Triazine-pyrimidine based molecular hybrids: Synthesis, docking studies and evaluation of antimalarial activity

Article abstract of DOI:10.1039/c4nj00978a

A series of novel triazine-pyrimidine hybrids have been synthesized and evaluated for their in vitro antimalarial activity. Some of the compounds showed promising antimalarial activity against both CQ-sensitive and CQ-resistant strains at micromolar level with a high selectivity index. All the compounds displayed better activity (IC₅₀ = 1.32-10.70 μM) than the standard drug pyrimethamine (>19 μM) against the chloroquine-resistant strain W2. All the tested compounds were nontoxic against mammalian cell lines. Further, docking studies of the most active compounds were performed on both wild type and quadruple mutant (N51I, C59R, S108N, I164L) PfDHFR-TS using Glide to analyse the interaction of the compounds with the binding site of the protein. The binding poses of compounds 14 and 19, having a high Glide XP score and the lowest Glide energies, show an efficient binding pattern similar to that of the DHFR substrate (dihydrofolate) in the wild type and mutant DHFR active site. The analysis of the pharmacokinetic properties of the most active compounds using ADMET prediction attests to the possibility of developing compound 14 as a potent antimalarial lead. This journal is

Full text of DOI:10.1039/c4nj00978a

NJC
View Article Online
View Journal | View Issue
PAPER
Triazine–pyrimidine based molecular hybrids:
synthesis, docking studies and evaluation of
antimalarial activity†
Cite this: New J. Chem., 2014,
38, 5087
Deepak Kumar,a Shabana I. Khan,b Prija Ponnana and Diwan S. Rawat*a
A series of novel triazine–pyrimidine hybrids have been synthesized and evaluated for their in vitro antimalarial
activity. Some of the compounds showed promising antimalarial activity against both CQ-sensitive and
CQ-resistant strains at micromolar level with a high selectivity index. All the compounds displayed better
activity (IC50
= 1.32–10.70 mM) than the standard drug pyrimethamine (419 mM) against the
chloroquine-resistant strain W2. All the tested compounds were nontoxic against mammalian cell lines.
Further, docking studies of the most active compounds were performed on both wild type and
quadruple mutant (N51I, C59R, S108N, I164L) PfDHFR-TS using Glide to analyse the interaction of the
compounds with the binding site of the protein. The binding poses of compounds 14 and 19, having a
high Glide XP score and the lowest Glide energies, show an efficient binding pattern similar to that
of the DHFR substrate (dihydrofolate) in the wild type and mutant DHFR active site. The analysis of
the pharmacokinetic properties of the most active compounds using ADMET prediction attests to the
possibility of developing compound 14 as a potent antimalarial lead.
Received (in Montpellier, France)
13th June 2014,
Accepted 11th August 2014
DOI: 10.1039/c4nj00978a
www.rsc.org/njc
mefloquine and related compounds were developed. Chloroquine
(CQ) has been the mainstay of malaria therapy for decades because
Introduction
Malaria is the third most infectious disease after tuberculosis of its efficacy, safety and low cost, until the emergence and spread
and HIV/AIDS, and affects over 100 countries in Africa, Asia and of CQ-resistance. Pyrimethamine-sulfadoxine (Fansidar) was
South America.1 Despite intensive efforts towards its eradica- one of the best therapeutic options after CQ, but was rendered
tion in the early 1960s, malaria remains a major public health ineffective in most malaria-endemic regions due to the spread of
problem to date. According to the World Health Organization resistance. Currently, natural endoperoxide artemisinin and its
(WHO) nearly 300–500 million people throughout the world semi-synthetic derivatives (artemether, arteether and artesunate)
become infected with malaria every year. The mortality rate is are the most potent and fast-acting antimalarials effective against
estimated to be around 1.1 million deaths per year, mostly resistant strains of P. falciparum. In order to combat the resistance
children under the age of five. Eighty percent of malaria cases problem, combination therapy has been introduced by the WHO,
worldwide occur in Africa; two thirds of the remaining cases are in which artemisinin and its analogue in combination with
found in six countries, of which India is one. There are four 4-aminoquinoline antimalarials are used to treat malaria. Although
major species of the malaria parasite, of which Plasmodium artemisinin combination therapy (ACT) is well-tolerated and is
falciparum causes the most virulent forms of malaria and is nearly 95% effective in treating malaria, its use is limited in some
responsible for more than 95% of malaria-related deaths. Due to the regions due to some serious issues such as the higher cost of
unavailability of effective vaccines, chemotherapy remains the only treatment and safety during pregnancy.2–7 In addition, resistance to
option for the treatment of malaria. After the discovery of quinine in artemisinin derivatives has also been reported in Southeast Asian
the late 1600s, a huge number of potent antimalarial agents countries and may continue to increase, subsequently making
such as chloroquine, amodiaquine, primaquine, pamaquine, malaria chemotherapy more complicated.5–8
Dihydrofolate reductase (DHFR) is one of the more well-defined
a Department of Chemistry, University of Delhi, Delhi-110007, India.
