Turk J Chem

(2015) 39: 306 – 316

c⃝ TUBITAK

doi:10.3906/kim-1402-37

Turkish Journal of Chemistry

http :// journa l s . tub i tak .gov . t r/chem/

Research Article

Molecular docking analysis and molecular dynamics simulation study of

ameltolide analogous as a sodium channel blocker

Maryam IMAN1, Atefeh SAADABADI2, Asghar DAVOOD2,∗

1Chemical Injuries Research Center, Baqiyatallah University of Medical Sciences, Tehran, Iran2Department of Medicinal Chemistry, Pharmaceutical Sciences Branch, Islamic Azad University, Tehran, Iran

Received: 20.02.2014 • Accepted/Published Online: 03.12.2014 • Printed: 30.04.2015

Abstract:Fifteen compounds related to ameltolide with sodium channel inhibitory activity were subjected to a molecular

docking study. The chemical structures of all compounds were built using the program HyperChem and conformational

studies were performed with a semiempirical method followed by the PM3 method. A docking study was performed

using the program AutoDock on all the compounds. To confirm the binding mode of inhibitors, molecular dynamics

simulations were performed using GROMACS 4.5.5, based upon the docked conformation of ameltolide. The docking

analyses indicated that these compounds interacted mainly with residues II-S6 and III-S6 of NaV1.2 by making hydrogen

bonds and (π − π) interactions with domains I, III, and IV in the channel’s inner pore. Our docking study reveals that

amide linker plays a major role in the drug–receptor interaction. The results of molecular dynamic simulations confirmed

the binding mode of ligands, the accuracy of docking, and the reliability of active conformations obtained by AutoDock.

Key words: Ameltolide, molecular dynamics simulation, sodium channel

1. Introduction

Epilepsy is a serious neurological disorder that 1% of the world’s population is affected by1,2 and among

them 30% of sufferers have uncontrolled seizures. Most antiepileptic drugs are associated with adverse effects,

such as sedation, ataxia, and weight loss (e.g., topiramate) or weight gain (e.g., valproate, tiagabine, and

vigabatrin). Rare adverse effects can be life threatening, such as rashes leading to Stevens–Johnson syndrome

(e.g., lamotrigine) or aplastic anemia (e.g., felbamate).2,3 Therefore, the development of safer and more effective

new antiepileptic drugs (AEDs) is necessary.1−5

Sodium channels are one of the best targets in the treatment of epilepsy. Neuronal voltage-gated sodium

channels (NVSCs) play an important role in the production and spread of action potentials in neurons and

other excitable cells. Therefore, NVSC blocking agents constitute a clinically important class of drugs used

in the treatment of neurological disorders. NVSCs usually include an alpha subunit that organizes the ion

conduction pore and one to two beta subunits.6 The alpha subunit has four repeat domains, marked I to IV,

each containing six membrane-spanning regions, marked S1 to S6. The family of sodium channels has nine

known members. The proteins of these channels are labeled Nav1.1 to Nav1.9.7,8

Docking studies are used at different stages in drug discovery such as in prediction of the docked structure

of a ligand–receptor complex and also to rank ligand molecules based upon their binding energy. Docking

∗Correspondence: adavood2001@yahoo.com

306

IMAN et al./Turk J Chem

protocols aid in elucidation of the most energetically favorable binding pose of a ligand to its receptor.

Our work is based on one of the potent AEDs, ameltolide, which previously was investigated by Eli Lilly,

originally emerging from the laboratories of Clark and coworkers.9−15 These groups of investigators isolated

the promising 4-aminobenzamide pharmacophore, which subsequently led to the fruitful design of several new

and potent anticonvulsant compounds. In this paper, we report the molecular modeling and drug–receptor

interaction profile of 15 compounds structurally related to ameltolide (Figure 1) that had been designed and

synthesized already. It was confirmed that this type of ligand acts as a sodium channel blocker,16 and so we

used a model of the open pore of the sodium channel as a receptor. This open pore model was developed

recently based on homology modeling of the crystal structures of the K channel.17,18

