Synthetic studies on hemicalide: Development of a convergent approach toward the C1-C25 fragment

Article abstract of DOI:10.1021/ol402077e

Synthetic studies on hemicalide, a recently isolated marine natural product displaying highly potent antiproliferative activity and a unique mode of action, have highlighted a reliable Horner-Wadsworth-Emmons olefination to create the C6-C7 alkene and a remarkable efficient Suzuki-Miyaura coupling to form the C15-C16 bond, resulting in the development of a convergent approach toward the C1-C25 fragment.

Full text of DOI:10.1021/ol402077e

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Synthetic Studies on Hemicalide:
Development of a Convergent Approach
toward the C1ÀC25 Fragment
Geoffroy Sorin,† Etienne Fleury,† Christine Tran,† Elise Prost,† Nicolas Molinier,‡
Franc-ois Sautel,‡ Georges Massiot,‡ Simon Specklin,§ Christophe Meyer,§
Janine Cossy,§ Marie-Isabelle Lannou,† and Janick Ardisson*,†
ꢀ
ꢀ
Faculte de Pharmacie, Universite Paris Descartes, CNRS UMR 8638, 4 avenue de
l’observatoire, 75270 Paris Cedex 06, France, CNRSÀPierre Fabre USR 3388,
ꢀ
Centre de Recherche & Developpement Pierre Fabre, 3 avenue Hubert Curien, 31035
Toulouse Cedex 01, France, and Laboratoire de Chimie Organique, ESPCI ParisTech,
CNRS UMR 7084, 10 rue Vauquelin, 75231 Paris Cedex 05, France
janick.ardisson@parisdescartes.fr
Received July 23, 2013
ABSTRACT
Synthetic studies on hemicalide, a recently isolated marine natural product displaying highly potent antiproliferative activity and a unique mode of
action, have highlighted a reliable HornerÀWadsworthÀEmmons olefination to create the C6ÀC7 alkene and a remarkable efficient SuzukiÀ
Miyaura coupling to form the C15ÀC16 bond, resulting in the development of a convergent approach toward the C1ÀC25 fragment.
The identification of structurally new and highly potent
bioactive lead compounds from marine sources has be-
come an active area of research.1 Although it could offer
considerable promise, investigations into the deep sea
environment remain rather scarce. Recently, researchers
of the CNRS-Pierre Fabre Laboratories joint unit
in association with the Institut de Recherche pour le
conducted immuno-cytochemistry studies indicated that
this compound acted by a unique mechanism that involved
the destabilization of the R/β microtubule network. The
planar structure of hemicalide 1, which was assigned by
extensive NMR studies, indicated a 46 carbon atom back-
bone comprised of 21 stereocenters and several remarkable
structural elements: a conjugated trienic acid (C1ÀC7), a
six-carbon polypropionate motif (C8ÀC13), a conjugated
diene (C14ÀC17), and R,β-dihydroxy (C19ÀC23) as well
as R-hydroxy (C37ÀC41) δ-lactones. Due to the extremely
limited supply of hemicalide (less than 1 mg was isolated),
neither derivatization nor degradation experiments could
be carried out and the configuration of the stereocenters
was not assigned. Intriguedby the architectural complexity
and promising bioactivity of hemicalide, we embarked on
determining the relative configuration and the total synth-
esis of this natural product. In previous work, a careful
comparison of the NMR data of appropriate diastereo-
meric model compounds with those of hemicalide allowed
us to assign the relative configuration of the C8À
C13 subunit3 as well as the R,β-dihydroxy δ-lactone
ꢀ
Developpement (IRD) collected the marine sponge Hemi-
mycale sp. in deep water around the Torres Islands
(Vanuatu). Bioassay-guided fractionation led to the iso-
lation of the new complex polyketide hemicalide 1 as
a potent mitotic blocker which displays high antiprolifera-
tive potency against a panel of human cancer cell lines at
subnanomolar concentrations.2 Additionally, the initially
†
ꢀ
Universite Paris Descartes.
‡ CNRSÀPierre Fabre USR 3388.
§ ESPCI ParisTech.
(1) For a recent review on marine natural products, see: Blunt, J. W.;
Copp, B. R.; Keyzers, R. A.; Munro, M. H. G.; Prinsep, M. R. Nat.
Prod. Rep. 2013, 30, 237–323.
(2) Carletti, I.; Massiot, G.; Debitus, C. WO 2011/051380A1 patent,
2011.
r
10.1021/ol402077e
XXXX American Chemical Society

