Synthesis of N-(5,7-diamino-3-phenyl-quinoxalin-2-yl)-3,4,5-substituted anilines and N-[4[(5,7-diamino-3-phenylquinoxalin-2-yl)amino]benzoyl]-l-glutamic acid diethyl ester: Evaluation of in vitro anti-cancer and anti-folate activities

Article abstract of DOI:10.1016/j.ejmech.2007.03.035

Several diamino quinoxalines were designed, synthesized and evaluated as anti-tumor agents. Two compounds showed the most potent cytotoxic activities against the leukemia CCRF-CEM cell line (GI₅₀ < 0.01 μM) and the ovarian cancer cell line OVCAR-4 (GI₅₀ = 0.03 μM), respectively, with comparable/better activities than Methotrexate (MTX). Docking calculations of the complexes of hDHFR with the most active compounds identified the binding mode of the described molecules with respect to MTX.

Full text of DOI:10.1016/j.ejmech.2007.03.035

Available online at www.sciencedirect.com
European Journal of Medicinal Chemistry 43 (2008) 189e203
http://www.elsevier.com/locate/ejmech
Preliminary communication
Synthesis of N-(5,7-diamino-3-phenyl-quinoxalin-2-yl)-3,4,5-
substituted anilines and N-[4[(5,7-diamino-3-phenylquinoxalin-
2-yl)amino]benzoyl]-L-glutamic acid diethyl ester: Evaluation
of in vitro anti-cancer and anti-folate activities
a
b
b
a
Paola Corona a, , Mario Loriga , M. Paola Costi , Stefania Ferrari , Giuseppe Paglietti
*
a Dipartimento Farmaco Chimico Tossicologico, Universita` degli Studi di Sassari, Via Muroni 23/A, 07100 Sassari, Italy
b Dipartimento di Scienze Farmaceutiche, Universita` degli Studi di Modena e Reggio Emilia, Via Campi 183, 41100 Modena, Italy
Received 1 December 2006; received in revised form 17 March 2007; accepted 26 March 2007
Available online 18 April 2007
Abstract
Several diamino quinoxalines were designed, synthesized and evaluated as anti-tumor agents. Two compounds showed the most potent
cytotoxic activities against the leukemia CCRF-CEM cell line (GI50 < 0.01 mM) and the ovarian cancer cell line OVCAR-4 (GI50 ¼ 0.03 mM),
respectively, with comparable/better activities than Methotrexate (MTX). Docking calculations of the complexes of hDHFR with the most active
compounds identified the binding mode of the described molecules with respect to MTX.
Ó 2007 Elsevier Masson SAS. All rights reserved.
Keywords: 5,7-Diaminoquinoxaline; Ovarian carcinoma; Leukemia; Anti-folate
1. Introduction
a new class of classical and non-classical anti-folate agents,
we replaced the pteridine and quinazoline nucleus present in
the most active anti-cancer agents Metotrexate (MTX), Trime-
trexate (TMQ) and Piritrexim (PTX) (Fig. 1) with the qui-
noxaline ring functionalised with a trifluoromethyl or amino
group, or both, that bioisosterically could modulate the affin-
ity for DHFR and TS enzymes. The wide range of compounds
synthesized and evaluated as anti-tumor agents at the National
Cancer Institute of Bethesda (NCI) allowed us to establish
some structureeactivity relationships, which can be summa-
rized as follows: (i) bioisosterism of quinoxaline with the
pteridine ring has shown to be profitable for the biological
activity; (ii) compounds bearing a trifluoromethyl group at po-
sition 6 or 7 of the heterocycle and a phenyl group at position
3 resulted to be more active than the unsubstituted ones; (iii)
the concomitant presence of both the trifluoromethyl group at
position 6 and amine group at position 8, whatever be the sub-
stituent present at positions 2 and 3, increases the anti-cancer
activity [3e17].
Despite continuous research efforts carried out on the dis-
covery of safer or more potent anti-cancer compounds from
a recent survey of these results [1] in the past decade only
two new agents have been introduced as anti-cancer endowed
with strong thymidilate synthase (TS) inhibitory activity
(Tomudex and Pemetrexed). However, a few quinoxaline
derivatives out of those described by us were reported as
anti-cancer agents and among these a recent patent of 7-fluoro
-4-hydrazinopyrrolo[1,2-a]quinoxalines as inhibitors of angio-
genesis has appeared [2] providing a new way to stop cancer
cell growth.
Contributions of our research group in this field have con-
sidered the design of structurally related pteridine compounds
as novel anti-cancer agents. Thus, with the aim to discover
* Corresponding author. Tel.: þ39079228764; fax: þ39079228720.
E-mail address: pcorona@uniss.it (P. Corona).
0223-5234/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2007.03.035

190
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
COOH
O
H
COOH
N
O
NH2
N
H
COOH
N
N
S
HN
H3C
N
CH3
O
N
N
CH3
COOH
N
H2N
N
Methotrexate
Raltitrexed
O
OCH3
O
COOH
NH2 CH3
HN
N
N
COOH
N
H
H3CO
H2N
N
H2N
N
N
Pemetrexed
Piritrexim
OCH3
OCH3
OCH3
NH2 CH3
N
N
H
H2N
N
Trimetrexate
Fig. 1. Chemical structure of some classical and non-classical anti-folates.
In our previous work, we have described compounds 1a and
reproduce the most bioisosterical substitution with the pyrim-
idine ring of the classical anti-folates MTX, TMQ and PTX.
At the same time, this replacement might also selectively
improve their affinity towards the human DHFR (hDHFR) en-
zyme [18e22] and maintain the same binding mode as MTX
in their X-ray crystal structure with hDHFR. With this in
mind, in this paper we have designed compounds 2aeg
(Fig. 3) for an evaluation of their anti-cancer activity and com-
parison of their properties with those of the previously inves-
tigated quinoxalines 1aei [3,5]. In addition the enzyme
inhibition activity against human TS (hTS) and DHFR was de-
termined to evaluate their sub-cellular selectivity and potency.
Along with the diamino compounds (2aeg), the 5,7-dinitro in-
termediates (11e17), were also tentatively subjected to enzy-
matic assay against three different species of TS such as,
Escherichia coli TS (EcTS), Lactobacillus casei TS, and hTS
and the results are presented here.
At the end we have afforded a preliminary investigation of
molecular modelling on compounds 2aed, which were se-
lected, based on their inhibitory activity against hDHFR en-
zyme. Their binding mode to the target enzyme was studied
using the molecular docking program Autodock 3.05 and the
interesting results suggest further X-ray crystal structure stud-
ies with hDHFR.
its analogues 1bei (Fig. 2) in which a phenyl ring was added
at position 3 of the heterocycle as a lypophilic hindering group
[3,5]. The screening performed at NCI has shown that many of
the quinoxaline derivatives exhibited promising in vitro total
growth inhibition and compound 1a (NSC 637404) (Fig. 2)
has been selected for in vivo screening [3]. Nevertheless, a cer-
tain number of the quinoxalines derivatives of classical and
non-classical type have shown anti-folate activity against
both murine and human DHFR [9,13].
Now, we extended our investigation taking into account that
phenyl ring substitution at position 3 yielded the best result for
in vitro anti-cancer activity and introduction of two amino
groups at 5 and 7 positions of the quinoxaline ring might
R1
OCH3
N
N
Ph
R2
R3
F3C
N
N
Ph
OCH3
R
N
H
N
H
OCH3
NH2
1b R = 6-CF3, 8-NH2; R1 = R2= OCH3; R3 = H
1c R = 6-CF3, 8-NH2; R1 = R3 = H R3 = OCH3
1d R = 6-CF3; R1 = R2 = R3 = OCH3
1a
1e R = 7-CF3; R1 = R2 = R3 = OCH3
1f R = 6-CF3; R1 = R2 = OCH3; R3 = H
1g R = 7-CF3; R1 = R2 = OCH3; R3 = H
1h R = 6-CF3; R1 = R2 = Cl; R3 = H
2. Results and discussion
O
COOEt
N
N
Ph
2.1. Chemistry
N
H
COOEt
F3C
N
H
The preparation of the compounds 2aeg was accomplished
according to the reactions described in Scheme 1. 5,7-Dinitro-
3-phenyl-2-chloroquinoxaline (3) underwent nucleophilic at-
tack by the corresponding anilines 4e10 given in Scheme 1
1i
Fig. 2. First series of quinoxaline non-classical anti-folates [3,5].

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
NH2 NH2
191
R1
O
COOEt
N
Ph
R2
N
Ph
N
H
COOE
t
H2N
N
N
H
R3
H2N
N
N
H
2a-f
2g
Compd
2a
R1
R2
OCH3
H
R3
H
H
OCH3
H
OCH3
OCH3 OCH3
OCH3 OCH3
2b
2c
OCH3
H
2d
2e
Cl
H
Cl
F
2f
H
Fig. 3. Designed compounds 2aef and 2g.
in refluxing 1-propanol to give the 5,7-dinitro-3-phenyl-2-ani-
lino substituted quinoxalines 11e17. In the case of ester 17,
ethanol was alternatively used and heating at 80 ꢀC was pro-
longed for 186 h. The desired 5,7-diamino-3-phenylquinoxa-
lines 2aeg were obtained by reduction with an excess of
hydrazine hydrate in ethanol and in the presence of 10% pal-
ladised charcoal.
5,7-Dinitro-3-phenyl-2-chloroquinoxaline (3) is a new
compound and has been prepared according to the reactions
outlined in Scheme 2. Picric acid (18) was converted with
POCl3 in the presence of N,N-dimethylaniline into the 2-
chloro-1,3,5-trinitrobenzene (19), which, once heated in
a sealed vessel at 150 ꢀC with a solution of ammonia saturated
with propanol, yielded the 2,4,6-trinitroanilina (20). Selective
reduction with H2S and diluted ammonia gave the 2-amino-
3,5-dinitrophenylamine (21) then cyclised regioselectively
into the isomer 5,7-dinitro-quinoxalin-2(1H )one (22) (yield
80%) with benzoyl formic acid (23) under acidic catalysis
[23e26]. A small amount (yield 2%) of the isomer 6,8-dini-
tro-quinoxalin-2(1H )one (24) was also obtained. This result
was in good accordance with what some of us reported in a pre-
vious paper where it was demonstrated that when the two
amino groups have different basicities, the ratio of the iso-
meric quinoxalinones is dependent on both the nature and
para/meta position of the substituent. The more basic amino
group undergoes protonation [26]. Unambiguous structure to
the pairs of isomers 22 and 24 was assigned on the basis of
NOE difference experiments carried out on isomer 22. In
fact, only in this case the proton at position 8 of 5,7-dinitro-
3-phenyl quinoxaline-2(1H )one (22) experienced NOE inter-
actions, greatly enhancing the resonance located at d 8.25 by
coupling with the amidic proton at position 1 (Fig. 4), whereas
for isomer 24 no NOE interactions between aromatic protons
and the amidic proton are possible and were not observed.
R1
R2
R3
Compd
11
OCH3
H
OCH3
H
H
OCH3
H
OCH3
OCH3
OCH3
Cl
12
R1
NO2
N
N
Ph
13
14
15
16
17
R2
R3
OCH3
OCH3
H
O2N
N
H
Cl
F
i
H
H
H
H
R1
11-17
NO2
CO-GlU-Et
ii for 17
R2
N
N
Ph
Cl
H2N
+
O2N
R3
iii
4-10
3
R1
R2
R3
Anilines
OCH3
H
OCH3
H
H
OCH3
H
4
R1
NH2
OCH3
OCH3
OCH3
Cl
5
R2
R3
N
N
Ph
6
7
OCH3
OCH3
H
H2N
N
H
Cl
F
8
H
H
H
H
9
2a-g
CO-GlU-Et
10
Scheme 1. Synthesis of 2aeg. Reagents and conditions: (i) 1-propanol, reflux for 1 he72 h. (ii) EtOH, reflux for 186 h. (iii) NH2eNH2/H2O, Pd/C, EtOH reflux for
10 mine2 h and 45 min.

