Design, synthesis, biological evaluation and computational investigation of novel inhibitors of dihydrofolate reductase of opportunistic pathogens

Article abstract of DOI:10.1016/j.bmc.2010.03.031

The present work deals with design, synthesis and biological evaluation of novel, diverse compounds as potential inhibitors of dihydrofolate reductase (DHFR) from opportunistic microorganisms; Pneumocystis carinii (pc), Toxoplasma gondii (tg) and Mycobacterium avium (ma). A set of 14 structurally diverse compounds were designed with varying key pharmacophoric features of DHFR inhibitors, bulky distal substitutions and different bridges joining the distal part and 2,4-diaminopyrimidine nucleus. The designed compounds were synthesized and evaluated in enzyme assay against pc, tg and ma DHFR. The rat liver (rl) DHFR was used as mammalian standard. As the next logical step of the project, flexible molecular docking studies were carried out to predict the binding modes of these compounds in pcDHFR active site and the obtained docked poses were post processed using MM-GBSA protocol for prediction of relative binding affinity. The predicted binding modes were able to rationalize the experimental results in most cases. Of particular interest, both the docking scores and MM-GBSA predicted ΔG_bind were able to distinguish between the active and low active compounds. Furthermore, good correlation coefficient of 0.797 was obtained between the IC₅₀ values and MM-GBSA predicted ΔG_bind. Taken together, the current work provides not only a novel scaffold for further optimization of DHFR inhibitors but also an understanding of the specific interactions of inhibitors with DHFR and structural modifications that improve selectivity.

Full text of DOI:10.1016/j.bmc.2010.03.031

Bioorganic & Medicinal Chemistry 18 (2010) 3187–3197
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Design, synthesis, biological evaluation and computational investigation
of novel inhibitors of dihydrofolate reductase of opportunistic pathogens
b
Seema Bag a, Nilesh R. Tawari a, Mariam S. Degani a, , Sherry F. Queener
*
a Department of Pharmaceutical Sciences and Technology, Institute of Chemical Technology, Matunga (E), Mumbai 400 019, India
b Department of Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, IN 46202, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The present work deals with design, synthesis and biological evaluation of novel, diverse compounds as
potential inhibitors of dihydrofolate reductase (DHFR) from opportunistic microorganisms; Pneumocystis
carinii (pc), Toxoplasma gondii (tg) and Mycobacterium avium (ma). A set of 14 structurally diverse com-
pounds were designed with varying key pharmacophoric features of DHFR inhibitors, bulky distal substi-
tutions and different bridges joining the distal part and 2,4-diaminopyrimidine nucleus. The designed
compounds were synthesized and evaluated in enzyme assay against pc, tg and ma DHFR. The rat liver
(rl) DHFR was used as mammalian standard. As the next logical step of the project, flexible molecular
docking studies were carried out to predict the binding modes of these compounds in pcDHFR active site
and the obtained docked poses were post processed using MM-GBSA protocol for prediction of relative
binding affinity. The predicted binding modes were able to rationalize the experimental results in most
Received 12 January 2010
Revised 12 March 2010
Accepted 13 March 2010
Available online 18 March 2010
Keywords:
Dihydrofolate reductase
2,4-Diaminopyrimidine
Docking
MM-GBSA
cases. Of particular interest, both the docking scores and MM-GBSA predicted
guish between the active and low active compounds. Furthermore, good correlation coefficient of 0.797
was obtained between the IC50 values and MM-GBSA predicted Gbind. Taken together, the current work
DGbind were able to distin-
D
provides not only a novel scaffold for further optimization of DHFR inhibitors but also an understanding
of the specific interactions of inhibitors with DHFR and structural modifications that improve selectivity.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
problem in this treatment. In addition, non-classical antifolates,
trimetrexate and piritrexim co-administered with leucovorin for
Immunocompromised hosts such as HIV positive people are of-
ten prone to life threatening infections caused by opportunistic
microorganisms including Pneumocystis carinii (pc), Toxoplasma
gondii (tg), and Mycobacterium avium (ma).1 Dihydrofolate reduc-
tase (DHFR), a NADPH dependent enzyme, catalyses the crucial
reaction of conversion of folic acid to dihydro and tetrahydro folic
acid (a cofactor involved in one carbon donation in purine and
pyrimidine de novo synthesis) both in mammals as well as micro-
organisms. Inhibitors of DHFR, which is essential for cell growth
and division, have been in clinical use for over 50 years for antican-
cer (methotrexate), antibacterial (trimethoprim) and antiprotozoal
(pyrimethamine) treatments. Current therapeutic treatment for
infections caused by pc and tg include co-administration of the
selective but weak DHFR inhibitor, trimethoprim or pyrimeth-
amine, in combination with sulfonamides to enhance potency.2
However, side effects related to sulfonamides is often a serious
host rescue are also used for the treatment of these infections.2
However, often, toxicity forces the termination of treatment. DHFR
inhibitors are yet to be clinically applied for the treatment of infec-
tions caused by ma. Thus, DHFR inhibitors with high potency and
selectivity are desired for the treatment of infections caused by
these pathogenic organisms.
The presence of overall structurally conserved DHFR in both hu-
mans and microorganisms makes the drug design process for antif-
olates challenging. Several laboratories continue to work towards
the search for potent and selective DHFR inhibitors against oppor-
tunistic microorganisms. Rosowsky et al. have reported structure
based drug design of several novel inhibitors with a bulky tricyclic
moiety at 6-position of the 2,4-diaminopteridine nucleus.3 In a
subsequent study, the 3D structure of the ternary complex of one
of these inhibitors, 6-((5H-dibenzo[b,f]azepin-5-yl)methyl)pteri-
dine-2,4-diamine, I, (Fig. 1) and NADPH with pcDHFR was analyzed
using crystallography, highlighting an unusual role of the di-
benz[b,f]-azepine moiety in positioning the diaminopteridine
within the active site (PDB code: IKLK).4 Attempts were made to
further enhance potency of this inhibitor, I, by replacing the pter-
idine ring with a pyrido[2,3-d]pyrimidine or quinazoline ring, or
Abbreviations: pc, Pneumocystis carinii; tg, Toxoplasma gondii; DHFR, dihydro-
folate reductase; HIV, human immunodeficiency virus; ma, Mycobacterium avium.
* Corresponding author. Tel.: +91 22 33612213; fax: +91 22 24145614.
E-mail address: ms.degani@ictmumbai.edu.in (M.S. Degani).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.03.031

3188
S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
compounds, 5a–b, 7a and 8c–d, were designed with various
bridges joining the distal hydrophobic substitution and 2,4-diami-
nopyrimidine ring system. Compounds, 6a–b, were designed with
conformationally constrained bridge which could possibly lock the
2,4-diaminopyrimidine nucleus and distal substitution in preferred
binding orientation. Compounds, 11a–b, were designed with
6-substituted 2,4-diaminopyrimidine nucleus.
Figure 1. Structures of some of the lipophilic DHFR inhibitors reported in literature.
2. Results and discussion
2.1. Synthesis
by changing the azepine ring nitrogen and –CH@CH– bridge by dif-
ferent isosteres. However, these studies did not yield encouraging
results.5 Furthermore, 20-position of the dibenz[b,f]azepine ring
was substituted with carboxylic acid side chain to improve potency
and selectivity.6 Another approach by Graffner-Nordberg et al. in-
cluded design, synthesis and computational affinity prediction of
ester soft drugs with lipophilic tricyclic ring, II, (Fig. 1) substituted
on 2,4-diaminoquinazoline nucleus as DHFR inhibitors.7
Our research group has been recently involved in the search for
novel DHFR inhibitors. Towards this goal we have reported molec-
ular modeling studies against maDHFR to get an insight into the po-
tential features required for maDHFR inhibition.8 Initially, we have
reported a series of extended biguanides and dihydrotriazines as
potential DHFR inhibitors against opportunistic microorganisms.9
Furthermore, based on virtual screening experiments, we have de-
signed and evaluated novel 2-hydrazino-pyrimidin-4(3H)-one
derivatives, lacking conventional 2,4-diaminopyrimidine nucleus
as inhibitors of DHFR from opportunistic microorganisms.10 While
several synthesized compounds showed promising in vitro activity
against microbial DHFR, many of these compounds were also equi-
potent inhibitors of mammalian DHFR thereby decreasing their
usefulness.