E-mail: dsrawat@chemistry.du.ac.in; Fax: +91-11-27667501;
Tel: +91-11-27667465
and explored targets in malarial chemotherapy. Pyrimethamine
and cycloguanil (Fig. 1) are potent DHFR inhibitors and are
clinically used for the treatment of P. falciparum malaria.9,10
Unfortunately, point mutations at certain amino acid residues in
the surroundings of the active site of P. falciparum DHFR have
resulted in resistance, compromising the clinical effectiveness of
b National Centre for Natural Products Research, University of Mississippi,
MS-38677, USA
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4nj00978a
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 5087--5095 | 5087

View Article Online
Paper
NJC
Fig. 1 DHFR inhibitor-based antimalarial drugs.
these drugs.11–13 Despite this, the folate pathway remains a
good target for malarial chemotherapy because the enzyme is
limited in its mutational capability, owing to the loss in enzyme
function. WR99210 (Fig. 1), having a flexible linker, is found to be
effective at nanomolar concentration even against the strains that
are highly resistant to other DHFR inhibitors. It is believed that
the exceptionally high activity of WR99210 is due to its highly
flexible nature, which helps to bind it into the active site of the
target.14 Unfortunately, WR99210 exhibits unacceptable gastro-
intestinal (GI) intolerance. Recently, a new DHFR inhibitor, P218
(Fig. 1), has been developed by BIOTEC pharmaceuticals.15,16
It inhibits blood stage growth of drug-resistant malarial parasites
with an IC50 value of 6 nM. It is also the most active antifolate
agent against the liver stage of P. yoelii (IC50 o 10 nM).17
Apart from this, a large number of structurally similar com-
pounds such as triazine18–20 and pyrimidine derivatives21–23 have
been synthesized, and some of these compounds have shown very
potent antimalarial activity against both CQ-sensitive and CQ-
resistant strains. Cycloguanil and pyrimethamine represent the
triazine and pyrimidine classes of compounds, respectively. To
combat the increasing resistance problem, there is an urgent need
to develop a potent, safe and cost-effective antimalarial agent. As
part of our ongoing malaria research programme,24–30 we became
interested in joining triazine and pyrimidine moieties together in a
single molecule, using a flexible linker to provide the molecule
with enough flexibility so that, like WR99210, it can easily fit into
the binding pocket of the target, and so may have a better anti-
malarial activity profile as a result. To the best of our knowledge,
this is the first report of covalent hybrids having both triazine and
pyrimidine pharmacophores. All the synthesized compounds were
characterized by various spectroscopic techniques.
Fig. 2 Design strategy for the synthesis of novel triazine–pyrimidine hybrids.
The disubstituted products 6 and 7 were obtained by the reaction
of triazine (5) with two equivalents of either morpholine or
diethylamine, respectively, at 0 1C to RT (o30 1C) in the presence
of K2CO3 using THF as the solvent (Scheme 2).32,33 Thereafter, the
resulting disubstituted triazines were reacted, by substitution of
the third chlorine, with various alkyl diamines to give trisubstituted
triazines (8–13) with a free terminal NH2 group (Scheme 2).34 This
reaction was carried out under reflux conditions in THF using
K2CO3 as a base. Finally, compounds 2–4 and 8–13, synthesized
as shown in Schemes 1 and 2, were coupled together in the
presence of K2CO3, using NMP as a solvent and under reflux
conditions, to give the desired triazine–pyrimidine hybrid molecules
14–31 (Scheme 3). All the compounds were purified by column
chromatography using MeOH–CHCl3 as eluent and characterized by
various spectroscopic techniques.
Biological activity
In vitro antimalarial activity
The antimalarial activity was determined by measuring plasmo-
dial LDH activity as described in the literature.35 A suspension
of red blood cells infected with the D6 or W2 strain of
P. falciparum (200 mL, with 2% parasitemia and 2% hematocrit
in RPMI 1640 medium supplemented with 10% human serum
and 60 mg mLÀ1 amikacin) was added to the wells of a 96-well
plate containing 10 mL of serially diluted test samples. The plate
Chemistry
The triazine–pyrimidine hybrids 14–31 (Fig. 2) were synthesized was flushed with a gas mixture of 90% N2, 5% O2, and 5% CO2
as shown in three different Schemes (1–3). Firstly, 2,4-dichloro- and incubated at 37 1C for 72 h in a modular incubation
pyrimidine (1) was treated with different secondary amines in chamber (Billups-Rothenberg, CA). Parasitic LDH activity
the presence of triethylamine in THF at 0 1C to RT to give was determined according to the procedure of Makler and
compounds 2–4 in high yield, with small amounts of the Hinrichs.36 20 mL of the incubation mixture was mixed with
regioisomer (Scheme 1).31 2,4,6-Trichloro-1,3,5-triazine shows 100 mL of the Malstatt reagent (Flow Inc., Portland, OR) and
temperature-dependent nucleophilic substitution reactions, with incubated at room temperature for 30 min. Twenty microliters
the first chlorine being replaceable at 0 1C, the second chlorine of a 1 : 1 mixture of NBT/PES (Sigma, St. Louis, MO) was then
at room temperature and the third at higher temperature. added and the plate was further incubated in the dark for 1 h.
