 steps in the first approach (34% overall yield) to 13 steps
to reach 30 from triflate 12. However, the overall yield
(18%) of the second approach was lower than we anticipat-
ed. The lengthy synthesis of epoxy sulfone 25, the moderate
yield of bromoketone 27, and the extra reductive workup at
the DDQ oxidation step made the second route impractical
as a means of acquiring a sufficient quantity of 30.
We then devised a new plan based on the modified retro-
synthetic analysis shown in Scheme 5. In this third approach,
we envisaged constructing the C ring by using ketyl radical
Scheme 6. Synthesis of the CD rings—the third approach: a) nBuLi,
THF, HMPA, ꢀ1008C, 94%; b) pTsOH·H2O, CH2Cl2, MeOH, RT, 94%;
c) MgBr2·OEt2, LiBr, CH2Cl2, ꢀ15 to 08C, 94%; d) DBU, CH2Cl2, 08C,
100% (d.r. 94:6); e) NaBH4, MeOH, CH2Cl2, ꢀ808C, 91%; f) methyl
propiolate, NMM, CH2Cl2, RT, 90%; g) PdCl2, CuCl, CH3CN, H2O, RT,
36 (44%), 37 (44%); h) Ac2O, py, DMAP, CH2Cl2, 86%; i) PdCl2, DMF,
H2O, RT, 73 %; j) K2CO3, MeOH, RT, 93%; k) ethyl propiolate, NMM,
CH2Cl2, RT, 97%; l) SmI2, THF, MeOH, 08C, 100%; m) TMSOTf, 2,6-lu-
tidine, 08C, CH2Cl2, 96%; n) DIBALH, CH2Cl2, ꢀ808C, 99%; o) 9, tolu-
ene, RT, 96%; p) DIBALH, CH2Cl2, ꢀ808C, 93%. DMAP=4-dimethyl-
Scheme 5. The third approach for the CD-ring fragment 20.
AHCTUNGTRENNGaUN minopyridine; NMM=N-methylmorpholine.
cyclization of ketone 31. The ketone would be accessed
from hydroxy olefin 32, which could be built up by using an
oxiranyl anion coupling reaction between the unsaturated
epoxy sulfone 33 and triflate 12. The requisite epoxy sulfone
33 was prepared from 4-butenal[24] and chloromethyl phenyl
sulfoxide as a mixture of racemic cis and trans-stereoisomers
by a three-step procedure (Scheme 6). Reaction of the oxir-
anyl anion derived from epoxy sulfone 33 with triflate 12
followed by removal of the TES group afforded hydroxy
epoxy sulfone 34 in 88% overall yield. The reaction of 34
with MgBr2·OEt2 proceeded without any trouble to provide
bromoketones in 94% yield, and the DBU-mediated cycli-
zation afforded ketone 35 quantitatively with 94:6 diastereo-
selectivity. Reduction of the ketone with NaBH4 furnished a
diastereoisomeric mixture of alcohols, from which the de-
sired alcohol 32 was isolated in 91% yield. Transformation
of the double bond to a methyl ketone proved to be more
difficult than anticipated. Wacker oxidation[25] of 32 gave the
desired hydroxy ketone 36 in a disappointing 44% yield.
The byproduct of this oxidation was hemiacetal 37, which
was also obtained in 44% yield. To prevent the undesired
hemiacetal formation, the hydroxy group was protected as a
b-alkoxy acrylate by reaction with methyl propiolate in the
presence of N-methyl morpholine, leading to 38 in 90%
yield. Wacker oxidation of 38, however, resulted in the same
outcome, yielding a 1:1 mixture of 36 and 37, due to the
rapid hydrolysis of the alkoxy acrylate group under acidic
conditions.
These disappointing results forced us to protect the hy-
droxy group with an appropriate protecting group that
would be tolerable under acidic conditions. To this end, the
hydroxy group of 32 was protected as an acetate. When ace-
Chem. Eur. J. 2010, 16, 7586 – 7595
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7589

Y. Mori et al.
tate 39 was subjected to Wacker oxidation under noncatalyt-
ic conditions,[26] the desired ketone 40 was produced in 73%
yield. Removal of the acetyl group followed by installation
of a b-alkoxy acrylate group afforded keto acrylate 31 in
90% overall yield. Treatment of 31 with SmI2 in the pres-
ence of MeOH[27] effected ketyl radical cyclization to afford
quantitatively bicyclic ester 41 as a single stereoisomer.
