Physicochemical properties of the modeled structure of astacin metalloprotease moulting enzyme...

Interdiscip Sci Comput Life Sci (2013) 5: 312–323

DOI: 10.1007/s12539-013-0182-9

Physicochemical Properties of the Modeled Structure of AstacinMetalloprotease Moulting Enzyme NAS-36 and Mapping the

Druggable Allosteric Space of Heamonchus contortus, Brugia malayiand Ceanorhabditis elegans via Molecular Dynamics Simulation

Om Prakash SHARMA1, Sonali AGRAWAL2,3, M. Suresh KUMAR1∗1(Centre of Excellence in Bioinformatics, School of Life Sciences, Pondicherry University, Pondicherry 605014, India)

2(Centre of Excellence in Bioinformatics, School of Biotechnology, Madurai Kamaraj University, Madurai 625021, India)3(Centre for Human Genetics, Bangalore 560100, India)

Received 7 June 2012 / Revised 6 December 2012 / Accepted 22 April 2013

Abstract: Nematodes represent the second largest phylum in the animal kingdom. It is the most abundantspecies (500,000) in the planet. It causes chronic, debilitating infections worldwide such as ascariasis, trichuriasis,hookworm, enterobiasis, strongyloidiasis, filariasis and trichinosis, among others. Molecular modeling tools canplay an important role in the identification and structural investigation of molecular targets that can act as a vitalcandidate against filariasis. In this study, sequence analysis of NAS-36 from H. contortus (Heamonchus contortus),B. malayi (Brugia malayi) and C. elegans (Ceanorhabditis elegans) has been performed, in order to identify theconserved residues. Tertiary structure was developed for an insight into the molecular structure of the enzyme.Molecular Dynamics Simulation (MDS) studies have been carried out to analyze the stability and the physicalproperties of the proposed enzyme models in the H. contortus, B. malayi and C. elegans. Moreover, the drugbinding sites have been mapped for inhibiting the function of NAS-36 enzyme. The molecular identity of thisprotease could eventually demonstrate how ex-sheathment is regulated, as well as provide a potential target ofanthelmintics for the prevention of nematode infections.Key words: Heamonchus contortus, Brugia malayi, Ceanorhabditis elegans, homology modeling, Molecular Dy-namics Simulation, NAS-36, principal component analysis (PCA).

1 Introduction

Lymphatic filariasis (LF) is a severely debilitating,disfiguring and stigmatizing disease ascribed as ne-glected tropical diseases (NTDs), caused by parasiticworms. It usually causes abnormal enlargement of thelimbs and the genitals (World Health Organization,2011). As per estimated 1.34 billion people in 81 coun-tries live in areas where filariasis is endemic and are atrisk of infection (World Health Organization, WHO).In 1993, following advances in diagnosis and treatmentthe Carter Centre’s International Task Force on DiseaseEradication classified LF as a “potentially eradicable”disease. Current drugs used for MDA implementationby national elimination program only clear microfilariaetemporarily without killing all adult worms (Bockarieand Deb, 2010). Thus the main problem regarding thechemotherapy of filariasis is that no safe and effectivedrug is available yet to combat the adult human filarial

∗Corresponding author.E-mail: [email protected]

worms (Ahmad and Srivastava, 2007). Therefore thereis an unequivocal call for the development of highly ef-ficient, complementary chemotherapeutical drug with amacrofilaricidal effect.

Extra cellular matrix (ECM), called the cuticle, beingtough but flexible exoskeleton maintains body shape,permits mobility via its attachments to muscles andprovides a protective barrier to the worm from externalenvironment (Page and Johnstone, 2007). The cuticle isa complex and versatile tissue made up of multiple lay-ers, having distinct levels of structural integrity. Theprincipal components of the cuticle are the collagens.Mutations in specific collagens and their processing en-zymes result in aberrant cuticle formation that leads todistinctive body shape phenotypes such as Dpy (dumpy,short and fat), Rol (roller), or Bli (blistered) (Theinet al., 2009). By investigating the life-cycle of worms,it came into light that the microfilariae molt twice incompetent insect vectors to become infective stage lar-vae (L3) that are infective to humans. L3 molt twice inthe human host and become adult worms that are re-

Interdiscip Sci Comput Life Sci (2013) 5: 312–323 313

productively active for years (Stepek et al., 2011). Themolt L3 to L4 stage, i.e. the degradation and escapeof third stage cuticle, is related to increased expres-sion of essential metalloprotease, suggesting a crucialrequirement of these enzymes in this process (Richeret al., 1992). The Nematode AStacin (NAS) metal-loprotease in free-living nematodes, C. elegans, havebeen identified to be essential to correct developmentand proper shedding of the cuticle (Davis et al., 2004).Particularly NAS-36, M12 family metalloprotease, en-codes functionally conserved enzyme, which are inde-pendently described as essential genes whose inactiva-tion disrupts molting (Stepek et al., 2011). NAS-36enzyme might also regulate the assembly of new cuticleby processing the precursors of particular extracellularmatrix proteins.

Recent advances including extensive genome search,sequence homology and complementation studies haveshown the characterized parasitic worms, Haemonchuscontortus and Brugia malayi are homologues of C. ele-gans (Stepek et al., 2011) particularly with regards tothe NAS-36 metalloprotease enzyme. This study hasdemonstrated that NAS-36 has a crucial function ofecdysis during cuticle development in free living as wellas in parasitic nematodes, therefore can be a key targetfor developing a newer control strategy with nematoci-dal effect.

Knowledge of the three-dimensional structures ofproteins opens the way to accelerate drug discovery andvaccine design (Sharma et al., 2011). Therefore in theabsence of crystal structure of NAS-36, we modeled thetertiary structure of the NAS-36 from H. contortus, B.malayi and C. elegans. The modeled structure was ver-ified using Molecular Dynamics Simulation (MDS) andvarious parameters to ensure the model structures arereliable and suitable for the further study. We have alsoidentified the druggable pockets for the target enzymein order to have a detailed insight into the molecularstructure of NAS-36.

