Homology modeling, molecular docking and MD simulation studiesto investigate role of cysteine protease from Xanthomonas campestrisin degradation of Aβ peptide

Maruti J. Dhanavade b, Chidambar B. Jalkute b, Sagar H. Barage c, Kailas D. Sonawane a,b,n

a Structural Bioinformatics Unit, Department of Biochemistry, Shivaji University, Kolhapur 416004, Maharashtra (M.S.), Indiab Department of Microbiology, Shivaji University, Kolhapur 416004, Maharashtra (M.S.), Indiac Department of Biotechnology, Shivaji University, Kolhapur 416004, Maharashtra (M.S.), India

a r t i c l e i n f o

Article history:Received 6 July 2013Accepted 25 September 2013

Keywords:Cathepsin BCysteine proteaseAlzheimer diseaseHomology modelingMolecular dynamics

a b s t r a c t

Cysteine protease is known to degrade amyloid beta peptide which is a causative agent of Alzheimer'sdisease. This cleavage mechanism has not been studied in detail at the atomic level. Hence, a three-dimensional structure of cysteine protease from Xanthomonas campestris was constructed by homologymodeling using Geno3D, SWISS-MODEL, and MODELLER 9v7. All the predicted models were analyzedby PROCHECK and PROSA. Three-dimensional model of cysteine protease built by MODELLER 9v7 showssimilarity with human cathepsin B crystal structure. This model was then used further for docking andsimulation studies. The molecular docking study revealed that Cys17, His87, and Gln88 residues of cysteineprotease form an active site pocket similar to human cathepsin B. Then the docked complex was refined bymolecular dynamic simulation to confirm its stable behavior over the entire simulation period. The moleculardocking and MD simulation studies showed that the sulfhydryl hydrogen atom of Cys17 of cysteine proteaseinteracts with carboxylic oxygen of Lys16 of Aβ peptide indicating the cleavage site. Thus, the cysteineprotease model from X. campestris having similarity with human cathepsin B crystal structure may be used asan alternate approach to cleave Aβ peptide a causative agent of Alzheimer's disease.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Cathepsin B is a cysteine protease which carries out a varietyof physiological and pathological processes [1–3]. It degradesproteins which are entering endolysosomal system [4]. Amyloidplaque formation is the pathological hallmark in the Alzheimer'sdisease [5,6]. The Aβ (1–42) peptides are the main constituents ofamyloid plaques in Alzheimer's disease [7–9]. The familial auto-somal dominant (FAD) mutations and improper proteolytic degra-dation increase Aβ (1–42) peptides deposition [10]. The C-terminal

truncation of Aβ (1–42) peptides by cathepsin B reduces amyloidbeta peptide deposition. Hence, the enhanced activity of endo-genous cathepsin B may decrease Aβ (1–42) peptide level inAD [11–14]. The reduced level of endogenous inhibitor cystatinC has been found useful to increase cathepsin B activity [15].

The enzymes such as IDE, NEP, ECE, ACE, MMP-9, and plasmin havea role in the Aβ peptide clearance [16–21]. The abnormal distribu-tion of cathepsin B and its endogenous inhibitor leads to differentneurodegenerative pathologies [22]. The reduced cathepsin Bimmunoreactivity was observed due to increase in its inhibitorconcentration that leads to the Aβ peptide deposition [23]. There-fore, low levels of endogenous cathepsin B inhibitors (secretorycystatin C and cystatin B) may increase cathepsin B activity andclearance of amyloid plaques [24].

There are various enzymes in human brains that decrease theload of Aβ peptides but the factors such as enzymatic loss throughgenetic mutations or nongenetic reasons like direct oxidativedamage or enhanced production of inhibitors may result in toabnormal Aβ catabolism [25]. Thus, further studies to searchenzymes from novel bacterial sources having potential to degradeAβ peptides are essential. The amyloid-degrading ability of Natto-kinase enzyme from Bacillus subtilis Natto has been reportedearlier [26]. Similarly, aminopeptidase from Streptomyces griseus

K565 also showed Aβ peptide degradation activity [27] which

suggests that the bacterial enzymes may be used for Aβ peptideclearance. The present work has been carried out to investigate therole of bacterial enzymes to degrade Aβ peptide that may befurther used for drug development studies in AD research.

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/cbm

Computers in Biology and Medicine

0010-4825/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.compbiomed.2013.09.021

Abbreviations: AD, Alzheimer's disease; NEP, neprilysin; IDE, insulin degradingenzyme; ACE, angiotensin converting enzyme; ECE, endothelin converting enzyme;MMP-9, matrix metalloprotease-9; Aβ peptide, amyloid beta peptide; MD,molecular dynamics; RMSD, root mean square deviation

n Corresponding author at: Structural Bioinformatics Unit, Department ofBiochemistry, Shivaji University, Kolhapur 416004, Maharashtra (M.S.), India.Tel.: þ91 9881320719, þ91 231 2609153; fax: þ91 231 2692333.