Elaboration of the ester to allylic alcohol 20 via 19 was car-
ried out in 85% overall yield by a four-step sequence involv-
ing: 1) protection of the tertiary alcohol with a TMS group,
2) reduction of the ester with DIBALH, 3) Wittig olefina-
tion with 9, and, finally, 4) the DIBALH reduction of the
unsaturated ester. This third approach to 20 has a favorable
overall yield of 34% for the 14 steps from 12, compared
with the first and second approaches.
Synthesis of the ABCD rings:[23] With an ample supply of
the CD-ring fragment 20, we then turned our attention to
the construction of the AB ring as shown in Scheme 7. The
B ring containing sterically congested 1,3-diaxial dimethyl
groups was synthesized according to the Nicolaouꢂs proto-
col.[28] Thus, epoxidation of allylic alcohol 20 with mCPBA
proceeded with a high diastereomeric ratio of 96:4 to afford
epoxide 42 in 96% yield. Subsequent oxidation with SO3·py
gave an aldehyde, which was subjected to methylenation to
afford epoxy olefin 43 in 75% overall yield. Removal of the
TMS group with one equivalent of nBu4NF at ꢀ108C fol-
lowed by exposure of the resulting hydroxy epoxide to
PPTS at 08C led to the B-ring vinyl alcohol 44 in 74% yield
over the two steps.
Construction of the A ring is a challenging issue, because
an a-axial hydroxy group at C(6) and a b-equatorial C3 side
chain at C(4) have to be installed stereoselectively. To intro-
duce the C(6) axial hydroxy group, Sasaki and Yamamoto/
Kadotaꢂs syntheses employed reduction of a hydroxy epox-
ide[10] or asymmetric allylation of an aldehyde,[11] in which
the hydroxy epoxide and the aldehyde were prepared from
the vinyl derivatives similar to 44. We envisioned that the
C(6) hydroxyl group and the C3-side chain could be intro-
duced by the reaction of epoxide 45 with 2-(3-benzyloxy-
propyl)-1,3-dithiane (46a). To this end, vinyl alcohol 44 was
subjected to the homoallylic hydroxy-directed epoxidation
with tert-butyl hydroperoxide in the presence of [VO-
Scheme 7. Synthesis of the ABCD-ring fragment 4: a) mCPBA,
NaHCO3, CH2Cl2, 08C, 96% (d.r. 96:4); b) SO3·py, Et3N, DMSO, 08C to
RT; c) Ph3P+CH3·Brꢀ, KHMDS, THF, ꢀ808C to RT, 87% (2 steps);
d) nBu4NF, THF, ꢀ108C, 100%; e) PPTS, CH2Cl2, 74%; f) tBuOOH,
[VOACHTUNTGRNEGN(U acac)2], CH2Cl2, 358C, 85% (d.r. 93:7); g) TESCl, Et3N, DMAP,
CH2Cl2, 96%; h) 46b, nBuLi, THF, 08C, 48 (99%); i) pTsOH·H2O,
MeOH, CH2Cl2, RT, 97%; j) PivCl, DMAP, py, CH2Cl2, 08C, 98%;
k) Hg
G
91%; m) DIBALH, CH2Cl2, ꢀ808C, 98%; n) NaH, BnBr, DMF, 86%;
o) nBu4NF, THF, RT, 97%. KHMDS=potassium hexamethyldisilazide,
mCPBA=meta-chloroperbenzoic acid, PivCl=pivaloyl chloride, PPTS=
pyridinium ptoluenesulfonate, TESCl=chlorotriethylsilane.
A
In this case, the desired hydroxy dithioacetal 48 was ob-
tained in 99% yield. The dithioacetal was then transformed
into b,d-dihydroxy ketone 49 in 85% overall yield in a
three-step sequence: removal of the TES and trityl groups
with pTsOH, protection of the primary alcohol as its piva-
loate, and hydrolysis of the dithioacetal group. The reduc-
tive cycloetherification of hydroxy ketone 49 was best ac-
complished with triethylsilane in the presence of SnCl4 at
ꢀ408C to proceed with complete stereochemical control,
leading to the ABCD-ring fragment 50 containing an equa-
torial side-chain group in 91% yield. Finally, reductive re-
moval of the pivaloyl group followed by dibenzylation of
the C(1) and C(6) hydroxy groups and desilylation afforded
the ABCD-ring diol 4 in 82% yield for the three steps.
