Synthesis of the C29-C44 portion of spongistatin 1 (altohyrtin A)

Article abstract of DOI:10.1021/jo0002801

Two synthetic approaches to the C29-C44 portion of spongistatin 1 (ahohyritin A) have been developed. The key step of the first approach relies on the Claisen rearrangement of glucal 18 to provide ester 20a. This intermediate was advanced to silyl enol ether 30, which was coupled under Mukaiyama aldol conditions with aldehyde 3. Cyclization of this aldol adduct completed our first synthesis of the C29-C44 portion of spongistatin 1, requiring 25 total steps and occurring in 2.4% yield over the longest linear sequence (21 steps). We have also developed a second-generation approach based on the C-glycosidation of glucal 43. Through equilibration of the corresponding C-glycosides 49a/b and 50a/b the desired C-glycoside (50a) was obtained in good yield. Aldol condensation of this ketone provided cyclization precursor 67, which undergoes acid-catalyzed ketalization to close the E-ring of the spongistatins. An oxidation/reduction protocol was employed to set the C37 stereocenter. Protection of the C37 carbonol and selective unmasking of the C44 carbonol completed our second generation synthesis. This approach requires 27 steps and occurred in 13-2% yield over the longest linear sequence (18 steps).

Full text of DOI:10.1021/jo0002801

J . Org. Chem. 2000, 65, 4145-4152
4145
Syn th esis of th e C29-C44 P or tion of Sp on gista tin 1 (Altoh yr tin A)
Grier A. Wallace, Robert W. Scott, and Clayton H. Heathcock*
Department of Chemistry, University of California, Berkeley, California 94720
heathcock@cchem.berkeley.edu
Received February 29, 2000
Two synthetic approaches to the C29-C44 portion of spongistatin 1 (altohyritin A) have been
developed. The key step of the first approach relies on the Claisen rearrangement of glucal 18 to
provide ester 20a . This intermediate was advanced to silyl enol ether 30, which was coupled under
Mukaiyama aldol conditions with aldehyde 3. Cyclization of this aldol adduct completed our first
synthesis of the C29-C44 portion of spongistatin 1, requiring 25 total steps and occurring in 2.4%
yield over the longest linear sequence (21 steps). We have also developed a second-generation
approach based on the C-glycosidation of glucal 43. Through equilibration of the corresponding
C-glycosides 49a /b and 50a /b the desired C-glycoside (50a ) was obtained in good yield. Aldol
condensation of this ketone provided cyclization precursor 67, which undergoes acid-catalyzed
ketalization to close the E-ring of the spongistatins. An oxidation/reduction protocol was employed
to set the C37 stereocenter. Protection of the C37 carbonol and selective unmasking of the C44
carbonol completed our second generation synthesis. This approach requires 27 steps and occurred
in 13.2% yield over the longest linear sequence (18 steps).
In 1993 and 1994, isolations of the spongistatins,1
altohyrtins,2 and cinachyrolide A3 were reported. These
structurally related macrocyclic lactones show extreme
cytotoxicity toward various tumor cell lines, thus making
them valuable tools for understanding cancer. Spong-
istatin 1 has been identified as “probably the best to date
in the NCI’s evaluation programs,”4 exhibiting 50%
growth inhibition against a range of tumor lines at
concentrations in the range of 10-10 to 10-12 mol/L.
Additionally, this material has shown potent activity
against a subset of highly chemoresistant tumor types.4
Spongistatin appears to disrupt microtubule assembly by
binding tubulin in the vinca alkaloid binding site.5 The
compound also has potent antifungal properties, inhibit-
ing the growth of many fungi, including strains resistant
to amphotericin B, ketoconazole, and flucyosine.6
chemotherapy. The problem is that they are available
from Nature in only minute amounts and there is
currently no practical way to farm sponges to obtain
larger quantities of the metabolites. We believe that, even
despite their great complexity, conventional organic
synthesis can provide multigram quantities of the spong-
istatins, and it is the goal of the current project to develop
an efficient total synthesis to provide several grams of
the natural product so that a phase I clinical trial can
be carried out.
