InsightsintotheMolecularMechanismsofProtein-Ligand...

Research ArticleInsights into the Molecular Mechanisms of Protein-LigandInteractions by Molecular Docking and Molecular DynamicsSimulation A Case of Oligopeptide Binding Protein

Yi Fu 123 Ji Zhao12 and Zhiguo Chen 3

1Wuxi Research Center of Environmental Science and Engineering Wuxi 214153 Jiangsu China2School of Internet of ings Engineering Wuxi City College of Vocational Technology Wuxi 214153 Jiangsu China3School of Internet of ings Engineering Jiangnan University Wuxi 214122 Jiangsu China

Correspondence should be addressed to Yi Fu fuyi_brightyahoocom

Received 29 July 2018 Accepted 16 October 2018 Published 4 December 2018

Guest Editor Xiaole Chen

Copyright copy 2018 Yi Fu et al +is is an open access article distributed under the Creative Commons Attribution License whichpermits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Protein-ligand interactions are a necessary prerequisite for signal transduction immunoreaction and gene regulation Protein-ligand interaction studies are important for understanding the mechanisms of biological regulation and they provide a theoreticalbasis for the design and discovery of new drug targets In this study we analyzed the molecular interactions of protein-ligandwhich was docked by AutoDock 42 software In AutoDock 42 software we used a new search algorithm hybrid algorithm ofrandom drift particle swarm optimization and local search (LRDPSO) and the classical Lamarckian genetic algorithm (LGA) asenergy optimization algorithms +e best conformations of each docking algorithm were subjected to molecular dynamic (MD)simulations to further analyze the molecular mechanisms of protein-ligand interactions Here we analyze the binding energybetween protein receptors and ligands the interactions of salt bridges and hydrogen bonds in the docking region and thestructural changes during complex unfolding Our comparison of these complexes highlights differences in the protein-ligandinteractions between the two docking methods It also shows that salt bridge and hydrogen bond interactions play a crucial role inprotein-ligand stability+e present work focuses on extracting the deterministic characteristics of docking interactions from theirdynamic properties which is important for understanding biological functions and determining which amino acid residues arecrucial to docking interactions

1 Introduction

Molecular docking methods are of utmost importance andhave been widely used in new drug design and discoveryprojects [1ndash3] Molecular docking methods can provide arelatively fast and economical alternative to standard ex-perimental techniques [4 5] +ey aim to predict the ex-perimental binding modes and affinities of small moleculeswithin the binding site of particular receptor targets Twoimportant goals in molecular docking are to find correctbinding poses and to accurately predict binding affinityMore accurate predictions of binding poses and bindingaffinities can suggest candidates for active compounds withhigher true positive rates and can considerably reduce ex-pensive experimental efforts [6] +e quality of molecular

docking depends on two factors the optimization searchmethod and the scoring function [7ndash9] An optimizationalgorithm mainly detects docking conformations withminimum binding energies +e scoring function is used toevaluate the results obtained from the search In this articlewe used AutoDock 42 software to perform moleculardocking In AutoDock 42 a new hybrid algorithm ofrandom drift particle swarm optimization with local search(LRDPSO) [10] and Lamarckian genetic algorithm (LGA)[11] were used as energy search algorithms Here randomdrift particle swarm optimization (RDPSO) algorithm is avariant of the particle swarm optimization algorithm (PSO)LGA is a combination of a genetic algorithm (GA) and alocal search (LS) and is one of the classical energy searchalgorithms in AutoDock 42 software

HindawiComputational and Mathematical Methods in MedicineVolume 2018 Article ID 3502514 12 pageshttpsdoiorg10115520183502514

In this study oligopeptide binding protein (OppA) [12]was used as the experimental object OppA can act as areceptor for peptide transport across the cell membrane andis a potential target in antibacterial drug design [13] It hasbroad specificity and can bind to a wide range of peptidesthat are 2ndash5 amino acid residues long Here we used thesequence Lys-Tyr-Lys as the ligand binding structure +eOppA structure (PDB code 1B58 [14]) consists of threedomains domain I contains residues 1ndash44 189ndash269 and487ndash517 domain II contains residues 45ndash188 and domainIII contains residues 270ndash486 Domains I and II have abilobate structure allowing the ligand to act as a flexiblehinge connecting the two lobes In the protein-ligandcomplex structure ligands are completely contained inthe protein interior [15] For LRDPSO and LGA we selectedthe two best solutions for 1B58 in terms of binding energy+e lowest binding energies in LRDPSO and LGA wereminus2509 kcalmol and minus1274 kcalmol respectively

Protein-ligand interactions play an important role in mostbiological processes such as signal transduction cell regula-tion and immune response [16 17] Studying protein-ligandinteractions continues to be very important in life sciencefields [18ndash21] +ere are variations in protein-ligand complexstructures due to different docking methods In this article wemainly focus on analyzing the binding interactions between aprotein and a ligand especially on the divergence of protein-ligand interactions which can help us understand and addresskey questions such as those related to the diversity of bindingaffinity and specificity We have performed molecular dy-namic (MD) simulations on molecular docking results at fourdifferent temperatures (ranging from 300K to 600K) to es-tablish a more reliable mechanism for illustrating ligand-protein interactions +e dynamic properties of complexeshave been compared in terms of residue flexibility bindingenergy salt bridge and hydrogen bond interactions in thebinding region and structural variations during unfolding atdifferent temperatures +is study provides a better un-derstanding of the specific interactions predicted by differentdocking methods and it also allows us to more precisely studya binding site or region to increase docking accuracy

2 Materials and Methods

21 Protein and Ligand Structure Preparation +e OppAstructure (PDB code 1B58 [14]) which was obtained from theRSCB protein data bank (httpwwwrcsborg) was used as areceptor of the experimental object+e sequence Lys-Tyr-Lyswhich contains 43 atoms was used as the ligand structure+edocking results were used as models for the MD simulation

22 Genetic Algorithm (GA) AutoDock has been appliedwith great success in the prediction of binding conforma-tions of protein-protein interactions peptide-antibodycomplexes and enzyme-inhibitor complexes Earlier ver-sions of AutoDock used simulated annealing as a searchmethod a subsequent version added the options of a GA aLS method and a combination of GA and LS which is calledLGA In LGA a GA is used for global searching and Solis

and Wets [22] is used as the LS method Each generation isfollowed by a LS which is performed on a user-definedproportion of the population

