3D-QSAR and molecular docking studies on HIV protease inhibitors [631780]

3D-QSAR and molecular docking studies on HIV protease inhibitors
Jianbo Tong*, Yingji Wu, Min Bai, Pei Zhan
College of Chemistry and Chemical Engineering, Shaanxi University of Science &Technology, Xi'an, 710021, PR China
article info
Article history:
Received 13 July 2016Received in revised form17 September 2016
Accepted 19 September 2016
Available online 20 September 2016
Keywords:
Cyclic-urea derivatives
3D-QSARCoMSIAMolecular dockingabstract
In order to well understand the chemical-biological interactions governing their activities toward HIV
protease activity, QSAR models of 34 cyclic-urea derivatives with inhibitory HIV were developed. The
quantitative structure activity relationship (QSAR) model was built by using comparative molecularsimilarity indices analysis (CoMSIA) technique. And the best CoMSIA model has r
cv2,rncv2values of 0.586
and 0.931 for cross-validated and non-cross-validated. The predictive ability of CoMSIA model was
further validated by a test set of 7 compounds, giving rpred2value of 0.973. Docking studies were used to
find the actual conformations of chemicals in active site of HIV protease, as well as the binding mode
pattern to the binding site in protease enzyme. The information provided by 3D-QSAR model and mo-
lecular docking may lead to a better understanding of the structural requirements of 34 cyclic-urea
derivatives and help to design potential anti-HIV protease molecules.
©2016 Elsevier B.V. All rights reserved.
1. Introduction
Acquired immunode ficiency syndrome (AIDS) is one of leading
causes of death worldwide, which may be induced by Human im-
munode ficiency virus (HIV) retrovirus [1]and progressive damage
the immune system manifesting in serious opportunistic diseases.The millions people living with HIV, and it's high actively spreading,
fatality rate brings large economic and social impacts [2]. Many
drugs are getting ineffective due to resistance from the mutation-
prone HIV. Hence, there is an urgent need to develop new drugs
with better therapeutic effect [3]. The HIV protease receptor (HIV
PR) is a key therapeutic target for the development of anti-HIV
inhibitors [4]for the treatment of AIDS as it plays an important
role in the maturation and replication of the virus. HIV PR is a small
enzyme; acting as a dimer of two 99-residue subunits; and is
tractable for structural and computational analyses [5].
Anti-AIDS drugs are classi fied into three main categories
including nucleoside reverse transcriptase inhibitors, non-
nucleoside reverse transcriptase inhibitors and protease in-
hibitors [6]. HIV protease is an important target in anti-HIV drug
therapy [7]. Based on present many crystallographic [8]and ener-
getic studies [9], HIV protease has become an attractive target for
the computer-aided drug design [10]. Quantitative structure ac-
tivity relationships (QSARs) is an universal method forchemometrics, which reveals the relationship between chemical
structures and their biological activities. As a supportive method for
drug design and prediction of drug activity, molecular docking is
also applied to understand the structural interaction between drug
and receptor [11].
Herein, the 3D-QSAR was performed by using the comparative
molecular similarity indices analysis (CoMSIA) technique [12].I n
present studies, CoMSIA was applied to investigating the relation-
ship between structure and activity on inhibiting HIV protease of 34
cyclic-urea derivatives [13]. CoMSIA-based 3D-QSAR study has
been carried out on cyclic-urea derivatives to determine the in-
fluence of steric, electrostatic, and hydrophobic fields of these
compounds on their HIV inhibitory activity. Furthermore, these
fields were mapped onto the inhibitors binding pocket of HIV
protease for the better understanding of these interactions. Seven
analogs of cyclic-urea derivatives were designed and synthesized as
potential HIV protease inhibitors, which is based on molecular
modeling with higher biological activity. AutoDock comparison
between the new designed compound N5 and the template com-
pound con firms the H-Bond between OH of sp
3hybridization of the
ligand and COOH group of ASP25 is crucial and bene fit higher
biological activity.
