C-glycosides to fused polycyclic ethers. A formal synthesis of (±)-hemibrevetoxin B

Article abstract of DOI:10.1021/jo001514j

This paper describes a formal total synthesis of the marine ladder toxin hemibrevetoxin B from Danishefsky's dienes. This approach couples the generation of C-glycosides from cyclic enol ethers with metathesis or acid-mediated annulation reactions. The result is a highly efficient synthesis of the tetracyclic ring system of hemibrevetoxin B.

Full text of DOI:10.1021/jo001514j

1380
J . Org. Chem. 2001, 66, 1380-1386
C-Glycosid es to F u sed P olycyclic Eth er s. A F or m a l Syn th esis of
(()-Hem ibr evetoxin B
J on D. Rainier,* Shawn P. Allwein, and J ason M. Cox
Department of Chemistry, The University of Arizona, Tucson, Arizona, 85721
rainier@u.arizona.edu
Received October 24, 2000
This paper describes a formal total synthesis of the marine ladder toxin hemibrevetoxin B from
Danishefsky’s dienes. This approach couples the generation of C-glycosides from cyclic enol ethers
with metathesis or acid-mediated annulation reactions. The result is a highly efficient synthesis of
the tetracyclic ring system of hemibrevetoxin B.
In tr od u ction
The marine ladder toxins comprise a family of red tide
toxins possessing highly complex architectures and very
interesting biological properties including neurotoxicity
and antimicrobial activity.1 Members include the breve-
toxins,2 ciguatoxins,3 maitotoxins,4 and gambieric acids,5
among others. Common to these agents is a highly
symmetrical fused polycyclic ether skeleton consisting of
(a) six- to nine-membered trans-fused ether rings;6 (b)
trans-syn-trans relative ring junction stereochemistry
about any ring;6 (c) ether linkages on vicinal ring junction
carbon atoms; (d) hydrogen or methyl substituents at the
ring junctions (see Figure 1). This high degree of sym-
metry has led many,7 including us,8 to believe that
iterative approaches might be the best way to synthesize
these architectures.9
F igu r e 1. Trans-syn-trans-fused polycyclic ethers present in
the marine ladder toxins.
Sch em e 1
Our general strategy to fused polycyclic ether ring
systems couples the synthesis of C-glycosides via cyclic
enol ether oxidations10 and carbon-carbon bond forming
(1) Shimizu, Y. Marine Natural Products; Academic Press: New
York, 1978; Vol. 1.
(2) (a) For the isolation of brevetoxin B, see: Lin, Y. Y.; Risk, M.;
Ray, M. S.; VanEnsen, D.; Clardy, J .; Golik, J .; J ames, J . C.; Nakanishi,
K. J . Am. Chem. Soc. 1981, 103, 6773. (b) For the total synthesis of
brevetoxin B, see: Nicolaou, K. C.; Rutjes, F. P. J . T.; Theodorakis, E.
A.; Tiebes, J .; Sato, M.; Untersteller, E. J . Am. Chem. Soc. 1995, 117,
10252. (c) For the isolation of brevetoxin A, see: Shimizu, Y.; Chou,
H.-N.; Bando, H.; Van Duyne, G.; Clardy, J . C. J . Am. Chem. Soc. 1986,
108, 514. (d) For the total synthesis of brevetoxin A, see: Nicolaou, K.
C.; Yang, Z.; Shi, G.-Q.; Gunzner, J . L.; Agrios, K. A.; Ga¨rtner, P.
Nature 1998, 392, 264.
(3) (a) Yasumoto, T.; Satake, M. J . Toxicology-Toxin Rev. 1996, 15,
91. (b) Scheuer, P. J . Tetrahedron 1994, 50, 3.
(4) (a) Zheng, W.; DeMattei, J . A.; Wu, J .-P.; Duan, J . J .-W.; Cook,
L. R.; Oinuma, H.; Kishi, Y. J . Am. Chem. Soc. 1996, 118, 7946. (b)
Nonomura, T.; Sasaki, M.; Matsumori, N.; Murata, M.; Tachibana, K.;
Yasumoto, T. Angew. Chem., Int. Ed. Engl. 1996, 35, 1675.
