Total synthesis of (-)-dictyostatin: Confirmation of relative and absolute configurations

Article abstract of DOI:10.1002/anie.200460593

Will the real dictyostatin please stand up? There were five finalists as stereostructures for the potent anticancer agent dictyostatin; ten, if one were to include enantiomers. A total synthesis of (-)-dictyostatin (1) has ended the decade-old masquerade

Full text of DOI:10.1002/anie.200460593

Communications
Natural Products Synthesis
Total Synthesis of (À)-Dictyostatin: Confirmation
of Relative and Absolute Configurations**
Youseung Shin, Jean-Hugues Fournier,
Yoshikazu Fukui, Arndt M. Brückner, and
Dennis P. Curran*
The macrolactone (À)-dictyostatin was isolated in small
quantities from a Maldives marine sponge Spongia sp. by
Pettit and co-workers in 1994.[1a] Dictyostatin strongly inhibits
the growth of human cancer cells with a GI50 in the 50 pm to
1 nm range, and in a subsequent patent it was assigned a
partial structure 1a shown.[1b] Wright and co-workers isolated
dictyostatin from Corallistidea sponges,[2] and showed that it
stabilized microtubules in the same way as discodermolide
(by preventing the depolymerization to tubulin heterodim-
ers). Studies on taxol-resistant cell lines and other experi-
ments also suggested that dictyostatin could be useful against
multidrug-resistant tumors. Despite these exciting results,
further study of dictyostatin has been retarded because it is
scarce and its structure is not known with certainty.
Dictyostatin is structurally related to discodermolide 2,[3]
and in 2002 we reported the syntheses of the first discoder-
molide–dictyostatin hybrid molecules and suggested that the
absolute configuration of dictyostatin should be inverted so
that it more closely resembles discodermolide.[4] The synthesis
of hybrids such as 3 taught us how to make the lower dienyl
ester chain of the dictyostatins, including the macrolactone
ring.
In further work, we began to doubt the relative stereo-
chemical assignments in 1a.[5] For example, we recently made
an isomer of dictyostatin with the structure 1c (“dictyosta-
tin 5”) that has discodermolide-like configurations in the
“upper” and “middle” fragments (solid boxes) and the
dictyostatin configurations proposed by Pettit and co-workers
in the “lower” fragment (dashed box). The resonances of the
protons from the “upper” and “middle” fragments resembled
those for dictyostatin very closely, but those of the “lower”
fragment (especially 3-H–5-H) were very different.[5] Like-
wise, we made 1d (“dictyostatin 2”),[5] which is one of the four
candidate structures originally proposed; this was not dic-
tyostatin either, and the resonances for its “lower” fragment
protons resembled those of 1c, not dictyostatin.
[*] Y. Shin, Dr. J.-H. Fournier, Dr. Y. Fukui, Dr. A. M. Brückner,
Prof. Dr. D. P. Curran
Very recently, Paterson and Wright conducted a detailed
1H NMR spectroscopic analysis to assign configurations to
dictyostatin fragments, and then used molecular modeling to
assemble the fragments.[6] They arrived at structure 1b in
which the 10 stereocenters that dictyostatin shares with
discodermolide all have the same relative configurations
(C14[7] of dictyostatin is not a stereocenter in discodermolide,
and was assigned the S configuration). We report herein the
synthesis of dictyostatin, and confirm that structure 1b is
correct. Concurrently, Paterson and co-workers report their
synthesis and have reached the same conclusion.[8]
Scheme 1 summarizes key elements of the strategic plan
to make 1b. Because of the uncertainties with configurations,
stereochemical flexibility is a key design element.[5] The
molecule breaks down into three key fragments as indicated
by the wavy lines in 1b. With the preparation of analogues in
mind, we intentionally designed a less convergent synthesis
Department of Chemistry
University of Pittsburgh
Pittsburgh, PA 15260 (USA)
Fax: (+1)412-624-9861
E-mail: curran@pitt.edu
[**] We thank the National Institutes of Health for funding this work, Y.S.
thanks the University of Pittsburgh for a Graduate Excellence
Fellowship, J.H.F. thanks FCAR, and A.M.B. thanks the German
Academic Exchange Service (DAAD) for postdoctoral fellowships.
We also thank Professor Billy Day for helpful discussions, Dr. Cherry
Herald for spectra of dictyostatin, Dr. Amy Wright for permission to
compare samples, and Prof. Ian Paterson for comparing the
samples and kindly agreeing to joint publication.
Supporting information for this article (comparison of 1H and
13C NMR spectroscopic data for natural and synthetic samples of
dictyostatin) is available on the WWW under http://www.ange-
wandte.org or fromthe author.
4634
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200460593
Angew. Chem. Int. Ed. 2004, 43, 4634 –4637

