Synthetic approaches to the microtubule-Sabilizing sponge alkaloid ceratamine A and desbromo analogues

Article abstract of DOI:10.1021/jo802322s

Two synthetic approaches to the microtubule-stabilizing ceratamine alkaloids are described. The first approach involved attempts to graft an aminoimidazole moiety onto an azepine ring to form partially hydrogenated versions of the unprecedented aromatic imidazo[4,5-d]azepine core of the ceratamines. This route ultimately failed because it was not possible to aromatize the partially hydrogenated ceratamine intermediates. A second approach started with tribromoimidazole that was sequentially metalated and functionalized to efficiently generate a key imidazole intermediate containing vinyl bromide and amide functionalities. An intramolecular Buchwald vinyl amidation reaction converted this key intermediate into a bicyclic imidazo[4,5-d]azepine that was at the same oxidation state as the aromatic core of the ceratamines. The 2-amino functionality present on the imidazole ring of the ceratamines was installed using a Buchwald/Hartwig amination reaction on a 2-chloroimidazole precursor. Deprotection and aromatization resulted in the first synthesis of desbromoceratamine A (55) and desmethyldesbromoceratamine A (60). An unanticipated addition of atmospheric oxygen was encountered during deprotection of the imidazole ring in the last step of the synthesis leading to C-11 oxygenated ceratamine analogues as byproducts. Evaluation of the synthetic ceratamines in a TG3 cell-based assay for mitotic arrest revealed that the C-14 and C-16 bromine substituents in ceratamine A (1) play a major role in the antimitotic potency of the natural product. The synthetic route to ceratamine analogues has provided sufficient quantities of desbromoceratamine A (55) for testing in mouse models of cancer.

Full text of DOI:10.1021/jo802322s

Synthetic Approaches to the Microtubule-Stabilizing Sponge
Alkaloid Ceratamine A and Desbromo Analogues
Matt Nodwell,† Alban Pereira,† Jenna L. Riffell,‡ Carla Zimmerman,‡ Brian O. Patrick,§
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 October 16, 2008
Two synthetic approaches to the microtubule-stabilizing ceratamine alkaloids are described. The first approach
involved attempts to graft an aminoimidazole moiety onto an azepine ring to form partially hydrogenated
versions of the unprecedented aromatic imidazo[4,5-d]azepine core of the ceratamines. This route ultimately
failed because it was not possible to aromatize the partially hydrogenated ceratamine intermediates. A second
approach started with tribromoimidazole that was sequentially metalated and functionalized to efficiently generate
a key imidazole intermediate containing vinyl bromide and amide functionalities. An intramolecular Buchwald
vinyl amidation reaction converted this key intermediate into a bicyclic imidazo[4,5-d]azepine that was at the
same oxidation state as the aromatic core of the ceratamines. The 2-amino functionality present on the imidazole
ring of the ceratamines was installed using a Buchwald/Hartwig amination reaction on a 2-chloroimidazole
precursor. Deprotection and aromatization resulted in the first synthesis of desbromoceratamine A (55) and
desmethyldesbromoceratamine A (60). An unanticipated addition of atmospheric oxygen was encountered
during deprotection of the imidazole ring in the last step of the synthesis leading to C-11 oxygenated ceratamine
analogues as byproducts. Evaluation of the synthetic ceratamines in a TG3 cell-based assay for mitotic arrest
revealed that the C-14 and C-16 bromine substituents in ceratamine A (1) play a major role in the antimitotic
potency of the natural product. The synthetic route to ceratamine analogues has provided sufficient quantities
of desbromoceratamine A (55) for testing in mouse models of cancer.
Introduction
products with widely divergent chemical structures has been
discovered, and the new anticancer drug ixabepilone based on
the epothilone scaffold has recently been approved for clinical
use.2,3 Nearly all of the known microtubule-stabilizing natural
products were initially discovered because they were potent
cytotoxins, and it was only subsequent investigation of their
mechanism of action that revealed their abilities to arrest cells
in mitosis and target tubulin. More recently, a cell-based assay
has provided a sensitive and robust method to directly screen
natural product extracts for the presence of compounds that
arrest cells in mitosis.4 Microscopic examination of tubulin in
cells that have been arrested by a hit in the assay can often
reveal if the active components are microtubule-depolymerizing
or -stabilizing agents. In this way, the assay can be used as an
The widespread clinical use of Vinca alkaloids, paclitaxel,
and synthetic analogues of these natural products for treating
cancer has provided unparalleled validation of tubulin as a
molecular target for developing new anticancer drugs. Horwitz’s
seminal discovery that paclitaxel stabilizes microtubules and
the subsequent success of paclitaxel in the clinic stimulated a
vigorous worldwide search for additional natural product
chemotypes possessing this important biological activity.1 As
a result, a substantial collection of microtubule-stabilizing natural
† Departments of Chemistry and EOS.
‡ Department of Biochemistry and Molecular Biology.
§ Department of Chemistry.
(1) Schiff, P. B.; Fant, J.; Horwitz, S. B. Nature 1979, 277, 665–667.
10.1021/jo802322s CCC: $40.75
Published on Web 01/07/2009
2009 American Chemical Society
J. Org. Chem. 2009, 74, 995–1006 995

