COMPARE analysis of the toxicity of an iminoquinone derivative of the imidazo[5,4-f]benzimidazoles with NAD(P)H:quinone oxidoreductase 1 (NQO1) activity and computational docking of quinones as NQO1 substrates

Article abstract of DOI:10.1016/j.bmc.2012.03.063

Synthesis and cytotoxicity of imidazo[5,4-f]benzimidazolequinones and iminoquinone derivatives is described, enabling structure-activity relationships to be obtained. The most promising compound (an iminoquinone derivative) has undergone National Cancer Institute (NCI) 60 cell line (single and five dose) screening, and using the NCI COMPARE program, has shown correlation to NQO1 activity and to other NQO1 substrates. Common structural features suggest that the iminoquinone moiety is significant with regard to NQO1 specificity. Computational docking into the active site of NQO1 was performed, and the first comprehensive mitomycin C (MMC)-NQO1 docking study is presented. Small distances for hydride reduction and high binding affinities are characteristic of MMC and of iminoquinones showing correlations with NQO1 via COMPARE analysis. Docking also indicated that the presence of a substituent capable of hydrogen bonding to the His194 residue is important in influencing the orientation of the substrate in the NQO1 active site, leading to more efficient reduction.

Full text of DOI:10.1016/j.bmc.2012.03.063

Bioorganic & Medicinal Chemistry 20 (2012) 3223–3232
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
COMPARE analysis of the toxicity of an iminoquinone derivative of the
imidazo[5,4-f]benzimidazoles with NAD(P)H:quinone oxidoreductase 1
(NQO1) activity and computational docking of quinones as NQO1 substrates
Vincent Fagan a, Sarah Bonham a, Michael P. Carty b, Patricia Saenz-Méndez c, Leif A. Eriksson d,
Fawaz Aldabbagh a,
⇑
a School of Chemistry, National University of Ireland Galway, University Road, Galway, Ireland
b Centre of Chromosome Biology, Biochemistry, School of Natural Sciences, National University of Ireland Galway, University Road, Galway, Ireland
c
9
i
Computational Chemistry and Biology Group, Facultad de Qu
mica, UdelaR, Montevideo 11800, Uruguay
d Department of Chemistry and Molecular Biology, University of Gothenburg, Göteborg 412 96, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
Synthesis and cytotoxicity of imidazo[5,4-f]benzimidazolequinones and iminoquinone derivatives is
described, enabling structure–activity relationships to be obtained. The most promising compound (an
iminoquinone derivative) has undergone National Cancer Institute (NCI) 60 cell line (single and five dose)
screening, and using the NCI COMPARE program, has shown correlation to NQO1 activity and to other
NQO1 substrates. Common structural features suggest that the iminoquinone moiety is significant with
regard to NQO1 specificity. Computational docking into the active site of NQO1 was performed, and the
first comprehensive mitomycin C (MMC)-NQO1 docking study is presented. Small distances for hydride
reduction and high binding affinities are characteristic of MMC and of iminoquinones showing correla-
tions with NQO1 via COMPARE analysis. Docking also indicated that the presence of a substituent capable
of hydrogen bonding to the His194 residue is important in influencing the orientation of the substrate in
the NQO1 active site, leading to more efficient reduction.
Received 26 February 2012
Revised 26 March 2012
Accepted 26 March 2012
Available online 3 April 2012
Keywords:
Anti-tumor
Bioreductive
DT-diaphorase
Heterocyclic compound
Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction
Imidazobenzimidazolequinones exist in two heterocyclic
arrangements: imidazo[4,5-f]benzimidazolequinones and imidazo
NAD(P)H:quinone oxidoreductase 1 (NQO1 and also known as
DT-diaphorase) is an enzyme which possesses a flavin prosthetic
group, and uses NADH or NADPH coenzymes with equal effi-
ciency.1,2 NQO1 is responsible for obligatory two-electron reduc-
tion of quinones to hydroquinones, as well as the reduction of
many other substrates including iminoquinones, nitro and azo-
compounds.3 It is often described as a detoxification enzyme since
it can override one-electron reductions that generate damaging
oxygen radicals by redox recycling,4 and has been linked with
superoxide scavenging (O2Åꢀ).5 Conversely, NQO1 bioreduction
can lead to activation of prodrugs, and can induce a cytotoxic or
cytostatic effect. Although NQO1 is expressed in normal tissues,
elevated levels are known to occur in many human tumor cell
lines.6–8 The cytotoxicity of well-known anti-tumor agents, mito-
mycin C (MMC) and EO9, show correlation to the levels of NQO1
in the cell lines used at the National Cancer Institute (NCI-60 tumor
cell line panel).6
[5,4-f]benzimidazolequinones. Examples of the former were first
reported by Schulz and Skibo. Dipyrrolo ring-fused compound 1a
was shown to be an excellent substrate for rat liver NQO1, with
high specificity towards melanoma cell lines (Fig. 1).9 In a subse-
quent article, molecular modeling studies showed that dipyrido
ring-fused analog 1b was capable of forming stable substrate com-
plexes with human NQO1.10 More recently, Fagan et al., reported
the synthesis and cytotoxicity evaluation of imidazo[5,4-f]ben-
zimidazolequinones, containing alicyclic and [1,4]oxazino fused
rings (Fig. 2).11,12 Cytotoxicity was assessed using two human can-
cer cell lines reported to possess high NQO1 activity; cervical
(HeLa),7 and prostate (DU-145)6 cancer cells, and a human normal
fibroblast cell line (GM00637).
The dipyridoimidazo[5,4-f]benzimidazolequinone 2b was mar-
ginally more potent towards the two human cancer cell lines than
its isomeric [4,5-f] analog 1b (Table 1). Iminoquinone 2a was the
most potent imidazobenzimidazole evaluated towards the prostate
cancer DU-145 cell line, being about 12 times more toxic towards
this cell line than towards the normal cell line.11 The DU-145 cell
line is one of the 60 cell lines in the Developmental Therapeutics
Program (DTP) at the National Cancer Institute (NCI).13 One quinone
⇑
Corresponding author. Tel.: +353 (0) 91 493120; fax: +353 (0) 91 495576.
E-mail address: fawaz.aldabbagh@nuigalway.ie (F. Aldabbagh).
0968-0896/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmc.2012.03.063

3224
V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
O
O
NO2
n
n
N
N
N
N
N
N
N
N
N
N
N
H2SO4, HNO3
°
80 C, 3 h
N
n
n
n
n
6
: n = 1 (85%)
7: n = 3 (92%)
1a: n = 1
O
1b: n = 2
(i), (ii), (iii)
N
N
N
N
6
Figure 1. Imidazo[4,5-f]benzimidazolequinones.
O
Z
O
3a
(76%)
m
N
N
N
N
N
N
N
N
O
NH
N
N
N
N
N
N
N
(iii)
(i), (ii)
n
3c
+
7
O
O
N
O
2a:
3a:
3b:
3c:
Z = NH
m, n = 1
O
2b:
Z = O
m = 1; n = 2
m, n = 3
3c (57%)
8
(26%)
O
N
O
Scheme 1. Synthesis of imidazo[5,4-f]benzimidazolequinones 3a and 3c. Reagents
and conditioins: (i) H2 (6 bar), Pd-C, MeOH, rt, 16 h; (ii) Frémy’s salt (potassium
nitrosodisulfonate (K2NO(SO3)2)), KH2PO4, rt, 1 h; (iii) 2 M HCl, rt, 15 min.
N
N
N
N
N
N
N
O
O
4
5
intermediates and their corresponding quinones. Under the same
conditions iminoquinone 2a was isolated in 96% yield from the
equivalent Frémy oxidation of the dipyridoimidazo[5,4-f]benzimid-
azole analog.11 In the case of the dipyrrolo analog, isolation of the
iminoquinone proved difficult, thus the mixture was hydrolyzed
using HCl (2 M) to give quinone 3a in 76% yield. The diazepino-fused
iminoquinone 8 was separated by chromatography in 26% yield, and
the remaining mixture hydrolyzed to give quinone 3c in 57% yield.
Unsymmetrical imidazo[5,4-f]benzimidazolequinone 4, contain-
ing one fused pyrido ring, was accessed according to Scheme 2.