2 Materials and methods

2.1 Protein sequence identification and analy-sis of NAS-36 from Haemonchus contor-tus, Brugia malayi and Caenorhabditis el-egans

Domain protein sequences of astacin metallopro-tease from Haemonchus contortus, Brugia malayi andCaenorhabditis elegans were searched and retrievedfrom NCBI GenBank (Benson et al., 2000). The ob-tained protein sequences of H. contortus, B. malayi andC. elegans were used as input sequences for Basic Lo-cal Alignment Search Tool (BLAST) and BLASTp wasperformed against the Protein Data Bank (PDB) usingour home made EpiBLAST tool (Sharma et al., 2012a)to find out the appropriate templates for the homol-

ogy modeling. Based on BLAST result, the most iden-tical and resembled crystal structure of proteins wasdownloaded from PDB for further analysis. ClustalX(Thompson et al., 2002) was used for the sequencealignment between the target and the template protein.

2.2 Homology modeling and structural analy-sis

The 3D structures of astacin metalloprotease from H.contortus, B. malayi and C. elegans were modeled usingeffective and comparative molecular modeling softwarenamed MODELLER9v9 (Sali et al., 1995). Best mod-eled structure was selected based on the Discrete Opti-mized Protein Energy (DOPE) scores (Eramian et al.,2006) defined by MODELLER. The modeled structureof the astacin metalloprotease was thoroughly analyzedwith the help of PDBsum generate (Laskowski, 2001).All the selected modeled structures were viewed andanalysed with the help of Chimera (Pettersen et al.,2004) and PyMOL (Grell et al., 2006) molecular visu-alization tools. Later, the structures were subjected tothe MDS using the GROMACS 4.5 (Van Der Spoel etal., 2005) package for the analysis of structural stabilityand flexibility.

2.3 Molecular Dynamics Simulations

The generated model of NAS-36 enzyme was usedfor performing MDS with the standard OPLS-AA (Op-timized Potentials for Liquid Simulations) force field(Xu et al., 2007). For MDS, we have used our earlierprotocol (Sharma et al., 2012b; Sharma et al., 2012c).The energy of the model protein was minimized us-ing the steepest descent approach realized in the GRO-MACS 4.5 package. Then, a 500 ps position restrain-ing simulation was carried out to restrain the proteinby a 1000 kJ/mol·A2 harmonic constraint to relieveclose contacts before the actual simulation. The Par-ticle Mesh Ewald (PME) method (Wang et al., 2010)was used for computing the long-range electrostatics, a1.4 A cutoff for van-der Waal (vdW) interactions, a 1.2A cutoff for Coulomb interaction with updates every10 steps, and the Linear Constraint Solver (LINCS)(Amiri et al., 2007) algorithm for covalent bond con-straints were used. The system was simulated at 300K.Berendsen’s temperature and pressure coupling methodwas used to regulate the temperature and pressure ofthe system. After restrained dynamics, the actual MDrun was set to 10 ns with the same parameters as men-tioned above. All analyses were performed using theprograms built within the GROMACS 4.5 and the coor-dinates were saved and analysed using XMGRACE soft-ware (http://rpmfind.net/linux/RPM/opensuse/11.4/x86 64/xmgrace-5.1.22-6.2.x86 64.html). The lowestpotential energy conformations were selected from the8ns MDS trajectory and further refined by energyminimization. The refined models were validated us-

314 Interdiscip Sci Comput Life Sci (2013) 5: 312–323

ing the Structural Analysis and Verification Server(http://nihserver.mbi.ucla.edu/SAVES/), which usesvarious tools, including PROCHECK (Amiri et al.,2007), ERRAT (Colovos and Yeates, 1993), and VER-IFY 3D (Eisenberg et al., 1997).

2.4 Root Mean Square Deviation and RootMean Square Fluctuations

To analyze the stability of the modeled structureof H. contortus, B. malayi andC. elegans Root MeanSquare Deviation was calculated using g rms command.Protein model consists of flexible as well as stiff re-gion. The flexibility of the proteins was computed bythe GROMACS 4.5 utility g rmsf.

2.5 Solvent Accessible Surface Area

Change in Solvent Accessible Surface Area (SASA)was calculated using a g sas command in GROMACS4.5. It computes hydrophobic, hydrophilic, and totalsolvent accessible surface areas (SAS) of a given so-lute as if it were represented alone in bulk solvent. Anincrease in SAS would represent the unfolding of theprotein to expose previously buried residues whereasdecrease in SAS would not represent contact betweenthe protein and surface with the subsequent exclusionof water but rather a compaction of the protein’s struc-ture with a reduction in its surface area.

2.6 Mapping druggable allosteric space

NMR and X-ray crystallographic structures often re-veal well defined binding pockets that accommodateendogenous ligands. However, sometimes the modelsproduced by these crystal structures hide other po-tential druggable sites which may arise due to thechange in the protein conformation. These bindingsites are called cryptic binding sites (Durrant and Mc-Cammon, 2011; Ivetac and McCammon, 2010). Inthis study, FTMAP (http://ftmap.bu.edu/) (Brenkeet al., 2009) was used to identify the cryptic bindingsites. Based on the fast Fourier transform correlation,FTMAP identifies the regions that can bind small probemolecules. For each structure, it performs rigid bodydocking with 16 small molecules as probes (ethanol, iso-propanol, isobutanol, acetone, acetaldehyde, dimethylether, cyclohexane, ethane, acetonitrile, urea, methy-lamine, phenol, benzaldehyde, benzene, acetamide andN , N -dimethylformamide). All the best 2000 poses foreach probe are retained for further minimization andrescoring. It uses CHARMM potential with the Analyt-ical Continuum Electrostatistics (ACE) model. Duringthe minimization it adopts Newton–Raphson methodto fix the protein atoms, while the other atoms of theprobe molecules are free to move. Further, it groups allthe minimized probe conformations into clusters usinga simple greedy algorithm (Tuffery et al., 2005). Con-sensus sites were determined based on the overlapping

clusters of different probes. The probes in the resultingset of CSs were used to describe the binding site.

