Computers in Biology and Medicine 43 (2013) 1334–1340

Contents lists available at ScienceDirect

Computers in Biology and Medicine

0010-48http://d

n CorrBiotecnBrazil. T

E-m

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

Analysis of the structure of calpain-10 and its interactionwith the protease inhibitor SNJ-1715

Ronaldo Correia da Silva a,b, Nelson Alberto N. de Alencar a,c, Cláudio Nahum Alves a,Jerônimo Lameira a,d,n

a Laboratório de Planejamento de Fármacos, Instituto de Ciências Exatas e Naturais, Universidade Federal do Pará, CEP 66075-110 Belém, PA, Brazilb Programa de Pós-Graduação em Genética e Biologia Molecular, Universidade Federal do Pará, CEP 66075-110 Belém, PA, Brazilc Programa de Pós-Graduação em Química, Universidade Federal do Pará, CEP 66075-110 Belém, PA, Brazild Instituto de Ciências Biológicas, Universidade Federal do Pará, CEP 66075-110 Belém, PA, Brazil

a r t i c l e i n f o

Article history:Received 28 February 2013Accepted 11 July 2013

Keywords:Calpain-10Homology modellingQM/MMMDInhibitor

25/$ - see front matter & 2013 Elsevier Ltd. Ax.doi.org/10.1016/j.compbiomed.2013.07.010

esponding author at: Universidade Federal doologia, Rua Augusto Corrêa, 01 – Guamá, 6607el.: +55 091 32018235.ail addresses: lameira@ufpa.br, jerolameira@y

a b s t r a c t

Calpain-10 (CAPN10) is a cysteine protease that is activated by intracellular calcium (Ca2+) and known tobe involved in diseases such as cancer, heart attack, and stroke. A role for the CAPN10 gene in diabetesmellitus type II was recently identified. Hyper activation of the enzyme initiates a series of destructivecycles that can cause irreversible damage to cells. The development of inhibitors may be useful astherapeutic agents for a number of calpainopathies. In this paper, we have used the homology modellingtechnique to determine the 3D structure of calpain-10 from Homo sapiens. The model of calpain-10obtained by homology modelling suggests that its active site is conserved among family members andthe main interactions are similar to those observed for μ-calpain. Structural analysis revealed that thereare small differences in the charge distribution and molecular surface of the enzyme. These differencesare probably less dependent on calcium for calpain-10 than they are for μ-calpain. In addition, the ionpair Cys�/His+ formation was observed using of Molecular Dynamics (MD) simulations that were basedupon hybrid quantum mechanical/molecular mechanical (QM/MM) approaches. Finally, the binding ofthe SNJ-1715 inhibitor to calpain-10 was investigated in order to further understand the mechanism ofinhibition of calpain-10 by this inhibitor at the molecular level.

& 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Calpains are non-lysosomal cysteine proteases [1] that areactivated by intracellular calcium (Ca2+). They convert increasesin the concentration of this ion into proteolytic signals andparticipate in various signal transduction pathways [2]. The 14calpain isoforms, which are characterised based on their func-tional domains, are found in multiple tissues [3].

Calpain-10, encoded by the CAPN10 gene, is present in alltissues, including those involved in diabetes, such as the liver,pancreas, and skeletal muscle [4]. It is found in abundance in themitochondria, and is involved in various processes such as apop-tosis and age related diseases, as well as the storage and release ofCa2+ [5]. An increase in the concentration of mitochondrial Ca2+

under different pathophysiological conditions can initiate a seriesof destructive cycles that can cause irreversible damage to cells.

ll rights reserved.

Pará, Faculdade de5-110 Belém, Pará,

ahoo.com.br (J. Lameira).

Therefore, the overexpression of this enzyme is responsible formitochondrial dysfunction [6].

In general, changes in calpain activity, due to either geneticchanges [7] or uncontrolled proteolysis resulting from overexpres-sion of the enzyme, results in diseases such Alzheimer's disease[8], type II diabetes [9], and some types of cancer [10,11].Strategies for the regulation of the proteolytic activity of calpainsare necessary, and include blocking the catalytic site of thebiomolecule. SNJ-1715 is a potent inhibitor of other calpain iso-forms with IC50 values ranging between 86 and 190 nm [12]. Thisinhibitor can exist in a free aldehyde form [13–15] or in a cyclichemiacetal form [16] (Fig. 1).

Homology modelling can be used to build the tertiary structureof a protein based on the primary structure. This technique, whichis based on the premise that protein structures are evolutionarilyconserved, provides structural information important for under-standing the function of biomolecules [17,18]. Recently, we haveused protein homology modelling [19], including the hybridQuantum Mechanics/Molecular Mechanics (QM/MM) method[20,21] and Molecular Dynamics (MD) [22], to study differentbiomolecular systems [23,24]. In this study, we have determinedthe structure of human calpain-10A through homology modelling.

