3D-QSAR pharmacophore modelling, virtual screening and docking studies for lead discovery of a novel scaffold for VEGFR 2 inhibitors: Design, synthesis and biological evaluation

Article abstract of DOI:10.1016/j.bioorg.2019.102988

A series of novel 6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives was successfully designed, synthesized and evaluated as a new chemical scaffold with vascular endothelial growth factor receptor (VEGFR 2) inhibitory activity. Compounds 6c and 6b showed enzyme inhibition of 97% and 87% at 10 μM, respectively, and exhibited potent dose-related VEGFR 2 inhibition with IC₅₀ values of 0.85 μM and 2.26 μM, respectively. The design of the 6,7-dihydro-5H-cyclopenta[d]pyrimidine scaffold was implemented via consecutive molecular modelling protocols prior to the synthesis and biological evaluation of the derivatives. First, sorafenib was docked in the binding site of VEGFR 2 to study its binding orientation and affinity, followed by the generation of a valid 3D QSAR pharmacophore model for use in the virtual screening of different 3D databases. Structures with promising pharmacophore-based virtual screening results were refined using molecular docking studies in the binding site of VEGFR 2. A novel scaffold was designed by incorporating the results of the pharmacophore model generation and molecular docking studies. The new scaffold showed hydrophobic interactions with the kinase front pocket that may be attributed to increasing residence time in VEGFR 2, which is a key success factor for ligand optimization in drug discovery. Different derivatives of the novel scaffold were validated using docking studies and pharmacophore mapping, where they exhibited promising results as VEGFR 2 inhibitors to be synthesized and biologically evaluated. 6,7-dihydro-5H-cyclopenta[d]pyrimidine is a new scaffold that can be further optimized for the synthesis of promising VEGFR 2 inhibitors.

Full text of DOI:10.1016/j.bioorg.2019.102988

Accepted Manuscript
3D-QSAR Pharmacophore Modelling, Virtual Screening and Docking Studies
for Lead Discovery of a Novel Scaffold for VEGFR 2 Inhibitors: Design, Syn-
thesis and Biological Evaluation
Mahitab K. Sobhy, Samar Mowafy, Deena S. Lasheen, Nahla A. Farag, Khaled
A.M. Abouzid
PII:
S0045-2068(19)30151-8
https://doi.org/10.1016/j.bioorg.2019.102988
102988
DOI:
Article Number:
Reference:
YBIOO 102988
Bioorganic Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
29 January 2019
14 May 2019
17 May 2019
Please cite this article as: M.K. Sobhy, S. Mowafy, D.S. Lasheen, N.A. Farag, K.A.M. Abouzid, 3D-QSAR
Pharmacophore Modelling, Virtual Screening and Docking Studies for Lead Discovery of a Novel Scaffold for
VEGFR 2 Inhibitors: Design, Synthesis and Biological Evaluation, Bioorganic Chemistry (2019), doi: https://
doi.org/10.1016/j.bioorg.2019.102988
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

3D-QSAR Pharmacophore Modelling, Virtual Screening and Docking Studies for Lead Discovery of a
Novel Scaffold for VEGFR 2 Inhibitors: Design, Synthesis and Biological Evaluation
a
b
Mahitab K. Sobhya,*,
, Samar Mowafy , Deena S. Lasheen , Nahla A.
nahla.farag@miuegypt.edu.eg
b,c
Faraga,*,
, Khaled A.M. Abouzid
nahla.farag@miuegypt.edu.eg
aPharmaceutical Chemistry Department, Faculty of Pharmacy, Misr International University, Km28 Cairo-Ismailia
Road (Ahmed Orabi District), Cairo-Egypt
bPharmaceutical Chemistry Department, Faculty of Pharmacy, Ain Shams University, Abbassia, 11566, Cairo, Egypt
cDepartment of Organic and Medicinal Chemistry, Faculty of Pharmacy, University of Sadat City, Sadat City , Egypt
*Corresponding author.
Graphical abstract
Highlights
6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives were synthesized and biologically evaluated for their
effect as VEGFR 2 inhibitors.
In order to rational new scaffold with potential VEGFR 2 inhibitory activity we perform consecutive protocols
of molecular modelling.
Studying the docking result of sorafenib in the binding site of VEGFR 2.
performing 3D-QSAR pharmacophore model generation protocol to generate a valid model for virtual
screening of different databases to generate new hits as VEGFR 2 inhibitors.
The structures with promising results of pharmacophore based virtual screening were then subjected to
molecular docking in the binding site of VEGFR 2, docking results were analysed and filtered into 3 HITs.
The structure of the 3 HITs, along with their pharmacophore mapping and docking results were used to
develop a new scaffold of VEGFR 2 inhibitors.
Abstract
A series of novel 6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives was successfully
designed, synthesized and evaluated as a new chemical scaffold with vascular endothelial growth
factor receptor (VEGFR 2) inhibitory activity. Compounds 6c and 6b showed enzyme inhibition of
97% and 87% at 10 µM, respectively, and exhibited potent dose-related VEGFR 2 inhibition with
IC50 values of 0.85 µM and 2.26 µM, respectively. The design of the 6,7-dihydro-5H-
cyclopenta[d]pyrimidine scaffold was implemented via consecutive molecular modelling protocols
prior to the synthesis and biological evaluation of the derivatives. First, sorafenib was docked in the
binding site of VEGFR 2 to study its binding orientation and affinity, followed by the generation of
a valid 3D QSAR pharmacophore model for use in the virtual screening of different 3D databases.
Structures with promising pharmacophore-based virtual screening results were refined using
molecular docking studies in the binding site of VEGFR 2. A novel scaffold was designed by
incorporating the results of the pharmacophore model generation and molecular docking studies.
The new scaffold showed hydrophobic interactions with the kinase front pocket that may be
attributed to increasing residence time in VEGFR 2, which is a key success factor for ligand
optimization in drug discovery. Different derivatives of the novel scaffold were validated using
1

docking studies and pharmacophore mapping, where they exhibited promising results as VEGFR 2
inhibitors to be synthesized and biologically evaluated. 6,7-dihydro-5H-cyclopenta[d]pyrimidine is
a new scaffold that can be further optimized for the synthesis of promising VEGFR 2 inhibitors.
1. Introduction
Vascular endothelial growth factor receptor (VEGFR) is a member of the receptor tyrosine
kinase (RTK) family. It is essential for the growth of blood vessels, either, new (vasculogenesis) or
pre-existing blood vessels (angiogenesis) [1]. VEGFR is activated by the binding of vascular
endothelial growth factors (VEGFs) to its extracellular domain, causing dimerization of the receptor
[2]. This conformational change makes the ATP binding site of the receptor available for ATP
binding, causing phosphorylation of the tyrosine residue of the receptor and subsequent activation
[2,3]. There are three subtypes of VEGFRs: VEGFR 1, 2 and 3. VEGFR 1 (Flt 1) is critical for the
regulation of macrophage and monocyte migration, VEGFR 2 (KDR) is responsible for vascular
endothelial cell’s normal and pathological developments, and VEGFR 3 (Flt 4) is responsible for
the development of lymphatic endothelial cells and the spreading of cancerous cells to lymph nodes
[4]. Since VEGFR 2 is the subtype responsible for angiogenesis and vasculogenesis [3], the
inhibition of VEGFR 2 will affect the blood supply to tumour cells, inhibiting their growth,
proliferation and metastasis [5,6]. Inhibition of the VEGFR signalling pathway is a crucial
therapeutic target for tumour inhibition. Targeting the neovascularization process of cancer cells
will promote their starvation and death in a short time due to their rapid growth rate.
VEGFR small molecule inhibitors are categorized into two types according to the conformation
of the receptor. Type I inhibitors compete with ATP molecules at the ATP binding site to bind to
the receptor in its active form, adopting the Asp-Phe-Gly (DFG)-in conformation and consequently
preventing the phosphorylation and activation of the receptor. In contrast, Type II inhibitors are
non-ATP competitors that bind and stabilize the inactive form of the receptor, the DFG-out
conformer, which is achieved through flipping of the DFG motif of the receptor to reveal an extra
hydrophobic pocket [6,7]. Several molecules have been approved by the FDA for the treatment of
different types of cancer [9]; for example, sorafenib is an oral multikinase dual-acting inhibitor that
affects tumour cells directly through inhibiting the rapidly accelerated fibrosarcoma
(RAF)/mitogen-activated protein kinase (MEK)/extracellular signal-regulated kinase (ERK)
signalling pathway or targeting the tumour vasculature by inhibiting platelet-derived growth factor
receptor-B (PDGFR-B) and VEGFR [9]. Sorafenib is a type II kinase inhibitor with a reported IC50
value of 0.057 µM for PDGFR-B and 0.09 µM for VEGFR 2 [10].
Several drug design studies were carried out with the aim of improving the potency and
pharmacokinetic properties of existing VEGFR inhibitors by structural modification or finding new
candidates for VEGFR inhibition. In this study, a novel 6,7-dihydro-5H-cyclopenta[d]pyrimidine
scaffold was designed using consecutive computational protocols. The potency and effectiveness of
the proposed derivatives as VEGFR 2 inhibitors were validated using molecular docking and
pharmacophore mapping protocols. Then, the proposed derivatives were synthesized. Their VEGFR
2 inhibitory activity was evaluated, and their dose-related IC50 values were measured and compared
to that of sorafenib as a reference compound.
2. Rationale
2.1. Molecular modelling studies
Computer-aided drug design (CADD) adopts different computational techniques to optimize
commercially available drugs into new ligands with potential biological activities against certain
2