(5) Nagai, H.; Torigoe, K.; Satake, M.; Murata, M.; Yasumoto, J . J .
Am. Chem. Soc. 1992, 114, 1102.
(6) Isolated examples of cis-fused ring junctions exist in this family
(cf. maitotoxin, see ref 4).
(7) For examples of iterative approaches to fused polyethers, see:
(a) Evans, P. A.; Roseman, J . D.; Garber, L. T. J . Org. Chem. 1996,
61, 4880. (b) Bowman, J . L.; McDonald, F. E. J . Org. Chem. 1998, 63,
3680. (c) Mori, Y. Chem. Eur. J . 1997, 3, 849. (d) Clark, J . S.; Kettle,
J . G. Tetrahedron Lett. 1997, 38, 123.
reactions with enol ether, olefin ring-closing metathesis
(RCM) or acid-mediated annulations (Scheme 1).8 The
result of this sequence is a homologous cyclic enol ether
that is ready for further elaboration.
While certainly pleased with their efficiency and flex-
ibility in model compounds, we were intent on demon-
strating the utility of our approaches to one of the
polyether natural products. With this in mind, we elected
(8) (a) Rainier, J . D.; Allwein, S. P. J . Org. Chem. 1998, 63, 5310.
(b) Rainier, J . D.; Allwein, S. P. Tetrahedron Lett. 1998, 39, 9601.
(9) For a pre-1995 review of synthetic strategies to fused polyether
rings, see: Alvarez, E.; Candenas, M.-L.; Pe´rez, R.; Ravelo, J . L.;
Mart´ın, J . D. Chem. Rev. 1995, 95, 1953.
(10) While dimethyl dioxirane is our reagent of choice for these
oxidations, we have also oxidized enol ethers using (a) NBS, H2O (see
ref 8a); (b) AD-mix (Rainier, J . D.; Allwein, S. P. Unpublished results).
10.1021/jo001514j CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/25/2001

Synthesis of (()-Hemibrevetoxin B
J . Org. Chem., Vol. 66, No. 4, 2001 1381
Sch em e 2
Sch em e 3
Con str u ction of th e Hem ibr evetoxin B A-Rin g.
Our hemibrevetoxin B efforts began with a hetero-Diels
Alder cycloaddition reaction to the A-ring (Scheme 3).
From the Danishefsky diene 12,14 a hetero-Diels Alder
cycloaddition with aldehyde 1315 provided dihydropyrone
14 in 92% yield.16,17 Attempted reduction of 14 using
L-selectride gave predominantly products resulting from
1,4-reduction. Fortunately, the reduction of 14 using
Luche’s conditions (NaBH4, CeCl3‚7H2O) gave the corre-
sponding allylic alcohol 15 having the desired C-3 stereo-
chemistry for hemibrevetoxin B.18,19 Because of its sen-
sitivity to chromatography and storage, 15 was carried
into the subsequent epoxidation reaction immediately
following its synthesis without purification. Epoxidation
with m-CPBA in methanol using the C-3 hydroxyl group
to direct the epoxidation reaction according to Rousseau’s
conditions16b gave diol acetal 16 in 65% yield for the two
steps. This protocol nicely complements the dimethyl-
dioxirane oxidation of hydroxyl protected variants of 15
where oxidation occurs from the face opposite the C-3
stereocenter. This reaction proved to be crucial as it
established both the desired C-4 stereochemistry and, in
an indirect fashion, the C-5 stereochemistry via a sub-
sequent C-C bond forming reaction. Differentiation of
the secondary hydroxyl groups in 16 was accomplished
through initial stannyl ketal formation followed by the
synthesis of the benzyl ether at the equatorial or more
exposed alkoxy group to provide 17.13b Acylation of the
remaining hydroxyl and incorporation of the anomeric
allylic group using allyltrimethylsilane and BF3‚Et2O16b
provided the hemibrevetoxin B A-ring as 18 in six steps
(40% overall yield) from 12.