E-mail addresses: kds_biochem@unishivaji.ac.in,kdsonawane19@gmail.com (K.D. Sonawane).

Computers in Biology and Medicine 43 (2013) 2063–2070

The three-dimensional model of the cysteine protease from abacterial source Xanthomonas campestris was developed by usingMODELLER 9v7 software. The steepest descent energy minimizationmethod was used to remove steric clashes. The structure wasvalidated through various online servers which confirm the structurequality. The sequence analysis and model comparison betweencysteine protease from X. campestris and human cathepsin B showedgood homology and resemblance in the active site pocket. Super-imposition of predicted structure of cysteine protease and humancathepsin B [28] showed a similar type of active site residues.Experimental results showed that the human cathepsin B cleavesAβ peptide at Lys16 [29]. Our molecular docking study revealed thatthe active site residue Cys17 of cysteine protease from X. campestris

interacts with Lys16 of Aβ peptide hence it may cleave Aβ peptidefrom carboxylic site of Lys16 similar to human cathepsin B [29].The molecular dynamic simulation of model and Aβ peptide complexverified stability of the docked complex. Therefore, the predictedcysteine protease model from X. campestris could be used to study Aβpeptide degradation activity which might be helpful to develop anew therapeutic approach for the AD patient's treatment.

2. Methods

2.1. Software and hardware

In order to get a good model of cysteine protease homologymodeling was done by using SWISS-MODEL [30], Geno3D [31] andMODELLER 9v7 [32]. The predicted models were evaluated by variousonline servers such as PROCHECK [33] for the Ramachandran plotquality evaluation, PROSA [34] for testing interaction energies andVERIFY-3D [35] for measuring the compatibility. A molecular dockingstudy was carried out between the best model of cysteine proteaseand Aβ peptide using AutoDock 4.2 [36]. The docked complex wasthen used for the MD simulation study. All the docking and simula-tion studies were done on HP workstation and Rack/Blade server.Interactive visualization and analysis of molecular structures werecarried out using chimera [37] and VMD [38].

2.2. Sequence alignment and homology modeling

The complete amino acid sequence of the cysteine proteasefrom X. campestris was retrieved from NCBI protein sequencedatabase (Accession no.—ZP_06488281.1). The BLAST programwas used to search suitable template available in the PDB. Thecrystal structure of human cathepsin B (PDB ID: 2IPP) [28] withresolution of 2.15 Å was used as a template to build a cysteineprotease model. The pairwise sequence alignment of the templatesequence (2IPP) and the cysteine protease sequence was doneusing the EMBOSS program [39].

Homology modeling of the cysteine protease was performed bythree homology modeling programs such as SWISS-MODEL [30],Geno3D [31] and MODELLER 9v7 [32]. For model building thealignment between template and query sequence was done by usingglobal dynamic programming with linear gap penalty functionavailable in MODELLER 9v7 [32]. The model was constructed bysatisfaction of spatial restraints, using its ‘automodel’ class. The modelwas generated using the python script command. The steepestdescent energy minimization was done to remove steric clashes [40].

2.3. Secondary structure analysis, model refinement and validation

Secondary structure analysis of the cysteine protease sequencewas done by Self-Optimized Prediction Method (SOPMA) Program[41]. The models obtained by SWISS-MODEL [30], Geno3D [31]

and MODELLER 9v7 [32] were validated by inspection of the Phi/Psi Ramachandran plot [42] obtained from PROCHECK analysis[33]. The model constructed by MODELLER 9v7 [32] was finallychosen for further investigations on the basis of geometry, 3Dalignment with the template and the results of PROCHECK [33]and PROSA [34] analyses. The PROSA [34] test was applied to checkthe energy criteria of predicted models in comparison with knownX-ray and NMR structures. The PROSA II energy plot was calcu-lated to check the interaction energies of all residues of thepredicted model. Further, the model quality was evaluated byVERIFY-3D [35]. The VERIFY-3D [35] compares model with itstemplate structure in terms of three-dimensional and sequencescore (3D–1D). The Ramachandran plot was obtained throughPROCHECK analysis [33]. The predicted model was then comparedwith a template structure (PDB ID: 2IPP) [28] by using thePDBeFOLD server [43] and chimera [37].

2.4. Active site prediction

The active site prediction in the refined model was done bysequence alignment. The sequences of cysteine protease fromX. campestris was aligned with sequence of human cathepsin B(PDB ID: 2IPP) [28] for which the active site is known. The 3Dalignment of cysteine protease and human cathepsin B (PDB ID:2IPP) was done by chimera [37] to compare the active site pocket.

