Synthesis of antimitotic analogs of the microtubule stabilizing sponge alkaloid ceratamine A

Article abstract of DOI:10.1021/ol7030284

Antimitotic analogs of the microtubule stabilizing sponge alkaloid ceratamine A (1) have been synthesized starting from tribromoimidazole. A key step In the synthesis Is the formation of the azepine ring via an intramolecular Buchwald coupling between a vinyl bromide and a N-methyl amide. This represents the first synthesis of a fully unsaturated imidazo[4,5,d]azepine. NMR data obtained for the synthetic ceratamine analogs has provided support for the structure assigned to the natural product.

Full text of DOI:10.1021/ol7030284

ORGANIC
LETTERS
2008
Vol. 10, No. 6
1051-1054
Synthesis of Antimitotic Analogs of the
Microtubule Stabilizing Sponge Alkaloid
Ceratamine A
Matt Nodwell,† Jenna L. Riffell,‡ Michel Roberge,*,‡ and Raymond J. Andersen*,†
Departments of Chemistry and Earth & Ocean Sciences, UniVersity of British
Columbia, VancouVer, B.C., Canada, V6T 1Z1 and Department of Biochemistry and
Molecular Biology, UniVersity of British Columbia, VancouVer, B.C.,
Canada, V6T 1Z3
randersn@interchange.ubc.ca
Received December 17, 2007
ABSTRACT
Antimitotic analogs of the microtubule stabilizing sponge alkaloid ceratamine A (1) have been synthesized starting from tribromoimidazole. A
key step in the synthesis is the formation of the azepine ring via an intramolecular Buchwald coupling between a vinyl bromide and a N-methyl
amide. This represents the first synthesis of a fully unsaturated imidazo[4,5,d]azepine. NMR data obtained for the synthetic ceratamine analogs
has provided support for the structure assigned to the natural product.
Ceratamines A (1) and B (2) are antimitoic alkaloids isolated
from the marine sponge Pseudoceratina sp. collected in
Papua New Guinea.1 The imidazo[4,5,d]azepine core het-
erocycle at the oxidation state found in the ceratamines
appears to have no precedent among known natural products
or synthetic compounds. Ceratamines stabilize microtubles
but do not bind to the taxol binding site.2 They produce an
unusual antimitotic phenotype characterized by the formation
of pillar-like tubulin structures that extend vertically from
the basal cell surface and span the entire thickness of the
arrested cells.
We embarked on a total synthesis of 1 to confirm the
assigned structure and generate additional quantities of the
natural product and analogs for further biological evaluation.
Our retrosynthetic analysis of 1 is shown in Scheme 1. At
the outset, the plan was to install the C-2 methylamine and
bromine atoms in the penultimate steps of the synthesis after
the heterocyclic core had been assembled (i.e., VIIb to 1),
recognizing that desmethylamine and desbromoceratamines
would also be useful for structure confirmation and biological
evaluation. The key step in the assembly of the imidazo-
[4,5,d]azepine core was to be an intramolecular Buchwald
coupling between a vinyl halide and an amide in V to give
VI.3 It was anticipated that removal of the protecting group
from the imidazole nitrogen in VI to give VIIa would set
the stage for tautomerization to VIIb. NMR data obtained
for the natural product 1 indicated that the fully unsaturated
† Chemistry and EOS.
‡ Biochemistry and Molecular Biology.
(1) Manzo, E.; van Soest, R.; Matainaho, L.; Roberge, M.; Andersen,
R. J. Org. Lett. 2003, 5, 4591-4594.
(2) Karjala, G.; Chan, Q.; Manzo, E.; Andersen, R. J.; Roberge, M.
Cancer Res. 2005, 65, 3040-3043.
10.1021/ol7030284 CCC: $40.75
© 2008 American Chemical Society
Published on Web 02/16/2008

