A Second-Generation Synthesis of the C1-C28 Portion of the Altohyrtins (Spongistatins)

Article abstract of DOI:10.1021/ja030316h

A practical second-generation synthesis of an advanced intermediate in our total synthesis of altohyrtin C (spongistatin 2) has been developed. A new approach to the C1-C15 (AB) portion features a vinyllithium addition to an aldehyde followed by a palladium-catalyzed allylic reduction to install the troublesome C13-C15 segment. Our general approach to the C16-C28 (CD) spiroketal has been retained, but some improvements have been made. Most notably, the kinetically controlled CD-spiroketalization reaction now proceeds in high yield with excellent diastereoselection. This new strategy uses the anti-aldol coupling used in our first-generation synthesis to join AB and CD fragments. A total of 9.6 g of intermediate 57 has been produced using this improved route.

Full text of DOI:10.1021/ja030316h

Published on Web 09/27/2003
A Second-Generation Synthesis of the C1-C28 Portion of the
Altohyrtins (Spongistatins)
Jed L. Hubbs and Clayton H. Heathcock*
Contribution from the Center for New Directions in Organic Synthesis,
Department of Chemistry, UniVersity of California, Berkeley, California 94720
Received May 27, 2003; E-mail: heathcock@cchem.berkeley.edu
Abstract: A practical second-generation synthesis of an advanced intermediate in our total synthesis of
altohyrtin C (spongistatin 2) has been developed. A new approach to the C1-C15 (AB) portion features a
vinyllithium addition to an aldehyde followed by a palladium-catalyzed allylic reduction to install the
troublesome C13-C15 segment. Our general approach to the C16-C28 (CD) spiroketal has been retained,
but some improvements have been made. Most notably, the kinetically controlled CD-spiroketalization
reaction now proceeds in high yield with excellent diastereoselection. This new strategy uses the anti-
aldol coupling used in our first-generation synthesis to join AB and CD fragments. A total of 9.6 g of
intermediate 57 has been produced using this improved route.
In 1993, the groups of Pettit, Kitagawa, and Fusetani
independently reported the isolation and characterization of a
group of remarkably active antitumor compounds from marine
sponges. Pettit reported the isolation of spongistatin 1 from a
Spongia sp. in the Indian Ocean,1 Kitagawa described the
isolation of altohyrtin A from the Okinawan sponge Hyrtios
altum,2 and Fusetani reported the isolation of cinachyrolide A
off the coast of the Japanese island Hachijo-jima from a sponge
of the genus Cinachyra.3 In addition to these initial reports, a
number of additional compounds in this family were reported
by Pettit4 and Kitagawa.5 In 1994, on the basis of Mosher’s
ester and circular dichroism analysis, Kitagawa and co-workers
established the absolute and relative configurations of the
altohyrtins.6 The antitumor activity of this family of molecules
compounds has been described as “probably the best to date in
the NCI’s evaluation programs.”7 Altohyrtin A (spongistatin 1),
the most active member, has an average IC50 of 0.13 nM against
the NCI’s panel of 60 cancer cell lines and is even more active
against certain cell lines in the panel.8 The altohyrtins act by
inhibiting tubulin polymerization and seem to bind at a unique
site on the tubulin dimer.8
The altohyrtins have attracted the attention of a number of
other research groups, and so far there have been five total
syntheses.9-13 In addition, a number of other groups have
reported efforts toward the total synthesis of the altohyrtins.14
The syntheses reported to date have been “traditional”, in the
sense that they have contributed a great deal to our understand-
(1) Pettit, G. R.; Cichacz, Z. A.; Gao, F.; Herald, C. L.; Boyd, M. R.; Schmidt,
J. M.; Hooper, J. N. A. J. Org. Chem. 1993, 58, 1302.
(2) Kobayashi, M.; Aoki, S.; Sakai, H.; Kawazon, K.; Kihara, N.; Sasaki, T.;
Kitagawa, I. Tetrahedron Lett. 1993, 34, 1993.
(3) Fusetanai, N.; Shinoda, K.; Matsunaga, S. J. Am. Chem. Soc. 1993, 115,
3977.
(4) (a) Pettit, G. R.; Cichacz, Z. A.; Gao, F.; Herald, C. L.; Boyd, M. R. J.
Chem. Soc., Chem. Commun. 1993, 1166. (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. (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. (d) 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.
(9) (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, 2741-2741. (c) Evans, D.
A.; Trotter, B. W.; Coˆte´, B.; Coleman, P. J.; Dias, L. C.; Tyler, A. N.
Angew. Chem., Int. Ed. Engl. 1997, 36, 2744-2747. (d) Evans, D. A.;
Trotter, B. W.; Coleman, P. J.; Coˆte´, B.; Dias, L. C.; Rajapakse, H. A.;
Tyler, A. N. Tetrahedron 1999, 55, 8671-8726.
