Synthesis by radical cyclization and cytotoxicity of highly potent bioreductive alicyclic ring fused [1,2-a]benzimidazolequinones

Article abstract of DOI:10.1002/chem.200601450

The key step in the synthesis of new five, six and seven-membered alicyclic ring [1,2-a]-fused bioreductive benzimidazolequinones was radical cyclisation. Six and seven-membered tributyltin hydride-mediated homolytic aromatic substitutions of nucleophilic

Full text of DOI:10.1002/chem.200601450

DOI: 10.1002/chem.200601450
Synthesis by Radical Cyclization and Cytotoxicity ofHighly Potent
Bioreductive Alicyclic Ring Fused [1,2-a]Benzimidazolequinones
Mary Lynch,[a] Sarah Hehir,[a] Paul Kavanagh,[a] Dónal Leech,[a] John OꢀShaughnessy,[a]
Michael P. Carty,*[b] and Fawaz Aldabbagh*[a]
Abstract: The key step in the synthesis
of newfive, six and seven-membered
alicyclic ring [1,2-a]-fused bioreductive
benzimidazolequinones was radical
cyclisation. Six and seven-membered
tributyltin hydride-mediated homolytic
aromatic substitutions of nucleophilic
N-alkyl radicals onto the benzimida-
zole-2-position occurred in high yields
(63–70%) when quaternising the pyri-
dine-like 3-N of imidazole with cam-
phorsulfonic acid and using large ex-
cesses of the azo-initiator, 1,1’-azobis-
(cyclohexanecarbonitrile), to supple-
ment the non-chain reaction. Elabora-
tion of benzimidazoles to the benzimi-
dazolequinones occurred in excellent
yields. The IC50 values for the cytotox-
icity of benzimidazolequinones towards
the human skin fibroblast cell line
GM00637 were in the nanomolar
range, as determined by using the MTT
assay. The benzimidazolequinones were
much more cytotoxic than indolequi-
4.4 times more cytotoxic than under
aerobic conditions, while mitomycin C
was 1.8 times more selective towards
hypoxia. The cyclopropane fused
pyrido
1a,2,3,9b-tetrahydro-1H-cyclopropa-
[3,4]pyrido[1,2-a]benzimidazole-5,8-
A
none
analogues.
1,2,3,4-
Tetrahydropyrido
R
A
ACHTREUNG
6,9-dione was the most potent com-
pound prepared being more than 300
times more cytotoxic than the clinically
used bioreductive drug, mitomycin C.
The latter benzimidazolequinone was
more potent under hypoxic conditions
(associated with solid tumors), being
dione was less cytotoxic and selective
than the five-membered ring analogue,
1,1a,8,8a-tetrahydrocyclopropa-
A
N
AHCTREUNG
lequinone by attaching methyl substitu-
ents onto the quinone moiety increased
reductive potentials and decreased cy-
totoxicity and selectivity towards hypo-
xia.
Keywords:
antitumor agents
·
benzimidazolequinones · radicals ·
synthetic methods
Introduction
bioreductive compounds.[2] The naturally occurring indole-
quinone, mitomycin C (1) is the archetypical clinically used
bioreductive antitumor agent. Mitomycin C owes its cyto-
toxicity to DNA-alkylation initiated by a single (SET) or
O2-deficient (anaerobic) hypoxic cells prevalent in solid
tumors may be selectively targeted by using bioreductive
compounds.[1] Quinone is the structural moiety commonly
associated with reductive activation, and is present in many
[a] M. Lynch, S. Hehir, P. Kavanagh, D. Leech, J. OꢀShaughnessy,
Dr. F. Aldabbagh
Department of Chemistry, National University of Ireland,
Galway (Ireland)
Fax : (+353)91-525700
E-mail: Fawaz.Aldabbagh@nuigalway.ie
[b] Dr. M. P. Carty
Department of Biochemistry, National University of Ireland,
Galway (Ireland)
Fax : (+353)91-495504
E-mail: Michael.Carty@nuigalway.ie
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
3218
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3218 – 3226

FULL PAPER
two-electron transfer, to give the semiquinone radical anion
or the hydroquinone, respectively. Although several en-
zymes can reduce 1, evidence indicates that the one-electron
reducing enzyme NADPH-cytochrome c reductase is in-
volved under hypoxic conditions. The reversibility of the
SET activation by oxygen has allowed researchers to use
bioreductive compounds including prodrugs to selectively
target hypoxia.[1–6] Alternatively, 1 can act as a substrate for
the obligate and non-reversible two-electron reduction
mediated by the enzyme DT-diaphorase.[2]
tumour activity exist,[9–11] but such compounds are excellent
substrates for DT-diaphorase and are not necessarily selec-
tive towards hypoxia.[9,10]
We nowreport the synthesis of novel benzimidazolequi-
nones 6–9 containing [1,2-a]-fused five-, six- and seven-
membered alicyclic rings, but minus the cyclopropane ring.
This allows us to assess the influence of the fused cyclopro-
pane and different sized [1,2-a] alicyclic rings on cytotoxici-
ty. The cytotoxicity of the dimethyl-substituted pyrrolo
A
a]benzimidazole-5,8-dione 7 and azepino[1,2-a]benzimida-
AHCTREUNG
In organic solvents, the reduction of 1, aziridinomitosene
(2) and synthetic analogues, such as cyclopropamitosene 3,
is known to proceed by the reversible SET process.[1–3] In
the case of 1, under aqueous conditions, both one and two-
electron reductions are thought to provide similar DNA al-
kylation pathways, and rapid disproportionation of the mito-
mycin semiquinone into the quinone and hydroquinone
under anaerobic conditions indicated that the hydroquinone
is the common intermediate.[7] Moreover many bioreductive
compounds are known to be toxic under oxygenated condi-
tions, due to either direct or indirect enzymatic reduction to
the hydroquinone or the formation of reactive oxygen spe-
zole-1,4-dione 9 will be presented in a subsequent paper.[12]
In the present paper, the effect of a five compared to a six-
membered fused [1,2-a] alicyclic ring on cytotoxicity is as-
sessed by using tetracycles 4 and 5, while the activity of the
cyclopropane ring is assessed by comparing the cytotoxicity
of 5 with 6. The influence of the quinone dimethyl substitu-
ents on reductive potentials and cytotoxicity is also evaluat-
ed by comparing the unsubstituted pyrido[1,2-a]benzimida-
U
zole-6,9-dione 6 with the 7,8-dimethyl-substituted analogue
8. Cytotoxicity is measured by using the MTT assay on
human skin fibroblast cells after treatment with benzimida-
zolequinones with selectivity deduced from assays under
aerobic and hypoxic conditions.
CÀ
cies, such as superoxide (O2 ), the product of the reaction
of the semiquinone radical anion with molecular oxygen.[1,2]
Moody and co-workers related selectivity of indolequi-
nones to redox potentials obtained by using cyclic voltam-
metry.[3b] Indolequinone 3 was shown to have a lower reduc-
tive potential in DMF (À1.395 V versus Ferrocene, Fc) than
1 (À1.421 V versus Fc) and greater cytotoxicity and selectiv-
ity towards hypoxia. Cyclopropamitosene 3 was reported to
be 34 times more cytotoxic under hypoxic than oxygenated
conditions, while 1 was only twice more cytotoxic under hy-
poxic conditions.[3] This is despite 3 not being capable of al-
kylating DNA at C-1, as the aziridine is replaced by a cyclo-
propane ring. The formation of the activated semiquinone
radical anion of 3 was speculated to induce radical ring
opening of the cyclopropane ring to give a highly reactive
radical capable of a cytotoxic effect through hydrogen ab-
straction from DNA leading to strand cleavage.[3,4] Chemical
evidence of cyclopropane ring opening was presented under
oxidative conditions for other cyclopropyl indolequinones,[4]
and Naylor et al.[5] showed that the 2-cyclopropyl substituent
in indolequinones was up to 2 orders of magnitude more ef-
fective towards hypoxia than the 2-isopropyl substituent.
More recently, we have reported the synthesis of new tet-
The key step in the synthesis of 6–9 is tributyltin hydride
(Bu3SnH)-mediated radical cyclisations of nucleophilic N-
alkyl radicals onto the benzimidazole 2-position activated by
quaternising the pyridine-like 3-N of imidazole with cam-
phorsulfonic acid (CSA) (Scheme 1). The addition of CSA
Scheme 1. Intramolecular homolytic aromatic substitution mechanism,
where R1 =OMe, R2 =H, n = 2 for synthesis of 6 and R1 =H, R2 =Me,
n = 1, 2, 3, respectively, for synthesis of 7–9 and AH=CSA.
to protonate the basic imidazole 3-N to facilitate intramolec-
ular 1,2-conjugate additions of alkyl radicals was first report-
ed by Gagosz and Zard for xanthate precursors by using a
peroxide initiator.[13] However, the radical cyclisations
herein are non-chain “oxidative” homolytic aromatic substi-
tutions carried out in the presence of the “reductant”
Bu3SnH.[14–16] The abstraction of the hydrogen atom or the
oxidation to regenerate aromaticity is nowunderstood to be
carried out by the azo-initiator and/or derived radical.[14,15]
Thus, such reactions should proceed in higher yields with
excess azo-initiator added, as the latter will provide radicals
for initiation as well as oxidative rearomatisation.