Fig. 1. The reversible intramolecular conversion of the cyclic hemiacetal form ofSNJ-1715 to the free aldehyde form is shown.

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–1340 1335

In addition, we have investigated the interaction between calpain-10 and the aldehyde form of the inhibitor SNJ-1715 using hybridQM/MM and molecular dynamics simulations.

2. Methods

2.1. Molecular modelling of calpain-10

The human calpain-10 amino acid sequence was obtained fromthe National Center for Biotechnology Information (NCBI) usingthe access code AAH04260 (GenBank). The sequence was identi-fied as one of the 46 members of the C2 family of proteasesaccording to the MEROPS database [1]. Subsequently, a stretch of315 sequence residues representing the functional domain wasidentified, isolated, and aligned in the PDB server via the BLASTalignment tool. The 3D structure selected as a template forconstructing the model of calpain-10 was μ-calpain from Rattus-norvegicus (PDB access code: 2G8E), which has a length of 326amino acid residues and a resolution of 2.2 Å by X-ray crystal-lography [16]. It has 36.73% identity and 50% similarity with thetarget sequence. The model of calpain-10 was validated based onthe stereochemical quality in the MolProbity server [25,26]. Thequality of folding was checked using Verify3D [27], and the systemenergy and quality of 3D alignment were monitored using AtomicNon-Local Environment Assessment (ANOLEA) [28,29] and RootMean Square Deviation (RMSD) [30], respectively.

2.2. Molecular dynamics simulations (MD)

The hybrid QM/MMmolecular dynamics simulations have becomethe method of choice for modelling reactions and interactions inbiomolecular systems [31]. In this report, the initial coordinates for theQM/MM MD calculations were taken from the crystal structure of theμ-calpain-SNJ-1715 complex and the model of human calpain-10 wasobtained by molecular homology modelling. Since the standard pKavalues of ionisable groups can be shifted by the local proteinenvironment, an assignment of the protonation states of all of theseresidues at pH 7 was carried out. The pKa values of the amino acidsresidues were determined with PROPKA 2.0, assuming the pH to be 7[32]. As a result, most residues were found at their standard protona-tion state and all protonated residues are located away from the activesite and did not strongly influence the quantum calculations.

The hydrogen atoms of the protein were added and theirpositions were relaxed by means of successive minimisation stepscombining different search algorithms [33]. In addition, a Ca2+ ion

was added to active site of the structure of calpain-10 model obtainedby homology, since this cation is present in template 2G8E. Afterward,the systemwas fully relaxed, but the peptide backbonewas restrainedwith a lower constant of 100 kJ mol�1 Å�2. The optimised proteinwas placed in a cubic box of pre-equilibrated water (80 Å side), usingthe principal axis of the protein–inhibitor complex as the geometriccentre. Afterwards, 100 ps of hybrid QM/MM Langevin-Verlet mole-cular dynamics (MD) at 300 K and a canonical thermodynamicensemble (NVT) were used to equilibrate the model.

The semiempirical AM1 Hamiltonian [34] was used to describethe QM part for SNJ-1715 and the residues Cys115 and His272 (inthe template), and residues Cys73 and His240 (in the target), bothwith 27 atoms (see Fig. S1 of Supplemental material) [31]. The restof the system, consisting of 63,505 atoms in the target and 63,685atoms in the template, was described using the OPLS-AA [35] andTIP3 [36] force fields for protein and water molecules, respectively,as implemented in the DYNAMO library [37]. All residues exceptthose atoms within 20 Å of the initial inhibitor were selected toremain stationary for the remaining calculations, thus making themodel computationally feasible. Interactions were applied using aswitching scheme, within a range radius from 14.5 to 16 Å.

Once the systems were pre-equilibrated, 3000 ps of QM/MMMD were run at a temperature of 300 K. The computed RMSD forthe protein during the last 400 ps always rendered a value below0.9 Å, while the RMS of the temperature along the differentequilibration steps was always lower than 2.6 K. The variationcoefficient of the potential energy during the dynamics simula-tions was never higher than 0.3%. The interaction calculationsbetween the SNJ-1715 and the enzymes were computed duringthe last 500 ps of the QM/MM MD simulations [38].

3. Results and discussion

3.1. Alignment and structural prediction

Alignment of the template and target sequences is a crucial step toensure the quality of the model when establishing a connectionbetween the model and the target sequence amino acids [39]. Fig. 2shows the alignment of the amino acid sequence of calpain-10 withthe sequence of μ-calpain performed using the ESPript 2.2 server [40],as well as the identification of residues that comprise the catalytic siteand the calcium binding site present in the enzyme loops. The modelobtained from the homology consists of 12 β-sheets, with seven inDomain I (β1–β7) and five in DII (β8–β12), and 10 α-helices, with sevenin DI (α1–α7) and three in DII (α08–α10; Fig. 3). We also identified theN- and C-terminal regions of the model. The analysis of the tertiarystructure showed that the model is classified as an α+β structure, withno supersecondary structures [41]. The model loops showing atomiccollision were moved in the search for the best spatial conformation.