Imidazo[5,4-f(4,5-f)]benzimidazole 9 was alkylated with 1.1 equiv
of 1-chlorobutane to give the mono-alkylated compound 10 in 59%
yield. Substitution of 1-butyl-5(7)H-imidazo[5,4-f(4,5-f)]benzimid-
azole 10 with 1-chloro-4-(phenylselenyl)butane gave approxi-
mately equal amounts of isomeric phenylselenides 11a and 11b in
83% combined yield. Following separation, 1,5-N-disubstituted phe-
nylselenide 11a was cyclized using a radical substitution protocol.11
The optimized yield of homolytic aromatic substitution product 12
(87%) was obtained by slow addition of 2.5 equiv of tributyltin hy-
dride (Bu3SnH) and 2,20-azobis(cyclohexanecarbonitrile) (ACCN),
while exposing the reaction mixture to air for part of the reaction
time. The slow addition of the initiator prevents radical reduction
by Bu3SnH, while exposure to air results in oxygenated radicals
which are thought to be involved in the final oxidative re-aromatiza-
tion step (hydrogen atom abstraction).14,15 The nitration of 12 gave a
mixture of nitro isomers which were difficult to separate. Therefore,
the mixture was hydrogenated and oxidized to give quinone 4 in 88%
yield.
Figure 2. Reported and new imidazo[5,4-f]benzimidazolequinones.
and two iminoquinones, including compound 2a, were submitted
for this more detailed in vitro cytotoxicity analysis by the DTP
NCI-60 screen, and we now present this data along with the results
of COMPARE analysis13 revealing interesting structure–activity
relationships (SARs) with regard to specificity towards NQO1. We
also establish SARs for the toxicity of fused alicyclic and nonfused
compounds, including the synthesis and evaluation of novel qui-
nones 3a, 3c, 4 and 5. The first comprehensive computational dock-
ing study of the clinical drug mitomycin C (MMC) into the human
NQO1 active site is provided, along with that of the imidazobenzim-
idazoles, in order to determine the substrate structural require-
ments for efficient NQO1 reduction, and to assess the relationship
between cytotoxicity and protein–ligand interactions.
2. Results and discussion
2.1. Synthesis
Dipyrrolo- and diazepino-fused quinones 3a and 3c were
prepared from previously reported dipyrrolo- and diazepino-imi-
dazo[5,4-f]benzimidazoles, respectively,11 according to Scheme 1.
Nitro compounds 6 and 7 were isolated in high yields (85% and
92%) by treatment with concentrated nitric and sulfuric acid.
Nitro compounds 6 and 7 were catalytically hydrogenated, and
Frémy oxidation of the resultant amine intermediates performed
under acidic conditions (pH 4), gave mixtures of iminoquinone
1,5-Dibutylimidazo[5,4-f]benzimidazolequinone
5 was pre-
pared in four steps according to Scheme 3. Imidazo[5,4-f(4,5-
f)benzimidazole 9 was dialkylated with 1-chlorobutane (2.2 equiv)
giving a separable ꢁ1:1 mixture of 1,5-dibutylimidazo[5,4-f]benz-
imidazole 13a and 1,7-dibutylimidazo[4,5-f]benzimidazole 13b in
88% overall yield. Nitration of 13a gave 4-nitroimidazo[5,4-f]benz-
imidazole 14 in 90% yield. This was hydrogenated and oxidized
with Frémy’s salt to give quinone 5 in 92% yield. No iminoquinone
was observed from the synthesis of quinones 4 or 5.
Table 1
Cytotoxicity evaluation using the MTT assay:IC50 values (l
M)a.
Entry
Compd
Cell Lines
GM00637
HeLa CCL-2
DU-145
1
2
3
4
5
6
7
8
MMC
1b11
2a11
2b11
3b12
3c
0.46 0.09
>5.0
3.63 0.52
>5.0
1.27 0.05
>5.0
2.12 0.04
1.05 0.15
0.27 0.16
3.33 0.47
1.55 0.35
1.67 0.05
0.95 0.10
1.78 0.19
1.28 0.05
0.44 0.04
0.17 0.02
1.73 0.23
0.30 0.01
1.99 0.39
0.79 0.05
4.46 0.16
1.35 0.11
1.34 0.06
2.2. Cytotoxicity against normal and cancer cell lines using the
MTT assay
4
5
a
IC50 = concentration of compound required for reduction of mean cell viability
to 50% of the control value after incubation for 72 h at 37 °C.
The cytotoxicity of new imidazo[5,4-f]benzimidazolequinones 3c,
4 and 5 was evaluated against human cancer cell lines: cervical (HeLa

V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
3225
H
N
N
H
N
N
N
N
H
N
N
N
SePh
N
N
N
N
N
NaH, Cl(CH2)4SePh
NaH, Cl(CH2)3CH3
SePh
+
DMF, 120 °C, 1 h
DMF, 120 °C, 1 h
N
N
N
N
11a
(43%)
10 (59%)
11b
(40%)
N
H
N
H
9
N
N
N
N
Bu3SnH, ACCN, (2.2 mLh-1)
PhMe, reflux, air
(i), (ii), (iii)
11a
4
(88%)
12 (87%)
Scheme 2. Synthesis of 3-butylimidazo[5,4-f]pyrido[1,2-a]benzimidazolquinone 4. Reagents and conditioins: (i) HNO3, H2SO4, 80 °C, 3 h; (ii) H2 (6 bar), Pd-C, MeOH, rt, 16 h;
(iii) Frémy’s salt, KH2PO4, rt, 1 h.
iminoquinone 8, were submitted for this more detailed cytotoxic-
ity evaluation.
The DTP NCI-60 screen is a two-stage process whereby 60 hu-
man tumor cell lines, representing nine major histological tissue
N
N
N
N
NaH, CH3(CH2)3Cl
13a (45%)
9
types, are initially treated with a single dose of 10 lM of the test
+
compound.13 The output of the single dose is reported as a mean
growth % graph, and a summary is presented in Figure 3 (the full
single dose mean growth % graph for each compound is given in
the Supplementary data accompanying this article).
DMF, 120 °C, 1 h
N
N
N
N
13b
(43%)
In agreement with the MTT assay evaluations (Table 1), imino-
quinone 2a displayed high activity towards the DU-145 cell line,
with activity towards melanoma and CNS cancers also observed.
Quinone 2b is inactive using the NCI screen. Increasing the size
of the fused alicyclic rings from pyrido to azepino resulted in a loss
in activity for iminoquinone 8, a trend also observed in the MTT as-
says results for the analogous quinone series.
Iminoquinone 2a gave a variable toxicity profile, and as a result,
was selected for assessment at the subsequent stage of five dose
testing. From these results the GI50 (concentration of compound re-
quired to inhibit cell growth by 50%) for compound 2a was ob-
tained for each cell line, and a summary is presented in Figure 4
(the full five dose screening results for compound 2a are given in
the Supplementary data accompanying this article). While imino-
quinone 2a was highly active towards DU-145, it was almost inac-
tive towards another prostate cancer cell line, PC-3, which
correlates with measured high and low NQO1 activity in the
DU-145 and PC-3 cell lines, respectively (Fig. 5). NQO1 has been
identified as a molecular target, and its activity in the NCI-60 cell
line panel has been quantified.6 The average NQO1 activity for each
cancer type is shown in Figure 5. When Figures 4 and 5 are
NO2
N
N
N
(i), (ii)
HNO3, H2SO4
80 °C, 3 h
13a
5
(92%)
N
14
(90%)
Scheme 3. Synthesis of 1,5-dibutylimidazo[5,4-f]benzimidazol-4,8-dione 5.
Reagents and conditioins: (i) H2 (6 bar), Pd-C, MeOH, rt, 16 h; (ii) Frémy’s salt,
KH2PO4, rt, 1 h.
CCL-2) and prostate (DU-145) cancer. For comparison purposes
cytotoxicity against a human normal fibroblast cell line (GM00637)
was also determined. Table 1 shows IC50 values of 3c, 4 and 5, along
with that of MMC, and previously reported imidazobenzimidazol-
equinones evaluated under the same conditions. The cytotoxicity
of dipyrroloimidazo[5,4-f]benzimidazolequinone 3a was not ob-
tained due to its poor solubility in organic and aqueous solvents.