The current study was aimed at synthesis and biological evalu-
ation of a diverse set of molecules followed by their molecular
modeling studies for the understanding of specific interactions of
the inhibitors with DHFR and structural modifications that im-
prove selectivity. In this light, based on the literature reports3–6
and understanding of DHFR active site, a set of 14 structurally di-
verse compounds was designed; with varying the nitrogen con-
taining heterocycle, bulky tricyclic distal substitutions or
different bridges joining the distal part and 2,4-diaminopyrimidine
nucleus. Thus, compounds, 3a–c, were specifically designed in or-
der to check the effect of addition of bulky tricyclic moiety on
our newly proposed pharmacophore; 2,4-diamino s-triazine for
inhibitors of DHFR from opportunistic microorganisms, which
structurally mimics the conventional 2,4-diaminopyrimidine phar-
macophore of DHFR. The 2,4-diamino groups on s-triazine are ex-
pected to form crucial H-bonding interactions with conserved
residues at active site of DHFR. Using pharmacophoric features of
trimethoprim, compounds 8a–b, were designed with the bulky
dibenz[b,f]azepine and dihydrodibenz[b,f]azepine moiety substi-
tuted on 2,4-diaminopyrimidine moiety. In an attempt to probe
steric and electronic tolerance of the DHFR binding pocket,
s-Triazine analogs 3a–c, were synthesized using the route
shown in Scheme 1. This involved N-alkylation of the tricyclic
nucleus (1a–c). The dihydroiminostilbene (1a) was reacted with
chloroacetonitrile using K2CO3 as a base and heating the reaction
mixture in presence of catalytic amount of PEG 6000 and water.11
The column purified product was then reacted with dicyandiamide
in presence of KOH, water using 2-methoxyethanol as solvent to
afford target compound 3a. N-Alkylation of 10H-phenothiazine
(1b) and 2-chloro-10H-phenothiazine (1c) were carried out using
NaH as base and DMSO/DMF as solvent.12 After column purifica-
tion, these were reacted with dicyandiamide, using 2-methoxyeth-
anol as solvent to synthesize the desired s-triazines (3b, 3c).13
The starting material 2,4-diamino-5-cyano pyrimidine (5,
Scheme 2) was synthesized by reacting guanidine and ethoxym-
ethylene malononitrile according to the method reported by Hu-
ber.14 Molecules 5a and 5b were synthesized from 2,4-diamino-
5-cyano pyrimidine (5, Scheme 2) using reductive alkylation in
presence of Raney nickel.15 2,4-Diamino-5-cyano pyrimidine (5,
Scheme 2) was converted to 2,4-diaminopyrimidine-5-carboxalde-
hyde (6, Scheme 2) using Raney nickel with 98–100% formic acid as
a solvent.14–17
Compounds 6a and 6b were synthesized from 2,4-diaminopyr-
imidine-5-carboxaldehyde (6, Scheme 2) by Claisen–Schmidt con-
densation. 2,4-Diamino-5-carboxaldhyde (6, Scheme 2) was
reduced to (2,4-diamonopyrimidine-5-yl) methanol (7, Scheme 2)
using NaBH4. NaBH4 (1.5 equiv) in 100 ml methanol were required
for reduction as 6 was sparingly soluble in solvent. It was observed
that more than 1 equiv of NaBH4 was required for the completion
of the reaction. Compound 7a was synthesized from (2,4-diamono-
pyrimidine-5-yl) methanol (7, Scheme 2) and naphthalene-2-sul-
fonyl chloride employing triethylamine as catalyst. 5-(Bromo-
methyl)pyrimidine-2,4-diamine (8, Scheme 2) was synthesized
from (2,4-diamonopyrimidine-5-yl) methanol (7, Scheme 2) using
HBr in glacial acetic acid.18 The product formed 8, was highly
unstable, therefore immediate azeotropic distillation to remove
solvent was followed by next reaction without characterization.
Synthesis of 8a–b involved N-alkylation of tricyclic ring system.
The reaction was performed under inert atmosphere with NaH in
dry solvents. Compounds, 8c–d, were synthesized by using reac-
tion of 5-(bromomethyl)pyrimidine-2,4-diamine (8, Scheme 2)
Scheme 1. Scheme for the synthesis of lipophilic s-triazine analogs.

S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
3189
Scheme 2. Scheme for the synthesis of 2,4-pyrimidine analogs.
Scheme 3.
with N-methylnaphthalen-1-amine19 and mercaptobenzoxazole.
Reaction of 4-chloro-2,6-diaminopyrimidine with the appropriate
sodium alkoxide (ROH, NaH, DMSO, 100 °C) afforded correspond-
ing O4-substituted pyrimidines, 11a–b, (Scheme 3).
pounds were also found to be selective inhibitors of microbial
DHFR compared to rat liver DHFR. In particular, compound, 8b,
inhibited all microbial DHFRs with some selectivity at concentra-
tion <1 lM.
Compounds, 5a–b, and 8c were designed by incorporating an
electronegative nitrogen atom as part of bridge joining the 2,4-dia-
minopyrimidine ring system and distal substitution. Compound,
2.2. Biology
The synthesized compounds were evaluated for their ability to
inhibit DHFR of pc, tg, ma and rat liver (rl).20,21 The rl DHFR was
used as mammalian standard as it shares high sequence similarity
to human DHFR. The data for the reported compounds (trimeth-
prim, I and II) are taken from literature reports3,7 for comparative
purpose. IC50 values for all analogs are given in Table 1.
5a, was able to inhibit pcDHFR at concentration of ꢀ363
ever, compound, 5b, with a bulky napthyl substitution at distal part
was found to be a better pcDHFR inhibitor with IC50 of ꢀ150 M.
M with
over ꢀ6-fold selectivity to that of rat liver DHFR. Interestingly, addi-
tion of a methyl group on the bridge nitrogen atom caused dramatic
increase in DHFR inhibitory activity as evident from compound, 8c,
lM. How-
l
Compound, 5b, also showed inhibition of tgDHFR at ꢀ20
l
Biological activity results revealed that the synthesized s-tri-
azine analogs, 3a–c, were less active against all DHFRs. These com-
pounds were able to inhibit pcDHFR at concentrations from ꢀ600
which showed inhibition of maDHFR at concentration < 1
with > 200-fold selectivity over rat liver DHFR.
lM
to 1200
lM. Only one compound, 3c, was able to inhibit tgDHFR
Compounds, 6a–b, were not found to be inhibitors of DHFR, one
at ꢀ400
l
M concentration. In contrast, the molecules, 8a–b, with
of the possible reasons for this behavior could be high conforma-
bulky tricyclic substitutions on 2,4-diaminopyrimidine ring system
tional restriction owing to presence of
bridge.
a,b-unsaturated carbonyl
were good inhibitors of microbial DHFRs. Furthermore, these com-

3190
S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
Table 1
Inhibitory concentration (IC50
lM) against rlDHFR, pcDHFR, tgDHFR and maDHFR and selectivity ratios versus rlDHFR by target compounds
Mol ID
MW
Activity (
l
M)
Selectivity
rl/tg
pc
tg
ma
rl
rl/pc
rl/ma
3a
3b
3c
5a
5b
6a
6b
7a
8a
8b
8c
318.4
322.4
356.8
245.3
265.3
270.3
360.4
330.4
315.4
317.4
279.3
273.3
306.1
345.4
—
1265.15
235.46
603.65
363.26
150.58
151.319 (4)
11.653 (2)
7.88
26.25
0.84
38.48
35.83
148.635 (0)
112.02
13
0.21
0.49
139.770 (10)
142.684 (15)
397.11
44.031 (7)
22.13
151.318 (15)
116.533 (10)
124.107 (13)
8.07
0.14
3.49
160.22
145.272 (10)
3.35
138.199 (9)
142.684 (4)
117.703 (5)
44.031 (0)
79.152 (9)
151.318 (14)
11.653 (4)
124.107 (15)
6.35
0.11
0.88
23.14
145.370 (12)
2.33
139.77 (0)
142.68 (16)
117.7 (8)
44.03 (0)
127.85
ND
ND
ND
ND
0.85
ND
ND
>15.73
>5
1.89
4.79
14.53
ND
ND
ND
ND
ND
5.78
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
127.42
178.130 (0)
124.107 (0)
130.00 (15)
1.60
184.26
520.65
148.635 (12)
119.861 (39)
180
4.4
0.21
ND
>16
11.48
52.74
3.25
ND
>35.77
65
102
0.31
>20
14.36
209.312
22.50
ND
>51.44
600
367
0.21
8d
11a
11b
TMPa
Ia
ND
14
21
0.43
2.80
0.043
0.67
0.30
0.012
1.0
—
—
IIa
Triplicate assays were performed as previously described.20,21
Number in parentheses indicates percent inhibition obtained at that concentration.
a
Data taken from Refs. 3,7.