biological targets. CADD is based on quantum mechanics and molecular modelling techniques and
is used to reduce the expenses and time consumed by rational drug design.
In the current study, a novel 6,7-dihydro-5H-cyclopenta[d]pyrimidine scaffold was designed
using different molecular modelling protocols: the molecular docking of sorafenib as a potent
VEGFR 2 inhibitor to understand binding affinity and orientation inside the binding site of VEGFR
2, 3D QSAR (Three Dimensional Quantitative Structure-Activity Relationship) pharmacophore
model generation and virtual screening of 3D databases using Accelrys Discovery Studio software
version 4.5 (DS 4.5, Accelrys Ltd., UK). The 3D QSAR pharmacophore model generation protocol
builds a pharmacophore model using a set of ligands with known activity to identify the essential
functional features contributing to their high potency and biological activity. The generated
pharmacophore model was used for the virtual screening of three 3D databases: the
DruglikeDiverse, MiniMaybridge, and scPDB databases. The selected hits of the virtual screening
were analysed and subjected to molecular docking into the crystal structure of VEGFR 2 with
protein data bank code 3WZE in complex with sorafenib as a reference inhibitor. The good
correlation between docking scores and pharmacophore fit values provided a reliable basis for
designing new VEGFR 2 inhibitors.
2.1.1. Molecular docking of sorafenib
A molecular docking protocol was implemented using the CDOCKER protocol under the
receptor ligand interaction protocol of Accelrys Discovery Studio 4.5. CDOCKER is a grid-based
molecular docking method that uses the CHARMm force field-based molecular dynamics (MD)
search algorithm to dock ligands into the binding site of the receptor. Different conformations of the
ligand are generated with different scoring functions, such as -CDOCKER_ENERGY (CHARMm
energy of the interaction energy and ligand strain) and -CDOCKER_INTERACTION_ENERGY
(interaction energy only).
The determination of the binding interactions and orientation of VEGFR 2 kinase inhibitors
was studied using the X-ray crystal structure of VEGFR 2 with sorafenib (PDB code 3WZE). The
molecular docking protocol (CDOCKER) of Accelrys Discovery Studio 4.5 was used, and the target
protein was prepared and the active binding site identified based on the positional coordinates of the
co-crystallized inhibitor. The results of the CDOCKER protocol were validated by re-docking of the
co-crystallized structure of the reference sorafenib inside the active site of VEGFR 2 (PDB code
3WZE). The root mean square deviation (RMSD) between the re-docked conformer and the co-
crystallized conformer of sorafenib is 0.257, (Fig. 1), which confirms the validity of the docking
protocol.
Molecular docking simulation provides insight into the binding interaction and affinity
between the compound and the receptor. Promising biological activity is indicated by lower
CDOCKER energy and similar binding interactions to that of the reference co-crystallized inhibitor.
A docking study of sorafenib resulted in a binding energy score of -48.92 with a total of ten
hydrogen bonds. The essential amino acid Cys919 located in the hinge region binds to the nitrogen
of the pyridine ring, the NH and the methyl group of the amide moiety. In addition, Glu917 in the
hinge region binds to the CH proton at the 6-position of the pyridine ring; the urea moiety binds to
the receptor through various hydrogen bonding interactions where both urea NH groups interact
with Glu885 in the C-helix; and the carbonyl group interacts with Asp1046 in the DFG motif and
with Cys1045. The fluoride atoms of the external trifluoromethyl moiety interact with His1026 and
Cys1045. Furthermore, the pyridine moiety forms hydrophobic interactions with a hydrophobic
pocket formed by amino acid residues Cys919, Leu1035, Ala866, Val848 and Leu840, and the
phenyl ring binds to another hydrophobic pocket through interactions with Lys868, Cys1045,
3

Val916, Phe1047 and Val848. In addition, the terminal phenyl ring forms hydrophobic bond with
Leu889, the meta trifluoromethyl substitution forms hydrophobic interactions with Val898, Ile1044,
His1026 and Leu1019, and the para chloro substituent exhibits hydrophobic interactions with
Leu1019 and Ile888 (Table 4). The resulting interactions resemble the reported binding mode of
sorafenib in the X-ray crystal structure of VEGFR 2 with PDB code 3WZE [8], (Fig. 2).
2.1.2.
3DQSAR pharmacophore model generation and validation
2.1.2.1. Assortment of anticancer library
A set of 25 VEGFR 2 inhibitors of known activity with IC50 values ranging from 0.027- 9.4 µM
were collected from the literature [11], (Fig. 3). Library ligands were prepared using the prepare
ligands protocol of Accelrys Discovery Studio 4.5. Ligand preparation is essential for energy
minimization, the correction of incorrect valences and the protonation of ligands.
2.1.2.2. 3D QSAR pharmacophore model generation
3D QSAR pharmacophore is an advanced common feature pharmacophore ligand-based
protocol of Discovery Studio 4.5 that is used to explore essential pharmacophore features from a set
of ligands with known activity against the biological target of interest. Pharmacophore model
generation is implemented using the HypoGen algorithm [8,9], which considers only features that
are common to active compounds, while common features among inactive compounds are excluded
[13]. For the generation of potential pharmacophore models, a feature-mapping protocol was
initiated to identify important chemical features imbedded within the training set compounds (T1-
15). Subsequently, a pharmacophore model was generated using twenty-five compounds that were
divided into a training set (T1-15) and a test set (S15-25), (Fig. 3). The features were driven by a
feature-mapping protocol: hydrogen bond acceptor (HBA), hydrogen bond donor (HBD),
hydrophobic (HYP) and ring aromatic (RA) features were selected, the IC50 of the compound was
set as the activity property of the model, the uncertainty property was changed from the default
value of 3.0 to 1.5, and validation was set as true with maximum omitted features of zero.
Accordingly, ten predictive pharmacophore model hypotheses were generated and validated.
.
2.1.2.3. Validation of pharmacophore models
The HypoGen algorithm generates different hypotheses, and the best-generated hypothesis is
chosen according to different parameters. The cost difference is calculated from the null cost and
the total cost of each hypothesis, where the total cost is the summation of three components: weight,
error and configuration cost. The weight component increases in a Gaussian form as the feature
weight deviates from an ideal value (2.0); the error component increases as the root mean squared
(RMS) difference between the estimated and experimental activities for the training set compounds
increases; and the configuration component is a constant cost that depends on the hypothesis space
being optimized. The fixed cost and null cost are another two theoretical costs that are measured by
the algorithm, where the fixed cost is the minimum possible cost of a model that fits all the data
perfectly, and the null cost is the maximum cost of a model devoid of any features. For the model to
be statistically significant, the difference between the null and fixed cost must be greater than that
between the total and fixed cost, and to generate a statistically reliable model above 90%; the
4

