Enantioselective total synthesis of brevetoxin A: Unified strategy for the B, E, G, and J subunits

Article abstract of DOI:10.1002/chem.200900776

Brevetoxin A is a decacyclic ladder toxin that possesses 5-, 6-, 7-, 8-, and 9-membered oxacycles, as well as 22 tetrahedral stereocenters. Herein, we describe a unified approach to the B, E, G, and J rings based upon a ring-closing metathesis strategy from the corresponding dienes. The enolate technologies developed in our laboratory allowed access to the precursor acyclic dienes for the B, E, and G medium-ring ethers. The strategies developed for the syntheses of these four monocycles ultimately provided multigram quantities of each of the rings, supporting our efforts toward the completion of a convergent synthesis of brevetoxin A.

Full text of DOI:10.1002/chem.200900776

FULL PAPER
DOI: 10.1002/chem.200900776
Enantioselective Total Synthesis of Brevetoxin A: Unified Strategy for the B,
E, G, and J Subunits
Michael T. Crimmins,* J. Michael Ellis, Kyle A. Emmitte, Pamela A. Haile,
Patrick J. McDougall, Jonathan D. Parrish, and J. Lucas Zuccarello[a]
Abstract: Brevetoxin A is a decacyclic ladder toxin that possesses 5-, 6-, 7-, 8-, and
9-membered oxacycles, as well as 22 tetrahedral stereocenters. Herein, we describe
a unified approach to the B, E, G, and J rings based upon a ring-closing metathesis
strategy from the corresponding dienes. The enolate technologies developed in our
laboratory allowed access to the precursor acyclic dienes for the B, E, and G
medium-ring ethers. The strategies developed for the syntheses of these four
monocycles ultimately provided multigram quantities of each of the rings, support-
ing our efforts toward the completion of a convergent synthesis of brevetoxin A.
Keywords: asymmetric synthesis ·
chiral auxiliaries · glycolate alkyla-
tion · ring-closing metathesis · total
synthesis
Introduction
ing 22 tetrahedral stereocenters. Brevetoxin A is a highly
toxic metabolite of Karenia brevis, which is known to cause
the infamous red tide phenomenon responsible for massive
fish kills as well as neurotoxic shellfish poisoning and bron-
chial irritation in humans.[3] The bioactivity of 1 is attributed
to strong binding to the a subunit of the voltage-sensitive
sodium ion channels, effecting an increase in the mean chan-
nel open time and inhibiting channel inactivation.[3a] The
landmark total synthesis of 1, reported in 1998 by Nicolaou
and co-workers,[4] stands as the only completed synthesis of
this captivating target.
A fundamental goal in our vision for the total synthesis of
1 was the incorporation of maximum convergency, particu-
larly in three major disconnections. The first disconnection
directed the simplification of the natural product into two
tetracyclic halves of similar complexity (2 and 3), which
would be united in a stereoselective Horner–Wittig reaction
precedented in the previous synthesis by Nicolaou and co-
workers (Scheme 1).[4,5] The Horner–Wittig coupling part-
ners 2 and 3 would be obtained from advanced fragments 4
and 5, respectively, which would be further disconnected
into two subunits each. Specifically, the BCDE fragment 4[6a]
would be prepared from the B and E ring subunits 6 and 7
through a novel convergent [X+2+X] strategy,[1c] and the
GHIJ subunit 5[6b] would be prepared in an analogous way
from the G and J ring subunits 8 and 9. The adjoining manu-
script details the convergent [X+2+X] strategy, as well as
the versatile endgame approach that ultimately led to the
completed total synthesis of 1.[7] Described herein is the de-
velopment of scalable routes to the four ring subunits 6–9,
The marine ecosystem is a source of a multitude of structur-
ally and biologically fascinating molecular metabolites, of
which the ladder ether toxins are some of the most complex
and intriguing small molecules ever discovered. The exqui-
site structures of marine polycyclic ether natural products,
which characteristically contain a linear series of trans-fused
ether rings of varying sizes from five to nine members, with
assorted methyl and hydroxyl substituents appended, have
captured the imagination of synthetic chemists for over two
decades. The development of novel technologies and strat-
egies for the preparation of ladder toxin natural products
has driven the total syntheses of a number of these targets.[1]
As a representative member of this class, the structure of
brevetoxin A (1), which was first elucidated in 1986 by Shi-
mizu and co-workers[2a,b] through X-ray analysis and inde-
pendently by Nakanishi and co-workers through spectro-
scopic studies,[2c] features 10 rings (including 5-, 6-, 7-, 8-,
and 9-membered oxacycles) fused in a linear array contain-
[a] Prof. M. T. Crimmins, Dr. J. M. Ellis, Dr. K. A. Emmitte,
Dr. P. A. Haile, Dr. P. J. McDougall, Dr. J. D. Parrish,
Dr. J. L. Zuccarello
Univeristy of North Carolina at Chapel Hill
Department of Chemistry
Chapel Hill, NC 27599-3290 (USA)
Fax : (+1) 919-962-2388
E-mail: crimmins@email.unc.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200900776.
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
9223

Scheme 2. Initial retrosynthetic analysis of the B ring 10. Bn=benzyl.
tioenriched starting material and the secondary alcohol of
the epoxide was protected as a benzyl ether to give the
known epoxy ether 14 in 91% yield (Scheme 3). The epox-
Scheme 1. Retrosynthetic analysis of 1. TBDPS=tert-butyldiphenylsilyl.
which were prepared through ring-closing metathesis
(RCM)[8] from acyclic diene precursors with stereodefined
a-carbon atoms in the ether linkage. In the case of the B, E,
and G rings, the stereogenicity of the a-carbon atoms would
be established through the implementation of enolate tech-
nology developed in our laboratory.
Scheme 3. First-generation B ring synthesis. Reagents and conditions:
a) (+)-diisopropyl tartrate, [TiACHTUNTGRENUNG(O-iPr)4], tBuOOH, CH2Cl2, 4 ꢁ molecular
Results and Discussion
sieves, À208C, 63%; b) NaH, BnBr, nBu4NI, THF, 91%; c) KCN,
LiClO4, CH3CN, 808C, 84%; d) NaOH, H2O, MeOH, 658C; e) LiAlH4,
Et2O, 35 8C, 78% for 2 steps; f) TIPSCl, imidazole, DMF, 89%; g) NaH,
BrCH2CO2H, THF, 88%; h) PivCl (Piv=pivaloyl), Et3N, (R)-5-lithio-4-
Synthesis of the B Ring: Efforts toward the total synthesis
of 1 commenced with the initial goal of preparing three of
the four medium-ring ethers present in the natural product,
namely, the B, E, and G rings. At the outset, the B ring 10
was targeted for assembly through a glycolate alkylation
strategy to set the stereogenic centers of the ether linkage,[9]
followed by RCM and hydrogenation to create the oxocane
(Scheme 2).[10]
isopropyloxazolidin-2-one, THF, 80%; i) NaN
ACHTUGNTRENN(NUG SiMe3)2, E-ICH2CH=
CHCH2OPMB (16), THF, PhMe, À78 to À458C, 80%; j) LiBH4, MeOH,
Et2O,
0 8C,
96%;
k) (COCl)2,
DMSO,
Et3N,
CH2Cl2;
l) (4-dIcr)2BCH2C(Me)=CH2 (dIcr=isocaranyl), Et2O, THF, À788C,
94% for two steps; m) Ac2O, pyridine, 4-dimethylaminopyridine
(DMAP), CH2Cl2, 92%; n) 2,3-dichloro-5,6-dicyano-1,4-benzoquinone
(DDQ), H2O, CH2Cl2, 08C, 85%; o) (+)-diethyl tartrate (DET), TiACHTUNGTRENNUNG(O-
iPr)4, tBuOOH, CH2Cl2, 4 ꢁ MS, À208C, 93%; p) K2CO3, MeOH, Et2O,
A Sharpless asymmetric epoxidation[11] of 1,4-pentadien-
3-ol (12) conveniently provided an epoxide 13 as an enan-
89%; q) NaIO4, THF, H2O; r) NaBH4, EtOH, 08C, 92% for two steps;
s) [RuACTHUGNTNER(NNUG =CHPh)Cl2ACHTUNRTGEG(NNUN Cy3P)CAHTNGURTEN(UNNG sIMes)], CH2Cl2, 408C, 70%.
9224
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234