2. Principle and methodology
2.1. Database and biological activity
The set of 34 cyclic-urea derivatives, inhibiting HIV protease was*Corresponding author.
E-mail address: jianbotong@aliyun.com (J. Tong).
Contents lists available at ScienceDirect
Journal of Molecular Structure
journal homepage: http://www.elsevier.com/locate/molstruc
http://dx.doi.org/10.1016/j.molstruc.2016.09.052
0022-2860/ ©2016 Elsevier B.V. All rights reserved.Journal of Molecular Structure 1129 (2017) 17 e22

compiled from the literature [13]. The Ki(nM) values of the HIV
protease inhibitor activity data of cyclic-urea derivatives were
converted into the negative logarithmic scale (log(1/ Ki)). That is, p Ki
values were used for calculation in this paper (p Ki¼- log Ki¼log(1/
Ki)). Based on the molecules structural diverse and activities, the set
was divided into the training set containing 27 compounds and the
test set containing 7 compounds. The experimental activity was
stemmed from the literature [13], while the predicted activity was
calculated with CoMSIA model for both training and test sets of
compounds. The CoMSIA model was developed using five
descriptor fields: steric, electrostatic, hydrophobic, and hydrogen-
bond acceptor and donor fields, the structures and activities of 34
cyclic-urea derivatives are listed in Table 1 .
2.2. Molecular modeling and alignment
Molecular alignment is the most important part of the 3D-QSAR
studies [14]. Molecular modeling calculations were performed by
SYBYL 8.1 [15] which was optimized using the Tripos Force Field
and Gasteiger-Hückel charges [16] with an energy charge of
0.001 kcal/mol, and the maximum iteration coef ficient of 1000. All
the 34 compounds have the same skeleton is shown in Fig. 1 . Then
the heptanuclear heterocyclic, a common substructure in the
molecules database, was used as template to align the compounds
set. The alignment model was used for CoMSIA studies. It is
assumed that they have the same action mechanism and binding
mode. These structures of 34 compounds were all subjected to
conformational search and have minimal energy. The compound 17was used as reference for the alignment of both the training set and
test set ( Fig. 2 ), because it is the highest biological activity for the
training set. So it is also considered that the molecular structure of
compound 17 is more stable.
2.3. CoMSIA methodology
To generate the CoMSIA descriptor fields, a 3D cubic lattice with
grid spacing of 2 Å in x, y and z directions, was created to encom-
pass the aligned molecules. CoMSIA descriptors were calculated
using the sp3carbon probe atom with a Van der Waals
radius ¼1.52 Å and a charge ¼ț1.0 to generate steric field energies
and electrostatic (Coulombic potential) fields with a distance-
dependent dielectric at each lattice point. All five force field prop-
erties (steric, electrostatic, hydrophobic, hydrogen-bond donor and
hydrogen-bond acceptor) of CoMSIA were determined at a 30-kcal/
mol energy cut-off, which means that energy fields greater than
30 kcal/mol are curtailed to that value, and thus can avoid in finite
energy values inside the molecule. The CoMSIA method was eval-
uated with the help of the probe atom [17]. Similarity indices ( AF,k)
are calculated at regularly spaced grid points for the pre-aligned
molecular by Equation (1).
Aq
F;kðjȚ¼X
iwprobe ;kwike/C0ar2
iq (1)
In the above equation, wprobe.k (radius ¼1 Å, charge ¼ț 1,
hydrophobicity ¼ț 1, hydrogen bond donating ¼ț 1, hydrogen
bond accepting ¼ț1) for the probe atom was placed at each grid
point to calculate the electrostatic, steric, hydrophobic, H-bondTable 1
Structures and predicted activities of 34 cyclic-urea derivatives.