difference between the null and total cost should be greater than or equal to 70 bits [14].
Accordingly, the best generated hypothesis is the model with the highest cost difference.
The generated pharmacophore model is validated by the measurement of its statistical
significance and its ability to estimate the biological activity of new compounds. Model validation
depends on the cost difference, accuracy of the predicted activity of the training set compounds
compared to their experimental activity, and mapping of sorafenib on the generated model as a
reference standard. Ten 3D QSAR pharmacophore models were generated with at least three
chemical features each (Table 1). Hypothesis 1 is the most statistically significant hypothesis
among the ten generated pharmacophore model hypotheses, and it has five features: one hydrogen
bond acceptor (HBA), three hydrophobic (HYP), and one ring aromatic (RA) (Fig. 4). Accordingly,
hypothesis 1 has the highest cost difference of 140.98, indicating a predictive power above 90%,
with the lowest (RMS) of 1.03 reflecting a high correlation between predicted and experimental
activity and a high correlation coefficient of 0.975 corresponding to the ability of the model to
predict the activity of the training set compounds (Table 2).
2.1.2.4. Pharmacophore mapping of sorafenib
The generated pharmacophore model was further validated by mapping an active reference
compound, sorafenib, into the generated model using the ligand mapping protocol of Accelrys
Discovery Studio 4.5 and scored a fit value of 6.19 (Fig. 5).
2.1.3. Virtual screening
Virtual screening is the adoption of computational techniques to identify potentially active
compounds from existing databases or virtual libraries. The screening was carried out using the
search and edit 3D database protocol of Accelrys Discovery Studio 4.5. Three 3D databases,
DruglikeDiverse, MiniMaybridge, and scPDB, containing a total of 12,930 compounds were
mapped into the generated pharmacophore model. The mapping of the three databases resulted in
2370 compounds with the features of the validated pharmacophore model (hypothesis 1). The best-
matched compounds were filtered and selected according to their fit values into 136 compounds
with fit values higher than that of sorafenib.
2.1.4. Docking studies of virtually screened compounds
The docking results of the 136 compounds of the pharmacophore-based virtual screening were
ranked according to their binding affinity and binding mode. The top three hits were KM10259,
CDI665171 and 2r4b_GW7. KM10259 has a fit value of 7.19 and the highest docking score of -
42.28. The urea NH groups form two hydrogen bonds with Glu885, and the urea carbonyl group
forms two hydrogen bonds with Cys1045 and Asp1046. In addition, hydrophobic interactions occur
between the furan ring and a hydrophobic pocket formed by residues Val848, Leu1035, Ala866 and
Cys919, between the terminal phenyl ring and Leu889, and thepara chloro substituent interacts
with Leu1019, Ile1044 and Val898 (Table 3).
The docking score of CDI665171 is -33.81, its fit value is 7.80, and it binds to the receptor
through hydrogen bonds similar to those of KM10259. Its hydrophobic interactions involve the
binding of the terminal methyl group to residues Leu1035, Cys919, Phe918 and Leu840, the
binding of the thiazole ring to residues Leu1035, Cys1045, Val916, Val848 and Ala866 and the
interaction of the phenyl ring attached to the thiazole ring with Lys868, Cys1045, Val899 and
5

Val916. In addition, hydrophobic interactions between the terminal phenyl ring and the Leu889
amino acid were observed (Table 3).
2r4b_GW7, with a fit value of 7.10 and a CDOCKER binding score of -35.08, binds to the
receptor through interaction of the N1 of the pyrimidine ring with Cys919 in the hinge region and
binding of the CH proton at the 2-position of the pyrimidine ring with Glu917. Cys919 and Glu917
are key amino acids for the hydrogen bonding of the adenine ring of ATP to the hinge region of
VEGFR 2 [15]. The hydrophobic interactions of 2r4b_GW7 involve the binding of the thienyl ring
to the kinase front pocket formed by residues Phe918, Leu840, Leu1035 and Ala866 and the
hydrophobic binding of the pyrimidine ring to residues Leu1035, Cys919 and Ala866. In addition,
the phenyl ring binds to Val916, Cys1045, Val848, Phe1047 and Lys868, the 3-chloro substituent
of the phenyl ring forms hydrophobic interactions with Lys868 and Val916, and the terminal phenyl
ring forms hydrophobic interactions with Leu889 (Table 3). The three hits showed consistency in
rendering the interactions of VEGFR 2 inhibitors.
2.2.
Design strategy
In studying the structure-activity relationship of reported VEGFR 2 inhibitors and analysing
their binding modes, three main features of type II inhibitors of VEGFR 2 were reported: most
inhibitors include a flat heteroaromatic ring system that binds to the NH of Cys919 in the hinge
region; a hydrogen bond donor and acceptor pair, which can be a urea or an amide group that binds
with Glu885 and Asp1046 in the enzyme DFG motif, where two hydrogen bonds are formed
between NH of the urea or amide group and Glu885 and another hydrogen bond is between the CO
group and Asp1046 [16]; and finally, the binding of the terminal aryl group in the newly created
allosteric hydrophobic pocket exposed upon flipping of the DFG loop, as shown in the inactive
conformation or DFG-out conformation of the enzyme [17].
Compound 2r4b_GW7 has a fit value of 7.10 and exhibits the 5 features (1HBA, 3HYP and
1RA) of the validated pharmacophore model, while sorafenib has a fit value of 6.19 and exhibits
only 4 features (1HBA, 3HYP), (Table 4). Molecular docking studies of the hit compound
2r4b_GW7 revealed a hydrogen bond between N1 of the thienyl pyrimidine moiety and Cys919
and a hydrogen bond between the CH proton at the 2-position of the pyrimidine ring and Glu917.
Both Cys919 and Glu917 are key amino acids for the hydrogen bonding of ATP to the hinge region
of VEGFR 2 [15]. The thienyl pyrimidine moiety forms additional hydrophobic interactions at the
front pocket of the kinase domain with residues Leu840, Ala866, Phe918 and Leu1035, (Table 4).
Hydrophobic interaction with the kinase front pocket complements the hydrogen bonding with the
hinge region, increasing the residence time [8], which is a key factor in improving potency in vivo
and enhancing the correlation between the in vitro and in vivo efficacy of drugs [18].
Utilizing the aforementioned essential structural features, bioisosteric modification strategies
and the results of the 3D QSAR pharmacophore model generation and molecular modelling
protocols, a proposed scaffold of 6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives was designed
and synthesized. The design of the new scaffold was based on the replacement of the N-
methylpicolinamide moiety of sorafenib with a 6,7-dihydro-5H-cyclopenta[d]pyrimidine moiety.
Bioisosteric replacement was carried out by the substitution of the oxygen linker with a classical
NH bioisostere, (Fig. 6). Introducing new substituents at different positions of an aromatic ring of a
hit compound may optimize the binding affinity of the drug with the receptor in addition to
conferring the physicochemical properties essential for drug distribution and metabolism. Different
derivatives of (I) were proposed with alternative lipophilic substituents at the terminal aromatic ring
6