Nodwell et al.
SCHEME 1
effective direct screen for microtubule-stabilizing natural prod-
ucts, and it enabled our discovery of caribaeoside,5 the cerata-
mines,6 and the nigricanosides.7
Ceratamines A (1) and B (2) are novel alkaloids isolated from
the marine sponge Pseudoceratina sp. collected in Papua New
Guinea.6 The imidazo[4,5-d]azepine core heterocycle at the
oxidation state found in the ceratamines has no precedent among
known natural products or synthetic compounds. Ceratamines
stabilize microtubules but do not bind to the taxol binding site,
and they produce a novel mitotic arrest phenotype characterized
by the presence of multiple pillar-like structures of tubulin.8
The ability of ceratamines to stabilize microtubules and elicit a
novel phenotype combined with their relative structural simplic-
ity and lack of chiral centers makes them attractive anticancer
drug leads. Scarcity of the natural products from the original
collection and the potential negative environmental impact of
recollecting additional source-sponge tissue created a need for
a synthetic source of the ceratamines and analogues to enable
further in vitro and in vivo evaluation of this new heterocyclic
microtubule-stabilizing pharmacophore. We embarked on a total
synthesis of the ceratamines to solve the supply problem, and
in a recent communication we disclosed a route to a C-11
hydroxy ceratamine analogue that showed in vitro antimitotic
activity.9 However, challenges associated with scaling up this
first generation route to a bioactive ceratamine and our failure
to make analogues that were not oxygenated at C-11 led us to
search for alternate methods to produce ceratamines. Herein,
we describe the complete details of our synthetic efforts to make
ceratamines, including description of a more efficient route to
the 2-aminoimidazo[4,5-d]azepine heterocyclic core and the
generation of new analogues lacking C-11 oxygenation that have
revealed significant information about the SAR for the antimi-
totic pharmacophore.
chromatographic behavior, and unexpected reactivities.10,11
Therefore, it was decided to introduce the 2-methylamino group
at a late stage in all of our synthetic approaches to ceratamines.
Our first approach to the imidazo[4,5-d]azepine ring system was
to build the five-membered 2-aminoimidazole ring onto a
suitably functionalized caprolactam ring and in the process
install three of the four nitrogen atoms of the ceratamines in
one operation (Scheme 1). We planned to construct partially
hydrogenated versions of the imidazo[4,5-d]azepine ring system
(i.e., III) and then take advantage of the putative aromaticity
of the fully dehydrogenated heterocycle as a driving force to
achieve the oxidation state found in the natural products.12
2-Aminoimidazoles have been formed in good yield via
condensation of guanidine derivatives with R-halo ketones or
R-pseudohalo ketones.13 In order to implement this approach
to ceratamine A, a suitably functionalized caprolactam such as
II shown in the retrosynthesis in Scheme 1 was required, and
it was anticipated that the R-halo ketone functionality in II could
be formed from the corresponding alkene in the caprolactam I.
The key caprolactams 12 and 13 were prepared as shown in
Scheme 2 as a first step in exploring the plan outlined in Scheme
1. Butenoic acid 3 was readily R-alkylated with benzylbromides 4
and 5 to give the carboxylic acids 6 and 7, and butenylamine 9
was routinely prepared from butenylbromide 8 (see Supporting
Information for experimental details). Coupling of butenylamine
9 with carboxylic acids 6 and 7 via their respective acid chlorides
gave amides 10 and 11 in good yield. Ring-closing metathesis was
carried out on 10 and 11 with Grubbs second-generation catalyst
to yield caprolactams 12 and 13, respectively.14 Model lactam 12
was converted to the R-tosylketone 16 in low yield by sequential
epoxidation, tosylation, and oxidation following a literature pre-
cedent (Scheme 3).15 Reaction of the pseudohalo ketone 16 with
guanidine in an attempt to form 17a gave an intractable mixture
of products. Attempts to form 17b by treatment of 16 with
N-acetylguanidine gave only starting material.
Results and Discussion
Molecules containing a 2-aminoimidazole motif are often
challenging to work with because of their poor solubility, poor
(2) Examples. (a) Eleutherobin: Lindel, T.; Jensen, P. R.; Fenical, W.; Long,
B. H.; Casazza, A. M.; Carboni, J.; Fairchild, C. R. J. Am. Chem. Soc. 1997,
119, 8744–8745. (b) Laulimalide: Mooberry, S. L.; Tien, G.; Hernandez, A. H.;
Plubrukam,A.;Davidson,B.S.CancerRes.1999,59,653–660. (c)Peloruside: Hood,
K. A.; West, L. M.; Rouwe, B.; Northcote, P. T.; Berridge, M. V.; Wakefield,
S. J.; Miller, J. H. Cancer Res. 2002, 62, 3356–3360. (d) Discodermolide: ter-Haar,
E.; Kowalski, R. J.; Hamel, E.; Lin, C. M.; Longley, R. E.; Gunasekera, S. P.;
Rosenkranz,H.S.;Day,B.W.Biochemistry1996,35,243–250. (e)Epothilone:Bollag,
D. M.; McQueney, P. A.; Zhu, J.; Koupal, L.; Liesch, J.; Goetz, M.; Lazarides,
E.; Woods, C. M. Cancer Res. 1995, 55, 2325–2333.
Further attempts to form the aminoimidazole ring from
guanidine began with the formation of halohydrin 18 (Scheme
4). Lactam 13 was reacted with N-bromosuccinimide and water
(3) Donovan, D.; Vahdat, L. T. Oncology 2008, 22, 408–416.
(4) Roberge, M.; Cinel, B.; Anderson, H. J.; Lim, L.; Jiang, X.; Xu, L.; Kelly,
M. T.; Andersen, R. J. Cancer Res. 2000, 60, 5052–5058.
(5) Cinel, B.; Roberge, M.; Behrisch, H.; van Ofwegen, L.; Castro, C. B.;
Andersen, R. J. Org. Lett. 2000, 2, 257–260.
(6) Manzo, E.; van Soest, R.; Matainaho, L.; Roberge, M.; Andersen, R. J.
Org. Lett. 2003, 5, 4591–4594.
(7) Williams, D. E.; Sturgeon, C. M.; Roberge, M.; Andersen, R. J. J. Am.
Chem. Soc. 2007, 129, 5822–5823.
(8) Karjala, G.; Chan, Q.; Manzo, E.; Andersen, R. J.; Roberge, M. Cancer