The following paper gives efficient preparations of a new
series of benzimidazolequinones containing [1,2-a]-fused ali-
cyclic rings (compounds 6–9), prepared by means of homo-
lytic aromatic substitutions onto the benzimidazole 2-posi-
tion. Structure activity/relationships are presented for benzi-
racyclic ring systems, pyrrolo and pyrido[1,2-a]benzimidazo-
A
lequinones 4 and 5 containing a fused cyclopropane ring
prepared by means of cycloaddition reactions induced by
thermolysis of Eschenmoser hydrazones.[8] Cyclopropa-
A
N
(À1.052 V versus Fc) and cyclopropa
C
ACHTREUNG
Chem. Eur. J. 2007, 13, 3218 – 3226
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3219

F. Aldabbagh, M. P. Carty et al.
midazolequinones 4–5, 6 and 8. This is part of our quest to
find more selective and potent antitumor agents.
compounds is avoided. Bowmanꢀs group have recently ex-
tended the ipso-substitution protocol to 2-(phenylsulfanyl)-
benzimidazole radical precursors attached to solid-phase
resins, and given an alternative synthesis of the precursors
from 2-chlorobenzimidazole and benzenethiols.[19]
Purification by using KF/silica-gel column chromatogra-
phy[21a] (later discussion) was found to be less successful for
purification of 12, and the Corey procedure of CsF/CsOH
(an anhydrous fluoride source) on silica gel was not attemp-
ted.[21b] We found that tin residues were efficiently eliminat-
ed by repeated addition of KF to the acidic imidazolium salt
solution with subsequent filtration of insoluble Bu3SnF.[22,23]
The basified free imidazole was purified by column chroma-
tography to give 12 in 70% yield with all traces of alkyltin
impurities eliminated, as indicated by analysis of 1H and
13C NMR spectra (Supporting Information), and supported
by clean NMR spectra and elemental analysis (within 0.2%
of the calculated C, H, N) for
Results and Discussion
Synthesis benzimidazolequinones 6–9: The synthesis of 6
began with the preparation of 4,7-dimethoxy-1H-benzimida-
zole (10) from the condensation of 3,6-dimethoxybenzene-
1,2-diamine with formic acid.[17] Benzimidazole 10 was alky-
lated by using less than one equivalent (0.75 equiv) of 1-
iodo-4-(phenylselanyl)butane and sodium hydride to avoid
dialkylation of the imidazole ring, giving radical precursor
11 in 71% yield (Scheme 2). Phenylselanides were used in
place of alkyl bromides and iodides due to their efficient
radical, but poor SN2 leaving group ability enabling facile
preparation of diazole radical precursors.[18,19]
target quinone 6. It is interest-
ing that the dimethoxy sub-
stituents seem not to impart
electronic conjugative dona-
tion effects onto the imida-
zole-2-position of 11 to de-
crease the efficiency of the
cyclisation, as no alkyl radical
reduction product was ob-
served from this reaction.
The
dimethoxybenzimida-
zole 12 was converted to qui-
none 6 in high yield (87%) by
Scheme 2. Preparation of target 6.
using hydrobromic acid-in-
duced demethylation to give
Radical cyclisation to give pyrido
A
the hydroquinone 13 in situ, which underwent facile oxida-
tion at room temperature with ferric chloride.[8] Quinones
6–9 were all found to be stable to repeated purification by
chromatography, and could be stored in a freezer without
any decomposition observed over periods of months.
The synthesis of benzimidazolequinones 7–9 began with
N-alkylation of commercially available 5,6-dimethylbenzimi-
dazole with 1-iodo-w-(phenylselanyl)alkanes to give radical
precursors 14–16 in 57–74% yield (Scheme 3). Radical cycli-
was carried out over 18 hours by using syringe-pump addi-
tion of Bu3SnH (2.5 equiv) and 1,1’-azobis(cyclohexanecar-
bonitrile) (ACN, 4.0 equiv) to a solution of 11 (1 equiv) and
CSA (1 equiv) heated under reflux in toluene. The addition
of further portions of ACN (1.5 equiv) during the reaction
was found to further improve yields of cyclisation products.
Furthermore, ACN was found to give higher product yields
than the azo-initiator AIBN (2,2’-azobis(isobutyronitrile))
used under the same reaction conditions, presumably be-
cause of its longer half-life allowing continual and slow gen-
eration of initiating radicals as required by the non-chain
1
1
mechanism of the reaction (ACN, t = ꢀ40 min, AIBN, t = <
2
2
2 min in toluene at 1108C[20]).
Despite the requirement for excess azo-initiator, the cycli-
sations described in this paper have several advantages over
the synthetic protocol described by Aldabbagh and Bowman
to access [1,2-a] alicyclic ring fused benzimidazoles by intra-
molecular N-(w-alkyl) radical ipso-substitutions of the 2-
phenylsulfanyl group.[18] Firstly, cyclisation yields are gener-
ally improved, secondly there is no requirement for the syn-
thetic steps used to prepare the intermediate 2-(phenylsulfa-
nyl)benzimidazole containing the necessary leaving group
for ipso-substitution and thirdly the use of malodorous thiol
Scheme 3. Preparation of radical precursors and cyclisations towards the
synthesis of targets 7–9.
3220
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3218 – 3226

A
FULL PAPER
sations using Bu3SnH and ACN were performed under iden-
tical conditions to the reaction that gave 12. The six-mem-
bered cyclisation gave the highest yield of cyclised product,
giving pyrido[1,2-a]benzimidazole 18a in 69% yield without
A
any radical reduction detected. The more difficult five- and
seven-membered cyclisations both gave radical reduction
products 17b and 19b in yields of 54 and 21%, as well as
cyclisation products 17a and 19a in yields of 32 and 63%,
respectively. Recovery yields of imidazole products were
good, and separation of cyclised from reduced products and
alkyltin residues was efficiently achieved. This was carried
out by using the direct column chromatography procedure
devised by Harrowven and Guy[21a] by purification of the
concentrated reaction mixture using gradient elution on
10% w/w of finely ground KF and 90% silica. The latter
procedure was an improvement on the cumbersome workup
procedure of repeated washing of the acidic imidazolium
salts with petrol to remove tin residues[18,19] and resulted in
improved overall recovery yields of diazoles (ꢀ70–85%).
Scheme 5. Preparation of targets 7–9.
exclusive product in 84% yield. Thus, we decided to use the
longer reaction time to nitrate 17a and 19a to exclusively
give dinitro compounds 22 and 23 in 90 and 63% yield, re-
spectively (Scheme 5). Hydrogenation in the presence of
Adamꢀs catalyst (PtO2) in ethanol at 40 psi gave the in situ
diamines that were readily oxidized by using a solution of
ferric chloride at room temperature to give the target qui-
nones 7–9 in high or quantitative yields.
The yield of azepino[1,2-a]benzimidazole 19a was particu-
A
larly impressive, as the equivalent seven-membered cyclisa-
tions by the ipso-substitution protocol only occurred in
yields of <17% for radical precursors without functionality
on the fused benzene ring.[18,19] The lowyield of the five-
membered cyclisation product 17a was consistent with simi-
lar literature radical reactions, and was presumably due to
strain associated with forming pyrrolo
C
Cytotoxicity ofbenzimidazolequinones 4, 5, 6 and 8 : By
using the MTT assay we have found that the treatment of
human skin fibroblast GM00637 cells with up to 0.005 mm of
cyclopropane fused compounds 4 and 5 under aerobic condi-
tions resulted in a similar effect (Figure 1). Under hypoxic
conditions, both 4 and 5 showed greater potency resulting in
hypoxic cytotoxicity ratios (HCR),[4] or efficiency towards
hypoxia, of 2.9 and 1.2, respectively (Table 1). However, 4
and 5 (as well as 6 and 8) were significantly more toxic re-
gardless of the environment than 1. The cytotoxicity of 1
was measured by using the same assay as for the benzimida-
zolequinones and the IC50 values for 1 (Table 1 and Support-
ing Information) are similar to those reported in the litera-
ture using Chinese hamster V79 cells, with an equivalent
The incorporation of methyl substituents on the fused
benzene part enabled conversion of cyclised compounds
17a–19a to quinones 7–9 by use of a nitration, reduction
and oxidation sequence.[24] Nitration was carried out by stir-
ring 17a–19a at 08C in a 50:50 mixture of concentrated sul-
furic and fuming nitric acid (Schemes 4 and 5). The first ni-
Scheme 4. Attempted nitrations of 18a.