(10) (a) Guo, J.; Duffy, K. J.; Stevens, K. L.; Dalko, P. I.; Roth, R. M.; Hayward,
M. M.; Kishi, Y. Angew. Chem., Int. Ed. 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. 1998, 37, 192-196.
(5) Kobayashi, M.; Aoki, S.; Sakai, H.; Kawazon, K.; Kihara, N.; Sasaki, T.;
Kitagawa, I. Chem. Pharm. Bull. 1993, 41, 989.
(6) Kobayashi, M.; Aoki, S.; Kitagawa, I. Tetrahedron Lett. 1994, 35, 1243.
(7) Pettit, G. R. J. Nat. Prod. 1996, 59, 812-821.
(8) Bai, R.; Taylor, G. F.; Cichacz, Z. A.; Herald, C. L.; Kepler, J. A.; Pettit,
G. R.; Hamel, E. Biochemistry 1995, 34, 9714.
9
12836
J. AM. CHEM. SOC. 2003, 125, 12836-12843
10.1021/ja030316h CCC: $25.00 © 2003 American Chemical Society

Synthesis of the C1−C28 Portion of the Altohyrtins
A R T I C L E S
Scheme 1
ing of organic synthesis and, in particular, to the unique synthetic
challenges presented by the altohyrtin skeleton, but they have
not provided significantly more altohyrtin than has been isolated
from the natural sourcesnone of the previous syntheses has
provided more than 4 mg of final product. It is our goal to
develop a total synthesis that is sufficiently efficient that we
can prepare multigram quantities of altohyrtin C (spongistatin
2).
improvement if we are to prepare gram quantities of altohyrtins.
In this and the following Article, we report a second-generation
synthesis of the C1-C28 segment that has yielded almost 10
grams of the intermediate and the incorporation of this building
block in a total synthesis of altohyrtin C (spongistatin 2).
One of the problems in our first-generation synthesis was
the long linear synthesis of compound 3, representing C1-C15.
Furthermore, nine of these 33 steps, proceeding in only 41%
overall yield, were required to add C14 and C15 and install the
methylene group at C13. Thus, one goal of our second-
generation synthesis was to develop a convergent synthesis of
the A/B spiroketal moiety, and another was to develop a better
way of incorporating C13-C15. Indeed, we had already
developed such a convergent route for the synthesis of the C/D
spiroketal 4 (see Scheme 1). Another goal that we had in mind
in developing a revised synthesis of C1-C28 was to make the
synthetic approaches to the A/B and C/D spiroketal subunits as
similar as possible, so that we can start with identical or
enantiomeric starting materials and minimize the number of
chemically different operations. To meet these objectives, we
identified compound 6 as our target and envisioned this
spiroketal precursor as arising from unification of aldehyde 7
with methyl ketone 8 in a manner similar to that used for the
synthesis of the C/D spiroketal in our first-generation synthesis.
We have previously reported first-generation syntheses of the
C1-C28 and C29-C44 building blocks (Scheme 1).15,16 Our
synthesis of the C1-C28 segment 5 required 60 total steps, a
longest linear sequence of 36 steps, and gave an overall yield
for this sequence of 0.86% (somewhat higher if one includes
recycling at several stages). The synthesis permitted us to
prepare several hundred milligrams of 5, but needed significant
(11) (a) Smith, A. B., III; Doughty, V. A.; Lin, Q.; Zhuang, L.; McBriar, M.
D.; Boldi, A. M.; Moser, W. H.; Murase, N.; Nakayama, K.; Sobukawa,
M. Angew. Chem., Int. Ed. 2001, 40, 191. (b) Smith, A. B., III; Lin, Q.;
Doughty, V. A.; Zhuang, L.; McBriar, M. D.; Kerns, J. K.; Brook, C. S.;
Murase, N.; Nakayama, K. Angew. Chem., Int. Ed. 2001, 40, 196. (c) Smith,
A. B., III; Doughty, V. A.; Sfouggatakis, C.; Bennett, C. S.; Koyanagi, J.;
Takeuchi, M. Org. Lett. 2002, 4, 783.
(12) (a) Paterson, I.; Oballa, R. M.; Norcross, R. D. Tetrahedron Lett. 1996,
37, 8581. (b) Paterson, I.; Gibson, K. R.; Oballa, R. M. Tetrahedron Lett.
1996, 37, 8585. (c) Paterson, I.; Keown, L. E. Tetrahedron Lett. 1997, 38,
5727. (d) Paterson, I.; Oballa, R. M. Tetrahedron Lett. 1997, 38, 8241. (e)
Paterson, I.; Wallace, D. J.; Gibson, K. R. Tetrahedron Lett. 1997, 38,
8911. (f) Paterson, I.; Chen, D. Y.-K.; Coster, M. J.; Acen˜a, J. L.; Bach,
J.; Gibson, K. R.; Keown, L. E.; Oballa, R. M.; Trieselmann, T.; Wallace,
D. J.; Hodgson, A. P.; Norcross, R. D. Angew. Chem., Int. Ed. 2001, 40,
4055-4060.
(13) (a) Crimmins, M. T.; Washburn, D. G. Tetrahedron Lett. 1998, 39, 7487.