ide with 93:7 diastereoselectivity, and the remaining hydroxy
group was protected as a TES ether to provide epoxide 45
in 82% yield over the two steps. Reaction of the epoxide
with the lithiodithiane derived from 46a under standard
conditions (nBuLi, THF, ꢀ208C, 2 h)[30] resulted in only
28% yield of the desired product. The lack of reactivity of
lithiated 46a was found to be due to decomposition of the
lithiodithiane during anion formation, which was confirmed
by the isolation of 2-allyl-1,3-dithiane. When the reaction
was carried out under in situ trapping conditions at 08C, an
unexpected product 47 was obtained in 77% yield.[31] To
avoid this undesired reaction, the trityl ether derivative 46b,
which has no benzylic protons, was reacted with epoxide 45.
7590
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7586 – 7595

Total Synthesis of Gambierol by Using Oxiranyl Anions
FULL PAPER
Synthesis of the ABCDEF rings: Elaboration of the ABCD
rings to the hexacyclic ABCDEF rings required constructing
seven and six-membered ring systems with sterically con-
gested 1,3-diaxial dimethyl groups. To this end, the requisite
building block 5 to be coupled to the ABCD-ring fragment
was prepared from 4-bromo-4-pent-2-enone (51) and (R)-
(ꢀ)-chloromethyl p-tolyl sulfoxide (52) (Scheme 8).[32]
Scheme 8. Preparation of epoxy sulfone 5: a) 52, LiHMDS, THF, ꢀ808C,
then 51, 53 (40%), 54 (40%); b) tBuOK, tBuOH, CH2Cl2, 08C, 55
(90%), 56 (83%); c) mCPBA, CH2Cl2, 08C, 94%. LDA=lithium diiso-
propylamide.
Chlorohydrins 53 and 54 were subjected to exposure to
tBuOK to afford the epoxy sulfoxides 55 and 56, respective-
ly, in good yields.
Scheme 9. Synthesis of the ABCDE-ring system 61: a) Tf2O, 2,6-lutidine,
ꢀ808C, CH2Cl2, then TESOTf, 92%; b) 5, nBuLi, THF, HMPA, ꢀ1008C,
91%; c) pTsOH·H2O, CH2Cl2, MeOH, RT, 91%; d) BF3·OEt2, CH2Cl2,
ꢀ808C, 98%; e) TMSCHN2, BF3·OEt2, CH2Cl2, ꢀ208C, 53%; f) nBu4NF,
THF, AcOH, RT, 97%; g) NaBH4, CH2Cl2, MeOH, ꢀ808C, 59%
(+isomer 41%). TMSCHN2 =trimethylsilyldiazomethane.
The stereochemistry of epoxides 55 and 56 was deter-
mined by 1H HMR spectroscopic analysis: the chemical
shifts of the allylic hydrogen atoms of the desired isomer 55
lay downfield relative to those of the isomer 56 due to the
anisotropic effect of the adjacent sulfinyl group. Oxidation
of the isomer 55 with mCPBA afforded the desired epoxy
sulfone 5 in 85% yield for the two steps. We also attempted
to prepare allylic or propargylic analogues of 5 from 4-pent-
2-enone or 4-pent-2-ynone, respectively. However, the reac-
tions of those ketones with the anion derived from 52 were
unsuccessful due to the facile enolization of the starting ke-
tones.
Construction of the E ring began with the anion coupling
of epoxy sulfone 5 with the ABCD-ring triflate 57 prepared
from diol 4 (Scheme 9). Thus, treatment of a mixture of 5
and 57 with nBuLi at ꢀ1008C followed by desilylation with
pTsOH afforded hydroxy epoxy sulfone 58 in 83% overall
yield. A sulfonyl-assisted 6-endo cyclization was carried out
with BF3·OEt2 to afford ketone 59 quantitatively. The
ketone was then subjected to a ring-expansion reaction with
trimethylsilyldiazomethane[16] in the presence of BF3·OEt2
followed by desilylation of the resulting a-trimethylsilyl
ketone with nBu4NF to furnish the desired seven-membered
ketone 60 in 51% yield for the two steps. Reduction of the
ketone with NaBH4 gave a 3:2 separable mixture of the de-
sired a-alcohol 61 and the undesired b-isomer in 59 and
41% yields, respectively. The latter isomer could be recycled
by oxidation and reduction, allowing the overall yield for
the 60!61 conversion to be raised to 89% after a three-
cycle operation.