Total Synthesis of Brevetoxin A
FULL PAPER
ide 14 was opened at its terminus with a cyanide ion and the
resultant nitrile was transformed to the diol 15 by hydrolysis
of the nitrile and reduction of the resulting acid. The pri-
mary alcohol was selectively protected as its triisopropylsilyl
(TIPS) ether and the secondary alcohol was alkylated with
sodium bromoacetate to produce a glycolic acid that was
converted to its mixed anhydride. The mixed anhydride was
utilized to acylate (R)-lithio-4-isopropyl-2-oxazolidinone,
providing glycolyloxazolidinone 11 and staging an alkylation
to stereoselectively install the C9 substituent.[9] In the event,
alkylation of the sodium enolate of glycolate 11 with iodide
16 afforded diene 17 in 80% yield as a single detectable dia-
stereomer (by 1H NMR spectroscopy). Reductive removal
of the auxiliary, oxidation[12] of the resultant alcohol to the
aldehyde, and Brown asymmetric methallylation[13] of the al-
dehyde provided stereodefined triene 18 in 90% yield for
the three-step sequence. Protection of the secondary alcohol
as an acetate ester and oxidative removal of the p-methoxy-
benzyl (PMB) ether preceded a Sharpless asymmetric epoxi-
dation[11] to provide diene 19 in high yield. RCM was at-
tempted using diene 19 in the presence of the Grubbs
second-generation catalyst [RuACHTNUGTRENN(NGU =CHPh)Cl2ACHTUNTRGEG(NNUN Cy3P)CAHTNUGTRENUN(GN sIMes)]
Scheme 4. Attempted incorporation of the C8 methyl. Reagents and con-
ditions: a) LiBH4, MeOH, Et2O, 0 8C, 96%; b) (COCl)2, DMSO, Et3N,
CH2Cl2; c) MeMgI, Et2O, 0 8C, 90% for 2 steps; d) (COCl)2, DMSO,
Et3N, CH2Cl2; e) (4-dIcr)2BCH2C(Me)=CH2, Et2O, THF, À788C, 89%
for two steps, 10:1 dr; f) DDQ, H2O, CH2Cl2, 08C, 85%; g) (+)-DET, Ti-
(Cy=cyclohexyl, sIMes=1,3-bis(2,4,6-trimethylphenyl)-2-
imidazolidinylidene),[8a,b] but the desired oxocene was not
observed.[14] Aware of precedent involving the facilitation of
medium-ring synthesis through RCM using cyclic con-
straints,[15] we chose to install a tetrahydrofuran as a tempo-
rary cyclic constraint. Base-induced cleavage of the acetate
of diene 19 led to spontaneous cyclization to form a tetrahy-
drofuran and the resultant 1,2-diol was oxidatively cleaved
to provide an aldehyde that was reduced to afford alcohol
20.[16] Gratifyingly, heating diene 20 at reflux in dichloro-
AHCTUNGTRENNUNG
2
N
CHTUNGTRENNUNG
AHCTUNGTRENNUNG
ACHTUNGTRENNUNGmethane with the Grubbs second-generation catalyst for
three days provided the targeted oxocene 21 in 70% yield,
along with 14% of the recovered diene 20.[17]
With this encouraging result, we set out to examine the
stereoselective introduction of the C8 methyl stereocenter.
Oxazolidinone 17 was converted in four straightforward
steps to methyl ketone 22 (Scheme 4). Exposure of ketone
22 to a Brown asymmetric methallylation delivered the terti-
ary alcohol 23 in excellent yield with a 10:1 diastereoselec-
tivity. Cleavage of the PMB ether then produced the re-
quired allylic alcohol 24. The allylic alcohol 24 was pro-
cessed under standard Sharpless conditions leading to in situ
formation of the desired tetrahydrofuran ring. Oxidative
cleavage of the product diol and subsequent reduction of
the intermediate aldehyde gave the alcohol 25. Unfortunate-
ly, unlike diene 20, RCM could not be induced by using
diene 25 or a variety of similar analogues to produce the ox-
ocene 26. This disappointing turn of events led to a reevalu-
ation of the strategy for the B ring synthesis and led to the
investigation of alternative routes.
Scheme 5. Aldol-based retrosynthetic analysis of the B ring 27.
and Brown methallylation to establish the C8 and C9 stereo-
centers in separate steps, we hoped to employ a glycolate
aldol reaction to introduce both centers stereoselectively in
a single transformation.[18] This would also allow the incor-
poration of a 1,3-diol that could be used to introduce the re-
quired temporary tether.
The chlorotitanium enolate of previously prepared glyco-
late 11 was treated with 3-methyl-3-butenal to afford the de-
sired syn product 28 in approximately 70% yield
(Scheme 6). Reduction of the chiral auxiliary and formation
of the cyclohexylidene acetal provided diene 29. With the
cyclic constraint in place, RCM[8a,b] provided oxocene 30 in
83% yield.[19] Replacement of the benzyl ether with a tert-
butyldimethylsilyl ether that would be compatible with the
reduction of the olefin was accomplished in two straightfor-
ward operations. Stereoselective hydrogenation of the C5–
In our revised approach to the B ring, we hoped to incor-
porate the C8 methyl group after the formation of the oxo-
cene (Scheme 5). To accommodate this approach we antici-
pated that a temporary, removable tether would be required
to facilitate the RCM. Rather than use a glycolate alkylation
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9225