5088 | New J. Chem., 2014, 38, 5087--5095
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Scheme 1
Scheme 2
Scheme 3
The reaction was then stopped by the addition of 100 mL of a 5% seeded into the wells of 96-well plates at a density of 25000 cells
acetic acid solution. The plate was read at 650 nm. Chloroquine per well and incubated for 24 h. Samples at different concentra-
and pyrimethamine were included in each assay as antimalarial tions were added and the plates were again incubated for 48 h.
drug controls. IC50 values were computed from the dose– The number of viable cells was determined by Neutral Red assay.
response curves. To determine the selectivity index of the anti- The IC50 values were obtained from dose–response curves.
malarial activity of the compounds, in vitro cytotoxicity of these
Docking studies
compounds against mammalian cells was also determined. The
assay was performed in 96-well tissue culture-treated plates as Antifolates act by inhibiting the dihydrofolate reductase activity
described earlier.37 VERO cells (monkey kidney fibroblasts) were of Plasmodium falciparum bifunctional enzyme dihydrofolate
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 5087--5095 | 5089

View Article Online
Paper
NJC
Table 1 In vitro antimalarial activity and cytotoxicity of triazine–pyrimidine
reductase-thymidylate synthase (PfDHFR-TS). The four point
mutations in codons 51, 59, 108, and 164 (N51I, C59R, S108N,
and I164L) have been found in the DHFR domain of the PfDHFR-
TS gene from the clinical isolates of dihydrofolate-resistant
parasite.38 In the present work we have studied the binding
pattern and ADMET properties of novel triazine–pyrimidine
hybrids with PfDHFR-TS. The 2D structures of all the compounds
hybridsa
P. falciparum
(D6 clone)
P. falciparum
(W2 clone)
Cytotoxicity
(VERO cells)
IC50 (mM)
IC50 (mM) SI
IC50 (mM) SI
Compound
14
15
1.18
4.41
6.21
49.80
3.88
2.15
4.54
3.67
4.32
10.29
3.05
2.22
3.46
1.28
1.54
4.85
1.57
2.50
0.04
0.01
48.53
1.32
7.48
8.83
49.80
4.60
2.62
5.52
5.86
4.60
47.62 410.07
42.21
41.52
1.00
41.30
41.07
1.00
42.06
43.53
41.72
41.58
41.96
49.78
49.50
49.80
49.52
49.26
49.52
49.26
49.02
were generated using ChemBioDraw Ultra 12.0 (www.cambridge 16
17
¨
soft.com) and the Ligprep module in Schrodinger was used to
18
19
42.45
44.30
42.09
42.52
42.08
41.03
43.40
44.53
43.00
47.88
46.36
42.08
46.24
43.80
41500
1820
generate energy-minimized 3D structures. Partial atomic charges
were computed using the OPLS_2005 force field. The correct 20
21
Lewis structure, tautomers and ionization states (pH 7.0 Æ 2.0)
22
for each of these ligands were generated and optimized with
23
410.70
7.39
1.00 410.70
¨
default settings (Ligprep 2.5, Schrodinger, LLC, New York, NY, 24
41.40 410.37
41.00 410.07
42.13 410.40
43.28 410.09
25
9.98
4.86
3.07
4.04
7.96
4.79
4.70
0.39
2012). The 3D crystal structures of wild type PfDHFR-TS (PDB ID:
3QGT; resolution 2.30 Å) and quadruple mutant (N51I + C59R +
26
27
S108N + I164L) PfDHFR-TS (PDB ID: 3QG2; resolution: 2.30 Å) 28
42.42
49.80
29
41.26 410.09
were retrieved from the protein data bank (www.rcsb.org). The
proteins were prepared for docking using the Protein Preparation
30
31
42.04
42.02
4150
—
49.80
49.52
460
18.2
¨
Wizard (Maestro 10.0 Schrodinger, LLC, New York, NY, 2012). CQ
b
Pyr
Water molecules within 5 Å of the protein structures were
considered. Bond order and formal charges were assigned and
hydrogen atoms were added to the crystal structure. To further
refine the structure, an OPLS-2005 force field parameter was used
to alleviate steric clashes and the minimization was terminated
when RMSD reached a maximum cutoff value of 0.30 Å.
a
IC50: the concentration that causes 50% growth inhibition; SI: selectivity
index (IC50 for cytotoxicity to VERO cells/IC50 for antimalarial activity).
Not active up to 19 mM.
b
and pyrimethamine as standard drugs. Cytotoxicity was deter-
The location of co-crystallized ligand pyrimethamine in both mined against VERO cells (Table 1). All compounds except for 16,
wild and mutant protein structures was used to choose the center 17 and 23 exhibited promising antimalarial activity with IC50
and size of the receptor grid, which was generated using Glide 5.8 values of o5 mM against the chloroquine-sensitive strain (D6).