Res. 2005, 65, 3040–3043.
(10) O’Malley, D. P.; Li, K.; Maue, M.; Zografos, A. L.; Baran, P. S. J. Am.
Chem. Soc. 2007, 129, 4762–4775.
(11) Meketa, M. L.; Weinreb, S. M.; Nakao, Y.; Fusetani, N. J. Org. Chem.
2007, 72, 4892–4899.
(12) Shengule, S. R.; Karuso, P. Org. Lett. 2006, 8, 4083–4084.
(13) Little, T. L.; Webber, S. E. J. Org. Chem. 1994, 59, 7299–7305.
(14) Deiters, A.; Martin, S. F. Chem. ReV. 2004, 104, 2199–2238.
(15) (a) Chochrek, P.; Wicha, J. Org. Lett. 2006, 8, 2551–2553. (b) Martin,
J. D.; Perez, C.; Ravelo, J. L. J. Am. Chem. Soc. 1985, 107, 516–518.
(9) Nodwell, M.; Riffell, J. L.; Roberge, M.; Andersen, R. J. Org. Lett. 2008,
10, 1051–1054.
996 J. Org. Chem. Vol. 74, No. 3, 2009

Microtubule-Stabilizing Sponge Alkaloid Ceratamine A
SCHEME 2
SCHEME 4
SCHEME 3
21 and 23. Recognizing that these compounds would provide
valuable insight into the pharmacophore of the ceratamines,
caprolactams 21 and 23 were methylated at the amide nitrogen
with MeI and K2CO3 to yield heterocycles 22 and 24, which
were evaluated for biological activity.
The failure of guanidine derivatives to condense with the
cyclic R-halo ketones 19 and 20 prompted us to explore alternate
methods to fuse the imidazole ring onto lactam 13. Substituted
guanidine derivatives have been prepared by the addition of
amines to cyanamides.17 We envisioned that treatment of an
appropriate vicinal bromoalkylcyanamide with ammonia would
give rise to a dihydro version of the 2-aminoimidazole ring of
the ceratamines and dehydrogenation of this compound would
give rise to the required aromatic imidazo[4,5-d]azepine.
Cyanamide can be added to alkenes through a bromonium ion,
and using this methodology the bromoalkyl cyanamide 25 was
prepared in moderate yield as outlined in Scheme 5.18 Treatment
of cyanamide 25 with ammonia in MeOH in a sealed tube
yielded the reduced ceratamine B analogue 26 in excellent yield.
Attempting this reaction with methylamine or benzylamine failed
to yield any of the desired substituted products. Completion of
in acetone to yield 18 in excellent yield. Oxidation of 18 with
3 equiv of Dess-Martin periodinane (DMP) yielded the
bromoenamide 19 in quantitative yield. The formation of 19
was advantageous, as it already possessed the requisite ∆6,7
olefin present in the ceratamines. This result was consistent with
a report that the use of a closely related hypervalent iodine
oxidant, iodoxybenzoic acid (IBX), is an efficient method for
the one-pot conversion of alcohols, ketones, and aldehydes to
R,ꢀ-unsaturated carbonyl compounds.16 Interestingly, oxidation
of 18 with IBX yields exclusively the C6-C7 saturated
compound 20. Multiple attempted reactions of these compounds
with guanidine derivatives yielded only elimination products
(17) Simig, G.; Lempet, K.; Tama´s, J.; Czira, G. Tetrahedron 1975, 31, 1195–
1200.
(16) Nicolaou, K. C.; Zhong, Y.-L.; Baran, P. S. J. Am. Chem. Soc. 2000,
122, 7596–7597.
(18) Zhang, C.; Yang, R.; Chen, L.; Zhong, B.; Yun, L.; Wang, H. J. Labelled
Compd. Radiopharm. 2004, 41, 583–590.
J. Org. Chem. Vol. 74, No. 3, 2009 997

Nodwell et al.
SCHEME 5
SCHEME 7
SCHEME 6
the synthesis of ceratamine B via this route was curtailed by
the almost total intractability and chemical instability of 26.
Attempts to further purify the compound via RP-HPLC or
oxidizeittoanaromaticceratamineresultedonlyindecomposition.
Our lack of success in grafting the 2-aminoimidazole ring
onto a preexisting caprolactam prompted us to change strategy
and explore avenues aimed at building a substituted caprolactam
onto a preexisting imidazole as outlined in Scheme 6. This
second approach made use of the fact that tribromoimidazoles
can be selectively and sequentially metalated leading to trisub-
stituted imidazole intermediates of general structure V.19 Once
again, the plan was to convert X (-SMe or Cl) to an amino
substituent near the end of the synthesis (X to 1) to avoid the
difficulties associated with handling 2-aminoimidazoles. The
cornerstone of this approach to the ceratamines was to be
the use of an intramolecular Buchwald vinyl amidation reaction
to create the lactam ring (VIII to IX).20 Implicit in our synthetic
plan was the expectation that the deprotected intermediate Xa
would spontaneously rearrange to the presumably more stable
aromatic isomer Xb. The initial planning was also based on
the assumption that the bromine atoms in ceratamines A (1)
and B (2) did not play a significant role in their antimitotic
activity, making the desbromo analogues just as attractive targets
as the natural products. This assumption stemmed from the
observation that many sponge metabolites contain the dibro-
momethoxyphenyl moiety present in the ceratamines but do not
show antimitotic activity. Therefore, at the outset it seemed
likely that the novel imidazo[4,5-d]azepine core was the
dominant feature of the ceratamine antimitotic pharmacophore.
The synthesis began with BOM-protected imidazole 27 prepared
by treatment of tribromoimidazole with BOMCl and K2CO3 in
DMF (Scheme 7). Metalation of 27 with 1 equiv of nBuLi,
followed by quenching with MeSSMe first introduced the 2-thi-
omethyl moiety.20 In the same pot, without workup, further
metalations and quenchings following literature precedent20b
installed the tributylstannane and aldehyde groups at C-5 and C-4,
respectively, to form 28. Cinnamic ester 30 was synthesized via
Cr(II)-mediated condensation of anisaldehyde with tribromoester
(19) (a) Lipshutz, B. H.; Hagen, W. Tetrahedron Lett. 1992, 33, 5865–5868.
(b) Groziak, M. P.; Wei, L. J. Org. Chem. 1991, 56, 4296–4300l. (c) Groziak,
M. P.; Wei, L. J. Org. Chem. 1992, 57, 3776–3780. (d) Meketa, M.; Weinreb,
S. Tetrahedron 2007, 63, 9112–9119.
(20) (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.
998 J. Org. Chem. Vol. 74, No. 3, 2009