The cytotoxicities of pyrrolo–pyrido 3b, dipyrido 2b and diaze-
pino 3c were compared in order to assess the influence that ring
size has on the cytotoxicity. Cytotoxicity clearly diminishes as
the size of the fused alicyclic ring increases across the series from
3b to 2b, to 3c. Diazepino-fused compound 3c is about six times
less toxic towards the DU-145 cell line than the pyrrolo–pyrido
analog 3b. An increase in cytotoxicity was observed towards all
three cell lines, when the extent of ring fusion was decreased from
the fully fused system 2b, to mono-fused pyrido compound 4, to
the nonfused compound 5 (containing no fused pyrido rings). Nev-
ertheless, MMC is the most toxic compound evaluated against all
three cell lines. The high potency of MMC may be attributed to
the fact that it possesses DNA alkylating functionalities that allow
it to form DNA-crosslinks, which induces apoptosis.16
2.3. Developmental Therapeutic Program (DTP) National Cancer
Institute (NCI) 60 human tumor cell line screen and COMPARE
analysis
Iminoquinone 2a was the most potent imidazobenzimidazole
evaluated towards the DU-145 cell line (Table 1), a cell line which
is also part of the DTP NCI-60 human tumor cell line screen. There-
fore, iminoquinone 2a, quinone analog 2b, and diazepino-fused
Figure 3. Summary of DTP NCI-60 single dose (10 lM) screening results; reported
as the average percent growth of each cancer type (with prostate cancer cell lines
shown separately) after treatment with compounds 2a, 2b and 8.

3226
V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
therefore, it is possible to determine which structural features ex-
ert greatest influence on the activity of these molecules. All three
molecules have the iminoquinone motif incorporated in some
way. All three are flat aromatic molecules with p-stacking capabil-
ities, and all have additional groups or positions with hydrogen
bonding capabilities. It seems that the iminoquinone moiety is
the most significant structural feature of compound 2a, and from
its correlation to NQO1, and the excellent correlations to other
NQO1 substrates, it appears that the iminoquinone moiety is very
significant with regard to NQO1 selectivity.
2.4. Computational docking
The NQO1 mechanism of action is described as a ping-pong
mechanism, as both NAD(P)H and the quinone substrate must
independently occupy the same binding site. NQO1 has a flavin
adenine dinucleotide (FAD) redox cofactor that remains bound to
the NQO1 protein during the catalytic cycle. The FAD cofactor is
reduced to FADH2 by NAD(P)H, and the resulting NAD(P)+ is lost.
The quinone substrate now enters the active site and is reduced
to a hydroquinone by FADH2. The substrate is then expelled as a
hydroquinone, regenerating the NQO1 enzyme (Scheme 4).1
Ten compounds were computationally docked into the active
site of human NQO1 (PDB code 1D4A).17,18 The in silico approach
allows us to gain insights into the structural requirements for
interactions between the quinones and NQO1, and can assist in
the design of NQO1 selective compounds. All molecular modeling
was performed using the molecular operating environment
(MOE) 2009.10 program.19 The NQO1 crystal structure has two
homodimers, each of which is made of two interlocked monomers.
The physiological unit of the quinone reductase is a dimer,20 which
was thus employed in the computational modeling. After docking
and relaxation/minimization of the structures (for details see Sec-
tion 4), the interaction energies (I.E.) were calculated (Table 2).
In all cases, the interaction between the ligand and the protein
lowers the total energy considerably, leading to stable complexes.
The most favorable docking energy was obtained for MMC
(ꢀ88.4 kcal molꢀ1). It should, however, be noted that reduction
by NQO1 is an activation step and does not directly result in cell
death. Therefore the calculated I.E. in Table 2 do not correlate well
with the observed cytotoxicity for many of the compounds. Fol-
lowing docking, models of selected ligand–protein complexes were
further refined by performing MD simulation and a final energy
minimization (rms gradient of 0.01 kcal molꢀ1 Åꢀ1) using the LigX
Figure 4. Summary of DTP NCI-60 five dose screening results for iminoquinone 2a;
reported as the average-logGI50 on each histological cancer type (with prostate
cancer cell lines shown separately). GI50 = concentration of compound required to
inhibit cell growth by 50%.
compared an overall trend is apparent. Generally, cytotoxicity in-
creases with increased levels of NQO1 activity, and this trend
was examined further using the COMPARE software.
The COMPARE program can be accessed through the DTP NCI
website (http://dtp.nci.nih.gov/), and can be used to correlate
mean graph data with known molecular target levels in the NCI-
60 panel (expressed as a pearson correlation coefficient). The data
produced can be used to support hypotheses concerning the com-
pound’s mechanism of action.13 COMPARE analysis was performed
using the five dose data for 2a. Compound 2a showed a moderate
to good correlation of 0.51 with NQO1 activity. The clinically used
drug MMC is known to be reductively activated by NQO1,16 and
has the highest correlation to NQO1 of the 171 standard agents
at the NCI. However, with only a moderate correlation of 0.43,
compound 2a compares favorably (Fig. 6). Compound 2a was also
compared to the ca. 43,000 synthetic compounds in the NCI data
base. The two highest correlations were obtained for compounds
15 and 16, with respective coefficients of 0.87 and 0.77 to imino-
quinone 2a. Compound 15 has a correlation of 0.64 to NQO1, the
second highest of all synthetic compounds in the NCI compound
database, while compound 16 has a correlation of 0.47 to NQO1.
Compounds 15 and 16 possess structural similarities to 2a, and
NH
O
OCONH2
OMe
H2N
Me
N
N
N
N
N
NH
O
O
MMC
2a
NSC: 26980
0.43 (NQO1)
NSC: 753790
0.51 (NQO1)
N
N
O
N
O
OH
H2N
O
O
15
16
NSC: 84167
0.87 (2a)
NSC: 94945
0.77 (2a)
0.64 (NQO1)
0.47 (NQO1)
Figure 5. Average NQO1 activity in each of the nine histological cancer types
represented in the DTP NCI-60 cell line panel (with prostate cancer cell lines shown
separately).
Figure 6. Results of COMPARE analysis of the five dose data for iminoquinone 2a
expressed as pearson correlation coefficients. Blue circle indicates the hydride
acceptor carbon in docking studies.

V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
3227
isoalloxazine
ring of FADH2
H
N
R
N
O
N
H
N
H
O
dCquinone-N5
H
H+
O
HO
H
O
OH
H+
OH
O
hydroquinone
quinone
substrate
Scheme 4. Transfer of hydride from FADH2 to a quinone in the NQO1 active site;
dCquinone-N5 indicated by dashed line.
Table 2
Interaction energies of heterocyclic substrates with NQO1
Entry
Complex with NQO1
Interaction energy (I.E.)a (kcal molꢀ1
)
Figure 7a. Molecular model of MMC docked into the active site of NQO1. Hydrogen
bonds are shown as dashed lines and the FAD cofactor is shown in yellow.
1
2
3
4
5
6
7
8
9
MMC
1b
2a
2b
3a
3b
3c
4
ꢀ88.4
ꢀ63.0
ꢀ64.0
ꢀ76.8
ꢀ53.6
ꢀ56.2
ꢀ61.9
ꢀ69.9
ꢀ65.1
ꢀ59.8
5
15
10
a
I.E. = Ecomplexꢀ(Eenz + Eligand).
interface in MOE.19 In these calculations the protein atoms far from
the substrate were constrained not to move, while receptor atoms
in the active site, and the quinone substrate in the active site, were
treated as flexible. The active site was defined as a distance of 8 Å
from the substrate. Several 2D and 3D properties were calculated
after energy minimization, including the binding affinity (the esti-
mated binding affinity reported in units of pKi), and the Cquinone-N5
distance (Table 3). The binding affinity is that of London
DG scoring
function reported in units of pKi. The affinity G scoring function
D
estimates the contribution to the free energy of binding using a
linear function of hydrophobic, ionic, hydrogen bond and metal
ligation terms.
For efficient reduction to occur the substrate–protein complex
must be stable (i.e., have a high binding affinity), and the substrate
must bind in a suitable orientation (i.e., the Cquinone-N5 distance
(Scheme 4) should be within 5 Å) in order to permit the hydride
transfer from the reduced FADH2 group to the quinone substrate.