¨
(Schrodinger, LLC, New York, NY, 2012) with default settings for all Four compounds (14, 27, 28 and 30) displayed potent antimalar-
parameters. The grid size was chosen to be sufficiently large so as ial activity with IC50 values ranging from 1.18 mM to 1.57 mM
to include all active site residues involved in substrate binding. against the CQ-sensitive strain and a high selectivity index, while
The cofactor NADH in the PfDHFR-TS wild and mutant structures other compounds showed moderate to good antimalarial activity.
was also considered as part of the receptor protein. All ligand con- None of these compounds showed any cytotoxicity towards
formers were docked to each of the receptor grid files (PfDHFR-TS mammalian kidney fibroblast (VERO cells). In the case of the
wild and mutant structures) using Glide extra precision (XP) mode. CQ-resistant strain, the compounds also showed significant
Default settings were used for the refinement and scoring.
activity (IC50 o10 mM). It is interesting to note that all the
compounds were found to be more active than the standard drug
pyrimethamine against the chloroquine-resistant strain (W2).
Compound 14 showed potent activity against both the strains.
The activity profile of these compounds against the
CQ-sensitive strain of P. falciparum clearly indicates that the com-
pounds having N-methyl or N-ethyl groups at the pyrimidine
nucleus were found to be more active than the compounds
having a morpholine ring (15 vs. 18 and 21, 16 vs. 19, 22).
Similarly, compounds having a diethyl amino group at the
triazine nucleus (24–31) were found to be more active than
the respective compounds having a morpholine ring at the
triazine nucleus (15–22), with the exception of compounds 14
and 23 where compound 14 was more active than compound 23.
Compounds having a diethyl group at the triazine nucleus
and N-methyl or N-ethyl groups at the pyrimidine nucleus
(27, 28, 30 and 31) were more potent than other compounds.
In silico ADMET prediction
The pharmacokinetic profile of compounds showing good anti-
malarial activity was predicted using the QikProp v3.5 program
¨
(Schrodinger, Inc., New York, NY, 2012). All the compounds were
prepared in neutralized form for the calculation of pharmaco-
¨
kinetic properties by QikProp using Schrodinger’s Maestro Build
module and LigPrep, saved in SD format. The QikProp program
utilizes the method of Jorgensen39 to compute pharmacokinetic
properties and descriptors such as the octanol/water partitioning
coefficient, aqueous solubility, brain/blood partition coefficient,
intestinal wall permeability and plasma protein binding.
Results and discussion
The triazine–pyrimidine hybrids were evaluated for their in vitro Although in the case of the CQ-resistant strain no uniform
antimalarial activity against both CQ-sensitive (D6 clone) and CQ- pattern was observed, compounds with a long carbon chain
resistant (W2 clone) strains of P. falciparum using chloroquine between the triazine and pyrimidine ring showed better activity
5090 | New J. Chem., 2014, 38, 5087--5095
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Table 2 Glide docking scores (kcal molÀ1) and docking energies of the most active molecules along with the reference compounds (pyrimethamine,
cycloguanil and WR99210) and dihydrofolate in the binding site of wild and mutant PfDHFR-TS
Docking results with wild PfDHFR
Docking results with mutant PfDHFR
XP G
score
van der Waals
energy
Coulomb
energy
Glide
energy
XP
H-bond
XP G
score
van der Waals
energy
Coulomb
energy
Glide
energy
Compounds
14
19
25
27
28
30
31
À6.25
À4.71
À2.13
À3.04
À3.22
À2.75
À1.98
À9.33
À9.04
À8.94
À4.84
À51.18
À46.48
À21.04
À28.90
À23.76
À32.65
À22.45
À52.14
À31.70
À30.12
À51.18
À6.65
À6.08
À1.98
À3.26
À2.32
À2.09
À0.97
À14.19
À15.51
À10.74
À6.91
À57.84
À46.31
À14.76
À28.78
À21.43
À26.54
À14.52
À64.84
À44.91
À38.55
À37.03
À1.99
À1.04
À0.12
À0.23
À0.26
À0.17
À0.18
À3.10
À2.64
À2.86
À2.09
À6.50
À6.1
À41.17
À29.81
À23.51
À28.08
À23.74
À29.54
À21.62
À43.68
À33.65
À34.30
À27.37
À9.5
À4.98
À3.42
À3.14
À3.26
À3.65
À3.05
À17.61
À12.06
À8.60
À8.07
À57.11
À39.09
À28.03
À37.65
À29.73
À32.54
À23.65
À61.30
À43.55
À46.6
À2.98
À3.92
À3.17
À3.36
À2.05
À11.00
À9.39
À8.95
À5.48
Dihydrofolate
Pyrimethamine
Cycloguanil
WR99210
À34.30
than compounds having a short carbon chain. Amongst all the active site of wild and mutant PfDHFR-TS are summarized
compounds, 14 was found to be the most potent compound with in Table 2 and Fig. 3 and 4. These docking results clearly
IC50 values of 1.18 mM against the chloroquine-sensitive strain indicate that the most active compounds in the study exhibited
(D6) and 1.32 mM against the chloroquine-resistant strain (W2). significant binding affinities towards the wild (Glide energy range
Molecular docking studies of the most active compounds À57.84 kcal molÀ1 to À14.52 kcal molÀ1) and quadruple mutant
(14, 19, 25, 27, 28, 30 and 31) were performed in the binding (Glide energy range À57.11 kcal molÀ1 to À23.65 kcal molÀ1
)
pocket of both the wild type PfDHFR-TS (PDB ID: 3QGT) and PfDHFR-TS structures, and the energy ranges are comparable to
quadruple mutant PfDHFR-TS (N51I, C59R, S108N, I164L, PDB the standard PfDHFR inhibitors (pyrimethamine, cycloguanil
ID: 3QG2) structures. The results of the docking studies and the and WR99210) and the native DHFR substrate, dihydrofolate
docked conformations of the best scoring ligands (14 and 19) in (Table 2).