Because of their complex structures and extreme
cytotoxicity, these natural products have attracted the
attention of many groups,7 and two groups have already
recorded complete total syntheses. Evans and co-workers
reported the first total synthesis of spongistatin 2 and
proved it to be identical to altohyrtin C (2).8 Kishi has
published the total synthesis of altohyrtin A, which is
identical to spongistatin 1 (1).9
At the beginning of our synthetic efforts, we decided
to focus on altohyrtin A (1), since it was the only structure
whose absolute and relative configuration was supported
by adequate data. Our strategy was designed around a
macrolactonization as the final key step. Further discon-
nection of the C28-C29 double bond revealed two
similarly complex fragments, which were to be joined
using a Wittig reaction. We have previously described
our synthesis of the C1-C28 portion of spongistatin 1
(altohyrtin A).10 The C29-C51 fragment was envisioned
These extremely active marine natural products pro-
vide a good example of the power of Nature to point us
in the direction of organic structures of potential use in
(1) (a) Pettit, G. R.; Cichacz, Z. A.; Herald, C. L.; Gao, F.; Boyd, M.
R.; Schmidt, J . M.; Hamel, E.; Bai, R. J . Chem. Soc., Chem. Commun.
1994, 1605-1606. (b) Pettit, G. R.; Herald, C. L.; Cichacz, Z. A.; Gao,
F.; Boyd, M. R.; Christie, N. D.; Schmidt, J . M. Nat. Prod. Lett. 1993,
3, 239-244. (c) Pettit, G. R.; Herald, C. L.; Cichacz, Z. A.; Gao, F.;
Schmidt, J . M.; Boyd, M. R.; Christie, N. D.; Boettner, F. E. J . Chem.
Soc., Chem. Commun. 1993, 1805-1807. (d) Pettit, G. R.; Cichacz, Z.
A.; Gao, F.; Herald, C. L.; Boyd, M. R. J . Chem. Soc., Chem. Commun.
1993, 1166-1168. (e) Pettit, G. R.; Cichacz, A. A.; Gao, F.; Herald, C.
L.; Boyd, M. R.; Schmidt, J . M.; Hooper, J . N. A. J . Org. Chem. 1993,
58, 1302-1304.
(2) (a) Kobayashi, M.; Aoki, S.; Kitagawa, I. Tetrahedron Lett. 1994,
35, 1243-1246. (b) Kobayashi, M.; Aoki, S.; Sakai, H.; Kawazoe, K.;
Kihara, N.; Sasaki, T.; Kitagawa, I. Tetrahedron Lett. 1993, 34, 2795-
2798.
(3) Fusetani, N.; Shinoda, K.; Matsunaga, S. J . Am. Chem. Soc.
1993, 115, 3977-3981.
(4) Pettit, G. R. J . Nat. Prod. 1996, 59, 812-821.
(7) For a recent review, see: Pietruszka, J . Angew. Chem., Int. Ed.
Engl. 1998, 37, 2629-2636.
(8) (a) Evans, D. A.; Coleman, P. J .; Dias, L. C. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2738-2741. (b) Evans, D. A.; Trotter, B. W.; Coˆte´,
B.; Coleman, P. J . Angew. Chem., Int. Ed. Engl. 1997, 36, 2741-2744.
(c) Evans, D. A.; Trotter, W. B.; Coˆte´, B.; Coleman, P. J .; Dias, L. C.;
Tyler, A. N. Angew. Chem., Int. Ed. Engl. 1997, 36, 2744-2747.
(9) (a) Guo, J .; Duffy, K. J .; Stevens, K. L.; Dalko, P. I.; Roth, R. M.;
Hayward, M. M.; Kishi, Y. Angew. Chem., Int. Ed. Engl. 1998, 37, 187-
192. (b) Hayward, M. M.; Roth, R. M.; Duffy, K. J .; Dalko, P. I.; Stevens,
K. L.; Guo, J .; Kishi, Y. Angew. Chem., Int. Ed. Engl. 1998, 37, 192-
196.
(5) (a) Bai, R.; Cichacz, Z. A.; Herald, C. L.; Pettit, G. R.; Hamel, E.
Mol. Pharmacol. 1993, 44, 757-766. (b) Bai, R.; Cichacz, Z. A.; Herald,
C. L.; Pettit, G. R.; Hamel, E. Proc. Annu. Meet. Am. Assoc. Cancer
Res. 1993, 34, 4: A2243. (c) Bai, R.; Taylor, G. F.; Cichaz, Z. A.; Herald,
C. L.; Kepler, J . A.; Pettir, G. R.; Hamel, E. Biochemistry 1995, 34,
9714-9721.