Genetic algorithm [23 24] is a population-based meta-heuristic algorithm which contains initial population gen-eration fitness function evaluation iteration and terminationcondition check off four steps In addition every iteration stepincludes selection crossover and mutation operations +epseudocode of GA is described in Algorithm 1

23 HybridAlgorithm of RDPSOand Local Search (LRDPSO)RDPSO [25] is derived from canonical PSO trajectoryanalysis [26] and the free electron model Particle behaviorin RDPSO is assumed to be similar to an electron moving ina metal conductor in an external electric field Particlemovement is thus the superposition of thermal and driftmotions which is based on a global search and local searchof the particle respectively

In a RDPSO with M individuals each individual istreated as a volumeless particle in the N-dimensional spaceVin (V1

in V2in VN

in) and Xin (X1in X2

in XNin)

are expressed as the velocity vector and the position vector ofparticle i at the nth iteration respectively

According to the above model the updated equations ofRDPSO can be expressed by the following equation

Vjin+1 α C

jn minusX

jin

11138681113868111386811138681113868

11138681113868111386811138681113868ϕjin + β p

jin minusX

jin1113872 1113873 (1)

Xj

in+1 Xj

in + Vj

in+1 (2)

Cjn

1M

1113874 1113875 1113944

M

i1P

j

in (3)

for i 1 2 M j 1 2 N where αgt 0 is a parametercalled the thermal coefficient and βgt 0 is another parametercalled the drift coefficient ϕj

in is a random numberϕj

in sim N(0 1) pjin is the local attractor of PSO algorithm Pj

in

is the personal best (pbest) position of particle i Cjn is defined

by the mean of the pbest positions of all particles called themean best (mbest) position +e pseudocode of RDPSO isdescribed in Algorithm 2

+e hybrid of the RDPSO algorithm with the Solis andWets algorithm together form the LRDPSO algorithm Herethe RDPSO algorithm is used as a global search algorithmand the Solis and Wets algorithm is used as a local searchalgorithm +e Solis and Wets algorithm can facilitatetorsional space search since it does not require gradientinformation about the local energy landscape [22 27] +eaddition of local search effectively maintains the diversity ofparticles and prevents premature convergence of the algo-rithm +erefore the effective combination of the RDPSOalgorithm and the Solis and Wets algorithm can provide agood global search ability and a rapid convergence ability+e pseudocode of LRDPSO is described in Algorithm 3

24 e Docking Experiment Settings In molecular dockingexperiments the prepared protein and ligand structureswere saved in the PDBQT file format +e AutoDockTools

2 Computational and Mathematical Methods in Medicine

(ADT) was used as molecular graphical visualization tool+e AutoDock package includes AutoGrid program andAutoDock program AutoGrid program is responsible for

the calculation of energy grid maps here a grid size was setto 60 times 60 times 60 points with a spacing of 0375 A AutoDockprogram is responsible for the conformation search and

(1) Initialize population P0(2) While (g lt maximum) do(3) Rg⟵recombine(Pg)(4) Mg⟵mutate(Rg)(5) Evaluate (Mg)(6) Pg+1⟵select(Mg cupRg)(7) End while

ALGORITHM 1 Pseudocode of GA

(1) Initialize the positions and velocities of all particles randomly(2) Evaluate the objective function value f(Xin)(3) Set the personal best position of each particle to its current position(4) Set n 0(5) While (termination condition false) do(6) n n + 1(7) Compute mean best position Cn according to equation (3)(8) For (i 1 to M)(9) Update the personal best position (Pin) and global best position (Gn)(10) For (j 1 to N)(11) V

jin+1 α|Cj

n minusXjin|ϕj

in + β(pjin minusX

jin)

(12) Xj

in+1 Xj

in + Vj

in+1(13) End for(14) Evaluate the objective function value f(Xin+1)(15) End for(16) End while

ALGORITHM 2 Pseudocode of RDPSO

(1) Initialize the positions and velocities of all the particles randomly(2) Evaluate the docking energy value f(Xin)(3) Set the personal best position of each particle to be its current position(4) Set n 0(5) While (termination condition false) do(6) n n + 1(7) Compute mean best position Cn according to equation (3)(8) For (i 1 to M)(9) Update the personal best position (Pin) and global best position (Gn)(10) Calculate the velocity Vin+1 using equation (1)(11) Calculate the position Xin+1 using equation (2)(12) Evaluate the docking energy value f(Xin+1)(13) End for(14) Apply the Solis and Wets to the best particle among all Xin+1(15) For (i 1 to M)(16) if Xin+1 better than Pin then(17) Pin+1 Xin+1(18) else(19) Pin+1 Pin(20) End for(21) End while(22) Return the best position among all Pin as optimal solution

ALGORITHM 3 Pseudocode of LRDPSO

Computational and Mathematical Methods in Medicine 3

energy evaluation here for LRDPSO and LGA algorithmsthe initial population was set to 50 individuals the numberof energy function evaluations was set to 25 times 105 andmaximum number of generations was set to 27000 +edetail setting information refers to reference [10] We fol-lowed the methods of Fu et al [10] In that article weconcentrate on discussing the design of LRDPSO algorithmand its application in protein-ligand docking

25 Molecular Dynamic (MD) Simulation Docked protein-ligand complexes were subjected to molecular dynamicsimulations using NAMD software [28] MD simulationswere performed using the CHARMM27 force field [29]Visual molecular dynamics (VMD) [30] was used to generatePSF files for both complexes Both complexes were solvated incubic water boxes containing transferable intermolecularpotential with 3 points (TIP3P) water molecules [31]+e boxsize was chosen so that there was a distance of 10 A betweenthe protein surface and the edges of the periodic box A 12 Acutoff distance was used to calculate short-range nonbondedinteractions+e particlemesh Ewald (PME) [32]methodwasused to calculate long-range electrostatic interactions +eSHAKE method [33] was used to constrain all bonds in-volving hydrogen atoms +e system first performed 10000steps of steepest descent with energy minimization +en theminimized system was used to perform simulations using anNVT ensemble +e NosendashHoover method [34] was used tomaintain a constant temperature +e simulated temperaturewas set in the range of 300K to 600K with an interval of100K +e simulation time for each simulated temperaturewas set to 10 ns +e time step of each simulation was set to2 fs Visualizations and data analysis were performed withVMD software