The active site residues are located at the α-4 α-helix (Cys73),between the β-8 and β-9 β-sheets (His238) and near the β-10β-sheet (Asn263). Despite the template and target sequencesbelonging to different calpain groups based on their structure(typical and atypical), there is enough sequence similarity in thecatalytic domain to allow for homology-based model analysis. Thissuggests that the active site has been preserved during evolution,which allows the conservation of protein function. The Cys73,His238, and Asn263 residues of the target are conserved in thetemplate (Cys115, His272, and Asn296, see Fig. S2a and b), which isin agreement with previous reports [16].

3.2. Model validation

The stereochemical quality of the model obtained by homologymodelling was analysed using the Molprobity software [25,26]. Most

Fig. 2. Target-template alignment. The green arrows point to C, H, and N residues within the catalytic site. The points within the blue rectangles indicate residues thatinteract with the Ca2+ in the loops. (For interpretation of the references to colour in this figure caption, the reader is referred to the web version of this article.)

Fig. 3. Model of the calpain-10 proteolytic core obtained by homology with the indicating the α-helix and β-sheet regions, the C, H, and N residues, and the SNJ-1715inhibitor docked at the active site.

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–13401336

target residues occupy more favoured regions of the Ramachandranplot, which represents the distribution of the Φ and ψ angles for eachamino acid residue. In the obtained model, 94.9% of the residueswere observed in favourable regions (Fig. S3).

The ANOLEA [28,29] was used to assess the packing quality ofthe model. The program performs energy calculations on a proteinchain, evaluating the non-local environment of each heavy atom inthe molecule. Negative energy values represent a favourableenergy environment for a particular amino acid, while positivevalues correspond to an unfavourable energy environment.According to ANOLEA, after loop modelling [42], the target pre-sented a very favourable energy environment (Fig. S4b).

The Verify3D software checks if the amino acids are in aplausible chemical environment, reflecting the quality of protein

folding. The results indicate that most of the residues are withinthe acceptable range (positive values). Thus, we can conclude that,for the most part, the structure that we built has the correctfolding (Fig. S5).

RMSD evaluates the structural similarity by overlapping thethree-dimensional structures and evaluating the Cα distance. AnRMSD value of 0.47 Å indicates that the proteins have a commonstructural homology, indicating conservation of internal structuresduring evolution (Fig. 4).

3.3. Electrostatic potential map

The work by Moldoveanu et al. [43] was essential to under-stand the role that Ca2+ plays in the function calpains. Before their

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–1340 1337

work, it was believed that the only metal-binding sites wereEF-hand motifs, absent in atypical calpains (including calpain-10). However, it was unexpectedly observed that two atoms of Ca2+

interact with different EF-hand motif structures existing in the DIand DII domains of calpain. Here, these structures were identifiedbetween the α3, α4 α-helix and near the β-5 β-sheet, in addition tobeing close to the α9 α-helix, between the β10 and β11 β-sheets.

In order to identify possible to interaction with calcium, webuilt an electrostatic potential map through the server PBEQ Solver[44]. This map calculates and displays the electrostatic potential ofa molecule by solving the Poisson–Boltzmann equation.

The identification and comparison of groups of residues thatcomprise the loops that coordinate Ca2+ showed that, in thetemplate, the looplocated in Domain I has two acidic residues(Asp and Glu) and two apolar residues (Val and Gly), while theloop in Domain II has four negative residues (two Glu and twoAsp) and one apolar residue (Met). For the target, the loop locatedin Domain I has one negative residue (Asp), one neutral residue

Fig. 4. Three dimensional structural alignment of the target (red), template (blue),and SNJ-1715 inhibitor (green). This alignment resulted in an RMSD of 0.47 Å. (Forinterpretation of the references to colour in this figure caption, the reader isreferred to the web version of this article.)

Fig. 5. Electrostatic potential map of the template (a) and target (b), and identification ofamino acids surface has a more acidic profile than their counterparts in the target. Thecolour in this figure caption, the reader is referred to the web version of this article.)

(Gln), and two apolar residues (Pro and Val), while the loop inDomain II has three negative residues (Glu), one neutral residue(Cys), and one apolar residue (Leu).