Fig. 3 2D and 3D docking poses showing the interactions of compound 14 in the binding site of (A) wild (PDB ID: 3QGT) and (B) mutant PfDHFR-TS
(PDB ID: 3QG2).
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 5087--5095 | 5091

View Article Online
Paper
NJC
Fig. 4 2D and 3D docking poses showing the interactions of compound 19 in the binding site of (A) wild (PDB ID: 3QGT) and (B) mutant PfDHFR-TS
(PDB ID: 3QG2).
Fig. 3 and 4 illustrate the predicted binding poses for compounds (dihydrofolate) binding and lies in the proximity of residues 51
14 and 19, showing hydrogen bonding along with p–p interactions and 59. C59R mutation does not cause any significant changes in
and van der Waals interactions, with the expected binding pattern as the protein structure and causes no close contacts with the inhibi-
observed for PfDHFR inhibitors and dihydrofolate in the wild type tors. N51I causes movement in the main chain atoms of residues
and mutant PfDHFR protein.24 Compound 14, showing the lowest 48–51. Moreover, the function of the residue Asp54 is preserved
binding energy (À57.84 kcal molÀ1) and a considerably high Glide in the mutant protein and is not affected by the two proximal
XP score (À6.50 kcal molÀ1) for mutant PfDHFR, binds deep in the mutations N51I and C59R.40 Further, I164L mutation causes shifts
DHFR binding site, forming a hydrogen bond between the linker NH in the residues 164–167 and affects the active site gap, causing steric
group of 14 and the carboxylate oxygen side chain of Asp54 in both interactions between Phe58 and small inhibitors such as pyrimeth-
wild type and mutant PfDHFR. The morpholine rings attached to amine and cycloguanil.40 Also, the p-chlorophenyl moiety in pyr-
the triazine moiety of the compound 14 lie in the opposite end of imethamine and cycloguanil causes steric interference with the side
the active site when bound to mutant PfDHFR, forming a charge- chain of Asn108 in the active site modified by the first mutation
mediated hydrogen bond between one of the morpholine ring S108N. In contrast, WR99210, endowed with a long and flexible side
oxygen heteroatoms and the side chain nitrogen atom of Arg122 chain, could avoid such steric interactions, leading to effective
(Fig. 3B). Further, a p–p interaction between the aromatic ring of binding with the mutant protein. The test compounds 14 and 19,
Phe58 and the triazine ring of the compound was observed. having a flexible linker similar to WR99210, form p–p interactions
A similar interaction pattern was observed for compound 14 with Phe58, thus avoiding a steric clash with the aromatic side
(Glide energy: À57.84 kcal molÀ1) in the binding site of wild type chain of Phe58. Furthermore, the oxygen atom in the morpholine
PfDHFR (Fig. 3A). Compound 19 was predicted to have a low binding side chain linked to the triazine nucleus of compounds 14 and 19
energy and high Glide score in wild type and mutant PfDHFR forms a H-bonding interaction with evolutionarily conserved
(Table 2), with a H-bonding pattern between the linker NH group Arg122. Such a charge-mediated interaction of Arg122 is important
of the compound and the carboxylate oxygen side chain of Asp54 and is observed with the a-carboxylate of DHFR substrate dihydro-
and a charge-mediated H-bond between the morpholine oxygen and folate. The binding of the morpholine oxygen atom with the side
the Arg122 side chain (Fig. 4B).
chain of Arg122 provides a rigid docking site by restricting the
The influence of quadruple mutations (N51I, C59R, S108N, mobility of the flexible linker between the two rings in the designed
I164L) in DHFR is attributed to the movement in the active site inhibitors. Several observations have shown that drug molecules
residues, and interferes in the inhibitor binding. The active site designed to occupy the surface volume of the native substrate of the
residue Asp54, forming a H-bond with test compounds 14 and 19, protein will be less susceptible to resistance occurring due to steric
has been reported to be crucial for inhibitor and substrate clashes in the mutated protein binding site.41,42
5092 | New J. Chem., 2014, 38, 5087--5095
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
Thus, it is desirable to explore the binding pattern of the novel
lead compound in the preliminary stages of drug design against
mutant proteins. The molecular overlay of the docking poses of the
active test compounds along with the reference molecules on the
dihydrofolate surface envelope clearly shows that the test com-
pounds occupy a similar volume to that of the protein substrate,
unlike the dihydrofolate, and avoid steric clash with the side chain
of Asn108 (Fig. 5). The results from the present study give us
important preliminary information to design novel compounds
with a similar scaffold that may lead to more active compounds,
which would be a scope for our future communications.