3 Results and Discussion

31 Molecular Docking Energy Analysis A semiempiricalfree energy scoring function was used to evaluate a dockedconformation in AutoDock 4 [11] +e total docked energyof the ligand and protein included two components whichare the intramolecular energy and the intermolecular energy+e intramolecular energy was evaluated for the transitionfrom the unbound state to the bound conformation of theligand and protein+e intermolecular energy was estimatedfor the combination of the ligand and the protein in theirbound conformation

+e force field consists of six pairwise evaluations (V)and an estimate of the conformational entropy lost uponbinding (ΔSconf )

ΔG VLminusLbound minusV

LminusLunbound1113872 1113873 + V

PminusPbound minusV

PminusPunbound1113872 1113873

+ VPminusLbound minusV

PminusLunbound + ΔSconf1113872 1113873

(4)

where P refers to the ldquoproteinrdquo and L refers to the ldquoligandrdquo ina protein-ligand docking calculation

Each of the pairwise energetic items includes the fol-lowing energy the first item is a Lennard-Jones 12-6 van derWaals interaction the second item is a 12-10 hydrogen bond

potential the third item is a coulombic electrostatic po-tential and the final item is a desolvation potential

V Wvdw 1113944ij

Aij

r12ij

minusBij

r6ij

⎛⎝ ⎞⎠ + WHbond 1113944ij

E(t)Cij

r12ij

minusDij

r10ij

⎛⎝ ⎞⎠

+ Welec 1113944ij

qiqj

ε rij1113872 1113873rij

+ Wsol 1113944ij

SiVj + SjVi1113872 1113873eminusr2

ij2σ21113872 1113873

(5)

We used the sequence Lys-Tyr-Lys which contains 43atoms as the ligand structure Table 1 shows a comparison ofthe best solutions obtained (out of 30 independent runs)from both LRDPSO and LGA for the OppA complexFigure 1 shows a comparison of each ligand atom bindingelectrostatic energy and van der Waals energy In general asindicated in Table 1 and Figure 1 the electrostatic and vanderWaals energy of the complex were lower when docked byLRDPSO than when docked by LGA In the moleculardocking predicted complex a lower binding energy wasassumed to be closer to the native state of the complex Forboth complexes the LRDPSO complex was energeticallymore stable than the LGA one given the obtained energyresults the lowest binding energy corresponded to minus-2509 kcalmol for the LRDPSO complex structure and minus-1274 kcalmol for the LGA one Both ligands were docked tothe site of protein 1B58 but the LRDPSO ligand confor-mation had a better docking position and a root mean squaredeviation (RMSD) of 063 A +e RMSD of the LGA dockingwas 200 A (Figure 2)

For LRDPSO and LGA we selected the two best solutionswith the lowest binding energy in analysis +e ligand con-formations in LRDPSO and LGA and the reference ligand arecompared in Figure 2 Figure 2(a) shows the best energysolution (the ligand in magenta) obtained by LRDPSO forprotein 1B58 and the reference ligand (in green) Figure 2(b)shows the solution selected from LGA and the referenceligand As shown the ligand has a better conformation inFigure 2(a) than in Figure 2(b) given that the ligand con-formation obtained by LRDPSO is very similar to that ofthe reference ligand +e RMSD scores of the ligand con-formations by LRDPSO and LGA were 063 A and 200 Arespectively

32 Structural Stability Analysis upon Ligand BindingWe assessed the residue RMSD to study the residue behaviorof the protein during the simulations In general a residuersquosRMSD value was considered to represent the local flexibilityof a protein It reflected the mobility of an atom during theMD simulation trajectory +erefore a higher residueRMSD value indicated higher mobility conversely a lowerresidue RMSD value indicates lower mobility

+e values of RMSD against each residue were calculatedfor both complexes by MD simulation at multiple tem-peratures +is presentation clearly highlighted the differ-ences in some residues of the complexes +e results areshown in Figure 3 for both complexes there were relativelysmall bump-like peaks for the structures at 300K As the


temperature increased some regions showed signicant in-creases in uctuation e curves observed for both com-plexes exhibited great similarity in their uctuation trendsOnly a few residues showed a great dierence in heat uc-tuations For example when the simulation temperature was300K the RMSD values of residue PHE253 were 1131 A and231 A corresponding to the LRDPSO and LGA dockingresults e RMSD values of residue PHE44 in the LRDPSOand LGA docking complexes were 230 A and 765 A re-spectively In the 500K simulation the region betweenARG41 and VAL164 had higher uctuation in the LRDPSOdocking complex than in the LGA docking result In additionfor both complexes the RMSD values of the residues asso-ciated with the ligand were relatively low even in the hightemperature simulations

33 Salt Bridge Analysis of the Binding Domain We dockedthe tripeptide into the OppA crystal structure e peptide

was bound to the central cleft that surrounded domain I andII In the case of the OppA complex there was a dierencebetween the results obtained by LRDPSO and LGA edierence can be explained by the use of dierent stochasticsearch algorithms Next we analyze concrete binding in-teractions and the changes in these interactions underthermal stress

In this article a salt bridge was dened according to thecriterion that the distances between any of the nitrogenatoms of basic residues and the oxygen atoms of acidicresidues were less than 4 A Figure 4 shows a salt bridgeinteraction associated with the ligand and the surroundingprotein residuese ligand can pack tightly into the bindingsite through a number of favorable salt bridge interactionswith the protein In the LRDPSO docking complex the LYS1of the ligand is anchored through a salt bridge interactionwith ASP419 Meanwhile the ligand LYS3 also formed a saltbridge interaction with GLU229 in the complex Two sig-nicant salt bridges (ASP419-LYS1 and GLU229-LYS3) were

ndash08ndash06ndash04ndash02

00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(a)

ndash12ndash1


00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(b)

Figure 1 Comparison of the binding electrostatic (a) and van derWaals (b) energy of the ligand atom corresponding to LRDPSO and LGA