MMC was found to bind with the highest affinity and a short
distance to the N5 of the FADH2 isoalloxazine ring, and therefore,
is calculated to be efficiently reduced by NQO1. A detailed repre-
sentation of the MMC-NQO1 active site is shown in Figure 7a.
Three residues hydrogen-bond to MMC (namely Gly193, His194
and Tyr126), holding it in a favorable position for reduction to oc-
cur. The His194 residue is reported to play an important role in the
interaction of quinone substrates with NQO1.21,22 The oxygen atom
of the carbamate (carbonyl) side chain of MMC hydrogen-bonds to
His194, influencing the position of MMC in the active site, so that
Figure 7b. Hydrophobic contact surface between MMC and NQO1.
Table 3
The affinity of selected heterocyclic substrates with NQO1 and the distance between
substrate and FADH2 for reduction
Entry
Complex with NQO1
Affinity (units of pKi)
dCquinone-N5 (Å)
1
2
3
MMC
15
2a
8.60
8.53
7.41
3.89
3.59
5.09
Figure 8a. Molecular model of compound 15 docked into the active site of NQO1.
Hydrogen bonds are shown as dashed lines and the FAD cofactor is shown in yellow.

3228
V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
Figure 8b. Hydrophobic contact surface between 15 and NQO1.
Figure 9b. Hydrophobic contact surface between 2a and NQO1.
Figure 9a. Molecular model of compound 2a docked into the active site of NQO1.
Hydrogen bonds are shown as dashed lines and the FAD cofactor is shown in yellow.
Figure 10. Molecular model of compound 2a in the active site of NQO1 after
hydride transfer. Hydrogen bonds are shown as dashed lines and the FAD cofactor is
shown in yellow.
the reactive carbon (indicated in Fig. 6) is located 3.89 Å from the
N5 of the FADH2. Further the Trp105 and Tyr128 residues provide
a pocket on one side of the active site towards which MMC displays
favorable hydrophobic contacts (Fig. 7b).
nitrogen by proton donation (Fig. 10). This model was obtained
through energy minimization employing the same force field and
definition of the active site (8 Å from the ligand). The charges of
the system and the parameters of the reduced ligand were derived
employing the semiempirical PM3 hamiltonian. More accurate
electronic events resulting from reduction can be evaluated by
quantum mechanics methods, but are beyond the scope of this
work.
The docking studies for compound 15 predicted a high affinity
and a short Cquinone-N5 distance (3.59 Å). Hydrogen bonding to
Tyr126 and His194, positions the hydride acceptor carbon of com-
pound 15 close to the N5 of the FADH2 group (Fig. 8a). The sub-
strate also shows favorable hydrophobic contacts with Trp105
and Tyr128, as indicated in Figure 8b. However, the calculations
show no interactions with the imine nitrogen of the iminoquinone
moiety prior to hydride transfer. Interestingly, the His194 forms a
hydrogen bond with the carboxyl side chain of 15 in a similar man-
ner to the carbamate group of MMC. This may indicate that the
efficiency of NQO1 reduction can be enhanced by the presence of
a good hydrogen bonding group (donor or acceptor) located four
to five bonds away from the (imino)quinone substructure.
Compound 2a differs from MMC and compound 15, in that it
lacks a polar group (carbamate or carboxylic acid side chain) and
thus, does not hydrogen bond to His194. This may account for its
lower affinity and longer Cquinone-N5 distance. Compound 2a inter-
acts through hydrogen bonding with Tyr126 and Tyr128 (Fig. 9a),
and presents favorable hydrophobic interactions with Pro68,
Ala68, Tyr126 and Tyr128 (Fig. 9b). Before reduction, Tyr128
hydrogen bonds to the imidazole nitrogen. Again the calculations
show no hydrogen bonding to the imine nitrogen of the iminoqui-
none of 2a. After hydride transfer, the phenolic hydroxyl group
of Tyr128 rotates to stabilize the resulting negatively charged
3. Conclusions
Cytotoxicity evaluations have shown that the expansion of the
fused alicyclic rings in imidazo[5,4-f]benzimidazolequinones and
iminoquinones decreases potency, while potency is increased by
replacing the fused pyrido rings by n-butyl substituents. The DTP
NCI-60 cell line screening results of dipyridoiminoquinone 2a sup-
ported the exceptionally high cytotoxicity towards the prostate
DU-145 cancer cell line, previously reported using the MTT assay.11
DTP NCI-60 screening also revealed high potency towards several
other tumor cell lines (including some melanoma and CNS cancer
cell lines). COMPARE analysis of the DTP NCI-60 screening results
for iminoquinone 2a show a moderate to good correlation of 0.51
to NQO1 activity, and excellent correlations to other synthetic
compounds that also show good correlations to NQO1 activity.
These compounds share structural similarities, and it seems that
the presence of an iminoquinone moiety can increase NQO1

V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
3229
specificity. Computational docking into the NQO1 active site was
performed, including the first comprehensive MMC-NQO1 docking
study. The binding affinity and Cquinone-N5 distance were calcu-
lated for compounds that showed correlation to NQO1 activity.
The presence of a good hydrogen bonding group located four to five
bonds away from the (imino)quinone substructure could lead to
more efficient NQO1 reduction through interaction with the
His194 residue. The absence of such a substituent on compound
2a may account for its low calculated binding affinity and large
Cquinone-N5 distance. Incorporating such a substituent in future
analogs of 2a may result in greater activity and specificity by
increasing the efficiency of reduction by NQO1. The computational
docking of iminoquinones showed no extra interactions with
NQO1 compared to their quinone counterparts. The greater
NQO1 selectivity displayed by the iminoquinones compared to
the quinones, may be due to the iminoquinones taking part in few-
er competitive biological interactions compared to the quinones.
Therefore, the results and conclusions reported here should con-
tribute to the future rational design of more active and selective
NQO1 substrates.
1452, 1409, 1346 (NO2), 1298, 1271, 1211, 1187, 1117, 1055; dH
2.67–2.74 (m, 2H, 2 or 8-CH2), 2.77–2.85 (m, 2H, 2 or 8-CH2),
3.11 (t, J 7.8, 2H, 3 or 9-CH2), 3.21 (t, J 7.7, 2H, 3 or 9-CH2), 4.23
(t, J 7.1, 2H, 1 or 7-CH2), 4.54 (t, J 7.2, 2H, 1 or 7-CH2), 7.72 (s,
1H, 11-H); dC 23.7, 24.4 (3,9-CH2), 25.8, 26.1 (2,8-CH2), 43.6, 48.0
(1,7-CH2), 106.5 (11-CH), 123.7, 124.8, 129.9, 140.2, 146.5 (all C),
164.1, 165.2 (3a,9a-C); HRMS (ESI): found MH+, 284.1150.
C14H14N5O2 requires, 284.1147.
4.1.1.2. 7-Nitro-2,3,4,5,10,11,12,13-octahydro-1H,9H-azepino[1,2-
a]azepino[10,20:1,2]imidazo[5,4-f]benzimidazole (7).
2,3,4,5,
10,11,12,13-Octahydro-1H,9H-azepino[1,2-a]azepino[10,20:1,2]imi-
dazo[5,4-f]benzimidazole gave the 7-nitro imidazo[5,4-f]benzimid-
azole 7 (0.219 g, 92%) as a yellow solid; Rf 0.17 (EtOAc–MeOH 9:1);
mp 278–280 °C; vmax (neat, cmꢀ1) 3029, 2938, 2923, 2849, 1545,
1520 (NO2), 1480, 1468, 1442, 1410, 1348 (NO2), 1305, 1254, 1215,
1190, 1135, 1088, 1045; dH 1.74–1.99 (m, 12H), 3.10 (t, J 5.3, 2H, 5
or 13-CH2), 3.16 (t, J 5.6, 2H, 5 or 13-CH2), 4.05–4.15 (m, 2H, 1 or 9-
CH2), 4.19 (t, J 4.9, 2H, 1 or 9-CH2), 7.64 (s, 1H, 15-H); dC 25.3, 25.4,
28.1, 28.6, 30.1, 30.4, 30.5, 30.8 (all CH2), 45.1, 46.8 (1,9-CH2), 103.2
(15-CH), 124.4, 125.1, 133.2, 133.9, 140.8 (all C), 161.0, 161.1 (5a or
13a-C); HRMS (ESI): found MH+, 340.1783. C18H22N5O2 requires,
340.1774.