The results of ADMET prediction by QikProp v3.543 are pre-
sented in Tables 3 and 4. Different pharmacokinetic parameters
of the compounds that showed good inhibitory potential in
malarial parasites were calculated. The most important of these
parameters together with their permissible ranges are listed in
Tables 3 and 4.
As a preliminary test of the drug-likeness of the compounds,
we calculated Lipinski’s rule of 5 using QikProp, which requires
compounds to have no more than 5 and 10 hydrogen bond donors
(donorHB) and acceptors (accptHB), respectively, molecular weights
(mol_MW) of less than 500 amu, and partition coefficients between
octanol and water (QPlogP(oct/wat)) of less than 5. Table 3 shows
the QikProp results for various parameters of Lipinski’s rule of 5. An
orally active compound should not have more than one violation of
these rules. In the present study, all the active test compounds
showed a number of violations of Lipinski’s rule of 5 less than the
maximum permissible value of 4, indicating that these active test
compounds are endowed with drug-like properties. All the com-
pounds showed 1 Lipinski’s rule of 5 violation except compound 19,
which showed 2 violations owing to it having mol_MW 4 500 and
accptHB 4 10. Prediction of oral drug absorption (Percent Human
Oral Absorption) was highly satisfactory for all the test compounds,
with the exception of compound 19, which showed a moderate
value. Studies have suggested that oral bioavailability is influenced
by a compound’s flexibility and can be measured by the number of
rotatable bonds (o15) and the polar surface area (70 Å2–200 Å2),
though it has been emphasized that this approach should be
considered with caution with respect to choice of descriptor
algorithm used and also because other factors can have a
significant influence on bioavailability.44 However, along with
Fig. 5 Superposition of the most active docked test compounds
(represented as grey sticks), pyrimethamine (blue sticks), cycloguanil (red
sticks), WR99210 (green sticks) and the PfDHFR substrate dihydrofolate
(yellow balls and sticks) bound to the binding site of quadruple mutant
(PDB ID: 3QG2), showing the fitting of the test compounds on the sub-
strate surface.
Table 3 Prediction of Lipinski’s ‘Rule of 5’ for the active test compoundsa
Compound mol_MW DonorHB AccptHB QPlogPo/w Rule of five
14
19
25
27
28
30
31
Pyr
Cg
472.55
513.64
472.63
471.65
485.68
485.68
499.70
248.71
253.73
2
2
2
2
2
2
2
4
5
12
12
9
9
9
9
9
3
3
2.127
2.56
1
2
1
1
1
1
1
0
0
4.532
4.358
4.361
4.728
4.667
1.809
0.888
a
All values calculated by QikProp v3.5; the explanations of the descriptors
are given in the text. Pyr = pyrimethamine, Cg = cycloguanil.
Table 4 Calculated ADMET properties
Percent Human
Oral Absorptiona QPPCacoa nm sÀ1
QPPMDCKa
QPlogHERGa
(480% high,
o25% poor)
(o25 poor,
QPlogBBa
(o25 poor, QPlogKhsaa (concern
QPlogS
(À6.5 to 0.5) (7.0–200.0) (0–15)
PSAa
#rotora
Compound
4500 great)
(À3.0 to 1.2) 4500 great) (À1.5 to 1.5) below À5)
14
19
25
27
28
30
31
85.123
61.96
100
87.557
88.046
90.327
89.246
1899.318
368.938
2146.387
484.098
514.064
523.024
476.685
412.287
111.854
À0.416
989.649
186.276
1129.502
249.85
266.608
271.635
245.718
468.849
126.604
À0.283
0.007
0.382
0.554
0.52
0.653
0.619
À0.243
À0.306
À4.251
À5.674
À4.903
À6.464
À5.977
À6.588
À5.988
À4.318
À4.578
À3.589
À3.557
À4.887
À5.163
À4.404
À5.496
À4.628
À2.978
À1.596
103.834
102.564
83.75
80.387
81.384
79.854
81.492
73.731
76.262
5
7
À0.32
À0.909
À0.598
À0.568
À0.635
À0.78
13
12
13
13
14
4
Pyrimethamine 84.346
Cycloguanil
À0.78
68.814
À0.17
2
a
Calculated using QikProp v3.5. Range/recommended values calculated for 95% of known drugs.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 5087--5095 | 5093

View Article Online
Paper
NJC
the polar surface area criterion, a total sum of H-bond donors Agricultural Research Service, Specific Cooperative Agreement
and acceptors criterion (r12) can be used, which is algorithm- No. 58-6408-1-603 is also acknowledged for partial support of
independent.44 In the present study, all the test compounds this work. The authors are also thankful to CIF-USIC, University
have a number of rotatable bonds o15 and the polar surface of Delhi, Delhi for NMR spectral data, RSIC, CDRI, Lucknow for
area falls satisfactorily within the permissible range (Table 4). mass data and Mr John Trott for technical support in anti-
Similarly, molecules obeying Lipinski’s rule of 5 could be more malarial activity testing at NCNPR.