(b) Crimmins, M. T.; Katz, J. D.; McAtee, L. C.; Tabet, E. A.; Kirincich,
S. J. Org. Lett. 2001, 3, 949-952. (c) Crimmins, M. T.; Katz, J. D. Org.
Lett. 2000, 2, 957-960. (d) Crimmins, M. T.; Katz, J. D.; Washburn, D.
G.; Allwein, S. P.; McAtee, L. F. J. Am. Chem. Soc. 2002, 124, 5661-
5663.
(14) For leading references to synthetic approaches from other laboratories, see
the list in ref 16, plus the following: (a) Zuev, D.; Paquette, L. A. Org.
Lett. 2000, 2, 679-682. (b) Terauchi, T.; Nakata, M. Tetrahedron Lett.
1998, 39, 3795-3798. (c) Lemaire-Audoire, S.; Vogel, P. Tetrahedron Lett.
1998, 39, 1345-1348. (d) Lemaire-Audoire, S.; Vogel, P. J. Org. Chem.
2000, 65, 3346-3356. (e) Zemribo, R.; Mead, K. T. Tetrahedron Lett.
1998, 39, 3895-3898. (f) Fernandez-Megia, E.; Gourlaouen, N.; Ley, S.
V.; Rowlands, G. J. Synlett 1998, 991-994. (g) Micalizio, G. C.; Pinchuk,
A. N.; Roush, W. R. J. Org. Chem. 2000, 65, 8730-8736. (h) Anderson,
J. C.; McDermott, B. P. Tetrahedron Lett. 1999, 40, 7135-7138. (i) Samadi,
M.; Munoz-Letelier, C.; Poigny, S.; Guyot, M. Tetrahedron Lett. 2000,
41, 3349-3353. (j) Terauchi, T.; Terauchi, T.; Sato, I.; Tsukada, T.; Kanoh,
N.; Nakata, M. Tetrahedron Lett. 2000, 41, 2649-2653. (k) Kary, P. D.;
Roberts, S. M. Tetrahedron: Asymmetry 1999, 10, 217-219. (l) Kim, H.;
Hoffmann, M. R. Eur. J. Org. Chem. 2000, 2195-2201. (m) Barrett, A.
G. M.; Braddock, D. C.; de Koning, P. D.; White, A. J. P.; Williams, D.
J. J. Org. Chem. 2000, 65, 375-380.
The improved route to intermediate 7 starts with (S)-malic
acid (Scheme 2) and utilizes a sequence similar to that
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12837

A R T I C L E S
Hubbs and Heathcock
Scheme 2 a
a (a) EtOH, SOCl2; (b) BMS, NaBH4, THF; (c) TIPSCl, imidazole, DMF; (d) tert-butylthio acetate, THF; (e) Et2BOMe, NaBH4, THF, MeOH; (f)
4-MeOPhCH(OMe)2, PPTS, CH2Cl2; (g) TBAF, THF; (h) 4-ClPhSO2Cl, Et3N, CH2Cl2; (i) NaI, NaHCO3, Na2SO3, MEK; (j) vinylmagnesium bromide,
Li2CuCl4, THF; (k) O3, MeOH, CH2Cl2, Bu3P.
Scheme 3 a
previously employed in the synthesis of the CD spiroketal.17
Diethyl (S)-malate, formed by esterification with thionyl chloride
in EtOH, was regioselectively reduced with borane-dimethyl
sulfide complex and catalytic sodium borohydride18 to produce
diol 10 in 83-94% yield. Diol 10 was protected as the primary
TIPS ether immediately after purification to avoid lactonization,
and the resulting compound was converted to the â-keto ester
by a mixed Claisen condensation with the lithium enolate of
tert-butyl acetate. The C3 ketone carbonyl was stereoselectively
reduced to 1,3-syn diol 11 by a chelation-controlled reduction
with diethyl methoxy borane and sodium borohydride.19 Treat-
a (a) NaH, BnBr, DMF; (b) isopropenylmagnesium bromide, Li2CuCl4,
THF; (c) TESCl, imidazole, DMF; (d) (i) O3, NaHCO3, MeOH, CH2Cl2;
(ii) DMS.
Scheme 4 a
ment of diol 11 with anisaldehyde dimethyl acetal and PPTS
afforded the p-methoxybenzylidine acetal which was desilylated
to alcohol 12. Alcohol 12 was homologated to aldehyde 7 by a
four-step sequence involving formation of the p-chloroben-
zenesulfonate,20 which was displaced with NaI in methyl ethyl
ketone to obtain iodide 13. The iodide was displaced by the
cuprate derived from vinylmagnesium bromide and catalytic Li2-
CuCl4, and the resulting alkene was carefully ozonized (Sudan
III dye indicator) to give aldehyde 7.21
Ketone 8 was prepared by the efficient four-step sequence
shown in Scheme 3. When (S)-glycidol benzyl ether (15)22 was
treated with isopropenylmagnesium bromide and Li2CuCl4, the
epoxide was smoothly opened by the vinyl cuprate to give
homoallylic alcohol 16. This material was treated with TESCl
and imidazole to protect the hydroxyl group, and the alkene
was ozonized to give ketone 8 in 97% yield from epoxy ether
15.