Angewandte
Chemie
Scheme 1. Summary of the synthetic strategy to make 1b. TBS=tert-
butyldimethylsilyl, PMP=p-methoxyphenyl, Tr=triphenylmethyl.
that introduces both dienes after coupling of the fragments.
Compound 4 became a key intermediate in this strategy and
was built from fragments 5, 6, and 7. In turn, the fragments
were made by building on prior work in the discodermolide
area,[3,9] including our own routes to simplified discodermo-
lide analogues.[10] A key element of the synthesis was our
decision to change from a Wittig approach[9] to the synthesis
Scheme 2. A) a) (MeO)2P(O)CH3, BuLi, À788C, 85%. B) a) LiH2NBH3,
THF, 96%; b) TBSCl, imidazole, 4-DMAP, 95%; c) DDQ, 99%;
d) SO3·py, Et3N, DMSO; e) CBr4, PPh3, 2,6-lutidine, 73% over two
steps; f) BuLi, THF, 95%. C) a) DIBAL-H, CH2Cl2, 97%; b) TrCl,
4-DMAP, pyridine; c) HF/py, pyridine, THF, 89% over two steps;
d) 1) SO3·py, Et3N, DMSO/CH2Cl2; 2) NaClO2, NaHPO4, 2-methyl-2-
butene, THF/H2O; e) NH(Me)(OMe)HCl, DCC, Et3N, DMAP, CH2Cl2,
73% (three steps). DIPEA=diisopropylethylamine, DMAP=dimethyl-
aminopyridine, DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone,
DMSO=dimethyl sulfoxide, DIBAL-H=diisobutylaluminum hydride,
DCC=dicyclohexylcarbodiimide.
À
of the C8 C9 Z alkene to a more reliable alkyne addition to
[11]
À
make the C7 C8 bond.
The syntheses of fragments 5, 6, and 7 are summarized in
Scheme 2. Fragment 5 was prepared from known Weinreb
ester 8[9a,b] by addition of LiCH2P(= O)(OMe)2. The C14
stereocenter of fragment 6 was introduced by conversion of
alcohol 9 into a sensitive iodide,[12] followed by direct Myers
alkylation.[13] This gave 10 in 87% yield as a single stereo-
isomer. Cleavage of the auxiliary group and protection of the
alcohol function were followed by removal of the PMB group,
oxidation, Corey–Fuchs olefination, and elimination to install
the alkyne on the other end of the backbone. Fragment 7 was
prepared by a straightforward sequence of functional-group
transformations from the ester 11.[14]
Still–Gennari olefination,[18] and removal of the PMB group
provided the E/Z dienyl ester 15.
The final key steps follow directly from our previous
syntheses.[4,5] Saponification, Yamaguchi lactonization,[19] and
global desilylation provided about 3 mg of 1b as a white solid
after purification by HPLC. Extensive high-field 13C and
1H NMR data are available for both natural samples, and our
data match these well.[20] Pettit and co-workers report that
natural dictyostatin has an optical rotation of À20 (c = 0.12,
MeOH),[1] whereas our sample has a rotation of À23 (c = 0.18,
MeOH). From these data, we conclude that dictyostatin has
structure 1b. The longest linear sequence of 34steps runs
through fragment 6 and provides 1b in 1% overall yield from
(2S)-3-hydroxy-2-methylpropionic acid methyl ester (Roche
ester).[9b]
The confirmation of the structure of dictyostatin clears the
way for further development in this exciting area. Despite
being a macrolactone, dictyostatin presumably has a similar
shape to discodermolide, and indeed we had even designed
macrocyclic discodermolide analogues before we were aware
of dictyostatin.[4] Armed with the knowledge of the structure
of dictyostatin, we now know that discodermolide/dictyosta-
tin hybrids such as 3 are much closer to dictyostatin than was
originally expected. Our present synthesis is flexible and is
The union of the fragments and the completion of the
synthesis are shown in Scheme 3. Metalation of alkyne 6 and
addition of Weinreb ester 7 provided an alkynyl ketone (not
shown). The asymmetric reduction of this ketone according to
Noyori and co-workers[11,15] followed by Lindlar hydrogena-
tion gave 12 as a single isomer. Protection of the secondary
alcohol followed by deprotection of the primary TBS ether,
oxidation, and Wittig–Horner coupling with 5 provided key
enone 4. The conjugated double bond of 4 was reduced with
nickel boride, and the ketone was then reduced with sodium
borohydride to give a readily separable mixture of epimers 13
(b/a 2.4:1).[16] The major b epimer was silylated, the PMP
acetal was opened, and the terminal diene was introduced by
a standard procedure used in the synthesis of discodermo-
lide[17] to give 14. Detritylation of 14 followed by oxidation,
Angew. Chem. Int. Ed. 2004, 43, 4634 –4637
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4635