4. Experimental section
4.1. General
4.1.2. General procedure for formation of quinones
The nitro-imidazo[5,4-f]benzimidazole(s) (0.6 mmol) and Pd-C
(5%, 20 mg) were stirred under 6 bar of H2 in MeOH (100 mL) at
room temperature for 16 h. The mixture was filtered, evaporated,
and KH2PO4 (0.2 M, 100 mL) added. A separate solution of Frémy’s
salt (0.483 g, 1.8 mmol) in KH2PO4 (0.2 M, 50 mL,) was added, and
the mixture stirred at room temperature for 1 h. Extracted with
CH2Cl2 (5 ꢂ 50 mL), and evaporated. HCl (2 M, 50 mL) was added
to the residue, and the solution stirred for 15 min (not required
for quinones 4 and 5). The acidic solution was neutralized with so-
lid Na2CO3 and extracted with CH2Cl2 (5 ꢂ 50 mL). The combined
organic extracts were dried (Na2SO4), evaporated, and the residue
purified by dry column vacuum chromatography with a gradient
elution of EtOAc and MeOH to give the quinones below:
All materials were obtained commercially, and the synthesis of
2,3,8,9-tetrahydro-1H,7H-pyrrolo[1,2-a]pyrrolo[10,20:1,2]imidazo
[5,4-f]benzimidazole and 2,3,4,5,10,11,12,13-octahydro-1H,9H-
azepino[1,2-a]azepino[10,20:1,2]imidazo[5,4-f]benzimidazole has
been reported using a one-pot double radical cyclization proto-
col.11 Thin layer chromatography (TLC) was carried out on alumin-
ium-backed plates coated with silica gel (Merck Kieselgel 60 F254).
Flash chromatography and dry column vacuum chromatography
were carried out using Merck Kieselgel 60H silica gel (particle size
0.040–0.060 mm) and Merck Kieselgel silica gel 60 (particle size
0.015–0.040 mm), respectively, using the specified eluents. Melt-
ing points were measured on a Stuart scientific melting point appa-
ratus SMP3. Infrared spectra were recorded using a Perkin–Elmer
spec 1 with ATR attached. NMR spectra were recorded using a Joel
GXFT 400 MHz instrument equipped with a DEC AXP 300 com-
puter workstation. Chemical shifts are reported in ppm, J values
are in Hz with Me4Si as internal standard, and CDCl3 as solvent.
All NMR assignments were supported by DEPT and 1H-13C NMR
2D spectra. High resolution mass spectroscopy was carried out
using electrospray ionization (ESI) on a Waters LCT Premier XE
spectrometer by manual peak matching. The precision of all accu-
rate mass measurements are better than 5 ppm. Hydrogenation
reactions were carried out using a ParrÒ 5500 series compact
reactor.
4.1.2.1. 2,3,8,9-Tetrahydro-1H,7H-pyrrolo[1,2-a]pyrrolo[10,20:1,2]
imidazo[5,4-f]benzimidazol-5,11-dione
dazo[5,4-f]benzimidazole 6 gave quinone 3a (0.122 g, 76%), as a yel-
low solid; Rf 0.55 (CH2Cl2–MeOH 9:1); mp >350 °C; vmax (neat, cmꢀ1
(3a).
5-Nitro-imi-
)
3316, 2990, 2957, 1663 (C@O), 1514, 1485, 1438, 1403, 1356, 1301,
1267, 1225, 1214, 1200, 1179, 1040; dH 2.69–2.76 (m, 4H, 2,8-CH2),
2.96 (t, 4H, J 7.6, 3,9-CH2), 4.29 (t, J 7.2, 4H, 1,7-CH2); dC 23.0
(3,9-CH2); 26.5 (2,8-CH2), 45.4 (1,7-CH2), 130.7, 145.9, 160.2 (all C),
171.8 (C@O); HRMS (ESI): found MH+, 269.1047. C14H13N4O2
requires, 269.1039.
4.1.2.2. 7-Imino-2,3,4,5,10,11,12,13-octahydro-1H,9H-azepino
[1,2-a]azepino[10,20:1,2]imidazo[5,4-f ]benzimidazol-15-one (8)
and 2,3,4,5,10,11,12,13-octahydro-1H,9H-azepino[1,2-a]azepi-
4.1.1. General procedure for nitration
The imidazo[5,4-f]benzimidazole (0.7 mmol) was heated in
concentrated H2SO4 (10 mL) and concentrated HNO3 (5 mL) at
80 °C for 3 h. The cooled mixture was diluted with water
(100 mL), neutralized with solid Na2CO3, and extracted with
CH2Cl2 (5 ꢂ 50 mL). The combined organic extracts were dried
(Na2SO4), evaporated and the residue purified by dry column vac-
uum chromatography with gradient elution of EtOAc and MeOH to
give the nitro-imidazo[5,4-f]benzimidazoles below:
no[10,20:1,2]imidazo[5,4-f]benzimidazol-7,15-dione (3c).
7-
Nitroimidazo[5,4-f]benzimidazole 7 gave a gave a ꢁ2.3:1 mixture
of iminoquinone 8 and quinone 3c, respectively. The mixture was
purified by dry column vacuum chromatography with gradient
elution of EtOAc and MeOH to give iminoquinone 8. The remaining
mixture was hydrolyzed with HCl to give quinone 3c.
Iminoquinone 8 (50 mg, 26%) as a yellow solid; Rf 0.60 (CH2Cl2–
MeOH 9:1); mp 260–265 °C decomp; vmax (neat, cmꢀ1) 2940, 2918,
2887, 2842, 1650, 1611, 1518, 1495, 1474, 1439, 1412, 1386, 1356,
1349, 1319, 1278, 1261, 1228, 1090, 1073, 1042, 1009; dH 1.63–
1.90 (m, 12H), 2.91 (t, J 5.7, 2H, 5 or 13-CH2), 2.97 (t, J 5.6, 2H, 5
or 13-CH2), 4.57–4.66 (m, 2H, 1 or 9-CH2), 4.76–4.88 (m, 2H, 1 or
9-CH2), 10.69 (s, 1H, NH); dC 25.2, 25.3, 28.4, 29.5, 29.6, 31.0 (all
4.1.1.1. 5-Nitro-2,3,8,9-tetrahydro-1H,7H-pyrrolo[1,2-a]pyrrolo
[10,20:1,2]imidazo[5,4-f]benzimidazole (6).
2,3,8,9-Tetrahy-
dro-1H,7H-pyrrolo[1,2-a]pyrrolo[10,20:1,2]imidazo[5,4-f]benzimid-
azole gave the 5-nitro imidazo[5,4-f]benzimidazole 6 (0.169 g,
85%) as a yellow solid; Rf 0.52 (CH2Cl2–MeOH 9:1); mp 245–
247 °C; vmax (neat, cmꢀ1) 3225, 1647, 1566, 1522 (NO2), 1484,

3230
V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
CH2), 45.9, 46.4 (1,9-CH2), 126.2, 131.9, 139.4, 156.4, 156.7, 156.9
white solid; Rf 0.42 (EtOAc–MeOH 9:1); mp 209–211 °C; vmax
(neat, cmꢀ1) 3093, 3032, 2960, 2930, 2875, 1734, 1516, 1491,
1448, 1370, 1354, 1289, 1223, 1198, 1166, 1117; dH (400 MHz,
CDCl3) 0.94 (t, J 7.4, 6H, CH3), 1.32–1.41 (m, 4H), 1.87–1.94 (m,
4H), 4.22 (t, J 7.1, 4H, NCH2), 7.73 (s, 2H, 4,8-H), 7.93 (s, 2H, 2,6-
H); dC (100 MHz, CDCl3) 13.7 (CH3), 20.1 (CH2), 31.6 (CH2), 45.1
(NCH2), 99.1 (4,8-CH), 131.6 (C), 141.7 (C), 144.3 (2,6-CH); HRMS
(ESI): found MH+, 271.1937. C16H23N4 requires 271.1923.