(a) (b)

Figure 2 Molecular docking results superposition of the predicted conformation (shown in magenta) and the native conformation (shownin green) (a) LRDPSO docking method and (b) LGA docking method

Table 1 Comparison of the docking energy of complexes from LRDPSO and LGA

Complex Binding freeenergy (kcalmol)

Vdw + Hbond + desolvenergy (kcalmol)

Electrostaticenergy (kcalmol)

Intermolecularenergy (kcalmol)

Internalenergy (kcalmol)

LRDPSO docking minus1680 minus1686 minus561 minus2247 minus262LGA docking minus1274 minus1451 minus390 minus1840 minus433Vdw van der Waals Hbond hydrogen bonds desolv desolvation


found in the binding position docked by LRDPSO Com-pared to the complex that was docked by LRDPSO there wasonly one salt bridge (ASP419-LYS1) located in the bindingregion in the LGA-docked complex

Figure 5 shows the changes in the salt bridge distancein the last 8 ns of the simulations In the dynamic simula-tion at 300K salt bridge ASP419-LYS1 was stable during the

simulation in both complexes In the 400K simulation thesalt bridges of both complexes experienced a short separationduring the simulation but they were mostly maintained ata distance of approximately 40 A In comparison salt bridgeASP419-LYS1 was found to be more stable in LRDPSOdocking than in LGA docking +e average distance ofASP419-LYS1 in the LRDPSO and LGA docking was 398 Aand 418 A respectively Along with the increase in thesimulated temperature the distance of the salt bridge alsochanged In the 500K simulation ruptures and restorations ofsalt bridge ASP419-LYS1 were observed along the wholesimulation process +e most obvious difference in this saltbridge between the two complexes was that salt bridgeASP419-LYS1 in the LGA docking was completely separatedduring the last 5 ns of the simulation +e disruption of saltbridge ASP419-LYS1 probably greatly weakened ligandbinding under extremely high temperatures

+e other salt bridge (GLU229-LYS3) was found in onlythe LRDPSO docking result Among the four differenttemperatures the salt bridge was the most unstable in the400K simulation and it was maintained within a shortdistance during only the first 400 ps of the simulation

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)

Figure 3 Plot of RMSD and the residue number of the docked complex in the MD simulated structures at different temperatures (a)LRDPSO-docked complex and (b) LGA-docked complex

Figure 4 An overview of the location of salt bridges in the bindingdomain of the LRDPSO-docked complex +e dashed line repre-sents a salt bridge interaction +e adjacent number is the corre-sponding distance of the salt bridge


Although the salt bridge plot showed several transientseparations in the 500K and 600K simulations the saltbridge remained at a short distance most of the time Finallythe salt bridge was completely separated after 65 ns ofsimulation at 600K

34 Hydrogen Bond Analysis upon Ligand BindingHydrogen bonds are another important factor that in-fluences protein stability Here a distance cutoff of 35 A andan angle cutoff of 30deg were applied in the hydrogen bondcalculation +e study showed that the ligand was entirelyburied in the interior of OppA +e main chain and sidechain of the ligand could form strong hydrogen bond in-teractions with the binding site residues of OppA

Figure 6 shows the hydrogen bond interactions associ-ated with the ligand and the surrounding protein residues inthe LRDPSO docking result LYS1 forms hydrogen bondswith the side chain of ASP419 and with the main chain ofCYS417 and LYS3 forms a hydrogen bond with the sidechain of ARG413 In both docking complexes hydrogenbonds are listed in Tables 2 and 3 Both tables also list theoccupancy time of hydrogen bonds in the binding regionduring the simulation from temperatures of 300K to 600KAs shown in the tables hydrogen bonds are a significantfactor that contributes to the stability of protein-ligandbinding interactions in both docking complexes Mean-while several hydrogen bond networks exist in the binding

region It can be seen that some hydrogen bonds (listed inthe tables) did not exist individually In the LRDPSOdocking result LYS1 of the ligand could form a hydrogen

02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)

Figure 5 Plot of salt bridge changes in the binding domain as a function of time (a) Salt bridge ASP419-LYS1 changes in the LRDPSO-docked complex during the simulation from 400K to 600K (b) salt bridge GLU229-LYS3 changes in the LRDPSO-docked complex duringthe simulation from 400K to 600K (c) comparison plot of salt bridge ASP419-LYS1 in both complexes during the 600K simulation

ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506

Figure 6 Critical residues of the hydrogen bonds in the bindingdomain of the LRDPSO-docked result Residues are shown as balland stick models and the ligand is shown in new cartoon secondarystructure style +e residues are shown in different colors yellowfor the hydrogen bond network of LYS1 of the ligand red for thehydrogen bond network of TYR2 of the ligand and green for thehydrogen bond network of LYS3 of the ligand


bond with the receptor residues ASP419 TRP416 CYS417LEU504 and ASN506 and TYR2 could form a hydrogenbond with GLU32 VAL34 and ARG404 LYS3 formed ahydrogen bond with ARG413 HSD371 and GLY415 +esehydrogen bond networks played a positive role instrengthening the binding effect between the protein andligand

+e average number of hydrogen bonds in both com-plexes is listed in Table 4 As the simulation temperatureincreased the protein fold structure was generally weakenedand the structure also became more distorted +ese effectsresulted in a concomitant decrease in the number of hy-drogen bonds at high temperatures Figure 7 indicates thenumber of hydrogen bond changes over simulation time+e number of hydrogen bonds was maintained to a certainextent during the simulations An interesting finding is thatthe number of hydrogen bonds did not decrease dramati-cally with increases in the simulated temperature When thetemperature increased to 600K the loss of hydrogen bondswas comparatively higher than at other temperatures +isfinding also indicates that the disruption of tertiary struc-tural folds is extremely prominent at a temperature of 600K

At the same time the occupancy time of hydrogen bondsalso decreased as the simulation temperature increased +eoccupancy time of hydrogen bonds in both complexes is listedover the simulation temperature range of 300K to 600K inTables 2 and 3 +e types of hydrogen bonds in the twocomplexes are almost the same Only a few hydrogen bondsare different For the hydrogen bonds that were formed fromthe same two amino acids the occupancy time varied indifferent complexes For the LRDPSO docking complex fourhydrogen bonds had a high occupancy time (gt50) in the600K simulation However only two hydrogen bonds in theLGA docking complex had a high occupancy time