tration carried out was on 18a, which resulted in a mixture
of the mono and dinitrated derivatives 20 and 21 in 15 and
44% yields, respectively (Scheme 4). This indicated that ni-
tration occurred first at the 6-position on the fused benzene
ring of 18a, as indicated by isolation of 20, and supported
by the 13C NMR spectra. Analogous selectivity in nitration
occurred by using 6-bromo-7-methyl-2,3-dihydro-1H-
Figure 1. Assessing the effect of the [1,2-a] alicyclic ring on cytotoxicity:
MTT assay of human skin fibroblast cells following treatment with
pyrrolo
[1,2-a]benzimidazole.[9] Although, 20 could probably
A
^
(
^
cyclopropapyrrolo
[1,2-a]benzimidazolequinone
4
,
)
and
be transformed to the quinone 8, we decided to extend the
reaction time from 3 to 12 h to enable a more efficient ap-
proach towards 8. This gave 6,9-dinitro analogue 21 as the
*
(
*
cyclopropapyrido[1,2-a]benzimidazolequinone
A
5
,
) under aerobic
(open symbols) and hypoxic (filled symbols) conditions for 24 h at 378C,
respectively.
Chem. Eur. J. 2007, 13, 3218 – 3226
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3221

F. Aldabbagh, M. P. Carty et al.
Table 1. Reductive potentials and cytotoxicity.[a]
Compound
Eredox [V]
versus Fc
IC50
(air)
IC50(N2)
HCR
]
[mmoldmÀ3
]
1
4
5
6
8
À1.421[3b]
À1.052[8]
À1.074[8]
À1.080
0.9
0.5
1.8
2.9
1.2
4.4
2.9
0.0069
0.0083
0.0070
0.0076
0.0024
0.0070
0.0016
0.0026
À1.168
[a] Reductive potentials were obtained by using cyclic voltammetry by
dissolving compounds in DMF, containing 0.1m tetrabutylammoniumper-
chlorate as electrolyte and 1 mm ferrocene (Fc) as reference. Eredox (Æ
0.010 V) calculated as (Epc +Epa)/2 from 100 mVsÀ1 cyclic voltammo-
grams. Epc =cathodic peak potential, Epa =anodic peak potential. Cyto-
toxicity was measured by using the MTT assay on human skin fibroblast
cells (GM00637). IC50 is the concentration required to reduce viability by
50% and it was obtained by drawing lines of best fit. Hypoxic cytotoxici-
ty ratios (HCR) were defined as the IC50 value for aerobic conditions di-
vided by the IC50 value for hypoxic conditions.
Figure 2. Assessing the effect of the fused cyclopropane ring on the cyto-
toxicity of 5: MTT assay of human skin fibroblast cells following treat-
*
,
*
) and
ment with cyclopropapyrido
A
&
pyrido[1,2-a]benzimidazolequinone 6 (&, ) under aerobic (open sym-
AHCTREUNG
bols) and hypoxic (filled symbols) conditions for 24 h at 378C, respective-
ly.
HCR of ꢀ2.[3] As cytotoxicity of mitomycin towards
GM00637 cells is similar to that reported with V79 cells, this
allows us to compare the cytotoxicity of the benzimidazole-
quinones used in the present study with published IC50
values for cyclopropamitosene 3 measured under the same
conditions as 1.[3] Indolequinone 3 has an impressive HCR
of 34, and the structurally related benzimidazolequinone 4 is
almost 700 and 90 times more potent under respective aero-
bic and hypoxic conditions.
The greater potency of 4 compared to 5 under hypoxic
conditions may be related to the greater strain in the pyrro-
Figure 3. Assessing the effect of the 7,8-dimethyl substituents on cytotox-
icity of 8: MTT assay of human skin fibroblast cells following treatment
lo- compared to the pyrido[1,2-a]benzimidazolequinone ring
A
system allowing more facile ring opening of the fused cyclo-
propane ring upon reductive activation. However, Moody
and co-workers observed an opposite trend in cytotoxicity
~
,
~
&
with pyrido
A
) under
aerobic (open symbols) and hypoxic (filled symbols) conditions for 24 h
at 378C, respectively.
when comparing the hydroxymethyl-substituted pyrrolo
A
a]indolequinone with the pyrido[1,2-a]indolequinone con-
AHCTREUNG
taining a fused cyclopropane ring with the latter showing
marginally greater cytotoxicity and HCR.[4] The hydroxy-
methyl-substituted cyclopropane fused indolequinones were
conversion of the prodrug quinone to the reductively acti-
vated biologically active forms. Compounds 6 and 8 exhibit-
ed very similar cytotoxicity under aerobic conditions, but 6
was ꢀ1.6 times more cytotoxic under hypoxic conditions
than 8. HCR values of 4.4 and 2.9 were obtained for 6 and
8, respectively (Table 1). In the latter cases, increasing the
reductive potential seems to slightly decrease cytotoxicity
and selectivity toward hypoxic conditions, which is in agree-
ment with general decreases in cytotoxicity with increasing
reductive potential observed with large sets of indolequino-
nes.[3b] As the most effective compound 6 lacks DNA dam-
aging functionality (such as a strained three-membered ring)
it is speculated that the cytotoxicity of the present benzimi-
dazolequinones may be largely due to reduction–reoxidation
cycles in the presence of O2 generating toxic oxygen species,
also less effective than 3 indicating an involvement of the
[4,6]
À
À
carbamate ( OCONH2 ) substituent in cytotoxicity.
To evaluate the role of the cyclopropane ring in the cyto-
toxicity of the pyrido
toxicity of 5 was compared with 6 (Figure 2). The pyrido-
[1,2-a]benzimidazolequinone lacking a cyclopropane ring 6
A
ACHTREUNG
was found to have a lower IC50 value than 5 under aerobic,
as well as hypoxic conditions (Table 1). Furthermore, 6 was
the most potent and selective of all the benzimidazolequi-
nones tested, being more than 120 and 310 times more cyto-
toxic under aerobic and hypoxic conditions, respectively,
than 1. Thus, it appears that the additional fused cyclopro-
pane ring is reducing the effectiveness of [1,2-a] alicyclic
ring-fused benzimidazolequinone 5 to act as a cytotoxin.
The role of the dimethyl substituents was assessed by
comparing the cytotoxicity of unsubstituted 6 with 7,8-di-
methyl-substituted quinone 8 (Figure 3). The inductive don-
ation of the methyl substituents was expected to increase re-
ductive potential (later discussion), and affect the rate of
CÀ
C [7,25]
such as H2O2, O2 and HO .
This may explain the low
HCR values for the benzimidazolequinones.
Cyclic voltammetry ofbenzimidazolequinones 6 and 8 : This
was carried out in DMF rather than in aqueous solution to
avoid the complexities of protonation.[3b,7] As only the one
and not the two-electron reduction can be reversed by
3222
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3218 – 3226

A
FULL PAPER
oxygen, we were only interested in the former as a guide to
selectivity toward hypoxia. The one-electron reductive po-
tentials are Eredox =À1.080 and À1.168 V versus Fc for the
functional groups. The higher cytotoxicity of the benzimida-
zolequinones compared to the indolequinones of compara-
ble structure may be related to their lower reductive poten-
tials. Future work will aim to establish precise enzymatic re-
duction and induction of DNA damage pathways, toxicity
toward cancer cell lines and the synthesis of further ana-
logues.
unsubstituted pyrido[1,2-a]benzimidazolequinone 6 and 7,8-
A
dimethyl analogue 8, respectively (Table 1). Thus, the induc-
tive effect of the dimethyl substituents increased the Eredox
of 8 by 88 mV (or 44 mV per methyl group). Compound 6
had a very similar reductive potential to pyrido[1,2-a]benz-
U
imidazolequinone 5 containing a fused cyclopropane ring.
The values of EpcÀEp/2 (ꢀ0.056 V) obtained at scan rates of
<100 mVsÀ1 for benzimidazolequinones 4, 5, 6 and 8 indi-
cated a reversible single-electron transfer with a stable semi-
quinone radical anion (Supporting Information). However,
this does not rule out the disproportionation of the latter to
the quinone and hydroquinone under aqueous conditions.[7]
Benzimidazolequinones are easier to reduce than the indole-
quinones due to the additional electronegative pyridine-like
nitrogen.
Experimental Section
Materials: All chemicals were purchased from Sigma–Aldrich, and all
solvents were purified and dried prior to use according to conventional
methods. Monitoring of reactions by TLC was performed by using alumi-
nium-backed plates coated with silica gel (Merck Kieselgel 60 F254).