R

NH

O CH3

CH3

R

NH

O

NH

CH3

CH3

CH3CH3

NH

S

CH3

CH3

NH

R

I-III IV-VI VII

N

O

O

R

CH3

CH3

N

O

O CH3

CH3 R

N

O

O

R

NH

CH3

CH3

CH3CH3

VIII-X and XIII, XIV XI, XII XV

Comp. R Comp. R Comp. R Comp. R Comp. R

I NO2 IV H VII NO2 X NH2 XIII OH

II NH2 V NO2 VIII H XI NO2 XIV Cl

III OH VI NH2 IX NO2 XII NH2 XV H

Figure 1. The structure of docked compounds I–XV.

2. Results and discussion

Flexible docking of all data sets used for the computational study19 was carried out on the active site of the

open pore of the sodium channel. The lowest energy and maximum number of conformations per cluster were

set as the criteria to predict the binding modes of the compounds. Our docking results (Figure 2a) indicate that

the CO–NH moiety of ameltolide and its benzanilides derivatives bind to the sodium channel in the trans form

and this form is stable in the molecular dynamic simulation (Figure 2b). Based on the procedure explained

in the experimental section, the binding affinity of the docked molecules was evaluated by binding energy,

docked energy, calculated inhibition constants (calc. K i), and hydrogen bonds in addition to the hydrophobic

interactions at the channel pocket. The poses of all compounds are presented in the Table.

307

IMAN et al./Turk J Chem

Figure 2. (a) Drug–receptor interaction after docking (before MD simulation) (b) Average structure based upon the

equilibration of the ameltolide in the 3 last ns, drug-receptor interaction after 10 ns MD simulation.

Table. Docking results of ameltolide analogues by using of AutoDock software (version 4).

Comp. BEa LEb Ki (µM)c IE d VEe EEf TIEg TEh UEi Log P DEj

I –6.81 –0.34 10.25 –7.7 –7.62 –0.08 –0.65 0.89 –0.65 3.70 –8.35II –5.81 –0.32 55.49 –6.4 –6.38 –0.02 –0.43 0.6 –0.43 2.96 –6.83III –5.69 –0.32 67.72 –6.28 –6.25 –0.04 –0.5 0.6 –0.5 3.46 –6.78IV –6.82 –0.36 9.99 –7.42 –6.28 –1.14 –0.2 0.6 –0.2 1.72 –7.62V –6.5 –0.3 17.07 –7.4 –7.42 0.02 –0.48 0.89 –0.48 1.68 –7.88VI –6.58 –0.33 15.0 –7.48 –6.11 –1.37 –0.16 0.89 –0.16 0.94 –7.64VII –5.42 –0.26 106.8 –6.91 –6.83 –0.08 –0.49 1.49 –0.49 5.03 –7.4VIII –5.69 –0.3 67.74 –5.99 –5.94 –0.05 –0.47 0.3 –0.47 3.61 –6.46IX –6.7 –0.3 12.23 –7.0 –7.01 0.01 –0.43 0.3 –0.43 3.56 –7.43X –6.07 –0.3 35.45 –6.37 –6.34 –0.03 –0.44 0.3 –0.44 2.82 –6.81XI –6.33 –0.29 22.86 –6.93 –6.7 –0.23 –0.47 0.6 –0.47 3.56 –7.4XII –5.87 –0.29 49.94 –6.46 –6.43 –0.03 –0.44 0.6 –0.44 2.82 –6.9XIII –5.91 –0.3 46.59 –6.21 –6.13 –0.08 –0.42 0.3 –0.42 3.32 –6.63XIV –5.94 –0.3 44.61 –6.23 –6.21 –0.02 –0.5 0.3 –0.5 4.13 –6.73XV –6.51 –0.31 17.0 –6.8 –6.69 –0.11 –0.29 0.3 –0.29 1.59 –7.09PHE –5.83 –0.31 53.37 –6.43 –6.38 –0.04 –0.71 0.60 –0.71 2.08 –6.74

aBE = The predicted binding energy (kcal/mol) is the sum of intermolecular energy and torsional free energy, bLE =

Ligand efficiency, c Inhibition constant (Ki) = exp (deltaG × 1000)/(Rcal × TK), where deltaG is the docking energy,

Rcalis 1.98719, and TK is 298.15, d IE = Intermolecular energy is sum of Vdw-hb-desolv-energy and Electrostatic-energy,

eVE =Vdw-hb-desolv energy, f EE = electrostatic energy, gTIE = total internal energy, hTE = torsional energy, iUE

= unbound energy, jDE = Docking energy is the sum of intermolecular energy and ligand’s internal energy.