192
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
NH2
Cl
O
NH2
O2N
NH2
O2N
NO2
OH
O2N
NO2
ii
iii
+
O
NO2
21
NO2
NO2
23
19
20
iv
OH
O2N
NO2
NO2
H
N
NO2
O
NO2
18
N
+
O2N
N
O2N
N
H
O
24
22
v
NO2
N
N
O2N
Cl
3
Scheme 2. Synthesis of 3. Reagents and conditions: (i) POCl3, N,N-dimethylaniline, room temperature for 15 min, (ii) 1-propanol saturated with gaseous NH3,
100 ꢀC for 3 h, (iii) EtOH, NH4OH, H2S, 45e50 ꢀC, 30 min, (iv) EtOH, HCl 4 N, reflux for 1 h 50 min, (v) POCl3, DMF anydrous, reflux for 2 h.
A further confirmation of this assignment came from the
1
examination of H NMR spectrum which showed downfield
shift by 0.67 of the proton peri to the amidic group in isomer
22 [27e30]. Eventually the 5,7-dinitroquinoxalin-2(1H )one
(22) was converted into the chloroquinoxaline (3) by treatment
with POCl3.
(TGI ¼ molar concentration of the compound leading to total
inhibition of net cell growth), and log10 LC50 value (LC50
¼
molar concentration of the compound leading to 50% net
cell death). Furthermore, a mean graph midpoint (MG_MID)
is calculated for each of the mentioned parameters, giving
an averaged activity parameter over all cell lines. The anti-
cancer activity of each compound is deduced from dosee
response curves.
2.2. Anti-cancer activity
The log10 GI50, log10 TGI and log10 LC50 values provided by
NCI for compounds 2aeg are reported in Tables 1 and 2. The
data showed that all the compounds exhibited moderate activity
against one or more human tumor cell lines. Compounds 2b and
2d have showed the most interesting values of cytotoxicity
(log10 GI50) against leukemia CCRCEM and ovarian cancer
OVCAR-4 cell lines, respectively. Moreover, different cancer
cell lines of the same tumor type possessed a variable response
to inhibition of growth in the presence of the new derivatives.
For example, the OVCAR-4 ovarian cancer cells were suscep-
tible to inhibition by 2d (log10 GI50 ¼ ꢁ7.50, log10 TGI ¼
ꢁ6.05, log10 LC50 ¼ ꢁ4.59), whereas other ovarian cancer
cell lines (such as IGROV1, OVCAR-3, OVCAR-5) showed
a much lower level of inhibition (log10 GI50 ranged from
ꢁ4.85 to ꢁ4.64). A similar situation has been highlighted in
the case of diverse leukemia cell lines, with compound 2b act-
ing selectively as a potent inhibitor (log10 GI50 < ꢁ8.00,
log10 TGI ¼ ꢁ6.78, log10 LC50 ¼ ꢁ6.01) against the CCRF-
CEM line, whereas other leukemia cell lines such as HL-60
(TB), K-562 and SR did not reach comparable TGI and LC50
levels. The RPMI-8226 cell line of this tumor type exhibited
a lower level of inhibition (log10 GI50 ¼ ꢁ5.90, log10 TGI ¼
ꢁ4.98, log10 LC50 ¼ ꢁ4.49).
Evaluation of anti-cancer activity of the diamino quinoxa-
lines synthesized (2aeg) was performed at the NCI. First,
all diamino quinoxalines were evaluated in primary anti-can-
cer assay at 10ꢁ4 M concentration against NCI e H460
(lung), MCF7 (breast), and SF-268 (CNS) cell lines. For
NCI criteria, compounds which reduce the growth of any
one of the cell lines to approximately 32% or less are passed
on for evaluation in the full panel of cell lines over a 5-log
dose range. Quinoxalines 2aeg which meet these criteria
were evaluated for their anti-cancer activity following the
known in vitro disease-oriented anti-tumor screening program
which is based upon use of a multiple panels of 60 human tu-
mor cell lines [31e33]. The anti-cancer activity of a tested
compound is given by three parameters for each cell line:
log10 GI50 value (GI50 ¼ molar concentration of the compound
that inhibits 50% net cell growth), log10 TGI value
NO2
N
O2N
N
H
O
1
8
H
In Table 3, we showed the mean graph midpoint values of
log10 GI50, log TGI50 and log10 LC50 for compounds 2aeg in
comparison with MTX. From these data, two compounds (2b
and 2e) emerged with an average log10 GI50 < ꢁ5.0.
5,7-dinitro-quinoxalin-2(1H)one (22)
Fig. 4. NOE correlations for compound 22.

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
193
Table 1
Inhibition of in vitro cancer lines by selected 5,7-diamino-3-phenyl-2-anilino-quinoxalines, 2aeda
b
d
Panel cell line
Response parameters: (A) log GI50 [M], (B) log TGIc [M], (C) log LC50 [M]
Compound 2a
Compound 2b
Compound 2c
Compound 2d
A
B
C
A
B
C
A
B
C
A
B
C
Leukemia
CCRF-CEM
HL-60(TB)
K-562
e
e
e
e
e
e
e
e
e
e
e
ꢁ4.92
ꢁ4.46
ꢁ4.53
ꢁ4.72
ꢁ4.72
e
<ꢁ8.00
ꢁ5.00
ꢁ4.56
ꢁ5.90
ꢁ5.25
e
ꢁ6.78
ꢁ6.01
e
e
e
e
ꢁ5.36
ꢁ5.18
ꢁ4.90
ꢁ4.91
ꢁ5.48
ꢁ5.59
ꢁ4.38
ꢁ4.63
ꢁ4.40
ꢁ4.49
ꢁ4.80
ꢁ4.74
e
e
e
e
ꢁ4.19
e
e
e
e
ꢁ4.69
ꢁ5.10
ꢁ5.57
ꢁ5.16
ꢁ4.15
ꢁ4.55
RPMI-8226
SR
MOLT-4
ꢁ4.98
ꢁ4.49
ꢁ4.06
ꢁ4.07
ꢁ4.14
ꢁ4.25
e
e
e
ꢁ4.66
e
e
e
e
e
e
Non-small cell lung cancer
A549/ATCC
EKVX
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
e
ꢁ4.56
ꢁ4.59
ꢁ4.56
ꢁ4.70
ꢁ4.62
ꢁ4.90
ꢁ4.63
ꢁ4.57
ꢁ4.83
ꢁ4.70
ꢁ4.88
ꢁ4.71
e
ꢁ4.45
ꢁ4.59
ꢁ4.46
e
ꢁ4.53
ꢁ4.75
ꢁ4.79
ꢁ5.56
e
e
ꢁ4.19
ꢁ4.27
e
ꢁ4.15
ꢁ4.32
ꢁ4.46
ꢁ4.68
e
e
HOP-62
HOP-92
ꢁ4.14
ꢁ4.04
ꢁ4.28
ꢁ4.01
ꢁ4.60
ꢁ4.08
ꢁ4.05
ꢁ4.55
e
e
e
NCI-H226
NCI-H23
NCI-H322M
NCI-H460
NCI-H522
Colon cancer
COLO 205
HCC-2998
HCT-116
HCT-15
ꢁ5.54
ꢁ4.56
ꢁ4.86
ꢁ5.10
ꢁ5.02
ꢁ4.48
ꢁ4.73
ꢁ4.58
ꢁ4.60
ꢁ4.69
ꢁ4.54
ꢁ4.24
e
e
e
e
e
e
e
ꢁ4.29
ꢁ4.10
ꢁ4.60
ꢁ4.61
ꢁ4.77
ꢁ4.74
ꢁ4.12
ꢁ4.10
ꢁ4.28
ꢁ4.27
e
e
e
e
e
ꢁ4.23
ꢁ4.28
ꢁ4.54
e
e
e
ꢁ4.28
ꢁ4.07
ꢁ4.80
ꢁ4.55
ꢁ4.83
ꢁ4.82
ꢁ4.62
ꢁ4.80
ꢁ4.65
ꢁ4.53
ꢁ4.27
ꢁ4.84
ꢁ4.61
ꢁ4.77
ꢁ4.70
ꢁ4.56
ꢁ4.88
ꢁ4.57
ꢁ4.50
ꢁ4.17
ꢁ4.70
ꢁ4.74
ꢁ4.83
ꢁ4.54
ꢁ4.60
ꢁ4.83
ꢁ4.81
ꢁ4.39
ꢁ4.40
ꢁ4.07
ꢁ4.07
ꢁ4.85
e
ꢁ4.57
e
ꢁ4.28
e
e
e
e
e
e
e
e
ꢁ4.55
ꢁ4.35
ꢁ4.37
ꢁ4.51
4.43
ꢁ4.28
ꢁ4.27
ꢁ4.53
ꢁ4.23
ꢁ4.81
ꢁ4.82
ꢁ4.48
ꢁ4.97
ꢁ4.62
ꢁ4.47
ꢁ4.34
ꢁ4.46
ꢁ4.12
e
e
e
e
e
e
e
HT29
KM12
SW-620
ꢁ4.05
ꢁ4.22
ꢁ4.22
ꢁ4.28
ꢁ4.49
ꢁ4.44
ꢁ4.13
ꢁ4.51
ꢁ4.14
ꢁ4.16
ꢁ4.08
ꢁ4.53
ꢁ4.09
e
e
e
e
CNS Cancer
SF-268
SF-295
e
e
e
e
e
ꢁ4.89
ꢁ4.50
e
ꢁ4.47
ꢁ4.84
ꢁ4.79
ꢁ4.51
ꢁ4.14
ꢁ5.25
ꢁ4.90
e
ꢁ4.24
ꢁ5.43
ꢁ4.64
ꢁ4.25
ꢁ4.23
ꢁ4.85
ꢁ4.77
ꢁ4.79
ꢁ4.45
e
ꢁ4.33
ꢁ4.34
ꢁ4.74
ꢁ4.80
ꢁ4.83
ꢁ4.48
e
ꢁ4.22
ꢁ4.37
e
e
e
SF-539
SNB-19
e
e
e
e
ꢁ4.43
ꢁ4.06
ꢁ4.54
ꢁ4.25
e
e
e
e
e
e
e
e
e
e
e
SNB-75
U251
ꢁ4.37
ꢁ4.53
ꢁ4.35
e
e
e
e
e
e
e
e
e
ꢁ4.26
ꢁ4.59
ꢁ4.61
Melanoma
LOX IMVI
MALME-3M
M14
e
e
e
e
ꢁ4.79
ꢁ4.82
ꢁ4.75
ꢁ4.81
ꢁ4.77
ꢁ4.82
ꢁ4.84
ꢁ4.77
ꢁ4.53
ꢁ4.55
ꢁ4.44
ꢁ4.50
ꢁ4.52
ꢁ4.54
ꢁ4.56
ꢁ4.52
ꢁ4.26
ꢁ4.27
ꢁ4.12
ꢁ4.20
ꢁ4.26
ꢁ4.27
ꢁ4.28
ꢁ4.26
ꢁ4.82
ꢁ5.36
ꢁ4.85
ꢁ4.92
ꢁ4.83
ꢁ4.98
ꢁ4.80
ꢁ5.13
ꢁ4.30
ꢁ4.61
ꢁ4.46
ꢁ4.20
ꢁ4.05
ꢁ4.52
ꢁ4.32
ꢁ4.63
ꢁ4.59
ꢁ4.80
ꢁ4.68
ꢁ4.57
ꢁ4.67
ꢁ4.91
ꢁ4.45
ꢁ4.87
ꢁ4.14
ꢁ4.45
ꢁ4.73
ꢁ4.89
ꢁ4.71
ꢁ4.80
ꢁ5.84
ꢁ4.89
ꢁ4.53
ꢁ4.87
ꢁ4.31
ꢁ4.55
ꢁ4.35
ꢁ4.45
ꢁ4.68
ꢁ4.09
ꢁ4.21
e
e
ꢁ4.07
ꢁ4.16
e
e
e
SK-MEL-2
SK-MEL-28
SK-MEL-5
UACC-257
UACC-62
Ovarian cancer
IGROV1
OVCAR-3
OVCAR-4
OVCAR-5
OVCAR-8
SK-OV-3
Renal cancer
786-0
ꢁ4.11
ꢁ4.20
e
e
ꢁ4.30
ꢁ4.06
ꢁ4.59
ꢁ4.27
ꢁ4.56
ꢁ4.23
e
e
e
e
e
ꢁ4.21
ꢁ4.56
ꢁ4.25
ꢁ4.50
ꢁ4.13
e
e
e
e
e
e
e
e
e
ꢁ4.81
ꢁ4.76
ꢁ4.98
ꢁ4.67
ꢁ4.86
ꢁ4.48
ꢁ4.45
ꢁ4.43
ꢁ4.66
ꢁ4.27
ꢁ4.51
ꢁ4.01
ꢁ4.08
ꢁ4.10
ꢁ5.57
ꢁ5.41
ꢁ4.75
ꢁ4.58
ꢁ5.21
ꢁ5.18
ꢁ4.77
ꢁ4.61
ꢁ4.52
ꢁ4.73
ꢁ4.53
ꢁ4.47
ꢁ4.42
ꢁ4.80
ꢁ4.85
ꢁ7.50
ꢁ4.67
ꢁ4.81
ꢁ4.64
ꢁ4.24
ꢁ4.43
ꢁ6.05
ꢁ4.33
ꢁ4.46
ꢁ4.24
e
ꢁ4.70
ꢁ4.00
ꢁ4.01
e
e
ꢁ4.33
ꢁ4.28
ꢁ4.59
e
e
e
e
e
e
ꢁ4.14
e
ꢁ4.17
ꢁ4.33
ꢁ4.40
ꢁ4.10
e
e
ꢁ4.06
ꢁ4.04
e
e
e
e
e
e
e
e
e
e
e
e
e
ꢁ4.80
ꢁ4.51
ꢁ4.22
ꢁ4.75
ꢁ5.31
ꢁ5.25
ꢁ4.83
ꢁ5.50
ꢁ5.06
ꢁ4.88
ꢁ5.05
ꢁ4.29
ꢁ4.65
ꢁ4.08
ꢁ4.14
ꢁ4.62
ꢁ4.93
ꢁ4.56
ꢁ4.74
ꢁ5.09
ꢁ4.78
ꢁ4.68
ꢁ4.55
ꢁ4.84
ꢁ4.50
ꢁ4.78
ꢁ4.81
ꢁ4.84
ꢁ4.69
ꢁ4.62
ꢁ4.82
ꢁ4.39
e
e
e
A498
ACHN
CAKI-1
ꢁ4.10
ꢁ4.25
e
e
e
ꢁ4.82
ꢁ4.45
ꢁ4.96
ꢁ4.69
ꢁ4.78
ꢁ4.78
ꢁ4.36
ꢁ4.37
ꢁ4.31
ꢁ4.46
ꢁ4.08
ꢁ4.18
ꢁ4.04
e
e
e
ꢁ4.18
ꢁ4.31
ꢁ4.36
RXF 393
SN12C
TK-10
ꢁ4.62
ꢁ4.24
ꢁ4.37
ꢁ4.47
ꢁ4.27
ꢁ4.73
ꢁ4.02
ꢁ4.08
e
e
e
e
e
e
e
ꢁ4.37
ꢁ4.35
ꢁ4.23
e
e
e
UO-31
ꢁ4.15
Prostate cancer
PC-3
DU-145
e
e
e
ꢁ4.68
ꢁ4.41
ꢁ4.32
ꢁ5.71
ꢁ4.64
ꢁ4.96
ꢁ4.24
ꢁ4.85
ꢁ4.35
ꢁ4.44
ꢁ4.03
ꢁ4.85
ꢁ4.66
ꢁ4.23
ꢁ4.13
e
e
e
e
e
e
(continued on next page)