Column chromatography by using silica gel was carried out with Merck
Kiesel 60 H silica. All reactions were conducted under a nitrogen atmos-
phere by using anhydrous solvents apart from those involving aqueous
solutions.
Instrumentation: Melting points were measured on a Stuart Scientific
melting point apparatus SMP3. Elemental analysis was carried out on a
Perkin–Elmer 2400 Series II analyser. IR spectra were acquired by using
a Perkin–Elmer Spec 1 with ATR attached.
Conclusion
A newseries of benzimidazolequinones containing [1,2- a]-
fused five, six and seven-membered alicyclic rings has been
prepared. The key synthetic step involves Bu3SnH-mediated
cyclisations of nucleophilic N-alkyl radicals onto the benzi-
midazole-2-position activated by quaternising the pyridine-
like 3-N of imidazole with camphorsulfonic acid. The substi-
All NMR spectra were recoded by using CDCl3 as solvent, and tetrame-
1
thylsilane as internal standard at room temperature. All H and 13C NMR
spectra were recorded at 400 and 100 MHz, respectively, by using a Jeol
GXFT 400 MHz instrument equipped with a DEC AXP 300 computer
1
work station. NMR assignments were supported by HMQC H–13C NMR
2D spectra and distortionless enhancement polarization transfer (DEPT).
Lowand high resolution electron impact (EI) and chemical ionisation
(CI) mass spectra were obtained on a Micro Mass GTC spectrometer.
The precision of this method for accurate mass measurement by manual
peak matching was better than 5 ppm. EPSRC National Mass Spectrome-
try Service (University of Swansea) carried out low resolution EI on the
Micromass Quattro II triple quadrupole instrument and high-resolution
mass spectrometry on the Finnigan MAT 900 XLT by manual peak
matching by using Electrospray (ESI) for compounds 6, 19a, 23, 7 and 8.
The precision of this method was better than 3 ppm. Elemental analysis
was carried out on a Perkin–Elmer 2400 Series II analyzer.
C
tuted aromatic product is formed upon loss of H , and is an
example of a formal oxidation in the presence of the reduc-
tant Bu3SnH. Optimised cyclisation yields required the addi-
tion of a large excess azo-initiator, 1,1’-azobis(cyclohexane-
carbonitrile) to maintain the non-chain reaction mechanism.
The six-membered cyclisations were regioselective giving
only the cyclised product in ꢀ70% yield. The five and
seven-membered cyclisations of 5,6-dimethyl N-[w-(phenyl-
selanyl)alkyl]benzimidazoles, respectively, gave 32 and 63%
yields of cyclised product with 54 and 21% Bu3SnH reduc-
tion products obtained. This newradical cyclisation protocol
compares favourably with literature radical cyclisations onto
the benzimidazole-2-position.
Absorbance was measured in the MTT assay by using a Wallac Victor2
1420 multilabel Counter (plate reader).
Syntheses:
[3,4]pyrrolo
The
preparation
of
1,1a,8,8a-tetrahydrocyclopropa-
A
ACHTREUNG
1H-cyclopropa
U
G
The IC50 values for benzimidazolequinone cytotoxicity to-
wards the human skin fibroblast cell line GM00637 were in
the nanomolar (10À9) range, as determined by using the
MTT assay. The most potent benzimidazolequinone tested
4,7-Dimethoxy-1-[4-(phenylseleno)butyl]-1H-benzimidazole (11): 4,7-Di-
methoxy-1H-benzimidazole (10) (1.39 g, 7.80 mmol) and sodium hydride
(0.25 g, 10.42 mmol) in THF (150 mL) were heated at reflux for 0.5 h. 1-
Iodo-4-(phenylselanyl)butane (2.00 g, 5.9 mmol) was added and heating
continued for an additional 2.5 h. The reaction was cooled, filtered
through Celite and evaporated to dryness. The resultant yellowresidue
was purified by column chromatography using silica gel as absorbent
with gradient elution of hexane and EtOAc to give 11 (1.63 g, 71%) as a
yellowoil. Rf =0.40 (EtOAc); 1H NMR: d=7.63 (s, 1H; Im-2-H), 7.43
(m, 2H; PhH), 7.22 (m, 3H; PhH), 6.54 (ABq, J=8.5 Hz, 2H; Ar-5-H,
Ar-6-H), 4.31 (t, J=7.0 Hz, 2H; NCH2), 3.97 (s, 3H; Me), 3.86 (s, 3H;
Me), 2.88 (t, J=7.1 Hz, 2H; CH2SePh), 1.96 (m, 2H; 2’-CH2), 1.69 ppm
(m, 2H; 3’-CH2); 13C NMR: d=146.0 (C), 142.0 (Im-2-CH), 141.6 (C),
136.0 (C), 132.6 (PhCH), 129.6 (C), 129.0 (PhCH), 126.8 (PhCH), 124.6
(C), 103.0 (ArCH), 101.6 (ArCH), 55.9 (Me), 55.5 (Me), 46.3 (NCH2),
31.5 (CH2SePh), 27.1 (2’-CH2), 26.8 ppm (3’-CH2); IR (neat): n˜ =963,
1060, 1093, 1153, 1183, 1226, 1260, 1373, 1490, 1517, 1580, 2840,
2926 cmÀ1; elemental analysis calcd (%) for C19H22N2O2Se: C 58.4, H 5.7,
N 7.2; found: C 58.5, H 5.7, N 7.1.
was the six-membered alicyclic ring-fused pyrido[1,2-a]ben-
A
zimidazole-6,9-dione 6, which was more than 100 times
more cytotoxic than the clinically used antitumour drug mi-
tomycin C (1) under aerobic conditions, and 300 times more
cytotoxic under hypoxic conditions. It was also more selec-
tive than 1, being 4.4 times more cytotoxic under hypoxic
than aerobic conditions (the cytotoxicity of 1 doubles under
hypoxic conditions). The fused cyclopropane ring of pyrido-
ACHTREUNG
A
8
AHCTREUNG
Chem. Eur. J. 2007, 13, 3218 – 3226
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3223

F. Aldabbagh, M. P. Carty et al.
6,9-Dimethoxy-1,2,3,4-tetrahydropyrido
Bu3SnH (1.15 g, 3.95 mmol) and ACN (1.54 g, 6.32 mmol) in PhMe
(55 mL) were added over 18 h by syringe pump to 11 (0.62 g,
A
(12):
1.70 ppm (m, 2H; 3’-CH2); 13C NMR: d=142.9 (C), 142.2 (Im-2-CH),
133.0 (PhCH), 132.1 (C), 131.0 (C), 130.0 (C), 129.2 (PhCH), 127.2
(PhCH), 120.5 (Ar-4-CH), 109.9 (Ar-7-CH), 42.5 (NCH2), 29.7 (2’-CH2),
27.3 (3’-CH2, CH2SePh), 20.7 (Me), 20.4 ppm (Me); IR (neat): n˜ =1023,
1215, 1326, 1436, 1473, 1493, 1570, 2933 cmÀ1; HRMS (CI): m/z: calcd for
C19H23N2Se: 359.1021; found: 359.1015 [M+H]+; HRMS (EI): m/z (%):
358 (20) [M]+, 201 (100), 200 (37), 159 (95), 91 (50), 78 (78).
1.58 mmol) and CSA (0.36 g, 1.58 mmol) in PhMe (95 mL) heated at
reflux. Further ACN (0.58 g, 2.37 mmol) in 10 mL of PhMe was divided
into three equal portions, and added after 1, 5 and 9 h to the reaction
mixture. The reaction was stirred at reflux for an additional 1 h, and
cooled. The diazole products were extracted with 2m HCl (75 mL), and
KF was added. The solution was gravity filtered, and further KF was
added until no more precipitate of Bu3SnF was formed. Na2CO3 was
added to the filtrate until it was pH 8. The free diazole was extracted
with CHCl3, dried (MgSO4) and evaporated to dryness and the residue
purified by column chromatography using silica gel as absorbent with
gradient elution of hexane and EtOAc to give 12 (0.26 g, 70%) as an
orange oil. Rf =0.23 (EtOAc); 1H NMR: d=6.49 (s, 2H; ArH), 4.42 (t,
J=5.9 Hz, 2H; 1-CH2), 3.94 (s, 3H; Me), 3.86 (s, 3H; Me), 3.06 (t, J=
6.4 Hz, 2H; 4-CH2), 2.04, 1.96 ppm (m, 4H; 2-CH2, 3-CH2); 13C NMR:
d=150.1 (4a-C), 145.4 (C), 141.8 (C), 134.5 (C), 125.3 (C), 102.2 (ArCH),
101.4 (ArCH), 55.7 (2”Me), 45.3 (1-CH2), 25.4 (all CH2), 23.0, 20.4 ppm;
IR (neat): n˜ =1091, 1157, 1193, 1260, 1311, 1402, 1424, 1449, 1523, 2837,
2932 cmÀ1; HRMS (EI): m/z: calcd for C13H16N2O2: 232.1212; found:
232.1215 [M]+; HRMS (EI): m/z (%): 232 (80) [M]+, 231 (100), 217 (97),
203 (25).