M. T. Crimmins et al.
recently developed in our laboratory.[21] The chlorotitanium
enolate of allyloxyglycolylimide 35 was treated with 3-
methyl-3-butenal to provide the alcohol 36 in 64% yield
and 9:1 dr, favoring the desired anti adduct (Scheme 7).[6c]
Removal of the chiral auxiliary and sequential protection of
Scheme 6. Second-generation B ring synthesis. Reagents and conditions:
a) TiCl4, iPr2NEt, CH2=C(Me)CH2CHO, CH2Cl2, À788C; b) LiBH4,
MeOH, Et2O, 0 8C, 58% for 2 steps; c) cyclohexanone, pyridinium p-tol-
uenesulfonate (PPTS), MgSO4, C6H6, 808C, 87%; d) [Ru
ACHTUNGTRNEN(UNG =CHPh)Cl2-
A
ACHTUNGTRENNUNG
91%; f) tert-butyldimethylsilyl trifluoromethanesulfonate (TBSOTf), 2,6-
lutidine, CH2Cl2, 90%; g) PPTS, EtOH, 788C, 72%; h) Ac2O, DMAP,
CH2Cl2, 87%; i) nBu4NF, THF, 100%; j) cyclohexanone, PPTS, MgSO4,
C6H6, 808C, 94%; k) Pd(OH)2/C, H2 (45 psi), EtOH, 90%, 1:1 dr.
C6 alkene was needed to establish the C6 methyl-bearing
stereocenter, but unfortunately the trisubstituted olefin of
oxocene 31 was unreactive under a variety of transition-
metal-catalyzed and transfer-hydrogenation conditions. Re-
moval of the cyclohexylidene acetal provided diol 32, which
was also unreactive toward reductive conditions, including
attempted hydrogenation by using the homogeneous Crab-
tree catalyst.[20] Postulating that the steric encumbrance of
the silyl protecting groups was deterring the reduction of
the olefin, the silyl ethers were replaced with a cyclohexyli-
dene acetal to deliver oxocene 33 in three routine steps.
Indeed, hydrogenation of oxocene 33 was possible in the
presence of the Pearlman catalyst, at elevated pressure, pro-
viding the targeted oxocane 34 in 90% yield; however, the
cyclohexylidene acetal was cleaved during this process and
the oxocane was isolated as a 1:1 diastereomeric mixture of
epimers at C6. It was clear that a new strategy needed to be
implemented for the B ring to allow for stereoselective es-
tablishment of C6 and the tertiary alcohol at C8 present in
1.
Scheme 7. First stereocontrolled completion of the B ring 46. Reagents
and conditions: a) TiCl4, (À)-sparteine, CH2=C(Me)CH2CHO, CH2Cl2,
À788C, 64%, 9:1 dr; b) LiBH4, MeOH, Et2O, 0 8C, 78%; c) TIPSCl, imi-
dazole, DMF, 95%; d) PMBBr, NaH, DMF, 86%; e) TiACHTUNGTRENNUNG(O-iPr)4,
nBuMgCl, Et2O, 0 8C, 86%; f) NaH, BrCH2CO2H, THF, DMF, 91%;
g) PivCl, Et3N, (R)-5-lithio-4-isopropyloxazolidin-2-one, THF, 90%;
h) NaN
MeOH, Et2O, 0 8C, 81%; j) (COCl)2, DMSO, Et3N, CH2Cl2; k) CH2=
CHMgBr, THF, 08C, 76% for 2 steps, 3:1 dr; l) [Ru(=CHPh)Cl2A(Cy3P)-
(sIMes)], CH2Cl2, 408C, 75%; m) [Ir(cod)(PCy3)(pyr)]·PF6, H2, CH2Cl2,
A
A
CHTUNGTRENNUNG
A
E
N
ACHTUNGTRENNUNG
À508C, 93%; n) Dess–Martin periodinane, CH2Cl2, 89%; o) MeMgCl,
Et2O, À788C, 89%; p) lithium di-tert-butylbiphenylilde (LiDBB), THF,
À788C, 98%; q) Dess–Martin periodinane, CH2Cl2, 88%; r) PPh3, p-
NO2C6H4CO2H, diethyl azodicarboxylate (DEAD), C6H6; iBu2AlH,
CH2Cl2, À788C, 50% for 2 steps.
At this stage, we adopted a third-generation approach to
the B ring that utilized the anti-glycolate aldol methodology
9226
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234

Total Synthesis of Brevetoxin A
FULL PAPER
the resultant diol provided diene 37. The allyl protecting
group was selectively removed in the presence of the 1,1-dis-
ubstituted olefin[22] and the resultant alcohol was trans-
formed to the glycolate 38 through the glycolic acid. Alkyla-
tion of the sodium enolate of glycolate 38 with benzyl iodo-
methyl ether (generated in situ from (BnO)2CH2 and trime-
thylsilyl iodide (TMSI)) proceeded in 93% yield and reduc-
tion of the oxazolidinone moiety delivered alcohol 39.[9]
Oxidation[12] of the primary alcohol followed by addition of
vinylmagnesium bromide gave dienes 40 and 41 in a 3:1 dia-
stereomeric ratio. RCM using the Grubbs second-generation
catalyst was possible in the absence of a cyclic constraint to
produce oxocenes 42 and 43, which were separable at this
stage. The minor component (42) of this mixture was trans-
formed to the desired oxocene 43 through a Mitsunobu in-
version.[23] The Crabtree catalyst proved remarkably effec-
tive for the reduction of the trisubstituted olefin of oxocene
43,[20] providing the oxocane 44 and its C6 epimer in a 3:1
ratio at 258C, an 8:1 ratio at À208C and a >19:1 ratio
(93% yield) at À508C. The allylic hydroxyl effectively di-
rects the hydrogenation of the C6–C7 alkene to the opposite
face of the ring.[19] Similar effects are observed in the direct-
ed Simmons–Smith cyclopropanation and directed epoxida-
tion of cyclooctenol.[24,25] Next, oxidation[26] of the C8 hy-
droxyl and subsequent treatment of the ensuing ketone with
methylmagnesium chloride provided the tertiary alcohol 45
in excellent yield and high diastereoselectivity (>19:1 dr).[27]
The benzyl ether of oxocane 45 was reductively removed[28]
in the presence of the more electron-rich PMB ether to
afford the primary alcohol, which was oxidized[26] to deliver
the targeted B ring aldehyde 46.
and incorporation of the C8 methyl group from the third-
generation route.
The sodium enolate of p-methoxybenzyloxyglycolylimide
49 was alkylated with methallyl iodide[9] and the resultant
product 50 underwent an unprecedented Claisen condensa-
tion with the lithium enolate of ethyl acetate to afford the
b-keto ester 51 directly (Scheme 9).[29] Exhaustive reduction
with lithium aluminum hydride followed by selective protec-
tion of the primary alcohol as its TIPS ether and finally oxi-
dation[12] of the secondary alcohol provided ketone 52.[30]
Chelation-controlled reduction of ketone 52 mediated by
Much had been learned regarding the required functional-
ity to facilitate the RCM and hydrogenation of the B ring
during the course of our studies. However, although a viable
approach to B ring aldehyde 46 was in place, a final im-
provement to this route was needed. Installation of C1 at an
early stage of the synthesis was desired to obviate the need
for a late-stage homologation at C2 of the B ring.
For the homologated B ring 47, we envisioned the instal-
lation of C1 through a Claisen addition and planned to then
intercept the previous strategy prior to the glycolate alkyla-
tion (Scheme 8).[6a] This would take advantage of the direct-
ed reduction of the C6=C7 double bond as before and uti-
lize the basic strategy for the construction of the oxocene
Scheme 9. Homologated
a) NaN
(SiMe3)2, CH2 =C(Me)CH2I, THF, À788C, 78%; b) LDA, EtOAc,
THF, À788C, 84%; c) LiAlH4, Et2O, 0 8C, 80%; d) TIPSCl, imidazole,
CH2Cl2, 100%; e) (COCl)2, DMSO, Et3N, CH2Cl2, 100%; f) Zn(BH4)2,
B ring synthesis. Reagents and conditions:
AHCTUNGTRENNUNG
ACHTUNGTRENNUNG
Et2O, À258C, 79%; g) NaH, BrCH2CO2H, THF, DMF, 90%; h) PivCl,
Et3N, (R)-5-lithio-4-isopropyl-oxazolidin-2-one, THF, 90%; i) NaN-
AHCTUNGTRENNUNG
A
N
ACHTUNGTRENNUNG
A
R
ACHTUNGTRENNUNG
o) (COCl)2, DMSO, Et3N, CH2Cl2, 95%; p) MeMgCl, Et2O, À788C,
88%; q) Raney Ni, H2, EtOH, 95%; r) Dess–Martin periodinane,
CH2Cl2, 85%; s) (COCl)2, DMSO, Et3N, CH2Cl2, 64%; t) [CeCl3]·7H2O,
NaBH4, MeOH, 82%.
Scheme 8. Key features of the strategy for the B ring 47.
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9227