likely to have good intestinal absorption or permeation, which is
confirmed by the predicted Caco-2 cell permeability (QPPCaco),
used as a model for the gut–blood barrier.45 QPPCaco predictions
for all the test compounds showed very good values except for
Notes and references
compounds 19, 27 and 31, which had moderately good values for
Caco-2 cell permeability, comparable to the value predicted
for the drug pyrimethamine. Further, QPlogKhsa, the prediction
for human serum albumin binding, was carried out for the test
compounds and all inhibitors, and the values were within the
expected range for 95% of known drugs (À1.5 to 1.5). Also, the
QikProp descriptor for the brain/blood partition coefficient
(QPlogBB) and the blood–brain barrier mimic MDCK cell perme-
ability (QPPMDCK) show satisfactory predictions for all the test
compounds and the reference compounds. In addition, the
aqueous solubility (QPlogS) parameter for the test compounds
was assessed and all the compounds were predicted to have
QPlogS values in the permissible range. Furthermore, the QPlo-
gHERG descriptor for the prediction of the IC50 value of HERG K+
channel blockage was predicted for the test compounds.
Compounds 14 and 25 were predicted to possess values in
1 J. Sachs and P. Malaney, Nature, 2002, 415, 680–685.
2 N. J. White, Lancet, 2010, 376, 2051–2052.
3 H. Noedl, Y. Se, K. Schaecher, B. L. Smith, D. Socheat and
M. M. Fukuda, N. Engl. J. Med., 2008, 359, 2619–2620.
4 R. Jambou, E. Legrand, M. Niang, N. Khim, P. Lim,
B. Volney, M. T. Ekala, C. Bouchier, P. Esterre, T. Fandeur
and O. Mercereau-Puijalon, Lancet, 2005, 366, 1960–1963.
5 A. M. Dondorp, S. Yeung, L. White, C. Nguon, N. P. J. Day,
D. Socheat and L. von Seidlein, Nat. Rev. Microbiol., 2010, 8,
272–280.
6 N. White, Philos. Trans. R. Soc., B, 1999, 354, 739–749.
7 N. Gargano, F. Cenci and Q. Bassat, Trop. Med. Int. Health,
2011, 16, 1466–1473.
8 N. J. White, Lancet, 2010, 376, 2051–2052.
9 H. C. Carrington, A. F. Crowther, D. G. Davey, A. A. Levi and
F. L. Rose, Nature, 1951, 168, 1080.
the permissible range comparable to the reference compounds 10 N. Alexis, J. Antimicrob. Chemother., 2006, 57, 1043–1054.
pyrimethamine and cycloguanil (Table 4).
11 A. Gregson and C. V. Plowe, Pharmacol. Rev., 2005, 57, 117–145.
12 Y. Yuthavong, J. Yuvaniyama, P. Chitnumsub, J. Vanichtanankul,
S. Chusacultanachai, B. Tarnchompoo, T. Vilaivan and
S. Kamchonwongpaisan, Parasitology, 2005, 130, 249–259.
13 I. M. Kompis, K. Islam and R. L. Then, Chem. Rev., 2005,
105, 593–620.
Conclusions
In summary, we report the synthesis, docking studies and evalua-
tion of the antimalarial activity of triazine–pyrimidine molecular
hybrids. The in vitro evaluation of these hybrids against the D6 and
W2 strains of P. falciparum revealed activity in the micromolar
range, with no cytotoxicity against VERO mammalian cell lines. The
active molecules were docked in the active site of wild type and
quadruple mutant PfDHFR-TS proteins to study the binding pat-
tern of the test molecules with DHFR. Compounds 14 and 19 were
found to show good binding with wild type and mutant DHFR
proteins, with an interaction pattern comparable to that of DHFR
inhibitors and the native DHFR substrate. Moreover, the test
compounds exhibited efficient binding with the mutant protein,
avoiding steric clashes resulting from amino acid mutations. The
calculated ADMET parameters for the test compounds indicated
good pharmacokinetic properties for compound 14, making it an
important candidate in the antimalarial drug discovery process.
14 K. H. Rieckmann, A. E. T. Yeo and M. D. Edstein, Trans.
R. Soc. Trop. Med. Hyg., 1996, 90, 568–571.
15 Y. Yuthavong, T. Vilaivan, S. Kamchonwongpaisan,
B. Tarnchompoo, C. Thongpanchang, C. Chitnumsub,
J. Yuvaniyama, D. Matthews, W. Charman, S. Charman,
L. Vivas and S. B. Katiyar, US Pat., 2009/0099220 A1, 2009.
16 Y. Yuthavong, B. Tarnchompoo, T. Vilaivan, P. Chitnumsub,
S. Kamchonwongpaisan, S. A. Charman, D. N. McLennan,
K. L. White, L. Vivas, E. Bongard, C. Thongphanchang,
S. Taweechai, J. Vanichtanankul, R. Rattanajak, U. Arwon,
P. Fantauzzi, J. Yuvaniyama, W. N. Charman and
D. Matthews, Proc. Natl. Acad. Sci. U. S. A., 2012, 109,
16823–16828.
17 M. Delves, D. Plouffe, C. Scheurer, S. Meister, S. Wittlin,
E. A. Winzeler, R. E. Sinden and D. Leroy, PLoS Med., 2012,
9, e1001169.
18 S. B. Katiyar, K. Srivastava, S. K. Puri and P. M. S. Chauhan,
Bioorg. Med. Chem. Lett., 2005, 15, 4957–4960.