5,6-Dimethyl-1-[5-(phenylseleno)pentyl]-1H-benzimidazole (16): The
procedure given for the preparation of 14 was followed; 5,6-dimethyl-
1H-benzimidazole (1.67 g, 11.42 mmol), sodium hydride (0.41 g,
17.11 mmol) and 5-iodo-1-(phenylselanyl)pentane (3.03 g, 8.58 mmol) in
DMF (200 mL) gave 16 (2.37 g, 74%) as a yellowsolid after column
chromatography. Rf =0.27 (EtOAc); m.p. 67–698C; 1H NMR: d=7.72 (s,
1H; Im-2-H), 7.56 (s, 1H; Ar-4-H), 7.44 (m, 2H; PhH), 7.24 (m, 3H;
PhH), 7.12 (s, 1H; Ar-7-H), 4.08 (t, J=7.1 Hz, 2H; NCH2), 2.85 (t, J=
7.3 Hz, 2H; CH2SePh), 2.38 (d, J=2.9 Hz, 6H; Me), 1.84 (m, 2H; 2’-
CH2), 1.71, 1.43 ppm (m, 4H; 3’-CH2, (4’)-CH2); 13C NMR: d=142.7 (C),
142.3 (Im-2-CH), 132.7 (2”C), 132.1 (PhCH), 131.1 (C), 131.0 (C), 129.2
(PhCH), 127.0 (PhCH), 120.4 (Ar-4-CH), 109.9 (Ar-7-CH), 44.9 (NCH2),
29.7 (4’-CH2), 29.4 (2’-CH2), 27.5 (CH2SePh), 27.0 (3’-CH2), 20.8 (Me),
20.4 ppm (Me); IR (neat): n˜ =1020, 1223, 1326, 1446, 1493, 1580, 2860,
2926, 3086 cmÀ1
found: 371.1108 [M]+; HRMS (EI): m/z (%): 189 (94), 156 (39), 68 (40).
6,7-Dimethyl-2,3-dihydro-1H-pyrrolo[1,2-a]benzimidazole (17a): Bu3SnH
; HRMS (EI): m/z: calcd for C20H24N2Se: 372.1105;
1,2,3,4-Tetrahydropyrido
A
ACHTREUNG
(1.36 g, 4.66 mmol) and ACN (1.82 g, 7.46 mmol) in PhMe (60 mL) were
added over 18 h by a syringe pump to 14 (0.64 g, 1.86 mmol) and CSA
(0.43 g, 1.86 mmol) in PhMe (60 mL) heated at reflux. Further ACN
(0.68 g, 2.79 mmol) in 9 mL of PhMe was divided into three equal por-
tions and added after 1, 5 and 9 h to the reaction mixture. The solution
was stirred at reflux for an additional 1 h, cooled and evaporated to dry-
ness to give a brown residue, which was purified by column chromatogra-
phy with 10% w/w of finely ground KF and 90% w/w silica gel as absorb-
ent and gradient elution with hexane and EtOAc to give in order of elu-
tion 6,7-dimethyl-1-propyl-1H-benzimidazole 17b (0.19 g, 54%), as a
yellowsolid and 17a (0.11 g, 32%) as a yellowpowder.
12 (0.57 g, 2.45 mmol) in 48% hydrobromic acid (20 mL) was heated at
reflux for 3 h. The reaction was cooled and evaporated to dryness. FeCl3
aq (0.7m, 20 mL) was added to the residue and stirred at room tempera-
ture overnight. The solution was extracted with CHCl3, the combined or-
ganic extracts dried (MgSO4) and evaporated to dryness to give a red
solid, which was purified by column chromatography using silica gel as
absorbent with gradient elution of hexane and EtOAc to yield 6 (0.43 g,
87%) as a yellowsolid. Rf =0.26 (EtOAc); m.p. 153–1558C; 1H NMR:
d=6.60 (ABq, J=10.5 Hz, 2H; 7-H, 8-H), 4.33 (t, J=6.1 Hz, 2H; 1-CH2),
3.01 (t, J=6.3 Hz, 2H; 4-CH2), 2.05, 1.98 ppm (m, 4H; 3-CH2, 2-CH2);
13C NMR: d=181.3 (C=O), 178.3 (C=O), 151.9 (4a-C), 141.4 (C), 136.2
(ArCH), 136.0 (ArCH), 130.1 (C), 45.5 (1-CH2), 25.0 (all CH2), 22.3,
19.8 ppm; IR (neat): n˜ =965, 1035, 1058, 1074, 1157, 1205, 1244, 1302,
1354, 1376, 1409, 1429, 1458, 1483, 2919, 1510, 1590, 1650 cmÀ1 (C=O);
HRMS (ESI): m/z: calcd for C11H11N2O2: 203.0815; found: 203.0818
[M+H]+; elemental analysis calcd (%) for C11H10N2O2: C 65.1, H 5.0, N
13.8; found: C 65.0, H 4.9, N 13.8.
Product 17a: Rf =0.10 (EtOAc); m.p. 159–1618C (lit[26] m.p. 177–1798C);
1H NMR: d=7.42 (s, 1H; Ar-5-H), 7.04 (s, 1H; Ar-8-H), 4.00 (m, 2H; 1-
CH2), 2.99 (m, 2H; 3-CH2), 2.66 (m, 2H; 2-CH2), 2.34 ppm (s, 6H; Me);
13C NMR: d=160.5 (Ar-3a-C), 147.6 (C), 131.0 (C), 130.7 (C), 130.4 (C),
119.8 (Ar-5-CH), 109.9 (Ar-8-CH), 42.8 (1-CH2), 26.2 (2-CH2), 23.6 (3-
CH2), 20.5 (Me), 20.4 ppm (Me); IR (neat): n˜ =1000, 1130, 1293, 1300,
1446, 1523, 1410, 2940 cmÀ1
.
5,6-Dimethyl-1-[3-(phenylseleno)propyl]-1H-benzimidazole (14): 5,6-Di-
methyl-1H-benzimidazole (1.50 g, 10.26 mmol) and sodium hydride
(0.40 g, 16.67 mmol) in DMF (120 mL) were heated at 808C for 0.5 h. 1-
Iodo-3-phenylselanylpropane (2.69 g, 8.27 mmol) was added and heating
continued for an additional 2.5 h. The reaction was cooled, filtered
through Celite and evaporated to dryness. The resultant brown residue
was purified by column chromatography using silica gel as absorbent
with gradient elution of hexane and EtOAc to give 14 (1.93 g, 68%) as a
Product 17b: Rf =0.33 (EtOAc); m.p. 81–838C; 1H NMR: d=7.74 (s,
1H; Im-2-H), 7.55 (s, 1H; Ar-4-H), 7.14 (s, 1H; Ar-7-H), 4.06 (t, J=
6.9 Hz, 2H; NCH2), 2.37 (d, J=8.2 Hz, 6H; Me), 1.85 (m, 2H; 2’-CH2),
0.91 ppm (t, J=7.3 Hz, 3H; CH2CH3); 13C NMR: d=142.4, 142.3, 132.1
(C), 132.0 (C), 130.9 (C), 120.3 (Ar-4-CH), 109.9 (Ar-7-CH), 46.8
(NCH2), 23.2 (2’-CH2), 20.8 (Me), 20.5 (Me), 11.5 ppm (CH2CH3); IR
(neat): n˜ =986, 1055, 1137, 1215, 1351, 1374, 1449, 1486, 1500, 1676, 2876,
2926 cmÀ1; elemental analysis calcd (%) for C12H16N2: C 76.5, H 8.6, N
14.9; found: C 76.2, H 8.5, N 14.7.
1
yellowoil. Rf =0.23 (EtOAc); H NMR: d=7.71 (s, 1H; Im-2-H), 7.54 (s,
1H; Ar-4-H), 7.46 (m, 2H; PhH), 7.25 (m, 3H; PhH), 7.10 (s, 1H; Ar-7-
H), 4.24 (t, J=6.7 Hz, 2H; NCH2), 2.80 (t, J=6.8 Hz, 2H; CH2SePh),
2.36 (s, 6H; Me), 2.20 ppm (m, 2H; 2’-CH2); 13C NMR: d=142.7 (C),
142.4 (Im-2-CH), 133.2 (PhCH), 132.2 (C), 131.2 (C), 129.4 (PhCH),
129.1 (C), 127.5 (PhCH), 120.5 (Ar-4-CH), 109.9 (Ar-7-CH), 44.2
(NCH2), 29.6 (2’-CH2), 24.3 (CH2SePh), 20.7 (Me), 20.4 ppm (Me); IR
(neat): n˜ =903, 1020, 1216, 1326, 1366, 1436, 1493, 1576, 1740, 2206,
2926 cmÀ1; HRMS (EI): m/z: calcd for C18H20N2Se: 344.0792; found:
344.0796 [M]+; HRMS (EI): m/z (%): 344 [M]+, 147 (34), 190 (100).