M. T. Crimmins et al.
ZnACHTUNGTRENNUNG(BH4)2 delivered the targeted one-carbon homologated
secondary alcohol 53.[31] From this point, a similar sequence
of transformations to those previously developed, with opti-
mizations for large-scale throughput, was employed. Glyco-
late 48 was prepared under standard conditions from alcohol
53 and the alkylation of imide 48 with benzyliodomethyl
ether proceeded efficiently.[9] Reduction of the alkylation
product afforded alcohol 54, which was oxidized to the cor-
responding aldehyde.[12] The aldehyde was exposed to vinyl-
magnesium bromide whereupon RCM again gave a mixture
of oxocenes 55 and 56 (3:1 dr) in just 2 h. Minor oxocene 56
was converted to the desired oxocene 55 through Swern oxi-
dation and Luche reduction.[32] As before, hydrogenation of
the allylic alcohol 55 with the Crabtree catalyst at À508C
provided the oxocane in 94% yield (>19:1 dr). Oxidation[12]
of the C8 alcohol and addition of methylmagnesium chlo-
ride to the resultant ketone gave alcohol 57 as a single ob-
servable diastereomer. Removal of the benzyl ether using
Raney nickel[33] and oxidation[26] of the primary alcohol de-
livered the desired homologated B-ring aldehyde 6 in 81%
yield over two steps. As a testament to the scalability of this
route, more than three grams of aldehyde 6 were prepared
by using this sequence in a single execution of the synthetic
sequence.
Synthesis of the E ring: The initial synthesis of the E ring
fragment 58 was based upon a glycolate alkylation[9] and
RCM strategy to prepare the functionalized oxonene
(Scheme 10).[10] An asymmetric glycolate alkylation and a
Scheme 11. Initial E ring synthesis. Reagents and conditions: a) NaN-
(SiMe3)2, CH2=CHCH2I, THF, À788C, 82%; b) NaBH4, H2O, THF, 95%;
c) (COCl)2, DMSO, Et3N, CH2Cl2; d) propionate 62, TiCl4, (À)-sparteine,
NMP, CH2Cl2, 92% for 2 steps; e) LiBH4, MeOH, Et2O, 0 8C, 91%;
f) TIPSCl, imidazole, DMF, 708C, 96%; g) NaH, BrCH2CO2H, THF,
81%; h) PivCl, Et3N, (R)-5-lithio-4-isopropyloxazolidin-2-one, THF,
Scheme 10. Key strategic elements for the E ring synthesis.
77%; i) NaN
N
thiazolidinethione-mediated aldol addition were envisioned
to establish three of the required stereocenters of the acyclic
diene precursor.
CH2Cl2, 96%; n) m-CPBA, CH2Cl2, À208C, 97%; o) [Ru
AHCTNUTGERGUN(NN PCy3)], CH2Cl2, 408C, 99%; p) HClO4, H2O, THF, 59%; q) NaIO4,
Alkylation of the sodium enolate of glycolate 60 with allyl
iodide, reduction to the primary alcohol, and oxidation
nBu4NHSO4, THF, H2O, 92%; r) LiAlH4, Et2O, 0 8C, 96%; s) NaH,
BnBr, DMF, THF, 408C, 75%; t) nBu4NF, THF, 96%.
under
Swern
conditions
afforded
aldehyde
61
(Scheme 11).[12] Addition of aldehyde 61 to the chlorotitani-
um enolate of (R)-N-propionyl-4-benzylthiazolidin-2-one 62
provided the aldol adduct 59 as a single detectable diaste-
reomer by 1H NMR spectroscopy.[34] Subsequent reductive
removal of the chiral auxiliary provided a 1,3-diol, which un-
derwent selective protection of the primary alcohol as its
TIPS ether to give secondary alcohol 63. Alkylation of the
secondary alcohol with sodium bromoacetate provided the
corresponding glycolic acid derivative, which was converted
to its mixed pivaloyl anhydride and exposed to (R)-5-lithio-
4-isopropyl-oxazolidin-2-one to produce the required glycol-
yl oxazolidinone 64. The C22 stereocenter was established
by alkylation of the glycolyl oxazolidinone with prenyl
iodide under standard conditions[9] to provide imide 65. Re-
duction of the oxazolidinone gave a primary alcohol which
was oxidized[12] to the aldehyde 66. Brown allylation[13] of al-
dehyde 66 gave a triene in 89% yield as a single detectable
9228
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234

Total Synthesis of Brevetoxin A
FULL PAPER
diastereomer, and acetylation of the secondary alcohol af-
forded acetate 67. The trisubstituted olefin 67 was selective-
ly epoxidized followed by exposure of the resultant diene to
the resultant triol was protected as the tris-p-methoxybenzyl
ether. Tetrabutylammonium fluoride mediated removal of
the primary silyl group gave alcohol 74. Although this se-
quence led to diene 74 in a shorter sequence of transforma-
tions, the subunit was lacking C24, which had been present
in our previous synthesis. Hoping to achieve a more practi-
cal synthesis of a homologated E ring, we once again revised
our approach to this medium-ring ether.
In the final plan for the E ring, as with the B ring (see
above, Scheme 9), we hoped to incorporate all of the carbon
atoms of the E ring from the outset, avoiding the need for a
late-stage homologation. Our third-generation approach to
the E ring[6a] commenced with an alkylation of the sodium
enolate of glycolate 49[9] with allyl iodide, followed by re-
ductive removal of the oxazolidinone and oxidation of the
resultant primary alcohol to deliver aldehyde 75
(Scheme 13).[12] A syn-aldol addition of the chlorotitanium
enolate of (R)-N-propionyl-4-benzylthiazolidin-2-thione (76)
to aldehyde 75 proceeded efficiently, generating the aldol
adduct 77.[34] Exposure of the aldol adduct 77 to sodium bo-
the Grubbs catalyst [RuACHTNUTRGENNUG(=CHPh)Cl2ACHTUNGTREN(NGUN Cy3P)2] to effect RCM,
which generated the oxonene 68 in 99% yield.[7] The epox-
ide was exposed to aqueous perchloric acid, resulting in hy-
drolytic opening of the epoxide accompanied by partial
cleavage of the primary silyl ether,[35] requiring conversion
back to the TIPS ether. Oxidative cleavage of the 1,2-diol
provided aldehyde 69. Simultaneous reduction of the alde-
hyde and the acetate with LiAlH4, protection of the resul-
tant diol as the bisACHTUNGTRENNUNG(benzyl) ether, and cleavage of the silyl
ether gave the desired E ring 70 in 96% yield.
Although a viable strategy for synthesis of the E ring had
been developed, we hoped to improve the protecting-group
strategy and eliminate the need for the somewhat cumber-
some oxidative cleavage of the prenyl substituent as a latent
oxygen functionality. As previously demonstrated in the B
ring synthesis (see above, Scheme 7), the anti-glycolate aldol
reaction developed in our laboratory is a useful means for
generating anti-1,2-diol units, which are commonly found in
ladder ether toxins.[21,6c] To that end, previously described al-
cohol 63 was transformed to the oxazolidinethione glycolate
71 (Scheme 12). The chlorotitanium enolate of glycolate 71
underwent an aldol reaction with 3-butenal in 50% yield
(89% based on recovered starting material (brsm) and
5:1 dr to provide the aldol adduct 72.[21] Reduction of the
oxazolidinethione moiety gave alcohol 73. At this stage, the
benzyl ether was cleaved with sodium naphthalenide and
AHCTUNGTREGrNNNU ohydride gave a diol, which upon selective protection of
the primary alcohol delivered the secondary alcohol 78. For-
mation of the glycolic acid and subsequent conversion to the
corresponding imide 79 set the stage for alkylation with bro-
moacetonitrile to diastereoselectively establish the C22 ste-
reocenter.[9] Removal of the auxiliary under reductive condi-
tions then provided alcohol 80 in 75% yield for two steps.
Oxidation[12] of alcohol 80 to the aldehyde preceded a sub-
strate-controlled stereoselective allylation to afford diene 81
in 80% yield (7:1 dr).[36,37] Treatment of diene 81 with hy-
drochloric acid at 658C in methanol served to hydrolyze the
nitrile to the carboxylic acid, which formed the g-lactone in
situ, and also cleaved the silyl and p-methoxybenzyl ethers
in a single operation in 85% yield. The resultant diol was
protected as the bis-TBS ether to give lactone 82. Com-
pound 82 was reduced with lithium aluminum hydride,
whereupon the resulting diol was transformed to the bis-
benzyl ether. Selective deprotection of the primary silyl
ether provided the targeted diene 7, with C24 in place.[38]
Synthesis of the G ring: The first-generation synthesis of the
G ring was inspired by the initial success of the asymmetric
glycolate alkylation/aldol/RCM approach to the E ring. In
the case of the G ring (83, Scheme 14), either an asymmetric
aldol or glycolate alkylation could potentially be used to set
À
the C26 stereocenter. Additionally, a latent C O bond at
C34 would be masked through the regioselective epoxida-
tion of a triene intermediate, followed by RCM. Thus, a key
intermediate would be complex glycolyl oxazolidinone 84,
with the stereocenter at C32 envisioned to arise from a
Sharpless kinetic resolution.
To investigate this approach, glycolyl oxazolidinone 84
was pursued, beginning with the addition of allylmagnesium
bromide to methacrolein (85) (Scheme 15). Resolution of
the resulting racemic mixture of secondary alcohols through
a Sharpless kinetic resolution provided epoxy alcohol 86
(98% ee at 43% conversion),[11b] which was protected as the
Scheme 12. Second-generation E ring synthesis. Reagents and conditions:
a) NaH, BrCH2CO2H, THF, 81%; b) PivCl, Et3N, (S)-5-lithio-4-benzyl-2-
oxazolidinethione, THF, 66%; c) TiCl4, (À)-sparteine, CH2=CHCH2CHO,
CH2Cl2, À788C, 50% (5:1 dr), 89% brsm; d) LiBH4, MeOH, Et2O, 0 8C,
80%; e) Na, naphthalene, THF, 08C, 98%; f) NaH, PMBBr, DMF;
g) nBu4NF, THF, 90% for 2 steps.
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9229