3 Results and discussion

3.1 Molecular modeling of NAS-36 astacinmetalloprotease

The protein sequences of the astacin metalloproteasefrom H. contortus (Accession No. ACZ64272.1), B.malayi (Accession No. ACZ64273.1), and C. elegans(Accession No. NP 492109.2) were retrieved from theNCBI protein database and BLAST search was per-formed against the PDB database. Based on the E-value and sequence identities of BLASTp result, thecrystallographic structures of Crayfish astacin (PDBID: 3LQ0, Resolution 1.45 A) (Guevara et al., 2010),Zn endopeptidease astacin (PDB ID: 1IAA, Resolution1.90 A) (Gomis-Ruth et al., 1994) and Astacus asta-cus (PDB ID: 1AST, Resolution 1.80 A) (Bode et al.,1992) proteins were selected as templates for the homol-ogy modeling of the respective proteins therefore down-loaded from the PDB database. The template and thetarget sequences of H. contortus, B. malayi and C. el-egans showed 55%, 55% and 57% similarity and 38%,38% and 40% identity with the e-value of 2e-19, 2e-21 and 2e-22 respectively. Thus, the sequence similar-ity between NAS-36 from H. contortus, B. malayi andC.elegans with 3LQ0, 1IAA and 1AST is reasonablefor selecting them as suitable templates for homologymodeling.

3.2 Molecular structure of astacin metalopro-tease

Our secondary structure prediction of H. contortus,B. malayi and C. elegans reveals six strands, eight he-lices, one β-hairpin (Ile108-Val110: Leu159-Thr161),eight γ-turns in H. contortus; six strands, six helices,one β-hairpin (Ile105-Val107:Leu157-Thr159), four γ-turns in B. malayi while seven strands, seven helices,two β-hairpin (Tyr59-Ser60: Arg68-Ser72 and Val106-Ile108: Val157-Thr159) and seven γ-turns in C. elegansthroughout the structures (Fig. 1). Total number ofhelices, strands and γ-turns has been summarized inthe Table 1. Our tertiary structure analysis has shownthat the astacin from H. contortus, B. malayi and C.elegans adopts a kidney shape, with a deep active sitecleft between its N-terminal and C-terminal domain.The sequence analysis of these three nematodes indi-cates that catalytic astacin domain (Pfam PF01400)of these nematodes also contains crucial Zn bindingsite (HExxHxxGFxHExxRxDRD) and the conservedmethionine-turn (SxMHY) (Stepek et al., 2011; Gue-vara et al., 2010; Bode et al., 1993; Stocker et al., 1995),which suggests the structural similarity between thesenematodes (Fig. 2).


H2

H7H5

S6 H6

S5H4

H8 H4 H6

H5 H3

S6

S1S3

S2

S4

H1H2

H3

S2

S1

S3S4

S5

H2H1

H4S6 S7S5

Beta-hairpin Beta-hairpin Beta-hairpin

Beta-hairpin

C-terminal C-terminal C-terminalN-terminal N-terminal N-terminal

S2

S1S3

S4H3

H1

(a) (b) (c)

Fig. 1 Three dimensional structures of homology models (a) H. contortus, (b) B. malayi and (c) C. elegans. All secondarystructure elements are labeled. “H” stands for helices and “S” stands for strands. The figure was prepared usingPyMol molecular visualization tool.

Table 1 Secondary structure analysis of H. contortus, B. malayi and C. elegans using PDBsum

SourceStrands Helices Turns

Start End Sequence Start End Sequence Start End Sequence

H. contortus Pro7 Phe12 PIKYRF Phe18 Val34 FYAVSNIIKAIRYWENV Thr4 Ala6 TTA

Glu38 Glu40 EFE Glu77 Cys79 ENC Asp47 Glu49 DNE

Phe51 Phe55 FIEFF Ala82 Gly94 AGVIEHEIGHAIG Glu49 Phe51 EDF

Gln71 Ser74 QGVS Ala104 Tyr107 AQSY Cys79 Lys81 CVK

Ile108 Val110 IKV Ser112 Phe114 SDF Asp122 Leu124 DFL

Leu159 Thr161 LIT Pro117 Tyr119 PSY Lys128 Ile130 KDI

Ile170 Gln172 IGQ Leu133 Leu135 LGL

Phe178 Ala187 FLDIETINKA Thr161 Asp163 TRD

B. malayi Pro6 Ile7 PI Ile17 Asn33 ILHISQILKALEIWQSN Glu43 Ser45 EAS

Tyr9 Phe11 YRF Thr81 Gly91 TIIHEVGHTLG Cys76 Arg78 CER

Lys37 Phe38 KF Glu95 Ser97 EQS Tyr164 Met166 YQM

Ile49 Phe51 IEF Pro99 Ala101 PDA Tyr186 Lys188 YCK

Ile105 Val107 ITV Ser118 Phe120 SEF

Leu157 Thr159 LLT Tyr177 Ala18 YNIKLINEA

C. elegans Pro6 Phe11 PIKYRF Phe17 Ser33 FYTISQIIAAIRFWEDS Ser42 Ser44 SDS

Thr37 Asn40 TFEN Met80 Leu91 MGVIEHEIGHAL Cys77 Lys79 CVK

Tyr49 Phe53 YIEFF Ala120 Tyr105 ALGY Ile108 Arg110 IER

Tyr59 Ser60 YS Arg110 Phe112 RDF Leu122 Arg124 LQR

Arg68 Ser72 RQGIS Ser119 Leu122 SDFL Arg124 Asp126 RDD

Val106 Ile108 VTI Ile168 Gln170 IGQ Asp126 Ile128 DEI

Val157 Thr159 VVT Phe176 Ala185 FYDVATINTA Asp151 Lys153 DQK

3.3 Molecular Dynamics Simulation study ofH. contortus, B. malayi and C. elegans

In order to evaluate the stability of the homologymodels, molecular dynamic calculations were carriedout for 10 ns. The average PDB structure was down-loaded from the system and has shown the tertiarystructure of the proteins with temperature factors. Ourinvestigation has revealed that most of the protein re-gion of H. contortus, B. malayi and C. elegans hascool region, which suggests their structural stability(Fig. 3). During the MDS we observed that one smallhelix (SDF, 112-114) had disappeared between 1 ns to3 ns, and then again reformed and been maintaining