to target hemibrevetoxin B.11 Architecturally, hemibreve-
toxin B consists of 4 heterocyclic rings and 10 stereo-
centers and is the simplest member of the marine ladder
toxin family.12 That hemibrevetoxin B had been synthe-
sized on four separate occasions was of added benefit;
our hemibrevetoxin B efforts would provide us with the
opportunity to gauge the effectiveness of our strategy
with respect to these other approaches.13
R et r osyn t h et ic An a lysis. Our analysis of hemi-
brevetoxin B is outlined in Scheme 2. As envisioned the
approach was to be linear, progressing from the six-
membered A-ring to the seven-membered D-ring. In
addition to providing an opportunity to test the effective-
ness of our chemistry, we were particularly attracted to
the notion that several of the transformations that were
needed to implement the strategy had either been
problematic in the model chemistry or had not been
examined at all. In particular, we were very concerned
with two aspects of the overall plan. First, we envisioned
that the stereocontrolled synthesis of a carbon glycoside
from a bicyclic enol ether having R-functionalization
might be problematic (e.g., 10). We had spent a consider-
able amount of effort examining a similar transformation
in our model chemistry with limited success.8a Another
source of concern was the synthesis of the C,D-ring
system. We had not previously demonstrated that our
approach was applicable to the stereoselective synthesis
of fused oxepanes.
(14) Danishefsky, S.; Kitahara, T.; Schuda, P. F. Org. Synth. 1983,
61, 147.
(15) McDonald, F. E.; Vadapally, P. Tetrahedron Lett. 1999, 40,
2235.
(16) Others have also utilized hetero-Diels Alder chemistry to the
hemibrevetoxin B A-ring. See ref 15 and (a) Gleason, M. M.; McDonald,
F. E. J . Org. Chem. 1997, 62, 6432. (b) Saleh, T.; Rousseau, G. Synlett
1999, 617.
(17) It is possible to generate 14 in enantiomerically enriched
form using Keck’s binol protocol. See refs 15, 16a, and Keck, G. E.; Li,
X.-Y.; Krishnamurthy, D. J . Org. Chem. 1995, 60, 5998.
(18) DIBAL also gave products from 1,2-reduction. Luche’s condi-
tions gave consistently higher yields.
(11) A portion of this work has been communicated. See Rainier, J .
D.; Allwein, S. P.; Cox, J . M. Org. Lett. 2000, 2, 231.
(12) Prasad, A. V. K.; Shimizu, Y. J . Am. Chem. Soc. 1989, 111, 6476.
(13) Hemibrevetoxin B syntheses with number of transformations:
(a) Nicolaou group: 54 synthetic steps from D-mannose. Nicolaou, K.
C.; Reddy, K. R.; Skokotas, G.; Sato, F.; Xiao, X.-Y.; Hwang, C.-K. J .
Am. Chem. Soc. 1993, 115, 3558. (b) Yamamoto group: 52 steps from
D-mannose. Kadota, I.; Yamamoto, Y. J . Org. Chem. 1998, 6597, 7.
(c) Nakata group: 61 steps from geranyl acetate. Morimoto, M.;
Matsukura, H.; Nakata, T. Tetrahedron Lett. 1996, 37, 6365. (d)
Mori group: 36 steps from tri-O-acetyl-D-glucal to a Yamamoto
intermediate, 43 steps overall. Mori, Y.; Yaegassi, K.; Furukawa, H.
J . Org. Chem. 1998, 63, 6200.
(19) Gemal, A. L.; Luche, J .-L. J . Am. Chem. Soc. 1981, 103, 5454.

1382 J . Org. Chem., Vol. 66, No. 4, 2001
Rainier et al.