to occupy the hydrophobic allosteric pocket. Substituents were either electron withdrawing, such as
chloro and trifluoromethyl groups, or electron donating, such as methyl and methoxy groups.
2.2.1.
Validation of the proposed scaffold of VEGFR 2 inhibitors
2.2.1.1. Pharmacophore mapping of the proposed VEGFR 2 inhibitors
The activity of the proposed scaffold as an effective VEGFR 2 inhibitor was validated through
mapping of the proposed derivatives into the validated pharmacophore model. The results showed
that the proposed derivatives 6a, 6b, 6c, 6d and 6e exhibited five features of the generated
pharmacophore model with fit values ranging from 6.63 to 9.21 (Table 5), which are higher than
that of the reference compound sorafenib, at 6.19.
2.2.1.2. Docking of the proposed derivatives of VEGFR 2 inhibitors
Molecular docking of the proposed derivatives is essential for the assessment of binding
affinity and binding interactions with key residues at the active binding site of VEGFR 2. The
proposed derivatives 6a-6e were docked inside the binding site of VEGFR 2 with PDB code 3WZE
with a similar binding mode. The N1 of the cyclopentapyrimidine ring forms two hydrogen bonds
with Phe918 and Cys919 at the hinge region, and the CH proton at the 2-position of the pyrimidine
ring forms a hydrogen bond with Glu917 at the hinge region. In addition, both urea NH groups
interact with Glu885 in the C-helix, and the carbonyl group of the urea moiety interacts with
Asp1046 and Cys1045 (Table 5). Moreover, compound 6e forms two additional hydrogen bonds
due to the interaction of the fluoride atoms of the terminal trifluoromethyl with His1026 and
Cys1045. Furthermore, the pyrimidine moiety forms hydrophobic interactions with a hydrophobic
pocket consisting of residues Cys919, Leu1035 and Ala866, and the phenyl ring binds to another
hydrophobic pocket through interactions with amino acids Lys868, Cys1045, Val916, Phe 1047 and
Val848. In addition, the terminal phenyl ring interacts hydrophobically with Leu889, and the
substituents of the terminal phenyl ring interact with different amino acids according to the type and
position of the substituent. The hydrophobic interactions of the proposed derivatives are similar to
those of the reference compound except for an additional hydrophobic interaction between the
cyclopentane moiety and the entrance region of the adenine binding site. The cyclopentane moiety
in derivatives 6a and 6e also forms additional hydrophobic interactions with Leu1035 and Leu840,
while compounds 6b, 6c and 6d bind with Leu1035, Leu840 and Phe918. Hydrophobic interaction
at the kinase entrance region might prolong the residence time and improve the binding kinetics of
VEGFR 2 receptor [18], (Fig. 7, Table 5).
3. Results and discussion
3.1. Chemistry
The synthetic route of the key intermediates and final compounds of the proposed scaffold is
illustrated in Schemes 1 and 2, respectively. The synthesis of the key intermediates (2a-e) was
carried out by reacting p-nitroaniline with different isocyanates in dry dichloromethane for 48 hours
at room temperature [19] to give compounds (1a–e), followed by the reduction of 4-nitrophenyl
substituted phenyl ureas (1a-e) using Pd/C in methanol [20] to give the corresponding amino
derivatives (2a-e).
The synthesis of the 6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives (6a-e) was achieved
through multistep reactions starting with a one-step Thorpe-Ziegler cyclization by reacting
adiponitrile and sodium hydride in toluene [21] to give 2-aminocyclopent-1-ene carbonitrile 3,
which was then cyclized by heating under reflux with formic acid and acetic anhydride for 48 hours
7

[22] to provide the 6,7-dihydro-5H-cyclopenta[d] pyrimidinone derivative 4. The chlorination of 4
was carried out by heating with phosphorus oxychloride [23]. The key intermediates (2a-e) were
then coupled with the chloro derivative 5 to give the corresponding 6,7-dihydro-5H-cyclopenta[d]
pyrimidine derivatives (6a-e), and the reaction was followed up using TLC.
3.2.
Biological analysis
Initial screening of all synthesized final compounds was carried out at a single dose of 10 µM
to evaluate their inhibitory activity against VEGFR 2 kinase. The percentage inhibition of 6,7-
dihydro-5H-cyclopenta[d]pyrimidine derivatives 6b and 6c was 87% and 97%, respectively. The
enzymatic activity of the other synthesized compounds ranged from weak to moderate inhibition.
Compounds 6d and 6e showed low inhibition despite their promising molecular modelling results
due to their poor solubility in DMSO (Table 7).
The inhibitory effect of the synthesized compounds on VEGFR 2 revealed that the
unsubstituted terminal phenyl ring in 6a resulted in weak enzyme inhibition. However, the addition
of different substituents, especially at the meta position of the terminal phenyl ring, resulted in a
higher enzyme inhibitory effect [4], as shown in compounds 6b and 6c due to the addition of the 3-
Cl-4-CH3 and 4-Cl-3-CF3 substituents. In addition, 6b and 6c exhibited high docking scores of -
34.28 and -29.29, respectively, and high pharmacophore mapping fit values of 6.63 for compound
6b and 9.21 for compound 6c. Synthesized compounds with high percentage inhibition (above
80%) against VEGFR 2 were subjected to five-dose testing to determine their IC50 value, and values
of 2.26 μM and 0.85 μM were obtained for compounds 6b and 6c, respectively (Table 8).
4. Conclusion
In summary, a novel VEGFR 2 inhibitor scaffold was designed using several molecular
modelling protocols. A valid 3D QSAR pharmacophore model was generated, followed by
pharmacophore-based virtual screening of three different databases. Virtually screened compounds
were filtered by docking into the VEGFR 2 structure with PDB code 3WZE. According to the
results of molecular docking studies, a new scaffold of VEGFR 2 was designed based on
substitution of the N-methylpicolinamide moiety of sorafenib with 6,7-dihydro-5H-
cyclopenta[d]pyrimidine.
Five different derivatives of the proposed scaffold were tested for validation using
pharmacophore mapping and molecular docking protocols. The proposed derivatives resulted in fit
values, binding energies and binding interactions that are comparable to those of sorafenib.
Following the molecular modelling results, the five proposed structures were synthesized and
subjected to biological evaluation against the VEGFR 2 enzyme. Compounds 6b and 6c exhibited
high percentage inhibition of 87% and 97%, respectively, at 10 µM. The compounds with the
highest enzyme inhibitory effect were tested for VEGFR 2 inhibition and demonstrated IC50 values
in the sub-micromolar range. The IC50 values of compounds 6c and 6b were 0.85 µM and 2.26 µM,
respectively. Therefore, compounds 6c and 6b represent a new molecular scaffold that can be
further optimized and used for the design of promising potent VEGFR 2 inhibitors.
5. Experimental
8