The analysis of the maps in Fig. 5a and b shows that the targetand the template have few differences in their charge distributionsand molecular surfaces. Even if the binding sites for Ca2+ areconserved among the proteases [43], we observed that the loops ofthe template have more acidic residues than the loops of thetarget. This may suggest that the 2G8E template, which is ratμ-calpain, has a higher affinity for this ion than human calpain-10.Different sensitivities to Ca2+ in calpain isoforms have beenreported in the literature [43,45,46]. It is possible that calpain-10is less dependent on Ca2+ than the typical calpains because of theabsence of an EF-hand motif, a change in the phosphorylation ofits structure [47], or even the electrostatic characteristics observedin this study. These differences may account for calpain-10'sdistinct physiological and pharmacological properties. However,further investigation is required to better understand the relation-ship between the metal and calpain-10.

3.4. Formation of the ionic pair Cys�/His+

The role of histidine residues in the catalytic triad has beenextensively described, and it has been suggested that histidinefunctions enzyme catalyst [48] as it keeps the thiolate anion in aneutral or weakly acid state [49].

Theoretical [50] and experimental [48,51,52] data indicate thatthere is an interaction between the Cys�/His+ ion pair, with a loss ofenzyme activity resulting from replacing the cysteine residue with aneutrally charged one [50]. This imidazole–thiolate interactive systemtransiently acylates the sulphur atom during catalysis [53]. It isnoteworthy that, unlike the serine proteases, the formation of theion pair is completely dependent on Ca2+ and does not depend of thepresence of a substrate for ionisation of the active site [54,55].

At the beginning of the simulation (Fig. S6a and b), weobserved the spontaneous transfer of the hydrogen atom of thecysteine residue (pKa 14) to the histidine residue (pKa 3.6) forming

the residues that interact with calcium (yellow letters). In comparison, the templatevaline residue of the target is not displayed. (For interpretation of the references to

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–13401338

the same Cys�/His+ ion pairs on the template and on the target(Fig. S7), as previously reported [50]. Therefore, these results areconsistent with the literature, demonstrating that this phenom-enon makes the chemical environment conducive to the onset ofenzymatic activity [48,51].

Table 1Average distances (Å) between the template or target atoms and the inhibitor SNJ-1715.

ResidueTemplate(Target)

AtomsResidue⋯SNJ-1715

Averagedistancestemplate

Averagedistancestarget

Experimentaldistance

C115 (C73) H⋯O21 1.97 (0.16) 3.28 (1.19) 3.24C115 (C73) SG⋯C20 2.26 (0.12) 2.53 (1.09) 1.63G207 (A167) O⋯H33 2.28 (0.20) 4 5 Å 2.95G207 (A167) H⋯O17 2.84 (0.27) 4 5 Å 2.75G208 (G168) O⋯H34 2.28 (0.20) 2.47 (0.33) 2.95G271 (F237) O⋯H31 2.64 (0.34) 4 5 Å 2.84

a Standard deviations (in Å) computed during the last 3000 ps are reported inparentheses.b The number of residues correspond to the Target are reported in parentheses infirst column.

3.5. Inhibitor–protein complex

The active site of proteases has been an attractive target for thedevelopment of inhibitors. Peptides, such as aldehydes, inhibitenzymatic activity through a reversible stable hemithioacetal bondwith a catalytic cysteine [56]. Inhibitors of cysteine proteases areuniquely designed to undergo attack by the sulphide anion in thetransition state analog regions. Other important inhibitor regions arethose capable of forming and stabilizing non-covalent interactionswith the active site of the enzyme.

After 3000 ps of hybrid QM/MM and MD simulations, structureanalysis was performed on the complex enzyme–inhibitor by com-paring experimental data obtained by crystallography with themodels submitted to MD. Cuerrier et al. [16] reported that the X-ray structure of μ-calpain complexed to the SNJ-1715 inhibitorshowed interactions with the Cys115, Gly207, Gly20,8 and Gly271residues, as well as nucleophilic attack of the thiol group of cysteineto the carbonyl carbon of the inhibitor. These experimental data weretaken as a reference for the MD simulation of the template and targetfor 3000 ps.

The average distances obtained between the residues of the activesite and SNJ-1715 were plotted. We found that the theoretical valueswere close to the experimental values, as shown in Fig. 6a and b andTable 1. The SNJ-1715 inhibitor forms hydrogen bonds with atoms ofthe Gly207, Gly208, and Gly271 residues in the template, and Gly168in the target. In addition, the hydrogen atom of the cysteine residueinteracts strongly with the O21 oxygen atom of the inhibitor,including a distance (1.97 Å) that was lower than when comparedto the experimental data (3.24 Å). A feature of the technique is thatthe crystallographic sample is obtained in its solid form, withoutdescribing the behaviour of the molecule over a period of time [57].This is in contrast to the computer simulation studies in which themovements of particles in a system are a function of time [58].