Sch em e 4
Sch em e 5
Con str u ction of th e Hem ibr evetoxin B A,B-Rin g
System . Having successfully synthesized the A-ring, we
investigated an enol ether, olefin ring-closing metathesis
(RCM) reaction to the B-ring (Scheme 4). Exposure of 18
to modified Takai conditions20 gave a mixture of acyclic
and cyclic enol ethers in 69% yield. As in our model
chemistry,8a the presence of a mixture of enol ethers was
not an obstacle as we simply exposed the mixture to the
Schrock molybdenum catalyst 521 and in the process
converted all of 20 into 19 in 93% yield.22,23 As the
synthesis of the A,B-ring system places the substituents
at C-1 and C-3 in an energetically disfavored 1,3-diaxial
relationship, this reaction serves as further evidence of
the power of the RCM protocol to fused polyethers.24
We believe that the formation of cyclic enol ether from
the Takai procedure is a result of an olefin metathesis-
carbonyl olefination reaction sequence.25,26 As proof of this
we subjected acyclic enol ether 228a to the Takai reaction
conditions and isolated olefin oligomers (Scheme 5). We
did not observe any formation of 23 in this experiment.
With 19 in hand, we were prepared to investigate the
critical B-ring C-glycoside forming chemistry (Scheme 6).
We were not overly concerned about the facial selectivity
in the oxidation reaction as we believed that the C-3
benzyloxy group on the A-ring would direct the approach
of dimethyldioxirane to the R-face of the enol ether.
However, we were concerned about the outcome of the
subsequent epoxide ring-opening reaction. This came
Sch em e 6
from the relatively low yields that we had observed for
this reaction in similarly substituted fused polyethers.8a
Undoubtedly, the low conversions in these instances were
related to the need to couple the nucleophile at the more
substituted, albeit more activated end of the epoxide. In
the event, we isolated 25 in 72% yield when 19 was
exposed to dimethyldioxirane27 at -60 °C followed by
propenylmagnesium chloride at 0 °C.28 Not only had
propenylmagnesium chloride added to C-7 but 25 is the
result of a C-8 reduction and a C-7 oxidation.29
We believe that 25 comes from a stereoselective epoxi-
dation to give 24. Subsequently, propenylmagnesium
chloride acts as a Lewis acid to give intermediate
oxonium ion 26 (Scheme 7). Oxonium 26 then undergoes
a syn-facial [1,2]-hydride migration to give ketone 27.
Addition of propenylmagnesium chloride to 27 gives the
observed product.30
(20) Takai, K.; Kakiuchi, T.; Kataoka, Y.; Utimoto, K. J . Org. Chem.
1994, 59, 2668.
(21) Schrock, R. R.; Murdzek, J . S.; Bazan, G. C.; Robbins, J .;
DiMare, M.; O’Regan, M. B. J . Am. Chem. Soc. 1990, 112, 3875.
(22) Fujimura, O.; Fu, G. C.; Grubbs, R. H. J . Org. Chem. 1994, 59,
4029.
As we had been able to successfully overcome the
problems associated with the formation of oxonium ions
in our model substrates through careful temperature
control,8b we were confident that the addition of pro-
penylmagnesium chloride to 24 at lower temperatures
(23) For reviews on RCM see: (a) Grubbs, R. H.; Miller, S. J .; Fu,
G. C. Acc. Chem. Res. 1995, 28, 446. (b) Schmaltz, H.-G. Angew. Chem.
Int. Ed. 1995, 34, 1833. (c) Schuster, M.; Blechert, S. Ang. Chem., Int.
Ed. Engl. 1997, 36, 2036. (d) Grubbs, R. H.; Chang. S. Tetrahedron
1998, 54, 4413. (e) Fu¨rstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012.
(24) The synthesis of the A-B ring system from a similar precursor
proved to be problematic in McDonald’s hemibrevetoxin work involving
alkynyl-tungsten cyclizations. See ref 16a.
(25) Grubbs has reported the use of the Tebbe reagent in an olefin
metathesis-carbonyl olefination reaction. (a) Stille, J . R.; Grubbs, R.
H. J . Am. Chem. Soc. 1986, 108, 855. (b) Stille, J . R.; Santarsiero, B.
D.; Grubbs, R. H. J . Org. Chem. 1990, 55, 843.
