Synthesis of the macrocyclic core of leiodermatolide

Article abstract of DOI:10.1021/ol2017388

The macrocyclic core (2) of the marine macrolide leiodermatolide (1) has been synthesized in 19 steps through a convergent strategy exploiting boron aldol methodology to install the requisite stereochemistry and a selective Stille coupling reaction for controlled fragment assembly, followed by a Yamaguchi macrolactonization and carbamate introduction at the C9-OH.

Full text of DOI:10.1021/ol2017388

ORGANIC
LETTERS
2011
Vol. 13, No. 16
4398–4401
Synthesis of the Macrocyclic Core of
Leiodermatolide
Ian Paterson,* Tanya Paquet, and Stephen M. Dalby
University Chemical Laboratory, Lensfield Road, Cambridge CB2 1EW, U.K.
ip100@cam.ac.uk
Received June 28, 2011
ABSTRACT
The macrocyclic core (2) of the marine macrolide leiodermatolide (1) has been synthesized in 19 steps through a convergent strategy exploiting
boron aldol methodology to install the requisite stereochemistry and a selective Stille coupling reaction for controlled fragment assembly,
followed by a Yamaguchi macrolactonization and carbamate introduction at the C9-OH.
Marine natural products continue to represent an im-
portant source of lead compounds for the development of
novel cancer chemotherapeutics,1 highlighted by the recent
FDA approval of Halaven (eribulin mesylate) for the
treatment of advanced breast cancer.1a Leiodermatolide
(1, Scheme 1) has recently emerged as a promising new
anticancer polyketide, isolated by the Wright group in
2008 from the lithistid sponge Leiodermatium sp., collected
off the coast of Florida.2 Crude sponge extracts displayed
activity in an antimitotic assay, and subsequent bioassay-
guided fractionation identified leiodermatolide as the ac-
tive component, which exhibited low nanomolar (<10 nM)
cytotoxicity against a range of human cancer cell lines.
This antimitotic activity appears to be mediated through
the disruption of tubulin dynamics, leading to cell-cycle
arrest in the G2/M phase. While the exact mechanism of
action ispresently unknown, itisclearly distinct from other
important antimitotic drugs such as the epothilones, tax-
anes, and Vinca alkaloids,3 marking outleiodermatolide as
a potential lead compound for the development of new
anticancer agents.
From a structural perspective, leiodermatolide features
a 16-membered macrolactone appended with a carbamate
and an unsaturated side chain terminating in a δ-lactone
ring, which together incorporate five alkenes, including a
(Z,Z)-diene and an (E,E)-diene, and nine stereocenters.2
Recently, we have used a combination of homo- and
heteronuclear NMR analysis, molecular modeling, and
DP4 NMR prediction methodology to determine the
stereostructure 1 for leiodermatolide,2a where the full
configuration of the C1ÀC16 macrocyclic and C20ÀC33
δ-lactone stereoclusters remains undefined. Synthetic stu-
dies by Maier and co-workers directed toward an earlier,
tentative stereostructure for leiodermatolide revealed in-
consistencies with this assignment.4 Herein, we report the
convergent synthesis of the C1ÀC16 macrocyclic core 2 of
leiodermatolide corresponding to stereostructure 1, whose
NMR homology with the natural product fully supports
our relative configurational assignment for this region of
the macrolide.
(1) (a) Ledford, H. Nature 2010, 468, 608. (b) Dalby, S. M.; Paterson,
I. Curr. Opin. Drug Discovery Dev. 2010, 13, 777. (c) Paterson, I.;
Anderson, E. A. Science 2005, 310, 451. (d) Cragg, G. M.; Grothaus,
P. G.; Newman, D. J. Chem. Rev. 2009, 109, 3012. (e) Nicolaou, K. C.;
Chen, J. S.; Dalby, S. M. Bioorg. Med. Chem. 2009, 17, 2290. (f) Yeung,
K.-S.; Paterson, I. Angew. Chem., Int. Ed. 2002, 41, 4632.
Our approach is outlined retrosynthetically in Scheme 1.
In order to resolve the remaining stereochemical ambigu-
ities, we aimed for late-stage incorporation of the δ-lactone
3, enabling faciledivergenceto bothpossiblediastereomers
(enantiomers) for comparison with the natural product.
(2) (a) Paterson, I.; Dalby, S. M.; Roberts, J. C.; Naylor, G. J.;
ꢀ
Guzman, E. A.; Isbrucker, R.; Pitts, T. P.; Linley, P.; Divlianska, D.;
Reed, J. K.; Wright, A. E. Angew. Chem., Int. Ed. 2011, 50, 3219. (b)
Wright, A. E.; Reed, J. K.; Roberts, J.; Longley, R. E. U.S. Pat. Appl.
Publ. (USA), US2008033035, 14 pp; Chem. Abstr. 2008, 148, 230103.
(3) Altmann, K. H.; Gertsch, J. Nat. Prod. Rep. 2007, 24, 327.
(4) (a) Rink, C.; Navickas, V.; Maier, M. E. Org. Lett. 2011, 13, 2334.
(b) Navickas, V.; Rink, C.; Maier, M. E. Synlett 2011, 191.
r
10.1021/ol2017388
Published on Web 07/14/2011
2011 American Chemical Society