7,8-Dimethyl-1,2,3,4-tetrahydropyrido[1,2-a]benzimidazole (18a): The
A
procedure given for the preparation of 17a was followed; Bu3SnH
(1.26 g, 4.33 mmol), ACN (2.33 g, 9.51 mmol) CSA (0.40 g, 1.73 mmol)
and 15 (0.62 g, 1.73 mmol) in PhMe (170 mL) gave 18a (0.24 g, 69%) as
an off-white powder after column chromatography. Rf =0.23 (EtOAc);
m.p. 161–1638C; 1H NMR: d=7.41 (s, 1H; Ar-6-H), 7.01 (s, 1H; Ar-9-
H), 3.97 (t, J=6.0 Hz, 2H; 1-CH2), 3.01 (t, J=6.3 Hz, 2H; 4-CH2), 2.34
(d, J=4.6 Hz, 6H; Me), 2.05 (m, 2H; 2-CH2), 1.96 ppm (m, 2H; 3-CH2);
13C NMR: d=150.9 (Ar-4a-C), 141.4 (C), 133.2 (C), 130.7 (C), 130.6 (C),
119.0 (Ar-6-CH), 109.2 (Ar-9-CH), 42.4 (1-CH2), 25.5 (4-CH2), 22.8 (3-
CH2), 20.9, 20.5, 20.4 ppm; IR (neat): n˜ =900, 1003, 1253, 1314, 1363,
1413, 1455, 1516, 1723, 2863, 2937 cmÀ1; HRMS (EI): m/z: calcd for
C13H16N2: 200.1388; found: 200.1106 [M]+; HRMS (EI): m/z (%): 200
[M]+ (100), 185 (99), 157 (40), 171 (41), 121 (9).
5,6-Dimethyl-1-[4-(phenylseleno)butyl]-1H-benzimidazole (15): The pro-
cedure given for the preparation of 14 was followed; 5,6-dimethyl-1H-
benzimidazole (1.46 g, 10.00 mmol), sodium hydride (0.36 g, 15.00 mmol)
and 4-iodo-1-(phenylselanyl)butane (2.56 g, 7.55 mmol) in DMF
(180 mL) gave 15 (1.54 g, 57%) as a yellowsolid after column chromo-
tography. Rf =0.28 (EtOAc); m.p. 74–768C; 1H NMR: d=7.70 (s, 1H;
Ar-2-H), 7.54 (s, 1H; Ar-4-H), 7.43 (m, 2H; PhH), 7.24 (m, 3H; PhH),
7.13 (s, 1H; Ar-7-H), 4.09 (t, J=7.1 Hz, 2H; NCH2), 2.88 (t, J=7.1 Hz,
2H; CH2SePh), 2.38 (d, J=2.9 Hz, 6H; Me), 1.99 (m, 2H; 2’-CH2),
2,3-Dimethyl-7,8,9,10-tetrahydro-6H-azepino[1,2-a]benzimidazole (19a):
T
The procedure given for the preparation of 17a was followed; Bu3SnH
(0.98 g, 3.38 mmol), ACN (1.81 g, 7.7 mmol), CSA (0.31 g, 1.35 mmol)
3224
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3218 – 3226

A
FULL PAPER
and 16 (0.50 g, 1.35 mmol) in PhMe (160 mL) gave in order of elution
after column chromatography 5,6-dimethyl-1-pentyl-1H-benzimidazole
(19b) (60 mg, 21%) as a yellowsolid and 19a (0.18 g, 63%) as a yellow
crystals
3.94 (m, 2H; 10-CH2), 3.14 (m, 2H; 6-CH2), 2.36 (d, J=14.1 Hz, 6H;
Me), 1.91 (m, 2H; 9-CH2), 1.81 ppm (m, 4H; 7-CH2, 8-CH2); 13C NMR:
d=161.8 (Ar-5a-C), 142.5 (C), 136.8 (C), 135.4 (C), 125.4 (C), 123.6 (C),
122.9 (C), 45.4 (10-CH2), 30.1 (9-CH2), 29.4 (6-CH2), 28.1, 25.0 (7-CH2, 8-
CH2), 15.0 (Me), 14.9 ppm (Me); IR (neat): n˜ =973, 1086, 1196, 1310,
1366 (NO2), 1526 (NO2), 1723, 2840, 2920 cmÀ1; HRMS (ESI): m/z: calcd
for C14H17N4O4: 305.1244; found: 305.1241 [M+H]+.
Product 19a: Rf =0.14 (EtOAc); m.p. 129–1318C; 1H NMR: d=7.44 (s,
1H; Ar-4-H), 7.03 (s, 1H; Ar-1-H), 4.10 (t, J=5 Hz, 2H; 10-CH2), 3.06
(t, J=5.6 Hz, 2H; 6-CH2), 2.37 (d, J=2.5 Hz, 6H; Me), 1.90 (m, 2H; 9-
CH2), 1.79 ppm (m, 4H; 7-CH2, 8-CH2); 13C NMR: d=156.7 (Ar-5a-C),
140.9 (C), 134.3 (C), 130.9 (C), 130.3 (C), 119.4 (Ar-4-CH), 109.1 (Ar-1-
CH), 44.5 (10-CH2), 31.0 (9-CH2), 30.2 (6-CH2), 28.9, 25.7 (7-CH2, 8-
CH2), 20.7 (Me), 20.3 ppm (Me); IR (neat): n˜ =910, 1003, 1183, 1313,
1416, 1463, 1516, 1616, 2933 cmÀ1; HRMS (CI): m/z: calcd for C14H19N2:
215.1543; found: 215.1540 [M+H]+; HRMS (ESI): m/z (%): 214 [M]+
(100), 199 (35), 185 (30), 159 (20), 77 (35).
Product 19b: Rf =0.27 (EtOAc); m.p. 74–758C; 1H NMR: d=7.76 (s,
1H; Im-2-H), 7.56 (s, 1H; Ar-4-H), 7.15 (s, 1H; Ar-7-H), 4.10 (t, J=
7.1 Hz, 2H; NCH2), 2.39 (d, J=9.2 Hz, 6H; Me), 1.86 (m, 2H; 2’-CH2),
1.33 (m, 4H, 3’-CH2, 4’-CH2), 0.88 ppm (t, J=7.0 Hz, 3H; CH2CH3);
13C NMR: d=142.5 (Im-2-CH), 142.4 (C), 132.3 (C), 131.8 (C), 130.8 (C),
120.3 (Ar-4-CH), 109.8 (Ar-7-CH), 45.0 (NCH2), 29.6, 28.9 (2’-CH2, 3’-
CH2), 22.2 (4’-CH2), 20.7 (Me), 20.3 (Me), 14.0 ppm (CH2CH3); IR
(neat): n˜ =910, 1003, 1183, 1313, 1416, 1463, 1516, 1616, 2933 cmÀ1; ele-
mental analysis calcd (%) for C14H20N2: C 77.7, H 9.3, N 13.0; found: C
77.8, H 9.5, N 12.9.
6,7-Dimethyl-2,3-dihydro-1H-pyrrolo[1,2-a]benzimidazole-5,8-dione (7):
N
Compound 22 (0.23 g, 0.83 mmol) and Adamꢀs catalyst (PtO2, 50 mg) in
ethanol (20 mL) was shaken under a 40 psi hydrogen atmosphere by
using a Parr apparatus for 18 h. The mixture was filtered through Celite
and evaporated to dryness. FeCl3 aq (0.7m, 30 mL) was added and stir-
ring continued for 12 h. Saturated NaOAc aq was added and the mixture
extracted with CHCl3. The combined organic extracts were dried
(MgSO4) and evaporated to dryness. The resultant orange solid was puri-
fied by column chromatography using silica gel as absorbent with a gradi-
ent elution of hexane and EtOAc to give 7 (0.18 g, 100%) as yellow
powder. Rf =0.32 (EtOAc); m.p. 235–2378C; 1H NMR: d=4.20 (t, J=
7.3 Hz, 2H; 1-CH2), 2.93 (t, J=7.8 Hz, 2H; 3-CH2), 2.71 (m, 2H; 2-CH2),
2.03 ppm (d, J=20 Hz, 6H; Me); 13C NMR: d=181.1 (C=O), 178.1 (C=
O), 160.4 (Im-3a-C), 145.9 (C), 141.0 (C), 138.7 (C), 129.3 (C), 45.0 (1-
CH2), 26.5 (3-CH2), 22.8 (2-CH2), 12.5 (Me), 11.9 ppm (Me); IR (neat):
n˜ =930, 1170, 1236, 1370, 1466, 1518, 1600, 1640 (C=O), 1663, 2920 cmÀ1
;
HRMS (ESI): m/z: calcd for C12H13N2O2: 217.0972; found: 217.0972
[M+H]+; HRMS (EI): m/z (%): 216 (100) [M]+, 187 (40), 159 (78), 53
(60); elemental analysis calcd (%) for C12H12N2: C 66.6, H 5.6, N 13.0;
found: C 66.2, H 5.3, N 13.1.