the structural conformation (Fig. 4). In B. malayi andC. elegans one β-hairpin sheet (B. malayi-LLT: 157-159and C. elegans-VVT: 157-159) has shown more fluctu-ation. In case of C. elegans, it has disappeared duringthe simulation period. All this conformation changesare marginal throughout the simulation, which suggestsits equilibrated conformation.

3.4 Model structure analysis and verification

The lowest potential energy (Fig. 5) of the pro-tein trajectory was analyzed to download the most es-tablished conformation of the modeled protein. Wefound that H. contortus, B. malayi and C. elegans hadlowest potential energy (PE) of −991331.68 kJ/mol,


HaemonchusCaenorhabditisBrugia



757372

149147147

191189189

. . . . . . . 160 . . . . . . . 170 . . . . . . . 180 . . . . . . . 180 . .

. . . 80 . . . . . . . 90 . . . . . . . 100 . . . . . . . 110 . . . . . . . 120 . . . . . . . 130 . . . . . . . 140 . . . . . . . 150

1 . . . . . . . 10 . . . . . . . . 20 . . . . . . . . 30 . . . . . . . . 40 . . . . . . . . 50 . . . . . . . . 60 . . . . . . . . 70 . . . . .

HEXXHXXGFXHEXXRXDRD SXMHY

191189189

. . . . . . . 160 . . . . . . . 170 . . . . . . . 180 . . . . . . . 180 . .

. . . 80 . . . . . . . 90 . . . . . . . 100 . . . . . . . 110 . . . . . . . 120 . . . . . . . 130 . . . . . . . 140 . . . . . . . 150

1. . . . . . . 10 . . . . . . . . 20 . . . . . . . . 30 . . . . . . . . 40 . . . . . . . . 50 . . . . . . . . 60 . . . . . . . . 70 . . . . .

HEXXHXXGFXHEXXRXDRD SXMHY

Fig. 2 Multiple sequence alignment of the astacin metalloprotease from H. contortus, C. elegans and B. malayi is shownin the figure. Amino acid sequence was aligned using ClustalX 2.1. Here, asterisks single denote fully conservedresidues, colons denote fully conserved strong groups, and periods denote fully conserved weaker groups.

N-terminalN-terminal

N-terminal

C-terminal(a) (c)(b)

C-terminal C-terminal

Hot regionCool region

Intermediate region

N-terminalN-terminal

N-term

C-terminal(a) (c)(b)

C-terminal C-termina

Hot regionCool region

Intermediate region

Fig. 3 Temperature factor of the average structure of modeled H. contortus, B. malayi and C. elegans is shown in the figures((a), (b) and (c)) respectively.

−1071257.25 kJ/mol and −1021674.62 kJ/mol respec-tively. It suggests that at this specified PE, the mod-eled structure is in the most stable conformation. Cor-responding protein structures were retrieved from thePDB database for further studies.

3.5 Root mean square deviation

The conformational flexibility of the modeled NAS-36 from H. contortus was estimated by RMSD values.The RMSD profile of H. contortus has exhibited mi-nor fluctuation of 0.1443 to 0.2910 nm, which suggeststhat the RMSD trajectory of NAS-36 enzyme from H.contortus is fairly stable throughout the MDS trajec-

tory period (Fig. 6(a)). RMSD profile of B. malayihas shown 0.1625 to 0.4289 nm fluctuation. Initially,it has shown minor deviation from 0.1625 to 0.300 nmand fluctuations up to 1800 ps and then tries to sta-bilize it up to 2500 ps. After 2500 ps once again itshows the sudden fluctuation and then maintains theRMSD throughout the MDS trajectory. C. elegans hasshown minor RMSD fluctuation of 1591 to 0.2705 nmthroughout the simulation trajectories. It is stabilizedafter 5000 ps and maintains it throughout the simula-tion trajectories. The RMSD graph (Fig. 6(a)) suggeststhat H. contortus, B. malayi and C. elegans have stableconformations.


Initial 1 ns 3 ns

Initial

Initial

LLT (157-159)

1 ns

7 ns

7 ns

1 ns

C.elegans

B.malayi

H.contortus

3 ns

5 ns 10 ns

5 ns 10 ns

3 ns

10 ns7 ns5 ns

VVT (157-159)

Initial 1 ns 3 ns

Initial

Initial

LT (157-159)

1 ns

7 ns

7 ns

1 ns

B.malayi

H.contortus

3 ns

5 ns 10 ns

5 ns 10 ns

3 ns

10 ns7 ns5 ns

VT (157-159)))))))

Fig. 4 Conformations of H. contortus, B. malayi and C.elegans during the MD simulation experiments for1000 ps.

H.contortusB.malayiC.elegans

−9.9E+05

−1.02E+06

−1.05E+06

Pot

enti

al e

nerg

y (k

J/m

ol)

−1.08E+060 2000 4000 6000

Time (ps)8000 10000


Fig. 5 Evaluation of the trajectory 10,000 ps for the cal-culation of potential energy.