194
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
Table 1 (continued)
Response parameters: (A) log GI50 [M], (B) log TGIc [M], (C) log LC50 [M]
Compound 2a Compound 2b Compound 2c
b
d
Panel cell line
Compound 2d
A
B
C
A
B
C
A
B
C
A
B
C
Breast cancer
MCF7
e
e
ꢁ4.83
ꢁ4.72
ꢁ4.94
ꢁ4.84
ꢁ4.78
ꢁ4.67
ꢁ4.87
ꢁ4.75
ꢁ4.55
ꢁ4.35
ꢁ4.60
ꢁ4.33
ꢁ4.52
ꢁ4.44
ꢁ4.50
ꢁ4.27
ꢁ4.28
ꢁ5.10
ꢁ4.89
ꢁ5.19
ꢁ4.96
ꢁ4.94
ꢁ4.89
ꢁ5.33
ꢁ5.35
ꢁ4.57
ꢁ4.13
ꢁ4.46
ꢁ4.20
ꢁ4.62
ꢁ4.44
ꢁ4.50
ꢁ4.64
ꢁ4.07
ꢁ4.69
ꢁ4.76
ꢁ5.31
ꢁ4.73
ꢁ4.79
e
ꢁ4.35
ꢁ4.37
ꢁ4.73
ꢁ4.30
ꢁ4.49
e
ꢁ4.01
ꢁ4.59
ꢁ4.84
ꢁ4.95
ꢁ4.82
ꢁ5.13
ꢁ4.98
ꢁ4.77
ꢁ5.22
ꢁ4.02
ꢁ4.36
ꢁ4.53
ꢁ4.11
ꢁ4.61
ꢁ4.50
ꢁ4.47
ꢁ4.56
e
e
e
NCI/ADR-RES
MDA-MB-231/ATCC
HS 578T
e
e
ꢁ4.26
ꢁ4.33
ꢁ4.11
e
e
e
MDA-MB-435
MDA-N
BT-549
ꢁ4.26
ꢁ4.22
ꢁ4.31
ꢁ4.20
ꢁ4.16
ꢁ4.03
ꢁ4.17
ꢁ4.02
ꢁ4.00
e
e
e
ꢁ4.12
ꢁ4.59
ꢁ4.84
ꢁ4.09
ꢁ4.42
e
e
e
T-47D
eCell line not tested.
a
Data obtained from the NCI’s in vitro disease-oriented human tumor cells screen (see Refs. [19e21]).
The log of the molar concentration that inhibits 50% net cell growth.
The log of the molar concentration giving total growth inhibition.
The log of the molar concentration leading to 50% net cell death.
The values of log GI50, log TGI or log LC50 >ꢁ4.00.
b
c
d
e
In Table 4, we represented the average of log10 GI50 for each
Compounds 2a, 2b and 2d are active against hDHFR with
Ki in the range 0.1e1.8 mM. The best compound is 2d. Com-
pounds 2b and 2d are selective towards hDHFR with respect
to hTS with selectivity indices (Ki hTS/Ki hDHFR) of 42
and 510, respectively. Compounds 2c, 2eeg are not active at
10 mM. Compound 2e is the best inhibitor towards hTS with
Ki of 0.9 mM, while 2a, 2b, 2d and 2g are moderate hTS
inhibitors showing Ki ranging between 6 and 51 mM. Com-
pounds 2e and 2g are selective towards hTS with respect to
hDHFR. Compounds 2c and 2f are not active at 10 mM.
From these data it appears that the methoxy groups contrib-
ute to the increasing affinity towards hDHFR (2a, 2b, 2d) with
respect to hTS. The substituition of the methoxy group with
the chlorine (2e) or fluorine (2f) increases the affinity for
hTS (2e), but loses it’s affinity for hDHFR.
Interestingly, among the 5,7-dinitro intermediates (11e17),
which were considered as precursors of the diamino deriva-
tives, compound 11 was by far the best inhibitor showing quite
the same Ki as 2e (Ki hTS ¼ 0.7 mM). Other compounds are
also active with Ki in the range of 2.1e36 mM.
Dinitro intermediates (11e15, 17) are more active with
respect to the corresponding diamino compounds (2aee, 2g)
except for 2e and 15. However, along these two series the
same trend is not observed. While the dichloro derivative
(2e) is the most active compound in the diamino series (Ki
hTS ¼ 0.9 mM), the dichloro dinitro intermediate (15) is the
worst compound of the dinitro series (Ki hTS ¼ 36 mM). Anal-
ogously while the dimethoxy derivatives in the dinitro series
(12, 13) show the same activity towards hTS (Ki hTS ¼ 14e15),
the two dimethoxy derivatives in the diamino series (2b,
2c) show very different Ki towards hTS (Ki hTS ¼ 25 and
no inhibition at 10 mM).
cancer type to put in evidence both activity and selectivity. All
but compound 2a showed particular selectivity against leukemia
cell lines. Best results were obtained with compounds 2b and
2e against leukemia cell lines (log10 GI50 ¼ 5.74 and 5.60,
respectively) whereas compound 2e recorded a MG_MID
log10 GI50 ¼ ꢁ5.08, which accounts for the highest micromolar
activity against all cell lines.
In Table 5, we compared the activities of the most active
compounds 2b and 2d, on leukemia (CCRF-CEM) and ovarian
(OVCAR-4) cell lines, respectively, with that recorded for
MTX. Compound 2b showed an anti-cancer activity against
leukemia (CCRF-CEM) better than that recorded for MTX
which is taken from Ref. [34].
The results thus obtained allowed some considerations
about structure and in vitro tumor inhibitory activity relation-
ships. It is evident that all but compound 2a are, on average,
very active against leukemia cell lines. When two methoxy
groups are present in the side chain, the best activity was re-
corded for the 30,40-disubstituted compound (2b), while in
the case of three methoxy groups (2d) the increase of lipophi-
licity produce an unfavourable effect on log10 GI50. Replace-
ment of the two methoxy groups in the anilino moiety (2b)
with chlorine (2e) seems to increase the anti-cancer activity
on the whole. Fluorine in 40 position slightly increases the ac-
tivity only on comparison with that of 2a and 2c.
When we compare the log10 GI50 of this series with that of
the corresponding previously described compounds that bear
the trifluoromethyl group in quinoxaline benzene counterpart
(Table 6), it is worth noting that in all cases the increase of po-
tency is due to the presence of the two amino groups [3,5].
In regard to the activity towards bacterial TS, compound 15
is the best with a Ki of 0.2 mM towards EcTS and being spe-
cies-specific with respect to hTS with a specificity index (Ki
hTS/Ki EcTS) of 180. Also, compounds 2e, 13 and 17 show
good activity towards EcTS (Ki EcTS ¼ 1e3 mM), but they
are not species-specific inhibiting hTS (Ki hTS ¼ 0.9e16).
2.3. Anti-folate activity
In Table 7, the enzyme inhibition activity profiles for com-
pounds 2aeg, 11e15, 17 towards hDHFR and a panel of three
TS enzymes are reported.