308

IMAN et al./Turk J Chem

The predicted binding and docked energies are the sum of the intermolecular energy and the torsional

free-energy penalty, and the docking ligand’s internal energy, respectively, and the inhibition constant (K i) is

calculated in AutoDock4 as follows:

Ki = exp(∆G× 1000)/(Rcal × TK), (1)

where ∆G is the docking energy, Rcal is 1.98719, and TK is 298.20−24 Our docking results reveal that, based

on the predicted binding energy, compounds IV, I, IX, VI, XV, V, XI, X, XIV, XIII, and XII with –6.82,

–6.81, –6.7, –6.58, –6.51, –6.5 , –6.33, –6.07, –5.94, –5.91, and –5.87 kcal/mol binding energy, respectively, are

more potent than phenytoin as a reference drug, with –5.83 kcal/mol binding energy. According to the K i ,

compounds IV, I, IX, XI, XV, V, XI, X, XIV, XIII, and XII with 9.99, 10.25, 12.23, 15.0, 17.0, 17.07, 22.86,

35.45, 44.61, 46.59, and 49.94 µM inhibition constant can inhibit the enzyme more efficiently when compared

to phenytoin with 53.37 µM inhibition constant. Compounds VII, III, VIII, and II have inhibition constants

and binding energy less than those of phenytoin. The relationship between binding energy and K i is shown in

Figure 3. This relationship is linear with R = 0.9, which means each compound with more binding energy has

a higher inhibition constant. This molecular docking shows in compounds I, VII, and IX that the oxygen of

NO2 forms hydrogen bonding with Ser84 of domain II-S6 (Figure 4). The oxygen of imide in compounds VIII

and XV and the OH of compound XIII form a hydrogen bonding interaction with the OH of Thr87 (Figures 5

and 6). In compounds IV and VI, the NH of amide and the NH of the piperidine ring form a hydrogen bonding

interaction with the Asn84 of domain III and Glu7B of domain II, respectively (Figure 7).

Figure 3. Relationships between binding energy and Ki

(inhibition constant) of compounds I–XV.

Figure 4. Docked structure of compound I in model

of sodium channel. Hydrogen bond (distance = 1.87 A)

between oxygen of NO2 group and Ser84 (binding energy:

–6.81 kcal/mol) is represented by a dashed green line.

Figure 5. Docked structure of compound XV in model of sodium channel; hydrogen bond between carbonyl group and

Thr87 (distance: 2.118 A, binding energy: –6.51 kcal/mol) is represented by a red line.

309

IMAN et al./Turk J Chem

Figure 6. Docked structure of compound XIII in model of sodium channel; hydrogen bond between OH moiety and

Thr87 (distance: 1.75 A, binding energy: –5.91 kcal/mol) is represented by a red line.

Figure 7. Docked structure of compound IV in model of sodium channel; hydrogen bonds between the NH of piperidine

ring and Glu7B, and NH of amide and Asn84 (distance: 1.931 and 1.945 A, respectively, binding energy: –6.82 kcal/mol)

are represented by a red line.