At this stage we envisioned two options for forming the F
ring: one was a vinyl radical cyclization route and the other
was a ketyl radical cyclization route (Scheme 10). The vinyl
radical approach proceeded through alkoxy acrylate 62,
which was prepared from 61 by a hetero-Michael reaction
with methyl propiolate in 95% yield. Treatment of 62 with
Bu3SnH in the presence of AIBN in toluene at reflux fur-
nished the olefinic ester 63 in 81% yield along with the de-
brominated product (15%). Dihydroxylation of 63 with
OsO4 provided the diol 64 in 88% yield as a single dia-
AHCTUNGTREGsNNNU tereoisomer. Selective mesylation of the primary alcohol
followed by reduction with LiEt3BH afforded diol 65 con-
taining 1,3-diaxial dimethyl groups in 94% overall yield.[33]
The primary alcohol 66 was then prepared by a two-step
procedure involving bis-triethylsilylation followed by selec-
tive monodesilylation in 89% overall yield.
The second approach to elaborate the F ring by ketyl rad-
ical cyclization began with the dehydrobromination of 61
with nBu4NF in DMF at 608C[34] to afford alkyne 67 in 99%
yield. Hydration of the terminal alkyne with a catalytic
[35]
amount of Hg
G
followed by reaction with methyl
propiolate provided keto acrylate 68 in 67% overall yield.
Treatment of 68 with SmI2 in the presence of MeOH in-
duced reductive cyclization to afford hydroxy ester 69 quan-
Chem. Eur. J. 2010, 16, 7586 – 7595
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7591

Y. Mori et al.
Scheme 10. Synthesis of the ABCDEF-ring triflate 72: a) methyl propiolate, NMM, CH2Cl2, RT, 95%; b) nBu3SnH, AIBN, toluene, 1108C, 81%;
c) OsO4, NMO, THF, H2O, RT, 88%; d) MsCl, Et3N, CH2Cl2, ꢀ108C, 96%; e) LiEt3BH, THF, 08C, 98%; f) TESOTf, 2,6-lutidine, CH2Cl2, 08C, 94%;
g) PPTS, MeOH, CH2Cl2, 08C, 95%; h) nBu4NF·3H2O, DMF, 608C, 99%; i) HgACHTUNGTRNE(UGN OTf)2, TMU, CH2Cl2, MeCN, H2O, RT, 82%; j) methyl propiolate,
NMM, CH2Cl2, RT, 82%; k) SmI2, THF, MeOH, 08C, 100%; l) TESOTf, 2,6-lutidine, CH2Cl2, 08C, 92%; m) DIBALH, CH2Cl2, 08C, 96%; n) o-
NO2PhSeCN, nBu3P, CH2Cl2, py, RT; o) H2O2, NaHCO3, THF, RT, 85% (2 steps); p) OsO4, NMO, THF, H2O, RT; q) NaIO4, MeOH, H2O, RT;
r) NaBH4, CH2Cl2, MeOH, RT, 98% (3 steps); s) pTsOH·H2O, CH2Cl2, MeOH, RT, 100%; t) Tf2O, 2,6-lutidine, CH2Cl2, ꢀ808C, then TMSOTf, 95%.
AIBN=2,2’-azobisisobutyronitrile, TMU=1,1,3,3,-tetramethylurea.
titatively as a single diastereoisomer. After protection of the
hydroxy group as a TES ether, the ester was reduced with
DIBALH to furnish alcohol 66 in 88% yield for the two
steps. The lack of need for the deoxygenation step (64!65)
made the second approach more practical. The elaboration
of alcohol 66 to diol 71 required the one-carbon diminution
of the side chain. Thus, alcohol 66 was converted to olefin
70 via o-nitrophenyl selenide in 82% overall yield.[36] Subse-
quent oxidative cleavage of the double bond was performed
by a three-step procedure involving dihydroxylation with
OsO4, oxidation with NaIO4, and NaBH4 reduction to afford
the primary alcohol. Removal of the TES group of the latter
compound with acid provided diol 71 in 98% overall yield
from 70. Triflation and trimethylsilylation of the diol in one
pot afforded triflate 72 in 95% yield, setting the stage for
the construction of the G and H rings.
yl (TBDPS) group with nBu4NF afforded the diol 75, which
was then subjected to one-pot triflation and silylation to fur-
nish triflate 76 in 91% yield for the three steps. The se-
quence described above was employed iteratively with equal
efficiency to construct the H-ring precursor 78. Thus, the
second coupling reaction of 76 with the oxiranyllithium of 6
provided, after desilylation, epoxy sulfone 77 in 94% yield.