Compounds, 7a and 8d, with electronegative bridges between
2,4-diaminopyrimidine ring system and distal substitution showed
modest DHFR inhibitory activity. Of particular interest, compound,
7a, was found to be selective for pcDHFR with approximately two-
fold increase in activity and good selectivity when compared to
trimethoprim.
Among the compounds, 11a–b, compound 11a was found to be
inactive in all species. However, compound, 11b, has shown inter-
esting activity of <5
selectivity.
lM against tg and ma DHFR with good
2.3. Computational studies
To understand the forms of interactions of synthesized com-
pounds with enzyme and to identify the binding modes of these
compounds, we created an accurate docking model for pcDHFR.
pcDHFR was selected as the enzyme of choice as the experimen-
tally determined crystal structure of the enzyme was available in
public domain. One of the most important considerations in dock-
ing simulations is selection of appropriate docking algorithm and
scoring function. To evaluate the effectiveness of Glide protocol
in flexible docking of DHFR inhibitors, the model of pcDHFR de-
rived from crystallographic data (PDB ID: 1KLK) extending to
2.3 Å resolution for a ternary complex containing a bound mole-
cule of NADPH and an DHFR inhibitor I, was selected as the docking
receptor. The model preparation and docking studies were carried
out using the protocol outlined in the experimental section. Since,
DHFR has a highly conserved ligand binding orientation (Fig. 2),
‘must match’ of at least one hydrogen bond criteria; with Ile123,
Ile10 and Glu32 residues was constrained during docking simula-
tions. Using flexible docking procedure in Glide-XP, the observed
Figure 2. Experimentally observed conserved geometry in the active site of DHFR
for compounds with a 2,4-diaminopyrimidine nucleus.
with the key active site residue Glu32, thereby rationalizing the
low experimentally observed activity for this compound. The s-tri-
azine ring in compounds, 3b–c, (Fig. 3b and c) showed additional
p–p stacking interaction with phenyl ring of Phe36 (centroid–cen-
troid distance 4.1 Å for 3b and 4.4 Å for 3c) at the bottom of active
site. However, in compound, 3b–c, (Fig. 3b and c) the s-triazine ring
was not orientated properly, possibly due to the presence of a one
carbon bridge joining the tricyclic moiety and s-triazine ring sys-
tem which is too short. Single hydrogen bonding interaction with
Glu32 in, 3b–c, could be the reason for better activity of these com-
pounds compared to 3a. However, the absence of the key bipolar
interaction with Glu32 as depicted in Figure 2, at the bottom of ac-
tive site was responsible for the low activity of, 3a–c.
crystallographic structure was reproduced with RMSD value of
0
0.65 ÅA indicating that this docking method is valid and can be used
for predicting binding modes of pcDHFR inhibitors. Once the dock-
ing algorithm was validated, studies were further extended for
molecular docking of synthesized molecules in the active site of
pcDHFR. The top scoring poses returned by XP scoring function
were analyzed using Maestro graphical interface.
In compounds, 3a–c and 8a–b, bulky tricyclic moiety showed
desirable hydrophobic interactions with the hydrophobic residues
Leu32, Ile65, Phe69, Leu25 and Ile33 in the distal part of active site.
However, compound, 3a, (Fig. 3a) did not show any interaction
In contrast to the bulky s-triazine derivatives, in compounds
8a–b (Fig. 3g), the 2,4-diaminopyrimidine ring orients properly in
the active site. The 4-amino group in these compounds is in-
volved in hydrogen bonding interaction with Ile123, while the

S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
3191
Figure 3. Predicted binding modes of the synthesized compounds in the active site of pcDHFR; (a) binding mode of compound 3a (off white color); (b) binding mode of
compound 3b (plum color); (c) binding mode of compound 3c (yellow color); (d) binding mode of compound 5a–b (5a, green color; 5b, yellow color); (e) binding mode of
compound (6a–b); (f) binding mode of compound 7a (green color); (g) binding mode of compound 8a–b (8a, off plum color; 8b, green color); (h) binding mode of compound
8c (green color); (i) binding mode of compound 8d (violet color); (j) binding mode of compounds 11a–b (11a, violet color; 11b, green color).
2-amino group is involved in hydrogen bonding interaction
with Glu32, thereby making these compounds better inhibitors
than 3a–c. Furthermore, dihydrodibenz[b,f]azepine analog, 8b,
showed proper orientation of the tricyclic moiety in the active
site.
Compounds, 5a–b and 8c, showed conserved interactions of
2,4-diaminopyrimidine ring system with Ile123, Ile10 and Glu32.
The methyl group of p-OCH3 substitution in compound, 5a,
(Fig. 3d) was involved in hydrophobic interaction with Phe69. In
compound, 5b, (Fig. 3d) the napthyl ring system showed additional

3192
S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
Fig. 3 (continued)
hydrophobic interactions with Leu32, Ile65, Phe69, Leu25 and
Ile33 residues of the active site. No direct interactions were ob-
served with the bridge; however it mainly affects the orientation
of the distal ring. In compound, 8c, (Fig. 3h) the methyl group on
bridge showed hydrophobic interaction with Leu25, in addition,
presence of this methyl group helps to orient the napthyl moiety
deeper in the hydrophobic pocket which can be attributed to ob-
served good activity of this compound.
Compounds, 11a–b, (Fig. 3j) showed H-bonding interactions of
2,4-diaminopyrimidine ring system with Ile123, Ile10 and Glu32.
However, key bipolar interaction with active site residue Glu32
was missing; this could be the reason for observed less activity.
Furthermore, 11b, forms several good hydrophobic interactions
with Leu25, Ile33, Phe36, Ile65 and Phe69 which explains the bet-
ter activity of 11b over 11a.
In order to predict the relative binding affinities of these com-
The reason for inactivity of, 6a–b, (Fig. 3e) could not be antici-
pated from the predicted binding modes as both compounds show
conserved interactions with Ile123, Ile10 and Glu32, in addition,
favorable hydrophobic interactions with the distal part were also
seen. As mentioned previously, high conformational restriction
could be the reason for inactivity of these compounds.
The extended electronegative bridges in compounds, 7a (Fig. 3f)
and 8d, (Fig. 3i) possess both the 2,4-diaminopyrimidine and distal
ring in correct orientation. In both compounds, 4-amino group
showed hydrogen bonding interactions with Ile123 and Ile10. In
particular, the 2-amino group and protonated N1 in both the com-
pounds showed bipolar conserved interaction with Glu32 of active
site which is essential requirement for DHFR inhibitory activity.
The napthyl ring of compound, 7a, showed good hydrophobic
interactions with Leu32, Ile65, Phe69, Leu25 and Ile33, thus mak-
ing this compound active and selective for pcDHFR.
pounds, Glide-XP scores and binding energy
DGbind values calcu-
lated using MM-GBSA protocol were compared with
experimental IC50 values. This comparison of docking scores,
D
Gbind and IC50 values clearly shows that the relative binding affin-
ities of two compounds, 6a–b, were overestimated by both meth-
ods. Also, the experimental IC50 of compound, 11a, was not exactly
determined. Hence, these compounds were considered as outliers
and were excluded from further regression analysis.
The predicted Glide-XP score values vary from ꢁ12.18 to
ꢁ5.75 kcal/mol (Table 2). Glide-XP scores showed a fair correlation
coefficient of 0.538 (Fig. 4) with the experimentally determined
IC50 values. Interestingly, the most inactive ligand, 11a, was also
predicted to be lowest active using Glide-XP score. Furthermore,
Glide-XP score was able to clearly distinguish between active and
inactive ligands. For instance, all the compounds with good activity
7a, 8a–d, I and trimethoprim, scored higher than the low active

S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
3193
Table 2
the scale of 10–225 kcal/mol. The most positive, deepest blue color
colored regions by the hydrogen of both amino groups and hydro-
gen of protonated N1, are in accordance with the H-bonding with
receptor active site. The least positive potential regions (deepest
red color), were seen in the vicinity of oxygen of sulfonyl group.