Table 2 Occupancy time of hydrogen bonds in the binding domain of the LRDPSO-docked complex at different temperatures

Donor Acceptor Occupancytime () (300K)

Occupancytime () (400K)



ARG413-side LYS3-side 100 100 100 100LYS1-main ASP419-side 100 100 9279 9239ARG404-side TYR2-side 100 7253 292 151ARG413-side LYS3-main 100 5461 761 6813LYS1-main TRP416-side 100 100 315 1351CYS417-main LYS1-main 8958 86 7786 6616LYS3-main GLY415-main 8935 837 6539 4606LYS1-side ASN506-side 8725 3796 319 246LYS1-main CYS417-main 8543 7636 3927 498TYR2-main GLU32-main 7958 5061 4556 2099LYS1-side ASP419-side 6658 6130 6463 3596VAL34-main TYR2-main 6390 6897 3423 3007LYS1-side LEU504-main 5915 2652 295 395LYS3-side HSD371-side 5233 061 016 104

Table 3 Occupancy time of hydrogen bonds in the binding domain of the LGA-docked complex at different temperatures





ARG413-side LYS3-side 100 100 100 8578ARG404-side TYR2-side 100 100 1158 58LYS1-main TRP416-side 100 100 6793 1429ARG413-side LYS3-main 100 100 9355 5171LYS1-main ASP419-side 9437 100 993 4638CYS417-main LYS1-main 9225 871 8022 4129LYS1-main CYS417-main 8562 6795 6167 2405LYS1-side ASP419-side 8387 8603 6223 3433LYS3-main GLY415-main 8048 6268 6162 308VAL34-main TYR2-main 7475 8047 2037 2776TYR485-side LYS3-side 7033 112 1612 34TYR2-main GLU32-main 6858 4452 584 219LYS1-side HSD161-side 6148 83 348 114

Table 4+e average number of hydrogen bonds in both complexesin different temperature simulations

ComplexAveragenumber(300K)

Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)

LRDPSO-dockedcomplex 208 200 182 155

LGA-docked complex 216 198 185 156


35 Structural Variation in Unfolding upon Ligand BindingUnder thermal stress protein local conformations usuallyundergo changes Due to the loss of interactions betweenresidues a regular secondary structure is often transformedinto an irregular secondary structure +e process ofunfolding under thermal pressure was not the same in thetwo complexes as these structures were obtained by differentdocking methods +ere are differences in structural vari-ation during unfolding Here an analysis of the time evo-lution of the secondary structure (Figure 8) can presentfurther structural variation information

Simulation reveals that the structures of both com-plexes are very stable when the simulation temperatureis 300 K In the case of the 400 K simulation onlyslight structural differences were observed for both com-plexes +ere was high similarity in the structures of thetwo complexes corresponding to the simulation resultsat 300 K and 400K For the LGA docking result thecomplex contained five α-helixes at residues VAL34-ASP42 PRO108-TYR115 ASP369-ILE376 TRP397-GLN406 and PRO423-ASN428 and it also contained fiveβ-sheets at residues PRO268-ILE277 LEU363-TYR365ASN394-GLU396 VAL411-CYS417 and ILE479-VAL486Compared to the LGA docking complex the LRDPSOdocking complex contained four α-helixes at residuesVAL34-ASP42 TYP112-GLN114 ASP369-ALA375 and

TRP397-GLN406 and it also contained six β-sheets at resi-dues PRO268-ILE277 LEU312-PRO313 LEU363-ASN366GLU393-GLN395 VAL411-CYS417 and ALA478-VAL486In the LRDPSO docking complex the structure of residuesPRO423-ASN428 was switched from an α-helix to a loopstructure relative to the LGA docking complex

+e structural fluctuations of both complexes weresignificantly more pronounced in the 500K simulation Atthe end of the simulation the LGA-docked complex con-tained four α-helixes and four β-sheets whereas theLRDPSO-docked complex contained four α-helixes (thesame as those in the LGA docking complex) and six β-sheetsCompared to the structures before the simulations some ofthe α-helixes and β-sheets were shortened among the regularsecondary structures in the high temperature simulationsFor regular secondary structures (α-helix and β-sheet)unfolding begins at the edges and associated turns becausethe center of these structures is mostly stronger than theiredges In addition the loops and turns begin to unfold tosome extent Despite these changes it appears that mostnative secondary structure elements remained present untilthe end of the simulation As shown in Figure 8 the structureof LGA-docked complex was looser than that of LRDPSO-docked complex due to unfolding at 500K simulation

At the higher temperature (600K) simulation the dom-inant structural change was that regular structures quickly

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)

Figure 7 Changes in the hydrogen bond number with respect to simulation time at four different temperatures Red represents theLRDPSO-docked complex and blue represents the LGA-docked complex (a) 300K (b) 400 K (c) 500K (d) 600K


became disordered which was accompanied by a loss ofmolecular contacts In the LRDPSO docking complexthree β-sheets (LEU270-GLU276 ALA412-CYS417 andTYR484-VAL486) and three helixes (ALA110-GLY116ALA375-ALA377 and PRO423-SER425) maintained stabil-ity until the end of simulation In the LGA docking complextwo β-sheets (TYR274-GLU276 and ALA412-ALA414) andfour helixes (SER107-TYR109 LEU370-ILE376 PRO423-ASN428 and ASN437-LYS442) were present Among thesehelixes ASN437-LYS442 was reformed by a loop after un-folding of the structure In other words the unfolding processgenerated new 3ndash10 helixes and α-helixes which originatedfrom areas that were initially coils and loops +e most im-portant point is that the ligand is completely out of thebinding position due to the unbinding of ligand and protein inLGA-docked complex