3.6 RMS fluctuation

RMS fluctuations of the backbone of Cα atoms aredepicted in the Fig. 6(b). RMS fluctuation study has

shown that the most flexible residues are located in theloops which connect the β-sheets to α-helices. It hasbeen noticed that γ-turn residues of H. contortus andB. malayi are slightly fluctuated. In H. contortus, atthe 128th position aspartic acid (Asp) and at the 129th

position lysine (Lys) there are slight fluctuations from0.3867 nm to 0.5556 nm (Asp128) and once again thereare decreases from 0.4992 nm (Lys129) to 0.3922 nm(Asp130). In case of B. malayi, at the 74th positionlysine, fluctuations from 0.1589 nm to 0.3858 nm andCys(76) from 0.2616 nm to 0.3288 nm of Glu(77) wereobserved. Overall residues fluctuated in the minimalrange of 0.0801 nm to 0.3858 nm. In case of C. elegans,there is no such remarkable fluctuation observed.

3.7 Radius of gyration

Radius of gyration (Rg) was calculated for H. contor-tus, B. malayi and C. elegans as shown in the Fig. 6(c).Initially the Rg value of H. contortus was 1.76 nm andit was stable up to the 2500 ps and thereafter therewas continuously negligible rise and fall of the Rg valuewas observed up to ∼8000 ps. It achieves the equilib-rium state at ∼8000 ps and maintains throughout thetrajectory. B. malayi and C. elegans both have startedat the Rg value of 1.725 nm. Initially, Rg value of B.malayi rises but at ∼6200 ps it stabilizes while C. ele-gans has stable Rg at ∼2500 ps to 8200 ps and then itsRg slightly rises. Overall the Rg value fluctuates from1.67 nm to 1.87 nm, which is negligible (0.27 nm). Ourresult clearly suggests that there are moderate confor-mational changes during the simulation period.

3.8 Solvent Accessible Surface Area (SASA)

The number of water molecules in contact with pro-tein residues is represented by the Solvent AccessibleSurface Area (SASA). It presents the changes in solventaccessible surface area (SAS). No significant change isseen in the SAS for any of the systems. This indicatesthat none of the protein surfaces causes the protein tounfold and/or spread out over the surface during thecourse of the simulation. Our investigation has shownthat 1662 out of 2971 atoms were classified as hydropho-bic in H. contortus but there are more hydrophilicresidues in the surface area of the protein (Fig. 6(d)).In B. malayi (Fig. 6(e)) and C. elegans (Fig. 6(f)),hydrophilicity of protein is maintained throughout thesimulation and varies from 78 to 73 nm2 in B. malayiand 70 to 77 nm2 area in C. elegans. In B. malayi,1709 out of 3009 atoms were classified as hydropho-bic whereas 1655 out of 2948 atoms were classified ashydrophobic in C. elegans, but they also have more hy-drophilic residues in surface area like H. contortus. OurMDS data for SASA suggests that H. contortus, B.malayi andC. elegans have more hydrophilic residuesin surface area, which made these proteins more hy-drophilic in nature. But this property does not affect





HydrophobicHydrophilicTotal



0.5

0.4

0.3

0.2

0.1

0

RM

SD (

nm)

RM

S fluc

tuat

ions

(nm

)

0 2000

(a) (b)

(c) (d)

(e) (f)

4000 6000 8000 10000Time (ps) Residue number

0

0.6

0.5

0.4

0.3

0.2

0.1

050 100 150 200

Rad

ius

of g

yrat

ion

(nm

)

0 2000

1.90

1.85

1.80

1.75

1.70

1.65

1.604000 6000 8000 10000

Time (ps)0 2000 4000 6000 8000 10000

Time (ps)

180

165

150

135

120

105

90

75

60

45

30

15

0

180

165

150

135

120

105

90

75

60

45

30

15

0

Are

a (n

m2 )

165

150

135

120

105

90

75

60

45

Are

a (n

m2 )

Are

a (n

m2 )

0 2000 4000 6000 8000 10000Time (ps)

0 2000 4000 6000 8000 10000Time (ps)

76-78 (γ-turn) 128-130(γ-turn)

B.malayiC.elegans

B.malayiC.elegans

76-78 (γ-turn) 128-130(γ-turn)

g

Fig. 6 Evaluation of the trajectory 10 000 ps for the calculation of (a) RMSD, (b) RMSF and (c) Radius of gyration. (d),(e) and (f) represent the solvent accessible surface for H. contortus, B. malayi and C. elegans respectively.

the stability of these proteins.

3.9 Principal component analysis

Principal component analyses (PCA) of the mod-eled structures were performed and diagonalized basedon a trajectory of 5001 frames with covariance matrixof 1719 frames for H. contortus, 1692 for B. malayiand 1701 for C. elegans. The dimension of the co-variance matrix is given by 573 backbone atoms forH. contortus, 564 for B. malayi and 567 for C. ele-gans, with eigenvalue sums of 14.9994 nm2, 11.8252

nm2 and 6.17144 nm2, respectively. Ploting eigenvaluesagainst the eigenvectors of each protein yielded steepfor eigenvalues curve (Fig. 7). Our investigation hasshown that most of the backbone motion was belowthe first 20 eigenvectors. To obtain the detail featureof the protein motion through the eigenvectors values,we projected the trajectory onto these individual eigen-vectors. The motion and projection of trajectory hasbeen shown in Fig. 8 through the Essential Dynamicsanalysis. Figs. 8(a), 8(b) and 8(c) show the motionsalong the first two, fifth, tenth, fifteenth and twentieth


Table 2 Ramachandran plot calculation on 3D models of H. contortus, B. malayi and C. elegans computed withthe PROCHECK program

Protein source Most favored region Additional allowed region Generously allowed region Disallowed region

H. contortus 77.4% 20.2% 2.4% 0.0%

B. malayi 81.9% 15.7% 2.4% 0.0%

C. elegans 86.1% 13.3% 0.6% 0.0%

Table 3 VERIFY 3D, ERRAT analysis and overall G-factor of the modeled structures of H. contortus, B.malayi and C. elegans

Protein source Verify 3D ERRAT Overall G factor

H. contortus 81.77% 93.642 0.64

B. malayi 83.07% 91.124 0.61

C. elegans 77.89% 85.976 −0.59


6

5

4

3

2

1

0

Eig

enva

lue

(nm

2 )