Comparison of experimental binding energy with docking scores and MM-GBSA
predicted binding affinities
DGbind against pcDHFR
Mol ID
pcDHFR activity
IC50 M)
XP
Evdwa
Ecoula
DGbind
(l
GScorea
3a
3b
3c
5a
5b
6a
6b
7a
8a
8b
8c
8d
11a
11b
TMP*
I*
1265.15
235.46
603.65
363.26
150.58
151.319 (4)
11.653 (2)
7.88
26.25
0.84
38.48
35.83
ꢁ6.29
ꢁ7.09
ꢁ22.66
ꢁ34.29
ꢁ31.23
ꢁ24.51
ꢁ28.72
ꢁ31.97
ꢁ35.94
ꢁ32.72
ꢁ33.25
ꢁ29.61
ꢁ27.77
ꢁ26.09
ꢁ40.10
ꢁ29.90
ꢁ30.34
ꢁ42.74
ꢁ2.92
ꢁ1.05
ꢁ2.26
ꢁ7.65
ꢁ6.82
ꢁ7.62
ꢁ5.77
ꢁ8.76
ꢁ4.44
ꢁ2.87
ꢁ6.14
ꢁ7.51
ꢁ7.19
ꢁ7.08
ꢁ7.42
ꢁ7.64
ꢁ15.95
ꢁ34.95
ꢁ32.24
ꢁ37.77
ꢁ36.57
ꢁ43.02
ꢁ52.01
ꢁ45.08
ꢁ34.43
ꢁ38.23
ꢁ42.30
ꢁ36.26
ꢁ40.84
ꢁ42.36
ꢁ41.50
ꢁ45.39
3. Conclusion
ꢁ7.26
ꢁ8.50
Fourteen novel, diverse compounds with varying pharmaco-
phoric features, bulky distal substitutions and different bridges
were designed and synthesized as DHFR inhibitors. The synthe-
sized compounds were evaluated against DHFR from opportunistic
microorganisms using in vitro assays. Some of the synthesized
compounds showed appreciable selective activity against DHFR
from opportunistic microorganisms. In particular compound, 7a,
was found to be selective for pcDHFR with approximately twofold
increase in activity and good selectivity when compared to tri-
methoprim. The observed experimental results were used to estab-
lish and understand binding modes of these compounds using
molecular modeling studies. Thus, the present study provides not
only novel scaffold for the further optimization of DHFR inhibitors
but also an understanding the specific interactions of inhibitors
with DHFR and structural modifications that improve selectivity.
ꢁ9.62
ꢁ10.17
ꢁ11.83
ꢁ10.82
ꢁ11.34
ꢁ9.26
ꢁ11.02
ꢁ9.24
148.635 (0)
112.02
13.00
ꢁ5.48
ꢁ7.85
ꢁ10.15
ꢁ12.19
0.21
*
TMP is trimethoprim and I is depicted in Fig. 1.
a
XP GScore is the predicted binding energy using Glide in kcal/mol; Evdw and
Ecoul are the vdW and electrostatic interaction energies between ligand and receptor
calculated using Glide-XP; Gbind is predicted binding affinity using MM-GBSA.
D
compounds 3a–c and 5a–b. Inspection of Evdw and Ecoul terms
which are van der Waals interaction energy and electrostatic inter-
action energy predicted by Glide-XP revealed that all ligands share
higher values of van der Waals interaction energy which is logical
considering the lipophilic distal part substitutions. Furthermore, a
striking observation was that, for all most active compounds (ex-
cept 8b) the electrostatic interaction energy term is higher than
the inactive compounds, which is in accordance with the con-
served bipolar interaction with Glu32 at bottom of active site in
the most active compounds.
In order to improve correlation between the observed IC50 val-
ues and predicted binding affinity, MM-GBSA post docking scoring
protocol was employed. Concisely, various energy components
which contribute to binding energy such as free energy, solvation
energy and surface area energy were calculated for the complex
holoenzyme, apoenzyme and free ligand; binding energy was cal-
culated as the sum of difference between energy of complex holo-
enzyme and sum of energy of apoenzyme and free ligand. The
results of binding affinity prediction using MM-GBSA are tabulated
in Table 2. Interestingly, the MM-GBSA predicted binding energy
4. Experimental
4.1. General considerations
Melting points were recorded on Thermomix Compbell elec-
tronics, having oil-heating system and were uncorrected. The
microwave reactions were carried out using CEM Focused Micro-
wave System in monomode, Model Discover. Analytical thin-layer
chromatography (TLC) was carried out on precoated plates SiO2
(Silica Gel 60, F 254, Merck). SiO2 (Silica gel 420, Merck) was used
for column chromatography using CombiFlashÒ RETRIEVEÒ sys-
tem. FTIR spectra were recorded on ‘Buck scientific infrared spec-
troscopy M500 spectrophotometer’ using KBr pellets. NMR
spectra were recorded on JEOL AL 300 MHz spectrometer or Bruker
DMX-500 spectrometer operating at 500 MHz with DMSO-d6 as
solvent; s = singlet, d = doublet, t = triplet, q = quartet, m = muti-
plet, br = broad. The mass spectrum was recorded on a Waters Q/
TOF Micromass spectrometer. All the reagents and chemicals used
were of ‘reagent grade’.
D
Gbind could clearly distinguish between active and inactive li-
4.2. Chemistry
gands. Also, improved correlation coefficient of 0.797 (Fig. 4) was
obtained between IC50 values and predicted binding affinity.
To understand the electronic distribution of molecule in the ac-
tive site the docked pose conformer of compound, 7a, was sub-
jected to single point energy calculation using density functional
theory (DFT) B3LYP/3-21G* level of basis set. The calculated molec-
ular electrostatic potential (MESP) map is depicted in Figure 5 on
4.2.1. 6-((10,11-Dihydro-5H-dibenzo[b,f]azepin-5-yl)methyl)-
1,3,5-triazine-2,4-diamine (3a)
A solution of 1a (1 g, 0.0051 mol), K2CO3 (0.71 g, 0.0051 mol),
PEG 6000 (125 mg, 12.5% w/w), water (1 ml) in chloroacetonitrile
(5 ml) was heated at 150 °C for 24 h. Column purification of the
Figure 4. Correlation between Glide-XP scores, MM-GBSA predicted binding affinities and experimentally determined IC50 values against pcDHFR.

3194
S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
Figure 5. MESP superimposed onto a surface of constant electron density (0.01 e/au3) for compound, 7a, showing the most positive potential region (deepest blue color) by
hydrogen of both amino groups and hydrogen of protonated N1 and the least positive potential regions (deepest red color) were seen in the vicinity of oxygen of sulfonyl
group.
reaction mixture yielded 0.38 g (31.7%) of 2a.11 Compound 2a
(0.38 g, 0.00162 mol), dicyandiamide (0.15 g, 0.0017 mol), was
stirred in 2-methoxyethanol (10 ml) in presence of KOH (0.25 g)
and water (0.2 ml). The resultant reaction mixture was subjected
to microwave irradiation at a power of 100 for 20 min to attain
the target temperature of 150 °C. TLC monitoring showed forma-
tion of product along with the unreacted starting material. The
reaction mixture was poured into cold water resulting in solid
precipitate. Filtration followed by washings with hot water
(2 ꢂ 20 ml) and with 20% EtOAc/n-hexanes mixture to remove
unreacted 2a, yielded 0.046 g (yield: 9%) of 6-((10,11-dihydro-
5H-dibenzo[b,f]azepin-5-yl)methyl)-1,3,5-triazine-2,4-diamine (3a,
Scheme 1) as brown solid. Mp = 227 °C. IR (KBr) Vmax cmꢁ1 3463,
3297, 3178, 3102, 1616, 1565, 1543, 1506, 1486, 1453, 1384,
1233; 1H NMR (DMSO-d6, 300 MHz): d 7.15–7.04 (m, 6H, Ar–H),
6.86–6.82 (t, 2H, Ar–H), 6.58 (s, 4H, D2O exchangeable –NH2),
4.56 (s, 2H, –CH2), 3.11 (s, 4H, ring –CH2–CH2–, overlapped with
H2O peak of DMSO-d6); 13C NMR (DMSO-d6, 75 MHz): d 167.0,
167.0, 167.0, 147.8, 133.4, 129.5, 126.1, 121.8, 120.2, 58.6, 32.4;
M/S 319.33 [M+1]+.