2.5. Molecular docking of cysteine protease and Aβ peptide

Molecular docking has been performed between cysteine proteasemodel and patch of Aβ peptide (10YDVHHNKLVFF20) from 1AML.pdb[44] by using AutoDock 4.2. [36]. The docked complex was thenminimized by the steepest descent (SD) method using Gromacs 4.0.4to remove hard contacts [40]. The Lamarckian genetic algorithm(LGA) has been used in this study. The active site was defined usingAuto Grid [36]. The grid size was set to 40�44�46 points with agrid spacing of 0.375 Ǻ centered on the selected flexible residuespresent in the active site of cysteine protease. The grid box has theentire binding site of the cysteine protease and provides enoughspace to the Aβ peptide for the translation and rotation. Step size of2 Ǻ for translation and 501 for rotation were chosen and themaximum number of energy evaluation was set to 25,00,000. Thus,10 runs were performed and for every independent run a maximumnumber of 27,000 GA operations were generated on a single popula-tion of 150 individuals. Finally, docked complex with lower bindingenergy has been selected for further analysis.

2.6. Molecular dynamic simulation of predicted model

Molecular dynamic simulation was performed by Gromacs4.0.4 program [40] by using OPLS-AA force field [45,46]. Thecomplex of cysteine protease and patch of Aβ peptide10YDVHHNKLVFF20 was used for MD simulation. The complexwas surrounded by SPC water molecules (water density 1.0) andsystemwas neutralized with three Naþ ions. The cubic type of boxwas used for MD simulation. The MD simulation system has a totalof 41,600 atoms. The solvated structure was minimized by thesteepest descent method for 15,000 steps at 300 K temperatureand constant pressure. Then the complex was equilibrated for 2 nsperiod. After equilibration production MD was run for 20 ns atconstant temperature and pressure. The LINCS algorithm [47] wasused to constrain the bond length. Periodic boundary conditionswere applied to the system. The electrostatics interactions werecalculated using a PME algorithm with 12 Å cut off [48]. Structuralcomparison between initial structure and final structure of MDsimulation was done using PDBeFOLD [43]. The dynamic runs

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–20702064

were then visualized using the Visual Molecular Dynamics (VMD)package [38]. The images after simulation were generated by usingchimera [37] and PyMOL. Three-dimensional structure of cysteineprotease was analyzed by various softwares such as RasMol [49],PyMOL (http://PyMOL.sourceforge.net/) and Chimera [37].

3. Result and discussion

3.1. Secondary structure analysis

The secondary structure analysis of cysteine protease per-formed by SOPMA [41] revealed that 28%, 48%, 19%, and 5% aminoacid residues are present in a helix, random coil, extended strandand turns respectively (Supplementary data Fig. 1). This analysisshows that random coil occupied the largest part of the proteinfollowed by alpha helix, extended strand and beta turns forcysteine protease which makes the protein stable. The SOPMA[41] result showed that the target sequence (Accession no.—ZP_06488281.1) was good for homology modeling.

3.2. Model building and validation

As per the earlier report enzyme production is easier frombacterial source than any other sources [50]. Hence, the purifiedenzyme from bacterial sources showing Aβ peptide degradationactivity could be useful in studies to develop effective therapeuticstrategy against Alzheimer's disease [26,27]. Thus, we used cysteineprotease from X. campestris to investigate Aβ peptide degradationactivity. The target sequence of cysteine protease (Accession no.:ZP_06488281.1) was aligned with the crystallographic structure ofhuman cathepsin B (PDB: 2IPP) [28] (Fig. 1). The cysteine proteasemodel was built by SWISS-MODEL [30], Geno3D [31], and MODELLER9v7 [32]. The predicted model quality was evaluated by programs likePROCHECK [33], PROSA [34] and VERIFY-3D [35].

The PROCHECK [33] analysis of a cysteine protease model builtby MODELLER 9v7 [32] shows that 74% residues are found in most

favored regions, 20% residues in allowed regions and 4% residuesin generously allowed regions (Supplementary data Fig. 2). Hence,the total 98% of residues have been found in the favored regionsand only 2% of residues are observed in the outer region suggest-ing the good model quality (Table 1). In addition the PROSA [34]tool was used to check the model quality by the Z score of thestructure. The Z score is indicative of overall model quality and isused to check whether the target structure is within the range ofscores typically found in native proteins of similar size. The Z scorefor a cysteine protease model was �4.31. The results obtained byPROSA [34] tool showed that the cysteine protease structure waswithin the acceptable range of X-ray and NMR studies (Fig. 2A).The PROSA [34] analysis of the model gives most of the residues tohave negative interaction energy and very few residues displayedpositive interaction energy (Fig. 2B). Overall the PROCHECK [33]and PROSA [34] analyses showed good result of the model built byMODELLER 9v7 [32] hence it is used for further analysis (Table 1).