The formation of a stable hemithioacetal bond was also observedbetween the sulphide anion and the carbonyl carbon of the inhibitor.These interactions block access to the active site and promote thestabilisation of the enzyme–inhibitor complex. They were alsoobserved in a study on leupeptin [59], confirming that our work isin agreement with other published findings [60,61]. Additionally,hemithioacetal bonding was also observed in the homology model,

Fig. 6. Schematic of the interactions between protein and

with an average distance of 2.53 Å, suggesting that calpain-10possesses a catalytic mechanism similar to other isoforms.

Further analysis allowed for visualisation of the structuralchanges in SNJ-1715 through intramolecular interactions betweenthe H46 hydrogen and the O21 oxygen of the compound, with anaverage distance of 2.46 Å (Fig. 7). The amplitude was 6.32–1.78 Åduring 3000 ps of MD (Fig. S8). The interaction occurred in anaqueous medium, simulating the surrounding biological environ-ment. In fact, this conformational change of the free aldehyde formto a cyclic hemiacetal form is reversible and necessary [13,15],since it is able to increase the bioavailability and cell permeabilityof the inhibitor by decreasing its high reactivity, which is char-acteristic of aldehydes, and makes the inhibitor more efficient [62].

During the MD simulations, the distances between the calciumions and the closest residues were kept frozen, since these ions areabout 20 Ǻ away from inhibitor. Each Ca2+ ion binds within a singledomain. These residues coordinate the structural alignment of theproteolytic core. This interaction provides a structural basis for thecooperativity between the two Ca2+ binding sites in calpain [43]. Thetemplate has a larger number of residues with acidic character, suchas aspartates (Asp185, Asp106, Asp331, Asp333, and Asp309) andglutamates (Glu302 and Glu333), which allow an octrahedral coor-dination to be formed by calcium and these residues (see Supple-mentary material, Fig. S9). On the target, the aspartate and glutamateresidues also (Asp62, Glu276, Glu298, and Glu298) show an impor-tant role for the octahedral coordination formed by calcium andclosest residues. The structural data presented here suggest thatthe protease core of calpain-10, which lacks both the EF-hand andC2-like domains, retains the minimal functional and structuralrequirements of a Ca2+-dependent cysteine protease [16]. Moreover,

inhibitor in (a) the template and (b) the target.

Fig. 7. Conformational changes in the SNJ-1715 inhibitor before (transparent) and after 3000 ps of MD. Carbon atoms are coloured in green, oxygen in red, nitrogen in blue,and sulphur in yellow. Key distances between O17, H33, H31, H48, O21, and C20 atoms and some residues are listed in Table 1. (For interpretation of the references to colourin this figure caption, the reader is referred to the web version of this article.)

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–1340 1339

the Ca2+-bound calpain-10 structure provides an ideal target for thedesign and testing of active site inhibitors that will help define thepathophysiological roles of the calpains.

4. Conclusion

The human calpain-10 enzyme model obtained through homologymodelling suggests that the active site of this enzyme is conserved andthe main interactions are similar to those observed for μ-calpain. Theamino acid mutations observed in the calcium binding site located inthe loop in Domain I are in accordance with the results of otherauthors see Ref. [43] that have proposed that calpain-10 is probablyless dependent on calcium than the typical calpains. Moreover, theCys�/His+ ion pair formation was observed in the target and in the2G8E template during the QM/MM MD simulations. Finally, structuralchanges were observed with the SNJ-1715 inhibitor during thesimulation, switching from the free aldehyde form to the reversiblecyclic hemiacetal form. We hope these results will be useful for thedesign of new inhibitors of calpain-10.

Conflict of interest statement

None declared.

Acknowledgments

We would like to thank the Conselho Nacional de Desenvolvi-mento Científico e Tecnológico (CNPq), Coordenação de Aperfei-çoamento de Pessoal de Nível Superior (CAPES) and Pró-Reitoriade Pesquisa e Pós-Graduação da Universidade Federal do Pará(PROPESP-UFPA) for their financial support.

Appendix A. Supporting information

Supplementary data associated with this article can be found in theonline version at http://dx.doi.org/10.1016/j.compbiomed.2013.07.010.

References

[1] N.D. Rawlings, F.R. Morton, A.J. Barrett, MEROPS: the peptidase database,Nucleic Acids Res. 34 (2010) 270–272.

[2] D.E. Goll, V.F. Thompson, H. Li, W.E.I. Wei, J. Cong, The Calpain system, Physiol.Rev. 83 (2003) 731–801.

[3] M.E. Saez, R. Ramirez-Lorca, F.J. Moron, A. Ruiz, The therapeutic potential ofthe calpain family: new aspects, Drug Discov. Today 11 (2006) 917–923.

[4] H. Sorimachi, K. Suzuki, The structure of calpain, J. Biochemistry 129 (2001)653–664.

[5] T. Ozaki, H. Tomita, M. Tamai, Sei-ichi. Ishiguro, Characteristics of mitochon-drial calpains, J. Biochem. 142 (2007) 365–376.