Starting material and reagents were purchased from Sigma Aldrich or Alfa-Aeser organics and
were used without further purification. Melting points were recorded using BUCHI B-540 apparatus
and were uncorrected. Reactions were monitored using thin layer chromatography (TLC) purchased
from Merck and performed on 0.255 mm silica gel plates, with visualization under U.V. light (254
nm). The hydrogenation process was carried out using hydrogenator (Parr Shaker) apparatus. FT-IR
1
spectra were recorded on a Perkin Elmer FT-IR spectrophotometer and H NMR spectra were
recorded on BRUCKER 400 MHz spectrophotometer at Ain Shams University. Mass spectra were
recorded on Thermo Scientific ISQ LT gas chromatograph mass spectrometer at the regional center
for Mycology and Biotechnology Al-Azhar University and on thermos Q-exactive orbitrap
instrument at laboratory for single cell mass spectrometry, quantitative biology center, Japan.
5.1. Chemistry
5.1.1. General procedure for the preparation of compounds (1a–e)
A solution of p-nitroaniline (1 g, 6 mmol: 1 equiv.) in dry dichloromethane (20 mL), was stirred
with the appropriate isocyanate (6 mmol: 1 equiv.) at room temperature for 48 hours. The resulting
mixture was filtered and the filtered solid was crystallized from ethanol to yield derivatives (1a–
e)[19].
5.1.1.1. 1-(4-Nitrophenyl)-3-phenylurea (1a)
Yield 40% as yellowish white crystals, m.p 226-228 °C, (as reported) [24].
5.1.1.2. 1-(3-Choloro-4-methylphenyl)-3-(4-nitrophenyl)urea (1b)
Yield 50% as yellow orange crystals, m.p 182-185 °C, (as reported) [10].
5.1.1.3. 1-(3-trifluoromethyl-4-chlorophenyl)-3-(4-nitrophenyl)urea (1c)
Yield 90% as pale yellow crystals, m.p 238-240 °C. FT-IR (ύ max, cm-1): 1596 (C=C), 1646 (C=N),
1717 (C=O), 3128 (CH aromatic), 3352 (NH).
5.1.1.4. 1-(3-Methoxyphenyl)-3-(4-nitrophenyl)urea (1d)
Yield 47% as yellow crystals, m.p 210-212 °C, (as reported) [24].
5.1.1.5. 1-(4-Nitrophenyl)-3-(3-(trifluoromethyl)phenyl)urea (1e)
Yield 60% as greenish yellow crystals, m.p 260 °C. FT-IR (ύ max, cm-1): 1575 (C=C), 1621 (C=N),
1720 (C=O), 3162 (CH aromatic), 3356 (NH).
5.1.2. General procedure for the preparation of compounds (2a–e)
Pd-C (0.1 g, 10%) was added to a solution of the appropriate nitro phenyl urea derivative (1a-e) (4
mmol) in methanol (100 ml), the mixture was stirred under H2 at room temperature, at 60 bar for 4
hours. The catalyst was removed by filtration over celite. Then the filtrate was concentrated and
dried to afford crystals of compounds (2a–e) which were then recrystallized from methanol [20].
5.1.2.1. 1-(4-Aminophenyl)-3-phenylurea (2a)
Yield 70% as white crystals, m.p 224-227 °C, (as reported) [25].
9

5.1.2.2. 1-(4-Aminophenyl)-3-(3-Choloro-4-methylphenyl)urea (2b)
Yield 80% as grey crystals, m.p 203-205 °C, (as reported) [10].
5.1.2.3. 1-(4-Aminophenyl)-3-(3-trifluoromethyl-4-chlorophenyl)urea (2c)
Yield 80% as grayish white crystals, m.p 187-190 °C. 1H NMR (400 MHz, DMSO-d6) δ 6.52-6.54
(d, J = 8 Hz, 1H, H2’), 7.07-7.09 (d, J = 8 Hz, 1H, H6’), 7.58-7.60 (d, J = 8 Hz, 1H, H5’’), 7.67-7.71
(m, 2H, H3’, H5’), 8.10-8.12 (d, J = 8 Hz, 1H, H6’’), 8.19-8.31 (m, 1H, H2”), 8.99 (s, 1H, NH,
exchangeable by D2O), 9.37 (s, 1H, NH, exchangeable by D2O) ), 9.44 (s, 1H, NH, exchangeable by
D2O), 9.64 (s, 1H, NH, exchangeable by D2O). Ms: (Mwt.: 329): m/z, 331 [M++2, (30%)], 329 [M+,
(100%)], 108 (77%), 80 (40%), 53 (24%).
5.1.2.4. 1-(4-Aminophenyl)-3-(3-Methoxyphenyl)urea (2d)
Yield 70% as white crystals, m.p 165-168 °C, (as reported) [10].
5.1.2.5. 1-(4-Aminophenyl)-3-(3-(trifluoromethyl)phenyl)urea (2e)
1
Yield 80% as grayish white crystals, m.p 147-149 °C. H NMR (400 MHz, DMSO-d6) δ 4.80 (s,
2H, NH, exchangeable by D2O), 6.51-6.53 (d, J = 8 Hz, 2H, H2’, H6’), 7.07-7.1 (d, J = 12 Hz, 2H,
H3’, H5’), 7.25-7.27 (d, J = 8 Hz, 1H, H4’’), 7.46-7.54 (m, 1H, H5’’), 7.48-7.50 (d, J = 8 Hz, 1H, H6’’),
8.01 (s, 1H, H2”), 8.27 (s, 1H, NH, exchangeable by D2O), 8.88 (s, 1H, NH, exchangeable by D2O).
Ms: (Mwt.: 295): m/z, 296 [M++1, (18%)], 295 [M+, (93%)], 108 (100%), 80 (55%), 53 (22%).
5.1.3. 2-aminocyclopent-1-ene carbonitrile (3)
A solution of adipodinitrile (10.8 g, 100 mmol) in toluene (100 ml) is added dropwise to a slurry of
sodium hydride (60% suspension in oil, 4.20 g, 105 mmol) in anhydrous toluene (30 ml). The
mixture was heated under reflux for 15 hours. The product was extracted with ethyl acetate (3x50
ml). Then recrystallized by dissolving in hot toluene (25 ml) and then adding petroleum ether (150
ml). The product was obtained as yellowish white solid (6.60 g, 61%), m.p 148 °C [21].
5.1.4. 6,7-dihydro-3H-cyclopenta[d]pyrimidin-4(5H)-one (4)
Acetic anhydride (46 mL, 492.38 mmol: 23.8 equiv.) was added portion wise to a stirred formic
acid (23 mL, 614.92 mmol: 29.8 equiv.) at 0 °C over 30 minutes. This was followed by addition of
compound (3) (2 g, 20 mmol: 1 equiv.) and stirring under reflux at 130 °C for 48 hours. The solvent
was evaporated under vacuum and the resultant solid was washed with diethyl ether, dried and
recrystallized from methanol. The product was separated as buff crystals (1.2 g, 60%), m.p 220 °C
[26].
5.1.5. 4-chloro-6,7-dihydro-5H-cyclopenta[d]pyrimidine (5)
A mixture of compound (4) (2 g, 0.01 mmol: 1 equiv.) and phosphorous oxychloride (29 mL, 278
mmol: 18.2 equiv.) was heated under reflux for 3 hours. The mixture was then slowly poured on
ice/water, neutralized using ammonia solution (33%, 50 mL), then extracted with ethyl acetate
(2x50 mL). The combined organic layer was separated and the solvent was evaporated under
vacuum to afford brown solid that was washed with diethyl ether to give the titled compound (5) as
brown crystals (1.5 g, 76%) which was used directly in the next reaction [27].
5.1.6. General procedure for the preparation of compounds (6a–e)
10