Accordingly, leiodermatolide would arise through cross-
coupling of macrocyclic vinyl bromide 2 with the suitably
functionalized vinyl metallic species 3/ent-3, forming the
C16ÀC19 diene.
rearrangement/alkylation would then install the (E)-γ,
δ-unsaturated ester of 4.
Scheme 2. Preparation of C12ÀC17 Bis-vinyl Halide 5
Scheme 1. Retrosynthetic Analysis of Leiodermatolide
Scheme 3. Preparation of C1ÀC11 Fragment 18
Macrocycle 2 would itself arise from vinyl stannane 4
and bis-vinyl halide 5, exploiting a regioselective macro-
lactonization and a chemoselective Stille coupling reaction
to differentiate the termini of 5 and generate the (10Z,12Z)-
diene.5 Notably, these maneuvers would be carried out
without protection of the C7,C9 diol, demanding regiose-
lective carbamate formation at C9. The preparation of
both 4 and 5 would utilize our versatile boron aldol
methodology to install the requisite stereochemistry. For
bis-vinyl halide 5, aldol coupling of lactate-derived ketone
66 with aldehyde 7 would set the C14,C15-anti-relation-
ship, while the conspicuous C6ÀC9 stereotetrad within 4
would be constructed through a boron aldol reaction of
chiral ketone 87,8 with aldehyde 9 containing a masked (Z)-
vinyl stannane.5 Subsequent chain extension through allylic
(5) Paterson, I.; Kan, S. B. J.; Gibson, L. J. Org. Lett. 2010, 12, 3724.
(6) (a) Paterson, I.; Wallace, D. J.; Velazquez, S. M. Tetrahedron
Lett. 1994, 35, 9083. (b) Paterson, I.; Wallace, D. J.; Cowden, C. J.
Synthesis 1998, 639.
(7) (a) Paterson, I.; Lister, M. A. Tetrahedron Lett. 1988, 29, 585. (b)
Paterson, I.; Razzak, M.; Anderson, E. A. Org. Lett. 2008, 10, 3295.
(8) (a) Paterson, I.; Goodman, J. M.; Lister, M. A.; Schumann, R. C.;
McClure, C. K.; Norcross, R. D. Tetrahedron 1990, 46, 4663. (b)
Paterson, I.; Florence, G. J.; Gerlach, K.; Scott, J. P.; Sereinig, N.
J. Am. Chem. Soc. 2001, 123, 9535.
As shown in Scheme 2, preparation of pivotal bis-vinyl
halide 5 commenced with the aldol reaction6 of the (E)-
boron enolate of ethyl ketone 6 (cHex2BCl, Et3N, Et2O,
0 °C) with known aldehyde 7,9 which provided anti adduct
10 (94%, dr >20:1).10 Silylation of the newly formed C15
Org. Lett., Vol. 13, No. 16, 2011
4399