[6] D.D. Arrington, T.R. Van Vleet, R.G. Schnellmann, Calpain 10: a mitochondrialcalpain and its role in calcium-induced mitochondrial dysfunction, Am. J.Physiol. – Cell Physiol. 291 (2006) 1159–1171.

[7] X.S. Puente, L.M. Sanchez, C.M. Overall, C. Lopez-Otin, Human and mouseproteases: a comparative genomic approach, Nat. Rev. Genet. 4 (2003)544–558.

[8] I. Ken-ichiro Kuwako, T. Nishimura, T.C. Uetsuki, K. Saido, Yoshikawa, Activa-tion of calpain in cultured neurons overexpressing Alzheimer amyloid pre-cursor protein, Mol. Brain Res. 107 (2002) 166–175.

[9] Y.E. Demirci H., M.A. Ergun, A.C. Yazici, C. Karasu, I. Yetkin, Calpain 10 SNP-44gene polymorphism affects susceptibility to type 2 diabetes mellitus anddiabetic-related conditions, Genet. Test 2 (2008) 305–309.

[10] M. Zatz, A. Starling, Calpains and disease, N. Engl. J. Med. 352 (2005)2413–2423.

[11] R. Moreno-Luna, A. Abrante, F. Esteban, M.A. González-Moles, M. Delgado-Rodríguez, M.E. Sáez, A. González-Pérez, R. Ramírez-Lorca, L.M. Real, A. Ruiz,Calpain 10 gene and laryngeal cancer: a survival analysis, Head Neck 33 (2011)72–76.

[12] Y. Shirasaki, H. Miyashita, M. Yamaguchi, J. Inoue, M. Nakamura, Exploration oforally available calpain inhibitors: Peptidyl α-ketoamides containing anamphiphile at P3 site, Bioorgan. Med. Chem. 13 (2005) 4473–4484.

[13] M. Nakamura, H. Miyashita, M. Yamaguchi, Y. Shirasaki, Y. Nakamura, J. Inoue,Novel 6-hydroxy-3-morpholinones as cornea permeable calpain inhibitors,Bioorgan. Med. Chem. 11 (2003) 5449–5460.

[14] S. Auvin, B. Pignol, E. Navet, M. Troadec, D. Carré, J. Camara, D. Bigg, P. Chabrier,Novel dual inhibitors of calpain and lipid peroxidation with enhanced cellularactivity, Bioorgan. Med. Chem. Lett. 16 (2006) 1586–1589.

[15] K.K.W. Wang, Po-wai Yuen, Development and therapeutic potential of calpaininhibitors, in: M.W.A.F.M. J. Thomas August, T.C. Joseph (Eds.), Advances inPharmacology, Academic Press, 1996, pp. 117–152.

[16] D. Cuerrier, T. Moldoveanu, J. Inoue, P.L. Davies, R.L. Campbell, Calpaininhibition by α-ketoamide and cyclic hemiacetal inhibitors revealed by X-raycrystallography, Biochemistry 45 (2006) 7446–7452.

[17] K.W. Nickerson, R.A. Day, Possible biological roles for metastable proteins,Biosystems 2 (1969) 303–306.

[18] M.M. Barry, C.D. Mol, W.F. Anderson, J.S. Lee, Sequencing and modeling of anti-DNA immunoglobulin Fv domains. Comparison with crystal structures, J. Biol.Chem. 269 (1994) 3623–3632.

R.C. da Silva et al. / Computers in Biology and Medicine 43 (2013) 1334–13401340

[19] A.H. Lima, P.R.M. Souza, N. Alencar, J. Lameira, T. Govender, H.G. Kruger, G.E.M. Maguire, C.N. Alves, Molecular modeling of T. rangeli, T. brucei gambiense,and T. evansi Sialidases in complex with the DANA inhibitor, Chem. Biol. DrugDes. 80 (2012) 114–120.

[20] P.R.M. Sousa, N.A.N. de Alencar, A.H. Lima, J. Lameira, C.N. Alves, Protein–ligand interaction study of CpOGA in complex with GlcNAcstatin, Chem. Biol.Drug Des. 81 (2013) 284–290.

[21] T. Rungrotmongkol, P. Decha, P. Sompornpisut, M. Malaisree, P. Intharathep,N. Nunthaboot, T. Udommaneethanakit, O. Aruksakunwong, S. Hannongbua,Combined QM/MM mechanistic study of the acylation process in furincomplexed with the H5N1 avian influenza virus hemagglutinin's cleavagesite, Proteins 76 (2009) 62–71.

[22] C.K. Regmi, Y.R. Bhandari, B.S. Gerstman, P.P. Chapagain, Exploring thediffusion of molecular oxygen in the red fluorescent protein mcherry usingexplicit oxygen molecular dynamics simulations, J. Phys. Chem. B 117 (2013)2247–2253.