4 Conclusion

In this article molecular docking and molecular dynamicsimulations were performed to provide insights into thestructural and dynamic characteristics of OppA-peptidebinding interactions +e analysis results reflected the re-action differences between proteins and ligands in the twodockingmethods+e results showed that although the typesof hydrogen bonds in the two complexes were nearly thesame the occupancy time of the same hydrogen bonds wasdifferent in the different complexes For salt bridges therewere two significant salt bridges in the LRDPSO dockingresult which were ASP419-LYS1 and GLU229-LYS3 In theLGA-docked structure there was only one stable salt bridgeASP419-LYS1 located in the binding region Based on these

findings the electrostatic and van der Waals energy werelower for the ligand docked by LRDPSO than for the liganddocked by LGA For structural variation under thermalstress the complex docked by LRDPSO was more stablethan the complex docked by LGA at high temperatures+is study provided a concrete difference in OppA-peptidebinding based on the two docking methods It revealed thatnonbonded interactions are a significant driving force inbiomolecular interactions and stability It also showed thatthe contribution of electrostatic interactions is an importantfactor in binding affinity differences +is study provides auseful guide for drug design and protein engineering andfuture design studies with the OppA system

Data Availability

+e data used to support the findings of this study are in-cluded within the article

Conflicts of Interest

+e authors declare that they have no conflicts of interest

Acknowledgments

+is study was supported by the National Natural ScienceFoundation of China under Grant no 61502203 the NaturalScience Foundation of Jiangsu Province under Grant noBK20150122 the Natural Science Foundation of the JiangsuHigher Education Institutions of China under Grant no17KJB520039 and the 333 High-Level Talents TrainingProject of Jiangsu Province

300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)

Figure 8 Snapshots from the thermal unfolding simulations for both complexes (a) LRDPSO-docked complex (b) LGA-docked complex+e receptor protein structure is represented in a new cartoon secondary structure style and the ligand is represented in a Vdw style


References

[1] E Yuriev J Holien and P A Ramsland ldquoImprovementstrends and new ideas in molecular docking 2012-2013 inreviewrdquo Journal of Molecular Recognition vol 28 no 10pp 581ndash604 2015

[2] E Lopez-Camacho M J G Godoy J Garcıa-NietoA J Nebro and J F Aldana-Montes ldquoSolving molecularflexible docking problems with metaheuristics a compara-tive studyrdquo Applied Soft Computing vol 28 pp 379ndash3932015

[3] E Di Muzio D Toti and F Polticelli ldquoDockingApp a userfriendly interface for facilitated docking simulations withAutoDock Vinardquo Journal of Computer-Aided Molecular De-sign vol 31 no 2 pp 213ndash218 2017

[4] S F Sousa N M Cerqueira P A Fernandes andM J Ramos ldquoVirtual screening in drug design and devel-opmentrdquo Combinatorial Chemistry amp High roughputScreening vol 13 no 5 pp 442ndash453 2010

[5] J C Pereira E R Caffarena and C N D Santos ldquoBoostingdocking-based virtual screening with deep learningrdquo Journalof Chemical Information and Modeling vol 56 no 12pp 2495ndash2506 2016

[6] W H Shin J K Kim D S Kim and C Seok ldquoGalaxyDock2proteinndashligand docking using beta-complex and global op-timizationrdquo Journal of Computational Chemistry vol 34no 30 pp 2647ndash2656 2013

[7] L Heo W H Shin M S Lee and C Seok ldquoGalaxySiteligand-binding-site prediction by using molecular dock-ingrdquo Nucleic Acids Research vol 42 no 1 pp W210ndashW2142014

[8] L Guo Z Yan X Zheng L Hu Y Yang and J Wang ldquoAcomparison of various optimization algorithms of protein-ligand docking programs by fitness accuracyrdquo Journal ofMolecular Modeling vol 20 no 7 pp 2251ndash2261 2014

[9] V Bazgier K Berka M Otyepka and P Banas ldquoExponentialrepulsion improves structural predictability of moleculardockingrdquo Journal of Computational Chemistry vol 37 no 28pp 2485ndash2494 2016

[10] Y Fu Z G Chen and J Sun ldquoRandom drift particle swarmoptimisation algorithm for highly flexible protein-liganddockingrdquo Journal of eoretical Biology vol 457 pp 180ndash189 2018

[11] G M Morris D S Goodsell R S Halliday et al ldquoAuto-mated docking using a Lamarckian genetic algorithm andan empirical binding free energy functionrdquo Journal ofComputational Chemistry vol 19 no 14 pp 1639ndash16621998

[12] H J Yoon H J Kim B Mikami Y G Yu and H H LeeldquoCrystal structure of a putative oligopeptide-binding peri-plasmic protein from a hyperthermophilerdquo Extremophilesvol 20 no 5 pp 723ndash731 2016

[13] T Wang and R C Wade ldquoComparative binding energy(COMBINE) analysis of OppA-peptide complexes to relatestructure to binding thermodynamicsrdquo Journal of MedicinalChemistry vol 45 no 22 pp 4828ndash4837 2002

[14] S H Sleigh P R Seavers A J Wilkinson J E Ladbury andJ R Tame ldquoCrystallographic and calorimetric analysis ofpeptide binding to OppA proteinrdquo Journal of Molecular Bi-ology vol 291 no 2 pp 393ndash415 1999

[15] M Maurer S B de Beer and C Oostenbrink ldquoCalculation ofrelative binding free energy in the water-filled active site ofoligopeptide-binding protein Ardquo Molecules vol 21 no 4p 499 2016

[16] T I Chandel M Zaman M V Khan et al ldquoA mechanisticinsight into protein-ligand interaction folding misfoldingaggregation and inhibition of protein aggregates an over-viewrdquo International Journal of Biological Macromoleculesvol 106 pp 1115ndash1129 2018

[17] V Linkuviene V O Talibov U H Danielson and DMatulisldquoIntroduction of intrinsic kinetics of protein-ligand in-teractions and their implications for drug designrdquo Journal ofMedicinal Chemistry vol 61 pp 2292ndash2302 2018

[18] Y Fukunishi and H Nakamura ldquoStatistical estimation ofthe protein-ligand binding free energy based on directprotein-ligand interaction obtained by molecular dynamicssimulationrdquo Pharmaceuticals vol 5 no 10 pp 1064ndash10792012

[19] L J Colwell ldquoStatistical and machine learning approaches topredicting protein-ligand interactionsrdquo Current Opinion inStructural Biology vol 49 pp 123ndash128 2018