The model of cysteine protease built by MODELLER 9v7 [32]was analyzed by VERIFY-3D [35]. The VERIFY-3D [35] programwasused to find the atomic model (3D) compatibility with its ownamino acid sequence (1D). The compatibility score above zero inthe VERIFY-3D [35] graph corresponds to acceptability side-chainenvironments (Fig. 3) for a cysteine protease model built byMODELLER 9v7 [32]. In VERIFY-3D [35] the vertical axis representsthe average 3D–1D protein score for each residue in a 21 residuesliding window that would be helpful for further validation of themodel. The VERIFY-3D [35] result shows that the model score(0.29) was found to be near to the template (0.47) (Fig. 5). Thepredicted structure of cysteine protease from X. campestris byMODELLER 9v7 [32] when compared with a template structure(2IPP) [28] using PDBeFOLD [43] and chimera [37] showed goodsimilarity (Fig. 4). Lower root mean square deviation (RMSD) valuei.e. 0.82 Å for the superimposed structure of the cysteine proteaseand human cathepsin B was given by PDBeFOLD [43] suggestinggood similarity between them. The sequence alignment andstructural comparison by superimposition between cysteine pro-tease model and human cathepsin B (2IPP) [27] show that theactive site pocket forming residues (Cys17, His87, and Gln88)

Fig. 1. 2D alignment of the cysteine protease sequence and the template (2IPP.pdb). Catalytic residues are indicated by arrows and shown in red color. (For interpretation ofthe references to color in this figure legend, the reader is referred to the web version of this article.)

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–2070 2065

(Fig. 5A and B) of cysteine protease model are almost like that ofhuman cathepsin B active site residues (Cys29, His110 and His111)(Fig. 5C). The analysis of models generated by SWISS-MODEL [30],Geno3D [31] and MODELLER 9v7 [32] shows that model generatedby MODELLER 9v7 [32] has good result (Table 1). Thereforethe model built by MODELLER 9v7 [32] was used for moleculardocking studies with Aβ peptide.

Table 1Stereo-chemical quality by PROCHECK and model evaluation through PROSA formodels of cysteine protease obtained by different programs.

Predicted models ofcysteine protease bydifferent programs

PROCHECK analysis showing residues atvarious regions

PROSAZ-score

Core(%)

Allowed(%)

Generously(%)

Disallowed(%)

Geno3D Model 1 63.1 29.0 6.2 1.7 �4.49Geno3D Model 2 59.7 31.2 7.4 1.7 �4.16Geno3D Model 3 58.0 31.8 7.4 2.8 �4.74Geno3D Model 4 60.8 29.0 6.2 4.0 �4.49Geno3D Model 5 53.4 39.2 4.0 3.4 �4.16SWISS-MODEL 70.7 22.1 2.9 4.3 �3.73MODELLER 74 19.8 4.0 2.2 �4.31

Fig. 2. Validation of cysteine protease model using ProSA. The black dot shows thesimilarity of model with X-ray and NMR structures.

Fig. 3. The 3D profile verified results of cysteine protease model (red) and template(green). (For interpretation of the references to color in this figure legend, thereader is referred to the web version of this article.)

Fig. 4. (A) Homology model of cysteine protease from Xanthomonas campestris.(B) 3D alignment between cysteine protease model shown in cornflower blue colorand human cathepsin crystal structure (2IPP) as a template shown in hot pink color.(For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–20702066

3.3. Molecular docking and MD simulation stabilityof cysteine protease

Molecular docking and MD simulation studies have been founduseful in earlier studies to explore protein–ligand interactions[51,52]. In the present study the active site residues of cysteineprotease model (Cys17, His87, Gln88) were kept flexible and then

docked with patch of Aβ peptide [10YDVHHNKLVFF20] (1AML.pdb).This docked complex shows hydrogen bonding interactionsbetween the hydrogen atom of –SH group of active site residueCys17 of cysteine protease enzyme with backbone carboxyl oxygenatoms of Lys16 and Leu17 of Aβ peptide with interatomic distances2.34 Å and 2.76 Å respectively (Fig. 6A and Table 2). The dockedcomplex was then simulated for a period of 20 ns. The analysis ofthe structure obtained after 20 ns simulation strengthened hydro-gen bonding interaction between the active site residue Cys17 ofcysteine protease enzyme and Lys16 of Aβ peptide by interatomicdistance 1.74 Å throughout the MD simulation (Fig. 6B andTable 2). This suggests that cysteine protease from X. campestris