Spiroketalization precursor 6 was prepared as shown in
a (a) LDA, then 7; (b) Dess-Martin, 1 equiv of H2O, THF, CH2Cl2; (c)
Scheme 4. The lithium enolate of ketone 8 reacted with aldehyde
7 to give a mixture of aldol diastereomers (17) in 80-88% yield.
These aldols were oxidized with the Dess-Martin reagent to
10:10:1 THF:CH3CN:12 N HCl.
give diketone 6, which was isolated as its enolic form in 80-
84% yield. The Dess-Martin oxidation reaction deserves special
comment. A survey of a variety of oxidizing reagents was
explored.23 Standard Dess-Martin conditions (periodinane, CH2-
Cl2)24 led to irreproducible yields (30-80%).25 In the course
of optimizing the reaction conditions, we noticed that higher
yields resulted when a heavy precipitate formed as the reaction
progressed. We hypothesized that this was an iodine-III species
(15) (a) Claffey, M. M.; Heathcock, C. H. J. Org. Chem. 1996, 61, 7646-
7647. (b) Hayes, C. J.; Heathcock, C. H. J. Org. Chem. 1997, 62, 2678-
2679. (c) Claffey, M. M.; Hayes, C. J.; Heathcock, C. H. J. Org. Chem.
1999, 64, 8267-8274.
(16) Wallace, G. A.; Scott, R. W.; Heathcock, C. H. J. Org. Chem. 2000, 65,
4145-4152.
(17) The enantiomer of 9 was used, see ref 15.
(18) Saito, S.; Ishikawa, T.; Kuroda, A.; Koga, K.; Moriwake, T. Tetrahedron
1992, 48, 4067.
(19) Chen, K.-M.; Hardtmann, G. E.; Prasad, K.; Repic, O.; Shapiro, M. J.
Tetrahedron Lett. 1987, 28, 155.
(20) It was necessary to use the p-chlorosulfonate instead of the more common
tosylate because the p-anisilidine acetal was hydrolyzed under the reaction
conditions with longer reaction times necessary to displace the tosylate
even in the presence of NaHCO3 and Na2SO3.
(23) Oxidizing reagents surveyed include Moffat-Swern; CrO3, pyridine; IBX;
PDC; TPAP, NMO; Dess-Martin, pyridine.
(24) (a) Dess, D. B.; Martin, J. C. J. Am. Chem. Soc. 1991, 113, 7277. (b) Dess,
D. B.; Martin, J. C. J. Am. Chem. Soc. 1983, 105, 4155.
(25) Problems with oxidation of aldol products under Dess-Martin conditions
have been noted previously: Meyer, S. D.; Schreiber, S. L. J. Org. Chem.
1994, 59, 7549.
(21) Use of tributylphosphine as the reducing agent minimized methanolysis of
the acetal in this step.
(22) Tse, B. J. Am. Chem. Soc. 1996, 118, 7094. Compound 15 is also
commercially available from the Aldrich Chemical Co.
9
12838 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Synthesis of the C1−C28 Portion of the Altohyrtins
A R T I C L E S
Scheme 5
Scheme 6
Scheme 7 a
and that, if it were to remain in solution, it might have a
deleterious effect by oxidizing the diketone product.26 When
the oxidation was carried out in 1:1 THF:CH2Cl2 with 1 equiv
of water, the amount of precipitate was maximized, and we were
able to obtain reproducibly excellent yields of the enolic
diketone 6. This observation is in contrast to the report of Meyer
and Schreiber, who advise the use of Dess-Martin periodinane
in methylene chloride under strictly anhydrous conditions for
the oxidation of â-hydroxy esters to â-keto esters.27 These
authors report that water accelerates Dess-Martin oxidations
and increases the amount of tricarbonyl compound, whereas our
conditions decelerate the oxidation and result in less tricarbonyl
compound.
a (a) MeLi, Et2O; (b) LiHMDS, 4-chloro-2-aminopyridyl triflimide, THF;
(c) Bu3SnH, LDA, CuCN, THF; (d) NBS, CH2Cl2.
group for the C12 primary alcohol, rather than the much less
convenient p-methoxyphenyl group. However, treatment of 22
with 5:5:1 THF, acetonitrile, and 6 N HCl gave 59% yield of
the desired spiroketal 18, accompanied by 30% yield of the
dihydropyrone 23. Further treatment of 23 with aqueous acid
did result in formation of more 18, but also resulted in some
decomposition, apparently due to hydrolysis of the tert-butyl
ester and dehydration of one or more of the secondary alcohols.
It was this result that prompted us to use a more acid-labile
benzylidine group for protection of the C3,C5 diol.