1599, 1460, 1394, 1247, 1212, 1097; 1H NMR (DMSO-d6,
300 MHz): d 7.73 (s, 1H, D2O exchangeable –NH2), 7.34 (s, 1H, D2O
exchangeable –NH2), 7.16–7.07 (m, 3H, Ar–H; 1H, D2O exchangeable
–NH2), 6.97–6.90 (m, 3H, Ar–H; 1H, D2O exchangeable –NH2), 6.68
(s, 3H, Ar–H), (s, 2H, –CH2); 13C NMR (DMSO-d6,75 MHz): d 169.9,
169.9, 169.8, 145.7, 143.6, 131.9, 127.6, 127.4, 126.5, 123.1, 122.0,
121.0, 120.4, 115.4, 115.1, 50.9; M/S 321.28 [Mꢁ1]+.
4.2.3. 6-((2-Chloro-10H-phenothiazin-10-yl)methyl)-1,3,5-
triazine-2,4-diamine (3c)
NaH (0.28 g, 0.0064 mol) was added to a solution of 1c (Scheme
1) (1 g, 0.0043 mol) in dry DMF (5 ml) and reaction mixture was stir-
red at room temperature for 20 min. Chloroacetonitrile (0.32 g,
0.0043 mol) was added and reaction mixture was heated on water
bath at 60–80 °C for 3 h, followed by stirring at room temperature
for 12 h. Column chromatography led to 0.20 g, (yield: 17.15%),
whitish pink solid of 2c.12 The solution of 2c (0.2 g, 0.00073 mol)
in methoxyethanol (10 ml) was treated with dicyandiamide
(62 mg, 0.00073 mol), KOH (0.2 g) and water (0.2 ml). The resultant
reaction mixture was heated at 100 °C for 12 h. Column chromatog-
raphy yielded 0.039 g (yield: 25%) of 6-((2-chloro-10H-phenothia-
zin-10-yl)methyl)-1,3,5-triazine-2,4-diamine (3c, Scheme 1) as
dark green solid. Mp = 197 °C; IR Vmax cmꢁ1 3467, 3307, 3100,
1652, 1627, 1559, 1544, 1312, 1255, 1209, 803; 1H NMR (DMSO-
d6, 300 MHz): d 7.03–7.01 (m, 3H, Ar–H; 1H, D2O exchangeable
–NH2), 6.91 (2H, Ar–H; 1H, D2O exchangeable –NH2), 6.75–6.65
(2H, Ar–H; 2H, D2O exchangeable –NH2), 4.65 (s, 2H, –CH2); 13C
NMR (DMSO-d6, 75 MHz): d 173.8, 166.9, 145.5, 143.4, 131.9,
127.6, 127.1, 126.2, 122.8, 121.8, 120.5, 120.0, 116.0, 115.9, 115.5,
53.8; M/S 357.24 [M+1]+, 359.19.
4.2.2. 6-((10H-Phenothiazin-10-yl)methyl)-1,3,5-triazine-2,4-
diamine (3b)
NaH (0.22 g, 0.0050 mol) was added to a solution of 1b (1 g,
0.0050 mol) in dry DMF (5 ml) and reaction mixture was stirred
at room temperature for 20 min. Chloroacetonitrile (0.38 g,
0.0050 mol) was added and reaction mixture was heated on a water
bath at 60–80 °C for 3 h followed by stirring at room temperature for
12 h. Column chromatography yielded 0.47 g, (yield: 40%), pale
white solid of 2b.11 The solution of 2b (0.2 g, 0.00084 mol) in
methoxyethanol (10 ml) was treated with dicyandiamide (0.071 g,
0.00084 mol), KOH (0.2 g) and water (0.2 ml).13 The resultant reac-
tion mixture was heated at 100 °C for 12 h. Column chromatography
yielded 0.139 g (yield: 51.48%) of 6-((10H-phenothi azin-10-yl)-
methyl)-1,3,5-triazine-2,4-diamine (3b, Scheme 1) as dark green so-
lid. Mp = 226–228 °C; IR (KBr) Vmax cmꢁ1 3460, 3218, 2911, 1647,
4.2.4. 2,4-Diaminopyrimidine-5-carbonitrile (5, Scheme 2)14–17
To the freshly prepared sodium ethoxide (sodium (1.0 g,
0.0438 mol) in 20 ml absolute ethanol) solution; guanidine nitrate
4, was added (5 g, 0.0409 mol) and the reaction mixture was
heated to reflux for 10–15 min, the clear solution immediately

S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
3195
turned turbid white. The solution was filtered and washed with
4.2.8. 5-((4-Methoxyphenylamino)methyl)pyrimidine-2,4-
diamine (5a)
10 ml absolute ethanol and the filtrate with combined washings
was used further. To the filtrate, ethoxymethylene malononitrile
(5 g, 0.049 mol) was added in portions for 15–20 min and the reac-
tion mixture was cooled to 10 °C and stirred for 8 h, after which it
was concentrated to dryness and the obtained residue was dis-
solved in glacial acetic acid and then given charcoal treatment.
The reaction mixture was filtered and the charcoal cake was
washed with 10 ml glacial acetic acid. The filtrate and washings
were collected together and cooled to 10 °C. On cooling yellow
crystals started precipitating out. To the cooled filtrate, 30% ammo-
nium hydroxide solution was added till it became neutral, which
led to further precipitation of solid product. The product was fil-
tered, washed with copious amount of cold water and then dried
to yield 3.5 g, (yield: 63%) of 2,4-diaminopyrimidine-5-carbonitrile
(5, Scheme 2) as yellow solid: mp = 318 °C, lit.14 mp = 317 °C; IR
(KBr) Vmax cmꢁ1 3430, 3336, 3104, 2202, 1652, 1586, 1548, 1475,
Raney nickel (1 g) and p-anisidine (0.2 g, 0.0017 mol) were
added to a solution of 5 (0.2 g, 0.00148) in glacial acetic acid
(20 ml) at room temperature. The resultant reaction mixture was
stirred for 2–6 h under H2 atmosphere.19 After completion of the
reaction (monitored by TLC) the reaction mixture was filtered,
washed with glacial acetic acid (10 ml). The filtrate and washings
were collected together and concentrated azeotropically using tol-
uene to dryness under reduced pressure at 50–55 °C. The obtained
slurry was treated with 1 N Na2CO3 to get free flowing solid. The
resultant mixture was cooled, filtered and washed with water till
the washings were neutral to pH paper. Solid was then washed
with 20% EtOAc/n-hexanes solution to remove any non-polar
impurity. The solid thus obtained was dried in oven to get 0.28 g,
71.79% brownish yellow solid. Mp = 215–217 °C; IR (KBr) Vmax
cmꢁ1 3420, 3212, 2952, 2835, 2929, 1683, 1602, 1527, 1501,
1460, 1286, 1246, 1206, 1178, 1108, 1030; 1H NMR (DMSO-d6,
300 MHz): d 8.76- 8.13 (m, 2H, Ar–H; 1H, –NH, D2O exchangeable),
7.15–6.63 (m, 4H, Ar–H; 2H, –NH, D2O exchangeable), 3.69 (s, 3H,
–OCH3, overlapped with H2O peak of DMSO-d6); 13C NMR (DMSO-
d6, 75 MHz): 162.6, 162.6, 161.8, 157.2, 144.1, 127.8, 121.8, 114.8,
114.3, 103.5, 55.1, 43.8; M/S 244.21 [Mꢁ1]+, 245.28.
1354, 1282 cmꢁ1 1H NMR (DMSO-d6, 500 MHz): d 8.18 (s, 1H,
;
–CH), 6.80–7.10 (br s, 4H, D2O exchangeable –NH2); M/S 136.04
[M+1]+.
4.2.5. 2,4-Diaminopyrimidine-5-carbaldehyde (6, Scheme 2)14–17
Raney nickel (2.5 g of 50% slurry with water) was added to a
solution of 5, (2.25 g, 0.0166 mol) in 98–100% formic acid
(20 ml) and the reaction mixture was refluxed for 3–4 h. After
which, the reaction mixture was filtered and washed with
10 ml formic acid. The filtrate and washings were collected to-
gether and concentrated under reduced pressure. The resulting
residue was stirred in 30% ammonia solution and cooled to get
free flowing solid product which was filtered and washed with
water followed by drying to get 1.8 g, (yield: 78.29%) of 2,4-dia-
minopyrimidine-5-carbaldehyde (6, Scheme 2) as off white solid.