The BF3-promoted cyclization of 77 resulted in the forma-
tion of octacyclic ketone 78 in 83% yield.
Conversion of the ketone 78 to the seven-membered H
ring ketone 79 proceeded smoothly with trimethylsilyldiazo-
methane in the presence of BF3·OEt2 (Scheme 12). We were
pleased to find that this ring enlargement was achieved in a
high 81% yield compared with the case of E-ring formation.
To then complete the H-ring functionalities, it remained to
introduce the C(28)=C(29) double bond and the C(30) terti-
ary alcohol. This transformation was carried out according
to the procedure employed in the previous syntheses.[10,11]
Thus, treatment of ketone 79 with LiHMDS in the presence
of TMSCl and Et3N afforded the corresponding enol silyl
ether, which was subjected to Saegusa oxidation[37] with Pd-
Synthesis of the octacyclic ring system of gambierol: With
the hexacyclic ring system 72 in hand, we then turned our
attention to the construction of the GH rings. Installation of
the G and H rings was based on the use of an oxiranyl
anion strategy (Scheme 11). Treatment of 6[15] with nBuLi in
the presence of triflate 72 at ꢀ1008C afforded epoxy sulfone
73 in 93% yield. Exposure of 73 to BF3·OEt2 caused detri-
methylsilylation and 6-endo cyclization to form the G-ring
ketone 74 in 91% yield. Stereoselective reduction of the
ketone with NaBH4 and removal of the tert-butyldiphenylsil-
AHCTUNGRTEG(NNUN OAc)2 in acetonitrile to provide enone 80 quantitatively.
Methylation with MeMgBr in toluene followed by removal
of the TBDPS group with nBu4NF afforded diol 81 in 93%
yield as a single diastereoisomer. At this point, we decided
to carry out an additional manipulation on the diol to reach
compound 82, an intermediate synthesized by Sasaki et al.[10]
7592
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7586 – 7595

Total Synthesis of Gambierol by Using Oxiranyl Anions
FULL PAPER
Scheme 12. Synthesis of the ABCDEFGH-ring diol 81: a) TMSCHN2,
BF3·OEt2, CH2Cl2, ꢀ808C; b) PPTS, CH2Cl2, MeOH, RT, 81% (2 steps);
c) LiHMDS, TMSCl, Et3N, THF, ꢀ808C; d) Pd
ACHTUNGRTEN(NNUG OAc)2, MeCN, RT, 98%
(2 steps); e) MeMgBr, toluene, ꢀ808C, 93%; f) nBu4NF, THF, RT, 100%;
g) TBSOTf, 2,6-lutidine, CH2Cl2, ꢀ808C, 86%.
yield. Subsequent oxidation of the alcohol with TPAP fur-
nished aldehyde 84, which was subjected to iodomethylena-
tion with iodomethyltriphenylphosphonium iodide[38] and
NaHMDS to afford the (Z)-vinyl iodide 85 in 63% yield for
the last two steps.
Scheme 11. Synthesis of ketone 78: a) nBuLi, THF, HMPA, ꢀ1008C,
93%; b) BF3·OEt2, CH2Cl2, 08C to RT, 91%; c) NaBH4, CH2Cl2, MeOH,
ꢀ808C, 100%; d) nBu4NF, THF, RT, 96%; e) Tf2O, 2,6-lutidine, CH2Cl2,
ꢀ808C, then TESOTf, 95%; f) 6, nBuLi, THF, HMPA, ꢀ1008C, 94%;
g) PPTS, CH2Cl2, MeOH, RT, 100%; h) BF3·OEt2, CH2Cl2, 08C to RT,
83%.