One way in which we hoped to improve on our first-
generation synthesis was the method for adding the “orphan”
stereocenter at C14. Because this stereocenter is not part of a
rigid ring, as are those in the spiroketal moiety, it must be
introduced by absolute stereocontrol using a chiral reagent or
by using a “chiral pool” building block. In our previous
synthesis, we introduced this center by use of an Evans
asymmetric aldol reaction. This necessitated a number of steps
to change a C13 hydroxy group into the required methylene
group and contributed greatly to the long linear sequence of
steps required to synthesize the full C1-C15 building block.
In our revised approach, we elected to use a chiral building
block that would carry both the necessary C13 methylene group
and the C14 stereocenter. Such a starting material is (S)-2-
methyl-3-hydroxypropanoic acid, which has previously been
converted into Weinreb amide 24 by Paterson and co-workers.30
As shown in Scheme 7, this material is converted to a methyl
ketone which is deprotonated by LiHMDS and the resulting
enolate treated with 4-chloro-2-aminopyridyl triflimide to obtain
vinyl triflate 25. The triflate was converted to the vinylstannane
by a copper-mediated coupling reaction, and this stannane was
readily converted to vinyl bromide 26 upon treatment with NBS.
Treatment of 6 with HCl in THF-CH3CN gave spiroketal
18 in 82% yield. As expected, only one diastereomer (R
configuration at the anomeric position) was produced in the
spiroketalization reaction because the most stable chair-chair
conformation of this diastereomer has both carbon substituents
equatorial and both oxygens axial with respect to the other ring,
thus profiting from a “double anomeric effect”.28
Success in the spiroketalization reaction depends strongly on
the correct choice of protecting groups for the C3,C5 diol and
C11 hydroxy group. In preliminary investigations, we explored
the use of a simple benzylidine group to protect the C3 and C5
syn diol. As shown in Scheme 5, this derivative (19)29 was
protected by a benzyl at C11, so that all three OH groups (C3,
C5, and C11) could be revealed at one time, prior to the
spiroketalization. After catalytic hydrogenolysis of the benzyl
ethers, the resulting triol was treated with HCl in aqueous
acetonitrile to get a mixture of dihydropyrone 20 and spiroketal
21. Attempts to convert 20 into the desired product by longer
treatment under acidic conditions sufficed only to destroy 20,
probably by dehydration of the C11 alcohol and by hydrolysis
of the tert-butyl ester. The best isolated yield of spiroketal 21
was 32%.
We also investigated compound 22,27 in which the C11
hydroxy group is protected as a triethylsilyl ether (Scheme 6).
In this case, both the TES and the benzylidine groups can be
removed by treatment with acid, so we were able to use a benzyl
(26) The oxidation of diketones by iodine-III species is a well-known process,
see for a review: Moriarty, R. M.; Prakash, O. Org. React. 1999, 54, 273.
We have also isolated the triketone product in a similar system: Claffey,
M. M.; Heathcock, C. H., unpublished results.
(27) Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549-7552.
(28) Deslongchamps, P.; Rowan, D. D.; Pothier, N.; Sauve´, G.; Saunders, J. K.
Can. J. Chem. 1981, 59, 1105.
(29) See the Supporting Information for the reactions used and yields obtained
in the preparation of 19 and 22.
(30) Paterson, I.; Florence, G. J.; Gerlach, K.; Scott, J. P.; Sereinig, N. J. Am.
Chem. Soc. 2001, 123, 9535.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12839

A R T I C L E S
Hubbs and Heathcock
Scheme 8 a
a (a) MeLi, CeCl3, THF; (b) Ac2O, DMAP, CH2Cl2; (c) TESOTf, 2,6-lutidine, CH2Cl2; (d) H2, Pd(OH)2, THF; (e) Moffat-Swern; (f) bromide 26 plus
2 equiv of t-BuLi, Et2O, then 29; (g) N-formylbenzotriazole, DMAP, CH2Cl2; (h) Pd2(dba)3, Bu3P, NH4CO2H, cyclohexane; (i) DDQ, CH2Cl2, H2O; (j)
Dess-Martin, CH2Cl2.
The remainder of the C1-C15 synthesis is summarized in
Scheme 8. We first converted the ketone of 18 to the tertiary
alcohol by the addition of methyllithium in the presence of
CeCl3. The secondary alcohol of the resulting diol was
selectively protected with acetic anhydride and DMAP to give
diester 27. The tertiary alcohol was protected with TESOTf,
and the benzyl ether was removed by hydrogenolysis to give
alcohol 28 in 72% yield over the four steps from 18. Alcohol
18 could be converted to aldehyde 29 under Moffat-Swern
conditions.31 Bromide 26 (prepared as shown in Scheme 7) was
lithiated with tert-butyllithium and the resulting vinyllithium
reagent was treated with aldehyde 29 to obtain alcohol 30 as a
1:1 mixture of diastereomers. The alcohol was formylated by
In this plan, we have made two significant changes from our
first-generation synthesis of the C16-C28 segment. First,
compound 34 has one less carbon in the chain than the
spiroketalization substrate previously employed. Second, we
have chosen a pivaloyl ester as the terminal protecting group
instead of the TBS ether used in the first-generation synthesis,
mainly because our previous experience with TBS ethers has
shown that they are somewhat labile to the hydrogenolysis
Katritzky’s method,32 and the resulting allylic formate was
reduced by Tsuji’s method,33 using Pd(PBu3)4 in the presence
of ammonium formate to give the reduction product 31 in 48%
yield from alcohol 28 as the only observed regioisomer.34 The
PMB protecting group in 31 could be removed under standard
conditions with DDQ35 to give alcohol 32 in 92% yield. This
alcohol was oxidized with Dess-Martin reagent to give
aldehyde 33. This compound is quite prone to double bond
conditions necessary to remove the benzyl and benzylidine
isomerization and is used in the aldol coupling without
groups. In addition, we thought that the similar steric environ-
purification. The conversion of alcohol 28 into alcohol 32
ment of the two terminal positions in 34 might make it difficult
proceeds in 44% overall yield, as compared to 41% overall yield
to selectively remove a TBS protecting group in the presence
for a similar conversion in our first-generation synthesis.