Acknowledgements
D.S.R. thanks the Council of Scientific and Industrial Research 19 A. Agarwal, K. Srivastava, S. K. Puri and P. M. S. Chauhan,
[No. 02(0049)/12/EMR-II] New Delhi, India for financial sup- Bioorg. Med. Chem. Lett., 2005, 15, 531–533.
port. D.K. is thankful to CSIR for the award of junior and senior 20 D. Gravestock, A. L. Rousseau, A. C. U. Lourens, S. S.
research fellowship. P.P. is thankful to CSIR for the award of
Research Associate. The United States Department of Agriculture,
Moleele, R. L. van Zyl and P. A. Steenkamp, Eur. J. Med.
Chem., 2011, 46, 2022–2030.
5094 | New J. Chem., 2014, 38, 5087--5095
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014

View Article Online
NJC
Paper
21 J. Morgan, R. Haritakul and P. A. Keller, Lett. Drug Des. 34 N. Kumar, S. I. Khan and D. S. Rawat, Helv. Chim. Acta, 2012,
Discovery, 2008, 5, 277–280. 95, 1181–1197.
22 N. Azas, P. Rathelot, S. Djekou, F. Delmas, A. Gellis, 35 M. Jain, S. I. Khan, B. L. Tekwani, M. R. Jacob, S. Singh,
C. Di Giorgio, P. Vanelle and P. Timon-David, Farmaco,
2003, 58, 1263–1270.
P. P. Singh and R. Jain, Bioorg. Med. Chem., 2005, 13,
4458–4466.
23 A. Agarwal, K. Srivastava, S. K. Puri and P. M. S. Chauhan, 36 M. T. Makler and D. J. Hinrichs, Am. J. Trop. Med. Hyg., 1993,
Bioorg. Med. Chem. Lett., 2005, 15, 1881–1883. 48, 205–210.
24 S. Manohar, S. I. Khan and D. S. Rawat, Bioorg. Med. Chem. 37 J. Mustafa, S. I. Khan, G. Ma, L. A. Walker and I. A. Khan,
Lett., 2010, 20, 322–325. Lipids, 2004, 39, 167–171.
25 D. S. Rawat, S. Manohar, and U. C. Rajesh, Indian Pat., 38 D. S. Peterson, W. K. Milhous and T. E. Wellems, Proc. Natl.
Application No 661/DEL/2012. Acad. Sci. U. S. A., 1990, 87, 3018–3022.
26 S. Manohar, S. I. Khan and D. S. Rawat, Chem. Biol. Drug 39 E. M. Duffy and W. L. Jorgensen, J. Am. Chem. Soc., 2000,
Des., 2011, 78, 124–136. 122, 2878–2888.
27 S. Manohar, U. C. Rajesh, S. I. Khan, B. L. Tekwani and 40 J. Yuvaniyama, P. Chitnumsub, S. Kamchonwongpaisan,
D. S. Rawat, ACS Med. Chem. Lett., 2012, 3, 555–559.
28 H. Atheaya, S. I. Khan, R. Mamgain and D. S. Rawat, Bioorg.
Med. Chem. Lett., 2008, 18, 1446–1449.
J. Vanichtanankul, W. Sirawaraporn, P. Taylor, M. D.
Walkinshaw and Y. Yuthavong, Nat. Struct. Biol., 2003, 10,
357–365.
29 N. Kumar, S. I. Khan, Beena, G. Rajalakshmi, P. Kumaradhas 41 N. M. King, M. Prabu-Jeyabalan, E. A. Nalivaika and C. A.
and D. S. Rawat, Bioorg. Med. Chem, 2009, 17, 5632–5638. Schiffer, Chem. Biol., 2004, 11, 1333–1338.
30 N. Kumar, S. I. Khan, H. Atheaya, R. Mamgain and D. S. 42 K. P. Romano, A. Ali, W. E. Royer and C. A. Schiffer, Proc.
Rawat, Eur. J. Med. Chem., 2011, 46, 2816–2827. Natl. Acad. Sci. U. S. A., 2010, 107, 20986–20991.
¨
31 J. Singh, R. K. S. Dhakarey, S. V. Singh, M. K. Suthar, 43 QikProp User Manual Copyright r 2013 Schrodinger, LLC.
J. K. Saxena and A. K. Dwivedi, Chem. Biol. Interface, 2012, 44 J. J. Lu, K. Crimin, J. T. Goodwin, P. Crivori, C. Orrenius,
2, 347–361.
L. Xing, P. J. Tandler, T. J. Vidmar, B. M. Amore, A. G. E.
Wilson, P. F. W. Stouten and P. S. Burton, J. Med. Chem.,
2004, 47, 6104–6107.
32 K. Doktorov, V. B. Kurteva, D. Ivanova and I. Timtcheva,
ARKIVOC, 2007, 15, 232–245.
33 V. B. Kurteva and C. A. M. Afonso, Green Chem., 2004, 6, 45 P. Artursson, K. Palm and K. Luthman, Adv. Drug Delivery
183–187. Rev., 2001, 46, 27–43.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2014
New J. Chem., 2014, 38, 5087--5095 | 5095