(6) Pettit, G. R. Patent Application WO97-US10200.
10.1021/jo0002801 CCC: $19.00 © 2000 American Chemical Society
Published on Web 06/02/2000

4146 J . Org. Chem., Vol. 65, No. 13, 2000
Wallace et al.
F igu r e 1.
Sch em e 1
Sch em e 2a
a
Key: (a) Na, NH3; (b) BPSCl, imid., DMF; (c) (4-methoxy)ben-
zyloxyacetic acid, DCC, DMAP, DMF; (d) LiHMDS, TMSCl, pyr.,
-78 °C to rt; (e) CH2N2.
as arising from an allylic rearrangement where the C45-
C58 olefin would originate in the C44-C45 position.
Disconnection across the R-â unsaturation leads to the
C29-C44 fragment of altohyrtin A. In this paper, we
describe our first- and second-generation approaches to
this key intermediate (Figure 1).
Retrosynthetic analysis of the EF bispyran ring system
(Scheme 1) suggested aldehyde 3 and ketone 4 as poten-
tial precursors. We planned to use a Claisen rearrange-
ment of glucal 5 as a key step in our synthesis of ketone
4. This would allow us to simultaneously set the C38 and
C39 stereocenters in a single synthetic operation.
As shown in Scheme 2, synthesis of the Claisen re-
arrangement precursor began with tribenzyl ether 6, pre-
viously described by our laboratory.11 The benzyl groups
of 6 were removed with a dissolving metal reduction, and
the resulting triol was protected as the primary tert-
butyldiphenylsilyl ether in 84% overall yield. Our at-
tempts to esterify the 3-OH of 7 selectively with (4-
methoxybenzyloxy)acetic acid12 were unsuccessful, pro-
viding a mixture of products. Instead, we chose to
bisesterify diol 7 with an excess of the PMB-protected
glycolic acid, providing 8 in 97% yield.
With the Claisen rearrangement precursor in hand we
could explore this key step of the synthesis. Treatment
of a premixed solution of 8, pyridine, and TMS-Cl with
excess LiHMDS resulted in formation of the bis-si-
lylketene acetal.13 Upon warming to room temperature,
this substrate underwent rapid Claisen rearrange-
(12) Abraham, D. J .; Kennedy, P. E.; Mehanna, A. S.; Patwa, D. C.;
Williams, F. L. J . Med. Chem. 1984, 27, 8, 967-978.
(13) Barrish, J . C.; Lee, H. L.; Mitt T.; Pizzolato, G.; Baggiolini, E.
G. Uskokovic, M. R. J . Org. Chem. 1988, 53, 4282-4295 and references
therein.
(10) Clafey, M. M.; Hayes, C. J .; Heathcock, C. H. J . Org. Chem.
1999, 64, 8267-8274.
(11) Scott, R. W.; Heathcock, C. H. Carbohydr. Res.1996, 291, 205-
208.

Synthesis of the C29-C44 Portion of Spongistatin 1
J . Org. Chem., Vol. 65, No. 13, 2000 4147
Sch em e 3
Sch em e 4
to substantially affect the ratio of 9a and 9b. At this
juncture, we took a purely empirical approach. The idea
was to replace the p-methoxybenzyl group with a p-
methoxyphenylethyl group. This change would presum-
ably produce a small but unpredictable effect on the
disastereoselectivity of the Claisen rearrangement. If the
effect happened to be beneficial, it was hoped that this
small beneficial effect would act to amplify the observed
2:1 diastereomeric ratio of 9a and 9b. Since we would be
introducing a new stereocenter, we thought we might find
ourselves with a double diastereoselection situation: with
one of the diastereomeric p-methoxyphenylethyl deriva-
tives amplifying the 2:1 ratio and the other attenuating
it. To investigate the feasibility of this approach, we
chose the easily prepared methylbenzyloxy glycolic acids
13a and 13b. Coupling of the potassium salt of com-
mercially available (S)- and (R)-phenethyl alcohols with
bromoacetic acid provided the desired acids in high yield
(Scheme 4).
Coupling of acids 13a and 13b with diol 7 was
accomplished with DCC and DMAP, providing the di-
esters 14a and 14b in high yield. The results of the
Claisen rearrangements of 8, 14a , and 14b are shown
in Scheme 5. The use of (R)-methylbenzyloxy ester caused
a decrease in the inherent selectivity of the rearrange-
ment, now slightly favoring the undesired ester 15b.