Microtubule-Stabilizing Sponge Alkaloid Ceratamine A
29.21 Stille coupling between 28 and 30 in toluene yielded 31 in
high yield.22 HMBC correlations observed between the C-5 carbon
resonance at δ 137.2 and both the alkene resonance at δ 8.10 and
the BOM methylene resonances at δ 5.00/5.08 in 31 confirmed
that the cinnamoyl residue was attached to C-5 of the imidazole
as drawn in Scheme 7. Z-Vinyl bromide 32 was synthesized by
reaction of the aldehyde 31 with (bromomethyl)triphenylphospho-
niumbromide and lithium hexamethyldisilazide in the presence of
HMPA.23 Saponification of the methyl ester 32, activation of the
acid as the HOBt ester 33, and displacement of the HOBt group
with excess methylamine yielded the secondary amide 34. The key
step in the synthesis was the copper(I)-catalyzed ring closure of
34 to form the enamide 35 following a procedure developed by
Buchwald.24 A dilute THF solution of 34 was treated with CuI,
Cs2CO3, and N,N′-dimethylethylenediamine at elevated tempera-
tures to yield the 5,7-bicyclic system 35 in excellent yield.
SCHEME 8
Enamide 35 was already at the oxidation state of the
ceratamines so the only remaining transformations required to
reach a desbromoceratamine were installation of the 2-amino
functionality, removal of the BOM protecting group, and
tautomerization to form the fully aromatic ceratamine hetero-
cyclic system. Unfortunately, the 2-methylthio group in 35
proved to be resistant to substitution by a variety of nitrogen
nucleophiles. Although oxidation of the methylthio group to
the sulfoxide/sulfone 36 could be carried out, albeit in very poor
yield due to competing oxidations, nucleophilic substitution of
the oxidized 2-methylthio group failed as well (Scheme 8). In
retrospect, this result was not surprising considering the
documented need for electron-withdrawing substituents at C-4
or C-5 to promote nucleophilic aromatic substitution in the
2-position of an imidazole.25
SCHEME 9
The failure of nucleophilic substitution as an approach to
install the 2-amino group shifted our attention to reactions
between an electrophilic azide and a C-2 carbanion.26 Compound
35 was reductively desulfurized with Ra-Ni to generate 37, the
precursor to the required imidazole C-2 carbanion (Scheme 8).
This reaction was unsatisfactory for a number of reasons. The
use of flammable Ra-Ni, large solvent volumes, adsorption of
the compound to the metal surface, and over-reduction leading
to low yields were all problems that characterized the transfor-
mation. Extensive experimentation with other desulfurizing
agents failed to improve the yield or the ease of conducting the
reduction. Nevertheless, a workable amount of 37 was generated,
and following deprotonation with MeLi, quenching with multiple
equivalents of TsN3, and catalytic reduction of the azido group
with H2, the desired primary amine 38 was produced relatively
cleanly. Attempted deprotection of the BOM group by hydro-
genation or HCl-catalyzed hydrolysis either failed or resulted
in decomposition. However, treatment of 38 with AlCl3 followed
by H2O workup gave the C-11 oxygenated ceratamine analogues
39 and 40 (Scheme 9).27
The addition of an oxygen atom at the C-11 position of the
ceratamine skeleton in 39 and 40 during AlCl3-mediated
deprotection of 38 was totally unexpected, especially since
AlCl3-mediated removal of the BOM group in the 2-methylthio
analog 35 and the 2-protio compound 37 did not result in
isomerization to the fully aromatic system or oxygenation at
the C-11 benzylic position (Scheme 9). It was clear that further
investigation of these unexpected results was required to
illuminate a solution to preventing the unwanted C-11 oxygen-
ation of 38. However, several low-yielding steps in the route
to 38 shown in Scheme 8 made it difficult to generate sufficient
material for mechanistic investigations of the AlCl3-mediated
deprotection or the eventual production of ceratamine analogues
for biological evaluation by this method, so we set out to develop
a more efficient synthesis of 38.
(21) Barma, D. K.; Kundu, A.; Zhang, H.; Mioskowski, C.; Falck, J. R. J. Am.
Chem. Soc. 2003, 125, 3218–3219.
(22) (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.
(23) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 2173–2174.
(24) (a) Jiang, L.; Job, G. E.; Klapars, A.; Buchwald, S. L. Org. Lett. 2003,
5, 3667–3669. (b) Toumi, M.; Couty, F.; Evano, G. J. Org. Chem. 2007, 72,
9003–9009.
(25) Jarosinski, M. A.; Anderson, W. K. J. Org. Chem. 1991, 56, 4058–
4062.
(26) (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. (c) Kawasaki, I.; Sakaguchi, N.; Fukushima, N.; Nikaido,
F.; Yamashita, M.; Ohta, S. Tetrahedron Lett. 2002, 43, 4377–4380.
(27) Groves, J. K.; Cundasawmy, N. E.; Anderson, H. J. Can. J. Chem. 1973,
51, 1089–1098.
J. Org. Chem. Vol. 74, No. 3, 2009 999

Nodwell et al.
SCHEME 10
Our second generation route to the ceratamine skeleton was
designed to exploit literature precedents suggesting that Pd(0)-
catalyzed amination of the chloroimidazole 48 would be an
efficient method to introduce the 2-aminoimidazole moiety.28
We also anticipated that the synthetic flexibility imparted by
the 2-chloro atom would allow for rapid structure diversification
at an advanced stage of the synthesis and facilitate the SAR
analysis of the ceratamine pharmacophore. Metalation of the
BOM-protected tribromoimidazole 27, followed by quenching
with hexachloroethane, yielded crystalline chloride 43 in high
yield (Scheme 10).29 Further metalations and quenchings in the
same pot installed the tributylstannyl and aldehyde groups to
C-5 and C-4,20b respectively, to afford stannane 44. The
cinnamic amide 45 was prepared in quantitative yield via
saponification of methyl ester 30, in situ activation of the
resulting acid, and displacement with excess methylamine.
Bromide 45 was coupled with stannane 44 in THF at room
temperature to yield aldehyde 46. Conducting the Stille reaction
in toluene at elevated temperatures scrambled the alkene
geometry in 46,30 which did not occur during the Stille reaction
between 28 and 30 (Scheme 7). Wittig olefination with
(bromomethyl)triphenylphosphoniumbromide and potassium
tert-butoxide in THF yielded the surprisingly insoluble Z-vinyl
bromide 47 in good yield after precipitation from CH2Cl2 and
chromatography of the remaining soluble materials. Ring closure
of 47 via a Cu(I)-catalyzed Buchwald enamide reaction yielded
the 2-chloroimidazo[4,5-d]azepine 48. The structure of this
compound was confirmed by single crystal X-ray diffraction
analysis (Supporting Information). A palladium-mediated ami-
nation was then carried out using an ammonia surrogate in an
attempt to form 38.31 Heating chloroimidazole 48, triphenylsi-
lylamine, and LiHMDS in the presence of Pd2(dba)3 and the
Buchwald ligand 2-dicyclohexylphosphino-2′,4′,6′-triisopropy-
lbiphenyl (XPhos) in toluene followed by acid hydrolysis of
the triphenylsilyl group gave the primary amine 38.32
With an efficient route to 38 in hand, we turned our attention
back to the formation of 39 and 40. The results shown in Scheme
9 suggested that oxygenation at the C-11 position of the
ceratamine system occurs only upon AlCl3-mediated deprotec-
tion of a system with an amino substituent at the 2-position of
the imidazole ring. Less effective electron-donating substituents
(MeS-, Cl-, and H-) do not lead to C-11 oxygenated products
(28) (a) Schoffers, E.; Olsen, P. D.; Means, J. C. Org. Lett. 2001, 3, 4221–
4223. (b) Wang, Z.; Rizzo, C. J. Org. Lett. 2001, 3, 565–568. (c) Wolfe, J. P.;
Tomori, H.; Sadighi, J. P.; Yin, J.; Buchwald, S. L. J. Org. Chem. 2000, 65,
1158–1174. (d) Hooper, M. W.; Utsunomiya, M.; Hartwig, J. F. J. Org. Chem.
2003, 68, 2861–2873.
(29) (a) Eriksen, B. L.; Vedsø, P.; Morel, S.; Begtrup, M. J. Org. Chem.
1998, 63, 12–16. (b) Eriksen, B. L.; Vedsø, P.; Morel, S.; Begtrup, M. J. Org.
Chem. 2001, 66, 8344–8348.
(30) Labadie, J. W.; Teuting, D.; Stille, J. K. J. Org. Chem. 1983, 48, 4634–
4642. (b) Echavarren, A. M.; Stille, J. K. J. Am. Chem. Soc. 1987, 109, 5478–
5486.
(31) (a) Lee, S.; Jørgensen, M.; Hartwig, J. F. Org. Lett. 2001, 3, 2729–
2732. (b) Huang, X.; Buchwald, S. L. Org. Lett. 2001, 3, 3417–3419.
(32) (a) Huang, X.; Anderson, K. W.; Zim, D.; Jiang, L.; Klapars, A.;
Buchwald, S. L. J. Am. Chem. Soc. 2003, 125, 6653–6655. (b) Billingsley, K.;
Buchwald, S. L. J. Am. Chem. Soc. 2007, 129, 3358–3366.
1000 J. Org. Chem. Vol. 74, No. 3, 2009