(26) Nicolaou has utilized the Tebbe and Petasis reagents to
fused polyether rings from olefins having pendant esters. He has
reported that this reaction proceeds via a carbonyl olefination, olefin-
olefin RCM sequence. See: (a) Nicolaou, K. C.; Postema, M. H. D.;
Yue, E. W.; Nadin A. J . Am. Chem. Soc. 1996, 118, 10335. (b) Nicolaou,
K. C.; Postema, M. H. D.; Claiborne, C. F. J . Am. Chem. Soc. 1996,
118, 1565.
(27) (a) Halcomb, R. L.; Danishefsky, S. J . J . Am. Chem. Soc. 1989,
111, 1, 6661. (b) Adam, W.; Bialas, J .; Hadjiarapoglou, L. Chem. Ber.
1991, 124, 2377.
(28) For the addition of propenylmagnesium chloride to glycal
epoxides, see ref 8a and: (a) Best, W. M.; Ferro, V.; Harle, J .; Stick,
R. V.; Tilbrook, D. M. G. Aust. J . Chem. 1997, 50, 463. (b) Evans, D.
A.; Trotter, B. W.; Coˆte´, B. Tetrahedron Lett. 1998, 39, 1709.
(29) Other nucleophiles (e.g., 49) provided similar rearrangement
products.
(30) The relative stereochemistry at C-7 and C-8 was established
from the corresponding NOE enhancements (see ref 11).

Synthesis of (()-Hemibrevetoxin B
J . Org. Chem., Vol. 66, No. 4, 2001 1383
Ta ble 1
Sch em e 7
entry
nucleophile
conditions
yield, %
1
2
3
MeMgBr
Me3Al
Me3Al
THF, -65 °C f rt
THF, -65 °C f rt
hexanes, -65 °C f rt
40a
50b
75
a
Major by-product (25%) was the tertiary alcohol resulting from
b
hydride migration. Major byproduct (20%) was the trans-syn-cis
would be more successful (eq 1). In the event, when
propenylmagnesium chloride was added to oxidized 24
diastereomer resulting from either direct addition to the epoxide
or from a nonstereoselective addition to the intermediate oxonium
ion.
Sch em e 8
at -60 °C, we isolated 28 in 35% yield along with the
hydride migration product 25 in 15% yield. To our
dismay, 28 had the desired hemibrevetoxin B connectiv-
ity but the undesired trans-syn-cis relative stereochem-
istry about the B-ring. Clearly, we had supressed hydride
migration but had not successfully overcome oxonium ion
formation. However, we were pleasantly surprised to find
that the low-temperature coupling of 24 with propenyl-
magnesium chloride was completely diastereoselective.
This result stands in contrast to our model work where
the addition of carbanions to oxonium ions typically gave
mixtures of diastereomers at the carbon glycoside bearing
stereocenter.8b
We were unable to determine the relative stereochem-
istry of alcohol 28 spectroscopically and therefore con-
verted it into the C-8 epimer of the hemibrevetoxin B
A-C ring system 29 (eq 2). From 28, annulation to 29
nucleophile to an oxonium ion having the C-ring carbon
atoms intact). This strategy required that we generate
an A,B ring system that contained the requisite C-ring
carbon atoms.
Our starting point for this approach was hydroxy-
acetal 17 (Scheme 8). Allyl incorporation using allylsilane
and TMSOTf gave 31 having the desired relative stereo-
chemistry in 89% yield.16b The incorporation of the
necessary C-ring carbon atoms was accomplished through
the esterification of the secondary alcohol with 32. Once
again, the A,B-ring system was generated by following
the same two step enol ether, olefin RCM protocol that
had been successful on 18. As before, the Takai protocol
provided a mixture of acyclic and cyclic enol ether that
was subsequently converted into bicyclic fused ether 34
using the Schrock catalyst 5 (8 steps, 34% overall yield
from 12). Interestingly, the much more robust Grubbs
imidazole catalyst 33 was equally effective in the conver-
sion of acyclic enol ether to 34.32
was accomplished in two steps by first converting the
secondary alcohol into the corresponding allylic ether
followed by olefin-olefin RCM using the Grubbs ruthe-
nium catalyst 30.31 The relative stereochemistry at the
B-C ring junction was ascertained from the indicated
NOE enhancements.