The appropriate 4-aminophenyl substituted phenyl urea (2a-e) (1 mmol: 1 equiv.) and TEA (0.3
mL, 2 mmol: 2 equiv.) were added to a solution of the 4-chloro cyclopenta[d]pyrimidine derivative
5 (0.25 g, 1 mmol: 1 equiv.) in ethanol (15 mL). The mixture was heated under reflux for 18–24
hours and the reaction was followed up using TLC. The resultant solid was collected by filtration,
washed with hot ethanol, allowed to dry and recrystallized from ethanol to give the titled
compounds (6a-e).
5.1.6.1. 1-(4-(6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-ylamino)phenyl)-3-phenylurea (6a)
The product was separated as yellowish brown powder (0.2 g, 33%), m.p > 300 °C. 1H NMR (400
MHz, DMSO-d6) δ 1.16-1.20 (t, 2H, cyclopentenyl H), 1.94-1.99 (m, 1H, cyclopentenyl H), 2.63-
2.65 (m, 1H, cyclopentenyl H), 2.72-2.77 (m, 1H, cyclopentenyl H), 3.07-3.11 (m, 1H,
cyclopentenyl H), 6.51-6.53 (d, J = 8 Hz, 2H, H2’, H6’), 6.90-7.01 (m, 1H, H4”), 7.08-7.10 (d, J = 8
Hz, 2H, H3’, H5’), 7.22-7.47 (m, 4H, H2”, H3”, H5’’, H6” ), 8.32 (s, 1H, pyrimidine H), 8.41 (s, 1H,
NH, exchangeable by D2O), 8.66 (s, 1H, NH, exchangeable by D2O), 8.90 (s, 1H, NH,
exchangeable by D2O). HRMS (ESI-MS) calcd: 346.16339 for C20H19N5O [M + H+]. Found:
346.16672.
5.1.6.2.1-(3-chloro-4-methylphenyl)-3-(4-(6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-
ylamino)phenyl)urea (6b)
1
The product was separated as yellowish brown powder (0.102 g, 17%), m.p 233-235 °C. H NMR
(400 MHz, DMSO-d6) δ 1.09-1.13 (m, 1H, cyclopentenyl H), 1.86-1.91 (m, 1H, cyclopentenyl H),
2.04-2.12 (m, 1H, cyclopentenyl H), 2.26 (s, 3H, Ph-CH3), 2.76-2.83 (m, 3H, cyclopentenyl H),
6.55-6.56 (d, 2H, H2’, H6’), 7.00-7.58 (m, 5H, ArH), 8.37 (s, 1H, pyrimidine H), 8.68 (s, 1H, NH,
exchangeable by D2O), 8.70 (s, 1H, NH, exchangeable by D2O), 8.81 (s, 1H, NH, exchangeable by
D2O). HRMS (ESI-MS) calcd: 394.14333 for C21H20N5OCl [M + H+]. Found: 394.14331.
5.1.6.3.1-(4-chloro-3-(trifluoromethyl)phenyl)-3-(4-(6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-
ylamino)phenyl)urea (6c)
The product was separated as yellow powder (0.11 g, 19%), m.p 288-290 °C. 1H NMR (400 MHz,
DMSO-d6) δ 1.22 (s, 1H, cyclopentenyl H), 1.98-2.06 (m, 2H, cyclopentenyl H), 2.76-2.79 (m, 3H,
cyclopentenyl H), 7.29 (s, 1H, ArH), 7.40-7.42 (d, J = 8 Hz, 1H, H5’’), 7.57-7.61 (m, 4H, ArH),
7.64-7.66 (d, J = 8 Hz, 1H, H6”), 8.15 (s, 1H, pyrimidine H), 8.71 (s, 1H, NH, exchangeable by
D2O), 9.37 (s, 1H, NH, exchangeable by D2O), 9.94 (s, 1H, NH, exchangeable by D2O). HRMS
(ESI-MS) calcd: 448.11507 for C21H17N5OClF3 [M + H+]. Found: 448.11562.
5.1.6.4.1-(4-(6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-ylamino)phenyl)-3-(3-methoxyphenyl)urea
(6d)
The product was separated as grayish white powder (0.13 g, 22%), m.p >300 °C. 1H NMR (400
MHz, DMSO-d6) δ 1.24 (s, 4H, cyclopentenyl H), 1.91 (s, 2H, cyclopentenyl H), 3.50 (S, 3H,
OCH3, below water of DMSO), 6.86-6.88 (d, J = 8 Hz, 1H, H2’), 6.96-6.97 (d, J = 4 Hz, 1H, H6’),
7.04-7.06 (d, J = 8 Hz, 1H, H4’’), 7.21-7.23 (d, J = 8Hz, 1H, H3’), 7.47-7.50 (d, J = 12Hz, 1H, H6’’),
7.62-7.65 (d, J = 12Hz, 1H, H5’’), 7.82-7.86 (d, 1H, H2’’), 7.93-7.94 (d, J = 4Hz, 1H, H5’), 8.15 (s,
1H, pyrimidine H), 8.65 (s, 1H, NH, exchangeable by D2O), 8.67 (s, 1H, NH, exchangeable by
D2O), 10.56 (s, 1H, NH, exchangeable by D2O); HRMS (ESI-MS) calcd: 376.17721 for
C21H22N5O2 [M + H+]. Found: 376.17709.
11

5.1.6.5.1-(4-(6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-ylamino)phenyl)-3-(3-
(trifluoromethyl)phenyl)urea (6e)
The product was separated as grayish white powder (0.13 g, 22%), m.p 283-285 °C. 1H NMR (400
MHz, DMSO-d6) δ 1.14 (s, 4H, cyclopentenyl H), 1.24 (s, 2H, cyclopentenyl H), 6.50-6.52 (d, J = 8
Hz, 2H, H2’, H6’), 7.08-7.10 (d, J = 8Hz, 2H, H4’’, H5’’), 7.22-7.24 (d, J = 8Hz, 2H, H3’, H5’), 7.46-
7.48 (d, J = 8 Hz, 1H, H6’’), 7.57 (s, 1H, H2’’), 7.99 (s, 1H, pyrimidine H), 9.01 (s, 1H, NH,
exchangeable by D2O), 8.84 (s, 1H, NH, exchangeable by D2O), 10.56 (s, 1H, NH, exchangeable by
D2O). HRMS (ESI-MS) calcd: 414.15404 for C21H19N5OF3 [M + H+]. Found: 294.08449 and
370.08269.
5.2. Biology
The structures of the synthesized compounds were submitted to BPS Bioscience, San Diego,
CA, USA (www.bpsbioscience.com) for a VEGFR 2 tyrosine kinase assay. The assay was carried
out using a Kinase-Glo Plus Luminescence Kinase Assay Kit (Promega). Kinase activity is
measured by quantitating the amount of ATP remaining in the solution after a kinase reaction. The
luminescent signal from the assay is correlated with the amount of ATP present and is inversely
correlated with the amount of kinase activity. Compounds 6a, 6b, 6c, 6d, and 6e were diluted in
10% DMSO, and 5 µl of the dilution was added to a 50 µl reaction so that the final concentration of
DMSO was 1% in all reactions.
All of the enzymatic reactions were conducted at 30 °C for 45 minutes. The 50 µl reaction
mixture contained 40 mM Tris, pH 7.4, 10 mM MgCl2, 0.1 mg/ml BSA, 1 mM DTT, 0.2 mg/ml
Poly (Glu, Tyr) substrate, 10 µM ATP and enzyme. After the enzymatic reaction, 50 µl of Kinase-
Glo Plus Luminescence Kinase Assay Solution (Promega) was added to each reaction, and the plate
was incubated for 15 minutes at room temperature. The luminescence signal was measured using a
BioTek Synergy 2 microplate reader.
References
[1]
[2]
[3]
C.S. Abhinand, R. Raju, S.J. Soumya, P.S. Arya, P.R. Sudhakaran, VEGF-A/VEGFR2 signaling
network in endothelial cells relevant to angiogenesis, J. Cell Commun. Signal. 10 (2016) 347–354.
F. Musumeci, M. Radi, C. Brullo, S. Schenone, Vascular Endothelial Growth Factor (VEGF)
Receptors: Drugs and New Inhibitors, J. Med. Chem. 55 (2012) 10797–10822.
V.A. Machado, D. Peixoto, R. Costa, H.J.C. Froufe, R.C. Calhelha, R.M. V Abreu, I.C.F.R.
Ferreira, R. Soares, M.R.P. Queiroz, Synthesis , antiangiogenesis evaluation and molecular docking
studies of 1-aryl-3- [( thieno [ 3 , 2- b ] pyridin-7-ylthio ) phenyl ] ureasꢀ: Discovery of a new
substitution pattern for type II VEGFR-2 Tyr kinase inhibitor, Bioorg. Med. Chem. 23 (2015)
6497–6509.
[4]
[5]
J.E. Daniel S. La, Julie Belzile, James V. Bready, Angela Coxon, Thomas DeMelfi, Nicholas Doerr,
A.M.L. Julie C. Flynn, Shaun R. Flynn, Russell F. Graceffa, Shawn P. Harriman, Jay F. Larrow,
M.M.W. Matthew W. Martin, Michael J. Morrison, Vinod F. Patel, Philip M. Roveto, Ling Wang,
J.-C.H. Douglas A. Whittington, Yohannes Teffera, Zhiyang Zhao, Anthony J. Polverino, Novel
2,3-Dihydro-1,4-Benzoxazines as Potent and Orally Bioavailable Inhibitors of Tumor-Driven
Angiogenesis, J. Med. Chem. 51 (2008) 1695–1705.
M. Rajagopalan, S. Balasubramanian, A. Ramaswamy, P. Prakash Mathur, Pharmacophore based
3D-QSAR modeling and free energy analysis of VEGFR-2 inhibitors, J. Enzyme Inhib. Med.
12