No. R/R0Exp.(p Ki)bPred.(p Ki)c
1C H 2C6H5 8.47 8.42
2 Me 5.30 5.303C H
2C6H4-4-CHMe 2 8.96 8.97
4aCH2CHMe 2 5.77 5.77
5 CH(Me)SMe 5.96 5.906C H
2-3-indolyl 6.24 6.22
7C H 2-Cy-C 6H11 7.55 7.57
8C H 2CH2C6H5 6.50 6.48
9C H 2-2-naphthyl 8.01 7.99
10aCH2-3-furanyl 8.08 8.07
11 CH 2C6H4-3-SMe 8.60 8.55
12 CH 2C6H4-4-SO 2Me 8.60 8.59
13aCH2C6H4-2-OMe 7.22 8.07
14 CH 2C6H4-2-OH 7.46 7.42
15 CH 2C6H4-3-OMe 8.33 8.32
16aCH2C6H4-4-OMe 8.07 8.06
17 CH 2C6H4-4-OH 8.96 8.97
18 CH 2C6H4-3-NH 2 8.55 8.50
19 CH 2C6H4-3-NMe 2 8.37 8.36
20 CH 2C6H4-4-NH 2 8.07 8.04
21 CH 2-4-pyridyl 7.66 7.73
22 3-(2,5-Me-pyrolyl)-CH 2C6H4 6.80 6.77
23 CH 2C6H5 8.72 8.64
24 CH 2CHMe 2 7.07 7.68
25aCHMe 2 6.60 6.61
26 CH(Me)SMe 5.60 5.6327 CH
2C6H4-4-F 8.24 8.22
28 CH 2C6H4-2-OMe 7.19 7.15
29aCH2C6H4-3-OMe 9.06 8.48
30 CH 2C6H4-3-OH 7.89 7.87
31 CH 2C6H4-4-OMe 8.54 8.55
32aCH2-naphthyl 8.37 8.06
33 CH 2C6H3-3,5-OMe 8.57 8.58
34 CH 2-2-thienyl 8.04 8.03
aTest set.
bExperimental activity.
cPredicted activity. Compounds 1-22 where ( P2/P20)¼benzyl, Compounds 23-34
where ( P2/P20)¼CH2-Cy-C3H5.
Fig. 1. Skeleton of 34 cyclic-urea derivatives.
Fig. 2. The alignment of all 34 compounds.J. Tong et al. / Journal of Molecular Structure 1129 (2017) 17 e22 18

donor and acceptor fields. The wikis the actual value of the physi-
cochemical property kof atom i. The riqis the mutual distance
between the probe atom at grid point qand atom iof the test
molecule. The ais attenuation factor with, default value of 0.3.
2.4. Partial least squares analysis
PLS method was used to linearly correlate the CoMSIA field to
the inhibitory activity values. For PLS models, CoMSIA descriptors
were used as independent variables and p kivalues were used as
dependent variables [18]. The cross-validated analysis was per-
formed using the leave-one-out (LOO) method in which one com-pound is removed and its activity is predicted using the model
derived from the rest of the dataset. So the optimal number of
components was determined by the Leave-One-Out (LOO) cross-
validation procedure. A final non-cross-validation analysis was
produced with the optimal number of components to obtain r
cv2,
which calculated by the following equation: 2. The predictive
ability of the 3D-QSAR model was determined from a test set
including 7 compounds ( Table 1 ). The criteria for the 3D-QSAR
model is that rcv2>0.5,rncv2>0.6. These molecules were aligned using
the same method as training set, and their activities were predicted
using the model generated by the training set. Some analogs with
high activity were used in the training and test set, the activity
values of which were also predicted. Finally new seven compounds
were designed and subsequently tested. In addition to the classical
test set, these new compounds were also included in the calcula-
tion of rpred2. Based on the test set molecules, the rpred2was calculated
by the following equation: 3.
r2
cv¼1/C0P
Y/C16
Ypred/C0Yactual/C172
P
YðYactual /C0YmeanȚ2(2)
Ypred,Yactual ,Ymean are predicted, experimental and mean value,
respectively.
r2
pred¼SD/C0PRESS
SD(3)
SD is the sum of the squared deviations between the biological
activity of the compounds in the test set and the mean biological
activity of the training set. PRESS is the sum of squared deviation
between predicted and actual activity values for each compound in
test set.