However, use of the (S)-methylbenzyloxy ester caused an
increase in the ratio to 6:1, favoring the desired diaste-
reomer 16a . Under the optimized reaction conditions, the
desired diastereomer could be isolated in 70% yield.
To explore the later steps of our synthesis, we had to
change the tert-butyldiphenylsilyl protecting group of 15a
to a protecting group that would withstand the depro-
tection of the methylbenzyl moiety. This was most
conveniently done early in the synthesis (Scheme 6). To
this end, glucal 6 was deprotected under dissolving metal
conditions and the resulting triol was protected as the
TIPS ether to provide 17 in 90% overall yield. The diol
was bis-acylated with acid 13b to provide the bis-ester
18 in 88% yield. Claisen-rearrangement and esterification
provided the desired ester 20a as the major product in
73% isolated yield.
ment.14,15 The crude reaction mixture was treated with
diazomethane to facilitate isolation of the methyl esters
10a and 10b in a 85% overall yield, as a 2:1 mixture of
diastereomers. If the reaction mixture was allowed to
remain at room temperature for more than 15 min, the
tandem Claisen product predominated.16 To elucidate the
relative stereochemistry of 10a and 10b, they were
separated by column chromatography, saponified to the
acids, and induced to lactonize with I2 and NaHCO3.17
With the structures rigidified as the 6/5 lactones 11a and
11b, NOE difference experiments were performed (Scheme
3).18 Irradiation of the bridgehead methyl group of 11b
produced a 14% enhancement in the signal of the proton
labeled Ha that was not observed when the bridgehead
methyl group of 11a was irradiated.
We were pleased that the Claisen rearrangement
provided the desired diastereomer as the major product.
However, the rather low selectivity deserves further
discussion. Literature data suggest that the presence of
a vinyl ether accelerates Claisen rearrangements on
olefins of this type. In addition, the presence of an oxy-
gen R to the carbonyl group also has an accelerating
effect.13,14,16 No examples could be found where both
of these situations were present simultaneously, as in
the present case. However, our data suggest that the
two accelerating effects are additive, providing
a
Claisen rearrangement that is very facile, but not very
selective.
At this point, we attempted to alter the reaction
conditions of the Claisen rearrangement to increase the
diastereoselectivity of this reaction. Attempts at pre-
chelation of the glycolic ester with lithium chloride in
order to affect the E/Z ratio of the silyl enol ether13 were
unsuccessful, providing esters 10a and 10b in the same
ratio. Addition of solvent additives such as HMPA, were
equally ineffective. Trapping the bisenolate of 8 with
either TES-Cl or TBS-Cl led to complex mixtures of
products. Returning to the original conditions and run-
ning the reaction at -78 °C was unsuccessful, providing
the same product ratio observed previously.
With ester 20a in hand we were in a position to explore
the later steps of our synthesis. Ester 20a was hydro-
borated with 9-BBN (Scheme 7). Unfortunately, reduction
of the methylbenzyloxyacetate of 20a occurred in prefer-
ence to hydroboration of the double bond, providing
alcohol 21. Under more forcing conditions triol 22 was
obtained. Other hydroborating reagents were also un-
satisfactory, producing complex mixtures of products.
This unforseen complication forced us to add unwanted
steps to keep the C41 and C42 hydroxyl groups (spong-
istatin numbering) differentially protected.
From the above results, it was evident that conven-
tional alterations in the reaction protocol were not going
(14) Wipf, P. Comprehensive Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Elmsford, 1991; Vol. 5, pp 857-862 and references
therein.
(15) Ireland, R. E.; Wipf, P.; Xiang, J .-N. J . Org. Chem. 1991, 56,
3572-3582.
Diester 20a was reduced with DIBAL-H in nearly
quantitative yield, providing diol 23 (Scheme 8). Protec-
tion of the primary alcohol as the TBS ether provided 24
in 90% yield. Compound 24 represented a convenient
(16) Curran, D. P.; Suh, Y.-G. Carbohydr. Res.1987, 171, 161-191.
(17) (a) Barnett, W. E.; Sohn, W. H. Tetrahedron Lett. 1972, 18,
1777-1780. (b) Bartlett, P. A.; Myerson, J . J . Am. Chem. Soc. 1978,
100, 3950-3952.