Microtubule-Stabilizing Sponge Alkaloid Ceratamine A
SCHEME 11
SCHEME 12
with aromatic imidazo[4,5-d]azepine core heterocycles (Scheme
9). The formal Michael addition of water to the C-11 position
of 38 upon N-1 deprotection could perhaps be rationalized by
the increased reactivity of the C-11 position upon C-4 ipso
protonation available only to a 2-aminoimidazole containing
compound in acidic media as shown in Scheme 11. However,
upon deprotection of 38 with AlCl3 followed by quenching with
MeOH, only alcohol 39 and ketone 40 were formed. Further-
more, quenching the deprotection of 38 with 18O-labeled water
gave no indication of incorporation of 18O into 39 or 40 as
observed by ESI-MS or MALDI-MS. These results demon-
strated that a simple Michael addition mechanism was not in
operation and that the oxygen atoms present at C-11 in 39 and
40 must have had an alternate origin.
Another possible source of the C-11 oxygen atom in 39 and
40 was from the BOM protecting group via some intramolecular
process that was occurring during the AlCl3-mediated depro-
tection sequence. While there was no obvious mechanistic
rationale for such a process, we nevertheless investigated this
possibility by scrambling the regiochemistry of the BOM group
as outlined in Scheme 12. Compound 48 was subjected to
standard deprotection with AlCl3 in CH2Cl2 followed by
hydrolysis. Without isolation, the resulting imidazole 49 was
treated with BOMCl and K2CO3 in DMF to yield a 1:1 mixture
of compound 48 and its regioisomer 50. The regioisomers were
separated, and 50 was aminated as described above to give 51.
When 51 was deprotected with AlCl3, only compounds 39 and
40 were isolated, eliminating the possibility that C-11 oxygen-
ation resulted from some intramolecular process involving the
BOM oxygen atoms. At this point it seemed most likely that
the C-11 oxygen atoms in 39 and 40 were coming from
atmospheric oxygen.
Since the synthetic goal was to produce ceratamine analogues
that were not oxygenated at C-11, multiple attempts were made
to deoxygenate 39 at C-11 via standard hydrogenolysis and
Barton deoxygenation procedures, but all resulted in decomposi-
tion. Similarly, attempts to bypass C-11 oxygenation by carrying
out a Buchwald/Hartwig amination on the already deprotected
chloride 49 with an ammonia surrogate also failed to yield
isolable products.
Next we sought to ameliorate the putative electron-donating
effects of the aminoimidazole by acylation in order to forestall
the oxygenation of C-11 during removal of the BOM group by
AlCl3. We reasoned that once the BOM protecting group had
been removed, the acyl substituent could be cleaved or
transformed under conditions where atmospheric oxygen was
excluded, which might result in isomerization and aromatization
to the ceratamine system without C-11 oxygenation.
Reaction of chloroimidazole 49 with N-methylformamide in
the presence of Cs2CO3 and Pd2(dba)3 using 2-di-tert-butylphos-
phino-3,4,5,6-tetramethyl-2′,4′,6′-triisopropyl-1,1-biphenyl 52 as
a ligand and refluxing in toluene for 24 h gave the N-
methylformamide derivative 53 in modest yield (Scheme 13).33
Exposure of 53 to anhydrous AlCl3 in CH2Cl2 cleanly yielded
54 as evidenced by TLC and HRESIMS. Although the 1H NMR
spectrum of 54 revealed a complicated dynamic behavior, likely
due to the presence of rotamers and/or isomers, it was clear
that tautomerization/oxidation had not occurred, and the com-
pound presumably exists as drawn. Attempts to remove the
formyl group in 54 by basic hydrolysis failed. However,
hydrolysis of 54 in 1,4-dioxane and 6 N HCl(aq) gave the C-11
alcohol 56 as the major product after brief heating, and for the
first time minor amounts of the desired desbromoceratamine
55 were also formed. After much experimentation, it was found
that passage of dry HCl(g) through a solution of 54 in 50% 1,4-
dioxane and water gave clean transformation of 54 to desbro-
moceratamine A (55) (Scheme 12) with the formation of only
small amounts of 56. The stream of HCl gas employed in this
reaction presumably purged atmospheric oxygen, resulting in a
dramatic reduction in the formation of the C-11 hydroxy
analogue 56 and formation of desbromoceratamine as the major
product, consistent with atmospheric oxygen being the source
of C-11 oxygenation.
Analysis of the NMR data recorded for 55 in DMSO-d6
revealed striking similarities with the NMR data recorded for
ceratamine A (1) in the same solvent.6 The spectra reveal two
interconverting forms for both 1 and 55 in a ∼3:1 ratio. COSY
data showed that the C-19 methyl groups in both 1 and 55 are
scalar coupled to the NH-18 proton in both forms, suggesting
that in each case they are slowly interconverting rotamers about
the C-2/N-18 bonds, consistent with the doubly vinylogous
(33) Ikawa, T.; Barder, T. E.; Biscoe, M. R.; Buchwald, S. L. J. Am. Chem.
Soc. 2007, 129, 13001–13007.
J. Org. Chem. Vol. 74, No. 3, 2009 1001