Fig. 5. (A) Active site pocket of cysteine protease showing cysteine residue (yellow)and histidine and glutamine residues (magenta) within pocket. (B) Active siteresidues of cysteine protease colored yellow. (C) Comparison between active siteresidues (cyan) of human cathepsin B crystal structure [2IPP] (hot pink) and activesite residue (yellow) of bacterial cysteine protease (cornflower blue). (For inter-pretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)

Fig. 6. Docked complex showing interactions between Cys17 of cysteine protease(yellow) with Lys16 and Leu17 of Aβ peptide (magenta) (A) before MD and (B) AfterMD. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

Table 2Interactions between active site residues of cysteine protease and amyloid betapeptide before and after MD simulation.

Interaction between Aβ peptide and activesite residues of cysteine protease before MD

Atomic distance (Å)

LYS16. O–CYS17. H 2.34LEU17. O–CYS17. HB1 2.76

Interaction between Aβ peptide and activesite residues of cysteine protease after MD

Atomic distance (Å)

LYS16. O–CYS17. HB1 1.74

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–2070 2067

could cleave Aβ peptide from carboxylic oxygen of Lys16 similar tothe cleavage site of human cathepsin B as reported experimentally[29] (Fig. 7A and B). However, the interaction between Cys17 ofcysteine protease and Leu17 of Aβ peptide was not maintainedduring MD simulation. The docking between human cathepsin Band Aβ peptide showed similar type of interactions between Cys29of human cathepsin B with Lys16 and Leu17 of Aβ peptide(Supplementary data Figs. 3 and 4).

The RMSD result shows stable behavior that can be seen fromFig. 8. The average RMSD value is 0.39 nm which indicates stablebehavior of cysteine protease and Aβ peptide complex over the20 ns simulation period (Fig. 8). The Radius of gyration (Rg) resultalso shows stable behavior during 20 ns simulation time (Fig. 9).The dynamics of the residues can be identified by RMSFs calcu-lated for the C-alpha atoms (Fig. 10). The RMSF results show smallfluctuations for the rigid structural elements and larger fluctua-tions for the ends and loops (Fig. 10). Generally, these results showthat the docked complex structure was stable over the entiresimulation time.

4. Conclusion

The present study revealed that cysteine protease fromX. campestris has an active site residues Cys17, His87, and Gln88similar to human cathepsin B (Cys29, His110, and His111) (Fig. 5C).The molecular docking and MD simulation studies showed thatthe hydrogen atom of sulfhydryl group of Cys17 of cysteineprotease forms hydrogen bonding interaction with backbonecarboxyl oxygen atom of Lys16 residue of Aβ peptide. This suggests

that the cysteine protease from X. campestris might cleave Aβ

Fig. 7. (A) Active site residue Cys17 (yellow) of cysteine protease interacting with Lys16 and Leu17 of Aβ peptide (magenta) before MD. (B) Active site residue Cys17 (yellow)of cysteine protease interacting with Lys16 of Aβ peptide (magenta) after MD. (For interpretation of the references to color in this figure legend, the reader is referred to theweb version of this article.)

Fig. 8. Showing RMSD of cysteine protease model.

Fig. 9. Showing Radius of gyration (Rg) of cysteine protease model.

Fig. 10. Showing RMSF of cysteine protease model.

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–20702068

peptide from carboxylic end of Lys16 similar to experimentalfindings of human cathepsin B [29]. Therefore, the predictedmodel of cysteine protease from bacterial source could be usefulin the studies of amyloid beta peptide degradation as well as todesign new therapeutic approach in the Alzheimer's disease.

5. Summary

Alzheimer's disease (AD) is a progressive neurodegenerativedisease that causes loss of cognitive function. Aβ peptides play acentral role in AD pathogenesis. Over production or inefficientclearance and defects in the proteolytic degradation results in Aβpeptide accumulation in AD brains. Hence, amyloid beta degrada-tion and clearance is a promising therapeutic approach in AD.Cathepsin B reduces Aβ levels by inducing C-terminal truncation

of Aβ (1–42). As cathepsin B enzyme is capable of degrading Aβpeptide. The enzyme similar to cathepsin B was found out frombacterial source. The bacterial cysteine protease from X. campestrissimilar to cathepsin B was targeted in this study. We build a three-dimensional structure of cysteine protease from X. campestris. Thepredicted structure of cysteine protease from X. campestris struc-ture shows active site pocket similar to human cathepsin B.Cysteine protease from X. campestris has active site residuesCys17, His87, and Gln88 similar to human cathepsin B (Cys29,His110, and His111). After that we docked the patch of Aβ peptidewith predicted cysteine protease structure. The active site residueCys17 of cysteine protease interacts with Lys16 of Aβ peptide

which suggests that it might play a role in Aβ peptide degradation.Thus, the predicted cysteine protease model might be useful infurther studies of Aβ peptide degradation as well as to design newnovel lead structures in the treatment of Alzheimer's disease.