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
195
Table 2
Inhibition of in vitro cancer lines by selected 5,7-diamino-3-phenyl-2-anilino-quinoxalines 2eega
b
d
Panel cell line
Response parameters: (A) log GI50 [M], (B) log TGIc [M], (C) log LC50 [M]
Compound 2e
Compound 2f
Compound 2g
A
B
C
A
B
C
A
B
C
Leukemia
CCRF-CEM
HL-60(TB)
K-562
e
e
e
e
e
e
e
e
e
ꢁ5.89
ꢁ5.59
ꢁ5.45
ꢁ5.39
ꢁ5.70
ꢁ5.59
e
e
e
e
ꢁ5.53
ꢁ5.02
e
e
e
e
e
e
e
e
ꢁ4.14
ꢁ4.95
ꢁ4.29
e
ꢁ5.30
e
e
e
e
e
e
e
MOLT-4
RPMI-8226
SR
ꢁ5.62
ꢁ5.50
e
ꢁ5.13
e
ꢁ5.31
ꢁ4.81
ꢁ5.33
ꢁ5.27
ꢁ4.46
e
e
e
Non-small cell lung cancer
A549/ATCC
EKVX
e
e
e
e
e
e
ꢁ4.97
ꢁ5.42
ꢁ4.76
e
ꢁ4.46
ꢁ4.79
ꢁ4.42
e
ꢁ4.76
ꢁ4.92
ꢁ4.42
ꢁ4.88
ꢁ4.75
ꢁ4.65
ꢁ4.77
ꢁ4.35
ꢁ4.70
ꢁ4.54
ꢁ4.71
ꢁ4.97
ꢁ4.57
ꢁ4.73
ꢁ4.38
ꢁ5.07
ꢁ4.56
ꢁ4.32
ꢁ4.07
e
ꢁ4.52
ꢁ4.12
e
e
HOP-62
HOP-92
ꢁ4.40
ꢁ4.53
ꢁ4.26
ꢁ4.09
e
e
e
ꢁ4.36
ꢁ4.32
ꢁ4.31
ꢁ4.09
e
NCI-H226
NCI-H23
NCI-H322M
NCI-H460
NCI-H522
Colon cancer
COLO 205
HCC-2998
HCT-116
HCT-15
ꢁ4.76
ꢁ4.72
ꢁ4.84
ꢁ5.80
ꢁ5.28
ꢁ4.42
ꢁ4.26
ꢁ4.52
ꢁ4.64
ꢁ4.66
ꢁ4.09
e
e
e
e
e
ꢁ4.26
e
ꢁ4.20
ꢁ4.25
ꢁ4.16
ꢁ4.45
ꢁ4.13
e
e
ꢁ4.58
ꢁ4.32
e
e
e
e
e
e
e
ꢁ4.93
e
ꢁ4.50
ꢁ4.07
ꢁ4.69
e
ꢁ4.25
ꢁ5.05
e
e
e
e
e
e
e
e
e
e
e
ꢁ4.97
ꢁ4.67
e
ꢁ4.28
ꢁ4.70
ꢁ4.90
ꢁ4.73
ꢁ4.92
ꢁ4.87
ꢁ4.82
ꢁ4.63
e
ꢁ4.29
e
e
e
e
e
ꢁ4.42
ꢁ4.45
ꢁ4.53
ꢁ4.41
HT29
KM12
SW-620
e
e
ꢁ4.17
e
e
e
e
ꢁ5.23
ꢁ4.88
ꢁ4.61
ꢁ4.29
ꢁ4.06
ꢁ4.14
ꢁ4.55
ꢁ4.39
e
e
e
e
CNS cancer
SF-268
SF-295
e
e
e
e
e
e
ꢁ4.92
ꢁ4.91
ꢁ4.77
ꢁ4.89
e
ꢁ4.54
ꢁ4.55
ꢁ4.45
ꢁ4.52
ꢁ4.17
ꢁ4.20
ꢁ4.14
ꢁ4.15
ꢁ4.88
ꢁ4.84
ꢁ4.96
ꢁ4.66
ꢁ4.68
ꢁ5.25
ꢁ4.39
ꢁ4.49
ꢁ4.61
ꢁ4.40
ꢁ4.38
ꢁ4.70
ꢁ4.63
ꢁ4.75
ꢁ4.56
ꢁ4.67
e
ꢁ4.14
ꢁ4.26
ꢁ4.14
ꢁ4.07
ꢁ4.30
ꢁ4.36
e
SF-539
SNB-19
e
SNB-75
U251
e
e
e
e
e
e
ꢁ5.33
ꢁ4.70
ꢁ4.36
ꢁ4.02
Melanoma
LOX IMVI
MALME-3M
M14
e
e
e
e
ꢁ5.36
ꢁ4.99
ꢁ4.83
ꢁ5.63
ꢁ4.79
ꢁ4.96
ꢁ4.94
ꢁ4.88
ꢁ5.32
e
ꢁ4.74
ꢁ4.31
ꢁ5.07
ꢁ5.55
ꢁ5.60
ꢁ5.11
ꢁ5.82
ꢁ5.18
ꢁ5.70
ꢁ5.45
e
ꢁ4.61
ꢁ4.22
ꢁ4.72
ꢁ4.35
ꢁ4.59
ꢁ4.55
ꢁ4.52
ꢁ4.24
e
e
ꢁ5.21
ꢁ5.23
ꢁ4.13
ꢁ5.52
ꢁ4.50
ꢁ5.31
ꢁ4.84
e
e
ꢁ4.79
ꢁ4.53
ꢁ4.69
ꢁ4.84
ꢁ4.82
ꢁ4.79
ꢁ4.36
ꢁ4.12
e
e
e
SK-MEL-2
SK-MEL-28
SK-MEL-5
UACC-257
UACC-62
Ovarian cancer
IGROV1
OVCAR-3
OVCAR-4
OVCAR-5
OVCAR-8
SK-OV-3
Renal cancer
786-0
ꢁ4.16
e
ꢁ4.41
ꢁ4.48
ꢁ4.54
ꢁ4.50
ꢁ4.12
ꢁ4.12
ꢁ4.26
ꢁ4.20
ꢁ5.23
e
ꢁ4.23
ꢁ4.16
ꢁ4.17
e
ꢁ4.19
e
e
e
e
e
e
e
e
ꢁ5.31
ꢁ5.53
ꢁ4.95
ꢁ4.55
ꢁ5.29
ꢁ4.86
ꢁ4.68
ꢁ5.00
ꢁ4.32
ꢁ4.31
ꢁ4.10
ꢁ4.50
ꢁ4.18
ꢁ4.78
ꢁ5.33
ꢁ4.87
ꢁ4.42
ꢁ4.91
ꢁ4.39
ꢁ5.20
ꢁ5.07
ꢁ4.57
ꢁ4.46
ꢁ4.47
ꢁ4.66
ꢁ4.05
ꢁ4.43
ꢁ4.75
ꢁ4.29
ꢁ4.50
e
e
e
ꢁ4.31
e
e
ꢁ4.07
ꢁ4.06
e
e
ꢁ4.52
ꢁ4.13
e
e
ꢁ4.15
ꢁ4.32
e
ꢁ4.83
ꢁ4.92
e
ꢁ4.43
ꢁ4.58
e
ꢁ4.03
ꢁ4.24
e
ꢁ4.76
ꢁ4.76
ꢁ4.99
ꢁ4.72
ꢁ5.06
ꢁ4.80
ꢁ4.76
ꢁ4.77
ꢁ4.27
ꢁ4.7
ꢁ4.75
e
ꢁ4.89
ꢁ4.56
ꢁ4.61
e
ꢁ4.48
ꢁ4.66
ꢁ4.43
ꢁ4.11
A498
ACHN
CAKI-1
ꢁ4.18
ꢁ4.31
ꢁ4.07
ꢁ4.07
ꢁ4.13
e
e
e
ꢁ4.65
ꢁ4.39
ꢁ4.54
ꢁ4.47
ꢁ4.45
ꢁ4.21
ꢁ4.33
e
e
ꢁ4.87
ꢁ5.21
ꢁ4.90
ꢁ4.86
ꢁ4.95
ꢁ4.53
ꢁ4.51
ꢁ4.50
ꢁ4.52
ꢁ4.44
ꢁ4.19
e
RXF 393
SN12C
TK-10
ꢁ4.07
e
ꢁ4.12
ꢁ4.12
e
ꢁ4.10
e
e
ꢁ4.18
ꢁ4.13
e
e
e
UO-31
Prostate cancer
PC-3
DU-145
e
e
e
e
ꢁ5.51
ꢁ4.78
ꢁ4.82
ꢁ4.49
ꢁ4.05
ꢁ4.20
ꢁ4.93
ꢁ4.92
ꢁ4.67
ꢁ4.35
ꢁ4.04
e
ꢁ4.55
ꢁ4.18
(continued on next page)

196
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
Table 2 (continued)
Response parameters: (A) log GI50 [M], (B) log TGIc [M], (C) log LC50 [M]
b
d
Panel cell line
Compound 2e
Compound 2f
Compound 2g
A
B
C
A
B
C
A
B
C
Breast cancer
MCF7
e
e
e
e
e
e
ꢁ5.30
ꢁ4.81
ꢁ5.24
ꢁ4.72
ꢁ5.67
e
ꢁ4.68
ꢁ4.26
ꢁ4.56
ꢁ4.29
ꢁ5.34
e
ꢁ4.12
ꢁ4.90
ꢁ4.62
ꢁ4.96
ꢁ4.72
ꢁ5.13
e
ꢁ4.53
ꢁ4.71
ꢁ5.05
ꢁ4.47
ꢁ4.92
e
e
NCI/ADR-RES
MDA-MB-231/ATCC
HS 578T
ꢁ4.24
ꢁ4.52
ꢁ4.23
ꢁ4.68
e
ꢁ4.20
ꢁ4.55
ꢁ4.12
ꢁ4.44
e
e
ꢁ4.07
ꢁ4.09
e
e
e
MDA-MB-435
MDA-N
BT-549
ꢁ5.01
e
ꢁ4.30
e
e
e
e
e
e
e
e
ꢁ4.92
e
ꢁ4.57
e
ꢁ4.20
e
ꢁ4.38
ꢁ5.78
e
T-47D
ꢁ4.94
ꢁ4.51
ꢁ4.08
ꢁ5.31
e Cell line not tested.
a
Data obtained from the NCI’s in vitro disease-oriented human tumor cells screen (see Refs. [19e21]).
The log of the molar concentration that inhibits 50% net cell growth.
The log of the molar concentration giving total growth inhibition.
The log of the molar concentration leading to 50% net cell death.
The values of log GI50, log TGI or log LC50 >ꢁ4.00.
b
c
d
e
Compounds 2c, 2g, 11, 12, 14 are not active towards EcTS at
position 3. Classical folate analogues MTX (1U72, 1DLS)
and their derivatives (1OHJ, 1OHK), pyrido[2,3-d]pyrimidine
derivatives (1DLR, 1KMV, 1KMS, 1PD8), furo[2,3-d]pyrimi-
dine derivatives (1HFQ, 1HFR, 1HFP), quinazoline derivatives
(1S3W)) bind to hDHFR forming a stacking interaction with
the nicotinamide ring of NADPH and also interact with
Thr7, Val8, Ala9, Leu22, Glu30, Phe31 and Phe34, whereas
the remaining part of the ligands are stretched along the active
site hole interacting with Phe31, Phe34, Gln35, Ile60, Pro61,
Asn64, Leu67 and Arg70.
10 mM. All compounds are moderately active towards LcTS
with Ki in the range of 1e36 mM. Compounds 2c and 2f are
species-specific with respect to hTS with specificity indices
(Ki hTS/Ki LcTS) of 32 and 63, respectively.
2.4. Molecular docking
With the aim to describe the binding mode of compounds
2aed to their predicted intracellular target, the enzyme
hDHFR, we attempted to rigidly superimpose compounds
2aed on the X-ray ternary complex hDHFReNADPHe
MTX [18e22]. The rigid matching failed due to geometry
differences between the quinoxaline/pteridine derivatives. No
X-ray structure of the 2,3-disubstitued quinoxaline derivatives
complexed with DHFR enzymes is available yet. Therefore we
suspected that the binding mode of the latter molecules could
be different with respect to the classical binding mode adopted
by the known pteridine derivatives, unless the enzyme does not
accommodate the molecules with a large conformational
change. This would be due to mainly the bulkiness of the
phenyl in position 2 and the methoxyphenyl fragment at
The complexes of compounds 2aed with hDHFR have
been predicted using AutoDock 3.05. AutoDock results clearly
show a preferred binding conformation for compounds 2aed
that correspond to the conformation of the best clusters (low-
est docked energy and most populated). AutoDock finds 31,
23, 55 and 60 clusters of binding conformations out of 100
runs for compounds 2a, 2b, 2c and 2d, respectively. The
best binding conformation groups together 37, 39, 14 and
19 out of 100 runs for compounds 2a, 2b, 2c and 2d, respec-
tively, with docked energies that range from ꢁ10.59 to
ꢁ10.86 kcal/mol. Compounds 2aed bind in the same bind-
ing site and variation in number and position of the substit-
uents on the phenyl ring does not alter their binding mode.
They bind in the active site of the enzyme in the area where
the PABA-Glu fragment of MTX also binds with the pyrazine
ring matching the benzene ring of the PABA fragment. The di-
amino quinoxaline ring is located perpendicular to the nicotin
ring and to part of the nicotin ribose of the NADPH and also
interacts with Phe31, Phe34, Thr56, Ser59 and Ile60, whereas
the two phenyls in positions 2 and 3 are oriented towards the
entrance of the active site and they interact with Phe31, Phe34,
Gln35, Ile60, Pro61, Asn64, Leu67 and Arg70. (Fig. 5).
The amine in position 7 forms two hydrogen bonds with the
side chain of Ser59 and the hydroxyl in position 2 of the ribose
ring bound to the nicotine part of the cofactor. The methoxy
groups in para and meta positions can hydrogen bind to
Arg28 and Asn64. These ligands form ringering interactions
Table 3
Log10 GI50, log10 TGI, log10 LC50 mean graph midpoints (MG_MD) of in vi-
tro cytotoxicity of compounds 2aeg against human tumor cell lines. In square
brackets the corresponding GI50, TGI and LC50 values are expressed in mM
Compound
log GI50 [mM]
log TGI [mM]
log LC50 [mM]
2a
2b
2c
ꢁ4.73 [18.6]
ꢁ5.04 [9.12]
ꢁ4.72 [19.0]
ꢁ4.90 [12.6]
ꢁ5.08 [8.32]
ꢁ4.82 [15.1]
ꢁ4.87 [13.5]
ꢁ6.33 [0.465]
ꢁ4.32 [47.9]
ꢁ4.36 [43.6]
ꢁ4.23 [58.9]
ꢁ4.38 [41.7]
ꢁ4.48 [33.1]
ꢁ4.34 [45.7]
ꢁ4.30 [50.1]
ꢁ3.77 [171]
ꢁ4.11 [77.6]
ꢁ4.07 [85.1]
ꢁ4.03 [93.3]
ꢁ4.07 [85.1]
ꢁ4.11 [77.6]
ꢁ4.08 [83.2]
ꢁ4.04 [91.2]
ꢁ3.61 [247]
2d
2e
2f
2g
MTXa
a
Values from NCI data base [34].