(18) We thank Dr. Anna Mapp for performing these experiments.

4148 J . Org. Chem., Vol. 65, No. 13, 2000
Wallace et al.
Sch em e 5
Sch em e 6a
Sch em e 8a
a
Key: (a) Na, NH3; (b) TIPSCl, imid., DMF; (c) 13b, DCC,
DMAP, DMF; (d) LiHMDS, TMSCl, pyr., -78 °C to rt (2-3 min);
(e) CH2N2.
a
Key: (a) DIBAL-H; (b) TBSCl, imid.; (c) Na, NH3; (d) PMBBr,
NaH; (e) (i) 9-BBN, (ii) [O]; (f) TIPSOTf, 2,6-lutidene.
Sch em e 7a
Sch em e 9a
a
Key: (a) 0.5% H2SO4, THF, H2O, (three iterations); (b) Dess-
Martin; (c) MeLi.
To properly functionalize 27, it was necessary to
deprotect the carbinol at C38 and convert the resulting
primary alcohol to the methyl ketone (Scheme 9). Depro-
tection of the primary TBS ether in the presence of the
primary TIPS ether19 was somewhat problematic, with
deprotection of the TIPS ether being competitive. To
obtain quantities of the primary alcohol 28 we settled
on the unoptimized procedure of treating bis-silyl ether
27 with catalytic sulfuric acid in THF/H2O, stopping the
reaction after partial conversion, and resubjecting the
recovered starting material. After three iterations, pri-
mary alcohol 28 was obtained in 63% yield along with
17% of recovered starting material. Primary alcohol 28
was oxidized to the aldehyde with Dess-Martin perio-
a
Key: (a) (i) 9-BBN, 50 °C, 10 h, (ii) [O]; (b) (i) 9-BBN, 80 °C,
48 h, (b) [O].
point to remove the (S)-methylbenzyl protecting group,
which was incompatible with several of the later steps
in our synthetic plan. Eventually, we planned to use the
p-methoxy version of this group, obviating this swap. The
methylbenzyl ether of 24 was removed with a dissolving
metal reduction in 74% yield, followed by the protection
of diol 25 as the bis-PMB ether, 26. Hydroboration of 26
proceeded to give a single diastereomer, which was
predicted by addition of the bulky reagent trans to the
allylic ether functionality at C42. Protection of the
resultant hydroxyl as the TIPS ether provided 27 in 60%
yield for the two-step procedure.
(19) Cunico, R. F.; Bedell, L. J . Org. Chem. 1980, 45, 4797-4798.

Synthesis of the C29-C44 Portion of Spongistatin 1
J . Org. Chem., Vol. 65, No. 13, 2000 4149
Sch em e 10a
Sch em e 11
a
uct 36, resulting from retro-Michael reaction of the
undesired enolate, was produced in small amounts
(Scheme 11). When this R,â-unsaturated ketone was
treated with K2CO3 in MeOH it cyclized to return ketone
35a with good stereochemical integrity. Upon extended
treatment under these conditions the C38 stereocenter
of 35a was equilibrated to a 6.5:1 mixture, favoring the
desired C38-(S) configuration. At this point, we carried
out molecular mechanics calculations of compounds of
this general structure and discovered, that, indeed, one
stereoisomer is favored over the other three. Of course,
the bulky side chain favors the equatorial position, and
there is precedent for this in analogous C-glycoside
chemistry.22 However, the C2 methyl group also has an
ordering effect, forcing the side chain to orient itself so
as to avoid potentially serious syn-pentane interactions
Key: (a) (i) LDA; (ii) TBSCl, HMPA; (b) 3, BF3‚OEt2, CH2Cl2;
(c) 0.5% H2SO4, THF, H2O.
dinane.20 The crude aldehyde was treated with methyl-
lithium to provide a diastereomeric mixture of alcohols,
which were oxidized a second time with Dess-Martin
periodinane to provide the desired ketone in 67% overall
yield. With ketone 29 in hand, we were prepared to
explore the aldol coupling with a suitably protected
aldehyde.