The oxygen of NO2 of the N-aryl part of compound XI forms a hydrogen bonding interaction with the

OH of Thr87. In compounds III and IX there is an efficient (π−π) interaction between the phenyl ring and the

aromatic ring of Phe84 of domain IV and Phe91 of domain III, respectively (Figure 8). In compound XIV, there

is a (π − π) interaction between the phenyl ring of phthalimide and Phe84 of domain IV and Tyr91 of domain

I (Figure 9). While phenytoin interacted with the domain IV of the Na channel, most of compounds I–XV

interacted mainly with the domains II-S6 and III-S6 of NaV1.2 by making hydrogen bonds and have a (π − π)

interaction with domains I, III, and IV in the inner pore of the channel (Figure 10). This docking analysis

reveals that amide linker plays a major role in the drug–receptor interaction because replacing it by thioamide

linker in compound VII results in significantly increasing the binding energy and K i , which agree with the

biological activity of these derivatives.16 Docking and pharmacological data of the subgroup of ameltolide and

phthalimide derivatives confirm that there is no significant reduction in the anticonvulsant effect between the

compounds with opening and closing pyrrolidine rings.

310

IMAN et al./Turk J Chem

Figure 8. Docked structure of compound IX in model of sodium channel; hydrogen bond between oxygen of NO2 group

and Ser84 (distance: 2.219 A, binding energy: –6.7 kcal/mol) is represented by a red line and (π−π) interaction between

phenyl ring and Phe91 of domain III is represented by a yellow cylinder.

Figure 9. Docked structure of compound XIV in model of sodium channel; (π − π) interactions between phenyl ring

of phthalimide and Phe84 of domain IV and Tyr91 of domain I are represented by a yellow cylinder, π − π interactions

are shown by a yellow cylinder.

Based on the pharmacological study, ameltolide (compound II) appears to be as potent as phenytoin

in interacting with the Na channel. The phthalimide (compound X) and nitro counterparts of ameltolide

(compound I) are several times more potent than ameltolide and phenytoin in binding tests and oral (maximal

electroshock seizure) MES tests in rats, which agree with the docking study of these compounds.16

2.1. Molecular dynamics simulation

To confirm the binding profile of ligands and to give an overall impression about the ameltolide derivatives,

ameltolide was subjected to 10 ns molecular dynamics (MD) simulations. The equilibration was monitored and

confirmed by examining the stability of the temperature, pressure, density, and potential energy of the system

as well as the root mean squared deviations (RMSDs) of the backbone atoms. The average backbone RMSD

of receptor was 0.6974 ± 0.0010 A with respect to the starting structure and the potential energy equilibrated

about an average of –556,254 ± 777.026 kJ/mol. Average temperature, pressure, and density were 300.001 T

(RMSD: 1.48133), 1.03143 bar (RMSD: 115.221), and 988.342 kg/m3 (RMSD: 2.302), respectively. All of the

311

IMAN et al./Turk J Chem

Figure 10. Right: docked structure of phenytoin in model of the open sodium channel (Nav1.2). The backbones of S6

α -helices of domains I-IV are shown by red, cyan, yellow, and magenta, respectively. Hydrogen bond (distance: 2.06 A

and binding energy: –5.83 kcal/mol) is formed between hydrogen of imide and Ser83 of domain IV-S6 . Left: docked

structure of compound I in model of sodium channel (one of ameltolide analogous) has a hydrogen bond with domain

II-S6.

plots show the normal oscillation behavior of the temperature, pressure, and density. The backbone RMSD of

protein (Figure 11a) indicates that the rigid protein structure equilibrates rather quickly after 1 ns. The drug

RMSD (0.1064 ± 0.00023, Figure 11b) indicates that the drug equilibrates after 3.5 ns, and the RMSD for

the drug reveals its mobility within the binding site in two phases. Figure 11 and the RMSD value (0.1064 ±0.00023) reveal that the ligands stayed within the binding site. It should be mentioned that the RMSD values of

ligands are a more reliable indicator of their mobility and here the low RMSD of ameltolide (0.1064 ± 0.00023)

over the 10 ns shows that ligands remained within the active site pocket. The very small standard deviation

of the RMSD demonstrates the stability of ameltolide within the active site. The number of hydrogen bonds

and the distance between the ameltolide and the receptor (Figures 12a and 12b) were analyzed using a distance

cutoff of 2.5 A and an angle cutoff of 60◦ . For receptor–drug complex, there is one important hydrogen bond

Figure 11. The results of MD simulation (a) The MD simulation time vs. RMSD of the backbone atoms (C, N, and

Cα) of protein. (b) The MD simulation time vs. RMSD of the ameltolide.