Debenzylation of 85 was now critical for the successful
completion of our route, as it needed to be executed in the
presence of the labile (Z)-vinyl iodide, cyclic allylic ether,
and TES ether functionalities. Upon considerable experi-
mentation, it was finally discovered that gentle heating of 85
with DDQ in the presence of water and diallyl ether in 1,2-
dichloroethane at 508C for 3 h smoothly promoted debenzy-
lation,[39] leading to, after removal of the TES ether, the de-
sired triol 2 in 78% yield. The above reaction conditions are
critical: extending the reaction time or elevated tempera-
tures resulted in decomposition and decreased the yield. Fi-
nally, Stille coupling of triol 2 with dienyl stannane 86[40] was
carried out by using Kadotaꢂs protocol[11] to provide gam-
bierol (1) in 68% yield. The spectroscopic and physical data
for synthetic gambierol were identical to those reported pre-
viously.[10,11]
Selective silylation of the primary hydroxy group as a TBS
ether led to compound 82. Indeed, the 1H and 13C NMR
spectroscopic data and the optical rotation of our synthetic
material matched those reported for 82, thus confirming the
structure and stereochemical assignments of 81.
Completion of the total synthesis: The last remaining chal-
lenge in the total synthesis of gambierol (1) required incor-
poration of the skipped triene side chain. This task was ac-
complished as summarized in Scheme 13. In the work previ-
ously performed by Sasakiꢂs and Kadotaꢂs groups,[10.11] the
robust C(1) and C(6) benzyl groups of 82 and the TBS de-
rivative of 79 were replaced by more easily removable silyl
groups or silyl and pivaloyl groups prior to installation of a
(Z)-vinyl iodide moiety, respectively. However, if debenzyla-
tion of 85 were feasible in the presence of the vinyl iodide
functionality, a straightforward route to reaching 2 would be
made possible. To this end, the primary alcohol 83 was pre-
pared from 81 by a two-step procedure involving bis-silyla-
tion followed by selective monodesilylation in 93% overall
Conclusion
Total synthesis of gambierol has been achieved by utilizing
an oxiranyl anion strategy in an iterative manner. The sali-
ent features of the route include: 1) direct carbon–carbon
bond formation on an oxirane ring and the subsequent sul-
Chem. Eur. J. 2010, 16, 7586 – 7595
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7593

Y. Mori et al.
[2] For recent reviews on the synthesis of marine polycyclic ethers, see:
a) M. Inoue, Chem. Rev. 2005, 105, 4379–4405; b) T. Nakata, Chem.
Rev. 2005, 105, 4314–4347; c) H. Fuwa, M. Sasaki, Curr. Opin. Drug
Discovery Dev. 2007, 10, 784–806.
[3] M. Satake, M. Murata, T. Yasumoto, J. Am. Chem. Soc. 1993, 115,
361–362.
[4] A. Morohashi, M. Satake, T. Yasumoto, Tetrahedron Lett. 1999, 40,
97–100.
[5] M. Inoue, M. Hirama, M. Satake, K. Sugiyama, T. Yasumoto, Toxi-
con 2003, 41, 469–474.
[6] H. Fuwa, N. Kainuma, K. Tachibana, C. Tsukano, M. Satake, M.
Sasaki, Chem. Eur. J. 2004, 10, 4894–4909.
[7] a) V. Ghiaroni, M. Sasaki, H. Fuwa, G. P. Rossini, G. Scalera, T. Ya-
sumoto, P. Pietra, A. Bigiani, Toxicol. Sci. 2005, 85, 657–665; b) V.
Ghiaroni, H. Fuwa, M. Inoue, M. Sasaki, K. Miyazaki, M. Hirama,
T. Yasumoto, G. P. Rossini, G. Scalera, A. Bigiani, Chem. Senses
2006, 31, 673–680.
[8] M. Louzao, E. Cagide, M. R. Vieytes, M. Sasaki, H. Fuwa, T. Yasu-
moto, L. M. Botana, Cell. Physiol. Biochem. 2006, 17, 257–268.
[9] a) E. Cuypers, A. Yanagihara, J. D. Rainier, J. Tytgat, Biochem. Bio-
phys. Res. Commun. 2007, 361, 214–217; b) K. T. LePage, J. D. Rain-
ier, H. W. B. Johnson, D. G. Baden, T. F. Murray, J. Pharmacol. Exp.
Ther. 2007, 323, 174–179; c) E. Cuypers, Y. Abdel-Mottaleb, I.
Kopljar, J. D. Rainier, A. L. Raes, D. J. Snyders, J. Tytgat, Toxicon
2008, 51, 974–983; d) F. Pietra, J. Phys. Org. Chem. 2008, 21, 997–
1001; e) I. Kopljar, A. J. Labro, E. Cuypers, H. W. B. Johnson, J. D.
Rainier, J. Tytgat, D. J. Snyders, Proc. Natl. Acad. Sci. USA 2009,
106, 9896–9901.
[10] a) H. Fuwa, M. Sasaki, M. Satake, K. Tachibana, Org. Lett. 2002, 4,
2981–2984; b) H. Fuwa, N. Kainuma, K. Tachibana, M. Sasaki, J.
Am. Chem. Soc. 2002, 124, 14983–14992.
[11] a) I. Kadota, H. Takamura, K. Sato, A. Ohno, K. Matsuda, Y. Yama-
moto, J. Am. Chem. Soc. 2003, 125, 46–47; b) I. Kadota, H. Taka-
mura, K. Sato, A. Ohno, K. Matsuda, M. Satake, Y. Yamamoto, J.
Am. Chem. Soc. 2003, 125, 11893–11899.
[12] a) H. W. B. Johnson, U. Majumder, J. D. Rainier, J. Am. Chem. Soc.
2005, 127, 848–849; b) U. Majumder, J. M. Cox, H. W. B. Johnson,
J. D. Rainier, Chem. Eur. J. 2006, 12, 1736–1746; c) H. W. B. John-
son, U. Majumder, J. D. Rainier, Chem. Eur. J. 2006, 12, 1747–1753.
[13] a) T. Saito, T. Takeuchi, M. Matsuhashi, T. Nakata, Heterocycles
2007, 72, 151–156; b) T. Saito, T. Nakata, Org. Lett. 2009, 11, 113–
116.
Scheme 13. Completion of the total synthesis: a) TESOTf, 2,6-lutidine,
CH2Cl2, 08C, 95%; b) PPTS, CH2Cl2, MeOH, 08C, 98%; c) TPAP, NMO,
CH2Cl2, RT, 93%; d) Ph3P+CH2I·Iꢀ, NaHMDS, THF, ꢀ808C to RT,
68%; e) DDQ, H2O, diallyl ether, DCE, 508C; f) pTsOH·H2O, CH2Cl2,
MeOH, RT, 78% (2 steps); g) 86, [Pd2ACHTNUGTRENN(GU dba)3]·CHCl3, PACHTGNUTREN(NGNU furyl)3, CuI,
DMSO, 408C, 68%. dba=dibenzylideneacetone, DCE=1,2-dichloro-
ethane. TPAP=tetra-n-propylammonium perruthenate, NMO=4-meth-
ylmorpholine-N-oxide.
[14] For a preliminary communication, see: H. Furuta, Y. Hasegawa, Y.
Mori, Org. Lett. 2009, 11, 4382–4385.
[15] Y. Mori, K. Yaegashi, H. Furukawa, J. Am. Chem. Soc. 1996, 118,
8158–8159.
[16] Y. Mori, K. Yaegashi, H. Furukawa, Tetrahedron 1997, 53, 12917–
12932.
[17] Y. Mori, K. Yaegashi, K. Iwase, Y. Yamamori, H. Furukawa, Tetra-
hedron Lett. 1996, 37, 2605–2608.
[18] Y. Mori, H. Hayashi, J. Org. Chem. 2001, 66, 8666–8668.
[19] Y. Mori, K. Yaegashi, H. Furukawa, Tetrahedron Lett. 1999, 40,
7239–7242.
fonyl-assisted 6-endo cyclization, 2) a ring-expansion ap-
proach to seven-membered ether rings, 3) a successful im-
plementation of debenzylation in the presence of labile
functional groups. Further application of this oxiranyl anion
coupling strategy to other marine polycyclic ethers is in
progress.
[20] a) K. C. Nicolaou, C.-K. Hwang, M. E. Duggan, J. Am. Chem. Soc.
1989, 111, 6682–6690; b) K. C. Nicolaou, D. A. Nugiel, E. Coula-
douros, C.-K. Hwang, Tetrahedron 1990, 46, 4517–4552.
[21] a) K. Miyashita, M. Ikejiri, H. Kawasaki, S. Maemura, T. Imanishi,
Chem. Commun. 2002, 742–743; b) K. Miyashita, M. Ikejiri, H. Ka-
wasaki, S. Maemura, T. Imanishi, J. Am. Chem. Soc. 2003, 125,
8238–8243.