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
197
Table 4
Log10 GI50 mean graph midpoint (MG-MID) recorded in each cancer typea for compounds 2aeg
Panel cell line
2a
2b
2c
2d
2e
2f
2g
Leukemia
ꢁ4.67
ꢁ4.66
ꢁ4.72
ꢁ4.70
ꢁ4.80
ꢁ4.76
ꢁ4.75
ꢁ4.54
ꢁ4.80
ꢁ4.73
ꢁ5.74
ꢁ4.92
ꢁ4.70
ꢁ4.89
ꢁ4.96
ꢁ5.12
ꢁ5.08
ꢁ5.17
ꢁ5.08
ꢁ5.04
ꢁ5.13
ꢁ4.58
ꢁ4.72
ꢁ4.69
ꢁ4.69
ꢁ4.55
ꢁ4.76
ꢁ4.60
ꢁ4.81
ꢁ4.72
ꢁ5.24
ꢁ4.79
ꢁ4.76
ꢁ4.69
ꢁ4.91
ꢁ5.21
ꢁ4.74
ꢁ4.75
ꢁ4.91
ꢁ4.90
ꢁ5.60
ꢁ5.07
ꢁ4.94
ꢁ4.96
ꢁ5.05
ꢁ5.08
ꢁ4.93
ꢁ5.14
ꢁ5.11
ꢁ5.08
ꢁ5.21
ꢁ4.72
ꢁ4.80
ꢁ4.88
ꢁ4.82
ꢁ4.78
ꢁ4.83
ꢁ4.92
ꢁ4.87
ꢁ4.82
ꢁ5.42
ꢁ4.69
ꢁ4.69
ꢁ4.66
ꢁ5.43
ꢁ4.74
ꢁ4.66
ꢁ4.51
ꢁ4.83
ꢁ4.87
Non-small cell lung cancer
Colon cancer
CNS cancer
Melanoma
Ovarian cancer
Renal cancer
Prostate cancer
Breast Cancer
MG-MID
a
Each cancer type represents the average of six to eight different cancer cell lines.
with Phe31, Phe34 and the nicotine ring of NADPH. More-
over, they interact with Leu22, Pro26, Gln35, Thr56, Ile60,
Pro61, Leu67, Arg70 and Val115 through van der Waals inter-
actions (Fig. 6).
assumption. However, further studies are in progress to verify
the effect of various substituents in position 2 and 3 of the 5,
7-diamino quinoxaline on the anti-tumor and anti-folate activ-
ities, and to avenue structural biology studies that could
provide more detailed information for structure-based devel-
opment of new molecules.
3. Conclusion
4. Experimental
4.1. Chemistry
Compounds 2a, 2cee and 2g exhibited better anti-cancer
activity with respect to the corresponding trifluoromethyl
analogues (1aei) (Table 6) [3,5]. Structureeactivity relation-
ships allow us to conclude that the presence of two amino groups
on the quinoxaline system increases the anti-proliferative
activity. Furthermore, compounds 2b and 2d, were the most po-
tent of all derivatives tested, and displayed a high degree of se-
lectivity against leukemia CCRF-CEM (log10 GI50 ¼ <ꢁ8.00,
log10 TGI ¼ ꢁ6.78, log10 LC50 ¼ ꢁ6.01) and OVCAR-4
(log10 GI50 ¼ ꢁ7.50, log10 TGI ¼ ꢁ6.05, log10 LC50 ¼ ꢁ4.59),
respectively (Table 1), and resulted even more active than MTX
in these cell lines (Table 5). The same compounds have shown
the most significant inhibitory activity against hDHFR:
Ki ¼ 0.6 mM for compound 2b and Ki ¼ 0.1 mM for compound
2d. The calculated complex between hDHFR and compounds
2aed suggested that the binding mode of these quinoxaline de-
rivatives are different with respect to that of MTX and this fact
4.1.1. General remarks
Melting points were carried out with a Kofler hot stage or
¨
Digital Electrothermal melting point apparatus and are uncor-
rected. Infrared spectra were recorded as nujol mulls on NaCl
plates with a Perkin-Elmer 781 IR spectrophotometer and are
expressed in n (cmꢁ1). UV spectra are qualitative and were re-
corded in nm for solutions in EtOH with a Perkin-Elmer
Lamba 5 spectrophotometer. Nuclear magnetic resonance
(1H NMR and NOE difference) spectra were determined in
CDCl3, DMSO-d6, CDCl3/DMSO-d6 (in the ratio 1:3) and
were recorded in a Varian XL-200 (200 MHz). Chemical shifts
(d scale) are reported in parts per million (ppm) downfield
also accounts for a better activity recorded for the intermediate
dinitro-derivatives in the enzyme assay. At this stage it is to be
observed that a relationship between the enzyme inhibitory
activity and anti-cancer activity for these compounds could
not be traced owing to incontrovertible selection of our
compounds affected at NCI.
Table 6
Anti-tumor activity (log10 GI50, log10 TGI, log10 LC50, mean graph midpoints
(MG-MID)) comparison between compounds 2a, 2cee, 2g and the corre-
sponding trifluoromethyl analogues (1aei) [3,5]. In square brackets the corre-
sponding GI50 values are expressed in mM
Compound
log10 GI50 [mM]
log10 TGI [mM]
log10 LC50 [mM]
2a
1c
2c
1b
1f
ꢁ4.73 [18.6 ]
ꢁ4.15 [70.8 ]
ꢁ4.72 [19.0]
ꢁ4.62 [24.0]
ꢁ4.47 [33.9]
ꢁ4.66 [21.9]
ꢁ5.08 [8.32]
ꢁ4.50 [31.6]
ꢁ4.90 [12.6]
ꢁ4.42 [38.0]
ꢁ4.83 [14.8]
ꢁ4.31 [49.0]
ꢁ4.87 [13.5]
ꢁ4.50 [31.6]
ꢁ4.32 [47.9]
ꢁ4.00 [100]
ꢁ4.23 [58.9]
ꢁ4.20 [63.1]
ꢁ4.08 [83.2]
ꢁ4.18 [66.1]
ꢁ4.48 [33.1]
ꢁ4.00 [100]
ꢁ4.38 [41.7]
ꢁ4.08 [83.2]
ꢁ4.23 [58.9]
ꢁ4.11 [77.6]
ꢁ4.30 [50.1]
ꢁ4.00 [100]
ꢁ4.11 [77.6]
ꢁ4.00 [100]
ꢁ4.03 [93.3]
ꢁ4.02 [95.5]
ꢁ4.03 [93.3]
ꢁ4.00 [100]
ꢁ4.11 [77.6]
ꢁ4.00 [100]
ꢁ4.07 [85.1]
ꢁ4.00 [100]
ꢁ4.06 [87.1]
ꢁ4.02 [95.5]
ꢁ4.04 [91.2]
ꢁ4.00 [100]
These results are encouraging since inhibition of the folate
enzymes seems to confirm the grounds for our original
Table 5
Comparison of cytotoxicity evaluations (log10 GI50) for compounds 2b and 2d
against selected tumor cell lines with MTX. In square brackets the correspond-
ing GI50 values are expressed in mM
1g
2e
1h
2d
1a
1d
1e
2g
1i
Compound
Cell line
log10 GI50 [mM]
2b
Leukemia CCRF-CEM
Ovarian cancer OVCAR-4
Leukemia CCRF-CEM
Ovarian cancer OVCAR-4
ꢁ8 [<0.01]
2d
ꢁ7.5 [0.0316]
ꢁ7.4 [0.0398]
ꢁ8.6 [0.00251]
MTXa
MTXa
a
Values from NCI [34].

198
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
Table 7
Inhibition constant (Ki) for compounds 2aeg and 5,7-dinitro intermediates
(11e17)
Compound
LcTS Ki
(mM)
EcTS Ki
(mM)
hTS Ki
(mM)
hDHFR
Ki (mM)
2a
2b
2c
2d
2e
2f
e
e
6
e
e
NIa
e
6
25
NIa
51
0.9
NIa
34
0.7
15
14
2.1
36
e
1.8
0.6
NIa
0.1
NIa
NIa
NIa
e
e
1
3
3
70
NIa
NIa
NIa
1
NIa
0.2
e
2g
11
12
13
14
15
16
17
MTX
a
27
8
6
e
Fig. 6. Predicted binding modes for compounds 2aed (coloured by atom with
carbon atoms in green, pink, orange and cyan, respectively, for compounds 2a,
2b, 2c and 2d) in the active site of hDHFR (only the interacting residues of the
protein, coloured by atom with carbon atoms in gray, are shown. For interpre-
tation of the references to colour in this figure legend, the reader is referred to
the web version of this article).
24
11
5
e
e
e
e
e
36
3
16
4.42 ꢃ 10ꢁ6b
0.040e0.063 mm (230e400 mesh ASTM). Elemental analyses
were performed on a Perkin-Elmer 2400 instrument at Labora-
No inhibition at 10 mM.
b
see Ref. [48].
`
torio diMicroanalisi, DipartimentodiChimica, UniversitadiSas-
sari, Italy, and results were within ꢂ0.4% of theoretical values.
from tetramethylsilane (TMS) as internal standar. Splitting
patterns are designated, as follows: s, singlet; d, doublet; t,
triplet; q, quadruplet; m, multiplet; br s, broad singlet; dd, dou-
ble doublet.
The assignment of exchangeable protons (OH and NH) was
confirmed by the addition of D2O. MS spectra were performed
on combined HP 5790-HP 5970 GC/MS apparatus or with
a combined Liquid Chromatograph-Agilent 1100 series Mass
Selective Detector (MSD). Analytical thin-layer chromatogra-
phy (TLC) was performed on Merck silica gel F-254. Pure
compounds showed a single spot in TLC. For flash chromatog-
raphy, Merck silica gel 60 was used with a particle size
4.1.2. Intermediates
2-Chloro-1,3,5-trinitrobenzene [23] (19) was prepared as
reported in the literature [23]; 2,4,6-trinitroaniline [24] (20)
and 2-amino-3,5-dinitro-phenylamine (21) are known in the
literature, but these intermediates were synthesized following
different synthetic routes as described below. Quinoxalin-
2(1H )ones 22, 24 and chloroquinoxaline 3 are new com-
pounds and were described below for the first time.
4.1.2.1. 2,4,6-Trinitroaniline [24] (20). A mixture of 2-chloro-
1,3,5-trinitrobenzene [23] (19) (2 g, 8.1 mmol) and a saturated
solution of NH3 in 1-propanol (32 mL) was heated in a sealed
tube at 100 ꢀC for 3 h. On cooling, the reaction mixture was
evaporated to dryness to give 1.1 g of 20, as solid, in 75%
yield; m.p. 189.5e191.5 ꢀC [24] (methanol/H2O).
4.1.2.2. 2-Amino-3,5-dinitrophenylamine [25] (21). Hydrogen
sulfide was introduced into a mixture of 2,4,6-trinitroaniline
[24] (20) (1 g, 4.37 mmol) and 30% aqueous solution
NH4OH (2.4 mL) in ethanol (100 mL) and heated under stir-
ring at 45e50 ꢀC for 30 min. Elemental sulfur was precipi-
tated during the introduction of H2S. The reaction mixture
was cooled, and the solvent was evaporated to dryness to
give 0.6 g of crude solid that was purified by flash chroma-
tography over silica gel using petroleum ether and ethyl ac-
etate as the eluent. 2-Amino-3,5-dinitrophenylamine [25]
(21) was obtained in 90% yield; m.p. 215e218 ꢀC [25]
(ethanol).
Fig. 5. Predicted binding modes for compounds 2aed (coloured by atom with
carbon atoms in green, pink, orange and cyan, respectively, for compounds 2a,
2b, 2c and 2d) in the active site of hDHFR (only the ribbon representation of
the protein, coloured in gray, is shown. The ribbon of key residues is coloured
in violet.). Coloured by atom with carbon atoms in yellow are the NADPH
cofactor and MTX (for interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article).
1H NMR (CDCl3/DMSO-d6): d 8.40 (1H, d, J ¼ 2.4 Hz),
8.64 (1H, d, J ¼ 2.4 Hz), 7.76 (2H, s, exc. with D2O), 7.61
(1H, d,), 5.51 (2H, s, exc. with D2O).
4.1.2.3. 5,7-Dinitro-3-phenylquinoxalin-2(1H )one (22) and 6,8-
dinitro-3-phenylquinoxalin-2(1H )one (24). To a mixture of