Scheme 1
Scheme 2
C-2 with n-butyllithium followed by addition of dimethyl-
disulfide to generate the C-2 thiomethylimidazole 4. In the
same pot without workup, a second equivalent of n-
butyllithium was added, resulting in metalation at C-5, and
this product was reacted with tributyltin chloride to give the
C-5 stannylated product 5. Finally, addition of a third
equivalent of n-butyllithium to 5 in the same pot gave
metalation at C-4, and this lithiated species was reacted with
DMF to generate the aldehyde 6.
Synthesis of the cinnamic acid ester moiety required for
the Stille coupling started with conversion of tribromoace-
tylchloride to the methyl ester 7 (Scheme 2). Condensation
of the tribromo ester 7 with anisaldehyde mediated by CrCl2
at rt gave the desired Z- methoxycinnamate 8 in excellent
yield.7 Stille coupling between the bromocinnamate 8 and
the tributylstannylimidazole 6 gave the desired E-cinnamate
9 in good yield (Scheme 3). The aldehyde 9 was converted
to the Z-vinyl bromide 11 using (bromomethyl)triphenylphos-
phoniumbromide and lithium hexamethyldisilazide.8 Hy-
drolysis of the ester in 11 with aqueous lithium hydroxide
gave the corresponding acid 12. The carboxylic acid 12 was
converted to the HOBt ester 13 by activation with DIPC
followed by addition of HOBt. Treatment of the ester 13
with excess methylamine in THF/CH2Cl2 generated the
N-methylamide 14 in good yield.
imidazo[4,5,d]azepine core had aromatic properties, leading
to the expectation that tautomeric equilibria for VII would
strongly favor VIIb. The vinyl halide in V was to be
generated from a Wittig reaction,3b,4 and the amide was to
come from the methoxycinnamic acid ester fragment III,
after it was added to the imidazole II via a Stille coupling
to give IV.5 The stannylated imidazole intermediate II was
to be prepared from commercially available tribrominated
imidazole I (P ) H) by sequential installation of BOM,
methylsulfide, tributyltin, and aldehyde functionalities.6
The synthesis started with N-1 protection of tribromoimi-
dazole by treatment with BOMCl and K2CO3 in DMF to
give 3 (Scheme 2). Tribromoimidazole 3 was sequentially
metalated and functionalized in a one-pot procedure follow-
ing a literature precedent.6 Thus, 3 was first metalated at
The key step in the synthesis was formation of the azepine
ring. This was accomplished using Buchwald methodology
involving a copper-catalyzed intramolecular coupling be-
tween the Z-vinylbromide and N-methyl amide functionalities
in 14 to give the enamide 15 (Scheme 4). With the core
skeleton of ceratamine A assembled, we turned our attention
to installing the amino functionality at C-2. Following a
Weinreb precedent, the methyl sulfide functionality in 15
was oxidized in poor yield to the corresponding sulfoxide
(3) (a) Jiang, L.; Job, G. E.; Klapars, A.; Buchwald, S. L. Org. Lett.
2003, 5, 3667-3669. (b) Meketa, M. M.; Weinreb, S. M. Org. Lett. 2007,
9, 853-855. (c) Toumi, M.; Couty, F.; Evano, G. J. Org. Chem. 2007, 72,
9003-9009.
(4) Gassman, P. G.; Lee, C. Tetrahedron Lett. 1989, 30, 2175-2178.
(5) (a) Farina, V.; Kapadia, S.; Krishnan, B.; Wang, C.; Liebeskind, L.
S. J. Org. Chem. 1994, 59, 5905-5911. (b) Sahli, S.; Stump, B.; Welti, T.;
Schweizer, W. B.; Diederich, F.; Blum-Kaelin, D.; Aebi, J. D.; Bo¨hm, H.
J. HeIV. Chim. Acta 2005, 88, 707-730. (c) Palmisano, G.; Santagostino,
M. Synlett 2003, 771-773. (d) Choshi, T.; Yamada, S.; Sugino, E.; Kuwada,
T.; Hibino, S. J. Org. Chem. 1995, 60, 5899-5904.
(6) (a) Lipshutz, B. H.; Hagen, W. Tetrahedron Lett. 1992, 33, 5865-
5868. (b) Groziak, M. P.; Wei, L. J. Org. Chem. 1991, 56, 4296-4300. (c)
Groziak, M. P.; Wei, L. J. Org. Chem. 1992, 57, 3776-3780. (d) Meketa,
M.; Weinreb, S. Tetrahedron 2007, 63, 9112-9119.
(7) Barma, D. K.; Kundu, A.; Zhang, H.; Mioskowski, C.; Falck, J. R.
J. Am. Chem. Soc. 2003, 125, 3218-3219.
(8) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 2173-2174.
1052
Org. Lett., Vol. 10, No. 6, 2008

Scheme 3
Scheme 4
16 with m-chloroperbenzoic acid. All attempts to displace
the sulfoxide in 16 with methyl amine in a sealed tube at
high temperatures to give 17 or with azide to give 18 failed
to give detectable products.
An alternate approach to introduction of an amine at C-2
started with the reductive removal of the methyl sulfide in
15 using Raney nickel to give 19.9 Deprotonation of 19 at
C-2 with methyl lithium, followed by treatment of the
carbanion with tosylazide10 and subsequent workup with H2O
and reduction of the azide with H2 and catalyst, proceeded
cleanly to give the amine 20 (Scheme 4). Removal of the
BOM protecting group from 20 with aluminum trichloride6d
gave the ceratamine A analog 21 having a hydroxyl at C-11
and a small amount of the corresponding C-11 ketone 22.
Similarly, removal of BOM from 15 and 19 with aluminum
trichloride6d gave the methylthio ceratamine analog 23 and
analog 24 lacking the C-2 amino functionality, respectively.
However, neither of the analogs 23 or 24 isomerized to give
the putative aromatic tautomers (i.e., VIIb).
The generation of a C-11 alcohol in the ceratamine analog
21 during the AlCl3 promoted deprotection of 20 and aqueous
workup was unexpected. Formally, the OH can be introduced
by Michael addition of water to the C-11/C-5/C-6 enamide
followed by air oxidation of the resulting dihydroceratamine
(Scheme 5). However, alcohol formation at C-11 is only
observed in the deprotection of 20 and not in the deprotection
of 15 or 19. Therefore, the process appears to depend on
the presence of an amino functionality at C-2, and it is
probably mechanistically more complex than shown. The
initial product from deprotection of 20 is colorless, too
unstable to isolate, and turns the bright yellow, characteristic
of the amino ceratamine chromophore in 21 and 22, upon
stirring in acidic water exposed to the air. This is consistent
with air oxidation of a dihydroceratamine intermediate as
shown in Scheme 5. Further air oxidation of 21 in the
presence of strong acid (HCl) as proposed in Scheme 5 can
lead to the C-11 ketone in 22. Attempts to remove the BOM
in 21 by hydrogenation or to remove the C-11 alcohol or
brominate the methoxyphenyl ring in 21 via various methods
have thus far all resulted in decomposition. These failures
demonstrate that the synthesis will have to be redesigned to
reach the natural products 1 and 2.
(9) (a) Back, T. G.; Baron, D. L.; Yang, K. J. Org. Chem. 1993, 58,
2407-2413. (b) Meketa, M. L.; Weinreb, S. M.; Nakao, Y.; Fusetani, N.
J. Org. Chem. 2007, 72, 4892-4899.
(10) (a) Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am.
Chem. Soc. 1990, 112, 4011-4030. (b) Lindel, T.; Hochgurtel, M. J. Org.
Chem. 2000, 65, 2806-2809.
1
The H NMR data for compounds 21, 22, 23, and 24
provided insight into the tautomeric equilibria and potential
aromaticity in these ceratamine analogs (Scheme 4). Of
Org. Lett., Vol. 10, No. 6, 2008
1053