However, the new route only requires five steps rather than nine
steps, resulting in significant savings in reagent costs. Most
importantly, it avoids the technically challenging Tebbe meth-
ylenation that was used to install the C13 methylene group in
our original route.
of a TIPS protecting group.36 To assemble the spiroketalization
precursor 34, we needed aldehyde 35 and ketone 36. Aldehyde
35 was prepared as we have previously described.15c
We started the synthesis of ketone 36 with (R)-glycidol (37),
as shown in Scheme 9. Protection of the alcohol and opening
of the epoxide with isopropenylmagnesium bromide in the
presence of catalytic Li2CuCl4 gave secondary alcohol 38.37
Benzylidine acetal 39 was prepared by treating monoether 38
with benzaldehyde dimethyl acetal and camphorsulfonic acid.
Reduction of 39 with diisobutylaluminum hydride resulted in
regioselective cleavage of the benzylidine group, leading to the
primary alcohol, which was protected as the pivaloyl ester.
Now that after we had developed a more efficient synthesis
of the AB spiroketal, we turned our attention to improving our
first-generation synthesis of the CD spiroketal moiety, ketone
4 (see Scheme 1). Our revised retrosynthetic plan is summarized
below:
(31) Tidwell, T. T. Org. React. 1990, 39, 297-572.
(32) Katritzky, A. R.; Chang, H. X.; Yang, B. Synthesis 1995, 503.
(33) (a) Tsuji, J.; Yamakawa, T. Tetrahedron Lett. 1979, 7, 613. (b) Tsuji, J.;
Shimizu, I.; Minami, I. Chem. Lett. 1984, 1017.
(36) Recall that our first-generation spiroketalization substrate had one more
carbon in the chain, making the TBS ether in that case significantly less
sterically hindered.
(37) Migration of the pivaloyl protecting group to the secondary alcohol was
observed during the epoxide opening when the more direct route from
glycidol pivalate was attempted.
(34) Care had to be taken to avoid overreduction to the alkane; TLC analysis
with 30% ethyl acetate in benzene separates the alkene from the overreduced
product.
(35) Oikawa, Y.; Yoshioka, T.; Yonemitsu, O. Tetrahedron Lett. 1982, 23, 885.
9
12840 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Synthesis of the C1−C28 Portion of the Altohyrtins
A R T I C L E S
Scheme 9 a
a (a) TBSCl, imidazole, DMF; (b) isopropenylmagnesium bromide, Li2CuCl4, THF; (c) PhCH(OMe)2, CSA, cat. MeOH, DMF; (d) DIBAL, CH2Cl2; (e)
PivCl, Et3N, DMAP, CH2Cl2; (f) OsO4, NaIO4, THF, H2O; (g) LDA, then 35; (h) Dess-Martin, H2O, CH2Cl2, THF; (i) H2, Pd(OH)2, THF; (j) ZnBr2,
CH2Cl2; (k) TBSOTF, 2,6-lutidine, CH2Cl2; (l) NaBH4, CeCl3, MeOH; (m) Cs2CO3, NaH, MeI.
Cleavage of the double bond under Johnson-Lemieux condi-
tions38 delivered ketone 36 in 49% yield from (R)-glycidol
without purification of any of the intermediates. Ketone 36 was
coupled to aldehyde 35 by a lithium enolate aldol reaction to
give hydroxy ketone 40 as a mixture of diastereomers. This
mixture was oxidized with the Dess-Martin conditions devel-
oped for the AB spiroketal (vide supra) to give the diketone,
which exists completely in the enolic form 34, in 78% yield.