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
199
2-amino-3,5-dinitrophenylamine (21) (0.3 g, 1.51 mmol) and
benzoyl formic acid (23) (0.36 g, 2.4 mmol) in ethanol
(15 mL), 1.5 ml of 4 M HCl was added. The mixture was re-
fluxed for 12 h. On cooling, the precipitate that formed was fil-
tered and washed with ethanol. The crude solid was purified by
silica gel column chromatography, eluting with a 8:2 mixture of
petroleum ether and ethyl acetate to yield the isomer 24. Further
elution with a 1:1 mixture of petroleum ether and ethyl acetate
gave the 5,7-dinitro-3-phenylquinoxalin-2(1H )one (22). Melt-
ing points, yields, analytical and spectroscopic data are reported
below.
5,7-Dinitro-3-phenylquinoxalin-2(1H )one (22) was ob-
tained at 75% yield; m.p. 317e319 ꢀC (CH3CN); TLC (petro-
leum ether/ethyl acetate 8:2): Rf 0.15; IR (nujol): n 3240, 1680,
1620, 1530, 1350 cmꢁ1; UV (EtOH): lmax 376, 297, 205 nm;
1H NMR (DMSO-d6): d 13.21 (1H, br s, exc. with D2O), 8.64
(1H, d, J ¼ 2.4 Hz), 8.33 (2H, dd, J ¼ 6.4 Hz and J ¼ 1.8 Hz),
8.25 (1H, d, J ¼ 2.4 Hz), 7.62e7.50 (3H, m); GC/MS: m/z 312
[M]þ. Anal. calcd. for C14H8N4O5: C, 53.85; H, 2.58; N,
17.94; found C, 53.65; H, 2.48; N, 17.84.
6,8-Dinitro-3-phenylquinoxalin-2(1H )one (24) was ob-
tained at 2% yield; m.p. 257e258 ꢀC (CH3CN); TLC
(petroleum ether/ethyl acetate 8:2): Rf 0.46; IR (nujol): n
3240, 1690, 1615, 1530, 1340 cmꢁ1; UV (EtOH): lmax
389, 282, 202 nm; 1H NMR (DMSO-d6): d 13.38 (1H, s,
exc. with D2O), 8.98 (1H, d, J ¼ 3 Hz), 8.92 (1H, d,
J ¼ 3 Hz), 8.35 (2H, dd, J ¼ 1.3 Hz and J ¼ 7.2 Hz), 7.65e
7.50 (3H, m); GC/MS: m/z 312 [M]þ. Anal. calcd. for
C14H8N4O5: C, 53.85; H, 2.58; N, 17.94; found C, 53.68;
H, 2.49; N, 17.83.
were purified by flash chromatography over silica gel using
petroleum ether and ethyl acetate as the eluent or by crystal-
lization as reported below. Yields, reaction conditions, melt-
ing points, analytical and spectroscopical data are reported as
follows.
4.1.2.5.1. 2-[(4-Methoxy-phenyl)amino]-3-phenyl-5,7-dinitro-
quinoxaline (11). This compound was obtained at 84% yield
starting from 3 (150 mg, 0.45 mmol) and 4-methoxy-phenyl-
amine (4) (16 mg, 1.35 mmol) after 8 h; m.p. 249e250 ꢀC
(from CH3CN); TLC (petroleum ether/ethyl acetate 8:2): Rf
0.31; IR (nujol): n 3390, 1620, 1550, 1350 cmꢁ1; UV
(EtOH): lmax 309, 202 nm; 1H NMR (CDCl3/DMSO-d6):
d 8.68 (1H, d, J ¼ 2.2 Hz), 8.63 (1H, s, exc. with D2O), 8.47
(1H, d, J ¼ 2.2 Hz), 8.06e7.90 (2H, m), 7.70 (2H, d,
J ¼ 9 Hz), 7.66e7.55 (3H, m), 6.94 (2H, d, J ¼ 9 Hz), 3.83
(3H, s); GC/MS: m/z 417 [M]þ. Anal. calcd. for
C21H15N5O5: C, 60.43; H, 3.62; N, 16.78; found C, 60.23;
H, 3.42; N, 16.58.
4.1.2.5.2. 2-[(3,5-Dimethoxy-phenyl)amino]-3-phenyl-5,7-
dinitro-quinoxaline (12). This compound was obtained at
80% yield starting from 3 (150 mg, 0.45 mmol) and 3,5-dime-
thoxy-phenylamine (5) (210 mg, 1.35 mmol) after 10 h; m.p.
253e254 ꢀC (from CH3CN); TLC (petroleum ether/ethyl ace-
tate 8:2): Rf 0.35; IR (nujol): n 3390, 1605, 1550, 1345 cmꢁ1
;
1
UV (EtOH): lmax 309, 205 nm; H NMR (CDCl3/DMSO-d6):
d 8.90 (1H, s, exc. with D2O), 8.71 (1H, d, J ¼ 2.2 Hz), 8.56
(1H, d, J ¼ 2.2 Hz), 7.97 (2H, dd, J ¼ 6 Hz and J ¼ 2.2 Hz),
7.70e7.60 (3H, m), 7.11 (2H, d, J ¼ 2.2 Hz), 6.30e6.25
(1H, m), 3.81 (6H, s); GC/MS: m/z 447 [M]þ. Anal. calcd.
for C22H17N5O6: C, 59.06; H, 3.83; N, 15.65; found C,
58.90; H, 3.63; N, 15.46.
4.1.2.4. 5,7-Dinitro-3-phenyl-2-chloro-quinoxaline (3). A sus-
pension of 22 (1.46 g, 4.67 mmol) in SOCl2 (26 mL) and 10
drops of DMF was heated under reflux until a clear solution
was obtained (10 min), and for an additional 90 min. The
SOCl2 was removed under reduced pressure, and the residue
was triturated with diethyl ether. A yellow solid of 3 (1.39 g)
was filtered off (90% yield); m.p. 158.0e160 ꢀC (C2H6O);
TLC (petroleum ether/ethyl acetate 8:2): Rf 0.56; IR (nujol):
n 1620, 1530, 1325 cmꢁ1; UV (EtOH): lmax 341, 241 nm;
1H NMR (CDCl3): d 9.14 (1H, d, J ¼ 2.6 Hz), 8.93 (1H, d,
J ¼ 2.6 Hz), 8.04 (2H, dd, J ¼ 8 Hz and J ¼ 1.2 Hz), 7.64e
7.50 (3H, m); GC/MS: m/z 330 [M]þ. Anal. calcd. for
C14H7ClN4O4: C, 50.85; H, 2.13; N, 16.94; found C, 50.75;
H, 2.10; N, 16.73.
4.1.2.5.3. 2-[(3,4-Dimethoxy-phenyl)amino]-3-phenyl-5,7-
dinitro-quinoxaline (13). This compound was obtained at
90% yield starting from 3 (150 mg, 0.45 mmol) and 3,4-dime-
thoxy-phenylamine (6) (210 mg, 1.35 mmol) after 8 h; m.p.
213e215 ꢀC (from C2H5O); TLC (petroleum ether/ethyl ace-
tate 6:4): Rf 0.44; IR (nujol): n 3390, 1605, 1535,
1350 cmꢁ1; UV (EtOH): lmax 309, 205 nm; 1H NMR
(CDCl3):
d
8.74 (1H, d, J ¼ 2.4 Hz), 8.52 (1H, d,
J ¼ 2.4 Hz), 8.00e7.88 (2H, dd, J ¼ 6 Hz and J ¼ 2.4 Hz),
7.71e7.60 (3H, m), 7.48 (1H, s, exc. with D2O), 7.39 (1H,
d, J ¼ 2.4 Hz), 7.20 (1H, dd, J ¼ 8.6 Hz and J ¼ 2.4 Hz),
6.91 (1H, d, J ¼ 8.6 Hz), 3.95 (3H, s), 3.92 (3H, s); LC/MS:
m/z 448 [M þ H], 470 [M þ Na], 486 [M þ K]. Anal. calcd.
for C22H17N5O6: C, 59.06; H, 3.83; N, 15.65; found C,
58.90, H, 3.50; N, 15.46.
4.1.2.5. General procedure for preparation of 5,7-dinitro-3-
phenyl-2-anilino substituted quinoxalines (11e17). A solution
of 3 (150 mg, 0.45 mmol) in 1-propanol (15 mL) was ob-
tained by heating under reflux. Then to the solution an excess
(1.35 mmol) of the required substituted anilines (4e10) was
added. Heating was maintained for different times as
reported below. On cooling, the resulting precipitates were
collected. In the case of the aminoester 10, nucleophilic dis-
placement of chlorine was carried out using ethanol to avoid
the possible trans-esterification of the diethyl ester (17) as
observed in other cases. The coloured precipitates formed
4.1.2.5.4. 2-[(3,4,5-Trimethoxy-phenyl)amino]-3-phenyl-5,7-
dinitro-quinoxaline (14). This compound was obtained at
95% yield starting from 3 (150 mg, 0.45 mmol) and 3,4,5-tri-
methoxy-phenylamine (7) (250 mg, 1.35 mmol) after 7 h;
m.p. 228e229 ꢀC (from CH3CN); TLC (petroleum ether/ethyl
acetate 6:4): Rf 0.38; IR (nujol): n 3390, 1600, 1545,
1345 cmꢁ1; UV (EtOH): lmax 312, 205 nm; 1H NMR
(CDCl3):
d
8.72 (1H, d, J ¼ 2.4 Hz), 8.53 (1H, d,
J ¼ 2.4 Hz), 7.98e7.86 (2H, m), 7.72e7.65 (3H, m), 7.50
(1H, s, exc. with D2O), 7.02 (2H, s), 3.92 (3H, s), 3.87
(3H, s); GC/MS: m/z 477 [M]þ. Anal. calcd. for