4.1.3.3.2. 1,7-Dibutylimidazo[5,4-f]benzimidazole (13b). (0.372 g,
43%): white solid; Rf 0.26 (EtOAc–MeOH 9:1); mp 86–87 °C; vmax
(neat, cmꢀ1) 3405, 3226, 3087, 2957, 2930, 2872, 1637, 1525,
1492, 1464, 11450, 1363, 1277, 1262, 1210, 1202, 1137, 1109,
1093, 1010; dH (400 MHz, CDCl3) 0.97 (t, J 7.4, 6H, CH3), 1.35–1.44
(m, 4H), 1.87–1.94 (m, 4H), 4.21 (t, J 7.1, 4H, NCH2), 7.21 (s, 1H, 8-
H), 7.92 (s, 2H, 2,6-H), 8.19 (s, 1H, 4-H); dC (100 MHz, CDCl3) 13.7
(CH3), 20.2 (CH2), 31.5 (CH2), 45.0 (NCH2), 88.3 (8-CH), 110.4 (4-
CH), 132.1 (C), 141.1 (C), 143.7 (2,6-CH); HRMS (ESI): found MH+,
271.1916. C16H23N4 requires 271.1923.
(all C), 173.0 (C@O); HRMS (ESI): found MH+, 324.1827.
C18H22N5O requires, 324.1824.
Quinone 3c (0.111 g, 57%), as a yellow solid; Rf 0.62 (CH2Cl2–
MeOH 9:1); mp 340–346 °C decomp; vmax (neat, cmꢀ1) 2950,
2919, 2886, 2842, 1654 (C@O), 1523, 1492, 1474, 1437, 1373,
1357, 1346, 1318, 1280, 1241, 1227, 1180, 1093, 1076, 1055,
1016; dH 1.69–1.94 (12H, m), 2.99 (4H, t, J 5.6, 5,13-CH2), 4.56–
4.64 (4H, m, 1,9-CH2); dC 25.0 (CH2), 28.2 (CH2), 29.4 (5,13-CH2),
30.9 (CH2), 46.0 (1,9-CH2), 130.5, 141.4, 157.8 (all C), 172.7
(C@O); HRMS (ESI): found MH+, 325.1652. C18H21N4O2 requires,
325.1665.
4.1.3. General procedure for alkylation
Imidazo[5,4-f(4,5-f)]benzimidazole 9 or 10 (3.2 mmol) and NaH
(84 mg, 3.5 mmol) in DMF (50 mL) were heated at 120 °C for
10 min. 1-Chlorobutane (0.322 g, 3.5 mmol for 10 and 0.653 g,
7.1 mmol for 13a and 13b) or 1-chloro-4-(phenylselenyl)butane
(0.868 g , 3.5 mmol) for 11a and 11b was added and the mixture
stirred for 1 h. The cooled mixture was evaporated, CH2Cl2 added,
and filtered. The filtrate was evaporated and the residue purified
by dry column vacuum chromatography with gradient elution of
EtOAc and MeOH to give the following imidazobenzimidazoles:
4.1.4. 3-Butyl-6,7,8,9-tetrahydroimidazo[5,4-f]pyrido[1,2-a]
benzimidazole (12)
To a solution of phenylselenide 11a (0.600 g, 1.4 mmol) in tolu-
ene (50 mL) at reflux under N2, was added a solution of 1,1’-
azobis(cyclohexanecarbonitrile) (ACCN) (0.855 g, 3.5 mmol) and
Bu3SnH (1.019 g, 3.5 mmol) in toluene (50 mL) using a syringe
pump at a rate of 2.2 mL hꢀ1. After ꢁ11.5 h the reaction was al-
lowed to proceed in the open air. The cooled mixture was extracted
with AcOH (80%, 5 ꢂ 20 mL), the combined extracts washed with
petroleum ether (3 ꢂ 20 mL), and evaporated. Na2CO3 solution
(5%, 50 mL) was added to the acidic residue, and extracted with
CH2Cl2 (5 ꢂ 50 mL). The combined organic extracts were dried
(Na2SO4), evaporated, and the residue purified by dry column vac-
uum chromatography with a gradient elution of EtOAc and MeOH
to give imidazobenzimidazole 12 (0.329 g, 87%) as a white solid; Rf
0.33 (EtOAc–MeOH 9:1); mp 238–240 °C; vmax (neat, cmꢀ1) 2924,
2856, 1529, 1477, 1447, 1358, 1314, 1259, 1189, 1119, 1084; dH
0.92 (t, J 7.4, 3H, CH3), 1.28–1.38 (m, 2H), 1.84–1.91 (m, 2H),
1.99–2.05 (m, 2H), 2.12–2.18 (m, 2H), 3.10 (t, J 6.4, 2H, 6-CH2),
4.11 (t, J 6.1, 2H), 4.20 (t, J 7.1, 2H), 7.58 (s, 1H), 7.60 (s, 1H),
7.89 (s, 1H, 2-H); dC 13.7 (CH3), 20.1, 20.9, 22.8 (all CH2), 25.8 (6-
CH2), 31.6, 42.6, 45.1 (all CH2), 97.7, 98.3 (4,11-CH), 131.5, 132.5,
140.6, 140.8 (all C), 143.5 (2-CH), 153.1 (5a-C); HRMS (ESI): found
MH+, 269.1759. C16H21N4 requires, 269.1766.
4.1.3.1.
(10).
1-Butyl-5(7)H-imidazo[5,4-f(4,5-f)]benzimidazole
(0.400 g, 59%), as a white solid; Rf 0.18 (EtOAc–MeOH
4:1); mp >350 °C; vmax (neat, cmꢀ1) 3072, 2935, 1577, 1508,
1476, 1446, 1436, 1371, 1284, 1225, 1210, 1136, 1115, 1059,
1021; dH 0.94 (t, J 7.4, 3H, CH3), 1.32–1.42 (m, 2H), 1.87–1.94 (m,
2H), 4.21 (t, J 7.1, 2H, NCH2), 7.68 (bs, 1H), 7.92 (bs, 1H), 7.95 (s,
1H), 8.15 (s, 1H), NH not observed; dC 13.7 (CH3), 20.2 (CH2),
31.6 (CH2), 45.2 (NCH2), 131.9 (C), 141.7 (C), 141.9 (CH), 144.1
(CH); HRMS (ESI): found MH+, 215.1291. C12H15N4 requires,
215.1297.
4.1.3.2. Phenylselenides 11a and 11b. 4.1.3.2.1. 1-Butyl-5-(4-
(phenylseleno)butyl)imidazo[5,4-f]benzimidazole (11a). (0.586 g,
43%), as a white solid; Rf 0.32 (EtOAc–MeOH 9:1); mp 287–
293 °C; vmax (neat, cmꢀ1) 3244, 2956, 1647, 1525, 1490, 1447,
1359, 1287, 1173, 1115, 1082; dH 0.94 (t, J 7.4, 3H, CH3), 1.31–
1.41 (m, 2H), 1.67–1.75 (m, 2H), 1.86–1.93 (m, 2H), 2.00–2.08
(m, 2H), 2.87 (t, J 7.1, 2H, CH2SePh), 4.17–4.23 (m, 4H, NCH2),
7.16–7.24 (m, 3H, Ph-H), 7.40–7.44 (m, 2H, Ph-H), 7.72 (s, 1H, 4
or 8-H), 7.74 (s, 1H, 4 or 8-H), 7.88 (s, 1H, 2 or 6-H), 7.95 (s, 1H,
2 or 6-H); dC 13.7 (CH3), 20.1 (CH2), 27.2 (CH2SePh), 27.3, 29.4,
31.5 (all CH2), 44.8, 45.1 (NCH2), 99.1, 99.2 (4,8-CH), 127.1 (Ph-
CH), 129.2 (Ph-CH), 129.8, 131.4, 131.6 (all C), 132.9 (Ph-CH),
141.6 (C), 141.7 (C), 144.1, 144.3 (2,6-CH); HRMS (ESI): found
4.1.5. 3-Butyl-6,7,8,9-tetrahydroimidazo[5,4-f]pyrido[1,2-a]
benzimidazol-4,11-dione (4)
A mixture of 4 and 11-nitroimidazo[5,4-f]benzimidazoles pre-
pared according to the general procedure for nitration, and sub-
jected to the general procedure for formation of quinones gave
quinone 4 (0.157 g, 88%), as a yellow solid; Rf 0.17 (EtOAc–MeOH
9:1); mp 216–217 °C; vmax (neat, cmꢀ1) 3190, 2955, 2871, 1678,
1656 (C@O), 1512, 1488, 1473, 1425, 1377, 1334, 1299, 1266,
1253, 1222, 1193, 1072, 1052; dH 0.92 (t, J 7.4, 3H, CH3), 1.29–
1.38 (m, 2H), 1.77–1.84 (m, 2H), 1.93–2.07 (m, 4H), 2.96 (t, J 6.4,
2H, 6-CH2), 4.31–4.35 (m, 4H), 7.60 (s, 1H, 2-H); dC 13.6 (CH3),
19.7, 19.8, 22.2 (all CH2), 25.1 (6-CH2), 32.4, 45.7, 47.0 (all CH2),
130.7 (2 ꢂ C), 142.2 (C), 142.5 (2-CH), 143.2 (C), 151.6 (C), 171.9
(C@O), 172.1 (C@O); HRMS (ESI): found MH+, 299.1497.