(C19ÀC23) bearing adjacent methyl-substituted stereo-
centers (C18 and C24) (Figure 1).4,5 Herein, we report
the results of our ongoing synthetic studies on hemicalide
which have enabled us to discover appropriate coupling
reactions to assemble three key subunits of the natural
product of value for a convergent approach toward the
C1ÀC25 fragment.
(5), the cis-1,2-diol at C21,C22 was installed by dihydroxy-
lation of an R,β-unsaturated δ-lactone 3 which in turn
arose from an intramolecular HÀWÀE olefination of the
keto-phosphonate 2.4 Despite considerable optimization,
β-elimination leading to enone 4 unavoidably took place
as a side reaction (17%). Another matter of concern was
that aldehyde 6, generated from compound 5, could not
be successfully engaged in a dibromomethylenation or
a Wittig-type reaction and underwent decomposition in-
stead (Scheme 2).
Scheme 2. Previous Synthesis of δ-Lactone Model 5
Figure 1. Structure of hemicalide 1 and relative configurations
of the C8ÀC13 and C18ÀC24 subunits.
With the goal of devising a convergent strategy toward
the C1ÀC25 subunit of hemicalide, two key disconnection
points were identified. As illustrated for one putative diaste-
reomer A, formation of the C15ÀC16 bond was envisaged by
a SuzukiÀMiyaura coupling between a trisubstituted alke-
nylboronate at C16 and a trisubstituted alkenyl iodide at C15
whereas formation of the C6ÀC7 olefin would be accom-
plished by a HornerÀWadsworthÀEmmons (HÀWÀE)
olefination. Therefore, the synthesis of the C1ÀC25 fragment
of hemicalide was planned from three key subunits: phos-
phonate D incoporating a (E,E)-dienoate (C1ÀC6), alkenyl
iodide C (C7ÀC15) containing six stereocenters, and alkenyl
boronate B comprising the R,β-dihydroxy δ-lactone moiety
(C16ÀC25) (Scheme 1).
Due to the difficulties encountered in the carbon chain
extension of aldehyde 6 at C17, a new strategy was devised
wherein an appropriate precursor of the C16ÀC17 tri-
substituted (E)-alkene was present before construction
of the lactone core. Thus, aldehyde (R)-8 possessing a
trisusbtituted alkenyl bromide, readily available from an
(S)-Roche ester,6 was engaged in a Dias allylation with the
chiral allylsilane (S)-9 (prepared from (R)-Roche ester).4
The reaction proceeded with high diastereoselectivity
(dr >95:5), and the resulting homoallylic alcohol 10 was
isolated in 62% yield as a 85:15 mixture of correspond-
ing (E) and (Z) diastereomers. After acylation of the
hydroxy group with bromoacetyl bromide, the presence
of the trisubstituted vinylic bromide allowed for the che-
moselective ozonolysis of the exo-methylene group7 to
provide ketone 11 albeit in moderate yield (50%). Inter-
estingly, as an alternative approach, an aldol reaction
involving methyl ketone 12 as a partner (prepared from
(R-Roche ester) was examined.8 Addition of the titanium
enolate derived from 12 to aldehyde (R)-8 proceeded with
satisfying diastereoselectivity (dr = 90:10) and led to
compound 11 (73%) after acylation of the hydroxy group
at C19 with bromoacetyl bromide, thereby avoiding the
Scheme 1. Retrosynthetic Analysis of the C1ÀC25 Subunit A
Synthesis of the C16ÀC25 subunit B was investigated
first. In our previously described synthetic approach to a
model compound for the C17ÀC25 subunit of hemicalide
(6) In our hands, the alcohol precursor of aldehyde (R)-8 was
obtained as an inseparable mixture of the two corresponding (E) and
(Z) olefins (dr = 85:15). (a) Cho, C.-G.; Kim, W.-S.; Smith, A. B., III.
Org. Lett. 2005, 7, 3569–3572. (b) Andrus, M. B.; Li, W.; Keyes, R. F.
J. Org. Chem. 1997, 62, 5542–5549.
(7) (a) Prior, W. A.; Giamalva, D.; Church, D. F. J. Am. Chem. Soc.
1983, 105, 6858–6861. (b) Paterson, I.; Paquet, T.; Dalby, S. M. Org.
Lett. 2011, 13, 4398–4401.
(3) Fleury, E.; Lannou, M.-I.; Bistri, O.; Sautel, F.; Massiot, G.;
Pancrazi, A.; Ardisson, J. J. Org. Chem. 2009, 74, 7034–7045.
(4) Fleury, E.; Sorin, G.; Prost, E.; Pancrazi, A.; Sautel, F.; Massiot,
G.; Lannou, M.-I.; Ardisson, J. J. Org. Chem. 2013, 78, 855–864.
(5) It is worth noting that the relative configuration of the C8ÀC13
subunit of hemicalide was later confirmed by Goodman and Smith using
Gauge-Inducing-Atomic-Orbital (GIAO) NMR calculations: Smith,
S. G.; Goodman, J. M. J. Am. Chem. Soc. 2010, 132, 12946–12959.
´
(8) Zambrana, J.; Romea, P.; Urpı, F.; Lujan, C. J. Org. Chem. 2011,
76, 8575–8587.
B
Org. Lett., Vol. XX, No. XX, XXXX