Acknowledgements
This research was supported by a Grant-in-Aid for Scientific Research
on Priority Areas (18032079) and for Scientific Research (C) (21590029)
from the Ministry of Education, Culture, Sports, Science, and Technology.
We thank Professors M. Sasaki (Tohoku Uni) and I. Kadota (Okayama
[22] P. Brougham, M. S. Cooper, D. A. Cummerson, H. Heaney, N.
Tompson, Synthesis 1987, 1015–1016.
1
University) for providing the H and 13C NMR spectra of gambierol.
[23] For a preliminary communication, see: H. Furuta, M. Hase, R.
Noyori, Y. Mori, Org. Lett. 2005, 7, 4061–4064.
[24] L. A. Paquette, T. M. Mitzel, J. Am. Chem. Soc. 1996, 118, 1931–
1937.
[1] For reviews on marine polycyclic ethers, see: a) T. Yasumoto, M.
Murata, Chem. Rev. 1993, 93, 1897–1909; b) M. Murata, T. Yasumo-
to, Nat. Prod. Rep. 2000, 17, 293–314; c) T. Yasumoto, Chem. Rec.
2001, 1, 228–242.
[25] S. Takai, N. Sawada, M. Isobe, J. Org. Chem. 2003, 68, 3225–3231.
7594
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 7586 – 7595

Total Synthesis of Gambierol by Using Oxiranyl Anions
FULL PAPER
[26] T. Mukaiyama, I. Shiina, H. Iwadare, M. Saitoh, T. Nishimura, N.
Ohkawa, H. Sakoh, K. Nishimura, Y. Tani, M. Hasegawa, K.
Yamada, K. Saitoh, Chem. Eur. J. 1999, 5, 121–161.
[27] G. Matsuo, N. Hori, T. Nakata, Tetrahedron Lett. 1999, 40, 8859–
8862.
[28] K. C. Nicolaou, C. V. C. Prasad, P. K. Somers, C.-K. Hwang, J. Am.
Chem. Soc. 1989, 111, 5330–5334.
[29] A. S. Rao in Comprehensive Organic Synthesis, Vol. 7 (Ed.: B. M.
Trost), Pergamon, Oxford, 1991, pp. 357–387.
[35] M. Nishizawa, M. Skwarczynski, H. Imagawa, T, Sugihara, Chem.
Lett. 2002, 12–13.
[36] P. A. Grieco, S. Gilman, M. Nishizawa, J. Org. Chem. 1976, 41,
1485–1486.
[37] Y. Ito, T. Hirao, T. Saegusa, J. Org. Chem. 1978, 43, 1011–1013.
[38] a) D. Seyferth, J. K. Heeren, G. Singh, S. O. Grim, W. B. Hughes, J.
Organomet. Chem. 1966, 5, 267–274; b) G. Stork, K. Zhao, Tetrahe-
dron Lett. 1989, 30, 2173–2174; c) W. R. F. Goundry, J. E. Baldwin,
V. Lee, Tetrahedron 2003, 59, 1719–1729.
[30] D. Seebach, E. J. Corey, J. Org. Chem. 1975, 40, 231–237.
[31] M. Matsushita, Y. Nagaoka, H. Hioki, Y. Fukuyama, M. Kodama,
Chem. Lett. 1996, 1039–1040.
[32] T. Satoh, T. Oohara, Y. Ueda, K. Yamakawa, J. Org. Chem. 1989,
54, 3130–3136.
[39] A. Hamajima, M. Isobe, Angew. Chem. 2009, 121, 2985–2989;
Angew. Chem. Int. Ed. 2009, 48, 2941–2945.
[40] E. Shirakawa, K. Yamasaki, H. Yoshida, T. Hiyama, J. Am. Chem.
Soc. 1999, 121, 10221.
[33] H. Fuwa, M. Sasaki, K. Tachibana, Tetrahedron 2001, 57, 3019.
[34] a) M. Okutani, Y. Mori, Tetrahedron Lett. 2007, 48, 6856–6859;
b) M. Okutani, Y. Mori, J. Org. Chem. 2009, 74, 442–444.
Received: February 25, 2010
Published online: May 18, 2010
Chem. Eur. J. 2010, 16, 7586 – 7595
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
7595