Scheme 4. Preparation of Macrolactone 2
alcohol (TESOTf, 2,6-lutidine) followed by DIBAL-
mediated reduction of the ketone and benzoate pro-
vided diol 11 (91%).
Completion of fragment 5 was then achieved through
11
oxidative cleavage of diol 11 using silica-supported NaIO4
toavoid competing desilylation, followed by StorkÀWittig
homologation (Ph3PCH2I2, NaHMDS, THF, À78 °C)
which proceeded to install the requisite (Z)-vinyl iodide 5
as a single detectable geometric isomer.12
Key to synthesizing the C1ÀC11 vinyl stannane coupling
partner 4 for 5 would be efficient control over installing the
C6ÀC9 syn,anti,syn-stereotetrad (Scheme 3). This was
achieved through a chiral ligand-mediated ((À)-Ipc2BOTf,
iPr2NEt) boron aldol reaction7,8 of ethyl ketone 8 with
aldehyde 9,5 which provided the desired syn adduct 12
(74%) as essentially a single diastereomer.10 The newly
formed C9 stereochemistry could then be relayed to C7
through 1,3-anti reduction under EvansÀTishchenko con-
ditions (SmI2, EtCHO, THF, À20 °C),13 which, following
in situ methanolysis of the ensuing propionate ester,
afforded the requisite stereotetrad as part of diol 13
(94%, dr >20:1). Protecting group manipulation invol-
ving bis-silylation (TBSOTf, 2,6-lutidine) and PMB ether
cleavage (DDQ) then provided alcohol 14.
for Claisen rearrangement to complete the C1ÀC5 region.
Accordingly, a sequence of oxidation and Grignard
additions provided 15 via alcohol 16. Unfortunately,
neither 15 nor the corresponding acetate derivative could
be coaxed into undergoing Claisen rearrangement under a
range of Ireland or Johnson-type conditions.15,16 Instead,
rather facile allylic rearrangement to the corresponding
primary allylic alcohol wasgenerally observed, forming the
Δ4 alkene with good levels of (E)-selectivity (>10:1). We
were able to exploit this behavior, however, through
conversion of allylic alcohol 15 to the corresponding
primary allylic bromide 17 according to the method of
Fuchter,17 involving treatment of the magnesium alkoxide
of 15 with TiBr4, which proceeded with excellent yield and
selectivity (98%, (E):(Z) >20:1). Finally, displacement of
the allylic bromide with the sodium anion of dimethylma-
lonate (87%) followed by Krapcho decarboxylation
(80%)18 smoothly afforded methyl ester 18, representing
the full C1ÀC11 sequence of leiodermatolide.
In preparation for coupling of the C1ÀC11 and C12À
C17 fragments, the vinyl stannane would have to be revealed
from vinyl dibromide 18. As shown in Scheme 4, selective
reductive debromination of the trans-bromide under
palladium-catalyzed conditions (Pd(PPh3)4, Bu3SnH)19
providedthecorresponding(Z)-vinylbromide (87%). This
At this point, we were able to contemplate incorporation
of the Δ4 (E)-trisubstituted alkene of leiodermatolide.
Following the failure of the corresponding terminal alkyne
to undergo Negishi carbometalation,14 a revised strategy
targeted tertiary allylic alcohol 15 as a potential substrate
was then subjected to desilylation with HF py to alleviate
3
(15) Rhoads, S. J.; Raulins, N. R. Org. React. 1975, 22, 1.
(16) These exploratory studies were carried out on the truncated
model substrate 15a:
(9) Hayes, C. J.; Sherlock, A. E.; Green, M. P.; Wilson, C.; Blake,
A. J.; Selby, M. D.; Prodger, J. C. J. Org. Chem. 2008, 73, 2041.
(10) (a) The configuration of the aldol adducts 10 and 12 was
determined by Mosher ester analysis. See the Supporting Information
for details. (b) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H.
J. Am. Chem. Soc. 1991, 113, 4092.
(11) Zhong, Y. L.; Shing, T. K. M. J. Org. Chem. 1997, 62, 2622.
(12) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 2173.
(13) Evans, D. A.; Hoveyda, A. H. J. Am. Chem. Soc. 1990, 112, 6447.
(14) (a) Negishi, E.; Van Horn, D. E.; Yoshida, T. J. Am. Chem. Soc.
1985, 107, 6639. (b) Rand, C. L.; Van Horn, D. E.; Moore, M. W.;
Negishi, E. J. Org. Chem. 1981, 46, 4093.
(17) Fuchter, M. J.; Levy, J.-N. Org. Lett. 2008, 10, 4919.
(18) (a) Krapcho, A. P.; Glynn, G. A.; Grenon, B. J. Tetrahedron
Lett. 1967, 8, 215. (b) Krapcho, A. P.; Jahngen, E. G. E.; Lovey, A. J.;
Short, F. W. Tetrahedron Lett. 1974, 15, 1091.
(19) (a) Uenishi, J.; Kawahama, R.; Yonemitsu, O.; Tsuji, J. J. Org.
Chem. 1996, 61, 5716. (b) Uenishi, J.; Kawahama, R.; Yonemitsu, O.;
Tsuji, J. J. Org. Chem. 1998, 63, 8965.
4400
Org. Lett., Vol. 13, No. 16, 2011