MH+, 427.1399. C22H27N4 Se requires, 427.1401.
80
4.1.3.2.2.
1-Butyl-7-(4-(phenylseleno)butyl)imidazo[4,5-f]benz-
imidazole (11b). (0.545 g, 40%), as a white solid; Rf 0.11 (EtOAc–
MeOH 9:1); mp 78–80 °C; vmax (neat, cmꢀ1) 3374, 3100, 2956,
2930, 2871, 1718, 1683, 1636, 1577, 1524, 1493, 1477, 1456,
1436, 1362, 1263, 1211, 1144, 1112, 1064, 1021; dH 0.94 (t, J 7.4,
3H, CH3), 1.30–1.40 (m, 2H), 1.66–1.73 (m, 2H), 1.83–1.90 (m,
2H), 1.99–2.06 (m, 2H), 2.87 (t, J 7.0, 2H, CH2SePh), 4.13–4.23 (m,
4H, NCH2), 7.15–7.23 (m, 4H, Ph-H & 8-H), 7.34–7.39 (m, 2H, Ph-
H), 7.81 (s, 1H, 2 or 6-H), 7.89 (s, 1H, 2 or 6-H), 8.20 (s, 1H, 4-H);
dC 13.7 (CH3), 20.1 (CH2), 27.1 (CH2), 27.2 (CH2SePh), 29.1, 31.5
(all CH2), 44.6 , 45.0 (NCH2), 88.5 (8-CH), 110.2 (4-CH), 127.1
(Ph-CH), 129.2 (Ph-CH), 129.6, 131.8, 132.0 (all C), 132.8 (Ph-CH),
141.1 (2 ꢂ C), 143.5, 143.7 (2,6-CH); HRMS (ESI): found MH+,
C16H19N4O2 requires, 299.1508.
4.1.6. 1,5-Dibutyl-4-nitroimidazo[5,4-f]benzimidazole (14)
Imidazo[5,4-f]benzimidazole 13a was subjected to the general
nitration procedure to give 14 (0.199 g, 90%): yellow solid; Rf
0.44 (EtOAc–MeOH 9:1); mp 155–156 °C; vmax (neat, cmꢀ1) 3109,
2949, 2929, 2870, 1763, 1526 (NO2), 1512, 1498, 1485, 1464,
1436, 1369 (NO2), 1349, 1333, 1317, 1280, 1230, 1200, 1191,
80
427.1386. C22H27N4 Se requires, 427.1401.
4.1.3.3. Dibutylimidazobenzimidazoles 13a and 13b. 4.1.3.3.1.
1,5-Dibutylimidazo[5,4-f]benzimidazole (13a). (0.389 g, 45%):

V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
3231
1125, 1087; dH (400 MHz, CDCl3) 0.86 (t, J 7.4, 3H, CH3), 0.93 (t, J
7.4, 3H, CH3), 1.15–1.25 (m, 2H), 1.30–1.39 (m, 2H), 1.58–1.66
(m, 2H), 1.85–1.92 (m, 2H), 4.26 (t, J 7.1, 2H, NCH2), 4.32 (t, J 7.2,
2H, NCH2), 7.97 (s, 1H), 7.99 (s, 1H, 8-H), 8.07 (s, 1H); dC
(100 MHz, CDCl3) 13.56 (CH3), 13.58 (CH3), 19.8, 20.1, 31.7, 32.3
(all CH2), 45.5 (NCH2), 47.8 (NCH2), 105.7 (8-CH), 123.4, 125.9,
131.7, 135.9, 143.7 (all C), 146.4, 147.5 (2 & 6-CH); HRMS (ESI):
found MH+, 316.1769. C16H22N5O2 requires 316.1774.
(20 lL, 5 mg/ml solution) was added, and the cells were incubated
for another 3 h. The supernatant was removed carefully by pipet-
ting. The resultant MTT formazan crystals were dissolved in
100 lL of DMSO and absorbance was determined on a plate reader
at 550 nm with a reference at 690 nm. Cell viability is expressed as
a percentage of the EtOH or DMSO-only treated value. Dose-
response curves were analyzed by nonlinear regression analysis
and IC50 values were estimated using GraphPad Prism software,
v. 5.02 (GraphPad Inc., San Diego, CA, USA).
4.1.7. 1,5-Dibutylimidazo[5,4-f]benzimidazol-4,8-dione (5)
4-Nitro-imidazo[5,4-f]benzimidazole 14 was subjected to the
general procedure for formation of quinones to give 4 (0.166 g,
92%): yellow solid; Rf 0.44 (EtOAc–MeOH 9:1); mp 213–214 °C;
vmax (neat, cmꢀ1) 3068, 2957, 2932, 2871, 1665 (C@O), 1515,
1489, 1459, 1385, 1336, 1217, 1185, 1065, 1033; dH (400 MHz,
CDCl3) 0.92 (t, J 7.4, 6H, CH3), 1.28–1.37 (m, 4H), 1.78–1.85 (m,
4H), 4.35 (t, J 7.2, 4H, NCH2), 7.64 (s, 2H, 2,6-H); dC (100 MHz,
CDCl3) 13.6 (CH3), 19.7 (CH2), 32.6 (CH2), 47.1 (NCH2), 131.0 (C),
142.9 (2,6-CH), 143.5 (C), 172.2 (C@O); HRMS (ESI): found MH+,
301.1656. C16H21N4O2 requires 301.1665; elemental analysis Calcd
(%) for C16H20N4O2: C 63.98, H 6.71, N 18.65. Found: C 63.60, H
6.85, N 18.31.
4.3. Computational methods
The crystallographic coordinates of human NQO1 (1D4A) were
obtained from the RCSB Protein Data Bank. The physiological dimer
was used throughout, and hydrogen atoms were added to the pro-
tein and the prosthetic group (FAD) based on their protonation
state at pH 7. The ligands were built in MOE.
All structures of ligands, enzyme model (NQO1 physiological di-
mer) and enzyme-ligand complexes were geometry minimized
using the AMBER99 molecular mechanics force field24 to within
an rms gradient of 0.05 kcal molꢀ1 Åꢀ1. Throughout, all systems
were surrounded by a distance dependent dielectric model of bulk
water (ESOL = 0, EELECTROSTATIC = qiqj/4pe0rij2). Where, ESOL represents
the implicit solvation energy calculated using the Generalized Born
model (GB/VI). To obtain a relaxed geometry of the initial receptor,
a short molecular dynamics (MD) simulations was performed, fol-
lowed by a final energy minimization step. The MD simulation of
the NQO1 dimer was performed using a canonical ensemble
(NVT), with initial temperature 150 K; heating for 50 ps; simula-
tion temperature 300 K; duration 500 ps; time step 1 fs; tempera-
ture response 1 ps; pressure response 0.5 ps and constraint
tolerance 1 ꢂ 10ꢀ9. After the MD simulation, a final energy minimi-
zation was performed. The resulting protein model was used in the
docking studies.