2.5. Molecular docking analysis
Molecular docking was performed by AutoDock 4.2 software
package [19]. The best junction mode between the receptor and the
ligand was found by using simulated annealing and genetic algo-
rithm. Then functional form (4) [20] was used to calculate bonding
free energy of the semi-empirical method to evaluate the binding
mode between receptor and the ligand.
DG¼DGvdwX
Aij
r12
ij/C0Bij
r6
ij!
țDGH/C0bondX
i;jEðtȚ
Cij
r10
ij!
țDGeleX
i;jqij
ε/C0rij/C1rijțDGtorNtorțDGsolX
i;j/C0SiVjțSjVi/C1e/C0
r2
ij/C14
2d2/C1
(4)
DGvdw,DGH-bond ,DGele,DGtorandDGsolare the semi-empiricalparameters obtained by fitting;
3. Results and discussion
3.1. CoMSIA analyses
Using different combinations of CoMSIA descriptor fields,
different CoMSIA models were developed. In this paper, a model
consisting of steric, electrostatic, hydrophobic, hydrogen-bond
acceptor and hydrogen-bond donor CoMSIA fields with signi fi-
cant r2ncv(0.931), r2cv(0.586), r2pred(0.973) and F (1009.45) was
selected for further analysis. The contributions of steric, electro-
static, hydrophobic, hydrogen-bond acceptor, hydrogen-bond
donor fields were 0.216, 0.310, 0.258, 0.085 and 0.131, respec-
tively. So the results of CoMSIA show that steric, electrostatic and
hydrophobic descriptor fields play a dominated role.
For this CoMSIA model, it has been found that the five different
descriptor fields are not totally independent to each other and the
dependencies of individual fields usually affect the statistical sig-
nificance of the models. All possible combinations of fields were
evaluated to determine the best predictive model. With the help of
a probe atom surrounding the molecules, molecular field analysis
finds the favourable or unfavourable interaction energies of aligned
molecules. These 3D color contour maps provide hints for the
modi fication required to design new molecules with improved
activity.
The above results suggest the model of CoMSIA-based 3D-QSAR
study is reasonable and also used to forecast the test set, to obtain
the predicted activities of training set and test set compounds. The
experimental and predicted p kivalues are displayed in Fig. 3 . It can
be seen that all points are located near the diagonal line and noobviously exceptional point was observed. The high r
ncv2,rcv2,rpred2
and F values indicated a good statistical correlation and reasonable
predictability of the CoMSIA model. The rpred2value of 0.973 and the
results of internal and external inspection showed that the model
was accurate and stable. So, the obtained QSAR model with good
exterior predictive capability was robust and can be used to design
and screen new compounds.
The best CoMSIA model was developed using five descriptor
fields: steric, electrostatic, hydrophobic, hydrogen-bond donor and
hydrogen-bond acceptor fields. The five descriptor fields contour
Fig. 3. Scatter plot between the experimental and predicted activity of 34 cyclic-urea
derivatives.J. Tong et al. / Journal of Molecular Structure 1129 (2017) 17 e22 19

maps of the highly active compound 17 are shown in Fig. 4 . The
green and yellow color contour maps show the favourable and
unfavourable steric interactions. The blue and red contour maps
indicate the favourable and unfavourable electrostatic interaction
with the molecules. The yellow and white contour maps indicate
the favourable and unfavourable hydrophobic interaction with the
molecules, cyan and blueviolet contour maps indicate the favour-
able and unfavourable hydrogen-bond donor interaction with the
molecules. The magenta and red contour maps indicate the
favourable and unfavourable hydrogen-bond acceptor interaction
with the molecules.