Mp = chars at 270–272 °C; IR (KBr) Vmax cmꢁ1 3424, 3384, 3142,
1635, 1594, 1506, 1407, 1236; 1H NMR (DMSO-d6, 500 MHz): d
9.42 (s, 1H, –CHO), 8.25 (s, –1H, –Het Ar–H), 7.59–7.80 (d, 2H,
D2O exchangeable –NH2), 7.20 (s, 2H, D2O exchangeable –NH2);
M/S 139.03 [M+1]+.
4.2.9. 5-((Naphthalen-1-ylamino)methyl)pyrimidine-2,4-dia-
mine (5b)
Raney nickel (1 g), 1-napthyl amine (0.25 g, 0.00178 mol), 2,4-
diaminopyrimidine-5-carbonitrile (0.2 g, 0.00148 mol) were used
following the same procedure as described for 5a, to yield 0.28 g,
77.77% greenish yellow solid. Mp = 214 °C; IR (KBr) Vmax cmꢁ1
3365, 3310, 3133, 1659, 1616, 1565, 1529, 1500, 1451, 1394,
1245 cmꢁ1 1H NMR (DMSO-d6, 300 MHz): d 8.86 (br s, 1H, –NH
;
D2O exchangeable), 8.48 (s, 1H, Ar–H), 8.23–8.17 (m, 2H, Ar–H),
7.92–7.90 (s, 1H, –NH), 7.74–7.71 (d, 1H, Ar–H), 7.52–7.47 (m,
3H, Ar–H), 7.17–7.14 (d, 1H, Ar–H), 6.76 (s, 2H, –NH D2O exchange-
able) (Bridge –CH2 peak got overlapped with H2O peak of DMSO-
d6); 13C NMR (DMSO-d6, 75 MHz): d 163.6, 163.1, 162.1, 159.7,
148.5, 133.6, 128.1, 127.8, 126.4, 125.8, 124.7, 123.0, 113.2,
103.7, 40.3; M/S 264.04 [Mꢁ1]+.
4.2.6. (2,4-Diaminopyrimidin-5-yl)methanol (7, Scheme 2)14–17
NaBH4 (0.41 g, 0.0108 mol) was added to a methanolic solution
of 6, (1 g, 0.0072 mol) and reaction mixture was stirred at room
temperature for 4–5 h. After which the clear reaction mixture
was concentrated to dryness and obtained pale white pasty mass
was stirred in 3 ml of cold water and refrigerated for 12 h to obtain
free flowing solid product. The solid obtained was filtered and
washed with 2 ml cold water, dried to get 0.56 g, (yield: 55.4%)
of (2,4-diaminopyrimidin-5-yl)methanol (7, Scheme 2) as greenish
yellow solid. decomposes at 263–265 °C, lit.14 = decomposes at
265 °C; M/S 141.1 [M+1]+.
4.2.10. 3-(2,4-Diaminopyrimidin-5-yl)-1-(4-methoxyphenyl)-
prop-2-en-1-one (6a)
0.5 ml of 10% NaOH solution was added to a solution of p-
methoxyacetophenone (0.1 g, 0.00072 mol) in 5 ml DMSO and
the reaction mixture was stirred at 40–50 °C for 60 min followed
by addition of 6 (0.1 g, 0.00072 mol). The resultant reaction mix-
ture was stirred at 40–50 °C for 8 h. TLC monitoring showed prod-
uct spot along with the spot of unreacted p-methoxyacetophene.
The reaction mixture was poured to the cold water. The precipi-
tated solid was filtered and washed with water till the filtrate
was neutral to pH paper. The solid obtained was purified by col-
umn chromatography to yield 29 mg, (15% yield) of pure product
as bright yellow solid. Mp = chars at 198 °C; IR Vmax cmꢁ1 3115,
2956, 1668, 1595, 1541, 1472, 1322, 1249, 1173; 1H NMR
(DMSO-d6, 300 MHz): d 9.09 (s, 1H, CH@CH), 8.20–8.18 (d, 2H,
Ar–H, 1H, –NH D2O exchangeable), 7.79–7.77 (s, 1H, Ar–H), 7.79–
7.77 (d, 1H, –CH@CH), 7.08 (m, 3H, Ar–H; 1H, 1H, –CH@CH; 2H,
–NH D2O exchangeable), 3.81 (s, 3H, –OCH3); 13C NMR (DMSO-
d6, 75 MHz): d 163.6, 163.4, 162.4, 161.4, 138.0, 130.7, 129.3,
114.8, 114.4, 55.5; M/S 271.22 [M+1]+.
4.2.7. 5-(Bromomethyl)pyrimidine-2,4-diamine (8, Scheme 2)18
Compound 7 (0.2 g, 0.0014 mol) was dissolved in glacial acetic
acid (4 ml) by heating the reaction mixture at 40–50 °C. This lem-
on yellow solution was cooled gradually to room temperature (in
case solid started precipitating out, 1–2 ml more of glacial acetic
acid was added to get a clear solution). To the clear solution,
3 ml of 30–32% HBr in acetic acid was added drop wise over a
period of 20 min. Solid precipitation was observed within addition
of first few drops of HBr solution, which disappeared on addition
of more HBr solution. After completion of addition of HBr solu-
tion, reaction mixture was heated at 60 °C for 2–3 h till TLC
showed completion of reaction. The reaction mixture was then
concentrated to dryness by azeotropic distillation using toluene
under reduced pressure at 50–55 °C. The product obtained as
hydrobromide salt being very unstable was used as such without
further purification.
4.2.11. 3-(2,4-Diaminophenyl)-1-(3-(4-methylbenzyloxy)phe-
nyl)prop-2-en-1-one (6b)
1-(3-(4-Methylbenzyloxy)phenyl)ethanone (0.2 g, 0.0015 mol),
6 (0.2 g, 0.0015 mol) were used following the same procedure as
described for 6a, to yield 94 mg, (18% yield) of pure product as yel-

3196
S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
lowish-brown solid. Mp = chars at 218 °C; IR Vmax cmꢁ1 3413,
3320, 3123, 2916, 1654, 1606, 1545, 1466, 1193, 1018; 1H NMR
(DMSO-d6, 300 MHz): d 9.15 (s, 1H, CH@CH), 8.28–8.26 (d, 1H,
CH@CH); 7.86 (m, 3H, Ar–H), 7.43–7.35 (m, 3H, Ar–H; 1H, –NH
D2O exchangeable), 7.21–7.18 (m, 3H, Ar–H; 3H, 2H, –NH D2O
exchangeable), 5.15 (s, 2H, –CH2), 2.29 (s, 3H, –CH3); 13C NMR
(DMSO-d6, 75 MHz): d 163.5, 162.2, 159.7, 158.8, 158.8, 139.7,
138.0, 137.0, 133.9, 129.0, 127.8, 120.0, 117.1, 115.2, 113.4,
112.6, 69.2, 20.7; M/S 359.2 [M+1]+.
15% of N-methylnaphthalen-1-amine19 which was dissolved in
5 ml dry THF at room temperature under N2 atmosphere for
5 min. To this clear solution NaH (92 mg, 0.00210 mol) was added
and the resultant reaction mixture was stirred at room tempera-
ture under N2 atmosphere for 2 h. To this reaction mixture freshly
prepared hydrobromide salt of 8, dissolved in dry N,N-dimethyl-
acetamide (15 ml) was added and the resultant reaction mixture
was stirred at room temperature for three days. TLC monitoring
showed formation of new spot along with unreacted starting mate-
rial. Column purification gave 110 mg, 19%, off white solid.
Mp = 221 °C; IR (KBr) Vmax cmꢁ1 3363, 2936, 1638, 1498, 1399,
1223, 761; 1H NMR (DMSO-d6, 300 MHz): d 8.36–8.29 (m, 2H,
Ar–H), 8.08 (s, 1H, Ar–H), 7.83–7.08 (m, 4H, Ar–H; 4H, –NH D2O
exchangeable), 6.86–6.77 (d, 1H, Ar–H), 3.12 (s, 3H, –CH3): 13C
NMR (DMSO-d6, 75 MHz): 164.2, 154.5, 139.7, 138.9, 131.4,
128.4, 128.3, 127.9, 126.8, 126.4, 124.8, 124.2, 119.7, 107.9, 48.7;
M/S: 280.10 [M+1]+.