M. T. Crimmins et al.
Scheme 15. First-generation synthesis of the G ring 83. Reagents and con-
ditions: a) CH2=CHCH2MgBr, Et2O, 35 8C, 80%; b) (À)-dicyclohexyl tar-
trate (DCHT), TiACTHUNTRGNE(UNG iPrO)4, tBuOOH, 4 ꢁ molecular sieves, CH2Cl2,
À208C, 43% (98% ee); c) BnBr, NaH, THF, 87%; d) Me2C=CHMgBr,
CuI, THF, 86%; e) BrCH2CO2H, NaH, THF, 87%; f) PivCl, Et3N, THF;
then (S)-5-lithio-4-benzyl-2-oxazolidinone (R1 =Bn) or (S)-5-lithio-4-iso-
propyl-2-oxazolidinone (R1 =iPr), À78 to 08C, 80 or 85%, respectively;
g) TiCl4, (À)-sparteine, acrolein, CH2Cl2, À78 to 08C, 25%; h) NaN-
Scheme 13. Third-generation E ring synthesis. Reagents and conditions:
a) NaN
THF, 93%; c) (COCl)2, DMSO, Et3N, CH2Cl2; d) propionate 76, TiCl4,
(À)-sparteine, NMP, CH2Cl2; e) NaBH4, H2O, THF, 78% for 3 steps;
f) TIPSCl, imidazole, CH2Cl2, 97%; g) NaH, BrCH2CO2H, THF, DMF,
92%; h) PivCl, Et3N, (R)-5-lithio-4-isopropyloxazolidin-2-one, THF,
(SiMe3)2, CH2=CHCH2I , THF, À788C, 76%; b) NaBH4, H2O,
E
H2O, 88%; j) Dess–Martin periodinane, CH2Cl2; k) CH2=CHMgBr, THF,
À788C, 80% for 2 steps; l) m-CPBA, CH2Cl2, À208C, 80%; m) [Ru(=
CHPh)Cl2A(PCy3)(sIMes)], CH2Cl2, 408C, 85%.
ACHTUNGTRENNUNG
C
ACHTUNGTRENNUNG
89%; i) NaN
ACHTUNGTRENNUNG
2
steps; k) (COCl)2, DMSO, Et3N, CH2Cl2; l) CH2=CHCH2SnACHTGNUTRENNUNG
AlMe3, CH2Cl2, À788C; m) HCl, MeOH, 658C, 69% for
3
benzyl ether 87. Copper-assisted Grignard addition to form
secondary alcohol 88, followed by O-alkylation with sodium
bromoacetate, produced glycolic acid 89. The glycolic acid
was first coupled to (S)-4-benzyl-2-oxazolidinone to examine
the aldol addition of the chlorotitanium enolate of imide
84a to acrolein, which would install both the C26 stereocen-
ter and the olefin required for subsequent RCM. Despite ef-
forts toward optimization, only a 25% yield of the desired
aldol adduct 90 could be obtained.
n) TBSCl, imidazole, DMF, 808C, 80%; o) LiAlH4, Et2O, À208C, 88%;
p) NaH, BnBr, DMF, 98%; q) HF·pyridine, THF, 72%.
On the other hand, coupling of glycolic acid 89 with (S)-4-
isopropyl-2-oxazolidinone allowed for the efficient alkyla-
tion of imide 84b with BnOCH2I to provide alkylation prod-
uct 91.[9] Reductive removal of the auxiliary with NaBH4
generated the primary alcohol, which underwent oxidation
by using Dess–Martin conditions to afford aldehyde 92. Ad-
dition of vinylmagnesium bromide to the aldehyde gave an
80% yield of the triene as a mixture of diastereomers, and
selective epoxidation of the trisubstituted olefin with m-
Scheme 14. Key features of the first-generation synthesis of the G ring
83.
9230
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234

Total Synthesis of Brevetoxin A
FULL PAPER
chloroperbenzoic acid (m-CPBA) gave epoxy diene 93.
Treatment of diene 93 with the Grubbs second-generation
catalyst then led to an 85% yield of oxocene 83.[8a,b] At-
tempts to perform the RCM with the Grubbs first-genera-
tion catalyst led to a lower yield (50%).
The first-generation synthesis of the G ring provided a
preliminary route to an important intermediate in the breve-
toxin synthesis and further demonstrated the utility of the
glycolate alkylation/RCM approach. However, the multi-
gram quantities of G ring needed for further development
of the total synthesis motivated the investigation of a
second-generation route optimized for scale up. In particu-
lar, a chelation-controlled Grignard addition leading to N-
glycolyl oxazolidinone 95 was envisioned as a more efficient
alternative to the Sharpless kinetic resolution for installation
of the C32 stereocenter (Scheme 16).
dition of lithiated tBuOAc to aldehyde ent-75 provided alco-
hols 97 as an inconsequential mixture of diastereomers and
exposure to LiAlH4 led to diols 98. Although this five-step
sequence was high yielding overall, our observation from
the B ring synthesis that N-glycolyl oxazolidinones readily
participate in Claisen addition reactions with lithiated esters
provided a means to streamline the preparation of diols 98
(see above, Scheme 9).[6a] The addition of the lithium eno-
late of tBuOAc to alkylation product 96 proceeded smooth-
ly to produce b-ketoester 99 in 85% yield. The reduction of
b-ketoester 99 with LiAlH4 was complicated somewhat by
the presence of enol tautomers, which led to enolate inter-
mediates that were resistant to reduction. Nonetheless, a
67% yield of diols 98 based on recovered starting material
was reproducible.
Upon selective protection of the primary alcohol of diols
98 as the TIPS ether and oxidation to the ketone under
Swern conditions (Scheme 18), the chelation-controlled ad-
dition of methylmagnesium chloride to ketone 100 afforded
tertiary alcohol 101 as a single isomer in excellent yield.
After O-alkylation with sodium bromoacetate and coupling
of the resulting glycolic acid with (S)-4-isopropyl-2-oxazoli-
Scheme 16. Key features of the second-generation synthesis of the G
ring.
The second-generation route to the G ring commenced
with an asymmetric alkylation of imide ent-49 (Scheme 17)
with allyl iodide.[9] In an earlier report,[6b] the alkylation
product 96 was treated with LiBH4 for reductive removal of
the auxiliary and the resultant alcohol was oxidized under
Swern conditions to the corresponding aldehyde ent-75. Ad-
Scheme 18. Second-generation synthesis of the G ring 8. Reagents and
conditions: a) TIPSCl, imidazole, CH2Cl2, 99%; b) (COCl)2, DMSO, di-
isopropyl ethylamine (DIEA), CH2Cl2, À78 to 08C, 94%; c) MeMgCl,
Et2O, À788C, 97%; d) BrCH2CO2H, NaH, DMF, 30 h, 84% (after 1 recy-
cle); e) PivCl, Et3N, THF; then (S)-5-lithio-4-isopropyl-2-oxazolidinone,
À78 to 08C, 86%; f) NaN
g) LiBH4, MeOH, Et2O, 0 8C, 70% for 2 steps; h) (COCl)2, DMSO, Et3N,
CH2Cl2, À78 to 08C, 99%; i) CH2 =CHMgBr, THF, 08C, 80%; j) [Ru(=
CHPh)Cl2A(PCy3)(sIMes)], CH2Cl2, 408C, 85%; k) o-
NO2C6H4SO2NHNH2, Et3N, dimethoxyethane (DME), 858C, 95%;
l) (COCl)2, DMSO, Et3N, CH2Cl2, À78 to 08C, 96%; m) iBu2AlH,
CH2Cl2, À788C, 98%.
G
Scheme 17. Preparation of diols 98. Reagents and conditions: a) NaN-
(SiMe3)2, CH2=CHCH2I, THF, À78 to À458C, 80%; b) LiBH4, MeOH,
N
ACHTUNGTRENNUNG
Et2O, 0 8C, 90%; c) (COCl)2, DMSO, Et3N, CH2Cl2, À78 to 08C, 97%;
d) tBuO2CCH2Li, THF, À788C, 85% from 96, 75% from ent-75;
e) LiAlH4, Et2O, 0 8C, 92% from 97, 67% from 99 brsm.
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9231