4.2. Cell culture and cytotoxicity evaluation
An SV40-transformed normal human skin fibroblast cell line
(repository number GM00637) was obtained from the National
Institute for General Medical Sciences (NIGMS) Human Genetic
Cell Repository (Coriell Institute for Medical Research, New Jersey,
USA). The HeLa cervical cancer cell line (repository number CCL-2)
was obtained from the American Type Culture Collection (ATCC).
The DU-145 prostate cancer cell line (ATCC repository number
HTB-81) was obtained from Professor R.W.G. Watson, School of
Medicine & Medical Sciences, University College Dublin, Ireland.
Possible active sites in the receptor model were identified by
using alpha site finder.25 Once defined, ligands were docked into
the enzyme retaining 500 poses using the alpha triangle placement
4.2.1. Cell culture
The SV40-transformed normal human skin fibroblast cell line
(GM00637) was grown in minimum essential media (MEM)
Eagle–Earle BSS supplemented with 15% heat-inactivated fetal
methodology with affinity
DG as scoring function as implemented
in MOE. In the docking studies, flexible ligand structures were gen-
erated using a Monte Carlo algorithm, whereas the receptor was
held fixed according to the minimized geometries.
bovine serum (FBS), penicillin-streptomycin, 2 mM L-glutamine,
2ꢂ essential and nonessential amino acids, and vitamins. HeLa-
CCL-2 cervical cancer cells were grown in Dulbecco’s Modified
Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated
Best scored binding orientations for each ligand were energy
minimized, subjected to MD simulation using the same settings
as above and finally energy minimized, using in all stages the AM-
BER99 force field. The potential energy of the final complex was
calculated (Ecomplex). The ligand was then moved away from the
protein, the noninteracting system energy minimized, and the po-
tential energy of the separated moieties calculated (Eenz + Eligand).
Subtracting this from Ecomplex gives the interaction energy (I.E.) re-
ported herein. Several properties were calculated for each best
scored complex, using the LigX procedure,19 such as the affinity,
fetal bovine serum, penicillin–streptomycin, 2 mM L-glutamine,
and MEM nonessential amino acids. DU-145 prostate cancer cells
were grown in RPMI-1640 medium supplemented with 10%
heat-inactivated fetal bovine serum, penicillin–streptomycin and
2 mM L-glutamine.
4.2.2. Cytotoxicity measurement
Cell viability was determined using the MTT colorimetric as-
say.23 Normal cells were plated into 96-well plates at a density
hydrogen bond interactions, van der Waals interactions and
interactions.
p–p
of 10,000 cells per well (200
over a period of 24 h. HeLa cells were plated into 96-well plates
at a density of 1000 cells per well (100 L per well) and allowed
to adhere over a period of 24 h. DU-145 cells were plated into
lL per well) and allowed to adhere
l
Acknowledgements
96-well plates at a density of 2000 cells per well (200
and allowed to adhere over a period of 24 h.
lL per well)
We thank the Irish Research Council for Science Engineering
and Technology funded by the National Development Plan for
awarding an Embark Scholar Award to Vincent Fagan. We also
gratefully acknowledge the financial support from National
University of Ireland Galway and Science Foundation Ireland for
a Research Frontiers Programme Award (07/RFP/CHEF227). We
are grateful to Professor William Watson (School of Medicine &
Medical Sciences, University College Dublin, Ireland) for the
DU-145 cell line.
Compound solutions were applied in ethanol (1% final concen-
tration in well), and the plates were incubated at 37 °C under a
humidified atmosphere containing 5% CO2 for 72 h. Control cells
were exposed to an equivalent concentration of ethanol alone.
MMC (sigma) solutions were applied in DMSO (1% final concentra-
tion in well) and the plates were incubated at 37 °C under a humid-
ified atmosphere containing 5% CO2 for 72 h. Control cells were
exposed to an equivalent concentration of DMSO alone. MTT

3232
V. Fagan et al. / Bioorg. Med. Chem. 20 (2012) 3223–3232
9. Schulz, W. G.; Skibo, E. B. J. Med. Chem. 2000, 43, 629.
Supplementary data
10. Suleman, A.; Skibo, E. B. J. Med. Chem. 2002, 45, 1211.
11. Fagan, V.; Bonham, S.; Carty, M. P.; Aldabbagh, F. Org. Biomol. Chem. 2010, 8,
3149.
12. Fagan, V.; Bonham, S.; McArdle, P.; Carty, M. P.; Aldabbagh, F. Eur. J. Org. Chem.
2012, 1967.
13. Shoemaker, R. H. Nature Rev. Cancer 2006, 6, 813.
14. Bowman, W. R.; Storey, J. M. D. Chem. Soc. Rev. 2007, 36, 1803.
15. Lynch, M.; Hehir, S.; Kavanagh, P.; Leech, D.; O’Shaughnessy, J.; Carty, M. P.;
Aldabbagh, F. Chem. Eur. J. 2007, 13, 3218.
Supplementary data (supplementary data includes 1H and 13C
NMR spectra of compounds, HPLC chromatograms, cell viability
plots, and full NCI, and COMPARE analysis data) associated with
this article can be found, in the online version, at http://
dx.doi.org/10.1016/j.bmc.2012.03.063.
16. Seow, H. A.; Penketh, P. G.; Baumann, R. P.; Sartorelli, A. C. Methods Enzymol.
2004, 382, 221.
References and notes
17. Faig, M.; Bianchet, M. A.; Talalay, P.; Chen, S.; Winski, S.; Ross, D.; Amzel, L. M.
Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 3177.
18. Hassani, M.; Cai, W.; Koelsch, K. H.; Holley, D. C.; Rose, A. S.; Olang, F.;
Lineswala, J. P.; Holloway, W. G.; Gerdes, J. M.; Behforouz, M.; Beall, H. D. J.
Med. Chem. 2008, 51, 3104.
19. Molecular Operating Environment, Version 2009.10, Chemical Computing
Group: Montreal, QC, 2009.
20. Bianchet, M. A.; Faig, M.; Amzel, L. M. Methods Enzymol. 2004, 382, 144.
21. Chen, S.; Wu, K.; Zhang, D.; Sherman, M.; Knox, R.; Yang, C. S. Mol. Pharmacol.
1999, 56, 272.
22. Huang, X.; Suleman, A.; Skibo, E. B. Bioorg. Chem. 2000, 28, 324.
23. Mosmann, T. J. Immunol. Methods 1983, 65, 55.
24. Wang, J.; Cieplak, P.; Kollman, P. A. J. Comput. Chem. 2000, 21, 1049.
25. Edelsbrunner, H.; Facello, M.; Fu, R.; Liang, J. Proceedings of the 28th Hawaii
International Conference on Systems Science, 1995.
1. Li, R.; Bianchet, M. A.; Talalay, P.; Amzel, L. M. Proc. Natl. Acad. Sci. U.S.A. 1995,
92, 8846.
2. Skelly, J. V.; Sanderson, M. R.; Suter, D. A.; Baumann, U.; Read, M. A.; Gregory, D.
S. J.; Bennett, M.; Hobbs, S. M.; Neidle, S. J. Med. Chem. 1999, 42, 4325.
3. Ross, D.; Siegel, D. Methods Enzymol. 2004, 382, 115.
4. Dinkova-Kostova, A. T.; Talalay, P. Arch. Biochem. Biophys. 2010, 501, 116.
5. Siegel, D.; Gustafson, D. L.; Dehn, D. L.; Han, J. Y.; Boonchoong, P.; Berliner, L. J.;
Ross, D. Mol. Pharmacol. 2004, 65, 1238.
6. Fitzsimmons, S. A.; Workman, P.; Grever, M.; Paull, K.; Camalier, R.; Lewis, A. D.
J. Natl Cancer Inst. 1996, 88, 259.
9
9
7. Bello, R. I.; Gómez-Diaz, C.; Navarro, F.; Alcain, F. J.; Villalba, J. M. J. Biol. Chem.
2001, 276, 44379.
8. Kolesar, J. M.; Pritchard, S. C.; Kerr, K. M.; Kim, K.; Nicolson, M. C.; McLeod, H.
Int. J. Oncol. 2002, 21, 1119.