Communications
Scheme 3. a) nBuLi, THF, 93%; b) S,S Noyori catalyst (20 mol%), iPrOH, 79%; c) Lindlar catalyst, H2 (balloon), toluene, 91%; d) TBSOTf, 2,6-luti-
dine, CH2Cl2, 99%; e) HF/py, pyridine, THF, 08C, 1 day, 67%; f) 1) Dess–Martin oxidation; 2) Ba(OH)2, 5, THF/H2O, 80% (two steps); g) NiCl2,
NaBH4, MeOH/THF, 76%; h) NaBH4, MeOH/THF, 70% (b), 29% (a); i) TBSOTf, 2,6-lutidine, CH2Cl2, 99%; j) DIBAL-H, CH2Cl2, 88%;
k) 1) Dess–Martin oxidation; 2) CH2 =CHCH(TMS)Br, CrCl2, THF; 3) NaH, THF, 89% (three steps); l) ZnBr2, CH2Cl2/MeOH, 69%; m) 1) Dess–
Martin oxidation; 2) (CF3CH2O)2P(O)CH2CO2Me, KHMDS, [18]crown-6, THF, 86% (two steps); n) DDQ, CH2Cl2/H2O, 88%; o) KOH (1n), EtOH/
THF; p) 2,4,6-trichlorobenzoyl chloride, Et3N, THF then 4-DMAP(10 equiv), toluene, 78% (two steps); q) HCl/MeOH (3n), THF, 55%. Tf=tri-
fluoromethanesulfonyl, TMS=trimethylsilyl, HMDS=hexamethyldisilazide.
[5] a) Y. Shin, unpublished results; syntheses of isomers of dictyo-
suitable for making other kinds of analogues as well.
statin will be described in a subsequent full paper; b) D. P.
Curran, Y. Shin, N. Choy, B. W. Day, R. Balachandran, C.
Increased quantities of dictyostatin are also needed for
testing. Although it has two more backbone carbon atoms,
dictyostatin lacks three of the stereocenters of discodermo-
lide, therefore it will probably be easier to make in the long
run. We are currently investigating ways to increase the
convergence of the current synthesis as a prelude to making
more dictyostatin.
Madiraju, T. Turner, WO 022552, 2004 [Chem. Abstr. 2004,
220327].
[6] I. Paterson, R. Britton, O. Delgado, A. E. Wright, Chem.
Commun. 2004, 632 – 633.
[7] The numbering system of discodermolide is imposed on
dictyostatin in this paper to facilitate comparisons.
[8] See preceding Communication in this issue: I. Paterson, R.
Britton, O. Delgado, A. Meyer, K. G. Poullennec, Angew. Chem.
2004, 116, 4729; Angew. Chem. Int. Ed., 2004, 43, 4629.
[9] a) A. B. Smith III, T. J. Beauchamp, M. J. LaMarche, M. D.
Kaufman, Y. P. Qiu, H. Arimoto, D. R. Jones, K. Kobayashi, J.
Am. Chem. Soc. 2000, 122, 8654– 8664; b) Weinreb amide 8 was
made in five steps from (2S)-3-hydroxy-2-methylpropionic acid
methyl ester (Roche ester) in 39% yield. Roche ester was
purchased from Aldrich or Mitsubishi-Rayon.
[10] a) N. Choy, Y. Shin, P. Q. Nguyen, D. P. Curran, R. Balachan-
dran, C. Madiraju, B. W. Day, J. Med. Chem. 2003, 46, 2846 –
2860; b) J. M. Minguez, S.-Y. Kim, K. A. Giuliano, R. Balachan-
dran, C. Madiraju, B. W. Day, D. P. Curran, Bioorg. Med. Chem.
2003, 11, 3335 – 3354.
[11] J. A. Marshall, M. J. Bourbeau, Org. Lett. 2003, 5, 3197 – 3199.
[12] a) R. D. Walkup, R. R. Kane, P. D. Boatman Jr, R. T. Cunning-