The coupling of ketone 29 and aldehyde 3 (synthesized
as described by Evans and co-workers8a) and transforma-
tion into the E-F ring system of the spongistatins is
shown in Scheme 10. Ketone 29 was converted to silyl
enol ether 30 by treatment with LiHMDS followed by
quenching of the enolate with TBS-Cl in HMPA. The
crude silyl enol ether 30 was condensed with aldehyde 3
using BF3‚OEt2 as the catalyst to provide the desired
Felkin-Ahn product 33 in 50% yield along with the
undesired anti-Felkin-Ahn product 32 (15%) and some
recovered ketone 29. The aldol adduct 33 was treated
with catalytic acid in THF/H2O, which removed the TES
group and caused spontaneous cyclization to the hemiket-
al 34.
Structure 34 represents the successful completion of
the C29-C44 portion of the spongistatin structure. The
synthesis required 25 total steps, with a 2.4% overall
yield for the longest linear sequence (21 steps). Although
we were pleased with the success of these efforts, we
realized that if our ultimate goal of producing gram-
quantities of the natural product were to be met a more
efficient synthesis was needed. We now describe a
serendipitous discovery that lead to a dramatic simpli-
fication of our synthesis and ultimately allowed us to
produce gram quantities of the C29-C44 portion of
spongistatin.
(inset). Calculations of a simplified model (P2 and P4
)
Me; P1 and P3 ) TMS) showed that conformation B is
favored over conformation A by approximately 0.4 kcal
mol-1, corresponding to an equilibrium ratio of ap-
proximately 2:1. Thus, the alkyloxy group prefers the
position anti to the anomeric hydrogen; this may be a
simple steric effect or there may be a polar component.
This observation suggested that we might be able to
introduce the alkyloxyacetone side chain and deal with
the C38 and C39 configuration in a very straightforward
manner.
To take advantage of this discovery, we altered our
synthetic strategy so as to incorporate a C-glycosidation
followed by an equilibration to set the C38 and C39
stereocenters. The new approach (Scheme 12) began with
commercially available tri-O-acetyl glucal, which was
converted into glucal 37 using Danishefsky’s procedure.23
Selective protection of the 3-OH of 37 was accomplished
with TES-Cl in CH2Cl2, providing bis-silyl ether 38 in
81% yield. Cyclopropanation of 38, directed by the
unprotected 4-OH, gave the desired R-isomer in 94%
yield.24,25 Removal of the TES ether was accomplished
While generating the lithium enolate of advanced
intermediate 35a,21 we noticed that fragmentation byprod-
(20) (a) Dess, D. B.; Martin, J . C. J . Am. Chem. Soc. 1991, 113, 7277.
(b) Ireland, R. E.; Liu, L. J . Org. Chem. 1993, 58, 2899.
(21) Scott, R. W. Doctoral Thesis, University of California at
Berkeley, 1997.
(22) Allevi, P.; Anastasia, M.; Ciuffreda, P.; Fiecchi, A.; Scala, A. J .
Chem. Soc., Perkin Trans. 1. 1989, 1275-1280.
(23) Danishefsky, S. J .; Gordon, D. M. J . Am. Chem. Soc. 1992, 114,
659-663.

4150 J . Org. Chem., Vol. 65, No. 13, 2000
Wallace et al.
Sch em e 12a
Sch em e 14a
a
Key: (a) NaOMe, MeOH; (b) TIPSCl, imid., DMF; (c). TESCl,
imid., Et3N, CH2Cl2; (d) Et2Zn, CH2I2, toluene; (e) CSA, H2O, THF;
(f) NaH, PMBCl, DMF; (g) PyrHBr3, pyridine, H2O, THF; (h)
Bu3SnH, AIBN, toluene, ∆.
a
Key: (a) TFAA, pyr., DMAP, CH2Cl2; (b) 46, ZnBr2; (c)
NaHMDS, THF; (d) KOH, MeOH.
of the anomeric hydroxyl group. Addition of an excess of
silyl enol ether 46 and ZnBr2 provided the desired
C-glycosides 49a /b and 50a /b as a mixture of diastere-
omers in good overall yield. The crude reaction mixture
was treated with NaHMDS to effect the R to â equilibra-
tion, providing ketones 50a and 50b as a 2:1 mixture of
diastereomers. Ketones 50a and 50b were separated by
column chromatography and the undesired isomer 50b
was equilibrated to the desired ketone 50a with KOH in
MeOH. Following three equilibration cycles, the desired
ketone 50a was obtained in 79% overall yield, based on
glucal 43.