The benzyl ether and benzylidine acetal were cleaved by
hydrogenolysis, and the resulting crude triol was treated with
anhydrous ZnBr2 in methylene chloride. Remarkably, R spiro-
ketal 41 was produced in greater than 80% yield as the only
observed diastereomer! This was quite a welcome result, because
the similar spiroketalization in our first-generation synthesis15c
had delivered the R and S spiroketals as a 7:1 mixture in a total
yield of about 60%. Without further purification, compound 41
was silylated with tert-butyldimethylsilyl triflate. In their
altohyrtin synthesis, Crimmins and co-workers found that a
similar ketone can be reduced selectively to the desired
equatorial alcohol with hydride reducing reagents.39 A quick
survey of such reducing agents revealed that ketone 41 is
selectively reduced under Luche conditions, giving the desired
equatorial isomer exclusively. This alcohol was then methylated
with MeI, Cs2CO3, and NaH to give compound 58.40 The five-
step sequence that converts diketone 34 to spiroketal 42 in 66%
yield is a significant improvement over the similar sequence
used in the first-generation synthesis, which gave a comparable
intermediate in only 29% yield.
a kinetic result. Yet why should the R spiroketal be the kinetic
product of ketalization? To understand this result, it is first
necessary to realize that the stereochemistry of nucleophilic
additions to cyclic oxocarbenium ions is subject to the same
stereoelectronic effect41 that operates in additions to cyclic
enones42 and imines.43 That is, when an alcohol adds to
oxocarbenium ion 43, the kinetic product for stereoelectronic
reasons should be 44 in which the new C-O bond is anti-
coplanar with one of the oxygen lone pairs, not 45, in which
the new bond is gauche to both lone pairs. Thus, if the
conformation is as shown in the following example, the newly
formed ketal bond will be trans to the substituent at C5:
The remarkable stereoselectivity observed in the C/D spiro-
ketalization reaction deserves comment. Previous studies15b have
shown that the R and S spiroketals are of comparable stability
and that extended treatment of the R spiroketal results in
isomerization to the S configuration. Therefore, the high
stereoselectivity observed in the conversion of 34 to 41 is clearly
Of course, the alternative half-chair conformation of oxocar-
benium ion 43 (shown above as 43a) could undergo axial
(41) Deslongchamps, P. Stereoelectronic Effects in Organic Chemistry. In
Organic Chemistry Series; Baldwin, J. E., Ed.; Pergamon Press: Oxford,
1989; Vol. 1.
(42) (a) Corey, E. J. Experentia 1953, 9, 329. (b) Corey, E. J. J. Am. Chem.
Soc. 1954, 76, 175. (c) Corey, E. J.; Sneen, R. A. J. Am. Chem. Soc. 1956,
78, 6229.
(38) Pappo, R.; Allen, D. S., Jr.; Lemieux, R. U.; Johnson, W. S. J. Org. Chem.
1956, 21, 478.
(39) Crimmins, M. T.; Katz, J. D. Org. Lett. 2000, 2, 957. They used L-Selectride
for their reduction.
(40) These unusual conditions were originally developed in a trial reaction where
Cs2CO3 and MeI did not effect the desired reaction so NaH was added to
the reaction mixture.
(43) (a) Stevens, R. V.; Lee, A. W. M. J. Chem. Soc., Chem. Commun. 1982,
102. (b) Stevens, R. V.; Lee, A. W. M. J. Chem. Soc., Chem. Commun.
1982, 103.
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12841

A R T I C L E S
Hubbs and Heathcock
Scheme 10
Scheme 11 a
addition of the alkoxy group, leading to conformer 45a, which
would eventually isomerize to the cis isomer 45. This possibility
is disfavored both by the fact that conformer 43a is probably
somewhat disfavored, relative to 43, and by the steric hindrance
that would result from interaction of the incoming axial OR
group with the R group.44 Another possibility for the observed
kinetic stereoselectivity in the spiroketalization process, sug-
gested by a reviewer, is that the zinc cation templates the R
configuration at the spiroketal center by chelation to the C25
axial hydroxyl, the spiroketal O, and one of the C17 pivalolyoxy
oxygens, along the lines suggested by Evans.9c However, the
stereoselectivity observed in the present case is much greater
than that seen by Evans, who used zinc chloride in methylene
chloride to equilibrate a 1:6 ratio of R and S spiroketals (obtained
by protic acid spiroketalization) to a 4.3:1 ratio.
a (a) DIBAL, CH2Cl2; (b) TsCl, Et3N, CH2Cl2; (c) NaI, MEK, reflux.
Scheme 12 a
Because we have chosen a spiroketalization substrate with a
carbonyl group at C21 (altohyrtin numbering), and because the
1,3-diketone exists completely in the intramolecularly hydrogen-
bonded form illustrated in Scheme 10, the first ketal bond to
form involves the C27 hydroxy group, leading to postulated
intermediate oxocarbenium ion 47. After ketonization of the
enol in 47, postulated intermediate 48, now freed of its
constraints, can undergo addition of the C19 hydroxy group.
The stereoelectronic principle now ensures that the C19 hydroxy
group will add trans to the C27 substituent, thus delivering the
R-configured spiroketal.45 We think that the higher stereose-
lectivity observed in the present case, relative to the very similar
one employed in our first-generation synthesis, is due to slower
equilibration of the kinetic product in the current case, possibly
because the deletion of one carbon in this case has brought the
pivaloyloxy group closer to the oxocarbenium ion, thus retarding
subsequent ketal isomerization.46
a (a) DIBAL, CH2Cl2; (b) Moffat-Swern; (c) Ph3PCH2, THF; (d) (i)
9-BBN; (ii) H2O2, NaOH; (e) Dess-Martin, CH2Cl2; (f) EtMgBr, THF;
(g) Dess-Martin, CH2Cl2.