Chem. 28 (2013) 1236–1246.
[6]
[7]
[8]
[9]
N.M.Y. Elsayed, R.A.T. Serya, M.F. Tolba, M. Ahmed, K. Barakat, D.A. Abou El Ella, K.A.M.
Abouzid, Design, synthesis, biological evaluation and dynamics simulation of indazole derivatives
with antiangiogenic and antiproliferative anticancer activity, Bioorg. Chem. 82 (2019) 340–359.
M.I. Shahin, D.A. Abou, E. Ella, N.S.M. Ismail, K.A.M. Abouzid, Design , synthesis and biological
evaluation of type-II VEGFR-2 inhibitors based on quinoxaline scaffold, Bioorg. Chem. 56 (2014)
16–26.
K. Okamoto, M. Ikemori-Kawada, A. Jestel, K. Von König, Y. Funahashi, T. Matsushima, A.
Tsuruoka, A. Inoue, J. Matsui, Distinct binding mode of multikinase inhibitor lenvatinib revealed
by biochemical characterization, ACS Med. Chem. Lett. 6 (2015) 89–94.
V.E. Kwitkowski, T.M. Prowell, A. Ibrahim, A.T. Farrell, R. Justice, S.S. Mitchell, R. Sridhara, R.
Pazdur, FDA Approval Summary: Temsirolimus as Treatment for Advanced Renal Cell Carcinoma,
Oncologist. 15 (2010) 428–435.
[10] M.A. Aziz, R.A.T. Serya, D.S. Lasheen, A.K. Abdel-az, A. Esmat, A.M. Mansour, A.N.B. Singab,
K.A.M. Abouzid, Discovery of Potent VEGFR-2 Inhibitors based on Furopyrimidine and
Thienopyrimidne Scaffolds as Cancer Targeting Agents, Sci. Rep. 6 (2016) 24460.
[11] Y. Oguro, N. Miyamoto, K. Okada, T. Takagi, H. Iwata, Y. Awazu, H. Miki, A. Hori, K.
Kamiyama, S. Imamura, Design , synthesis , and evaluation of 5-methyl-4-phenoxy-5 H -pyrrolo- [
3 , 2- d ] pyrimidine derivativesꢀ: Novel VEGFR2 kinase inhibitors binding to inactive kinase
conformation, Bioorg. Med. Chem. 18 (2010) 7260–7273.
[12] W.Y.C. Chen, P.Y. Chen, C.Y.C. Chen, J.G. Chung, Exploring 3D-QSAR Pharmacophore Mapping
of Azaphenanthrenone derivatives for mPGES-1 inhibition Using HypoGen Technique, in: 2008
IEEE Symp. Comput. Intell. Bioinforma. Comput. Biol., 2008: pp. 207–213.
[13] L. H, S. J, H. R, HypoGen: An Automated System for Generating Predictive 3D Pharmacophore
Models, Pharmacophore Perception, Dev. Use Drug Des. 2 (2000) 171–187.
[14] S. Sakkiah, S. Thangapandian, S. John, Y. Jung, K. Woo, 3D QSAR pharmacophore based virtual
screening and molecular docking for identification of potential HSP90 inhibitors, Eur. J. Med.
Chem. 45 (2010) 2132–2140.
[15] K. Sanphanya, S. Phowichit, S.K. Wattanapitayakul, V. V. Fokin, O. Vajragupta, Novel VEGFR-2
kinase inhibitor identified by the back-to-front approach, Bioorg Med Chem Lett. 23 (2013) 2962–
2967.
[16] A. Ghith, K.M. Youssef, N.S.M. Ismail, K.A.M. Abouzid, Design, synthesis and molecular
modeling study of certain VEGFR-2 inhibitors based on thienopyrimidne scaffold as cancer
targeting agents, Bioorg. Chem. 83 (2019) 111–128.
[17] J. Dietrich, C. Hulme, L.H. Hurley, The design, synthesis, and evaluation of 8 hybrid DFG-out
allosteric kinase inhibitors: A structural analysis of the binding interactions of Gleevec ®,
Nexavar®, and BIRB-796, Bioorganic Med. Chem. 18 (2010) 5738–5748.
[18] E. V Schneider, J. Böttcher, R. Huber, K. Maskos, L. Neumann, Structure – kinetic relationship
study of CDK8 / CycC specific compounds, PNAS. 110 (2013) 8081–8086.
[19] Y. Dai, K. Hartandi, Z. Ji, A.A. Ahmed, D.H. Albert, J.L. Bauch, J.J. Bouska, P.F. Bousquet, G.A.
Cunha, K.B. Glaser, C.M. Harris, D. Hickman, J. Guo, J. Li, P.A. Marcotte, K.C. Marsh, M.D.
Moskey, R.L. Martin, A.M. Olson, D.J. Osterling, L.J. Pease, N.B. Soni, K.D. Stewart, V.S. Stoll,
P. Tapang, D.R. Reuter, S.K. Davidsen, M.R. Michaelides, Discovery of N - ( 4- ( 3-Amino-1 H -
indazol-4-yl ) phenyl ) - N ′ - ( 2-fluoro-5-methylphenyl ) urea ( ABT-869 ), a 3-Aminoindazole-
Based Orally Active Multitargeted Receptor Tyrosine Kinase Inhibitor, J. Med. Chem. 50 (2007)
1584–1597.
[20] T. Suzuki, M.N.A. Khan, H. Sawada, E. Imai, Y. Itoh, K. Yamatsuta, N. Tokuda, J. Takeuchi, T.
Seko, H. Nakagawa, N. Miyata, Design, Synthesis, and Biological Activity of a Novel Series of
13