M. T. Crimmins et al.
dinone, alkylation of the N-acyloxazolidinone 102 with
BnOCH2I once again performed well, even on larger scale:
over 100 g of alkylation product 103 has been prepared.
Similar to before, reductive removal of the chiral auxiliary
from alkylation product 103, oxidation of the resulting alco-
hol under Swern conditions, and addition of vinylmagnesium
bromide to aldehyde 104 provided the allylic alcohols as a
mixture of diastereomers. Subsequent exposure of the diene
to the Grubbs second-generation catalyst gave oxocene 94
in high yield at up to 10 mm concentration.[8a,b]
ACHTUNGTRENNUNG
To properly functionalize the G ring, we had originally re-
ported[6b,d] the use of the Crabtree catalyst[20] under a hydro-
gen atmosphere for the hydrogenation of the endocyclic
olefin. However, upon scale up, these conditions led to a
somewhat disappointing 70% yield of the desired oxocanes
105, with loss of the PMB protecting group and epimeriza-
tion at C26 by alkene isomerization accounting for undesira-
ble portions of the mass balance. Alternatively, hydrogena-
[39]
tion with o-NO2C6H4SO2NHNH2 and Et3N was found to
cleanly afford the oxocanes 105 in 96% yield without any
significant side reactions. For convergence of the C27 diaste-
reomers, oxidation to the ketone under Swern conditions
followed by treatment with iBu2AlH provided the alcohol 8
as a single isomer. This completed the second-generation
synthesis of G ring intermediate 8 in 16 steps, with an over-
all yield of 12%.
Scheme 20. First-generation Synthesis of the J Ring 106. Reagents and
conditions: a) TiCl4, DIEA, CH2Cl2, À788C; then trans-cinnamaldehyde
or trans-2-hexenal, 76% (9:1 dr) or 64% (7:1 dr), respectively;
b) NaBH4, H2O, THF, 80% (R=Ph), 82% (R=nPr); c) TIPSCl, imida-
zole, DMF, 92% (R=Ph), 73% (R=nPr); d) BrCH2CO2H, NaH, THF,
83% (R=Ph), 87% (R=nPr); e) PivCl, Et3N, THF; then (R)-5-lithio-4-
isopropyl-2-oxazolidinone, À78 to 08C, 80% (R=Ph), 67% (R=nPr);
Synthesis of the J ring: In formulating a synthetic plan for
the J ring 106, it was reasoned that a RCM approach would
be of particular interest because the dihydropyran produced
would be the required substrate for dihydroxylation to in-
stall the necessary syn diol (Scheme 19). In our initial ap-
f) NaN
90% (R=nPr); g) NaBH4, H2O, THF, 86%; h) BnBr, NaH, THF, 99%;
i) [Ru(=CHPh)Cl2A(PCy3)2], CH2Cl2, 408C, 95%; j) OsO4, N-methylmor-
(SiMe3)2, CH2=CHCH2I, THF, À78 to À458C, 28% (R=Ph),
A
CHTUNGTRENNUNG
pholine N-oxide (NMO), THF, H2O, 78% (3:1 dr).
would undergo a RCM, the phenyl substituent on the olefin
was inconsequential. Reductive removal of the auxiliary
with NaBH4 and selective protection of the resultant pri-
mary alcohol 110a produced TIPS ether 111a. After O-alky-
lation of the secondary alcohol with sodium bromoacetate
and coupling of the corresponding glycolic acid to (R)-4-iso-
propyl-2-oxazolidinone, the alkylation of imide 112a was at-
tempted.[9] Unfortunately, alkylation with allyl iodide took
place in only 28% yield, with the major side reaction involv-
ing 2,3-Wittig rearrangement of the sodium enolate.[42]
To circumvent the low yield of the glycolate alkylation, it
was hypothesized that replacement of the phenyl group on
the olefin with an aliphatic chain would reduce the rate of
the competing 2,3-Wittig rearrangement. To this end, trans-
2-hexenal was used as the aldehyde in the acetate aldol re-
action (108 to 109b, Scheme 20). In this case, a 64% yield
of the aldol adduct was obtained, with a slightly diminished
diastereoselectivity (7:1). After the same four-step sequence
as before, the glycolate alkylation was attempted once
again. Pleasingly, a 90% yield of the alkylation product
113b was obtained.
Scheme 19. Key features of the first-generation synthesis of the J ring
106.
proach to the diene substrate 107 for RCM, an asymmetric
acetate aldol and a glycolate alkylation were envisioned as
À
key stereodetermining and C C bond-forming steps.
The original route to the J ring drew upon interest in our
laboratory on aldol reactions involving the titanium enolates
of N-acetylthiazolidinethione auxiliaries. Though we have
recently shown that the mesityl-substituted auxiliary of this
type undergoes aldol reaction with a variety of aldehydes in
high yield and stereoselectivity,[40] at the outset of the J ring
synthesis, the isobutyl-substituted N-acetylthiazolidinethione
auxiliary was known to provide reasonable selectivity levels
through the Urpꢂ protocol, particularly with a,b-unsaturated
aldehydes.[10b,41] Specifically, the aldol reaction of N-acetyl-
Conversion of alkylation product 113b to J ring 107 was
then accomplished in four additional steps. Upon reductive
9232
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234