Conflict of interest statement

All authors have no conflict of interest.

Acknowledgment

MJD is thankful to Department of Science and Technology, NewDelhi, for providing fellowship as research assistance under thescheme DST-PURSE. CBJ is grateful to Shivaji University, Kolhapur,for providing fellowship in the form of Departmental ResearchFellow (DRF). KDS is thankful to the Department of Biotechnology,New Delhi, for the financial support under the scheme DBT-IPLSsanctioned to Shivaji University, Kolhapur. Authors are thankful toComputer Centre, Shivaji University, Kolhapur, for providing thecomputational facility.

Appendix A. Supplementary material

Supplementary data associated with this article can be foundin the online version at http://dx.doi.org/10.1016/j.compbiomed.2013.09.021.

References

[1] M. Knop, H.H. Schijer, S. Rupp, D.H. Wolf, Vacuolar/lysosomal proteolysis:proteases, substrates, mechanisms, Current Opinion in Cell Biology 5 (1993)990–996.

[2] J.S. Mort, D.J. Buttle, Cathepsin BInternational Journal of Biochemistry and CellBiology 29 (1997) 715–720.

[3] M.Y. Khan, S.K. Agarwal, S. Ahmad, Structure-activity relationship in buffalospleen cathepsin B, Journal of Biochemistry 111 (1992) 732–735.

[4] H. Braak, E. Braak, Neuropathological stageing of Alzheimer-related changes,Acta Neuropathologica 82 (1991) 239–259.

[5] J. Hardy, D.J. Selkoe, The amyloid hypothesis of Alzheimer's disease: progressand roblems on the road to therapeutics, Science 297 (2002) 353–356.

[6] G.G. Glenner, C.W. Wong, Alzheimer's disease: initial report of the purificationand characterization of a novel cerebrovascular amyloid protein, Biochemicaland Biophysical Research Communications 120 (1984) 885–890.

[7] G.K. Gouras, H. Xu, J.N. Jovanovic, J.D. Buxbaum, R. Wang, P. Greengard,N.R. Relkin, S. Gandy, Generation and regulation of beta-amyloid peptidevariants by neurons, Journal of Neurochemistry 71 (1998) 1920–1925.

[8] C.L. Masters, G. Simms, N.A. Weinman, G. Multhaup, B.L. McDonald,K. Beyreuther, Amyloid plaque core protein in Alzheimer disease and Downsyndrome, Proceedings of the National Academy of Sciences of the UnitedStates of America 82 (1985) 4245–4249.

[9] R.E. Tanzi, L. Bertram, Twenty years of the Alzheimer's disease amyloidhypothesis: a genetic perspective, Cell 120 (2005) 545–555.

[10] D.J. Selkoe, Clearing the brain's amyloid cobwebs, Neuron 32 (2001) 177–180.[11] S. Mueller-Steiner, Y. Zhou, H. Arai, E.D. Roberson, B. Sun, J. Chen, X. Wang,

G. Yu, L. Esposito, L. Mucke, L. Gan, Antiamyloidogenic and neuroprotectivefunctions of cathepsin B: implications for Alzheimer's disease, Neuron 51(2006) 703–714.

[12] T. Saido, M.A. Leissring, Proteolytic degradation of amyloid b-protein, ColdSpring Harbor Perspectives in Medicine (2012) a006379–6397.

[13] C. Wang, B. Sun, Y. Zhou, A. Grubb, L. Gan, Cathepsin B degrades amyloid-β inmice expressing wildtype human amyloid precursor Protein, Journal ofBiological Chemistry 287 (2012) 39834–39841.

[14] J.S. Mort, D.J. Buttle, Molecules in focus: cathepsin B, International Journal ofBiochemistry and Cell Biology 29 (1997) 715–720.

[15] E.A. Eckman, D.K. Reed, C.B. Eckman, Degradation of the Alzheimer's amyloidbeta peptide by endothelin-converting enzyme, Journal of Biological Chem-istry 276 (2001) 24540–24548.

[16] N. Iwata, S. Tsubuki, Y. Takaki, K. Shirotani, B. Lu, N.P. Gerard, C. Gerard,E. Hama, H.J. Lee, T.C. Saido, Metabolic regulation of brain Aβ by neprilysin,Science 292 (2001) 1550–1552.