steric hindrance that initially had prevented further func-
tionalization of the vinyl bromide. After substantial in-
vestigation, vinyl bromide 19 could be converted to (Z)-
vinyl stannane 4 in moderate yield (33%) under modified
WulffÀStille conditions (Pd(PPh3)2Cl2, (Me3Sn)2, Li2CO3,
THF, 50 °C).20 This compound proved rather unstable
with respect to proto-destannylation and was thus sub-
mitted immediately for Stille coupling with bis-vinyl halide
5. Under conditions utilized recently in the context of our
chivosazole synthesis5 (Pd(PPh3)4, CuTC, Ph2PO2NBu4,
DMF, 0 °C),21 a smooth and completely regioselective
cross-coupling was achieved, generating the (10Z,12Z)-
diene of 20 in high yield (88%).
is appropriately functionalized for direct cross-coupling
with either antipode of a C17ÀC33 δ-lactone side-chain
fragment to complete the total synthesis. Spectroscopic
and chiroptical comparisons of these diastereomers with
the natural product data should finally unambiguously
determine the complete stereochemistry of leioderma-
tolide. Work toward this end will be reported in due
course.
With the complete C1ÀC17 skeleton in place, attention
now focused on closure of the macrolactone. Accordingly,
basic ester hydrolysis (LiOH) followed by acid-mediated
desilylation (CSA) provided the truncated seco-acid 21
(54%). In the event, we were delighted to observe com-
pletely regioselective macrolactonization under Yamagu-
chi conditions,22 forming the desired 16-membered macro-
cycle in high yield (95%), presumably reflecting a favorable
conformational preorganization of the substrate.
Completion of the macrocyclic portion of leiodermato-
lide finally required regioselective installation of the car-
bamate at C9. This was predicted based on our previous
experience in derivatizing the natural product which had
indicated a highly crowdedenvironment aboutC7.2a Inthe
event however, treatment of the macrocyclic diol with a
slight excess (1.1 equiv) of Cl3C(O)NCO at À78 °C followed
by hydrolytic workup23 provided only a ca. 3:2 mixture
(87% combined yield) of the C7 and C9 monoacylated
products, from which the natural product-like C9 carba-
mate 2 was isolated in 35% yield after HPLC
purification.
Figure 1. NMR comparison for macrocycle 2 with leioderma-
tolide (1).
Acknowledgment. We thank Clare College, Cambridge
(fellowship to S.M.D.), the Commonwealth Scholarship
Commission and St John’s College, Cambridge (scholar-
ship to T.P.), Dr Guy Naylor (Cambridge) for preliminary
experimental studies, Dr Amy Wright (Harbor Branch
Oceanographic Institute) for helpful discussions, and the
EPSRC National Mass Spectrometry Service (Swansea) for
mass spectra.
Detailed NMR comparisons for macrocycle 2 with the
natural product data for the C1ÀC16 region2a were carried
out (Figure 1). While some deviation at C15, C16, and C30
is to be expected for this macrocyclic truncate, all other 1H
resonances fell within (0.03 ppm of the corresponding
natural product values, and 13C resonances within (1.0
ppm, including particular homology for the C6ÀC9
region. Good correlation was also observed for the respec-
3
tive JH,H coupling constants about the macrocycle, and
Supporting Information Available. Experimental de-
tails and spectroscopic data for new compounds,
and details of NMR comparisons of synthetic macro-
cycle 2 with natural leiodermatolide. This material
is available free of charge via the Internet at http://
pubs.acs.org.
key NOE enhancements indicative of the proposed stereo-
chemistry.24 Together, these comparisons provide a strong
indication of the validity of our stereochemical assignment
for this region of leiodermatolide.2a
In conclusion, we have prepared the macrocyclic core of
leiodermatolide in 19 steps from 8. Notably, macrocycle 2
(22) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M.
Bull. Chem. Soc. Jpn. 1979, 52, 1989.
(23) (a) Kocovsky, P. Tetrahedron Lett. 1986, 27, 5521. (b) Paterson,
I.; Florence, G. J.; Gerlach, K.; Scott, J. P. Angew. Chem., Int. Ed. 2000,
39, 377.
(20) Wulff, W. D.; Peterson, G. A.; Bauta, W. E.; Chan, K.-S.;
Faron, K. L.; Gilbertson, S. R.; Kaesler, R. W.; Yang, D. C.; Murray,
C. K. J. Org. Chem. 1986, 51, 277.
€
(21) Furstner, A.; Funel, J. A.; Tremblay, M.; Bouchez, L. C.; Nevado,
C.; Waser, M.; Ackerstaff, J.; Stimson, C. C. Chem. Commun. 2008, 2873.
(24) See the Supporting Information for detailed comparisons.
Org. Lett., Vol. 13, No. 16, 2011
4401