7,8-Dimethy-6-nitro-1,2,3,4-tetrahydropyrido
A
(20)
and 7,8-dimethyl-6,9-dinitro-1,2,3,4-tetrahydropyrido
AHCTREUNG
(21): Compound 18a (0.22 g, 1.10 mol) in a 50:50 mixture of concentrated
H2SO4 and fuming concentrated HNO3 (20 mL) was stirred at 08C for
3 h. The solution was basified with NH4OH to pH 12 and extracted with
CHCl3 (3”50 mL). The organic extracts were combined, dried (Na2SO4)
and evaporated to dryness. The resultant yellowsolid was purified by
column chromatography using silica gel as absorbent with gradient elu-
tion of hexane and EtOAc to give in order of elution 21 (0.14 g, 44%) as
a yellowcrystals and 20 (40 mg, 15%) as a yellowsolid.
7,8-Dimethyl-1,2,3,4-tetrahydropyrido[1,2-a]benzimidazole-6,9-dione (8):
G
The procedure given for the preparation of 7 was followed; 21 (0.18 g,
0.64 mmol) and Adamꢀs catalyst (PtO2, 50 mg) in ethanol (20 mL) was
shaken under 40 psi hydrogen for 18 h followed by FeCl3 oxidation to
give 8 (0.13 g, 85%) as yellowcrystals after column chromatography.
Rf =0.25 (ethyl acetate); m.p. 205–2078C; 1H NMR: d=4.32 (m, 2H; 1-
CH2), 2.99 (m, 2H; 4-CH2), 2.06 ppm (m, 10H; 2-CH2, 3-CH2, Me);
13C NMR: d=181.5 (C=O), 178.6 (C=O), 151.5 (Im-4a-C), 141.6 (C),
141.1 (C), 140.4 (C), 129.9 (C), 45.5 (1-CH2), 25.0 (4-CH2), 22.3 (2-CH2),
19.8 (3-CH2), 12.5 (Me), 12.1 ppm (Me); IR (neat): n˜ =915, 1037, 1199,
2925, 1229, 1265, 1307, 1416, 1579, 1644 cmÀ1 (C=O); HRMS (ESI): m/z:
calcd for C13H14N2O2: 230.1055; found: 230.1066 [M]+; elemental analysis
calcd (%) for C13H14N2O2: C 67.9, H 6.1, N 12.2; found: C 67.7, H 6.1, N
12.1.
Product 20: Rf =0.20 (EtOAc); m.p. 186–1888C; 1H NMR: d=7.22 (s,
1H; Ar-9-H), 4.05 (t, J=6.3 Hz, 2H; NCH2), 3.09 (t, J=6.3 Hz, 2H; 4-
CH2), 2.41 (s, 3H; Me), 2.34 (s, 3H; Me), 2.12 (m, 2H; 2-CH2), 2.01 ppm
(m, 2H; 3-CH2); 13C NMR: d=153.7 (Ar-4a-C), 141.2 (C), 134.4 (C),
134.1 (C), 131.6 (C), 123.2 (Ar-6-C), 112.5 (Ar-9-CH), 42.9 (1-CH2), 25.7
(4-CH2), 22.6 (2-CH2), 21.1 (3-CH2), 20.6 (8-Me), 14.9 ppm (7-Me); IR
(neat): n˜ =1027, 1165, 1276, 1309, 1370 (NO2), 1418, 1449, 1516 (NO2),
2922 cmÀ1
.
2,3-Dimethyl-7,8,9,10-tetrahydro-4H-azepino[1,2-a]benzimidazole-1,4-
H
Product 21: Rf =0.70 (EtOAc); m.p. 175–1768C; 1H NMR: d=3.95 (t, J
6.0 Hz, 2H; 1-CH2), 3.13 (t, J=6.3 Hz, 2H; 4-CH2), 2.36 (s, 3H; Me),
2.32 (s, 3H; Me), 2.08 (m, 2H; 2-CH2), 2.00 ppm (m, 2H; 3-CH2);
13C NMR: d=155.9 (Ar-4a-C), 142.0 (C), 136.9 (C), 136.0 (C), 125.1 (C),
123.5 (C), 123.4 (C), 43.7 (1-CH2), 26.0 (4-CH2), 22.4 (2-CH2), 19.7 (3-
CH2), 15.1 (Me), 14.9 ppm (Me); IR (neat): n˜ =1036, 1250, 1306, 1373
(6H)-dione (9): The procedure given for the preparation of 7 was fol-
lowed; 23 (0.31 g, 1.02 mmol) and Adamꢀs catalyst (PtO2, 50 mg) in etha-
nol (20 mL) were shaken under 40 psi hydrogen for 18 h followed by
FeCl3 oxidation to give 9 (0.25 g, 100%) as yellowcrystals, after column
chromatography; Rf =0.37 (EtOAc); m.p. 129–1318C; 1H NMR: d=4.56
(m, 2H; 10-CH2), 2.99 (t, J=5.7 Hz, 2H; 6-CH2), 2.03 (d, J=16 Hz, 6H;
Me), 1.88 (m, 2H; CH2), 1.79 (m, 2H; CH2), 1.70 ppm (m, 2H; 8-CH2);
13C NMR: d=181.4 (C=O), 179.2 (C=O), 157.9 (Im-5a-C), 140.7 (C),
140.2 (C), 139.9 (C), 130.3 (C), 45.9 (10-CH2), 30.9 (9-CH2), 29.3 (6-CH2),
28.2 (7-CH2), 25.0 (8-CH2), 12.5 (Me), 12.2 ppm (Me); IR (neat): n˜ =
1046, 1200, 1306, 1410, 1516, 1636, 2933 cmÀ1; HRMS (EI): m/z: calcd for
C14H16N2O2: 244.1212; found: 244.1206 [M]+; elemental analysis calcd
(%) for C14H16N2O2: C 68.8, H 6.6, N 11.5; found: C 68.8, H 6.6, N 11.5.
(NO2), 1410, 1506 (NO2), 2913 cmÀ1
; HRMS (EI): m/z: calcd for
C13H14N4O4: 290.1015; found: 290.1043 [M]+; HRMS (EI): m/z (%): 290
(26) [M]+, 284 (65), 226 (24), 199 (30), 189 (88), 121 (100).
6,7-Dimethyl-5,8-dinitro-2,3-dihydro-1H-pyrrolo[1,2-a]benzimidazole
A
(22): Compound 17a (0.18 g, 0.96 mmol) in a 50:50 mixture of concen-
trated H2SO4 and fuming concentrated HNO3 (20 mL) was stirred at 08C
for 12 h. The reaction was basified by using NH4OH to pH 12 and ex-
tracted with CHCl3. The organic extracts were combined, dried (Na2SO4)
and evaporated to dryness to give 22 (0.24 g, 90%) as an orange solid.
Biological studies
Cell culture: The SV40-transformed normal human skin fibroblast cell
line (GM00637I) was grown in minimum essential media eagle-earle BSS
(MEM) supplemented with 15% uninactivated fetal bovine serum
(FBS), penicillin/streptomycin, 2”essential and non-essential amino acids
and vitamins, with 2 mm l-glutamine. The cells were incubated at 378C
under a humidified atmosphere containing 5% CO2.
1
Rf =0.67 (EtOAc); m.p. 163–1658C; H NMR: d=4.13 (t, J=7.1 Hz, 2H;
1-CH2), 3.10 (t, J=7.1 Hz, 2H; 3-CH2), 2.75 (m, 2H; 2-CH2), 2.42 (s, 3H;
Me), 2.37 ppm (s, 3H; Me); 13C NMR: d=165.2 (Ar-3a-C), 140.9 (C),
125.6 (C), 125.2 (C), 123.2 (overlapping C), 45.4 (1-CH2), 26.1 (2-CH2),
23.4 (3-CH2), 15.7 (Me), 15.1 ppm (Me); IR (neat): n˜ =810, 1000, 1300
(NO2), 1523 (NO2), 2933 cmÀ1
.