Nodwell et al.
SCHEME 13
SCHEME 14
amide nature of the 2-N-methylamino substituents and the C-6
carbonyl in 1 and 55. Variable temperature 1H NMR experiments
conducted on 55 in DMSO-d6 showed a coalescence of the two
forms at 400 °C (Supporting Information).
Desbromoceratamine A (55) was treated with 2 equiv of
N-bromosuccinimide in DMF or 2 equiv of Br2 in HOAc in an
attempt to complete the synthesis of ceratamine A (1). However,
both of these reaction conditions yielded only the 9-monobro-
minated product 57, and use of a larger excess of bromine led
to very complex mixtures from which none of the desired
tribromoceratamine could be isolated (Scheme 13).
In an attempt to circumvent the long reaction times and
relatively low yields involved in the transformation of 49 to
55, the deactivating formyl group was added onto the more
readily formed primary amine 38. Reaction of 38 with acetic
formic anhydride in THF for 24 h afforded 58 in good yield
(Scheme 14).34 Formamide 58 can be readily alkylated with
MeI and K2CO3 in DMF to form 53, which represents an
improved method to make this critical desbromoceratamine
precursor. Removal of the BOM group from 58 with AlCl3 gave
formamide 59. Unfortunately, reduction of the N-formyl group
in 59 with BH3/THF followed by acidic aqueous workup again
yielded alcohol 56 exclusively. However, desmethyldesbro-
moceratamine A (60) was cleanly formed when a stream of dry
HCl(g) was bubbled through a solution of formamide 59 in 50%
1,4-dioxane and water.
a 2-amino substituent are exposed to strong acid, they can undergo
ipso protonation at C-4 of the imidazole ring to give XII.37
Analogues of XI with methylthio, proton, chloro, or N-formyl
substituents at C-2 are less basic and not expected to as readily
undergo this type of C-4 protonation. Loss of a proton from the
3-NH in XII gives the neutral species XIII. Reaction of XIII with
atmospheric oxygen generates the radical XIVa, which is greatly
stabilized by the aromatic imidazo[4,5-d]azepine core resonance
structure XIVb. Reaction of XIVb with the hydroperoxide radical
formed during the hydrogen abstraction step (XIII to XIVa),
perhaps in a cage reaction, can lead to the hydroperoxide XV that
can decompose to give the C-11 hydroxy (39 or 56) or keto (40 or
61) derivatives formed as the major products under most conditions.
Alternatively, radical XIVb can abstract a hydrogen atom from
some source to give the desired nonoxygenated analogues 55 and
60. The formation of 55 and 60 may in fact result from a radical
chain reaction where the proton source is the intermediate XIII.
This mechanistic proposal provides an explanation for the variable
yields of 39 and 55 that are formed when a dioxane/H2O solution
of 54 is treated with dry HCl gas. The relative amounts of oxygen
and compound 54 present in the reaction mixture should determine
(34) Zhang, Y. J. Labelled Compd. Radiopharm. 2002, 45, 1055–1064.
(35) Manes, L. V.; Bakus, G. J.; Crews, P. Tetrahedron Lett. 1984, 25, 931–
934.
(36) (a) van Vliet, M. C. A.; Mandelli, D.; Arends, I. W. C. E.; Schuchardt,
U.; Sheldon, R. A. Green Chem. 2001, 3, 243–246. (b) Leffler, J. E.; Miller,
D. W. J. Am. Chem. Soc. 1977, 99, 480–483.
Scheme 15 presents a mechanistic rationalization for the forma-
tion of C-11 oxygenated species during the synthesis of desbro-
moceratamines. When intermediates of general structure XI bearing
(37) Abou-Jneid, R.; Ghoulami, S.; Martin, M.-T.; Huu Dau, E. T.; Travert,
N.; Al-Mourabit, A. Org. Lett. 2004, 6, 3933–3936.
1002 J. Org. Chem. Vol. 74, No. 3, 2009

Microtubule-Stabilizing Sponge Alkaloid Ceratamine A
SCHEME 15
whether pathway A to C-11 oxygenated products or pathway B to
nonoxygenated products will dominate.
previously reported,9 but it does not generate a large percentage
of cells arrested in mitosis compared with ceratamine A (1) and
its activity falls off at higher concentrations. Desbromocerata-
mine A (55) shows the most promising activity of the synthetic
analogues tested. At the highest concentration evaluated (50 µg/
mL), 55 shows significant mitotic arrest (∼20%), while 60,
which differs from desbromoceratamine A (55) simply by loss
of the methyl substituent on the 2-amino group, showed no
activity at the concentrations tested. The data in Figure 1
demonstrates that although the C-14 and C-16 bromine sub-
stituents in ceratamine A (1) are not essential for antimitotic
activity, they make significant contributions to maximum
potency (1 versus 55). Similarly, it is apparent that methylation
of the 2-amino functionality is required for effective antimitotic
activity (55 versus 60).
Biological Activities of Synthetic Ceratamine Analogues.
Synthetic analogues 22, 24, 26, 39, 40, 41, 42, 55, 56, and 60
as well as natural ceratamine A (1) were evaluated for their
ability to arrest cells in mitosis in the TG3 cell-based assay.4
Compound 22, which lacks the aminoimidazole ring in cerata-
mine A (1), showed evidence of weak mitotic arrest with a very
low percentage of cells arrested in mitosis at concentrations
above 40 µg/mL, and the dihydro analogue 24 and the partially
reduced ceratamine 26 were not active (data not shown). These
results indicated that the aminoimidazole ring is a required
element in the ceratamine antimitotic pharmacophore and
suggested that a fully planar (i.e., aromatic) imidazoazepine
heterocycle is also essential. Both conclusions were supported
by the observation that the analogues 41 and 42 lacking both
the 2-N-methylamino substituent and an aromatic imida-
zoazepine heterocycle were completely inactive.
Conclusion
The synthetic route shown in Scheme 16 has been developed
for the preparation of the antimitotic compound desbromocerata-
mine A (55). This work features the first synthesis of an aromatic
imidazo[4,5-d]azepine heterocycle, and it revealed an unexpected
reaction between atmospheric oxygen and an imidazo[4,5-d]azepine
precursor to the core of the ceratamines. The reaction with oxygen
only occurred in precursors having an amino substituent at C-2 of
the imidazole ring. A mechanistic rationalization (Scheme 15)
suggests that the amino substituent is required to make the
imidazole ring basic enough to undergo C-4 ipso protonation, which
triggers the reaction with atmospheric oxygen and aromatization.
The cornerstone of this first synthetic route to ceratamine analogues
is the use of an intramolecular Buchwald vinyl amidation reaction
to form the azepine ring, and the end game of the synthesis makes
the first use of Buchwald/Hartwig amination to prepare 2-ami-
noimidazoles from a 2-chloroimidazole precursor. The discovery
that the C-14 and C-16 bromine substituents in ceratamine A (1)
Figure 1 shows the TG3 assay data for ceratamine A (1), 39,
40, 56, 60, and desbromoceratamine A (55). The C-11 hydroxy
analogue 39 shows antimitotic activity at ∼10 µg/mL as
FIGURE 1. Antimitotic activity of natural ceratamine A (1) and the
synthetic analogues 39, 40, 56, 60, and desbromoceratamine A (55)
evaluated in a TG3 cell-based assay.4
J. Org. Chem. Vol. 74, No. 3, 2009 1003

Nodwell et al.
SCHEME 16
added, and the solution was stirred at -78 °C for 10 min and then
placed in a cold-water bath. This solution was stirred for 10 min
and then recooled to -78 °C. nBuLi (5.91 mL, 1.6 M in hexanes,
9.46 mmol) was slowly added, and the solution was stirred for 20
min. DMF (2.0 mL) was then slowly added, the solution stirred at
-78 °C for 10 min, and then the reaction was allowed to slowly
warm to room temperature. The reaction was stirred for 10 min at
room temperature and then quenched by addition of water. The
reaction mixture was extracted into EtOAc, and the organic phase
was dried over Na2SO4, filtered, and concentrated. The crude
product was purified by column chromatography (hexanes/EtOAc)
to yield stannane (44) (3.16 g, 5.85 mmol, 62%) as a colorless oil.
1H NMR (400 MHz, CD2Cl2) δ 0.89 (t, J ) 7.3 Hz, 9H), 1.24 (m,
5H), 1.33 (m, 7H), 1.53 (m, 6H), 4.54 (s, 2H), 5.43 (s, 2H), 7.35
(m, 5H), 9.81 (s, 1H); 13C NMR (100 MHz, CD2Cl2) δ 12.2, 14.4,
28.2, 29.9, 71.5, 76.6, 128.5, 129.0, 129.4, 137.5, 137.8, 144.8,
150.4, 187.5; ESIMS [M + Na]+ calcd for C24H37N2O235Cl116SnNa
559.1459, found 559.1447.
2-Bromo-3-(4-methoxy-phenyl)-N-methyl-acrylamide (45). Ester
(30) (10.95 g, 40.4 mmol) was dissolved in 30 mL of THF, 5 mL
of H2O, and 10 mL of MeOH. LiOH-H2O (3.40 g, 80.8 mmol)
was added, and the reaction was stirred vigorously at room
temperature for 2 h. The reaction was then acidified with 1 M HCl
and extracted into EtOAc. The organic layer was dried over Na2SO4,
filtered, and concentrated to dryness. The resulting solid was
redissolved in 70 mL of CH2Cl2. HOBt (8.2 g, 60.6 mmol) and
DMAP (∼10 mg) were added, and the solution was cooled to
0 °C. DIPC (9.38 mL, 60.6 mmol) was slowly added, and the
solution was stirred for 10 min at 0 °C. The reaction was warmed
to room temperature and stirred for an additional 1 h. The reaction
was then concentrated to dryness. The residue was then redissolved
contribute in a major way to the potency of the antimitotic activity
is noteworthy and was completely unanticipated. Efforts are
currently ongoing in our laboratory to use this important SAR
finding to guide the design and synthesis of new analogues that
capture the potency of the natural product. Evaluation of desbro-
moceratamine A (55) in mouse xenograft models of human cancer
are underway, and the results will be reported elsewhere.
Experimental Section
Additional details are available in Supporting Information.
1-Benzyloxymethyl-4,5-dibromo-2-chloro-1H-imidazole (43). Tri-
bromide 27 (6.97 g, 16.4 mmol) was dissolved in 60 mL of dry
THF. The solution was cooled to -78 °C, and nBuLi (10.3 mL,
1.6 M in hexanes, 16.4 mmol) was added slowly. The solution was
stirred cold for 20 min, and then a solution of C2Cl6 (4.27 g, 18.04
mmol) in 10 mL of dry THF was added over a period of 6 min.
The reaction was stirred at -78 °C for 10 min and then warmed to
room temperature for an additional 20 min. The reaction was
quenched with addition of H2O, and the product was extracted into
EtOAc. The organic layer was dried over Na2SO4, filtered, and
concentrated. An initial 4.64 g (12.1 mmol, 74%) of the chloride
could be recovered by crystallization from cold hexanes, and flash
chromatography (hexanes/EtOAc) of the remaining residue yielded
an additional 1.24 g (3.3 mmol, 20%) of product (94% yield
overall).
1-Benzyloxymethyl-2-chloro-5-tributylstannanyl-1H-imidazole-
4-carbaldehyde (44). Chloride 43 (3.60 g, 9.46 mmol) was dissolved
in 36 mL of dry THF and cooled to -78 °C. nBuLi (5.91 mL, 1.6
M in hexanes, 9.46 mmol) was slowly added, and the solution was
stirred cold for 20 min. Bu3SnCl (2.55 mL, 9.46 mmol) was then
1004 J. Org. Chem. Vol. 74, No. 3, 2009