Although we were pleased with the success of this
strategy at providing ketone 50a , we were surprised at
the difference in equilibrium ratios of ketones 50a and
50b (2:1) compared to those seen with ketones 35a and
35b (6.5:1). Because we had changed both the protecting
groups on the C38 and C41 carbinols we were unable to
determine which change had caused the deterioration in
the ratio of 50a and 50b. To better understand the fac-
tors defining the thermodynamics in this system we
decided to synthesize a series of C-glycosides changing
the protecting groups on the C38 and C41 carbinols
individually.
The bis TIPS-protected 2-methyl glucal 54, needed for
our study, was synthesized in an analogous fashion to
43.28 Likewise, the syntheses of the chiral ketone coupling
fragments are analogous to that of ketone 45. Unfortu-
nately, formation of chiral silyl enol ether 52a was not
as regioselective as the analogous achiral silyl enol ether
46 (Scheme 15). In addition, our inability to readily
separate the undesired C-glycosides, resulting from reac-
tion of silyl enol ether 52b, forced us to adopt an
alternative route to the desired C-glycosides. We decided
to utilize the chiral bis-silyl ketene acetal 53a as a
surrogate for 52a . Treatment of acid 13b with LiHMDS
in the presence of TMSCl furnished the desired si-
lylketene acetal 53b in good yield.
Sch em e 13a
a
Key: (a) NaH, chloro-2-methyl-2-propene, DMF; (b) OsO4,
NaIO4, H2O/THF; (c) LiHMDS, TMSCl, THF.
with catalytic acid in THF/H2O in near-quantitative yield.
Diol 40 was then protected as the bis-PMB ether by
treatment with NaH and PMBCl in DMF/THF. Treat-
ment of 41 with pyridinium hydrobromide perbromide
in THF/H2O resulted in ring opening to the bromide 42
in 87% yield. Reduction of bromide 42 with Bu3SnH
provided the desired hemiacetal in good yield.
Synthesis of the alkyloxyacetone coupling partner is
shown in Scheme 13. p-Methoxybenzyl alcohol was
alkylated with 3-chloro-2-methylpropene, and the result-
ing alkene 4426 was oxidized to ketone 45 with OsO4 and
NaIO4. Conversion to the desired thermodynamic silyl
enol ether was accomplished by treating ketone 45 with
LiHMDS in the presence of TMS-Cl.27
With 2-methyl glucal 43 and silyl enol ether 46 in
hand, we were prepared to explore our C-glycosidation/
equilibration strategy (Scheme 14). After considerable
experimentation, a one-pot procedure for the anomeric
activation and C-glycosidation of 43 was developed.
Treatment of glucal 43 with trifluoroacetic anhydride in
the presence of pyridine and DMAP resulted in acylation
(24) Cope, A. C.; Moon, S.; Park, C. H. J . Am. Chem. Soc. 1962, 84,
4843-4849.
(25) Charette, A. B.; Cote, B. J . Org. Chem. 1993, 58, 933-936.
(26) Frejd, T.; Wennerberg, J . Acta Chem. Scand.1998, 52, 95-99.
(27) The double bond geometry of silyl enol ether 46 was determined
through NOE difference experiments performed by Dr. Shannon Chi.
(28) Wallace, G. A. Doctoral Thesis, University of California at
Berkeley, 2000.

Synthesis of the C29-C44 Portion of Spongistatin 1
J . Org. Chem., Vol. 65, No. 13, 2000 4151
Sch em e 15
Sch em e 17
Sch em e 16a
Sch em e 18a
a
Key: (a) (i) TFAA, pyr., DMAP, CH2Cl2; (ii) 59, ZnBr2; (b)
Me(MeO)NH‚HCl, DCC, HOAT, Et3N, DMF; (c) MeMgCl, THF.
Following the C-glycosidation procedure described
earlier, glucal 54 and silyl ketene acetal 53 were con-
densed to furnished acids 55 as a mixture of four
diastereomers. The diastereomeric mixture was con-
verted into the corresponding Weinreb amides. At this
point, it was convenient to separate the R- and â-isomers
by column chromatography. Treatment of the mixture of
â-Weinreb amides, 57a /b, with MeMgCl provided the
desired ketones, 58a and 58b.With the desired ketones
in hand, the equilibration studies were conducted where
each pair of structural isomers was separated and
individually equilibrated to the same diastereomeric
mixture, as determined by 1H NMR of the crude reaction
mixtures. The results of the equilibrations are shown in
Scheme 17. From the results, it is clear that both the
protecting groups at C41 and C38 have moderate effects
on the equilibrium ratio. Comparing entries 1 and 2, an
increase in the size of the C41 protecting group ac-
companied an increase in the ratio of the desired R-(S)-
isomer. We believe this shift in the equilibrium, although
small, arises from a perturbation of the pyranose ring
geometry. In addition, changing the C38 protecting group
from p-methoxybenzyl to (S)-methylbenzyl also shows an
increase in the ratio of the desired R-(S)-isomer (entries
1 and 3). Combined, these two effects are additive,
providing a 6.5:1 ratio of the R-(S)-isomer (entry 4).