Cytotoxicity measurement: MTT assay: Growth inhibition was deter-
mined by using the MTT colorimetric assay. Cells (GM00637I) were
plated in 96-well plates at a density of 50000 cells/well and allowed to
adhere over a period of 48 h. Benzimidazolequinone solutions (5 mL)
were applied in DMSO (2.5% concentration in well), and the plates
were incubated at 378C under a humidified atmosphere containing 5%
2,3-Dimethyl-1,4-dinitro-7,8,9,10-tetrahydro-6H-azepino[1,2-a]benzimida-
A
zole (23): The procedure given for the preparation of 22 was followed;
19a (0.10 g, 0.47 mmol) in a 50:50 mixture of concentrated H2SO4 and
fuming concentrated HNO3 (20 mL) gave after workup 23 (90 mg, 63%)
as a yellowcrystals. Rf =0.77 (EtOAc); m.p. 129–1318C; 1H NMR: d=
Chem. Eur. J. 2007, 13, 3218 – 3226
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3225

F. Aldabbagh, M. P. Carty et al.
CO2 for 24 h (aerobic conditions). MTT (100 mg) was added and the cells
were incubated for another 3 h. The supernatant was then removed care-
fully by pipetting. The resultant MTT formazan crystals were dissolved in
100 mL of DMSO and absorbance was determined on a plate reader at
550 nm with reference at 690 nm. Under hypoxic conditions 96-well
plates were incubated in a hypoxic chamber (0.5% O2, 5% CO2, 2% H2,
92.5% N2) at 378C under a humidified atmosphere for 3 h before addi-
tion of the benzimidazolequinone solutions. The plates were not removed
from the chamber until after incubation with MTT was complete. Each
data point given in figures represents the mean of at least three inde-
pendent experiments Æ one standard deviation. IC50, the drug concentra-
tion required to reduce viability to 50%, was obtained from plots of cell
viability (%) versus concentration, by drawing lines of best fit. Cell via-
bility is expressed as a percentage of the DMSO-only treated control
value.
[5] M. A. Naylor, M. Jaffar, J. Nolan, M. A. Stephens, S. Butler, K. B.
Patel, S. A. Everett, G. E. Adams, I. J. Stratford, J. Med. Chem.
1997, 40, 2335–2346.
[6] M. A. Naylor, E. Swann, S. A. Everett, M. Jaffar, J. Nolan, N. Rob-
ertson, S. D. Lockyer, K. B. Patel, M. F. Dennis, M. R. L. Stratford,
P. Wardman, G. E. Adams, C. J. Moody, I. J. Stratford, J. Med.
Chem. 1998, 41, 2720–2731.
[7] B. M. Hoey, J. Butler, A. J. Swallow, Biochemistry 1988, 27, 2608–
2614.
[8] a) J. OꢀShaughnessy, D. Cunningham, P. Kavanagh, D. Leech, P.
McArdle, F. Aldabbagh, Synlett 2004, 2382–2384; b) J. OꢀShaugh-
nessy, F. Aldabbagh, Synthesis 2005, 1069–1076.
[9] a) W. G. Schulz, I. Islam, E. B. Skibo, J. Med. Chem. 1995, 38, 109–
118; b) E. B. Skibo, I. Islam, W. G. Schulz, R. Zhou, L. Bess, R.
Boruah, Synlett 1996, 297–309; c) E. B. Skibo, S. Gordon, L. Bess,
R. Boruah, M. J. Heileman, J. Med. Chem. 1997, 40, 1327–1329;
d) A. Suleman, E. B. Skibo, J. Med. Chem. 2002, 45, 1211–1220.
[10] C. Flader, J. Liu, R. F. Borch, J. Med. Chem. 2000, 43, 3157–3167.
[11] a) I. Antonini, F. Claudi, G. Cristalli, P. Franchetti, M. Grifantini, S.
Martelli, J. Med. Chem. 1988, 31, 260–264; b) A. K. Singh, J. W.
Lown, Anti-Cancer Drug Des. 2000, 15, 265–275.
[12] S. Hehir, M. Lynch, L. OꢀDonovan, M. P. Carty, F. Aldabbagh, un-
published results.
[13] F. Gagosz, S. Z. Zard, Org. Lett. 2002, 4, 4345–4348.
[14] A. L. J. Beckwith, V. W. Bowry, W. R. Bowman, E. Mann, J. Parr,
J. M. D. Storey, Angew. Chem. 2004, 116, 97–100; Angew. Chem.
Int. Ed. 2004, 43, 95–98.
Electrochemistry
Cyclic voltammagrams were recorded at scan rates 20, 50, 100 and
200 mVsÀ1 by using an EcoChemie Autolab with PGSTAT12 potentiostat
controlled by GPES software. Quinones 6 and 8 were dissolved in DMF,
containing 0.1m tetrabutylammoniumperchlorate as electrolyte and 1 mm
ferrocene (Fc) as reference. Cyclic voltammetry was recorded at a plati-
num disk electrode in a single compartment electrochemical cell (2 mL
volume) containing Ag/AgCl reference electrode and a platinum wire
counter electrode. All measurements were carried out at RT. The plati-
num disc electrode was previously polished on a microcloth pad in
0.05 mm alumina slurry, and nitrogen was bubbled through solvents.
[15] P. T. F. McLoughlin, M. A. Clyne, F. Aldabbagh, Tetrahedron 2004,
60, 8065–8071.
[16] a) M. A. Clyne, F. Aldabbagh, Org. Biomol. Chem. 2006, 4, 268–
277; b) F. Aldabbagh, M. A. Clyne, Lett. Org. Chem. 2006, 3, 510–
513.
[17] L. Weinberger, A. R. Day, J. Org. Chem. 1959, 24, 1451–1454.
[18] a) F. Aldabbagh, W. R. Bowman, Tetrahedron Lett. 1997, 38, 3793–
3794; b) F. Aldabbagh, W. R. Bowman, Tetrahedron 1999, 55, 4109–
4122.
Acknowledgements
The authors would like to thank the National University of Ireland,
Galway and Kerry County Council for Postgraduate Fellowships for
Mary Lynch and the Irish Research Council for Science, Engineering &
Technology (IRCSET) for an Embark Scholarship Award to Sarah
Hehir. The authors would also like to thank National Centre for Biomed-
ical Engineering Science (NCBES) and the EPSRC National Mass Spec-
trometry Service, Swansea for mass spectra and Drs. Adrienne Gorman
and Afshin Samali (Biochemistry, NUI, Galway) for use of the hypoxic
chamber.
[19] S. M. Allin, W. R. Bowman, R. Karim, S. S. Rahman, Tetrahedron
2006, 62, 4306–4316.
[20] K. W. Dixon in Polymer Handbook (Eds.: J. Brandrup, E. H. Immer-
gut, E. A. E. Grulke), Wiley, NewYork, 4th ed., 1999, pp. 1–76.
[21] a) D. C. Harrowven, I. L. Guy, Chem. Commun. 2004, 1968–1969;
b) B. S. Edelson, B. M. Stoltz, E. J. Corey, Tetrahedron Lett. 1999, 40,
6729–6730.
[22] J. E. Leibner, J. Jacobus, J. Org. Chem. 1979, 44, 449–450.
[23] M. Lynch, PhD thesis, National University of Ireland, Galway, 2007.
[24] I. Antonini, G. Cristalli, P. Franchetti, M. Grifantini, S. Martelli,
Heterocycles 1982, 19, 2313–2317.
[25] a) J. W. Lown, H. -h. Chen, Can. J. Chem. 1981, 59, 390–395; b) L.
Durse, S. Rajagopalan, H. M. Eliot, J. M. Covey, B. K. Sinha, Cancer
Res. 1990, 50, 648–652.
[26] A. R. Freedman, D. S. Payne, A. R. Day, J. Heterocycl. Chem. 1966,
3, 257–259.
[1] a) R. H. J. Hargreaves, J. A. Hartley, J. Butler, Front. Biosci. 2000, 5,
e172–180; b) M. A. Naylor, P. Thomson, Mini-Rev. Med. Chem.
2001, 1, 17–29; c) M. Jaffar, N. Abou-Zeid, L. Bai, I. Mrema, I.
Robinson, R. Tanner, I. J. Stratford, Curr. Drug Delivery 2004, 1,
345–350.
[2] S. E. Wolkenberg, D. L. Boger, Chem. Rev. 2002, 102, 2477–2495.
[3] a) C. J. Moody, N. OꢀSullivan, I. J. Stratford, M. A. Stephens, P.
Workman, S. M. Bailey, A. Lewis, Anti-Cancer Drugs 1994, 5, 367–
372; b) A. S. Cotterill, C. J. Moody, R. J. Mortimer, C. L. Norton, N.
OꢀSullivan, M. A. Stephens, N. R. Stradiotto, E. Swann, I. J. Strat-
ford, J. Med. Chem. 1994, 37, 3834–3843.
[4] C. J. Moody, C. L. Norton, A. M. Z. Slawin, S. Taylor, Anti-Cancer
Drug Des. 1998, 13, 611–634.
Received: October 11, 2006
Published online: January 2, 2007
3226
www.chemeurj.org
ꢁ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 3218 – 3226