InFig. 4 a, green contours indicate regions where steric bulk
groups increase the activity, while yellow contours indicate regions
where steric bulk groups decrease the activity. There is one green
contour around Rposition, which can be explained by compound 3
and 2. Because of the different steric bulk groups it was found that
the activity values of compound 3 and 2 are higher than that of 17.
There is another green contour around R0position similar to the R
position. In this area bulk groups increase the activity, which can be
explained by compound 29 and 25. The yellow contour near P2/P20
indicate that the steric occupancy with more bulky groups in this
region will decrease the activity, which can be explained by com-
pounds 24 and 4.
For electrostatic map in Fig. 4 b, blue contours indicate regions
where electron positive groups increase activity, and red contours
indicate regions where electron negative groups increase activity.
There are blue contours near R0and P20, where the electron positive
groups occupation will increase the activity, while there are red
contours around Rand P2, where the electron negative groups
occupation will increase activity. It is shown that there is a large
blue contour in C 2- and C 3- benzyl substituent for P2orP20, where
electron positive groups would increase activity.
In hydrophobic map Fig. 4 c, yellow contours suggest that a hy-
drophobic substituent may favor activity, while white contours
disfavor its activity. There are two white contours in the R/R0sectionof the benzene ring, where hydrophobic groups would decrease
activity, while there is a yellow area in OH group of C 4- benzyl
substituent, where hydrophobic groups would increase activity.
For H-bond donor contour maps ( Fig. 4 d), there are two small
cyan contours around the OH groups of C 4-benzyl substituent that
hydrogen-bond donors in this region favor the activity, like com-
pound 16 and 17. However, for H-bond acceptor contour maps
(Fig. 4 e), there are two big magenta contours around the groups on
C1- and C 4- heptanuclear heterocyclic substituents it shows that
hydrogen-bond acceptors in this region favor the activity.
3.2. Drug molecular design
Based on above CoMSIA model, the structure-activity relation-
ships of cyclic-urea derivatives were obtained, that is the values of
pKihave a close relationship with the fields of steric, electrostatic,
hydrophobic. Compound 29 of the highest activity in the 34 cyclic-
urea derivatives was selected for optimization, some compounds
with higher activity based on compound 29 as template can be
designed. Seven newly designed molecules and their predictive
activity were shown in Table 2 .
Compared the seven newly designed compound with com-
pound 29, it shows that the substitution with -OCHMe 2, -OMe,
-C6H5, -CHMe 2groups for compounds 1 e4 generate better pre-
dicted activity, predicted activity of these new compounds are
lower than experimental data of compound 29, but it is higher than
that of the original compounds published in literature [13]. Because
the increased biological activity stems from of the increased steric
interaction on R/R0positions. It can also observed that the increased
electronegativity on R/P2positions contributed to higher biological
activity. It is shown that the new compound N2 possesses lower
biological activity compared to the original compound 27 and 30,
because the new compound N2 has lower electronegativity on R/P 2
positions. Therefore, according to the information provided by
CoMSIA, we can obtain higher activity drug molecules by changing
Fig. 4. CoMSIA contour maps for the highly active compound (17) of the training set. (For interpretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)J. Tong et al. / Journal of Molecular Structure 1129 (2017) 17 e22 20

the structure of the compounds.
3.3. Docking study
The docking studies were carried out to explore the interaction
mechanism between inhibitors and the receptor [21]. The infor-
mation of binding pocket of a receptor for its ligand is veryimportant for drug design. Docking was performed using AutoDock
4.2 software package. 3D co-crystallized structure of HIV protease
was taken from the Research Collaboratory for Structural Bioin-
formatics (RCSB) Protein Data Bank (PDB ID: 1AJV) [22]. For the
docking study, water molecules and original ligand were removed
from the protein. The polar hydrogens and united atom Kollman
charges were added for the protein so as to assign appropriate
ionization states to both acidic and basic amino acid residues. In
this experiment, the compound 29 template and N7 newly
designed compounds were studied by molecular docking. AutoGrid
was carried out for the preparation of the grid map using a grid box
with a npts (number of points in xyz) ¼(42, 40, 40) Å box, which
encloses the original ligand (NMB), and the box spacing was
0.375 Å. For compound 29 template and newly designed compound
N5, grid center was designed at dimensions (x, y, z): (11.955, 21.957,
4.646) and (12.055, 22.550, 4.992), respectively. The compounds
were constructed using ChemDraw 8.0 software, which were con-
verted to 3D structures viaing energetically minimized. Then the
compounds were saved as PDB file format. Lamarckian Genetic
Algorithm (LGA) was used as ligand conformation search process
and the other parameters were default. The interactions of complexHIV protein-ligand conformations, including hydrogen bonds and
the bond lengths were analyzed using PyMol.