previous ozonolysis step. At this stage, a SmI2-induced
intramolecular Reformatsky reaction, proceeding under
neutral conditions,9 followed by dehydration, provided a
remarkably efficient entry to the R,β-unsaturated δ-lactone
13 (79%). Subsequent chemo- and diastereoselective syn-
dihydroxylation afforded δ-lactone 14 in quantitative yield
and excellent diastereoselectivity (dr >95:5) (Scheme 3).10
organolithium generated from the (E)-alkenyl iodide 1713
to Weinreb amide 16 afforded β-silylenone 18 in almost
quantitative yield (99%). Cleavage of the TMS ether at
C11 led to a β-hydroxyketone which underwent diastereo-
selective reduction with Zn(BH4)2 (dr = 90:10) to afford
the syn-1,3-diol 19 (87%) (Scheme 4).14
Scheme 4. Synthesis of the C7ÀC16 Subunit 20
Scheme 3. Synthesis of the C16ÀC25 Subunit 15
The choice of the protecting groups for 1,3-diol 19 was
crucial. When 19 was protected as di-TES ether, subse-
quent iododesilylation under standard conditions did not
readily proceed. Use of a cyclic ketal proved to be essential
toperformthisreaction. Astheacidicremoval ofacetonide
at the end of the synthesis could be problematic,15 it
became apparent that the most successful route would
involve protection of 19 as a di-(tert-butyl)-silylene ketal,
with final deprotection with fluoride sources under mild
conditions.16,17 Thereby, after protection of 19 as a cyclic
silylene ketal, iododesilylation, and chemoselective cleav-
age of the TBS ether at C7, the (E)-alkenyl iodide 20
corresponding to the C7ÀC16 subunit of hemicalide was
obtained (6 steps from Weinreb amide 16, 49% overall
yield) (Scheme 4).
As the preparation of phosphonate 21, corresponding
to the C1ÀC6 subunit D, has been previously achieved
from 2-trimethylsilylethyl sorbate using a chemoselective
cross-metathesis with allyl bromide followed by an Arbuzov
reaction,3 the assembly of the three subunits B, C, andDwas
investigated.
The primary alcohol 20 was oxidized with Dess-Martin
periodinane (DMP), and aldehyde 21 was engaged in an
HÀWÀE olefination with the lithium salt of phosphonate
22 (Scheme 5). The resulting (E,E,E)-triene 23 was obtained
Protection of the 1,2-diol as a bis-triethylsilyl ether
(77%)11 followed by a palladium-catalyzed borylation of
the trisubstituted alkenyl bromide 14 with bis(pinacolato)-
diboron12 eventually delivered, in 82% yield, the trisub-
stituted (E)-alkenylboronate 15 armed for further func-
tionalization at both terminii.
The preparation of the C16ÀC25 fragment has been
achieved in 7 steps from methyl ketone 12, implying a total
of 10 steps from the (R)-Roche ester (24% overall yield).
The synthesis of the C7ÀC15 subunit C was then under-
taken from Weinreb amide 16. We have previously described
the preparation of this latter compound from the (S)-
Roche ester by iterative aldol reactions, during the synth-
esis of model compounds for the assignment of the relative
configuration of the C8ÀC13 subunit.3 Addition of the
(13) Nicolaou, K. C.; Piscopio, A. D.; Bertinato, P.; Chakraborty,
T. K.; Minowa, N.; Koide, K. Chem.;Eur. J. 1995, 1, 318–333.
(14) The relative configuration of the C13 stereocenter was con-
firmed by examination of the 13C NMR spectrum of 19. A characteristic
low 13C NMR chemical shift (δ = 4.1 ppm) was noticed for the sole C12
methyl group: this value is consistent with a synÀsyn stereotriad (see
Supporting Information). (a) Hoffmann, R. W.; Weidmann, U. Chem.
Ber. 1985, 118, 3980–3992. (b) See ref 3.
(9) (a) Molander, G. A.; Etter, J. B. J. Am. Chem. Soc. 1987, 109,
6556–6558. (b) Molander, G. A.; Brown, G. A.; Storch de Gracia, I.
J. Org. Chem. 2002, 67, 3459–3463.
(10) As for lactone 5,4 the relative configuration of lactone 14 was
confirmed by NOESY (see Supporting Information).
(15) Bock, M.; Dehn, R.; Kirschning, A. Angew. Chem., Int. Ed.
2008, 47, 9134–9137.
(11) At this stage, the benzyl ether at C25 can be cleaved in high yield
in the presence of DDQ (CH2Cl2, reflux) which may be useful to achieve
coupling with the C26ÀC46 subunit.
(16) Colobert, F.; Choppin, S.; Ferreiro-Mederos, L.; Obringer, M.;
Luengo Arratta, S.; Urbano, A.; Carreno, M. C. Org. Lett. 2007, 9,
4451–4454.
(17) Baker, T. M.; Edmonds, D. J.; Hamilton, D.; O’Brien, C. J.;
Procter, D. J. Angew. Chem., Int. Ed. 2008, 47, 5631–5633.
(12) Matsumura, D.;Takarabe, T.;Toda, T.;Hayamizu, T.; Sawamura,
K.; Takao, K.-i.; Tadano, K.-i. Tetrahedron 2011, 67, 6730–6745.
Org. Lett., Vol. XX, No. XX, XXXX
C