312

IMAN et al./Turk J Chem

between NH2 of ameltolide and the oxygen atom in residue Asn88. Figures 2a and 2b show the drug–receptor

interactions before and after 10 ns MD simulation; in both of them there is an efficient hydrogen bond between

Asn88 and drug.

Figure 12. The results of MD simulation (a) The MD simulation time vs. number of hydrogen bond deviation between

ameltolide and receptor. (b) Intermolecular distance from NH2 of ameltolide to oxygen atom of Asn88.

Figure 13. Whole protein of sodium channel that was made by homology modeling, a) top view of secondary structure

b) side view of secondary structure, different domains are shown in different colors.

In analyzing the drug–enzyme complex, our observation reveals that the complex was stable to the

simulation conditions and ameltolide remained in the active site pocket (Figures 2a and 2b). The results of

MD simulations confirmed the binding mode of ligands, the accuracy of docking, and the reliability of active

conformations obtained by AutoDock.

The docking analyses indicated that these compounds interacted mainly with residues II-S6 and III-S6 of

NaV1.2 by making hydrogen bonds and (π–π) interactions with domains I, III, and IV in the channel’s inner

pore. Our docking analysis reveals that amide linker plays a major role in drug–receptor interactions. It is

noteworthy that the MD simulation study corroborates the accuracy of the results of docking analysis.

313

IMAN et al./Turk J Chem

3. Experimental

3.1. Molecular modeling

The molecular modeling and docking procedures were based on our previous articles.25−28 X-ray crystallography

revealed that ameltolide exists in the trans configuration but using molecular modeling it was suggested that

the CO–NH moiety of ameltolide and its derivatives bind to their biological target in their cis configuration.29

Because of the greater reliability of the crystallographic results and small energy difference between the cis and

trans forms (3 kcal/mol),29 in the flexible docking study the trans form was considered and also the amide

bond was considered as a rotatable bond to create both the cis and trans forms in docking calculations. The

chemical structures of inhibitors (Figure 1) were constructed using HyperChem software (version 7, Hypercube

Inc.). Conformational analysis of the favorite compounds was executed through semiempirical molecular orbital

calculations (PM3) by utilizing the software HyperChem.

Total energy gradient was assessed using the Polak–Ribiere (conjugate gradient) algorithm as a root mean

square (RMS) value, until the RMS gradient was 0.01 kcal mol−1 . The gradient (G) is the rate of change (first

derivative) of total energy (TE) with regard to movement of each atom in the x, y, and z directions for atoms

from 1 to n. HyperChem reports this value for geometry optimization and single point calculations. An RMS

gradient of zero means that the structure is at a local minimum or saddle point in the potential energy surface,

not necessarily at the structure and state of the lowest energy (global minimum).

Energy minimization alters molecular geometry to lower the energy of the system and yields a more stable

conformation. The generation of new starting conformations for energy minimization uses random variation of

dihedral angles. Rotation is used for acyclic bond dihedral angles. As the minimization progresses, it searches

for a molecular structure in which the energy does not change with miniscule changes in geometry. This means

that the derivative of the energy with respect to all Cartesian coordinates, called the gradient, is near zero. If

small changes in geometric parameters raise the energy of the molecule, the conformation is relatively stable,

and this is referred to as a minimum. If the energy lowers by small changes in one or more dimensions, but not

in all dimensions, it is a saddle point. A molecular system can have many minima. The one with the lowest

energy is called the global minimum and the rest are referred to as local minima. Among all energy minima

conformers, the global minima of compounds were applied in docking calculations and the resulting geometry

was transferred into AutoDock (version 4.2), which was developed by Arthur J Olson’s Chemometrics Group.19

The structure of the docked conformer of compound II is shown in Figure 1.

3.2. Docking

Docking calculations were executed using AutoDock (version 4.2). A model of the open pore of the Na channel

was utilized as a receptor (Figures 13a and 13b). This open pore model was developed based on a homology

model of the crystal structures of the K channel.17,18 The model constructed by homology with K channel

structures was advisedly successful in accounting for inner pore residue interactions with local anesthetics and

anticonvulsant drugs like phenytoin. The desired compounds were docked in to the active site as well as

phenytoin, which was acting as our reference drug for validation of our procedure.