aminoimidazoazepine ring of 21, based on HMBC and
HSQC data, were in excellent agreement with the corre-
sponding assignments for 1.11 The similarity between the 1H
and 13C NMR assignments for the synthetic analog 21 and
the natural product 1 provides strong support for the structure
assigned to the natural product. Significant downfield shifts
of the H-8, H-9 and N-Me resonances in 1, 21, and 22
compared with their counterparts in 23 and 24 can be
attributed to the presence of a ring current, suggesting that
the aminoimadzoazepine core in the ceratamines 1, 21, and
22 has aromatic character.
Scheme 5
Ceratamine analogs 21 and 22 both showed antimitotic
activity with potency close to ceratamines A and B in a cell-
based assay (Supporting Information), demonstrating that
Me-19 and the bromine atoms in 1 are not required for
activity.2 The analogs 23 and 24 were inactive, suggesting
that either one or both of the aromatic imidazo[4,5,d]azepine
heterocyclic core or an amino substituent at C-2 are required
for antimitotic activity.
In summary, the synthesis of ceratamines 21 and 22 has
confirmed the structures assigned to the natural products 1
and 2, provided a route for the preparation of antimitotic
ceratamine analogs for further biological evaluation, and
revealed preliminary SAR for this novel microtubule-
stablizing pharmacophore. The synthesis, which utilizes an
intramolecular Buchwald reaction to assemble the azepine
ring in the ceratamines, represents the first preparation of
an aromatic imidazo[4,5,d]azepine heterocyle.
particular interest were the resonances assigned to the
enamide protons H-8 and H-9 and the protons on the benzylic
carbon C-11. The H-8 enamide protons in 23 and 24 have
chemical shifts (CD2Cl2) of δ 6.23 and 6.27, the H-9 protons
have shifts of δ 6.05 and 6.03, and the H-11 protons appear
as a singlets at δ 7.10 and 7.05, respectively, consistent with
the presence of an E ∆5,11 alkene. The H-8 and H-9
resonances (CD3OD) are shifted significantly downfield to
δ 7.81 and 6.59 in 21 and to δ 6.65 and 7.89 in 22,
respectively, close to the corresponding shifts in ceratamine
A (1) {DMSO-d6 δ 7.73 (H-8), 6.42 (H-9)}. The N-methyl
resonances have shifts of δ 3.12 and 3.06 in 23 and 24
compared with shifts of δ 3.65 and 3.66 in 21 and 22, and
Acknowledgment. Financial support was provided by
NSERC (R.J.A.), NCIC (R.J.A.), and CIHR (M.R.).
Supporting Information Available: Experimental details
and NMR spectra for compounds 6, 9, 11, 13, 14, 15, 19,
20, 21, 22, 23, and 24. This material is available free of
charge via the Internet at http://pubs.acs.org.
1
δ 3.56 in 1. H-11 appears as a singlet at δ 6.34 in the H
OL7030284
NMR spectrum of 21 and shows an HSQC correlation to a
13C resonance at δ 75.3, consistent with a C-11 carbinol
methine instead of a ∆5,11 alkene. 13C assignments for the
(11) 21: δ 176.4 (C-2), 160.5 (C-4), 126.8 (C-5), 165.7 (C-6), 145.6
(C-8), 102.3 (C-9), 173.2 (C-10); 1: δ 175.6 (C-2), 160.5 (C-4), 121.3
(C-5), 164.1 (C-6), 142.9 (C-8), 100.4 (C-9), 169.7 (C-10).
1054
Org. Lett., Vol. 10, No. 6, 2008