[17] W.Q. Qiu, D.M. Walsh, Z. Ye, K. Vekrellis, J. Zhang, M.B. Podlisny, M.R. Rosner,A. Safavi, L.B. Hersh, D.J. Selkoe, Insulin-degrading enzyme regulates extra-cellular levels of amyloid β-protein by degradation, Journal of BiologicalChemistry 273 (1998) 32730–32738.

[18] H.M. Tucker, M. Kihiko, J.N. Caldwell, S. Wright, T. Kawarabayashi, D. Price,D. Walker, S. Scheff, J.P. McGillis, R.E. Rydel, S. Estus, The plasmin system isinduced by and degrades amyloid-β aggregates, Journal of Neuroscience 20(2000) 3937–3946.

[19] K.J. Yin, J.R. Cirrito, P. Yan, X. Hu, Q. Xiao, X. Pan, R. Bateman, H. Song, F.F. Hsu,J. Turk, Matrix metalloproteinases expressed by astrocytes mediate extracel-lular amyloid-beta peptide catabolism, Journal of Neuroscience 26 (2006)10939–10948.

[20] K. Zou, H. Yamaguchi, H. Akatsu, T. Sakamoto, M. Ko, K. Mizoguchi, J.S. Gong,W. Yu, T. Yamamoto, K. Kosaka, Angiotensin converting enzyme convertsamyloid beta-protein 1–42 (Aβ1–42) to Aβ1–40, and its inhibition enhancesbrain Abeta deposition, Journal of Neuroscience 27 (2007) 8628–8635.

[21] N. Cimerman, M.T. Prebanda, B. Turk, T. Popovic, I. Dolenc, V. Turk, Interactionof cystatin C variants with papain and human cathepsins B, H and L, Journal ofEnzyme Inhibition 14 (1999) 167–174.

[22] K. Ii, H. Ito, E. Kominami, A. Hirano, Abnormal distribution of cathepsinproteinases and endogenous inhibitors (cystatins) in the hippocampus ofpatients with Alzheimer's disease, parkinsonism–dementia complex on Guam,and senile dementia and in the aged, Virchows Archiv A Pathological Anatomyand Histology 423 (1993) 185–194.

[23] A. Deng, M.C. Irizarry, R.M. Nitsch, J.H. Growdon, G.W. Rebeck, Elevation ofcystatin C in susceptible neurons in Alzheimer's disease, American Journal ofPathology 159 (2001) 1661–1668.

[24] B. Sun, Y. Zhou, B. Halabisky, I. Lo, S.H. Cho, S. Mueller-Steiner, X. Wang,A. Grubb, L. Gan, Cystatin C–Cathepsin B axis regulates amyloid beta levels andassociated neuronal deficits in an animal model of Alzheimer's disease,Neuron 60 (2) (2008) 247–257.

[25] D.S. Wang, D.W. Dickson, J. Malter, β-Amyloid degradation and Alzheimer'sdisease, Journal of Biomedicine and Biotechnology (2006) 1–12.

[26] R.L. Hsu, K.T. Lee, J.H. Wang, Y. Lily, L. Lee, P. Rita, Y. Chen, Amyloid-degradingability of nattokinase from Bacillus subtilis Natto, Journal of Agricultural andFood Chemistry 57 (2009) 503–508.

[27] C. Yoo, K. Ahn, J.E. Park, M.J. Kim, S.A. Jo, An aminopeptidase from Strepto-myces sp. KK565 degrades beta amyloid monomers, oligomers and fibrils,FEBS Letters 584 (2010) 4157–4162.

[28] D. Musil, D. Zucic, D. Turk, R.A. Engh, I. Mayr, R. Huber, T. Popovic1, V. Turk,T. Towatari, N. Katunuma, W. Bode, The refined 2.15 A X-ray crystal structureof human liver cathepsin B: the structural basis for its specificity, EMBOJournal 9 (1991) 2321–2330.

[29] E.A. Mackay, A. Ehrhard, M. Moniatte, C. Guenet, C. Tardif, C. Tarnus,O. Sorokine, B. Heintzelmann, C. Nay, J.M. Remy, J. Higaki, A.V. Dorsselaer,J. Wagner, C. Danzin, P. Mamont, A possible role for cathepsins D, E, and Bin the processing of P-amyloid precursor protein in Alzheimer's disease,European Journal of Biochemistry 244 (1997) 414–425.

[30] T. Schwede, J. Kopp, N. Guex, M.C. Peitsch, SWISS-MODEL: an automatedprotein homology-modeling server, Nucleic Acids Research 31 (2003)3381–3385.

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–2070 2069

[31] C. Combet, M. Jambon, G. Deléage, C. Geourjon, Geno3D: automatic compara-tive molecular modelling of protein, Bioinformatics 18 (2002) 213–214.