Because of the stereoselective nature of the carbon-
carbon bond forming reaction, we believed that a poten-
tial solution to our C-8 stereochemical problem might
come from simply reversing the order of the carbon-
carbon bond forming sequence (i.e., adding a methyl
With 34 in hand, we investigated the oxidation and
subsequent formation of the requisite carbon-carbon
bond using methyl nucleophiles (Table 1). While both
MeMgBr and Me3Al gave 35 having the desired trans-
(32) In contrast, the first generation Grubbs ruthenium catalyst 30
(see eq 2) was not successful in the conversion to 34. For examples of
the use of 30 in enol ether-olefin RCM, see: (a) Sturino, C. F.; Wong,
J . C. Y. Tetrahedron Lett. 1998, 63, 9623. (b) Clark, S. J .; Hamelin, O.
Angew. Chem., Int. Ed. 2000, 39, 372.
(31) (a) Schwab, P.; France, M. B.; Ziller, J . W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2039. (b) Delgado, M.; Tetrahedron
Lett. 1997, 38, 6299. (c) Clark, J . S.; Hamelin, O.; Hufton, R.
Tetrahedron Lett. 1998, 39, 8321.

1384 J . Org. Chem., Vol. 66, No. 4, 2001
Rainier et al.
Ta ble 2
Sch em e 9
Sch em e 10
product yield, %
entry nucleophile
conditions
40 42 + 43
41
1
2
3
4
5
6
allyMgCl
-65 °C, 3 h
64a
-
45
-
-
-
vinylMgCl -78 °C f rt, 2 h
45b
vinylMgCl -60 °C f 50 °C, 48 h 40b
40
14
vinylMgCl -50 °C f -40 °C, 48 h 60b
14
44
-
various temperatures
MgCl2, 65 °C
-
-
60-90c
73
-
a
b
40a : R′ ) CH2CHdCH2. 40b: R′ ) CHdCH2. c A variety of
reaction conditions resulted in mixtures of 41, 42, and/or 43 in
syn-trans stereochemistry, Me3Al proved to be the re-
agent of choice. Optimized conditions involved the use
of excess Me3Al at low temperature in hexanes to provide
35 in 75% yield. Thus by simply reversing the order of
carbon-carbon bond formation at C-8, we were able to
stereoselectively generate both the trans-syn-cis as well
as the trans-syn-trans hemibrevetoxin B B-ring isomers.
These experiments clearly demonstrate the flexibility of
our approach to these ring systems.33
We believe that 35 results from the ionization of
epoxide 36 and the intramolecular transfer of a methyl
group to the resulting oxonium ion (e.g., 37, Scheme 9).34
One would expect nonpolar solvents to favor the forma-
tion of 37 and consequently products from intramolecular
delivery.
Con str u ction of th e Hem ibr evetoxin B A-C Rin g
System . Exposure of 35 to PPTS, pyridine, and heat
provided 39 in 66% yield (Scheme 10). This highly
efficient annulation reaction proceeds via the initial
formation of mixed cyclic acetal 38 followed by the in situ
elimination of methanol.8b,35
With 39 in hand, we were reasonably confident
that we would be able to carry out a stereoselective
epoxidation reaction on the â-face of the C-ring opposite
the angular methyl group. In the event, oxidation with
dimethyldioxirane followed by the addition of propenyl-
magnesium chloride gave 40a as a single isomer in
64% yield (eq 3). As determined from NOE difference
overall yields ranging from 60 to 90%.
acetal Grignard 44 (Table 2). If successful, this reac-
tion would allow us to take advantage of the same
single-flask acid-mediated cyclization/elimination reac-
tion that had been successful for the synthesis of the
C-ring. Unfortunately, we were unable to couple 44 with
41. In fact, when the temperature of the addition reaction
was kept below -60 °C, we recovered 41 intact. Surpris-
ingly, epoxide 41 was even partially stable to silica gel
chromatography! When the temperature of the addition
reaction was allowed to rise above -30 °C, we isolated a
mixture of C-ring ketone 42 (from hydride migration)
and/or tertiary alcohol 43 from Grignard addition to
the ketone. Interestingly, propenylmagnesium chloride
appears to be ideally suited for addition to 41 as even
vinylmagnesium chloride gave predominantly recovered
41, 42, and/or 43 unless it were allowed to react for
extended periods of time at low temperature (entries
2-4).36,37 As a demonstration of its propensity to re-
arrange, we were able to isolate ketone 42 in 73% yield
when 41 was stirred with MgCl2 at 65 °C for 4 h (entry 6).