Docking was done using AutoDock 4.2; in order to assign the perfect grid of each ligand, grid box values

were obtained by trial and error and from previous research.20−23 Grid maps with 60 × 60 × 60 points were

made and the grid point spacing was 0.375 A.24 The Lamarckian genetic algorithm (LGA), considered one of

the best docking methods available in AutoDock, was adopted to perform the molecular docking studies. The

314

IMAN et al./Turk J Chem

parameters for LGA were defined as follows: a maximum number of 250,000 energy evaluations, a maximum

number of generations of 27,000, and mutation and crossover rates of 0.02 and 0.8, respectively. Pseudo-Solis

and Wets parameters were used for the local search, and 300 iterations of the Solis and Wets local search were

imposed. Both AutoGrid and AutoDock computations were performed on Cygwin and 100 independent docking

runs were performed for each phthalimide. Final docked conformations were clustered using a tolerance of 1

A RMSD and the docking log (dlg) files were analyzed using AutoDock Tools, the graphical user interface of

AutoDock. The docked conformations of each ligand were ranked into clusters based on the binding energy and

the top ranked conformations were visually analyzed. Hydrogen bonding and hydrophobic interactions between

docked potent agents and macromolecules were analyzed using AutoDock Tools (version 1.50).

3.3. Molecular dynamics simulation studies

To confirm the binding mode of inhibitors, the MD simulations were performed using GROMACS 4.5.530,31

based on the docked conformation of ameltolide and whole protein. A united-atom GROMOS96 43A1 force

field was used for protein parameters and using the PRODRG server, and a GROMOS87/GROMOS96 force

field was used to generate a starting topology for ameltolide. One Na+ ion was added to neutralize the system,

and then solvated in a cubic box of spc water molecules.

Before the MD simulation, the complexes were subjected to 50,000 steps of energy minimization to

relieve any geometric strain and close intermolecular contacts. To begin real dynamics, the solvent and ions

were equilibrated around the protein. Equilibration was conducted in two phases: NVT and NPT. In the first

phase, with a weak constraint to the system (10 kcal/mol), the system was conducted under an NVT ensemble

(constant number of particles, volume, and temperature) and was gradually heated from 0 to 300 K in 100

ps and then equilibrated for 100 ps at 300 K. In the second phase of equilibration, the system was conducted

under an NPT ensemble (constant number of particles, pressure, and temperature) and equilibrated for 100 ps

at 300 K. Then a 10 ns MD simulation was performed using the periodic boundary conditions, with constant

temperature and pressure (1 bar at 300 K). The output trajectories were recorded every 2 ps for the purpose of

subsequent analysis. The equilibration was monitored by examining the stability of the temperature, pressure,

potential energy, and density of the system as well as the RMSD of the backbone atoms.32,33

Acknowledgments

We are grateful to the Azad University for financial support for this research, and to Prof Arthur J Olson and

Prof A Fozzard for their kindness in offering us the AutoDock 4.2 program and the homology model of the

sodium channel.

References

1. Walker, M. C.; Sander, J. W. Seizure 1996, 5, 165–169.

2. Iman, M.; Saadabadi, A.; Davood, A. Iran. J. Basic. Med. Sci. 2013, 16, 1016–1021.

3. Bourgeois, B. F. Arch. Neurol. 1998, 55, 1181–1183.

4. De Sarro, G.; Di Paola, E. D.; Gareri, P.; Gallelli, L.; Scotto, G.; De Sarro, A. Eur. J. Pharmacol. 1999, 368,

149–159.

5. Meldrum, B. Epilepsy Res. 2002, 50, 33–40.

6. Yuesheng, W.; Jones, P. J.; Batts, T. W.; Landry, V.; Patel, M. K.; Brown, M. L. Bioorg. Med. Chem. 2009, 17,

7064–7072.