Human Sirtuin-2-Selective Inhibitors, J. Med. Chem. 55 (2012) 5760–5773.
[21] Y. Nurlela, A.J. Minnaard, S. Achmad, D. Wahyuningrum, A Long Chain Alcohol as Support in
Solid Phase Organic Synthesis, Sains Malaysiana. 40 (2011) 977–983.
[22] M.S. Coumar, M.T. Tsai, C.Y. Chu, B.J. Uang, W.H. Lin, C.Y. Chang, T.Y. Chang, J.S. Leou, C.H.
Teng, J.S. Wu, M.Y. Fang, C.H. Chen, J.T.A. Hsu, S.Y. Wu, Y.S. Chao, H.P. Hsieh, Identification,
SAR studies, and X-ray Co-crystallographic analysis of a novel furanopyrimidine aurora kinase a
inhibitor, ChemMedChem. 5 (2010) 255–267.
[23] J.I. Pyo, S.H. Lee, C.S. Cheong, A facile synthesis of some substituted furopyrimidine derivatives,
J. Heterocycl. Chem. 43 (2006) 1129–1133.
[24] L.S. Reddy, S.K. Chandran, S. George, N.J. Babu, A. Nangia, Crystal Structures of N-Aryl-N′-4-
Nitrophenyl Ureas: Molecular Conformation and Weak Interactions Direct the Strong Hydrogen
Bond Synthon, Cryst. Growth Des. 7 (2007) 2675–2690.
[25] M. Uno, H.S. Ban, W. Nabeyama, H. Nakamura, De novo Design and synthesis of N-benzylanilines
as new candidates for VEGFR tyrosine kinase inhibitors, Org. Biomol. Chem. 6 (2008) 979–981.
[26] A. Gangjee, Y. Zhao, S. Raghavan, C.C. Rohena, S.L. Mooberry, E. Hamel, Structure–Activity
Relationship and in Vitro and in Vivo Evaluation of the Potent Cytotoxic Anti-microtubule Agent
N-(4-Methoxyphenyl)-N,2,6-trimethyl-6,7-dihydro-5H-cyclopenta[d]pyrimidin-4-aminium
Chloride and Its Analogues As Antitumor Agents, J. Med. Chem. 56 (2013) 6829–6844.
[27] M. Adel, R.A.T. Serya, D.S. Lasheen, K.A.M. Abouzid, Identification of new pyrrolo[2,3-
d]pyrimidines as potent VEGFR-2 tyrosine kinase inhibitors: Design, synthesis, biological
evaluation and molecular modeling, Bioorg. Chem. 81 (2018) 612–629.
Fig. 1. 3D image of the superimposition of the re-docked conformer of sorafenib over the co-crystalized conformer (colored
yellow) with RMSD value of 0.257. (For interpretation using colors, the reader is referred to the web version of this article).
Fig. 2. [A] The reported 2D image of binding interactions of sorafenib in the x-ray crystal structure of VEGFR 2 with PDB code
3WZE [8], [B] 3D image of sorafenib inside 3WZE using Accelrys Discovery Studio 4.5; docking score is -48.92 along with the
hydrogen and hydrophobic interaction with the key residues of the receptor. (For interpretation using colors, the reader is referred
to the web version of this article).
Fig. 3. The library of T1-15 training set and S15-25 test set of known active VEGFR 2 inhibitors used for pharmacophore model
generation along with their IC50 values (µM) indicated in parenthesis.
Fig. 4. The best generated Pharmacophore model with three features: hydrogen bond acceptor (HBA) colored green, ring aromatic
(RA) colored orange and hydrophobic (HYP) colored cyan. With their inter-feature distances in angstroms and angles displayed.
Fig. 5. Sorafenib as reference compound mapped into the best pharmacophore model with fit value of 6.19.
Fig. 6. Illustration of binding of sorafenib and the proposed scaffold (I) to the binding regions of VEGFR 2.
Fig. 7. 2D image of hydrogen bonds and hydrophobic interactions of sorafenib and proposed derivatives 6a and 6d with the kinase
domain of VEGFR 2.
Scheme 1. Synthesis of 4-aminophenyl substituted phenyl ureas reagents and conditions: (a) Phenyl isocyanates, DCM, rt, 48 hrs;
(b) H2, Pd/C, MeOH, rt, 4 hrs.
Scheme 2. Synthesis of 6,7-dihydro-5H-cyclopenta[d]pyrimidine derivatives reagents and conditions: (a) NaH, toluene, 15 hrs;
(b) HCOOH, acetic anhydride, reflux, 48 hrs; (c) POCl3, reflux, 3 hrs; (d) Ethanol, TEA, reflux, 18-48 hrs.
14

Table 1. Tabular column illustrating the details of the ten hypotheses generated using hypoGen algorithm.
Cost
difference
Correlation
coefficient
Maximum
fit
Hypothesis
Total cost
RMS
Featuresb
1
2
59.35
66.56
74.56
74.86
74.95
75.82
76.17
76.54
78.23
78.97
140.98
133.78
125.78
125.48
125.39
124.52
124.16
123.80
122.10
121.37
1.03
1.42
1.76
1.77
1.76
1.74
1.78
1.80
1.89
1.89
0.97
0.95
0.92
0.92
0.92
0.92
0.92
0.92
0.91
0.91
HBA, HYP, HYP, HYP, RA
HBA, HYP, HYP, HYP
HBD, HYP, HYP, HYP
HBD, HYP, HYP, HYP
HYP, HYP, HYP, HYP, HYP
HBA, HBA, HYP, HYP, HYP
HBA, HBD, HYP, RA
10.80
7.36
7.81
7.90
8.83
7.08
6.23
6.39
7.34
6.41
3
4
5
6
7
8
HBD, HYP, HYP, HYP
HBA, HYP, HYP, RA
9
10
HBD, HYP, RA, RA
babbreviations used for features; hydrogen bond acceptor (HBA), hydrophobic (HYP) and ring aromatic (RA).
Table 2. The experimental and predicted activity values of the training set compounds according to hypothesis 1 along with their
fit values.
Training set
Experimental activity IC50
nM
Predicted activity IC50 nM
Fit value
T1
T2
T3
T4
T5
T6
T7
940
870
340
4.1
32
688.787
1,299
6.505
6.230
6.800
8.397
8.206
8.596
8.389
349.804
8.839
13.729
5.586
5.3
4.4
8.998
15

T8
19
2.7
33
12.989
3.086
8.230
8.854
7.823
7.287
8.478
8.176
8.861
8.107
T9
T10
T11
T12
T13
T14
T15
33.111
113.757
7.335
110
7.1
30
14.713
3.033
3.7
14
17.234
Table 3. Docking score, hydrogen and hydrophobic interactions of sorafenib and the three top hits with VEGFR 2 (pdb code
3WZE), along with their pharmacophore mapping and fit values inside the best-generated pharmacophore model.
Binding
interactions
CDOCKER
energy
Pharmacophore
mapping
Fit
value
Compound
Sorafenib
H
N
H
N
CF3
Cl
O
O
O
N
N
H
-48.92
6.19
KM10259
H
H
O
N
N
S
O
Cl
-42.28
7.16
16

Br
H
H
CDI665171
O
N
N
N
O
S
O
NH
-33.81
7.80
2r4b_GW7
F
Cl
O
HN
N
S
N
-35.08
7.10
Table 4. Docking results of sorafenib and the three top hits KM10259, CDI665171 and 2r4b_GW7 in the active site of
VEGFR 2 (3WZE) showing hydrogen bonds and hydrophobic interactions.
fx1
Table 5. Docking score, hydrogen bonds and hydrophobic interactions of sorafenib and the proposed derivatives with VEGFR 2
(pdb code 3WZE), along with their pharmacophore mapping and fit values inside the best generated pharmacophore model.
Compound
CDOCKER
energy
Pharmacophore
mapping
Fit
value
interactions
17

H
N
H
Sorafenib
N
CF3
Cl
O
O
O
N
N
H
-48.92
6.19
NH
NH
C
O
6a
HN
N
N
-29.49
9.05
6b
NH
NH
Cl
C
O
HN
N
N
-34.28
6.63
18

NH
NH
CF3
Cl
C
O
6c
HN
N
N
-29.29
9.21
6d
NH
NH
OCH3
C
O
HN
N
N
-31.94
6.63
19

6e
NH
NH
CF3
C
O
HN
N
N
-34.00
6.7
Table 6. Docking results of sorafenib and the three top hits KM10259, CDI665171 and 2r4b_GW7 in the active site of
VEGFR 2 (3WZE) showing hydrogen bonds and hydrophobic interactions.
fx2
Table 7. VEGFR 2 enzyme inhibition of the synthesized compounds at 10µM.
Compounds
R
% inhibition
6a
6b
6c
H
5
87
97
0
3-Cl-4-CH3
4-Cl-3-CF3
3-OCH3
3- CF3
6d*
6e*
1
*these compounds are sparingly dissolved at 10mM in 100% DMSO.
Table 8. The IC50 value of Synthesized compounds with high VEGFR 2 percent inhibition.
Compounds
6b
VEGFR 2 IC50 (µM)
2.26
0.85
0.09
0.09
6c
Staurosporine
Sorafenib
20