4.2.12. (2,4-Diaminopyrimidin-5-yl)methyl naphthalene-2-
sulfonate (7a)
Naphthalene-2-sulfonyl chloride (0.32 g, 0.00143 mol) was
added to a solution of 7 (0.2 g, 0.00143 mol) in 1 ml DMSO and tri-
ethylamine (0.144 g, 0.00143 mol) and the reaction mixture was
stirred at room temperature for 8 h. After completion of reaction
(monitored by TLC), reaction mixture was added to water and ex-
tracted with 80% chloroform/methanol mixture. Column purifica-
tion gave 70 mg (yield 15%) of product as white solid. Mp 195–
196 °C; IR (KBr) Vmax cmꢁ1 3342, 1656, 1544, 1483, 1270, 1174,
1122, 1070; 1H NMR (DMSO-d6, 300 MHz): d 8.74 (s, 1H, Ar–H),
8.21–8.14 (m, 2H, Ar–H), 7.99–7.72 (m, 3H, Ar–H; 2H, –NH D2O
exchangeable), 7.69–7.52 (m, 2H, Ar–H; 2H, –NH D2O exchange-
able), 4.17 (s, 2H, –CH2 overlapped with H2O peak of DMSO-d6);
13C NMR (DMSO-d6, 75 MHz): d 162.5, 162.4, 152.0, 140.1, 133.9,
132.7, 131.6, 129.2, 127.4, 126.4, 124.0, 109.8, 56.4; M/S 331.43
[M+1]+.
4.2.16. 5-((Benzo[d]oxazol-2-ylthio)methyl)pyrimidine-2,4-
diamine (8d)
NaH (62 mg, 0.00141mol) was added to a solutionof 2-mercapto-
benzoxazole (0.21 g, 0.00139 mol) in 5 ml dry THF under the N2
atmosphere. Freshly prepared hydrobromide salt of 8, dissolved in
dry N,N-dimethylacetamide (15 ml) was added and the resultant
reaction mixture was stirred at room temperature for three days.
TLC monitoring showed formation of new spots along with unre-
acted 2-mercaptobenzoxazole, at this stage reaction mixture was
purified using column chromatography, to yield 110 mg, 29%, off
white solid. Mp = 124–126 °C; IR (KBr) Vmax cmꢁ1 3346, 3187,
1683, 1532, 1505, 1456, 1256, 1236, 1212, 1133; 1H NMR (DMSO-
d6, 300 MHz): d 7.83 (s, 1H, Ar–H), 7.63 (2H, Ar–H;), 7.31 (s, 2H, –
Ar–H), 6.58 (s, 2H, –NH D2O exchangeable), 5.98 (s, 2H,
–NH D2O exchangeable), 4.34 (s, 2H, –CH, overlapped with H2O peak
of DMSO-d6); 13C NMR (75 MHz): 172.1, 162.8, 162.0, 157.0, 151.1,
141.2, 124.6, 124.2, 118.2, 110.1, 100.6. 30.8; M/S: 274.33 [M+1]+.
4.2.13. 5-((5H-Dibenzo[b,f]azepin-5-yl)methyl)pyrimidine-2,4-
diamine (8a)
NaH (50 mg, 0.011 mol) was added to a solution of iminostil-
bene (0.18 g, 0.00093 mol) in 5 ml dry THF and the reaction mix-
ture was stirred at room temperature under N2 atmosphere for
2 h. To this reaction mixture freshly prepared hydrobromide salt
of 5-(bromomethyl) pyrimidine-2,4-diamine 8, dissolved in dry
N,N-dimethylacetamide (15 ml) was added and the resultant reac-
tion mixture was stirred at room temperature for three days.3,5 TLC
monitoring showed formation of new spots along with unreacted
iminostilbene. Column purification of the crude product yielded
60 mg, 20.5%, greenish yellow solid. Mp = 116–117 °C; IR (KBr)
Vmax cmꢁ1 3341, 3185, 3019, 3061, 2926, 2853, 1659, 1607,
1487, 1456, 1425, 1290, 761; 1H NMR (DMSO-d6, 300 MHz): d
7.75 (s, 1H, Ar–H), 7.30–6.97 (m, 8H, Ar–H), 6.82 (s, 2H, –CH@CH),
6.20 (s, 2H, –NH D2O exchangeable), 5.73 (s, 2H, –NH D2O
exchangeable), 4.61 (s, 2H, –CH2); 13C NMR (DMSO-d6, 75 MHz):
162.9, 162.3, 156.7, 149.3, 133.1, 131.8, 128.8, 123.6, 120.7,
101.5, 48.9; M/S: 316.04 [M+1]+.
4.2.17. 6-(3,4,5-Trimethoxybenzyloxy)pyrimidine-2,4-diamine
(11a)
NaH (0.20 g, 0.0086 mol) was added to a solution of 9 (1 g,
0.0069 mol) and 10a (1.37 g, 0.0069 mol), in dry DMSO (5 ml)
and reaction mixture was heated at 100 °C for 5–6 h. TLC monitor-
ing showed formation of new spot along with unreacted starting
material. Column purification yielded 1.22 g, (yield: 57%), white
solid of 11a (Scheme 3). Mp = 166–168 °C; 1H NMR (DMSO-d6,
300 MHz): d 6.71 (s, 2H, Ar–H), 6.04–5.93 (d, 4H, D2O exchangeable
–NH2), 5.09 (s, 3H, Ar–H; –CH2), 3.75 (s, 6H, two –OCH3), 3.63 (s,
3H, –OCH3), 13C NMR (DMSO-d6,75 MHz): d169.7, 166.0, 162.8,
152.7, 136.9, 133.0, 105.4, 76.2, 66.2, 59.9, 55.8; M/S 307.04
[M+1]+.
4.2.14. 5-((10,11-dihydro-5H-dibenzo[b,f]azepin-5-yl)methyl)-
pyrimidine-2,4-diamine (8b)
NaH (55 mg, 0.00125mol), dihydroiminostilbene (0.25g,
0.00128 mol) were used following the same procedure as de-
scribed for 8a, to yield 26 mg, 6.4%, greenish yellow solid.
Mp = 105 °C; IR (KBr) Vmax cmꢁ1 3320, 3185, 2926, 1648, 1601,
1524, 1492, 1342, 756; 1H NMR (DMSO-d6, 300 MHz): d 8.25 (s,
2H, –NH D2O exchangeable), 7.69–6.63 (8H, Ar–H; 4H, –NH D2O
exchangeable), 3.52 (s, 2H, –CH, overlapped with DMSO peak),
2.92 (s, 4H, –CH); M/S: 318.50 [M+1]+.
4.2.18. 6-(2-(5H-Dibenzo[b,f]azepin-5-yl)ethoxy)pyrimidine-
2,4-diamine (11b)
NaH (0.20 g, 0.0086 mol) was added to a solution of 9 (1 g,
0.0069 mol) and 10b (1.64 g, 0.0069 mol), in dry DMSO (5 ml) wer-
sxe used following the same procedure as described for 10a, to
yield 0.9 g, 37.8%, white solid 11b (Scheme 3). Mp = 122–124 °C;
1H NMR (DMSO-d6, 300 MHz): d 7.3–7.0 (m, 8H, Ar–H), 6.7 (s,
2H, Ar–H), 5.97–5.82 (d, 4H, D2O exchangeable –NH2), 4.97 (s,
1H, Ar–H), 4.09–4.00 (d, 4H, –CH@CH–), 13C NMR (DMSO-
4.2.15. 5-((Methyl(naphthalen-1-yl)amino)methyl)pyrimidine-
2,4-diamine (8c)
d6,75 MHz): d 165.9, 165.8, 162.8, 150.2, 133.3,131.9, 129.1,
129.0, 123.4, 120.3, 76.1, 62.1, 49.0 M/S 346.07 [M+1]+.