As is shown in Fig. 5 , for compounds N1-7 the estimated binding
energy (
DG) decreases with the increase of predicted biological
activity, the values of DG were observed in Table 2 . The estimated
binding energy correlates with the predicted biological activity
values rendering a good monotonicity, thus demonstrating that the
molecular docking calculation has certain reliability.
It is generally believed that the small molecule drug contact
with active amino acid residues of target enzyme to inhibit the
activity of enzyme. Small molecules center for the active site was
calculate by docking. The docked conformations showed that all
compounds bind to the active residues in the prede fined hydro-
phobic binding pocket. As can be seen from Fig. 6 compound 29
forms hydrogen bonding (H-bond) with ASP25, and ILE50 in the
binding sites. The docked compound 29 was found to have H-bond
of 1.9 nm between OH of sp3hybridization of the ligand and C ]O
group of ASP25 and H-bond of 2.0 nm between OH of sp3hybrid-
ization of ligand and OH group of ASP25, as well as hydrogen bond
of 1.8 nm between C ]Oo fs p2hybridization of ligand and NH 2
group of ILE50. In the docked compound N5, it was found there are
H-Bond of 2.0 nm between OH of sp3hybridization of the ligandTable 2
Newly designed molecular structures and predictive activity.
No. R/R0Pred.(p Ki) DG
N1 CH 2C6H4-3-OCH(CH 3)2 8.551 /C010.05
N2 CH 2C6H3-3,4-OMe 7.934 /C09.12
N3 CH 2C6H4-3-C 6H5 8.085 /C010.59
N4 CH 2C6H4-3-CHMe 2 8.542 /C011.47
N5 CH 2C6H4-3-CHMe 2 8.963 ¡12.20
N6 CH 2C6H4-4-OCHMe 2 8.71 /C011.67
N7 CH 2C6H3-2,4-OH 6.956 /C010.47
Compounds N5 eN7 where ( P2/P20)¼benzyl, Compounds N1 eN4 where ( P2/
P20)¼CH2eCyeC3H5.
Fig. 5. The diagram between estimated energy of binding and biological activity.
Fig. 6. The docking interaction pattern of HIV protease (1AJV) active residues with ligands. (a)The interactions of compound 29 and HIV protease residues (b ) The interactions of N5
and HIV protease residues.J. Tong et al. / Journal of Molecular Structure 1129 (2017) 17 e22 21

and COOH group of ASP25. Although the docked compound N5 has
one H-bond, it possesses lower binding free energy comparing to
compound 29. The binding free energy of compounds N5 and 29
are/C012.20 kcal/mol, -10.63 kcal/mol, respectively. Thus, H-bond
interaction between OH of sp3hybridization of the ligand and
COOH group of ASP25 played a major role in the combination of
drugs and receptor. Due to the role of these residues, compound 29
showed a certain inhibitory activity of HIV protease. From molec-
ular docking model, it was found that in the docking process of
ligand and receptor, the formation of hydrogen bonds between the
compound and receptor determined the activity of inhibitor. The
positions of the active site has an important role [23]. As a common
bonding between drug molecules and biological macromolecules
receptor, hydrogen bonding makes the combination of drug mol-
ecules to the enzyme more stable and increase the drug activity
[24]. The docking results agreed well with the observed biological
activity datas, which showed that these docking conformations
were desirable to analyse the drug models.