Total Synthesis of Brevetoxin A
FULL PAPER
removal of the chiral auxiliary, the resulting primary alcohol
was protected as benzyl ether 114. As expected, RCM with
the Grubbs catalyst afforded the dihydropyran in excellent
yield,[8] and Upjohn dihydroxylation[43] occurred in 78%
overall yield with a diastereoselectivity of 3:1 to produce J
ring 106. The facial bias is presumably imparted by the
bulky TIPS group. Alternatively, the Sharpless asymmetric
dihydroxylation protocol[44] gave slightly improved selectivi-
ty (4:1), but in only 60% overall yield.
The first-generation approach to the J ring confirmed the
effectiveness of the RCM/dihydroxylation sequence for
completing the ring core, but was limited in terms of
throughput by the moderate yield and tedious separation of
diastereomers associated with the acetate aldol. Further-
more, the C43 and C44 carbon atoms would need to be
added through subsequent manipulations, further lengthen-
ing the synthesis. Thus, a second-generation synthesis target-
ing J ring 115 possessing the C42–C44 side chain was formu-
lated (Scheme 21). It was also postulated that the acetate
Scheme 22. Second-generation synthesis of the J ring 115. a) TBDPSCl,
imidazole, CH2Cl2, 08C to RT, 95%; b) Me3S+IÀ, BuLi, THF, 08C, 86%;
c) (EtO)2CHCH2CH=CH2, PPTS, C6H6, 608C, 35 mm Hg, 87%; d) [Ru
ACHTUNGTRENNUNG(=
CHPh)Cl2A(PCy3)2], CH2Cl2, 408C, 95%; e) RuCl3, NaIO4, H2O, EtOAc,
CTHUNGTRENNUNG
CH3CN, 75%; f) CDI, THF, 658C, 90%; g) CH2=CHCH2SiMe3, trime-
thylsilyl trifluoromethanesulfonate (TMSOTf), CH3CN, À108C, 87%
(dr>10:1); h) K2CO3, MeOH, 76%; i) PPTS, (MeO)2CMe2, 92%; j) 9-
BBN, THF; NaOH, H2O2 (aq.), 90%.
Scheme 21. Key features of the second-generation synthesis of the J ring
115.
Conclusion
aldol and glycolate alkylation steps could be replaced by an
epoxide opening and a Hosomi–Sakurai reaction, respective-
ly. This approach would involve the mixed acetal intermedi-
ate 116, which could be rapidly prepared through transace-
talization.
Scalable approaches to the B, E, G, and J rings of 1 have
been developed, allowing the preparation of multigram
quantities of each of these oxacycles. RCM was utilized in
each case for the cyclization event from the corresponding
acyclic diene. Enolate methodologies developed in our labo-
ratory were exploited to introduce eight of the 22 stereocen-
ters present within 1. The routes described have proven suit-
able for supplying sufficient quantities of the B, E, G, and J
rings to support our efforts toward a highly convergent syn-
thesis of the decacyclic natural product, the completion of
which[48] is reported in the accompanying manuscript.[7]
To begin the second-generation synthesis of the J ring,
known alcohol 118[45] was prepared in two steps from (R)-
glycidol (117) (Scheme 22). The smooth conversion of alco-
hol 118 to mixed acetal 116 was then accomplished through
a transacetalization reaction under mild conditions. RCM
with the Grubbs catalyst proceeded in excellent yield as
before, and the resulting dihydropyran was dihydroxylated
stereoselectively with the RuCl3/NaIO4 oxidation system
(8:1 dr),[46] with the bulky silyl protecting group once again
providing the presumed facial bias. Protection of diol 119 as
the carbonate using 1,1’-carbonyldiimidazole (CDI) preced-
ed the key Hosomi–Sakurai reaction,[47] which delivered
pyran 121 in good yield and high diastereoselectivity (dr>
10:1) via putative transition state 120. Anticipating the need
for a more robust diol protecting group, the carbonate
group was removed under basic conditions and the diol was
reprotected as acetonide 122. Hydroboration of the terminal
Acknowledgements
Financial support of this work by the National Institute for General Med-
ical Sciences (GM60567) is acknowledged with thanks.
[1] a) K. C. Nicolaou, M. O. Frederick, R. J. Aversa, Angew. Chem.
2008, 120, 7292; Angew. Chem. Int. Ed. 2008, 47, 7182; b) H. Fuwa,
M. Sasaki, Curr. Opin. Drug Discovery Dev. 2007, 10, 784; c) M.
Inoue, Chem. Rev. 2005, 105, 4379; d) T. Nakata, Chem. Rev. 2005,
105, 4314; e) I. Kadota, Y. Yamamoto, Acc. Chem. Res. 2005, 38,
423; f) M. Sasaki, H. Fuwa, Synlett 2004, 1851; g) P. A. Evans, B. De-
louvriꢃ, Curr. Opin. Drug Discovery Dev. 2002, 5, 986; h) F. P.
Marmsꢄter, F. G. West, Chem. Eur. J. 2002, 8, 4346; i) Y. Mori,
olefin with 9-borabicycloACTHNURTGNEU[GN 3.3.1]nonane (9-BBN) then deliv-
ered primary alcohol 115 exclusively. In testament to the
scalability of this route, over 21 g of alcohol 115 was pre-
pared through this sequence.
Chem. Eur. J. 2009, 15, 9223 – 9234
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
9233