The solution of 1-napthylamine (2 g, 0.0139 mol), methyl iodide
(2.18 g, 0.0139 mol) and K2CO3 (2.31 g, 0.0167 mol) in acetonitrile
(30 ml) was stirred at 40–50 °C in dark for three days. TLC showed
two spots along with unreacted 1-napthylamine spot. Reaction
mixture was purified by column chromatography to yield 0.33 g,
4.3. Biological assay
The synthesized compounds were evaluated for their ability to
inhibit DHFR from pc, tg, ma and rl using a continuous spectropho-

S. Bag et al. / Bioorg. Med. Chem. 18 (2010) 3187–3197
3197
the ligand and apoenzyme. Corrections for entropic changes were
not applied.
tometric assay measuring oxidation at 340 nM of NADPH at 37 °C
under conditions of saturating substrate and cofactor as previously
described.20,21
4.4.3. MESP calculations
MESP computation was carried out using Jagur,28 version 7.5,
Schrödinger, LLC, New York, NY, 2008 installed on Intel Pentium
4 workstation using single point energy calculation at hybrid den-
sity functional theory with Becke’s three-parameter exchange po-
tential and the Lee-Yang-Parr correlation functional (B3LYP),
using basis set 3-21G* level on the docked pose of 7a.
4.4. Computational studies
4.4.1. Molecular docking
Docking studies were carried out using standard Glide molecu-
lar docking protocol implemented within Maestro molecular mod-
eling suit by Schrödinger, LLC, New York, NY, 2008 installed on
Intel Pentium 4 computer.22
The protein structure PDB code: 1KLK4 which is the ternary
Acknowledgments
complex of enzyme, NADPH and I ( Fig. 1) with resolution of
0
Seema Bag is thankful to University Grand Commission (UGC),
India and Nilesh R. Tawari is thankful to Department of Biotechnol-
ogy (DBT), India for financial support.
2.3 ÅA was retrieved from Protein Data Bank (www.rcsb.com)
and used for docking simulations. All water molecules were de-
leted, bond orders and charges of cofactor NADPH and 6-((5H-
dibenzo[b,f]azepin-5-yl)methyl)pteridine-2,4-diamine, I ( Fig. 1),
were assigned properly in the protein preparation step, nitrogen
N1 of inhibitor was protonated and given a +ve charge. A grid
which is the representation of shape and properties of receptor
using several different sets of fields was generated. The structure
of I (Fig. 1) was randomized using Ligprep23 and used for the cog-
nate docking studies. The docking study was carried out using ex-
tra precision (XP) mode with ‘must match’ of at least one
hydrogen bond criteria; with Ile123, Ile10 and Glu32 as these
are the conserved residues in most of the DHFRs and H-bonding
with these residues is one of the essential criteria for antifolate
activity.
Supplementary data
Supplementary data (1H NMR and 13C NMR spectra of all target
molecules) associated with this article can be found, in the online
version, at doi:10.1016/j.bmc.2010.03.031.
References and notes
1. Gangjee, A.; Yang, J.; Queener, S. F. Bioorg. Med. Chem. 2006, 14, 8341.
2. Klepser, M. E.; Klepser, T. B. Drugs 1997, 53, 40.
3. Rosowsky, A.; Cody, V.; Galitsky, N.; Fu, H.; Papoulis, A. T.; Queener, S. F. J. Med.
Chem. 1999, 42, 4853.
4. Cody, V.; Galitsky, N.; Luft, J. R.; Pangborn, W.; Rosowsky, A.; Queener, S. F. Acta
Crystallogr., Sect. D 2002, 58, 946.
5. Rosowsky, A.; Fu, H.; Queener, S. F. J. Heterocycl. Chem. 2000, 37, 921.
6. Rosowsky, A.; Fu, H.; Chan, D. C. M.; Queener, S. F. J. Med. Chem. 2004, 47, 2475.
7. Graffner-Nordberg, M.; Kolmodin, K.; Åqvist, J.; Queener, S. F.; Hallberg, A. Eur.
J. Pharm. Sci. 2004, 22, 43.
8. Bag, S.; Tawari, N. R.; Degani, M. S. QSAR Comb. Sci. 2009, 28, 296.
9. Bag, S.; Tawari, N. R.; Queener, S. F.;. Degani, M. S., J. Enzyme Inhib. Med. Chem.,
in press. doi: 10.3109/14756360903179443.
10. Degani, M. S. Bag, S.; Bairwa, R.; Tawari, N. R.; Queener, S. F., J. Heterocycl.
Chem., in press. doi: 10.1002/jhet.216.
The structures of all synthesized compounds were constructed
using MarvinSketch version 5.2 from ChemAxon Ltd24 nitrogen
N1 of all compounds was protonated and given a +ve charge. A un-
ique low-energy 3D, protonated structure was generated for each
ligand with the help of Ligprep,23 where appropriate hydrogens
were added to all structures and were subsequently subjected to
energy minimization using the OPLS-2005 force field with a con-
stant dielectric of 1.0. All ligands were docked using same protocol
as described for compound I.
11. Gabriele, B. A. Process for Preparation of Opiramol. EP1 468 2007, 992, B1.
12. Reddy, G. O.; Sarma, M. R.; Chandrasekhar, B.; Charan, C. J. A Process for the
Preparation of 2-[Phenothiazin-10-yl] Ethyl methane sulfonate, WO 02/40460,
2002.
4.4.2. Binding affinity prediction using MM-GBSA
The assessment of molecular docking was carried out using a
post scoring approach, MM-GBSA.25 Maximum of top five docked
poses returned using Glide-XP score were minimized using the lo-
cal optimization feature26 in Prime, and the energies were calcu-
lated using the OPLS-AA force field27 and the GBSA continuum
model.27 For each molecule the best scoring pose predicted by
MM-GBSA was selected for comparison with the experimental
13. Simons, J. K.; Saxton, M. R. Org. Synth. 1963, Collective Vol. 4, 78.
14. Huber, W. J. Am. Chem. Soc. 1943, 65, 2222.
15. Gangjee, A.; Adair, O. O.; Queener, S. F. J. Med. Chem. 2003, 46, 5074.
16. Wyss, P. C.; Gerber, P.; Hartman, P. G.; Hubschwerlen, C.; Locher, H.; Marty, H.
P.; Stahl, M. J. Med. Chem. 2003, 46, 2304.
17. Otzen, T.; Wempe, E. G.; Kunz, B.; Bartels, R.; Yvetot, G. L.; Hänsel, W.; Schaper,
K. J.; Seydel, J. K. J. Med. Chem. 2004, 47, 240.
18. Piper, J. R.; Johnson, C. A.; Maddry, J. A.; Malik, N. D.; McGuire, J. J.; Otter, G. M.;
Sirotnak, F. M. J. Med. Chem. 1993, 36, 4161.
19. Gangjee, A.; Guo, X.; Queener, S. F.; Cody, V.; Galitsky, N.; Luft, J. R.; Pangborn,
W. J. Med. Chem. 1998, 41, 1263.
IC50 values. The binding free energy
DGbind was estimated as
20. Broughton, M. C.; Queener, S. F. Antimicrob Agents Chemother. 1991, 35, 1348.
21. Chio, L. C.; Queener, S. F. Antimicrob Agents Chemother. 1993, 37, 1914.
22. Glide, version 5.0, Schrödinger, LLC, New York, NY, 2008.
23. LigPrep, version 3.0, Schrödinger, LLC, New York, NY, 2008.
24. MarvinSketch version 5.2 ChemAxon Ltd, 2008.
25. Nu, H.; Kalyanaraman, C.; Irwin, J. J.; Jacobson, M. P. J. Chem. Inf. Model. 2006,
46, 243.
26. Jacobson, M. P.; Pincus, D. L.; Rapp, C. S.; Day, T. J.; Honig, B.; Shaw, D. E.;
Friesner, R. A. Proteins 2004, 55, 351.
D
Gbind
¼
D
EMM
þ
D
Gsolv
þ
DGSA
where
D
EMM is the difference in energy calculated using OPLS force
field between the complexed holoenzyme structure and the sum of
the energies of the ligand and apoenzyme, Gsolv is the difference in
the GBSA solvation energy of the complexed holoenzyme and the
sum of the solvation energies for the ligand and apoenzyme protein,
D
27. Yu, Z.; Jacobson, M. P.; Friesner, R. A. J. Comput. Chem. 2006, 27, 72.
28. Jaguar, version 8.0 Schrödinger, LLC, New York, NY, 2008.
and
DGSA is the difference in the surface area energy for the com-
plexed holoenzyme and the sum of the surface area energies for