200
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
C23H19N5O7: C, 57.86; H, 4.01; N, 14.67; found C, 58.16; H,
3.66; N, 14.59.
recrystallization or by flash chromatography over silica gel
as reported below.
4.1.2.5.5. 2-[(3,4-Dichloro-phenyl)amino]-3-phenyl-5,7-dini-
tro-quinoxaline (15). This compound was obtained at 44%
yield starting from 3 (150 mg, 0.45 mmol) and 3,4-dichloro-
phenylamine (8) (210 mg, 1.35 mmol) after 7 h; the compound
was purified by silica gel column chromatography eluting with
a 8:2 mixture of petroleum ether/ethyl acetate; m.p. 291e
293 ꢀC (from CH3CN); TLC (petroleum ether/ethyl acetate
8:2): Rf 0.46; IR (nujol): n 3390, 1605, 1535, 1350 cmꢁ1; UV
(EtOH): lmax 312, 205 nm; 1H NMR (CDCl3/DMSO-d6):
d 9.51 (1H, br s, exc. with D2O), 8.74 (1H, d, J ¼ 2.4 Hz),
8.64 (1H, J ¼ 2.4 Hz), 8.12 (1H, d, J ¼ 2.4 Hz), 8.01e7.90
(3H, m), 7.68e7.60 (3H, m), 7.50 (1H, d, J ¼ 8.8 Hz); GC/
MS: m/z 456 [M]þ. Anal. calcd. for C20H11Cl2N5O4: C,
52.65; H, 2.43; N, 15.35; found C, 52.43; H, 2.20; N, 15.22.
4.1.2.5.6. 2-[(4-Fluoro-phenyl)amino]-3-phenyl-5,7-dinitro-
quinoxaline (16). This compound was obtained at 56% yield
starting from 3 (150 mg, 0.45 mmol) and 4-fluoro-phenyl-
amine (9) (150 mg, 1.35 mmol) after 1 h; the compound was
purified by silica gel column chromatography eluting with
a 8:2 mixture of petroleum ether/ethyl acetate; m.p. 259e
263 ꢀC (from CH3CN); TLC (petroleum ether/ethyl acetate
4.1.3.1. 2-[(4-Methoy-phenyl)amino]-3-phenyl-5,7-diaminoqui-
noxaline (2a). This compound was obtained in 90% yield start-
ing from 11 (150 mg, 0.36 mmol), hydrazine hydrate (150 mg,
3 mmol) and palladised charcoal (15 mg) in ethanol (81 mL) af-
ter 2 h and 45 min under reflux; m.p. 159e160 ꢀC (from diethyl
ether/petroleum ether); TLC (petroleum ether/ethyl acetate 4:6):
Rf 0.34; IR (nujol): n 3331, 3201, 1623, cmꢁ1; UV (EtOH): lmax
1
396, 289, 228, 203 nm; H NMR (CDCl3): d 7.75 (2H, dd,
J ¼ 1.8 Hz and J ¼ 7.8 Hz), 7.61 (2H, d, J ¼ 9 Hz), 7.60e
7.49 (3H, m), 6.88 (2H, d, J ¼ 9 Hz and 1H partially obscured,
exc. with D2O), 6.36 (1H, d, J ¼ 1.8 Hz), 6.11 (1H, d,
J ¼ 1.8 Hz), 4.30 (2H, br s, exc. with D2O), 3.80 (3H, s); GC/
MS: m/z 357 [M]þ. Anal. calcd. for C21H19N5O: C, 70.57; H,
5.36; N, 19.59; found C, 70.21; H, 5.20; N, 15.40.
4.1.3.2. 2-[(3,5-Dimethoxy-phenyl)amino]-3-phenyl-5,7-diami-
no-quinoxaline (2b). This compound was obtained at 71%
yield starting from 12 (250 mg, 0.56 mmol), hydrazine hydrate
(260 mg, 5.2 mmol) and palladised charcoal (25 mg) in etha-
nol (125 mL) after 1 h under reflux; m.p. 148e150 ꢀC (from
CH4O þ H2O); TLC (petroleum ether/ethyl acetate 4:6): Rf
0.49; IR (nujol): n 3420, 3360, 1600, 1200, 1070 cmꢁ1; UV
(EtOH): lmax 293, 223, 204 nm; 1H NMR (DMSO-d6):
d 8.08 (1H, s, exc. with D2O), 7.88e7.78 (2H, dd,
J ¼ 6.6 Hz and J ¼ 1.2 Hz), 7.62e7.42 (3H, m), 7.12 (2H, d,
J ¼ 2 Hz), 6.12 (1H, t), 6.10 (1H, d, J ¼ 2 Hz), 6.03 (1H, d,
J ¼ 2 Hz), 5.64 (2H, br s, exc. with D2O), 5.57 (2H, br s,
exc. with D2O), 3.73 (6H, s); GC/MS: m/z 387 [M]þ. Anal.
calcd. for C22H21N5O2: C, 68.20; H, 5.46; N, 18.08; found
C, 68.18; H, 4.26; N, 18.00.
8:2): Rf 0.34; IR (nujol): n 3362, 1609, 1538, 1348 cmꢁ1
;
UV (EtOH): lmax 312, 2063, 222, 205 nm; 1H NMR
(CDCl3/DMSO-d6): d 9.27 (1H, br s, exc. with D2O), 8.67
(1H, d, J ¼ 2.4 Hz), 8.58 (1H, d, J ¼ 2.4 Hz), 8.00e7.90
(2H, m), 7.89e7.80 (2H, m), 7.65e7.57 (3H, m), 7.12
(2H, t); GC/MS: m/z 405 [M]þ. Anal. calcd. for
C20H12FN5O4: C, 59.26; H, 2.98; N, 17.28; found C, 59.02;
H, 2.70; N, 17.00.
4.1.2.5.7. N-[4(5,7-Dinitro-3-phenyl-quinoxalin-2-ylamino)-
benzoyl]glutamic acid diethyl ester (17). This compound was
obtained at 61% yield using ethanol as solvent and starting
from 3 (150 mg, 0.45 mmol) and p-amino-benzoyl-glutamic
acid diethyl ester (10) (430 mg, 1.35 mmol) after 186 h; the
compound was purified by silica gel column chromatography
eluting with a 6:4 mixture of petroleum ether/ethyl acetate;
m.p. 174e175 ꢀC (from C2H6O); TLC (petroleum ether/ethyl
acetate 6:4): Rf 0.32; IR (nujol): n 3400, 3300, 1720, 1640,
1600, 1520, 1360 cmꢁ1; UV (EtOH): lmax 428, 321, 204 nm;
1H NMR (CDCl3): d 8.90 (1H, d, J ¼ 2.4 Hz), 8.60 (1H, d,
J ¼ 2.4 Hz), 8.00e7.82 (4H, m and 1H exc. with D2O),
7.82e7.65 (5H, m), 7.16 (1H, d, exch. with D2O), 4.88e4.72
(1H, m), 4.26 (2H, q), 4.14 (2H, q), 2.62e2.09 (4H, m), 1.32
(3H, t), 1.25 (3H, t); GC/MS: m/z 616 [M]þ. Anal. calcd. for
C30H28N6O9: C, 58.44; H, 4.58; N, 13.63; found C, 58.62; H,
4.20; N, 13.60.
4.1.3.3. 2-[(3,4-Dimethoxy-phenyl)amino]-3-phenyl-5,7-diami-
no-quinoxaline (2c). This compound was obtained at 90%
yield starting from 13 (300 mg, 0.67 mmol), hydrazine hydrate
(495 mg, 10 mmol) and palladised charcoal (230 mg) in etha-
nol (24 mL) after 30 min under reflux; the compound was pu-
rified by silica gel column chromatography eluting with a 4:6
mixture of petroleum ether/ethyl acetate; m.p. 181e182 ꢀC
(from CH4O þ H2O); TLC (petroleum ether/ethyl acetate
4:6): Rf 0.61; IR (nujol): n 3420, 3370, 1625, 1200,
1070 cmꢁ1; UV (EtOH): lmax 400, 290, 225, 202 nm; 1H
NMR (DMSO-d6): d 7.90 (1H, s, exc. with D2O), 7.79 (H,
dd, J ¼ 6.4 Hz and J ¼ 1.4 Hz), 7.55 (1H, d, J ¼ 2.2 Hz),
7.53e7.44 (3H, m), 7.26 (1H, dd, J ¼ 6.6 Hz and
J ¼ 2.2 Hz), 6.87 (1H, d, J ¼ 8.8 Hz), 6.03 (1H, d,
J ¼ 2.2 Hz), 5.96 (1H, d, J ¼ 2.2 Hz), 5.58 (1H, br s, exc.
with D2O), 5.49 (2H, br s, exc. with D2O), 3.74 (3H, s),
3.72 (3H, s); LC/MS: m/z 388 [M þ H]. Anal. calcd. for
C22H21N5O2: C, 68.20; H, 5.46; N, 18.08; found C, 67.24;
H, 5.10; N, 18.07.
4.1.3. General procedure for preparation of 5,7-diamino-3-
phenyl-2-anilino substituted quinoxalines (2aeg)
A mixture of 11e17 (0.24e0.67 mmol) and an excess of
hydrazine hydrate (2.9e10 mmol) in ethanol (15e127 mL),
in the presence of 10% palladised charcoal was refluxed for
the time reported below. The desired compounds were ob-
tained by filtration of the ethanolic solution and evaporation
in vacuo of the solvent. The crude solids were purified by
4.1.3.4. 2-[(3,4,5-Trimethoxy-phenyl)amino]-3-phenyl-5,7-dia-
mino-quinoxaline (2d). This compound was obtained at 46%
yield starting from 14 (150 mg, 0.31 mmol), hydrazine hydrate

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
201
(145 mg, 2.9 mmol) and palladised charcoal (15 mg) in etha-
nol (60 mL) after 1 h under reflux; m.p. 175e177 ꢀC (from
CH4O þ H2O); TLC (petroleum ether/ethyl acetate 4:6): Rf
0.41; IR (nujol): n 3425, 3350, 1600, 1200, 1070 cmꢁ1; UV
(EtOH): lmax 398, 292, 202 nm; 1H NMR (DMSO-d6):
d 8.01 (1H, s, exc. with D2O), 7.79 (H, dd, J ¼ 6.4 Hz and
J ¼ 1.4 Hz), 7.60e7.40 (3H, m), 7.27 (2H, s), 6.05 (1H, d,
J ¼ 2.2 Hz), 5.99 (1H, d, J ¼ 2.2 Hz) 5.60 (2H, br s, exc.
with D2O), 5.53 (2H, br s, exc. with D2O), 3.75 (6H, s),
3.62 (3H, s); GC/MS: m/z 417 [M]þ. Anal. calcd. for
C23H23N5O3: C, 66.17; H, 5.55; N, 16.78; found C, 66.00;
H, 5.32; N, 16.57.
(2H, d, J ¼ 1.6 Hz), 4.24 (2H, q), 4.11 (2H, q), 4.00 (2H, br
s, exc. with D2O), 2.60e2.03 (4H, m), 1.31 (3H, t), 1.22
(3H, t); LC/MS: m/z 557 [M þ H]. Anal. calcd. for
C30H32N6O5: C, 64.73; H, 5.79; N, 15.10; found C, 64.67;
H, 5.62; N, 15.01.
4.2. Pharmacology
Compounds 2aeg were evaluated for in vitro activity
against sixty human tumor cell lines, derived from nine cancer
types (leukemia, non-small cell lung, colon, CNS, melanoma,
ovarian, renal, prostate, and breast cancers). For each com-
pound, doseeresponse curves against each cell line were mea-
sured at a minimum of five concentrations at 10-fold dilutions.
A protocol of 48 h continuous drug exposure was used, and
a sulforhodamine B (SRB) protein assay has been used to es-
timate cell viability or growth [35].
4.1.3.5. 2-[(3,4-Dichloro-phenyl)amino]-3-phenyl-5,7-diamino-
quinoxaline (2e). This compound was obtained at 54% yield
starting from 15 (200 mg, 0.44 mmol), hydrazine hydrate
(320 mg, 6.4 mmol) and palladised charcoal (20 mg) in etha-
nol (15 mL) after 10 min under reflux; the compound was pu-
rified by silica gel column chromatography eluting with a 6:4
mixture of petroleum ether/ethyl acetate; m.p. 192e193 ꢀC
(from CH4O); TLC (petroleum ether/ethyl acetate 6:4): Rf
0.4; IR (nujol): n 3435, 3333, 1633 cmꢁ1; UV (EtOH): lmax
4.3. Enzymology assays
All the reagents used for enzyme purification and enzyme
kinetics were of purity grade. 6-R-N5,N10-Methylentetrahydro-
folate is a generous gift of Eprova (Schaffhausen, Switzer-
land). The hDHFR, hTS, EcTS, LcTS plasmids and strains
were supplied by Dr. D.V. Santi (University of California,
San Francisco, USA; present address Kosan, San Francisco),
and Dr. G.F. Maley (New York State Department of Health,
New York). The enzymatic assays have been run on spectro-
photometer UVevis, Beckman DU 640, equipped with a ther-
mostat circulating bath HAAKE F3C.
The expression and purification of LcTS, EcTS, hTS and
hDHFR was done following known procedures as already
reported [36e40]. The enzyme preparations were >95% ho-
mogeneous as visualized by SDS (sodium dodecyl sulfate)-
polyacrylamide gel electrophoresis. The purified enzymes
were stored at ꢁ80 ꢀC in 10 mM phosphate buffer, pH 7.0,
and 0.1 mM EDTA.
1
301, 221, 203 nm; H NMR (CDCl3/DMSO-d6): d 8.22 (1H,
d, J ¼ 2.4 Hz), 7.76 (2H, dd, J ¼ 1.8 Hz and J ¼ 7.4 Hz),
7.62e7.50 (3H, m), 7.47 (1H, d, J ¼ 2.4 Hz), 7.35 (1H, t),
6.38 (1H, d, J ¼ 2.2 Hz), 6.20 (1H,d, J ¼ 2.2 Hz), 5.11 (2H,
br s, exc. with D2O), 4.66 (2H, br s, exc. with D2O); LC/
MS: m/z 396 [M þ H]. Anal. calcd. for C20H15Cl2N5: C,
60.62; H, 3.82; N, 17.67; found C, 60.42; H, 3.53; N, 17.57.
4.1.3.6. 2-[(4-Fluoro-phenyl)amino]-3-phenyl-5,7-diamino-qui-
noxaline (2f). This compound was obtained at 77% yield start-
ing from 16 (150 mg, 0.37 mmol), hydrazine hydrate (250 mg,
5 mmol) and palladised charcoal (15 mg) in ethanol (18 mL)
after 10 min under reflux; m.p. 158e161 ꢀC (from
CH4O þ H2O); TLC (petroleum ether/ethyl acetate 1:1): Rf
0.42; IR (nujol): n 3430, 3310, 1630 cmꢁ1; UV (EtOH):
1
lmax 392, 291, 225, 202 nm; H NMR (CDCl3/DMSO-d6):
d 7.80e7.65 (7H, m), 7.60 (2H, d, J ¼ 6.2 Hz), 7.02 (2H, t),
6.35 (1H, d, J ¼ 2.6 Hz), 6.14 (2H, d, J ¼ 2.6 Hz), 4.72 (2H,
br s, exc. with D2O), 4.20 (2H, br s, exc. with D2O); GC/
MS: m/z 345 [M]þ. Anal. calcd. for C20H16FN5: C, 69.55;
H, 4.67; N, 20.28; found C, 69.44; H, 4.57; N, 20.12.
The TS activity was determined spectrophotometrically by
steady-state kinetic analysis, following the increasing absor-
bance at 340 nm due to the oxidation reaction of N5,N10-meth-
ylenetetrahydrofolate (mTHF) to dihydrofolate (DHF) [41].
One milliliter of reaction solution was formed by standard as-
say buffer pH 7.4, dUMP (100 mM), 6(R,S)-1-CH2CH4-folate
(140 mM) (SIGMA), enzyme (0.07 mM). Assays were per-
formed at 20 ꢀC in the standard assay buffer formed by TES
(N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid)
(50 mM) at pH 7.4, MgCl2 (25 mM), formaldehyde
(6.5 mM), EDTA (1 mM), and 2-mercaptoethanol (75 mM).
DHFR catalyzes the NADPH-dependent reduction of dihy-
drofolate (DHF) to tetrahydrofolate (THF). The assays were
run spectrophotometrically by measuring the decrease in ab-
sorbance at 340 nm upon DHF reduction at 25 ꢀC. One en-
zyme unit is defined as 1 nmol of DHF reduced per minute.
The reaction mixture was formed by 50 mM TES, pH 7.0,
EDTA 1 mM, 1-mercaptoethanol 75 mM, NADPH 100 mM,
DHF 58.03 mM [42].
4.1.3.7. N-[4-[(5,7-Diamino-3-phenyl-quinoxalin-2-yl)amino]-
benzoyl]glutamic acid diethyl ester (2g). This compound
was obtained at 76% yield starting from 17 (150 mg,
0.24 mmol), hydrazine hydrate (170 mg, 3.5 mmol) and palla-
dised charcoal (15 mg) in ethanol (18 mL) after 30 min under
reflux; the compound was purified by silica gel column chro-
matography eluting with a 4:6 mixture of petroleum ether/
ethyl acetate; m.p. 164e165 ꢀC (from CH4O); TLC (petro-
leum ether/ethyl acetate 4:6): Rf 0.24; IR (nujol): n 3430,
3310, 1730, 1640, 1600 cmꢁ1; UV (EtOH): lmax 379, 295,
208 nm; 1H NMR (CDCl3): d 7.81 (4H, s), 7.73 (2H, d,
J ¼ 7 Hz), 7.62e7.50 (3H, m), 7.18 (1H, s, exc. with D2O),
6.99 (1H, d, exc. with D2O), 6.43 (1H, d, J ¼ 1.6 Hz), 6.17