0 20 40 60Eigenvaector index

Eigenvalues of the covariance matrix

80 100

Fig. 7 Eigenvalues for H. contortus, B. malayi and C. el-egans are shown in decreasing order of magnitudeand obtained from the backbone coordinate covari-ance matrix as a function of the eigenvector index.

eigenvectors obtained from the backbone co-ordinatecovariance matrix for H. contortus,B. malayi and C. el-egans for the 10 ns simulation respectively. Figs. 8(d),8(e), and 8(f) projects the trajectory onto the planesdefined by the tenth and twentieth eigenvectors fromthe backbone co-ordinates matrix for H. contortus, B.malayi and C. elegans respectively. The eigenvectorprojections of each protein onto the plane of the back-bone motions are correlated and are in expected range.The result indicates that there is no high projection ob-served far from the diagonal. This data suggests thatthe modeled structures are exhibiting prolonged stabil-ity throughout the 10 ns simulation period.

The structural analysis and verification server haveshown that the modeled structures of H. contortus, B.malayi andC. elegans have overall good structural ge-ometry. None of the residues is found in the disallowedregion (Table 2). ERRAT tool was used to calculate thestatistics of non-bonded interactions between differentatom types and plot the value of the error function ver-sus position of a 9-residue sliding window, by a compari-

son with statistics from highly refined structures (Table3). The data suggests that the modeled structures ofH. contortus is a high resolution quality model whereasthe model structures of B. malayi and C. elegans haveaverage overall quality factors with approximately res-olution of 2.5 A to 3.0 A and Verify-3D structure anal-ysis has shown that 81.77%, 83.07% and 77.89% of theresidues have an average 3D-1D score higher than 0.2respectively (Table 3). The result strongly suggeststhat our modeled structures are appropriate for anyfurther study in future.

3.10 Druggable allostaric binding sites

It has been shown by several investigators that dueto the different receptor backbone conformation lig-and may bind to the newly formed binding sites whichare often difficult to identify through visual or com-putational examination of previously solved structures(Frembgen-Kesner and Elcock, 2006; Pargellis et al.,2002). A recent example is a crystal structure of p38

MAP kinase, where due to the 10 A movement ofPhe169, side chain creates a new drug binding site(Frembgen-Kesner and Elcock, 2006). This crypticdrug binding site may play an important role in thestructure based drug design through the allosteric inhi-bition of the protein activity. We have identified sevencryptic drug binding sites in H. contortus, ten in B.malayi and eleven in C. elegans. The details of thesecryptic binding sites are shown in Fig. 9 and summa-rized in Table 4. The inhibition of this molting enzymeNAS-36 may block the normal cuticle ecdysis in freeliving parasitic nematodes.

4 Conclusions

Astacin metalloprotease enzyme (NAS-36) plays animportant role in the cuticle ecdysis in C. elegans andit also has crucial role in the development of H. con-tortus and B. malayi. Therefore, it represents an ideal


0 2000

(a) (b)

(c) (d)

(e) (f)

4000 6000 8000 10000Time (ps)

6

4

2

0

−2

−4

−60 2000 4000 6000 8000 10000

Time (ps)

6

4

2

0

−2

−4

−6

0 2000 4000 6000 8000 10000Time (ps)

6

4

2

0

−2

−4

−6

1.0

0.5

0

−0.5

−1.0

ev_1ev_2ev_5ev_10ev_15ev_20

ev_1ev_2ev_5

−2 −1 0 1 2Projection on 10 th eigenvector (nm)

Pro

ject

ion

on 2

0 th

eig

enve

ctor

(nm

)P

roje

ctio

n on

20

th e

igen

vect

or (

nm)

Pro

ject

ion

on 2

0 th

eig

enve

ctor

(nm

)

1.0

0.5

0

−0.5

−1.0−2 −1 0 1 2

Projection on 10 th eigenvector (nm)

Pro

ject

ion

on e

igen

vect

ors

(nm

)

Pro

ject

ion

on e

igen

vect

or (

nm)

Pro

ject

ion

on e

igen

vect

or (

nm)

1.0

0.5

0

−0.5

−1.0−2 −1 0 1 2

Projection on 10 th eigenvector (nm)

ev_10ev_15ev_20

ev_1ev_2ev_5

ev_10ev_15ev_20

ev_2ev_5ev_10ev_15ev_20

Fig. 8 Motion and projection of trajectory is shown through essential dynamics analysis. (a), (b) and (c) show the motionsalong the first two, fifth, tenth, fifteenth and twentieth eigenvectors obtained from the backbone coordinate covariancematrix for H. contortus, B. malayi and C. elegans respectively. In (d), (e) and (f), projection of the trajectory ontothe planes is defined by the tenth and twentieth eigenvectors from the backbone coordinates matrix for H. contortus,B. malayi and C. elegans respectively.

drug target for the parasitic nematodes and antifilar-ial drugs. In the present study, homology models ofthe astacin domain from the three closely related ne-matodes namely H. contortus, B. malayi and C. ele-gans were generated to get an insight into the molecularstructure of the NAS-36 enzyme. Molecular dynamicsstudy was performed for 10 ns to examine the physicalproperties and the structural stability of the enzyme.Our MDS has revealed that all the modeled structureis exhibiting marginal conformation changes. In H. con-tortus, motif helix (ENC, 77-79) is providing structuralstability to maintain its tertiary structure. The modelshave been verified using Ramachandran plot, ERRAT,

Verify 3D etc., which suggests that our modeled struc-ture is reliable. We have mapped the drug binding sites.As to our knowledge, there is no published work regard-ing the physical properties and the drug binding sitesof the NAS-36 enzyme. We hope that our models willinspire new experimental efforts in this area, specifi-cally, the drug binding sites identified by our in silicoapproach can be useful for the suitable inhibitor design.