[32] A. Sali, T.L. Blundell, Comparative protein modelling by satisfaction of spatialrestraints, Journal of Molecular Biology 234 (1993) 779–815.

[33] R.A. Laskowaski, M.W. McArther, D.S. Moss, J.M. Thornton, PROCHECK aprogram to check sterio-chemical quality of a protein structures, Journal ofApplied Crystallography 26 (1993) 283–291.

[34] M. Wiederstein, M.J. Sippl, ProSA-web: interactive web service for therecognition of errors in three-dimensional structures of proteins, NucleicAcids Research 35 (2007) W407–W410.

[35] D. Eisenberg, R. Luthy, J.U. Bowie, VERIFY3D: assessment of protein modelswith three-dimensional profiles, Methods in Enzymology 277 (1997) 396–404.

[36] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson, AutoDock4 and AutoDockTools4: automated docking with selectivereceptor flexibility, Computional Chemistry Journal 30 (2009) 2785–2791.

[37] E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt,E.C. Meng, T.E. Ferrin, UCSF Chimera—a visualization system for exploratoryresearch and analysis, Journal of Computational Chemistry 25 (2004)1605–1612.

[38] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, Journalof Molecular Graphics 14 (1996) 33–38.

[39] P. Rice, I. Longden, A. Bleasby, EMBOSS: the european molecular biology opensoftware suite, Trends in Genetics 16 (6) (2000) 276–277.

[40] V.D. Spoel, E. Lindahl, B. Hess, G. Groenhof, A.E. Mark, H.J. Berendsen,GROMACS: fast, flexible, and free, Journal of Computational Chemistry26 (16) (2005) 1701–1718.

[41] C. Geourjon, G. Deléage, SOPMA: significant improvements in protein second-ary structure prediction by consensus prediction from multiple alignments,Computer Application in Biosciences 11 (6) (1995) 681–684.

[42] S.C. Lovell, I.W. Davis, W.B. Arendall, P.I.W. de Bakker, J.M. Word, M.G. Prisant,J.S. Richardson, D.C. Richardson, Structure validation by Calpha geometry: phi,psi and Cbeta deviation, Proteins: Structure, Function and Genetics 50 (2002)437–450.

[43] E. Krissinel, K. Henrick, Secondary-structure matching (SSM), a new tool forfast protein structure alignment in three dimensions, Acta Crystallography D60 (2004) 2256–2268.

[44] H. Sticht, P. Bayer, D. Willbold, S. Dames, C. Hilbich, K. Beyreuther, R.W. Frank,P. Rosch, Structure of amyloid A4-(1–40)-peptide of Alzheimer's disease,European Journal of Biochemistry 233 (1995) 293–298.

[45] W.L. Jorgensen, D.S. Maxwell, J. TiradoRives, Development and testing ofthe OPLS all-atom force field on conformational energetics and propertiesof organic liquids, Journal of the American Chemical Society 118 (1996)11225–11236.

[46] G.A. Kaminski, R.A. Friesner, J. Tirado-Rives, W.L. Jorgensen, Evaluation andreparametrization of the OPLS-AA force field for proteins via comparison withaccurate quantum chemical calculations on peptides, Journal of PhysicalChemistry B 105 (2001) 6474–6487.

[47] B. Hess, H. Bekker, H.J.C. Berendsen, J.G.E.M. Fraaije, LINCS: a linear constraintsolver for molecular simulations, Journal of Computational Chemistry 18(1997) 1463–1472.

[48] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, Asmooth particle mesh Ewald method, Journal of Chemical Physics 103 (1995)8577–8593.

[49] R.A. Sayle, E.J. Milner-White, RASMOL: biomolecular graphics for all, Trends inBiochemical Sciences 20 (9) (1995) 374.

[50] R. Jayaraj, P.M. Smooker, So you need a protein—a guide to the production ofrecombinant proteins, Open Veterinary Science Journal 3 (2009) 28–34.

[51] C.B. Jalkute, S.H. Barage, M.J. Dhanavade, K.D. Sonawane, Molecular dynamicssimulation and molecular docking studies of angiotensin converting enzymewith inhibitor lisinopril and amyloid beta peptide, Protein Journal 32 (2013)356–364.

[52] G. Tseng, K.D. Sonawane, Y.V. Korolkova, M. Zhang, J. Liu, E.V. Grishin, R.H. Guy,Probing the outer mouth structure of the HERG channel with peptidetoxin footprinting and molecular modeling, Biophysical Journal 92 (2007)3524–3540.

M.J. Dhanavade et al. / Computers in Biology and Medicine 43 (2013) 2063–20702070