While not readily apparent at the present time, the
difference in reactivity between propenylmagnesium
chloride and the other nucleophiles may be related to the
ability of propenylmagnesium chloride to form a six-
(33) The oxidation protocol can also provide flexibility. For example,
while the oxidation of 23 (Scheme 5) with dimethyldioxirane gave
products resulting from reaction on the â-face, the oxidation of 23 with
osmium tetraoxide gave products resulting from selective oxidation of
the R-face (Rainier, J . D.; Allwein, S. P. Unpublished results).
(34) We have exploited the facility with which both aluminum and
boron carry out the intramolecular transfer to glycal anhydrides in
the synthesis of C-glycosides. See: Rainier, J . D.; Cox, J . M. Org. Lett.
2000, 2, 2707.
(35) For the use of PPTS in the elimination of methanol from mixed
acetals, see: (a) Corey, E. J .; Kang, M.-C.; Desai, M. C.; Ghosh, A. K.;
Houpis, I. N. J . Am. Chem. Soc. 1988, 110, 649. (b) Kozikowski, A. P.;
Ghosh, A. K. J . Org. Chem. 1985, 50, 3017.
(36) Evans has also observed that propenylmagnesium chloride
provides the most efficient couplings with glycal epoxides. See ref
28b.
experiments on the corresponding C-11 acetate, 40a
had the desired hemibrevetoxin C-ring stereochemistry.11
While satisfied with the formation of 40a from 39, a
more efficient strategy to the hemibrevetoxin B ring
system would involve the coupling of epoxide 41 with
(37) We were also unsuccessful in our attempts to couple epoxide
41 with homoallylmagnesium chloride.

Synthesis of (()-Hemibrevetoxin B
J . Org. Chem., Vol. 66, No. 4, 2001 1385
nucleophilic addition to the ketone, and/or epoxide. The
presence of a mixture of C-15, C-16 diastereomers in the
F igu r e 2. Proposed propenyl Grignard intermediate.
Sch em e 11
C-glycoside formation was not surprising or of concern
as we intended to destroy the oxygen-bearing center by
converting it into the corresponding ketone. We were
optimistic that any undesired C-16 diastereoisomer that
remained after oxidation could be converted into the
desired isomer through a base-induced equilibration
reaction.
Our inability to couple nucleophiles other than allyl
with the epoxide from 46 is perplexing in light of our
ability to couple the epoxide from the simpler oxepane
48 with a variety of nucleophiles including acetal mag-
nesium bromide 49 (eq 5).42 Clearly, these coupling
membered transition structure by coordinating to the
cyclic ether oxygen as depicted in Figure 2.38
Con str u ction of th e Hem ibr evetoxin B Tetr a -
cycle. We were able to overcome our inability to couple
41 with 44 by converting the propenyl addition product
40a into the corresponding tetracycle using a similar
sequence of reactions to those that had been successful
in our synthesis of the A-C epimeric substrate 29 (eq
2). Namely, allyl ether formation and bis-olefin RCM
using the Grubbs ruthenium catalyst gave 45 (Scheme
11).39 Olefin isomerization using Wilkinson’s catalyst40
gave the hemibrevetoxin A-D ring system 46 in 67%
overall yield for the three transformations. Most impres-
sively, the sequence of events that have been described
here gave the tetracyclic core of hemibrevetoxin B along
with 8 of the 10 stereocenters in 14 steps (7.1% overall
yield) from the Danishefsky diene.14
With the hemibrevetoxin B tetracyclic unit in hand,
we proceeded to transform it into an intermediate that
Mori had generated during his formal synthesis of
hemibrevetoxin B (e.g., 55).13d The addition of propenyl-
magnesium chloride to the epoxide from the oxidation of
46 gave coupled product 47 as a mixture of three isomers
in a 3:1:1 ratio in 84% overall yield after acylation of the
resulting alcohol (eq 4). As with the C-ring couplings
described above, our attempts to couple oxidized 46 with
other anions41 resulted in the isolation of mixtures of
ketone from hydride migration, tertiary alcohol from
reactions are much more complicated than we had
initially imagined. Apparently, this is particularly true
in tri- and tetracyclic ring systems.