315

IMAN et al./Turk J Chem

7. Goldin, A. L. Annu. Rev. Physiol. 2001, 63, 871–894.

8. Catterall, W. A.; Goldin, A. L. Waxman, S. G. Pharmacol. Rev. 2005, 57, 397–409.

9. Clark, C. R.; Wells, M. J. M.; Sansom, R. T.; Norris, G. N.; Dockens, R. C.; Ravis, W. R. J. Med. Chem. 1984,

27, 779–782.

10. Clark, C. R.; Sansom, R. T.; Lin, C. M.; Norris, G. N. J. Med. Chem. 1985, 28, 1259–1262.

11. Clark, C. R.; Lin, C. M.; Sansom, R. T. J. Med. Chem. 1986, 29, 1534–1537.

12. Clark, C. R.; Davenport, T. W. J. Med. Chem. 1987, 30, 1214–1218.

13. Clark, C. R. Epilepsia 1988, 29, 198–203.

14. Leander, J. D.; Robertson, D. W.; Clark, C. R.; Lawson, R. R.; Rathbun, R. C. Epilepsia 1988, 29, 83–90.

15. Leander, J. D.; Lawson, R. R.; Robertson, D. W. Neuropharmacology 1988, 27, 623–628.

16. Vamecq, J.; Lambert, D.; Poupaert, J. H.; Masereel, B.; Stables, J. P. J. Med. Chem. 1998, 41, 3307–3313.

17. Lipkind, G. M.; Fozzard, H. A. Mol. Pharmacol. 2005, 68, 1611–1622.

18. Lipkind, G. M.; Fozzard, H. A. Biochem. 2000, 39, 8161–8170.

19. Morris, G. M.; Goodsell, D. S.; Pique, M. E.; Lindstrom, W. L.; Huey, R.; Forli, S. et al. J. Comput. Chem. 2009,

30, 2785–2791.

20. Cosconati, S.; Marinelli, L.; Lavecchia, A.; Novellino, E. J. Med. Chem. 2007, 50, 1504–1513.

21. Storici, P.; De Biase, D.; Bossa, F.; Bruno, S.; Mozzarelli, A.; Peneff, C. J. Biol. Chem. 2004, 279, 363.

22. Huey, R.; Morris, G. M. Using AutoDock 4 with AutoDocktools: a tutorial. La Jolla, CA, USA: The Scripps Research

Institute, Molecular Graphics Laboratory; 2008, pp. 54–56.

23. Davood, A.; Nematollahi, A.; Iman, M.; Shafiee, A. Arch. Pharm. Res. 2009, 32, 481–487.

24. O’Reilly, A. O.; Khambay, B. P. S.; Williamson, M. S.; Field, L. M.; Wallace, B. A. Biochem. J. 2006, 396,

255–263.

25. Davood, A.; Iman, M. Med. Chem. Res. 2011, 20, 955–961.

26. Davood, A.; Iman, M.; Nematollahi, A.; Shafiee, A. Med. Chem. Res. 2012, 21, 325–332.

27. Davood, A.; Shafaroodi, H.; Amini, M.; Nematollahi, A.; Shirazi, M.; Iman M. Med. Chem. 2012, 8, 953–963.

28. Davood, A.; Iman, M. Turk. J. Chem. 2013, 37, 119–133.

29. Philip, A. E.; Poupaert, J. H.; Cheve, G.; Muccioli, G.; Lambert, D.; McCurdy, C. R. Med. Chem. Res. 2007, 16,

130–135.

30. Hess, B.; Kutzner, C.; Spoel, D van der.; Lindahl, E. J. Chem. Theory. Comput. 2008, 4, 435–447.

31. Kerrigan, J. E. Gromacs Drug/Enzyme complex solvation tutorial. IST/Academic Computing Services RWJMS,

University of Medicine and Dentistry of NJ, Piscataway, NJ, USA, 1997.

32. Cakan, S.; Akdogan, E. D. Turk. J. Chem. 2013, 37, 449–463.

33. Yuan, M.; Luo, M.; Song, Y.; Xu, Q.; Wang, X.; Cao, Y.; Bu, X.; Ren, Y.; Hu, X. Bioorg. Med. Chem. 2011, 19,

1189–1196.

316