ham, Tetrahedron Lett. 1990, 31, 7587 – 7609; b) R. D. Walkup,
J. D. Kahl, R. R. Kane, J. Org. Chem. 1998, 63, 9113 – 9116;
c) alcohol 9 was made in five steps from Roche ester in 47%
yield.
Received: May 7, 2004
Published Online: July 1, 2004
Keywords: antitumor agents · conformation analysis ·
.
macrocycles · natural products · total synthesis
[1] a) G. R. Pettit, Z. A. Cichacz, F. Gao, M. R. Boyd, J. M. Schmidt,
J. Chem. Soc. Chem. Commun. 1994, 1111 – 1112; b) G. R. Pettit,
Z. A. Cichacz, Isolation and structure of dictyostatin,
WO 5430053, 1995 [Chem. Abstr. 1995,733500].
[2] a) A. E. Wright, J. L. Cummins, S. A. Pomponi, R. E. Longley,
R. A. Isbrucker, Dictyostatin compounds for stabilization of
microtubules, WO 0162239, 2001 [Chem. Abstr. 2001, 635882];
b) R. A. Isbrucker, J. Cummins, S. A. Pomponi, R. E. Longley,
A. E. Wright, Biochem. Pharmacol. 2003, 66, 75 – 82.
[13] A. Myers, B. Y. Yang, H. Chen, L. McKinstry, D. J. Kopecky, J.
Gleason, J. Am. Chem. Soc. 1997, 119, 6496 – 6511.
[14] The enantiomer of the methyl ester of 6 is known: a) P. Phukan,
S. Sasmal, M. E. Maier, Eur. J. Org. Chem. 2003, 1733 – 1740;
b) M. B. Andrus, A. B. Argade, Tetrahedron Lett. 1996, 37,
5049 – 5052; c) ester 11 was made in six steps from 1,3-propane-
diol in 35% yield.
[3] a) M. Kalesse, ChemBioChem 2000, 1, 171 – 175; b) D. C. Myles,
Annu. Rep. Med. Chem. 2002, 37, 125 – 132; c) I. Paterson, G. J.
Florence, Eur. J. Org. Chem. 2003, 2193 – 2208; d) S. J. Mickel,
G. H. Sedelmeier, D. Niederer, R. Daeffler, A. Osmani, K.
Schreiner, M. Seeger-Weibel, B. Berod, K. Schaer, R. Gamboni,
Org. Process Res. Dev. 2004, 8, 92 – 100, and succeeding papers in
that issue.
[4] Y. Shin, N. Choy, R. Balachandran, C. Madiraju, B. W. Day, D. P.
Curran, Org. Lett. 2002, 4, 4443 – 4446.
[15] K. Matsumura, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am.
Chem. Soc. 1997, 119, 8738 – 8739.
4636
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4634 –4637

Angewandte
Chemie
[16] The nonselective reduction was expressly used to make both
isomers because the configuration at this center was at issue.
[17] I. Paterson, G. J. Florence, K. Gerlach, J. P. Scott, N. Sereinig, J.
Am. Chem. Soc. 2001, 123, 9535 – 9544.
[18] W. C. Still, C. Gennari, Tetrahedron Lett. 1983, 24, 4405 – 4408.
[19] J. Inanaga, H. Kuniko, S. Hiroko, T. Katsuki, M. Yamaguchi,
Bull. Chem. Soc. Jpn. 1979, 52, 1989 – 1994.
[20] Dr. Robert Britton, Cambridge University, kindly recorded
500 MHz NMR spectra of our sample and a natural sample
provided by Dr. Amy Wright, Harbor Branch Oceanographic
Institute. These spectra are in the Supporting Information.
Angew. Chem. Int. Ed. 2004, 43, 4634 –4637
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4637