Acknowledgements

Om Prakash Sharma is grateful to the Coun-cil of Scientific & Industrial Research (CSIR),


C-terminalC-terminal

N-terminalN-terminalSite 1

Site 4Site 10

Site 9

Site 3Site 7

Site 5

Site 1

Site 7Site 3

Site 6

Site 4

(a) H.contortus (b) B.malayi (c) C.elegans

Site 2

Site 6Site 8

Site 5Site 2 Site 5 Site 2

Site 1

Site 3

Site 6Site 9

Site 4Site 11

N-terminalC-terminal

Site 7

Site 10Site 8

Fig. 9 (a), (b) and (c) define the cryptic drug binding regions of H. contortus, B. malayi and C. elegans. (a), (b), and (c)are shown in cartoon format, and cryptic binding site is shown in mesh format, the respective protein is in solidsurface format. The figure was built using PyMOL.

Table 4 Cryptic binding site of H. contortus, B. malayi and C. elegans

Protein source Active site residues at cryptic binding site

H. contortus Site 1: Asp180, Thr183, Glu98, Ser141, Trp96

Site 2: Gly171, Ile170, Gly146

Site 3: His144, Tyr145, Ser168, Arg173, Val142, Leu139, Ser141

Site 4: Val84, Glu88

Site 5: His97

Site 6: Tyr145, Glu88, Ser62

Site 7: Asp122

B. malayi Site 1: Tyr186, Leu90

Site 2: Gly80, Thr81, Arg171

Site 3: Tyr116, Glu119

Site 4: Trp2, Leu90

Site 5: Gly169, His142, Val140,

Site 6: Phe148

Site 7: His88, Leu92

Site 8: Glu85, His88

Site 9: Glu127

Site 10: Thr129, Phe131, Val133

C. elegans Site 1: His95, Glu96, Gln97, Tyr143

Site 2: His142, Gly144, Phe148, Ile168

Site 3: Gln97, Asp120, Phe121, Gln123

Site 4: Trp94, Asp129, Leu131, Gly132, Ile133, Ser139

Site 5: Ser76, Val82, Glu86

Site 6: Ser98, His95,Asp125

Site 7: Arg171,Asp178, Val82, Glu84

Site 8: Cys58, Ser60, Glu86,His95

Site 9: Leu131, Gly132, Ile133, Glu96

Site 10: Val140, Tyr177

Site 11: Leu91


India for the Senior Research Fellowship (SRF)(09/559/(0085)/2012/EMR-I). The research carried onin the laboratory of the Centre for Excellence in Bioin-formatics, Pondicherry University, India was funded bythe Department of Information Technology (DIT) andthe Department of Biotechnology (DBT), Governmentof India, New Delhi, India.

References

[1] Ahmad, R., Srivastava, A.K. 2007. Biochemical com-position and metabolic pathways of filarial wormsSetaria cervi: Search for new antifilarial agents. JHelminthol 81, 261-280.

[2] Amiri, S., Sansom, M.S., Biggin, P.C. 2007. Molec-ular dynamics studies of AChBP with nicotine andcarbamylch-oline: The role of water in the bindingpocket. Protein Eng Des Sel 20, 353-359.

[3] Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Os-tell, J., Rapp, B.A., Wheeler, D.L. 2000. GenBank.Nucl Acid Res 28, 15-18.

[4] Bockarie, M.J., Deb, R.M. 2010. Elimination of lym-phatic filariasis: Do we have the drugs to complete thejob? Curr Opin Infect Dis 23, 617-620.

[5] Bode, W., Gomis-Ruth, F.X., Huber, R., Zwilling, R.,Stocker, W. 1992. Structure of astacin and implicationsfor activation of astacins and zinc-ligation of collage-nases. Nature 358, 164-167.

[6] Bode, W., Gomis-Ruth, F.X., Stockler, W. 1993.Astacins, serralysins, snake venom and matrix met-alloproteinases exhibit identical zinc-binding environ-ments (HEXXHXXGXXH and Met-turn) and topolo-gies and should be grouped into a common family, the’metzincins’. FEBS Lett 331, 134-140.

[7] Brenke, R., Kozakov, D., Chuang, G.Y., Beglov, D.,Hall, D., Landon, M.R., Mattos, C., Vajda, S. 2009.Fragment-based identification of druggable ’hot spots’of proteins using Fourier domain correlation tech-niques. Bioinformatics 25, 621-627.

[8] Colovos, C., Yeates, T.O. 1993. Verification of proteinstructures: Patterns of nonbonded atomic interactions.Protein Sci 2, 1511-1519.

[9] Davis, M.W., Birnie, A.J., Chan, A.C., Page, A.P.,Jorgensen, E.M. 2004. A conserved metalloproteasemediates ecdysis in Caenorhabditis elegans. Develop-ment 131, 6001-6008.

[10] Durrant, J.D., McCammon, J.A. 2011. Molecular dy-namics simulations and drug discovery. BMC Biol 9,71.

[11] Eisenberg, D., Luthy, R., Bowie, J.U. 1997. VER-IFY3D: Assessment of protein models with three-dimensional profiles. Methods Enzymol 277, 396-404.

[12] Eramian, D., Shen, M.Y., Devos, D., Melo, F., Sali,A., Marti-Renom, M.A. 2006. A composite score forpredicting errors in protein structure models. ProteinSci 15, 1653-1666.

[13] Frembgen-Kesner, T., Elcock, A.H. 2006. Computa-tional sampling of a cryptic drug binding site in a pro-tein receptor: Explicit solvent molecular dynamics andinhibitor docking to p38 MAP kinase. J Mol Biol 359,202-214.

[14] Gomis-Ruth, F.X., Grams, F., Yiallouros, I., Nar, H.,Kusthardt, U., Zwilling, R., Bode, W., Stocker, W.1994. Crystal structures, spectroscopic features, andcatalytic properties of cobalt(II), copper(II), nickel(II),and mercury(II) derivatives of the zinc endopeptidaseastacin. A correlation of structure and proteolytic ac-tivity. J Biol Chem 269, 17111-17117.