Microtubule-Stabilizing Sponge Alkaloid Ceratamine A
1
in THF and treated with MeNH2 (50 mL, 2.0 M in THF, 100 mmol)
for 10 min. The reaction was then extracted into EtOAc and washed
with 2 × 1 M HCl and 2 × saturated NaHCO3. The organic layer
was dried over Na2SO4, filtered, and concentrated. The crude
product was purified by column chromatography (hexanes/EtOAc)
to yield 45 (10.3 g, 38.4 mmol, 95%) as a colorless oil that
crystallized upon standing. 1H NMR (400 MHz, CDCl3) δ 2.96 (d,
J ) 4.8 Hz, 3H), 3.85 (s, 3H), 6.83 (s, br, 1H), 6.94 (d, J ) 8.7
Hz, 2H), 7.80 (d, J ) 8.7 Hz, 2H), 8.27 (s, 1H); 13C NMR (100
MHz, CDCl3) δ 28.4, 56.3, 113.3, 114.8, 127.5, 132.9, 137.9, 161.7,
164.2; ESIMS [M + Na]+ calcd for C11H12NO279BrNa 291.9949,
found 291.9940.
into a solution of 48 in toluene. H NMR (300 MHz, CD2Cl2) δ
3.27 (s, 3H), 3.77 (s, 3H), 4.35 (s, 2H), 4.57 (d, J ) 11.5 Hz, 1H),
4.90 (d, J ) 11.5 Hz, 1H), 6.13 (d, J ) 9.0 Hz, 1H), 6.19 (d, J )
9.0 Hz, 1H), 6.77 (d, J ) 8.6 Hz, 2H), 6.94 (d, J ) 8.6 Hz, 2H),
7.28 (m, 6H); 13C NMR (75 MHz, CD2Cl2) δ 38.2, 55.4, 70.9, 73.2,
107.9, 114.4, 121.5, 123.5, 126.9, 127.9, 128.0, 128.4, 129.1, 130.5,
134.9, 136.1, 136.2, 136.9, 160.4, 169.3; ESIMS [M + Na]+ calcd
for C24H22N3O335ClNa 458.1247, found 458.1242.
2-Amino-3-benzyloxymethyl-4-(4-methoxy-benzylidene)-6-meth-
yl-4,6-dihydro-3H-imidazo[4,5-d]azepin-5-one (38) from Chloride
48. Chloride 48 (60 mg, 0.14 mmol), triphenylsilylamine (46.3 mg,
0.17 mmol), Pd2(dba)3 (13 mg, 0.014 mmol), and XPhos (CAS
564483-17-7; 16.2 mg, 0.034 mmol) were combined under Ar in
4.0 mL of dry toluene. LiHMDS (182 µL, 1.0 M in toluene, 0.182
mmol) was added, and the solution was heated to 100 °C for 1 h.
The reaction was cooled to room temperature and diluted with
EtOAc. Then, 1.0 M HCl was added, and the biphasic mixture was
stirred rapidly for 10 min. The aqueous layer was basified with
NaHCO3 and extracted into EtOAc. The organic layer was dried
over Na2SO4, filtered, and concentrated. The crude product was
purified by column chromatography (EtOAc/MeOH) to yield 38
(40 mg, 0.098 mmol, 69%) as a yellow solid. 1H NMR (600 MHz,
CD2Cl2) δ 3.26 (s, 3H), 3.76 (s, 3H), 4.19 (d, J ) 11.5 Hz, 1H),
4.23 (d, J ) 11.5 Hz, 1H), 4.45 (s, br, 2H), 4.52 (s, 2H), 6.03 (d,
J ) 8.9 Hz, 1H), 6.07 (d, J ) 8.9 Hz, 1H), 6.77 (d, J ) 8.9 Hz,
2H), 6.97 (d, J ) 8.9 Hz, 2H), 7.18 (m, 3H), 7.28 (m, 3H); 13C
NMR (150 MHz, CD2Cl2) δ 38.7, 55.8, 70.8, 72.6, 109.1, 114.6,
119.2, 122.5, 127.8, 128.1, 128.5, 128.7, 128.9, 130.9, 131.9, 134.0,
137.2, 152.3, 160.3, 170.0; ESIMS [M + Na]+ calcd for
C24H24N4O3Na 439.1746, found 439.1735.
2-(3-Benzyloxymethyl-2-chloro-5-formyl-3H-imidazol-4-yl)-3-(4-
methoxy-phenyl)-N-methyl-acrylamide (46). Stannane 44 (2.86 g,
5.30 mmol), amide 45 (1.58 g, 5.83 mmol), Pd(PPh3)4 (673 mg,
0.58 mmol), and CuI (544 mg, 2.86 mmol) were combined in 50
mL of dry THF and stirred at room temperature for 20 h. The
reaction was then concentrated to dryness, redissolved in CH2Cl2,
and filtered through Celite. The resulting solution was then
concentrated, and the crude product was purified by column
chromatography (hexanes/EtOAc) to yield 46 (1.70 g, 3.86 mmol,
1
73%) as a slightly yellow solid. H NMR (400 MHz, CD2Cl2) δ
2.82 (d, J ) 4.8 Hz, 3H), 3.77 (s, 3H), 4.51 (s, 2H), 5.08 (d, J )
10.7 Hz, 1H), 5.22 (d, J ) 10.7 Hz, 1H), 6.24 (s, br, 1H), 6.78 (d,
J ) 8.8 Hz, 2H), 6.98 (d, J ) 8.8 Hz, 2H), 7.18 (m, 2H), 7.81 (m,
2H), 8.07 (s, 1H), 9.67 (s, 1H); 13C NMR (100 MHz, CD2Cl2) δ
27.8, 56.3, 72.4, 74.5, 115.4, 118.3, 126.8, 128.6, 129.1, 129.4,
132.7, 137.2, 137.6, 137.9, 138.9, 144.8, 162.3, 166.1, 184.5;
ESIMS [M + Na]+ calcd for C23H22N3O435ClNa 462.1197, found
462.1182.
2-[3-Benzyloxymethyl-5-(2-bromo-vinyl)-2-chloro-3H-imidazol-
4-yl]-3-(4-methoxy-phenyl)-N-methyl-acrylamide (47). (Bromom-
ethyl)triphenylphosphonium bromide (1.85 g, 4.24 mmol) was
slurried in 50 mL of dry THF. Potassium tert-butoxide (476 mg,
4.24 mmol) was added, and the resulting yellow solution was stirred
at room temperature for 10 min. The ylide solution was then cooled
to -78 °C, and a 10 mL THF solution of aldehyde 46 (933 mg,
2.12 mmol) was added slowly. The reaction was stirred at -78 °C
for 30 min and then allowed to warm to room temperature for an
additional 20 min. The reaction was quenched with addition of
water, and the product was extracted with EtOAc and then with 2
× CH2Cl2. The combined organic phases were dried over Na2SO4,
filtered, and concentrated. The resulting crude residue was dissolved
in 20 mL of CH2Cl2 and allowed to sit for 10 min. The resulting
precipitate was filtered off and washed with 5 mL of CH2Cl2 to
yield 47 (0.5 g, 0.97 mmol, 46%). An additional 370 mg (0.72
mmol. 34%) could be isolated from the remaining residue by
column chromatography (hexanes/EtOAc) (80% yield overall). 1H
NMR (400 MHz, CD2Cl2) δ 2.74 (d, J ) 5.0 Hz, 3H), 3.78 (s,
3H), 4.46 (s, 2H), 5.00 (d, J ) 10.8 Hz, 1H), 5.16 (d, J ) 10.8 Hz,
1H), 5.72 (s, br, 1H), 6.31 (d, J ) 8.3 Hz, 1H), 6.73 (d, J ) 8.3
Hz, 1H), 6.80 (d, J ) 8.9 Hz, 2H), 7.04 (d, J ) 8.9 Hz, 2H), 7.16
(m, 2H), 7.80 (m, 3H), 8.07 (s, 1H); 13C NMR (100 MHz, CD2Cl2)
δ 27.6, 56.3, 72.0, 74.2, 107.1, 115.3, 118.5, 122.8, 127.1, 128.4,
128.9, 129.4, 129.7, 132.8, 135.5, 137.0, 137.4, 144.6, 162.2, 166.4;
ESIMS [M + Na]+ calcd for C24H23N3O335Cl79BrNa 538.0509,
found 538.0499.
3-Benzyloxymethyl-2-chloro-4-(4-methoxy-benzylidene)-6-meth-
yl-4,6-dihydro-3H-imidazo[4,5-d]azepin-5-one (48). Bromide 47
(682 mg, 1.31 mmol), CuI (50 mg, 0.262 mmol), and Cs2CO3 (853
mg, 2.62 mmol) were combined in 40 mL of dry THF. N,N′-
Dimethylethylenediamine (57 µL, 0.524 mmol) was then added,
and the solution was heated to 70 °C for 20 h. The reaction was
cooled to room temperature and concentrated to dryness. The
residue was redissolved in CH2Cl2 and filtered through Celite. The
resulting solution was concentrated to dryness again and purified
by column chromatography (hexanes/EtOAc) to yield 48 (486 mg,
1.11 mmol, 85%) as a pale yellow solid. A single crystal suitable
for X-ray diffraction analysis was grown by infusion of MeOH
N-[1-Benzyloxymethyl-8-(4-methoxy-benzylidene)-6-methyl-7-
oxo-1,6,7,8-tetrahydro-imidazo[4,5-d]azepin-2-yl]-N-methyl-forma-
mide (53). Chloride 48 (208 mg, 0.48 mmol), Cs2CO3 (234 mg,
0.72 mmol), Pd2(dba)3 (22 mg, 0.024 mmol), and 52 (58 mg, 0.12
mmol) were combined in 4.5 mL of dry toluene. N-Methylforma-
mide (43 mL, 0.72 mmol) was added, and the slurry was refluxed
for 24 h. The reaction was then cooled to room temperature and
concentrated to dryness. The residue was redissolved in CH2Cl2,
filtered through Celite, and concentrated to give an orange oil. The
crude product was purified by column chromatography (hexanes/
EtOAc) to yield 53 (159 mg, 0.34 mmol, 72%) as a slightly yellow
1
oil. H NMR (600 MHz, CD2Cl2) δ 3.17 (s, 3H), 3.29 (s, 3H),
3.76 (s, 3H), 4.28 (s, 2H), 4.53 (d, J ) 11.3 Hz, 1H), 4.59 (d, J )
11.3 Hz, 1H), 6.16 (d, J ) 9.1 Hz, 1H), 6.21 (d, J ) 9.1 Hz), 6.75
(d, J ) 8.3 Hz, 2H), 6.88 (d, J ) 8.3 Hz, 2H), 7.18 (m, 2H), 7.30
(m, 4H), 8.31 (s, 1H); 13C NMR (150 MHz, CD2Cl2) δ 31.9, 37.6,
54.8, 70.4, 72.1, 107.5, 113.6, 121.1, 121.8, 126.3, 127.4, 127.5,
127.9, 128.7, 129.9, 133.9, 136.0, 145.6, 159.9, 161.7, 168.9;
ESIMS [M + K]+ calcd for C26H26N4O439K 497.1591, found
497.1601.
4-(4-Methoxy-benzyl)-6-methyl-2-methylamino-6H-imidazo[4,5-
d]azepin-5-one or Desbromoceratamine A (55). Formamide 53 (90
mg, 0.19 mmol) was dissolved in 10 mL of CH2Cl2. Anhydrous
AlCl3 (25 mg, 1.9 mmol) was added, and the slurry was stirred
rapidly for 30 min. The solution was quenched with careful addition
of satd NaHCO3, and the crude reaction was extracted into EtOAc.
The organic phase was dried over Na2SO4, filtered, and concentrated
to yield 54. ESIMS [M + Na]+ calcd for C18H18N4O3Na 361.1277,
found 361.1275. The resulting product was used without further
purification. The crude 54 was then dissolved in 25 mL of 1,4-
dioxane and diluted with 25 mL of deionized water. Ar was then
bubbled through the solution for 30 min. The solution was cooled
to 0 °C, and anhydrous HCl(g) was bubbled through the solution
for 1 min. The solution was then allowed to sit for an additional 2
min. The pH of the solution was then brought to neutral with
addition of 5 M NaOH and then further basified with NaHCO3.
The reaction was extracted into EtOAc, and the bright yellow
organic phase was dried over Na2SO4 and concentrated. The crude
product was purified by flash chromatography (CH2Cl2/MeOH) to
J. Org. Chem. Vol. 74, No. 3, 2009 1005