Now that the difficult transformations were behind us, we
were ready to convert compound 42 to ethyl ketone 4. Our
original plan for carrying out this transformation is shown in
Scheme 11. In a manner similar to our conversion of iodide 13
to aldehyde 7 (see Scheme 2), we planned to couple iodide 49
with a vinyl cuprate 50 to obtain an alkene that could be
oxidatively cleaved to give ketone 4. To this end, the pivaloyl
protecting group in 41 was removed by reduction with diisobu-
tylaluminum hydride, and the resulting primary alcohol was
converted to the tosylate with tosyl chloride and Et3N. Unfor-
tunately, all attempts to convert this sulfonate to the necessary
iodide failed because the oxygen of the methyl ether readily
displaces the leaving group at C18 to give the tricyclic ether
53.
(44) For a similar example and detailed analysis of the situation, see: Pothier,
N.; Goldstein, S.; Deslongchamps, P. HelV. Chim. Acta 1992, 75, 604-
620.
(45) The same stereoelectronic preference operates in the formation of the A/B
spiroketal 18 (Scheme 4). In that case, axial addition of the C11 hydroxy
group should give the observed S spiroketal because addition should be
trans to the substituent at C3, which in this case has the R configuration.
(46) For examples in which proximal heteroatom substituents stabilize glycosides
with respect to acid-catalyzed hydrolysis, see: (a) Roush, W. R.; Lin, X.-
F. J. Am. Chem. Soc. 1995, 117, 2236. (b) Overend, W. G.; Rees, C. W.;
Sequeira, J. S. J. Chem. Soc. 1962, 3429. (c) Buncel, E.; Bradley, P. R.
Can. J. Chem. 1967, 45, 515.
To avoid this problem, a more circuitous route to ketone 4
was employed (Scheme 12). After the pivaloyl group was
removed by reduction with diisobutylaluminum hydride, the
resulting alcohol was oxidized under Moffat-Swern conditions,
and the aldehyde was subjected to Wittig methylenation to
obtain alkene 54 in 76% yield from 41. This alkene was
9
12842 J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003

Synthesis of the C1−C28 Portion of the Altohyrtins
A R T I C L E S
Scheme 13 a
Aldehyde 33 was added to the resulting (E)-enolate to give 67%
yield of the desired anti-aldol product 56, which was readily
acetylated with acetic anhydride and DMAP to give diacetate
57 in 93% yield. The spectroscopic properties of this compound
were identical to those produced by our first-generation
synthesis. As in our previous report,15c the major aldol product
is accompanied by a small amount (about 7%) of an isomeric
aldol resulting from enolization of ketone 4 at C4 rather than
C2. Both Evans9d and Patterson12f reported a 9:1 ratio of aldols
in a similar coupling of a C1-15 aldehyde with the dicyclo-
hexylboron enolate of a C16-29 ketone in which the substrates
differed by remote protecting groups. These groups both refer
to a “9:1 diastereomeric ratio” in the aldol reaction, but neither
provided definitive evidence that the minor aldol was a
diastereomer rather than a regioisomer.
The revised synthetic route described in this Article requires
a total of 62 steps, with a longest linear sequence of 35 steps.
We have prepared a total of 9.6 g of compound 57 by this route.
The total synthesis of altohyrtin C (spongistatin 2) is described
in the accompanying paper.
a (a) (c-hex)2BCl, Et3N, hexanes; (b) Ac2O, DMAP, CH2Cl2.
converted to the terminal alcohol by hydroboration with 9-BBN
followed by oxidation under standard conditions. Oxidation of
the resulting alcohol with Dess-Martin periodinane gave
aldehyde 55. Addition of ethylmagnesium bromide gave a
mixture of secondary alcohol diastereomers that were oxidized
under Dess-Martin conditions to give our ketone 4 in 83% yield
from alkene 54. Ketone 4 was prepared in 9.0% overall yield
for the 21 linear steps starting from (R)-malic acid.
Acknowledgment. This work was supported by a research
grant from the United States Public Health Service (AI15027).
The Center for New Directions in Organic Synthesis is supported
by Bristol-Myers Squibb as a Sponsoring Member and Novartis
as a Supporting Member.
Now that we had our requisite aldehyde 33 and ketone 4, we
were ready to carry out the critical anti-aldol reaction. This
reaction is now well precedented and proceeded without event
as shown in Scheme 13. Ethyl ketone 4 was converted into the
(E)-enolate by reaction with (c-hex)2BCl and triethylamine.
Supporting Information Available: Experimental procedures
and characterization for all new compounds (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA030316H
9
J. AM. CHEM. SOC. VOL. 125, NO. 42, 2003 12843