4. Conclusions
In this paper, 3D-QSAR and molecular docking studies were
performed on a series of 34 HIV protease inhibiting drug of cyclic-
urea derivatives. The 3D-QSAR model was built by using CoMSIA
method. The obtained model was calculated with fivefields (steric,
electrostatic, hydrophobic, hydrogen donor and hydrogen acceptor
descriptors), which had the highest cross-validated rcv2(0.586) and
the non-cross-validated PLS analysis had a conventional rncv2(0.931)
and was considered as the best CoMSIA model. In addition, the 3D-
QSAR results suggested that the steric interaction on R/R0groups
and the electronegativity on R/P2groups may enhance the inhibi-
tory activity of enzyme. Molecular docking approach was employed
to study the relationship between drug ligands and macromolec-
ular receptor. The docking results indicated that the ligands would
form hydrogen bonding interactions with ILE50 and ASP25 of the
protein receptor. More importantly, the hydrogen bonding between
the OH of sp3hybridization of the ligand and COOH group of ASP25
of the protein receptor is important for its high activity. These re-
sults demonstrated the power of a combined docking/QSAR
approach to explore the probable binding conformations of com-
pounds at the active sites of the protein target, and can also help to
design and screen new compounds to obtain new HIV protease
inhibitors with high activities.
Acknowledgements
This work was supported by the National Natural Science Funds
of China (21475081) (21275094), the Natural Science Foundation of
Shaanxi Province of China (2015JM2057), and the Graduate Inno-
vation Fund of Shaanxi University of Science and Technology.
References
[1]H.F. Mohammad, H. Afsane, G. Sajjad, QSAR prediction of HIV-1 protease
inhibitory activities using docking derived molecular descriptors, J. Theor.Biol. 369 (2015) 13 e22.
[2]M.A. Islam, T.S. Pillay, Simpli fied molecular input line entry system-based
descriptors in QSAR modeling for HIV-protease inhibitors, Chemom. Intell.
Lab. 153 (2016) 67 e74.
[3]B.R. Meher, S. Patel, Chapter eight-structural a nd dynamical aspects of HIV-1
protease and its role in drug resistance, Adv. Protein Chem. Str. 92 (5) (2013)299e324.
[4]A. Wlodawer, M. Miller, M. Jaskolski, B.K. Sathyanarayana, E. Baldwin,
I.T. Weber, L.M. Selk, L. Clawson, J. Schneider, S.B. Kent, Conserved folding in
retroviral proteases: crystal structure of a synthetic HIV-1 protease, Science
245 (1989) 616 e621.
[5]M.C. Sharma, Structural requirements of N-aryl-oxazolidinone-5-carboxamide
derivatives for anti-HIV protease activity using molecular modelling tech-niques, J. Taibah Univ. Sci. 8 (2) (2013) 111 e123.
[6]P.J. Ala, E.E. Huston, R.M. Klabe, P.K. Jadhav, P.Y. Lam, C.H. Chang, Counter-
acting HIV-1 protease drug resistance: structural analysis of mutant proteases
complexed with XV638 and SD146, cyclic urea amides with broad speci fic-
ities, Biochemistry 37 (43) (1998) 15042 e15049 .
[7]A. Wlodawer, J. Vondrasek, Inhibitors of HIV-1 protease: a major success of
structure-assisted drug design, Annu. Rev. Bioph. Biomol. 27 (1) (1998)
249e284.
[8]S. Bihani, A. Das, V. Prashar, J.L. Ferrer, M. Hosur, X-ray structure of HIV-1
protease in situ product complex &dagger, U. S. Natl. Libr. Med. Institutes
Health 74 (3) (2009) 594 e602.
[9]C.H. Shen, Y.C. Chang, J. Agniswamy, R.W. Harrisona, I.T. Webera, Conforma-
tional variation of an extreme drug resistant mutant of HIV protease, J. Mol.