[23] G. Moraes, V. Azevedo, M. Costa, A. Miyoshi, A. Silva, V. da Silva, D. de Oliveira,M.F. Teixeira, J. Lameira, C.N. Alves, Homology modeling, molecular dynamicsand QM/MM study of the regulatory protein PhoP from Corynebacteriumpseudotuberculosis, J. Mol. Model. 18 (2012) 1219–1227.

[24] A.D.S. Guimarães, E.M.S. Dorneles, G.I. Andrade, A.P. Lage, A. Miyoshi,V. Azevedo, A.M.G. Gouveia, M.B. Heinemann, Molecular characterization ofCorynebacterium pseudotuberculosis isolates using ERIC-PCR, Vet. Microbiol.153 (2011) 299–306.

[25] V.B. Chen, W.B. Arendall, J.J. Headd, D.A. Keedy, R.M. Immormino, G.J. KapralL.W. Murray, J.S. Richardson, D.C. Richardson, MolProbity: all-atom structurevalidation for macromolecular crystallography, Acta Crystallogr. D 66 (2010)12–21.

[26] I.W. Davis, L.W. Murray, J.S. Richardson, D.C. Richardson, MolProbity: structurevalidation and all-atom contact analysis for nucleic acids and their complexes,Nucleic Acids Res. 32 (2004) 615–619.

[27] R. Luthy, J.U. Bowie, D. Eisenberg, Assessment of protein models with three-dimensional profiles, Nature 356 (1992) 83–85.

[28] F. Melo, E. Feytmans, Assessing protein structures with a non-local atomicinteraction energy, J. Mol. Biol. 277 (1998) 1141–1152.

[29] F. Melo, D. Devos, E. Depiereux, E. Feytmans, ANOLEA: a www server to assessprotein structures, Proc. Int. Conf. Intell Syst. Mol. Biol. 5 (1997) 187–190.

[30] R.A. Engh, R. Huber, Accurate bond and angle parameters for X-ray proteinstructure refinement, Acta Crystallogr. A 47 (1991) 392–400.

[31] H.M. Senn, W. Thiel, QM/MM methods for biomolecular systems, Angew.Chem. Int. Ed. 48 (2009) 1198–1229.

[32] D.C. Bas, D.M. Rogers, J.H. Jensen, Very fast prediction and rationalization ofpKa values for protein–ligand complexes, Proteins 73 (2008) 765–783.

[33] H.B. Richard, L. Peihuang, A Limited-Memory, Algorithm for bound-constrained optimization, SIAN J. Sci. Comput. 16 (1995) 1190–1208.

[34] M.J.S. Dewar, E.G. Zoebisch, E.F. Healy, J.J.P. Stewart, Development and use ofquantum mechanical molecular models. 76. AM1: a new general purposequantum mechanical molecular model, J. Am. Chem. Soc. 107 (1985)3902–3909.

[35] J. Tirado-Rives, W.L. Jorgensen, The OPLS force field for proteins. Energyminimizations for crystals of cyclic peptides and crambin, J. Am. Chem. Soc.110 (1988) 1657–1666.

[36] W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, Com-parison of simple potential functions for simulating liquid water, J. Chem.Phys. 79 (1983) 926.

[37] M.J. Field, M. Albe, C. Bret, F. Proust-De Martin, A. Thomas, The dynamo libraryfor molecular simulations using hybrid quantum mechanical and molecularmechanical potentials, J. Comput. Chem. 21 (2000) 1088–1100.

[38] J. Lameira, C.N. Alves, V. Moliner, S. Marti, R. Castillo, I. Tunon, Quantummechanical/molecular mechanical molecular dynamics simulation of wild-type and seven mutants of CpNagJ in complex with PUGNAc, J. Phys. Chem. B114 (2010) 7029–7036.

[39] K.W. Lee, J.M. Briggs,Molecular, modeling study of the editing active site ofEscherichia coli leucyl-tRNA synthetase: two amino acid binding sites in theediting domain, Proteins 54 (2004) 693–704.

[40] P. Gouet, E. Courcelle, D.I. Stuart, F. Métoz, ESPript: analysis of multiplesequence alignments in PostScript, Bioinformatics 15 (1999) 305–308.

[41] A. Lesk, Introduction to Bioinformatics, Oxford University Press, 2002.[42] A. Fiser, R.K. Do, A. Šali, Modeling of loops in protein structures, Protein Sci. 9

(2000) 1753–1773.[43] T. Moldoveanu, C.M. Hosfield, D. Lim, J.S. Elce, Z. Jia, P.L. Davies, A Ca2+ switch

aligns the active site of Calpain, Cell 108 (2002) 649–660.[44] W. Im, D. Beglov, B. Roux, Continuum solvation model: computation of

electrostatic forces from numerical solutions to the Poisson–Boltzmannequation, Comput. Phys. Commun. 111 (1998) 59–75.