[20] S Lee and M G Barron ldquo3D-QSAR study of steroidal andazaheterocyclic human aromatase inhibitors using quantita-tive profile of protein-ligand interactionsrdquo Journal ofCheminformatics vol 10 no 1 p 2 2018

[21] W Becker K C Bhattiprolu N Gubensak and K ZanggerldquoInvestigating protein-ligand interactions by solution nuclearmagnetic resonance spectroscopyrdquo ChemPhysChem vol 19no 8 pp 895ndash906 2018

[22] F J Solis and R J B Wets ldquoMinimization by random searchtechniquesrdquoMathematics of Operations Research vol 6 no 1pp 19ndash30 1981

[23] G Jones P Willett R C Glen A R Leach and R TaylorldquoDevelopment and validation of a genetic algorithm forflexible dockingrdquo Journal of Molecular Biology vol 267 no 3pp 727ndash748 1997

[24] E J Gardiner P Willett and P J Artymiuk ldquoProtein dockingusing a genetic algorithmrdquo Proteins Structure Function andGenetics vol 44 no 1 pp 44ndash56 2001

[25] J Sun X Wu V Palade W Fang and Y Shi ldquoRandom driftparticle swarm optimization algorithm convergence analysisand parameter selectionrdquoMachine Learning vol 101 no 1ndash3pp 345ndash376 2015

[26] M Clerc and J Kennedy ldquo+e particle swarm-explosionstability and convergence in a multidimensional complexspacerdquo IEEE Transactions on Evolutionary Computationvol 6 no 1 pp 58ndash73 2002

[27] H M Chen B F Liu H L Huang et al ldquoSODOCK swarmoptimization for highly flexible protein-ligand dockingrdquoJournal of Computational Chemistry vol 28 no 2 pp 612ndash623 2007

[28] J C Phillips R Braun W Wang et al ldquoScalable moleculardynamics with NAMDrdquo Journal of Computational Chemistryvol 26 no 16 pp 1781ndash1802 2005

[29] A D MacKerell D Bashford M Bellott et al ldquoAll-atom empirical potential for molecular modelingand dynamics studies of proteins and nucleic acidsrdquoJournal of Physical Chemistry B vol 102 no 18pp 3586ndash3616 1998

[30] W Humphrey A Dalke and K Schulten ldquoVMD visualmolecular dynamicsrdquo Journal of Molecular Graphics vol 14no 1 pp 33ndash38 1996

[31] W L Jorgensen J Chandrasekhar J D MaduraR W Impey and M L Klein ldquoComparison of simple po-tential functions for simulating liquid waterrdquo Journal ofChemical Physics vol 79 no 2 pp 926ndash935 1983

[32] T Darden D York and L Pedersen ldquoParticle mesh Ewaldan N log (N) method for Ewald sums in large systemsrdquo


Journal of Chemical Physics vol 98 no 12 pp 10089ndash100921998

[33] J P Ryckaert G Ciccotti and H J C Berendsen ldquoNu-merical integration of the cartesian equations of motion of asystem with constraints molecular dynamics of n-alkanesrdquoJournal of Computational Physics vol 23 no 3 pp 327ndash3411977

[34] W G Hoover ldquoCanonical dynamics equilibrium phase-spacedistributionsrdquo Physical Review A vol 31 no 3 pp 1695ndash1697 1985


Stem Cells International

Hindawiwwwhindawicom Volume 2018


MEDIATORSINFLAMMATION

of

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Immunology ResearchHindawiwwwhindawicom Volume 2018

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine


Behavioural Neurology

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary andAlternative Medicine

Volume 2018Hindawiwwwhindawicom

Submit your manuscripts atwwwhindawicom

In this study oligopeptide binding protein (OppA) [12]was used as the experimental object OppA can act as areceptor for peptide transport across the cell membrane andis a potential target in antibacterial drug design [13] It hasbroad specificity and can bind to a wide range of peptidesthat are 2ndash5 amino acid residues long Here we used thesequence Lys-Tyr-Lys as the ligand binding structure +eOppA structure (PDB code 1B58 [14]) consists of threedomains domain I contains residues 1ndash44 189ndash269 and487ndash517 domain II contains residues 45ndash188 and domainIII contains residues 270ndash486 Domains I and II have abilobate structure allowing the ligand to act as a flexiblehinge connecting the two lobes In the protein-ligandcomplex structure ligands are completely contained inthe protein interior [15] For LRDPSO and LGA we selectedthe two best solutions for 1B58 in terms of binding energy+e lowest binding energies in LRDPSO and LGA wereminus2509 kcalmol and minus1274 kcalmol respectively

Protein-ligand interactions play an important role in mostbiological processes such as signal transduction cell regula-tion and immune response [16 17] Studying protein-ligandinteractions continues to be very important in life sciencefields [18ndash21] +ere are variations in protein-ligand complexstructures due to different docking methods In this article wemainly focus on analyzing the binding interactions between aprotein and a ligand especially on the divergence of protein-ligand interactions which can help us understand and addresskey questions such as those related to the diversity of bindingaffinity and specificity We have performed molecular dy-namic (MD) simulations on molecular docking results at fourdifferent temperatures (ranging from 300K to 600K) to es-tablish a more reliable mechanism for illustrating ligand-protein interactions +e dynamic properties of complexeshave been compared in terms of residue flexibility bindingenergy salt bridge and hydrogen bond interactions in thebinding region and structural variations during unfolding atdifferent temperatures +is study provides a better un-derstanding of the specific interactions predicted by differentdocking methods and it also allows us to more precisely studya binding site or region to increase docking accuracy

2 Materials and Methods

21 Protein and Ligand Structure Preparation +e OppAstructure (PDB code 1B58 [14]) which was obtained from theRSCB protein data bank (httpwwwrcsborg) was used as areceptor of the experimental object+e sequence Lys-Tyr-Lyswhich contains 43 atoms was used as the ligand structure+edocking results were used as models for the MD simulation

22 Genetic Algorithm (GA) AutoDock has been appliedwith great success in the prediction of binding conforma-tions of protein-protein interactions peptide-antibodycomplexes and enzyme-inhibitor complexes Earlier ver-sions of AutoDock used simulated annealing as a searchmethod a subsequent version added the options of a GA aLS method and a combination of GA and LS which is calledLGA In LGA a GA is used for global searching and Solis

and Wets [22] is used as the LS method Each generation isfollowed by a LS which is performed on a user-definedproportion of the population