[15] Grell, L., Parkin, C., Slatest, L., Craig, P.A. 2006. EZ-Viz, a tool for simplifying molecular viewing in Py-MOL. Biochem Mol Biol Educ 34, 402-407.

[16] Guevara, T., Yiallouros, I., Kappelhoff, R., Bissdorf,S., Stocker, W., Gomis-Ruth, F.X. 2010. Proenzymestructure and activation of astacin metallopeptidase. JBiol Chem 285, 13958-13965.

[17] Ivetac, A., McCammon, J.A. 2010. Mapping the drug-gable allosteric space of G-protein coupled recep-tors: A fragment-based molecular dynamics approach.Chem Biol Drug Des 76, 201-217.

[18] Laskowski, R.A. 2001. PDBsum: Summaries and anal-yses of PDB structures. Nucl Acid Res 29, 221-222.

[19] Laskowski, R.A., Rullmannn, J.A., MacArthur, M.W.,Kaptein, R., Thornton, J.M. 1996. AQUA andPROCHECK-NMR: Programs for checking the qualityof protein structures solved by NMR. J Biomol NMR8, 477-486.

[20] Page, A.P., Johnstone, I.L. 2007. The cuticle. Worm-Book 19, 1-15.

[21] Pargellis, C., Tong, L., Churchill, L., Cirillo, P.F.,Gilmore, T., Graham, A.G., Grob, P.M., Hickey, E.R.,Moss, N., Pav, S., Regan, J. 2002. Inhibition of p38MAP kinase by utilizing a novel allosteric binding site.Nat Struct Biol 9, 268-272.

[22] Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch,G.S., Greenblatt, D.M., Meng, E.C., Ferrin, T.E. 2004.UCSF Chimera - a visualization system for exploratoryresearch and analysis. J Comput Chem 25, 1605-1612.

[23] Richer, J.K., Sakanari, J.A., Frank, G.R., Grieve, R.B.1992. Dirofilaria immitis: Proteases produced by third-and fourth-stage larvae. Exp Parasitol 75, 213-222.

[24] Sali, A., Potterton, L., Yuan, F., van Vlijmen, H.,Karplus, M. 1995. Evaluation of comparative proteinmodeling by MODELLER. Proteins 23, 318-326.

[25] Sharma, O.P., Jadhav, A., Hussain, A., Kumar,M.S. 2011. VPDB: Viral Protein Structural Database.Bioinformation 6, 324-326.

[26] Sharma, O.P., Das, A.A., Krishna, R., Kumar, M.S.,Mathur, P.P. 2012a. Structural Epitope Database(SEDB): A web-based database for the epitope, andits intermolecular interaction along with the tertiarystructure information. J Proteomics Bioinform 5, 084-089.


[27] Sharma, O.P., Pan, A., Hoti, S.L., Jadhav, A., Kan-nan, M., Mathur, P.P. 2012b. Modeling, docking, sim-ulation, and inhibitory activity of the benzimidazoleanalogue against β-tubulin protein from Brugia malayifor treating lymphatic filariasis. Med Chem Res 21,2415-2427.

[28] Sharma, O.P., Vadlamudi, Y., Liao, Q., Strodel, B.,Muthuvel, S.K. 2012c. Molecular modeling, dynamics,and an insight into the structural inhibition of cofactorindependent phosphoglycerate mutase isoform 1 fromWuchereria bancrofti using cheminformatics and mu-tational studies. J Biomol Struct Dyn 31, 765-778.

[29] Stepek, G., McCormack, G., Birnie, A.J., Page, A.P.2011. The astacin metalloprotease moulting enzymeNAS-36 is required for normal cuticle ecdysis in free-living and parasitic nematodes. Parasitology 138, 237-248.

[30] Stocker, W., Grams, F., Baumann, U., Reinemer,P., Gomis-Ruth, F.X., McKay, D.B., Bode, W. 1995.The metzincins - topological and sequential relationsbetween the astacins, adamalysins, serralysins, andmatrixins (collagenases) define a superfamily of zinc-peptidases. Protein Sci 4, 823-840.

[31] Thein, M.C., Winter, A.D., Stepek, G., McCormack,G., Stapleton, G., Johnstone, I.L., Page, A.P. 2009.Combined extracellular matrix cross-linking activity of

the peroxidase MLT-7 and the dual oxidase BLI-3 iscritical for post-embryonic viability in Caenorhabditiselegans. J Biol Chem 284, 17549-17563.

[32] Thompson, J.D., Gibson, T.J., Higgins, D.G. 2002.Multiple sequence alignment using ClustalW andClustalX. Curr Protoc Bioinformatics 23, 1-22.

[33] Tuffery, P., Guyon, F., Derreumaux, P. 2005. Improvedgreedy algorithm for protein structure reconstruction.J Comput Chem 26, 506-513.

[34] Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof,G., Mark, A.E., Berendsen, H.J. 2005. GROMACS 4.5:Fast, flexible, and free. J Comput Chem 26, 1701-1718.

[35] Wang, H., Dommert, F., Holm, C. 2010. Optimizingworking parameters of the smooth particle mesh Ewaldalgorithm in terms of accuracy and efficiency. J ChemPhys 133, 034117.

[36] World Health Organization. 2011. Working to over-come the global impact of neglected tropical diseases -Summary. Wkly Epidemiol Rec 86: 113-120.

[37] Xu, Z., Luo, H.H., Tieleman, D.P. 2007. Modifying theOPLS-AA force field to improve hydration free ener-gies for several amino acid side chains using new atomiccharges and an off-plane charge model for aromaticresidues. J Comput Chem 28, 689-697.

Physicochemical properties of the modeled structure of astacin metalloprotease moulting enzyme...

Documents

Transcript of Physicochemical properties of the modeled structure of astacin metalloprotease moulting enzyme...