[45] H.A. Wells, A. Calpain. In: A. Wells, A. Huttenlocher (Eds.). Handbook of CellSignaling 02, Califórnia Handbook of Cell Signaling, San Francisco, vol. 141,2003, pp. 105–111.

[46] M.H. Nakagawa K., H. Sorimachi, K. Suzuki, Dissociation of m-calpain subunitsoccurs after autolysis of the N-terminus of the catalytic subunit, and is notrequired for activation, J. Biochem. 130 (2001) 605–611.

[47] A. Glading, R.J. Bodnar, I.J. Reynolds, H. Shiraha, L. Satish, D.A. Potter, H.C. Blair,A. Wells, Epidermal growth factor activates m-Calpain (Calpain II), at least inpart, by extracellular signal-regulated kinase-mediated phosphorylation, Mol.Cell Biol. 24 (2004) 2499–2512.

[48] F. Talfournier, N. Colloc'h, Jean-Paul Mornon, G. Branlant, Comparative study ofthe catalytic domain of phosphorylating glyceraldehyde-3-phosphate dehy-drogenases from bacteria and archaea via essential cysteine probes and site-directed mutagenesis, Eur. J. Biochem. 252 (1998) 447–457.

[49] K. Brocklehust, Enzyme mechanisms, R. Soc. Chem. 140 (1987) 158.[50] A. Cartier, D. Brown, B. Maigret, S. Boschi-Muller, S. Rahuel-Clermont,

G. Branlant, Modelling the active site of glyceraldehyde-3 phosphate dehy-drogenase with the LSCF formalism, Theor. Chem. Acc. 101 (1999) 241–245.

[51] L. Polgár, Ion-pair formation as a source of enhanced reactivity of the essentialthiol group of d-glyceraldehyde-3-phosphate dehydrogenase, Eur. J. Biochem.51 (1975) 63–71.

[52] A. Soukri, A. Mougin, C. Corbier, A. Wonacott, C. Branlant, G. Branlant, Role ofthe histidine-176 residue in glyceraldehyde-3-phosphate dehydrogenase asprobed by site-directed mutagenesis, Biochemistry 28 (1989) 2586–2592.

[53] K. Brocklehust, F. Willenbrock, E. Salih, Cysteine proteinases. In: A. Neuberger,K. Brocklehust (Eds.), Hydrolytic Enzymes, vol. 16, 1987, pp. 39–158.

[54] B. Wiederanders, G. Kaulmann, K. Schilling, Functions of propeptide parts incysteine proteases, Curr. Protein Pept. Sci. 4 (2003) 309–326.

[55] S.S. Mellor G.W., D. Kowlessur, E.W. Thomas, K. Brocklehurst, Catalytic-sitecharacteristics of the porcine calpain II 80 kDa/18 kDa heterodimer revealedby selective reaction of its essential thiol group with two-hydronic-state time-dependent inhibitors: evidence for a catalytic site Cys/His interactive systemand na ionizing modulatory group, Biochem. J. 290 (1993) 75–83.

[56] I.O. Donkor, A survey of calpain inhibitors, Curr. Med. Chem. 7 (2000)1171–1188.

[57] J.C. Kendrew, G. Bodo, H.M. Dintzis, R.G. Parrish, H. Wyckoff, D.C. Phillips, Athree-dimensional model of the myoglobin molecule obtained by X-rayanalysis, Nature 181 (1958) 662–666.

[58] J. Norberg, L. Nilsson, Advances in biomolecular simulations: methodologyand recent applications, Q. Rev. Biophys. 36 (2003) 257–306.

[59] T. Moldoveanu, R.L. Campbell, D. Cuerrier, P.L. Davies, Crystal structures ofCalpain-E64 and leupeptin inhibitor complexes reveal mobile loops gating theactive site, J. Mol. Biol. 343 (2004) 1313–1326.

[60] Z. Li, A.C. Ortega-Vilain, G.S. Patil, D.L. Chu, J.E. Foreman, D.D. EvelethJ.C. Powers, Novel peptidyl α-keto amide inhibitors of calpains and othercysteine proteases, J. Med. Chem. 39 (1996) 4089–4098.

[61] S. Mehdi, M.R. Angelastro, J.S. Wiseman, P. Bey, Inhibition of the proteolysis ofrat erythrocyte membrane proteins by a synthetic inhibitor of calpain,Biochem. Biophys. Res. Commun. 157 (1988) 1117–1123.

[62] J. Wang, B. Pignol, Pierre-Etienne Chabrier, T. Saido, R. Lloyd, Y. Tang,M. Lenoir, Jean-Luc Puel, A novel dual inhibitor of calpains and lipidperoxidation (BN82270) rescues the cochlea from sound trauma, Neurophar-macology 52 (2007) 1426–1437.