As with tricyclic epoxide 41, the hydride migration
chemistry of 52 is facile. We were able to isolate ketone
53 in 74% yield when epoxide 52 was exposed to MgCl2
at room temperature for 6 h (eq 6).
(38) As depicted, the addition reaction should require 2 equiv
of propenylmagnesium chloride. We typically use an excess of the
carbon nucleophile (3-5 equiv) and have not carried out the control
experiments to determine whether 2 equiv is necessary.
(39) Both Crimmins and Clark have utilized bis-olefin RCM to
generate oxepenes. See: (a) Clark, J . S.; Kettle, J . G. Tetrahedron Lett.
1997, 38, 127. (b) Crimmins, M. T.; Choy, A. L. J . Org. Chem. 1997,
62, 7548.
(40) Corey, E. J .; Suggs, J . W. J . Org. Chem. 1973, 38, 3224.
(41) We have been unsuccessful in our attempts to couple epoxide
52 with BrMgCH2CH2CH2OTBS and ClMgCH2CH2CHdCH2.

1386 J . Org. Chem., Vol. 66, No. 4, 2001
Rainier et al.
Sch em e 12
diastereomers under the assumption that any undesired
C-16 stereoisomer could be epimerized. Surprisingly,
the attempted isomerization of 56 using DBU was un-
successful in our hands.43 Fortunately we were able to
transform all of 56 into 55 by turning to NaOEt in
EtOH. This sequence efficiently provided (()-55 that was
identical in all respects to Mori’s published data.13d
Con clu sion
To conclude, we have accomplished a formal total
synthesis of (()-hemibrevetoxin B by employing sequen-
tial enol ether oxidations, carbon nucleophile additions,
and annulations. These efforts validate the notion that
C-glycoside-based approaches to fused polyethers are
concise while maintaining a high degree of flexibility.
Undoubtedly, these strategies will make fused polyethers
more readily available than they have been to date.
Ack n ow led gm en t. Dedicated with great respect
and admiration to Professor Bob Bates upon his retire-
ment from teaching at The University of Arizona. We
would like to thank Mr. Brett Howard and Ms. Cathy
Ai for their assistance in the synthesis of large quan-
tities of 12-17. The authors are extremely grateful to
the National Institutes of Health, General Medical
Sciences (GM56677), Research Corporation, and the
donors of the Petroleum Research Fund, administered
by the American Chemical Society, for support of this
research. We would also like to thank Dr. Neil J acobsen
and Dr. Arpad Somagyi for help with NMR and mass
spectroscopy experiments, respectively.
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
mental details for compounds 14, 15, 19, 25, 28, 33, 38, 45.
Complete experimental details and spectroscopic data for
compounds 16-18, 28, 31, 33, 34, 39-42, 44, 46-49, 52-54.
This material is free of charge via the Internet at http://
pubs.acs.org.
The conversion of propenyl adduct 47 into the Mori
intermediate 55 is outlined in Scheme 12. Hydroboration
of the alkene, TBDPS ether formation, ester hydrolysis,
and oxidation gave a mixture of 55 and 56 in a 2:1 ratio,
respectively. As mentioned above we had gone through
this sequence of transformations with a mixture of C-16
J O001514J
(42) Oxidized 48 also coupled with propenylmagnesium chloride in
an unoptimized 52% yield.
(43) Others have had more success with this protocol. For example,
see ref 13a.