Nodwell et al.
yield 55 (15 mg, 0.048 mmol, 25%, 2 steps) as a bright yellow
solid. Compound 56 was also isolated from the product mixture
(see below). The NMR showed a 3:1 mixture of isomers. Isomer 1
(major): 1H NMR (600 MHz, DMSO) δ 3.07 (s, 3H), 3.53 (s, 3H),
3.66 (s, 3H), 4.26 (s, 2H), 6.39 (d, J ) 10.0 Hz, 1H), 6.76 (d, J )
8.4 Hz, 2H), 7.28 (d, J ) 8.4 Hz, 2H), 7.70 (d, J ) 10.0 Hz, 1H),
8.51 (m, 1H); 13C NMR (150 MHz, DMSO) δ 29.2, 35.2, 43.7,
54.9, 100.3, 113.3, 123.8, 129.9, 132.7, 142.4, 157.4, 160.0, 164.1,
4-[Hydroxy-(4-methoxy-phenyl)-methyl]-6-methyl-2-methylamino-
6H-imidazo[4,5-d]azepin-5-one (56). Compound 58 (20 mg, 0.045
mmol) was dissolved in 1 mL of CH2Cl2. AlCl3 (60 mg, 0.45 mmol)
was added, and the resulting blood red slurry was stirred rapidly at
room temperature for 30 min. The reaction was quenched by careful
addition of saturated NaHCO3 and extracted into EtOAc. The
organic phase was dried over Na2SO4, filtered, and concentrated to
yield 59. ESIMS [M + Na]+ calcd for C17H16N4O3Na 347.1120,
found 347.1128. This product was used without further purification.
The crude compound 59 was then dissolved in 2.0 mL of dry THF.
BH3-THF (175 µL, 1.0 M in THF, 0.175 mmol) was added, and
the solution was stirred for 1 h at room temperature. The reaction
was then carefully quenched with the addition of 1.0 M HCl. The
solution was basified with the addition of saturated NaHCO3 and
extracted into EtOAc. The organic phase was dried over Na2SO4,
filtered, and concentrated. The crude product was purified by
column chromatography (CH2Cl2/MeOH) to yield 56 (5.0 mg, 0.015
mmol, 33%) as a bright yellow solid. The NMR showed a 3:1 ratio
1
169.7, 175.4. Isomer 2 (minor): H NMR (600 MHz, DMSO) δ
3.06 (s, 3H), 3.56 (s, 3H), 3.66 (s, 3H), 4.21 (s, 2H), 6.52 (d, J )
9.7 Hz, 1H), 6.75 (d, J ) 8.6 Hz, 2H), 7.22 (d, J ) 8.6 Hz, 2H),
7.83 (d, J ) 9.7 Hz, 1H), 8.59 (m, 1H); 13C NMR (150 MHz,
DMSO) δ 29.3, 35.4, 43.9, 54.9, 100.5, 113.3, 123.1, 129.8, 132.4,
143.2, 157.3, 160.6, 163.9, 170.0, 176.1; ESIMS [M + Na]+ calcd
for C17H18N4O2Na; 333.1327, found 333.1319.
8-Bromo-4-(4-methoxy-benzyl)-6-methyl-2-methylamino-6H-
imidazo[4,5-d]azepin-5-one (57). To a solution of 55 (1.8 mg, 0.006
mmol) in 0.5 mL of HOAc was added a solution of Br2 (0.0116
mmol, 1.0 M in HOAc, 11.6 µL). The solution was stirred at room
temperature for 1 h, and then the solvent was removed under
vacuum. The crude compound was purified by column chroma-
tography (CH2Cl2/MeOH) to yield 57 (1.0 mg, 0.0026 mmol, 43%)
as a bright yellow solid. The NMR showed a 3:4 mixture of
1
of rotamers. Only the major isomer is described. H NMR (600
MHz, DMSO) δ 3.06 (d, J ) 5.0 Hz, 3H), 3.55 (s, 3H), 3.69 (s,
3H), 6.26 (d, J ) 11.3 Hz, 1H), 6.48 (d, J ) 9.9 Hz, 1H), 6.80 (d,
J ) 8.6 Hz, 2H), 7.16 (s, br, 1H), 7.38 (d, J ) 8.6 Hz, 2H), 7.80
(d, J ) 9.9 Hz, 1H), 8.91 (d, J ) 5.0 Hz, 1H); 13C NMR (150
MHz, DMSO) δ 29.3, 43.7, 54.9, 72.6, 100.6, 113.2, 124.4, 127.3,
136.8, 143.6, 158.0, 163.8, 170.4, 174.5; ESIMS [M + Na]+ calcd
for C17H18N4O3Na 349.1277, found 349.1279.
1
rotamers. Isomer 1: H NMR (600 MHz, CD2Cl2) δ 3.25 (d, J )
5.3 Hz, 3H), 3.62 (s, 3H), 3.73 (s, 3H), 4.23 (s, 2H), 6.74 (d, J )
8.6 Hz, 2H), 7.36 (d, J ) 8.6 Hz, 2H), 7.92 (s, 1H); 13C NMR
(150 MHz, CD2Cl2) δ 30.1, 36.4, 45.1, 55.2, 97.4, 113.7, 126.4,
129.1, 130.8, 143.9, 158.5, 163.2, 164.0, 167.8, 174.9. Isomer 2:
1H NMR (600 MHz, CD2Cl2) δ 3.38 (s, 3H), 3.73 (s, 3H), 3.74 (s,
3H), 4.38 (s, 2H), 6.79 (d, J ) 8.9 Hz, 2H), 7.41 (d, J ) 8.9 Hz,
2H), 8.22 (s, 1H); 13C NMR (150 MHz, CD2Cl2) δ 30.5, 35.9, 46.3,
55.2, 99.1, 114.2, 126.8, 130.9, 148.1, 158.5, 162.8, 163.0, 163.8,
166.9, 167.6; ESIMS [M + Na]+ calcd for C17H17N4O279BrNa;
411.0433, found 411.0429.
2-Amino-4-(4-methoxy-benzyl)-6-methyl-6H-imidazo[4,5-d]azepin-
5-one (60). Compound 58 (25 mg, 0.056 mmol) was dissolved in
1 mL of CH2Cl2. AlCl3 (74 mg, 0.56 mmol) was added, and the
resulting blood red slurry was stirred rapidly at room temperature
for 30 min. The reaction was quenched by careful addition of
saturated NaHCO3 and extracted into EtOAc. The organic phase
was dried over Na2SO4, filtered, and concentrated to yield 59.
ESIMS [M + Na]+ calcd for C17H16N4O3Na 347.1120, found
347.1128. This product was used without further purification. Crude
59 was dissolved in 5 mL of 1,4-dioxane and diluted with 5 mL of
deionized water. HCl(g) was bubbled through the solution for 1 min,
and the solution was allowed to sit for an additional 2 min. The
pH of the solution was brought to neutral with the addition of 5 M
NaOH and then further basified by the addition of NaHCO3. The
reaction was extracted with EtOAc, and the aqueous phase was
saturated with NaCl and extracted again with EtOAc. The combined
organic phases were dried over Na2SO4, filtered, and concentrated.
The crude product was purified by column chromatography
(CH2Cl2/MeOH) to yield 60 (2.5 mg, 0.0084 mmol, 15%) as a bright
yellow solid. Trace amounts of 39 were also present in the reaction
mixture. 1H NMR (400 MHz, DMSO) δ 3.53 (s, 3H), 3.66 (s, 3H),
4.22 (s, 2H), 6.43 (d, J ) 9.8 Hz, 1H), 6.74 (d, J ) 8.7 Hz, 2H),
7.21 (d, J ) 8.7 Hz, 2H), 7.74 (d, J ) 9.8 Hz, 1H), 8.04 (s, br,
1H), 8.17 (s, br, 1H); 13C NMR (100 MHz, DMSO) δ 36.8, 45.2,
56.4, 101.8, 114.8, 125.0, 131.3, 133.9, 144.0, 158.8, 162.2, 165.4,
171.6, 177.7; ESIMS [M + H]+ calcd for C16H17N4O2 297.1352,
found 297.1354.
N-[1-Benzyloxymethyl-8-(4-methoxy-benzylidene)-6-methyl-7-
oxo-1,6,7,8-tetrahydro-imidazo[4,5-d]azepin-2-yl]-formamide (58).
To a slurry of sodium formate (215 mg, 3.16 mmol) in 3 mL of
dry THF was added acetyl chloride (150 mL, 2.1 mmol). The slurry
was then heated to 45 °C for 4 h. The slurry was cooled to room
temperature, and a solution of 38 (40 mg, 0.1 mmol) in 1 mL of
THF was added, followed by a catalytic amount of DMAP. The
solution was then stirred at room temperature for 24 h. Hydrochloric
acid (1 M) was added, and the solution was then stirred for 10
min. The solution was basified with the addition of saturated
NaHCO3, and the crude reaction mixture was extracted into EtOAc.
The organic phase was dried over Na2SO4, filtered, and concen-
trated. The crude product was purified by column chromatography
1
to yield 58 (36 mg, 0.08 mmol, 80%) as a light yellow solid. H
NMR (600 MHz, DMSO) δ 3.22 (s, 3H), 3.73 (s, 3H), 4.26 (s,
2H), 4.31 (d, J ) 11.5 Hz, 1H), 5.03 (d, J ) 11.5 Hz, 1H), 6.14
(d, J ) 9.1 Hz, 1H), 6.30 (d, J ) 9.1 Hz, 1H), 6.85 (d, J ) 8.0 Hz,
2H), 6.91 (d, J ) 8.0 Hz, 2H), 7.17 (m, 3H), 7.26 (m, 3H), 8.82 (s,
1H), 10.88 (s, 1H); 13C NMR (150 MHz, DMSO) δ 37.8, 55.2,
69.9, 71.2, 107.6, 114.3, 120.9, 126.6, 127.6, 127.7, 128.2, 129.0,
130.2, 132.0, 134.0, 136.9, 144.3, 159.6, 162.9, 168.6; ESIMS [M
+ Na]+ calcd for C25H24N4O4Na; 467.1695, found 467.1694.
Methylation of 58 To Yield 53. A solution of 58 (4.6 mg, 0.01
mmol), K2CO3 (2.7 mg, 0.02 mmol), and MeI (1.0 mL, 0.016 mmol)
in 0.5 mL of DMF was stirred at room temperature for 24 h. The
solution was extracted into EtOAc and washed with 3 × H2O. The
organic phase was dried over Na2SO4, filtered, and concentrated.
The crude product was purified by column chromatography
(hexanes/EtOAc) to yield 53 (4.0 mg, 0.009 mmol, 90%) as a dull
yellow solid.
Acknowledgment. Financial support was provided by NS-
ERC (R.J.A.), NCIC (R.J.A.), and CIHR (M.R.).
Supporting Information Available: Experimental synthetic
methods, 1H and 13C NMR spectra for all new synthetic
compounds, and details of the X-ray diffraction analysis for
compound 48. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO802322S
1006 J. Org. Chem. Vol. 74, No. 3, 2009