However, the effect of the C38 protecting group is
sensitive to the absolute configuration of the methylben-
zyl group (compare entries 3 and 5). It appears from entry
a
Key: (a) (i) Bu3BOTf, (iPr)2EtN, (ii) 5-trimethylacetoxypen-
tanal, (iii) H2O2; (b) TESCl, imid., DMF; (c) LiSEt, THF; (d) Et3SiH,
Pd/C, CH2Cl2.
5 that the effect of the C38 protecting group is the
combination of several subtle factors that are not obvious
on first inspection.
Due to the added steps needed to prepare ketones such
as 58a and our overall protecting group strategy we
decided to continue with the synthesis of the C29-C44
portion of spongistatin with ketone 50a . To increase the
convergency of the later part of our synthesis, we altered
our synthesis of the C29-C35 aldehyde. The synthesis
of the required aldehyde is shown in Scheme 18. Using
Evans’ oxazolidinone chemistry, 5-trimethylacetoxy-pen-
tanal was condensed with acylated oxazolidinone 62.
Protection of the secondary alcohol of 63 as the TES ether
occurred in 95% yield. Removal of the Evan’s auxiliary
was accomplished with lithium ethylthiolate, providing
thioester 65 in 93% yield. Reduction to desired aldehyde
was accomplished with Et3SiH and Pd/C, providing the
desired aldehyde 66 in high yield.
Coupling of aldehyde 66 and ketone 50a is shown in
Scheme 19. Although the Mukaiyama aldol reaction had

4152 J . Org. Chem., Vol. 65, No. 13, 2000
Wallace et al.
Sch em e 19a
Sch em e 20a
a
Key: (a) TBSCl, KHMDS, THF; (b) TBAF, THF.
along with a small amount of the undesired alcohol 68
(19% yield), which could be recycled through the above
procedure.
Having developed an efficient method to set the C35
stereocenter, we protected the axial alcohol of 70 as the
TBS ether in good yield (Scheme 20). Removal of the
primary silyl ether was accomplished with TBAF, provid-
ing alcohol 72 in 81% yield. Bispyran 72 represents our
second generation synthesis of the C29-C44 portion of
spongistatin. This route requires 27 total steps and occurs
in a much improved 13.2% yield over the longest linear
sequence (18 steps). Through our efforts we have been
able to produce gram-quantities of alcohol 72. Investiga-
tions on appending the C45-C51 side chain and the
completion of the total synthesis of spongistatin 1 are
under way and will be reported in due course.
a
Key: (a) (i) LiHMDS, THF, (ii) 66, (iii) K2CO3; (b) Ph3PHBr,
MeOH/THF; (c) Dess-Martin periodinane, pyr., CH2Cl2; (d) L-
Selectride, THF.
provided the desired diastereomer as the major product,
our inability to increase the selectivity of this reaction
lead us to explore an alternate approach to set the C35
stereocenter. Under carefully optimized conditions, the
lithium enolate of ketone 50a was condensed with
aldehyde 60 to provide a diastereomeric mixture of aldol
adducts 67, in 88% yield. Removal of the TES silyl ether
and cyclization was accomplished with catalytic Ph3PHBr
in THF/MeOH, providing alcohols 68 and 70 in 87% yield
as a 3.4:1.0 ratio. At this point, it was possible to separate
the two diastereomers by column chromatography. The
undesired diastereomer 68 was oxidized with Dess-
Martin periodinane to the ketone 69 in high yield.
Reduction of the ketone from the less hindered face with
L-selectride provided the desired alcohol 70 (79% yield)
Ack n ow led gm en t. This work was supported by a
research grant from the United States Public Health
Service (AI15027).
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures and characterization for all new compounds reported
in this paper. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O0002801