Graph. Model 62 (2015) 87 e96.
[10] G. Klebe, Recent developments in structure-based drug design, J. Mol. Med. 78
(5) (2000) 269 e281.
[11] A. Maryam, F. Ramezani, M. Elyasi, H.S. Aliabadi, M. Amanlou, A study on
quantitative structure-activity relationship and molecular docking of metal-
loproteinase inhibitors based on L-tyrosine scaffold, DARU J. Pharm. Sci. 23 (1)
(2015) 1 e10.
[12] G. Klebe, U. Abraham, Comparative Molecular Similarity Index Analysis
(CoMSIA) to study hydrogen-bonding properties and to score combinatoriallibraries, J. Comput. Aid. Mol. Des. 13 (1999) 1 e10.
[13] R. Garg, S.P. Gupta, H. Gao, M.S. Babu, A.K. Debnath, C. Hansch, Comparative
quantitative structure-activity relationship studies on anti-HIV, Drugs. Chem.
Rev. 99 (1999) 3525 e3601 .
[14] M.V. Damre1, R.P. Gangwal1, G.V. Dhoke, M. Lalit, D. Sharma, K. Khandelwal,
A.T. Sangamwar, 3D-QSAR and molecular docking studies of amino-
pyrimidine derivatives as PknB inhibitors, J. Taiwan Inst. Chem. E 45 (2)(2014) 354 e
364.
[15] SYBYL 8.1, Tripos 1699 South Hanley Road, St. Louis, Missouri, 63144.
[16] J. Gasteiger, M. Marsill, Iterative partial equalization of orbital
electronegativith-a rapid access to atomic charges, Tetrahedron 36 (1980)
3219e3228 .
[17] V. Murugesan, Y.S. Prabhakar, S.B. Katti, CoMFA and CoMSI A studies on
thiazolidi n-4-one as anti-HIV-1 agents, J. Mol. Graph. Model 27 (2009)
735e743.
[18] S. Patel, B. Patel, H. Bhatt, 3D-QSAR studies on 5-hydroxy-6-oxo-1,6-
dihydropyrimidine-4- carboxamide derivatives as HIV-1 integrase in-
hibitors, J. Taiwan Inst. Chem. E 59 (2016) 61 e68.
[19] M.F. Sanners, R. Huey, S. Dallakyan, et al., Autodock4.2[CP], The Scripps
Research Institute, Jupiter FL, 2001 .
[20] R. Huey, G.M. Morris, A.J. Olson, D.S. Goodselle, Software news and update a
semiempirical free energy force field with charge-based desolvation,
J. Comput. Chem. 28 (6) (2007) 1145 e1152 .
[21] U. Singh1, R.P. Gangwal1, G.V. Dhoke, R. Prajapati, M. Damre, A.T. Sangamwar,
3D-QSAR and molecular docking analysis of (4-piperidinyl)-piperazines as
acetyl-CoA carboxylases inhibitors, Arab. J. Chem. (2012) .
[22] K. Backbro, S. Lowgren, K. Osterlund, J. Atepo, T. Unge, J. Hulten, N.M. Bonham,
W. Schaal, A. Karlen, A. Hallberg, Unexpected binding mode of a cyclic sul-
famide HIV-1 protease inhibitor, J. Med. Chem. 40 (6) (1997) 898 e902.
[23] Y.T. Li, Y.L. Liu, B.Z. Shi, G.X. Wang, G.Z. Liang, Molecular docking technology
for maleic imide class analysis of characteristics of the role of GSK – 3 alphainhibitors, J. Mol. Sci. 29 (4) (2013) 265 e275.
[24] J.B. Tong, X. Zhao, L. Zhong, J. Chang, M.L. Li, Phenyl pyrrole class aromatase
inhibitors of 3D-QSAR and molecular docking studies, J. Mol. Sci. 30 (5) (2014)
403e409.J. Tong et al. / Journal of Molecular Structure 1129 (2017) 17 e22 22