202
P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
[3] M. Loriga, M. Fiore, P. Sanna, G. Paglietti, Farmaco 50 (1995) 289.
Stock solutions of the inhibitors were prepared in DMSO
(dimethyl sulfoxide) and stored at ꢁ20 ꢀC until use. The
inhibition pattern for all compounds was determined by
steady-state kinetic analysis of the dependence of enzyme
activity on folate concentration at varying inhibitor concen-
trations. All compounds showed competitive inhibition. Ki
values were obtained from the linear least-squares fit of
the residual activity as a function of inhibitor concentration,
using suitable equations for competitive inhibition [43].
Each experiment was repeated at least three times, and no
individual measurement differed by more than 20% from
the mean. The effect of increasing DMSO concentration in
the TS assay mixture was studied, and it was observed
that no change in TS activity was seen at concentrations
up to 8% DMSO (personal communication).
[4] M. Loriga, M. Fiore, P. Sanna, G. Paglietti, Farmaco 51 (1996) 559.
[5] M. Loriga, S. Piras, P. Sanna, G. Paglietti, Farmaco 52 (3) (1997) 157.
[6] M. Loriga, P. Moro, P. Sanna, G. Paglietti, Farmaco 52 (8e9)
(1997) 531.
[7] G. Vitale, P. Corona, M. Loriga, G. Paglietti, Farmaco 53 (1998) 139.
[8] G. Vitale, P. Corona, M. Loriga, G. Paglietti, Farmaco 53 (1998) 150.
[9] P. Corona, G. Vitale, M. Loriga, G. Paglietti, M.P. Costi, Farmaco 53
(1998) 480.
[10] G. Vitale, P. Corona, M. Loriga, G. Paglietti, Farmaco 53 (1998) 594.
[11] P. Corona, G. Vitale, M. Loriga, G. Paglietti, Farmaco 55 (2000) 77.
[12] S. Piras, M. Loriga, G. Paglietti, Farmaco 57 (2002) 1.
[13] M. Loriga, S. Piras, G. Paglietti, M.P. Costi, A. Venturelli, Farmaco 58
(2003) 51.
[14] S. Alleca, P. Corona, M. Loriga, G. Paglietti, R. Loddo, V. Mascia,
B. Busonera, P. La Colla, Farmaco 58 (2003) 639.
[15] S. Piras, M. Loriga, G. Paglietti, Farmaco 59 (2004) 189.
[16] P. Corona, G. Vitale, M. Loriga, G. Paglietti, P. La Colla, G. Collu,
G. Sanna, R. Loddo, Eur. J. Med. Chem. 41 (2006) 1102e1107.
[17] S. Piras, M. Loriga, A. Carta, G. Paglietti, M.P. Costi, S. Ferrari, J. Het-
erocycl. Chem. 43 (2006) 541.
4.4. Molecular docking
[18] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat,
H. Weissig, I.N. Shindyalov, P.E. Bourne, Nucleic Acids Res. 28
(2000) 235.
The program AutoDock 3.0.5 was used for docking [44].
The protein structure of hDHFR was taken from the crystal
structure of the ternary complex with NADPH and MTX
(1U72), available in the PDB [18e20,22]. Polar hydrogen
atoms were added to the protein with the WHATIF program
[45]. Solvation parameters and atomic partial charges were as-
signed to the protein using the programs q.kollua and addsol,
included in the AutoDock program suite. Polar hydrogen
atoms were added and atomic charges were assigned to the co-
factor using the AutoDockTools program [46]. The ligand
(2aed) structures were built with the ChemOffice suite of pro-
grams [47]. Atomic partial charges were assigned using the
AutoDockTools program. A grid with sides of 20.25, 24 and
[19] H.M. Berman, K. Henrick, H. Nakamura, Nat. Struct. Biol. 10 (12)
(2003) 980.
[20] http://www.pdb.org/
[21] W.S. Lewis, V. Cody, N. Galitsky, J.R. Luft, W. Pangborn,
S.K. Chunduru, H.T. Spencer, J.R. Appleman, R.L. Blakley, J. Biol.
Chem. 270 (1995) 5057.
[22] V. Cody, J.R. Luft, W. Pangborn, Acta Crystallogr. D. Biol. Crystallogr.
61 (2005) 147.
[23] E.T. Borrows, J.C. Clayton, B.A. Hems, A.G. Long, J. Chem. Soc.
(1949) 190.
`
[24] R.J.W. Le Fevre, D.D. Moir, E.E. Turner, J. Chem. Soc. (1927) 2330.
[25] R. Nietzki, H. Hagenback, Chem. Ber. 30 (1897) 543.
[26] M. Loriga, A. Nuvole, G. Paglietti, J. Chem. Res., Synop. (1989) 202;
J. Chem. Res., Miniprint (1989) 1553.
˚
20.25 A long, centered on the active site of the enzyme was
[27] G.O. Dudek, Spectrochim. Acta 19 (1963) 691.
defined. This grid is large enough to cover the active site of
the enzyme. A Lamarckian genetic algorithm was used to gen-
erate 100 bound conformations for each ligand. All single
bonds of the ligands were allowed to rotate during the calcu-
lation. The ga_pop_size parameter was set at 200. All other
settings were kept as default. On the basis of the clustering his-
togram output from the AutoDock program, the lowest energy
conformation of the most populated low energy cluster was
selected.
[28] C. MacLean, E.L. Mackor, Mol. Phys. 3 (1960) 223.
[29] R.H. Martin, N. Defay, F. Greets-Evrard, Tetrahedron 20 (1964) 1505.
[30] S.N. Bannore, J.L. Bose, A.A. Thakar, M.S. Wadia, Indian J. Chem. 13
(1975) 609.
[31] M.R. Boyd, Am. Assoc. Cancer Res. 30 (1989) 652.
[32] A. Monks, D. Scudiero, P. Skehan, R. Shoemaker, K. Poull, D. Vistica,
C. Hose, J. Langley, P. Cronise, A. Vaigro-Wolff, M. Gray-Goodrich,
H. Cambell, M. Boyd, J. Natl. Cancer Inst. 83 (1991) 757.
[33] J.N. Weinstein, T.G. Myers, P.M. O’Connor, S.H. Friend, A.J. Fornance Jr.,
K.W. Kohn, T. Fojo, S.E. Bates, L.V. Rubinstein, N.L. Anderson,
J.K. Buolamwini, W.W. van Osdol, A.P. Monks, D.A. Scudiero,
E.A. Sausiville, D.W. Zaharevitz, B. Bunow, V.N. Viswanadhan,
G.S. Johnson, R.E. Wittes, K.D. Paull, Science 275 (1997) 343.
[34] <http://dtp.nci.nih.gov/>.
Acknowledgments
[35] P. Skehan, R. Storeng, D. Scudiero, A. Monks, J. Mc Mahon, D. Vistica,
J.T. Warren, H. Bokesh, S. Kenney, M.R. Boyd, J. Natl. Cancer Inst. 82
(1990) 1107.
The authors are grateful to the staff of the Department of
Health and Human Services, National Cancer Institute, Be-
thesda, MD, USA, for the anti-cancer screening. This work
was supported by MIUR (Italian Ministry of University and
Research) Cofin04_2004030405_004 CSTMPL and Cofin04
Paglietti.
[36] J.T. Kealey, D.V. Santi, Prot. Expr. Purif. 4 (1992) 380.
[37] G.F. Maley, F. Maley, J. Biol. Chem. 263 (1988) 7620.
[38] V. Davisson, W. Sirawaraporn, D.V. Santi, J. Biol. Chem. 264 (1989)
9145.
[39] J.G. Dann, G. Ostler, R.A. Bjur, W.R. King, P. Scudder, G.C.K. Roberts,
S.V. Burgen, Biochem. J. 157 (1976) 559.
References
[40] M.P. Costi, A. Gelain, D. Barlocco, S. Ghelli, F. Soragni, F. Reniero,
T. Rossi, A. Ruberto, C. Guillou, A. Cavazzuti, C. Casolari, S. Ferrari,
J. Med. Chem. 49 (20) (2006) 5958.
[1] I.M. Kompis, K. Isalam, R.L. Then, Chem. Rev. 105 (2005) 593 and
references cited therein.
[41] A.L. Pogolotti Jr., P.V. Danenberg, D.V. Santi, J. Med. Chem. 29
(1986) 478.
[42] T.D. Meek, E.P. Garvey, D.V. Santi, Biochemistry 24 (1985) 678.
[2] N. Nemati, U.S. Pat. Appl. Pubbl. US 2006 235 034. Chem. Abstr. 145
(2005) 920.

P. Corona et al. / European Journal of Medicinal Chemistry 43 (2008) 189e203
203
[43] I.H. Segel, Enzyme Kinetics: Behaviour and Analysis of Rapid Equilib-
rium and Steady-state Enzyme Systems, John Wiley and Sons, Inc., New
York, 1975, 105.
[45] G. Vriend, J. Mol. Graph. 8 (1990) 52.
[46] Molecular Graphics Laboratory, The Scripps Research Institute.
[47] Cambridge Soft Corporation.
[44] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart,
R.K. Belew, A.J. Olson, J. Comput. Chem. 19 (1998) 1639.
[48] J.I. DeGraw, P.H. Christie, H. Tagawa, R.L. Kisliuk, Y. Gaumont,
F.A. Schmid, F.M. Sirotnak, J. Med. Chem. 29 (1986) 1056.