M. T. Crimmins et al.
Chem. Eur. J. 1997, 3, 849; j) E. Alvarez, M.-L. Candenas, R. Pꢃrez,
J. L. Ravelo, J. D. Martꢂn, Chem. Rev. 1995, 95, 1953.
Doi, K. Shimawaki, Y. Ohtaniuchi, A. Takemura, S.-i. Souma, A.
Murai, Synlett 2002, 1496.
[2] a) Y. Shimizu, H. N. Chou, H. Bando, G. Vanduyne, J. C. Clardy, J.
Am. Chem. Soc. 1986, 108, 514; b) Y. Shimizu, H.-N. Chou, H.
Bando, G. Van Duyne, J. C. Clardy, J. Chem. Soc. Chem. Commun.
1986, 1656; c) J. Pawlak, M. S. Tempesta, J. Golik, M. G. Zagorski,
M. S. Lee, K. Nakanishi, T. Iwashita, M. L. Gross, K. B. Tomer, J.
Am. Chem. Soc. 1987, 109, 1144.
[3] a) S. Michelliza, W. M. Abraham, H. M. Jacocks, T. Schuster, D. G.
Baden, ChemBioChem 2007, 8, 2233; b) J. P. Naar, L. J. Flewelling,
A. Lenzi, J. P. Abbott, A. Granholm, H. M. Jacocks, D. Gannon, M.
Henry, R. Pierce, D. G. Baden, J. Wolny, J. H. Landsberg, Toxicon
2007, 50, 707; c) B. Kirkpatrick, L. E. Fleming, L. C. Backer, J. A.
Bean, R. Tamer, G. Kirkpatrick, T. Kane, A. Wanner, D. Dalpra, A.
Reich, D. G. Baden, Harmful Algae 2006, 5, 526; d) L. E. Fleming,
L. C. Backer, D. G. Baden, Environ. Health Perspect. 2005, 113, 618,
and references therein.
[20] The Crabtree catalyst is known to be highly effective for hindered
olefins with allylic and homoallylic alcohols, see: R. H. Crabtree, H.
Felkin, T. Fillebeen-Khan, G. E. Morris, J. Organomet. Chem. 1979,
168, 183.
[21] M. T. Crimmins, P. J. McDougall, Org. Lett. 2003, 5, 591.
[22] J. Lee, K. Cha, Tetrahedron Lett. 1996, 37, 3663.
[23] O. Mitsunobu, M. Yamada, T. Mukaiyama, Bull. Chem. Soc. Jpn.
1967, 40, 935.
[24] A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307.
[25] Oxocene 42 was also subjected to hydrogenation by using the Crab-
tree catalyst and provided the opposite configuration at C6.
[26] D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48, 4155.
[27] The configuration of the stereocenters at C4, C6, and C8 were estab-
1
lished by using NOESY two-dimensional H NMR spectroscopic an-
alyis.
[4] a) K. C. Nicolaou, Z. Yang, G.-Q. Shi, J. L. Gunzner, K. A. Agrios,
P. Gꢄrtner, Nature 1998, 392, 264; b) K. C. Nicolaou, M. E. Bunnage,
D. G. McGarry, S. Shi, P. K. Somers, P. A. Wallace, X.-J. Chu, K. A.
Agrios, J. L. Gunzner, Z. Yang, Chem. Eur. J. 1999, 5, 599; c) K. C.
Nicolaou, P. A. Wallace, S. Shi, M. A. Ouellette, M. E. Bunnage,
J. L. Gunzner, K. A. Agrios, G.-q. Shi, P. Gꢄrtner, Z. Yang, Chem.
Eur. J. 1999, 5, 618; d) K. C. Nicolaou, G.-q. Shi, J. L. Gunzner, P.
Gꢄrtner, P. A. Wallace, M. A. Ouellette, S. Shi, M. E. Bunnage,
K. A. Agrios, C. A. Veale, C.-K. Hwang, J. Hutchinson, C. V. C.
Prasad, W. W. Ogilvie, Z. Yang, Chem. Eur. J. 1999, 5, 628; e) K. C.
Nicolaou, J. L. Gunzner, G.-q. Shi, K. A. Agrios, P. Gꢄrtner, Z.
Yang, Chem. Eur. J. 1999, 5, 646.
[28] a) P. K. Freeman, L. L. Hutchinson, J. Org. Chem. 1980, 45, 1924;
b) R. E. Ireland, M. G. Smith, J. Am. Chem. Soc. 1988, 110, 854.
[29] Typically, an oxazolidinone is converted to the corresponding ester
or acid chloride that is then utilized in a Claisen condensation.
[30] Stereoselective reductions of b-keto esters to b-hydroxy esters using
ZnACTHUNGRTNEUNG(BH4)2 or Noyori hydrogenation conditions are known; however,
efforts to carry out these types of transformations on b-keto ester 51
proved unsuccessful.
[31] T. Oishi, T. Nakata, Acc. Chem. Res. 1984, 17, 338.
[32] A. L. Gemal, J. L. Luche, J. Am. Chem. Soc. 1981, 103, 5454.
[33] K. Horita, T. Yoshioka, T. Tanaka, Y. Oikawa, O. Yonemitsu, Tetra-
hedron 1986, 42, 3021.
[5] a) L. Horner, H. Hoffmann, H. G. Wippel, G. Klahre, Chem.Ber.
1959, 92, 2499; b) A. D. Buss, S. Warren, J. Chem. Soc. Perkin Trans.
1 1985, 2307, and references therein.
[34] a) M. T. Crimmins, B. W. King, E. A. Tabet, K. Chaudhary, J. Org.
Chem. 2001, 66, 894; b) M. T. Crimmins, K. Chaudhary, Org. Lett.
2000, 2, 775.
[6] a) M. T. Crimmins, P. J. McDougall, J. M. Ellis, Org. Lett. 2006, 8,
4079; b) M. T. Crimmins, J. L. Zuccarello, P. A. Cleary, J. D. Parrish,
Org. Lett. 2006, 8, 159; c) M. T. Crimmins, P. J. McDougall, K. A.
Emmitte, Org. Lett. 2005, 7, 4033; d) M. T. Crimmins, P. A. Cleary,
Heterocycles 2003, 61, 87.
[35] The hydrolysis gave 59% of the diol and 27% of the triol (resulting
from primary ether cleavage). The triol could be reprotected in
74% yield to merge the material.
[36] The stereoinduction in this transformation is predictable based upon
a non-chelated Felkin–Anh model. Allyltrimethylsilane was unreac-
tive in the presence of trimethylaluminum. Allylmagnesium bromide
was non-selective (1:1 dr). Brown allylation provided the product in
>19:1 dr, but in only 30% yield.
[37] a) M. T. Crimmins, M. G. Stanton, S. P. Allwein, J. Am. Chem. Soc.
2002, 124, 5958; b) G. E. Keck, D. E. Abbott, Tetrahedron Lett.
1984, 25, 1883.
[38] 16% of the corresponding diol resulting from cleavage of both silyl
ethers was also obtained and could be recycled to the bis-silyl ether
in 95% yield.
[39] A. G. Myers, B. Zheng, M. Movassaghi, J. Org. Chem. 1997, 62,
7507; M. Movassaghi, J. Org. Chem. 1997, 62, 7507.
[40] M. T. Crimmins, M. Shamszad, Org. Lett. 2007, 9, 149.
[41] a) Y. Nagao, Y. Hagiwara, T. Kumagai, M. Ochiai, T. Inoue, K. Ha-
shimoto, E. Fujita, J. Org. Chem. 1986, 51, 2391; b) A. Gonzalez, J.
Aiguade, F. Urpi, J. Vilarrasa, Tetrahedron Lett. 1996, 37, 8949;
c) M. T. Crimmins, P. Siliphaivanh, Org. Lett. 2003, 5, 4641.
[42] T. Nakai, K. Mikami, Chem. Rev. 1986, 86, 885.
[43] V. VanRheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976, 17,
1973.
[7] M. T. Crimmins, J. L. Zuccarello, P. J. McDougall, J. M. Ellis, Chem.
Eur. J. 2009, 9235–9244.
[8] a) T. M. Trnka, R. H. Grubbs, Acc. Chem. Res. 2001, 34, 18; b) M.
Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1, 953;
c) R. H. Grubbs, S. Chang, Tetrahedron 1998, 54, 4413; d) P. Schwab,
M. B. France, J. W. Ziller, R. H. Grubbs, Angew. Chem. 1995, 107,
2179; Angew. Chem. Int. Ed. Engl. 1995, 34, 2039.
[9] M. T. Crimmins, K. A. Emmitte, J. D. Katz, Org. Lett. 2000, 2, 2165.
[10] a) M. T. Crimmins, K. A. Emmitte, Synthesis 2000, 899; b) M. T.
Crimmins, K. A. Emmitte, Org. Lett. 1999, 1, 2029.
[11] a) M. Nakatsuka, J. A. Ragan, T. Sammakia, D. B. Smith, D. E. Ueh-
ling, S. L Schreiber, J. Am. Chem. Soc. 1990, 112, 5583; b) Y. Gao,
R. M. Hanson, J. M. Klunder, S. Y. Ko, H. Masamune, K. B. Sharp-
less, J. Am. Chem. Soc. 1987, 109, 5765.
[12] A. J. Mancuso, S. Huang, D. Swern, J. Org. Chem. 1978, 43, 2480.
[13] H. C. Brown, R. S. Randad, K. S. Bhat, M. Zaidlewicz, U. S. Racher-
la, J. Am. Chem. Soc. 1990, 112, 2389.
[14] Earlier investigations of the RCM of an analogous substrate lacking
a cyclic constraint had also proven unsuccessful.
[15] S. J. Miller, S.-H. Kim, Z.-R. Chen, R. H. Grubbs, J. Am. Chem. Soc.
1995, 117, 2108.
[44] E. N. Jacobsen, I. Marko, W. S. Mungall, G. Schroeder, K. B. Sharp-
less, J. Am. Chem. Soc. 1988, 110, 1968.
[16] For a review on manipulation of glycidol-related epoxy alcohols,
see: R. M. Hanson, Chem. Rev. 1991, 91, 437.
[45] Previous synthesis: S. C. Bergmeier, D. M. Stanchina, J. Org. Chem.
1999, 64, 2852.
[17] The remainder of the isolated material (10%) for the RCM repre-
sented cross-metathesis between the monosubstituted alkene of
diene 20 and styrene.
[18] a) M. T. Crimmins, B. W. King, E. A. Tabet, J. Am. Chem. Soc. 1997,
119, 7883; b) M. T. Crimmins, A. L. Choy, J. Org. Chem. 1997, 62,
7548; c) M. T. Crimmins, J. She, Synlett 2004, 1371.
[46] P. H. J. Carlsen, T. Katsuki, V. S. Martin, K. B. Sharpless, J. Org.
Chem. 1981, 46, 3936.
[47] A. Hosomi, Y. Sakata, H. Sakurai, Tetrahedron Lett. 1984, 25, 2383.
[48] M. T. Crimmins, J. L. Zuccarello, J. M. Ellis, P. J. McDougall, P. A.
Haile, J. D. Parrish, K. A. Emmitte, Org. Lett. 2009, 11, 489.
[19] For the synthesis of a similar oxocene intermediate by RCM and
subsequent directed hydrogenation, see: K. Fujiwara, Y. Koyama, E.
Received: March 25, 2009
Published online: July 31, 2009
9234
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 9223 – 9234