Genetic algorithm [23 24] is a population-based meta-heuristic algorithm which contains initial population gen-eration fitness function evaluation iteration and terminationcondition check off four steps In addition every iteration stepincludes selection crossover and mutation operations +epseudocode of GA is described in Algorithm 1

23 HybridAlgorithm of RDPSOand Local Search (LRDPSO)RDPSO [25] is derived from canonical PSO trajectoryanalysis [26] and the free electron model Particle behaviorin RDPSO is assumed to be similar to an electron moving ina metal conductor in an external electric field Particlemovement is thus the superposition of thermal and driftmotions which is based on a global search and local searchof the particle respectively

In a RDPSO with M individuals each individual istreated as a volumeless particle in the N-dimensional spaceVin (V1

in V2in VN

in) and Xin (X1in X2

in XNin)

are expressed as the velocity vector and the position vector ofparticle i at the nth iteration respectively

According to the above model the updated equations ofRDPSO can be expressed by the following equation

Vjin+1 α C

jn minusX

jin

11138681113868111386811138681113868

11138681113868111386811138681113868ϕjin + β p

jin minusX

jin1113872 1113873 (1)

Xj

in+1 Xj

in + Vj

in+1 (2)

Cjn

1M

1113874 1113875 1113944

M

i1P

j

in (3)

for i 1 2 M j 1 2 N where αgt 0 is a parametercalled the thermal coefficient and βgt 0 is another parametercalled the drift coefficient ϕj

in is a random numberϕj

in sim N(0 1) pjin is the local attractor of PSO algorithm Pj

in

is the personal best (pbest) position of particle i Cjn is defined

by the mean of the pbest positions of all particles called themean best (mbest) position +e pseudocode of RDPSO isdescribed in Algorithm 2

+e hybrid of the RDPSO algorithm with the Solis andWets algorithm together form the LRDPSO algorithm Herethe RDPSO algorithm is used as a global search algorithmand the Solis and Wets algorithm is used as a local searchalgorithm +e Solis and Wets algorithm can facilitatetorsional space search since it does not require gradientinformation about the local energy landscape [22 27] +eaddition of local search effectively maintains the diversity ofparticles and prevents premature convergence of the algo-rithm +erefore the effective combination of the RDPSOalgorithm and the Solis and Wets algorithm can provide agood global search ability and a rapid convergence ability+e pseudocode of LRDPSO is described in Algorithm 3

24 e Docking Experiment Settings In molecular dockingexperiments the prepared protein and ligand structureswere saved in the PDBQT file format +e AutoDockTools







jin+1 α|Cj

n minusXjin|ϕj

in + β(pjin minusX

jin)

(12) Xj

in+1 Xj

in + Vj













PminusPbound minusV




(4)




V Wvdw 1113944ij

Aij

r12ij

minusBij

r6ij

⎛⎝ ⎞⎠ + WHbond 1113944ij

E(t)Cij

r12ij

minusDij

r10ij

⎛⎝ ⎞⎠

+ Welec 1113944ij

qiqj

ε rij1113872 1113873rij

+ Wsol 1113944ij


ij2σ21113872 1113873

(5)











00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(a)

ndash12ndash1


00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(b)


(a) (b)














0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)








02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of
























jin+1 α|Cj

n minusXjin|ϕj

in + β(pjin minusX

jin)

(12) Xj

in+1 Xj

in + Vj













PminusPbound minusV




(4)




V Wvdw 1113944ij

Aij

r12ij

minusBij

r6ij

⎛⎝ ⎞⎠ + WHbond 1113944ij

E(t)Cij

r12ij

minusDij

r10ij

⎛⎝ ⎞⎠

+ Welec 1113944ij

qiqj

ε rij1113872 1113873rij

+ Wsol 1113944ij


ij2σ21113872 1113873

(5)











00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(a)

ndash12ndash1


00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(b)


(a) (b)














0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)








02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of


























PminusPbound minusV




(4)




V Wvdw 1113944ij

Aij

r12ij

minusBij

r6ij

⎛⎝ ⎞⎠ + WHbond 1113944ij

E(t)Cij

r12ij

minusDij

r10ij

⎛⎝ ⎞⎠

+ Welec 1113944ij

qiqj

ε rij1113872 1113873rij

+ Wsol 1113944ij


ij2σ21113872 1113873

(5)











00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(a)

ndash12ndash1


00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(b)


(a) (b)














0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)








02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of
























00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(a)

ndash12ndash1


00204

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43

Ligand atom

Ener

gy (k

calm

ol)

LRDPSOLGA

(b)


(a) (b)














0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)








02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of























0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

300K400K

500K600K

(a)

300K400K

500K600K

0

5

10

15

20

25

30

35

17 27 37 54 112 161 202 229 247 268 278 306 316 332 370 382 400 410 420 430 440 479 489 506Residue number

RMSD

(Aring)

(b)








02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of























02468

101214

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

400K500K600K

(a)

400K500K600K

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

(b)

0

5

10

15

20

25

0 1000 2000 3000 4000 5000 6000 7000 8000Time (ps)

Dist

ance

(Aring)

LRDPSOLGA

(c)


ARG404

GLY415ARG413

GLU32

TRP416

ASP419

LEU504CYS417

VAL34

HSD371

ASN506




















Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of




































Averagenumber(400K)

Averagenumber(500K)

Averagenumber(600K)









0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of
























0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(a)

0

50

100

150

200

250

300

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(b)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(c)

0

50

100

150

200

250

0 1000 2000 3000 4000 5000 6000Time (ps)

Hyd

roge

n bo

nd n

umbe

r

LRDPSOLGA

(d)




4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of




















4 Conclusion



Data Availability




Acknowledgments


300K 400K 500K 600K

(a)

300K 400K 500K 600K

(b)



References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of



















References










































of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of



























of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of



















InsightsintotheMolecularMechanismsofProtein-Ligand...

Documents

Transcript of InsightsintotheMolecularMechanismsofProtein-Ligand...