in excellent yield (92%), and the alkenyl iodide at C15
underwent a remarkably smooth SuzukiÀMiyaura cou-
pling with the previously synthesized alkenyl pinacol bor-
onate 15 (C16ÀC25 subunit), catalyzed by Pd(PPh3)4 and in
the presence of aqueous TlOEt as the base.18 Under these
conditions, compound 24 possessing the desired trisubsti-
tuted (E,E)-C14ÀC17 diene of hemicalide was isolated in
very high yield (97%) (Scheme 5).
utilized to deprotect the 1,3-diol at C11,C13 (Scheme 5).
The synthesis of compound 25 corresponding to one
putative stereoisomer of the C1ÀC25 fragment of hemi-
calide was thus successfully completed.
It was interesting to compare the NMR spectroscopic
data of hemicalide with those of compound 25 which
should constitute one of the two possible diastereomers
of the C1ÀC25 subunit based on our previous stereoche-
mical assignment studies.3,4,19 Since the natural product
was isolated as a carboxylate salt, significant chemical shift
differences were observed in the trienic region (for H3 and
carbons C1ÀC7), as previously noted.3 For the C8ÀC24
Scheme 5. Synthesis of the C1ÀC25 Subunit of Hemicalide 25
1
segment, the H and 13C NMR chemical shifts of com-
pound 25 were in good agreement with those of hemicalide
(|Δδ| e 0.12 ppm and |Δδ| e 1.2 ppm, respectively). In
particular, differences were insignificant in the C8ÀC17
region, but the presence of a benzyl ether at C25 in
compound 25 does not obviously allow for a more relevant
comparison with the natural product at this stage.
In conclusion, we have developed a highly convergent
approach to one possible diastereomer of the C1ÀC25
subunit of hemicalide. The strategy relies on formation of
the C6ÀC7 olefin by an HÀWÀE olefination and creation
of the C15ÀC16 bond by a SuzukiÀMiyaura coupling.
Our results demonstrate the remarkable efficiency of the
SuzukiÀMiyaura coupling for construction of the C14ÀC17
conjugated diene of hemicalide with highly functionalized
partners and strongly encourage its use as a possible endgame
in the synthesis of hemicalide (or stereoisomers), a program
currently underway in our laboratories.
Acknowledgment. We thank Pierre Fabre Laboratories
and the CNRS for a doctoral fellowship for E.F. and the
CNRS for postdoctoral fellowships for G.S. and S.S.
Based on our selection of silyl protecting groups, the
final deprotection could be accomplished in two steps
using tris(dimethylamino)sulfonium difluoro trimethyl sil-
icate (TASF) (DMF, rt) to cleave the TES ethers and the
Supporting Information Available. Experimental pro-
cedures and full analyses of 13C and 1H NMR spectra for
all compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
2-trimethylsilyl ester first, and then buffered HF Py was
3
(18) (a) Miyaura, N.; Yamada, K.; Suginome, H.; Suzuki, A. J. Am.
Chem. Soc. 1985, 107, 972–980. (b) Ghidu, V. P.; Wang, J.; Wu, B.; Liu,
Q.; Jacobs, A.; Marnett, L. J.; Sulikowski, G. A. J. Org. Chem. 2008, 73,
4949–4955. (c) Wu, B.; Liu, Q.; Sulikowski, G. A. Angew. Chem., Int. Ed.
2004, 43, 6673–6675. (d) Francais, A.; Leyva, A.; Etxebarria-Jardi, G.;
Ley, S. V. Org. Lett. 2010, 12, 340–343.
(19) Tables of comparison are provided in the Supporting
Information.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX