Inhibition of chikungunya virus by picolinate that targets viral capsidprotein

Rajesh Sharma a, Benazir Fatma a, Amrita Saha b, Sailesh Bajpai c, Srinivas Sistla c,Paban Kumar Dash b, Manmohan Parida b, Pravindra Kumar a, Shailly Tomar a,n

a Department of Biotechnology, Indian Institute of Technology Roorkee, Roorkee 247667, Indiab Division of Virology, Defence Research and Development Establishment, Gwalior 474002, Indiac GE Healthcare Life Sciences, John F Welch Technology Centre, EPIP, Phase 2, Whitefield Road, Bangalore 560048, India

a r t i c l e i n f o

Article history:Received 3 June 2016Returned to author for revisions26 August 2016Accepted 29 August 2016Available online 9 September 2016

Keywords:Chikungunya virusCapsid proteinProtein-protein interactionsPicolinic acidPyridine ring, Surface Plasmon resonanceAntiviralQuantitative real-time RT-PCR

a b s t r a c t

The protein-protein interactions (PPIs) of the transmembrane glycoprotein E2 with the hydrophobicpocket on the surface of capsid protein (CP) plays a critical role in alphavirus life cycle. Dioxane basedderivatives targeting PPIs have been reported to possess antiviral activity against Sindbis Virus (SINV),the prototype alphavirus. In this study, the binding of picolinic acid (PCA) to the conserved hydrophobicpocket of capsid protein was analyzed by molecular docking, isothermal titration calorimetry (ITC),surface plasmon resonance (SPR) and fluorescence spectroscopy. The binding constant KD obtained forPCA was 2.1�10�7 M. Additionally, PCA significantly inhibited CHIKV replication in infected Vero cells,decreasing viral mRNA and viral load as assessed by qRT-PCR and plaque reduction assay, respectively.This study is suggestive of the potential of pyridine ring compounds as antivirals against alphavirusesand may serve as the basis for the development of PCA based drugs against alphaviral diseases.

& 2016 Elsevier Inc. All rights reserved.

1. Introduction

The re-emergence of the chikungunya disease caused by CHIKVpossesses a serious worldwide threat to the public health (Cavriniet al., 2009). The main vectors responsible for the transmission ofCHIKV are Aedes albopictus and Aedes aegypti. CHIKV infection ischaracterized by acute fever, rashes and musculoskeletal pain inhumans (Morrison, 2014). Apart from these symptoms, CHIKV fe-ver is also characterized by persistent arthralgia (Larrieu et al.,2010). Clinical manifestations of CHIKV fever includes lethal he-patitis, lymphopenia, severe dermatological lesions, encephalitisin adults and fetal transmission during pregnancy causing neo-natal encephalopathy (Sourisseau et al., 2007). The time betweenthe mosquito bite to the start of the symptoms ranges from 2 to 12days. This disease was first reported from the serum of a febrilepatient in 1952 following an outbreak on the Makonde plateau(Ross, 1956). This place is the border area near Mozambique andTanzania. Since its first report in Africa, numerous outbreaks

occurred in other parts of the world. In Kenya, major outbreakoccurred in 2004 that had spread this disease in Comoro, Mada-gascar, Mayonette, La Reunion islands, South East Asia and WestAfrica (Sergon et al., 2008). An unexpected intrusion of the CHIKVinfection in La Reunion Island in 2005 that affected one-third ofthe total population shocked the world (Nimmannitya et al., 1969;Padbidri and Gnaneswar, 1978; Sergon et al., 2007; Shah et al.,1964). In India itself, about more than 1.4 million people wereaffected with CHIKV epidemic in 2006 (Mavalankar et al., 2008).Prior to 2013, CHIKV infection had been reported in Asia, Africa,Europe, and the Indian and Pacific Oceans. Currently this diseasehas been identified in about more than 60 countries and territories(http://www.cdc.gov/chikungunya/geo/index.html). In 2008, USNational Institute of Allergy and Infectious Diseases (NIAID) haslisted this virus as category C priority pathogen (Thiboutot et al.,2010; Wang et al., 2008).

CHIKV belongs to the genus Alphavirus that consists of single-stranded positive-sense RNA viruses. These are transmitted bymosquitoes and infect wide range of animals including mammals.Alphavirus genus belongs to Togaviridae family and includes manymedically relevant human and animal viruses including CHIKV,Ross River virus (RRV), Venezuelan Equine Encephalitis virus(VEEV), Western Equine Encephalitis virus (WEEV) etc. The genuscontains at least 29 members that are able to infect various ver-tebrates such as humans, horses, fish, birds and rodents. These

Contents lists available at ScienceDirect

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

Virology

http://dx.doi.org/10.1016/j.virol.2016.08.0290042-6822/& 2016 Elsevier Inc. All rights reserved.

Abbreviations: CHIKV, Chikungunya virus; NC, Nucleocapsid; CVCP, CHIKV capsidprotein; cdE2, Cytoplasmic domain of E2 Glycoprotein; PCA, Picolinic acid; ITC,Isothermal Titration Calorimetry; SPR, Surface Plasmon Resonance; CP, CapsidProtein; PPIs, protein-protein interactions

n Corresponding author.E-mail addresses: shailfbt@iitr.ac.in, shaiprav@gmail.com (S. Tomar).

Virology 498 (2016) 265–276

members have been currently classified into eight complexesbased on antigenic and genetic similarities (Powers et al., 2001;Weaver et al., 2012). Alphaviruses enter the host cells via receptor-mediated endocytosis followed by genome replication in the cy-toplasm (Jose et al., 2009). The alphavirus virion is approximately70 nm in diameter and has host cell derived lipid membranewhich is embedded with 80 spikes in T¼4 icosahedral symmetry.Each spike is composed of trimers of E1 and E2 glycoproteinheterodimers (Mukhopadhyay et al., 2006; Soonsawad et al., 2010;Strauss and Strauss, 1994). The RNA genome (�11.7 Kb) is sur-rounded by nucleocapsid core (NC) that consists of 240 copies ofCP. The virion genomic RNA with a 5′ methylated cap and apolyadenylated tail is directly translated into viral replication en-zymes in the host cell cytoplasm on entry (Baron et al., 1996;Vasiljeva et al., 2000). Alphavirus infection generally causes ar-thralgia, rash, headache, fever and arthritis. However, some of thepathogenic species like VEEV, EEEV, WEEV cause encephalitis andinfections can also be life-threatening. Alphaviruses are classifiedinto New and Old world viruses depending upon the mechanismby which they down regulate the transcription of host cell anddisease complications. Old World viruses like CHIKV, Semliki For-est virus (SFV) and SINV etc. employ nsp2 protein for down reg-ulation of the cellular transcription and cause arthralgia. Newworld viruses like VEEV, WEEV etc. employ capsid protein forshutting off the host cell transcription and cause encephalitis(Garmashova et al., 2007; Hahn et al., 1988).

The CP of alphaviruses is a multifunctional protein. It consistsof two domains, the RNA binding N-terminal domain and theC-terminal protease domain (Choi et al., 1991; Hong et al., 2006).The N-terminal domain, which is highly disordered is involved inbinding to the genomic RNA, PPIs that lead to the CP dimerizationand the inhibition of host transcription (Owen and Kuhn, 1996;Perera et al., 2001). Similarly, the N-terminal domain of alphavirusnsp4 protein is disordered and involved in PPIs (Rupp et al., 2011;Tomar et al., 2006). The C-terminal domain of CP is a serine pro-tease that possesses a chymotrypsin-like fold. It has autoproteo-lytic activity and cleaves itself from the N-terminus of the struc-tural polyprotein and thus, plays a critical role in the processing ofstructural polyprotein and the viral life cycle. After the cis-clea-vage, the C-terminal tryptophan residue of CP remains bound tothe active site and blocks further CP protease activity (Choi et al.,1991). However, recently the in-vitro trans-cleavage activity of theC-terminal truncated Aura Virus capsid protease (AVCP) andCHIKV capsid protease (CVCP) using synthetic fluorogenic peptidecontaining the CP protease cleavage site has been reported (Ag-garwal et al., 2014; Aggarwal et al., 2015).

The mature virion surrounded by two concentric shells consistsof 240 copies of each capsid, E1and E2 protein. The outer shellcomprises of E1 and E2 glycoproteins and inner shell comprises ofCP (Cheng et al., 1995; Strauss and Strauss, 1994). The space be-tween the two concentric shells is filled with the host derived lipidbilayer through which the transmembrane helix of the glycopro-teins penetrates (Mancini et al., 2000; Tang et al., 2011). The cryo-electron microscopy (cryo-EM) and crystallographic studies fromvarious alphaviruses provide crucial details about the distribution,structures and possible interactions of the cytoplasmic tail of E1and E2 glycoproteins and CP (Kostyuchenko et al., 2011; Zhanget al., 2002). The molecular interaction of CP with the cytoplasmicdomain of E2 glycoprotein (cdE2) facilitates the budding of virionsfrom the plasma membrane of infected host cell (Jose et al., 2012).To analyze this molecular interaction, the homology models ofAVCP cdE2 glycoprotein was generated and fitted into the cryo-EMdensity map of VEEV along with crystal structure of AVCP-dioxanecomplex. Fitting analysis revealed the hydrophobic contacts of theconserved Pro405 of Aura virus cdE2 with the hydrophobic pocketof AVCP. The Pro405 is present at the loop region of helix-loop-

helix structure of E2-cytoplasmic domain (Aggarwal et al., 2012;Lopez et al., 1994; Owen and Kuhn, 1997). It was postulated thatthe dioxane present in the CP-hydrophobic pocket of AVCP-diox-ane complex structurally mimics the pyrrolidine ring of Pro405residue of cdE2 (Aggarwal et al., 2012). This suggested that het-erocyclic ring compounds similar to dioxane and its derivativescould bind in the hydrophobic pocket of CP and disrupt its PPIswith cdE2. PCA, a pyridine containing compound is known to haveantiviral, antimicrobial, cytotoxic and apoptotic properties (Corrêaet al., 2011; Fernandez-Pol et al., 2000). Previous studies have re-ported the antiviral properties of PCA against Human Hepatitis Bvirus, Dengue virus and SINV. PCA was found to inhibit the SINVreplicons at various concentrations tested (Fernandez-Pol andFernandez-Pol, 2008). The chemistry of nitrogen containing het-erocyclic rings can be exploited for design and synthesis of drugmolecules. Additionally, drug molecules based on pyridine ringsuch as isoniazid, the antituberculosis agent are already availablein the market (Preziosi, 2007). Recent findings have provided somebreakthroughs in finding anti-CHIKV chemotherapeutic agents likeInterferons, Ribavirin, Mercaptopurine etc. (Jadav et al., 2015). Buttill date, no effective drug or vaccine is available in the market forthe treatment of CHIKV disease.

In this study, it is postulated that the binding of picolinic acid toCVCP hydrophobic pocket might inhibit CP-cdE2 interactions seenin cryo-EM alphavirus structures (Aggarwal et al., 2012; Tang et al.,2011; Zhang et al., 2011). Molecular docking, isothermal titrationcalorimetry (ITC), surface plasmon resonance (SPR) and fluores-cence studies have been carried out to characterize the binding ofpicolinic acid to the hydrophobic pocket of CVCP. Our resultsclearly demonstrate the antiviral properties of PCA against CHIKV,and thus support the postulate that the binding of PCA to thehydrophobic pocket of CVCP blocks cd-E2-capsid interaction.Hence, this study illustrates the potential to use pyridine ring fordesign and development of pharmaceutically active antiviralcompounds to combat chikungunya disease.

2. Materials and methods

2.1. 3D model generation

Homology modeling was done for CVCP protein (residue range:113–261), in following steps: (a) selection of template from proteindata bank (PDB), (b) sequence template alignment, (c) building ofmodel, (d) refinement and validation. NCBI BLAST search tool wasused to perform the template search. The crystal structure ofSemliki Forest virus capsid protein (PDBID: 1VCP) was the first hitwith 93% sequence identity to CVCP (Choi et al., 1997). Clustalomega program was used for multiple sequence alignment of thequery sequence with the template sequence (http://www.ebi.ac.uk/Tools/msa/clustalo/). 1VCP was used as a template to generatethree dimensional structure of CVCP. Several models were gener-ated and the model with the lowest DOPE score was further as-sessed using PROCHECK and ERRAT (http://nih server.mbi.ucla.edu/SAVES/). The model with the least number of residues in thedisallowed region was further evaluated for energy minimizationusing Swiss-Pdb Viewer 4.01 (http://spdbv.vital-it.ch/).

2.2. Molecular docking

AutoDock 4.2 was used for carrying out docking studies ofsmall molecules into CVCP modeled structure. Docking was con-ducted using Windows 2007 on a HPxw8400 workstation. Themodeled CVCP protein structure was selected for the dockingstudies. All the water molecules were removed from the proteinstructure, and non-polar hydrogens and Kollman charges (Kollman

R. Sharma et al. / Virology 498 (2016) 265–276266

charges ¼9) were added using AutoDock MGL Tools. The proteinwas saved in .pdbqt file format. The ligands for the CVCP proteindocking studies were retrieved from PubChem Compound Data-base. The name of the ligands and their molecular formulas aregiven in Table 1. The ligands were downloaded from PubChem inSDF format and then converted in to PDB format using Open Babelsoftware. For the ligand preparation, same procedure was em-ployed for all the ligands. The Grid parameters were same. Hy-drogens atoms and Gasteiger charges were added. For dioxane,proline and picolinic acid the Gasteiger charges added were�0.0002, �0.0001 and �1.0001, respectively. After these changesthe ligands were saved in .pdbqt file format. The atomic potentialgrid map was calculated by AutoGrid 4 with a spacing of 0.375 Åand the Grid box dimensions were 30 Å�30 Å�30 Å. The centerpoint co-ordinates set was X¼8.726, Y¼115.84 and Z¼36.188.Lamarckian Genetic algorithm in combination of grid based energyevaluation method was used for docking. The number of total GAruns was increased from 10 to 50. Other docking parameters wereset as default. For analysis, the configuration with the lowestbinding energy was considered.

2.3. Virus

An Indian isolate of Chikungunya virus, DRDE-06 (GenBankaccession no: EF-210157) belonging to East Central and SouthAfrican (ECSA) genotype maintained at Virology Division, DRDE,Gwalior was used in the present study for the expression of re-combinant capsid (Aggarwal et al., 2015).

2.4. Expression and purification of CVCP protein

2.4.1. Expression of CVCP proteinChemically competent cells of Escherichia coli strain Rosetta

(DE3) were transformed with recombinant plasmid containingN-terminal His-tagged CVCP capsid protein gene (residues 106-261). The amino acid sequence for the affinity tag and the TEVprotease cleavage site at the N-terminus of the protein wasMGSSHHHHHHSSENLYFQGHM. The optimization of expressionconditions for CVCP was done to produce the protein in solubleform. These transformed cells were plated on LB agar plate con-taining 50 μg/ml kanamycin and 35 μg/ml chloramphenicol, andincubated at 37 °C overnight. Single colony was picked and in-oculated in LB broth supplemented with the desired antibiotics,

grown at 37 °C overnight with constant stirring. The overnightgrown culture was used to raise 1 L secondary culture. The sec-ondary culture was grown till the OD at 600 nm (OD600) reachedto �0.8 and afterwards the protein expression was induced using0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). After in-duction with IPTG, culture was grown at 37 °C for 4 h with con-stant agitation. Induced cells were harvested by centrifugation at8000g for 15 min at 4 °C and analyzed on 15% SDS-PAGE (Sodiumdodecyl sulfate-polyacrylamide gel electrophoresis) to confirm theprotein expression and solubility.

2.4.2. Purification of CVCP proteinThe CVCP protein was purified using cell pellet from 1 L culture.

Frozen pellets were thawed and re-suspended in 25 ml of lysisbuffer (50 mM Tris-HCl pH 7.6, 10 mM imidazole and 100 mMNaCl) on ice. The re-suspended cells were disrupted using theFrench Press (Constant Systems Ltd, Daventry, England) and cen-trifuged at 16,000xg for 45 min at 4 °C to separate the supernatantand pellet. The soluble protein was purified by IMAC (Immobilizedmetal assisted chromatography) and size exclusion chromato-graphy. The Ni-NTA (Nitrilotriacetic acid) agarose beads (Qiagen,USA) were pre-equilibrated with the binding buffer (50 mM Tris-HCl pH 7.6, 100 mM NaCl) and were loaded with the clarified su-pernatant into the column. The supernatant was incubated withthe pre-equilibrated Ni-NTA agarose beads for 45 min at 4 °C.Consecutive three washes were made with varied concentrationsof Imidazole (W1¼30 mM Imidazole, W2¼50 mM Imidazole andW3¼100 mM Imidazole) along with the binding buffer. Finally theN-terminal His-tagged protein was eluted using 250 mM Imida-zole. The eluted fractions were pooled and incubated with TEVprotease for His-tag cleavage during dialysis against the dialysisbuffer (50 mM Tris-HCl pH 7.6, 20 mM NaCl) at 4 °C for 12 h. Aftercleavage of His-tag, protein samples were reloaded onto Ni-NTAcolumn to remove uncleaved His-tagged protein and His-taggedTEV protease. All the fractions were collected and analyzed using15% SDS-PAGE. The bands of purified protein with and withoutHis-tagged TEV protease is shown in Fig. 1A. For the efficient re-covery of purified protein after dialysis, a dialysis membrane of10 kDa MWCO was used (Thermo Scientific). The flow-throughcontaining His-tag cleaved and purified protein was concentratedto �5 mg/ml using a 3 kDa cutoff Amicon Ultra-15 concentrator(Millipore, Bedford, Massachusetts, USA). Concentrated proteinwas loaded onto HiLoad 16/60 Superdex 75 pg size-exclusionchromatography column (GE Healthcare) using 50 mM Tris-HCl pH7.6, 20 mM NaCl as buffer (filtered and degassed). AKTA purifier(GE Healthcare) purification system kept at 4 °C operated at a flowrate of 0.5 ml/min was used to conduct the gel-filtration chroma-tography step. Gel-filtration fractions were collected and analyzedusing 15% SDS-PAGE. The fractions containing protein of interestwere pooled and again concentrated to �5 mg/ml. The molecularweight marker proteins bovine serum albumin (66 kDa), ovalbu-min (45 kDa), trypsin (23 kDa) and lysozyme (14 kDa) were loadedand run on the size-exclusion HiLoad 16/60 Superdex 75 columnto determine the void volume of the column and to construct thestandard curve. Further, the standard curve was used for esti-mating the molecular weight of the purified proteins. The UVabsorbance spectroscopy at a wavelength of 280 nm was used forestimation of protein concentration using an extinction coefficientof 23,950 M�1 cm�1 for CVCP protein.

2.5. Circular dichroism spectroscopy

In the absence of the atomic structure of CVCP, the secondarystructure of CVCP (residue range: 106–261) was determined usingcircular dichroism (CD) spectroscopy. Protein samples were bufferexchanged into 20 mM potassium phosphate buffer at pH 7.6 and

Table 1List of the compounds with their empirical formula and molecular structure used indocking in to CVCP hydrophobic pocket.

Ligand Molecular formula Molecular structure

Dioxane C4H8O2

Proline C5H9NO2

Picolinic Acid (PCA) C6H5NO2

R. Sharma et al. / Virology 498 (2016) 265–276 267

protein concentration used for CD measurements was 0.1 mg/mlfor each protein sample. The spectra were recorded using PCcontrolled Applied Photophysics (Model Chirascan, UK) spectro-polarimeter in a quartz cell of 0.1 cm from 180 to 260 nm of wa-velength with a bandwidth of 1 nm under continuous nitrogenpurge. Three scans were recorded with a response time 0.25 s perpoint at 25 °C and afterwards the average of these scans was done.The baseline data was also collected for the buffer without proteinand subtracted from the averaged spectrum of protein. The ana-lysis of the data was done using the Dichroweb web server byCDSSTR method (Sreerama and Woody, 2000).

2.6. Isothermal titration calorimetry

Isothermal Titration Calorimetry experiments were conductedusing MicroCal iTC200 (GE Healthcare) to determine the thermo-dynamic parameters of dioxane and picolinic acid binding to CVCPprotein. The titrations were performed at 25 °C in buffer contain-ing 50 mM Tris (pH7.6), 20 mM NaCl. The picolinic acid used in thetitration against CVCP was purchased from Sigma Aldrich (Cat. No.P42800). Picolinic acid (PCA) at the concentration of 12 mM

(Syringe) was titrated in to CVCP protein 50 mM (Cell). Whereasthe concentration of the dioxane used for the titrations was 1 mM(Syringe) against the CVCP protein 50 mM (Cell). The various setparameters used to perform the isothermal titrations of dioxaneand picolinic acid with purified CVCP are shown in Table 2. The ITCdata obtained was subsequently analyzed by MicroCal Origin7.0 software from MicroCal. The ITC data was integrated, dilutioneffects were corrected and concentration was normalized beforethe curve fitted to a one site binding model by using MicroCalOrigin 7.0 software.

2.7. Surface plasmon resonance

All the SPR experiments were performed using BIACORE T200(GE Healthcare, USA) instrument with research grade CM5 sensorchips (Biacore AB, Uppsala, Sweden). The analysis and samplecompartment temperature was set 25 °C for all the binding andkinetic experiments.

2.7.1. Buffer preparationAll the SPR binding and kinetic studies were carried out in

Fig. 1. SDS PAGE (15%) profile of CVCP (residue range: 106–261) purification. Lane 1, molecular-weight markers (kDa); Lane 2, Purified CVCP lane with His-tag (uncleaved)and Lane3, Purified CVCP lane without His-tag (cleaved). (B) Modeled structure of Chikungunya virus capsid protein (CVCP): Cartoon diagram of CVCP showing helix (redcolor), β- sheets (yellow color) and loops (green color) (C) ProSA energy profile of CVCP modeled structure. The protein folding energy is in good agreement with the ProSAplot (Z-score �6.57). (D) The cartoon representation of superimposition of CVCP homology model and SFCP template. CVCP and SFCP are shown in green and skyblue color,respectively.

R. Sharma et al. / Virology 498 (2016) 265–276268

Phosphate buffer saline (PBS). The PBS buffer, 10X (BR-1006–72)was purchased from GE-Healthcare. Distilled deionized water wasfiltered using 0.22 mm Millipore filters and degassed using Milli-pore Degassing unit to avoid microbubbles formation. As theanalytes studied for binding and kinetic studies were small mo-lecules, therefore 0.5% Tween20 (P9416-Molecular Biology grade)as a surfactant was added to the PBS 1Xbuffer. The final con-centration of 1X PBSþP includes (20 mM NaH2PO4-Na2HPO4,150 mM NaCl, 0.5% (v/v) surfactant Tween20). This buffer wasused for diluting the analyte stock solutions followed by pre-paration of samples and buffer blanks.

2.7.2. Immobilization of CVCP protein on CM5 sensor chipThe Amine Coupling Kit was purchased from GE Healthcare (GE

Healthcare, BR-1000–50). The dextran surface of the CM5 sensorchip was first activated with 1:1 mixture of 0. M 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and 0.1 M N-hydro-xysuccinimide (NHS) with a flow rate of 10 ml/min to form reactivesuccinimide esters. For immobilization of ligand on the sensor chipsurface, the CVCP protein (100 mg/ml) was dissolved in low saltbuffer (0.1 M Sodium acetate buffer) with pH 4.5. After thistreatment ligand acquired positive charge and got pre-con-centrated onto negatively charged carboxymethyl dextran matrixof CM-Chip that further increase the efficiency of amine coupling.The CVCP protein was injected and flowed across the activatedsensor surface with flow rate of 10 ml/min. After the successfulimmobilization of ligand on the sensor surface, all the remainingactive succinimide esters were blocked using ethanolamine. Theligand (CVCP protein) was immobilized on flow cell 2 (FC-2) andthe final immobilization level was 4500 RU. For all the SPR mea-surements, the flow cell 1 was blank immobilized (without CVCP)and used as a reference. The stability of CVCP protein on the FC-2surface was confirmed by five washes with regeneration buffer(10 mM Glycine-HCL, pH 2.5). The flow rate was kept constantthroughout the binding (10 ml/min) and kinetics experiment(45 ml/min). The data was collected by Biacore control softwareand analysis was done with T200 evaluation software.

2.8. Fluorescence spectroscopy

All the fluorescence experiments were carried out using a Hi-tachi F-4600 fluorescence spectrophotometer. The quartz cuvetteof dimensions 10 mm�4 mm was used for all the fluorescencemeasurements, recorded at 25 °C. To measure the intrinsic fluor-escence of the CVCP, the protein sample (25 mM) were excited atwavelength of 280 nm and the emission spectra were recorded at

wavelength of 295–550 nm. The excitation and emission slit widthof 2.5 nm was used for recording the spectra. Each spectra ob-tained was average of three scans. The buffer containing 50 mMTris, pH¼7.6 and 20 mM NaCl was used for preparing the proteinsamples. The effect of picolinic acid (PCA) on the CVCP intrinsicfluorescence was studied. The emission spectra was recorded uponaddition of PCA at different concentrations up to 5 mM (0.5 mM,1.0 mM, 1.5 mM, 2.0 mM and 5.0 mM) to the CVCP (25 mM), afterincubation for 15 min. The spectra obtained were compared withCVCP (protein without PCA) spectra and analyzed for quenching.

2.9. Cells and virus propagation

The Vero cells (African green monkey kidney cell line) wasobtained from National Centre for Cell Science (NCCS), Pune, Indiaand were maintained in minimum essential medium (MEM)supplemented with 10% heat inactivated fetal bovine serum (FBS),80 U gentamicin, 2 mM L-glutamine and 1.1 g/L sodium bicarbo-nate in humidified 37 °C, 5% CO2 incubator. An Indian isolate ofChikungunya virus (strain DRDE-07; GenBank: EU372006) be-longing to East Central and South African (ECSA) genotype main-tained at Virology Division, DRDE, Gwalior was used for antiviralstudies. Virus was propagated using standard virus adsorptiontechniques and titer was determined by plaque assay, expressed asplaque forming units per ml (PFU/ml). The virus stocks werestored in �80 °C until further use.

2.9.1. Cytotoxicity testingThe cytotoxic effect of the compound was determined by using

neutral red dye uptake (NRDU) assay in Vero cells. Vero cells wereseeded onto 96-well culture plates at cell density of 1�104 cellsper well and allowed for overnight incubation for cell attachment.Serial two-fold dilution of compound was prepared in MEM from8 mM to 62.5 mM. The plates were incubated in 37 °C, 5% CO2 for72 h. After 72 h incubation, neutral red solution was added intoeach well and incubated for one hour at 37 °C. The absorbance wasmeasured at 540 nm. The maximum nontoxic dose (MNTD) wasdetermined. The TUNEL assay was carried out using in situ CellDeath Detection Kit, Fluorescein (Roche Applied Sciences, Ger-many) with standard protocol to detect apoptosis induced by PCA(16 mM �2 mM) in Vero cells, counterstained with 4′,6-diamidi-no-2-phenylindole (DAPI). This was further verified by Caspase3 apoptosis assay, counterstained with Hoechst 33342 to identifynuclear morphology that allows comparison of apoptotic nucleiwith surrounding normal nuclei that were observed micro-scopically at 36 hpi.

2.9.2. Antiviral assayVero cells were seeded onto 24-well culture plates at cell

density of 1�105 cells per well and incubated overnight. The cellswere infected with CHIKV at MOI of 0.01 except for the cell con-trol. Compound was added 24 h before infection (in pre-treat-ment), 0 h after infection (in simultaneous treatment) and 2 hafter viral adsorption (in post-treatment) to reach a final con-centration of 2 mM. The plates were incubated in 37 °C, 5% CO2 for48 h. The infected supernatant was collected at 24 and 48 hpi forquantitation of both CHIKV RNA and infectious virus by qRT-PCRand plaque assay respectively.

2.9.3. Quantitative real-time RT-PCR assayQuantitative real-time RT-polymerase chain reaction (qRT-PCR)

assay was used to determine the CHIKV load. The CHIKV RNA wasextracted from the supernatant by using QIAamp viral RNA kit(Qiagen, Germany) according to the manufacturer's protocol andthe extracted RNAwas stored at �80 °C until analysis. An in-houseqRT-PCR utilizing SS III Platinum one step qRT-PCR kit (Invitrogen,

Table 2The various set parameters used to perform the Isothermal titrations of dioxaneand picolinic acid with purified CVCP.

Parameters Dioxane Picolinic acid

Parameter Setting Setting

Total number of injections 30 28Cell temperature 25 °C 25 °CReference power 8 ucal/s 8 ucal/sInitial delay 70 s 70 sSyringe concentration 1 mM 12 mMCell concentration 0.05 mM 0.05 mMStirring speed 1000 rpm 1000 rpmVolume of Ist Injection 0.2 μl 0.2 μlDuration of Ist injeiction 0.4 s 0.4 sVolume after Ist injection 1.2 ml 1.4 mlDuration after Ist injection 1.7 s 2.8 sInjection spacing 280 s 240 sFilter period 5 s 5 sFeedback mode/gain High High

R. Sharma et al. / Virology 498 (2016) 265–276 269

USA) in Mx3005 P system (Stratagene, USA) with primers differ-entiating CHIKV E1 segment was performed as described earlier(Agarwal et al., 2013). Standard curve for quantitation was pre-pared by qRT-PCR using serial dilution of known copies number ofin vitro transcribed RNA.

2.9.4. Plaque assayThe inhibitory effect of the compound on CHIKV infection was

also determined by plaque reduction assay. Vero cells were seededonto 24-well culture plates incubated overnight. The 10-fold serialdilution of infected culture supernatant of compound treated-CHIKV cells along with virus control was prepared. The mediumwas discarded and the cell monolayer was infected with differentdilutions and incubated for 1.5 h for virus adsorption. The cellmonolayer was then overlaid with overlay medium containing1.25% methylcellulose and further incubated at 37 °C, 5% CO2 for3 days. The overlay medium was discarded and the plate wasrinsed gently with PBS and fixed with chilled methanol beforestaining with crystal violet and the reduction of viral plaquescompared to control was calculated.

3. Results

3.1. 3D Homology model

Homology search of CVCP capsid protein gave the hits oftwenty four sequences against PDB database. The crystal structureof Semliki Forest virus capsid protein (SFCP) was the first hit with93% sequence identity (PDB �1VCP). The 3D homology model ofCVCP capsid protein was generated using 1VCP as template. Thecartoon representation of the CVCP 3D homology model wasshown in Fig. 1B. The Ramachandran plot analysis of the modeledCVCP 3D structure by PROCHECK, showed 97.3% residues arepresent in the core region, 2.7% in allowed region, and no residuein disallowed region. ProSA (https://prosa.services.came.sbg.ac.at/prosa.php) analysis further confirmed the validation of the mod-eled structure as the folding energy of protein is in good agree-ment with the ProSA plot Fig. 1C. The superimposition of the finalselected 3D structure model of CVCP with the template 1VCPshowed a root mean square deviation (RMSD) of 0.2 Å (Fig. 1D).

As expected, the tertiary structure of CVCP exhibited similargeometry to that of crystal structure of SFCP (1VCP). The predictedstructure of C-terminal domain of Chikungunya virus capsid pro-tein (113–261) consists of two β-barrel subdomains similar tocrystal structure of other alphavirus CP's, reported till date. Thesubdomain I of CVCP (residue range: 113–261) consists of five β-strands and two helices unlike the Semliki Forest virus capsidprotein (residue range: 119–267) subdomain I that consists of sixβ-strands and same number of helices (Choi et al., 1997). Helix α-1is present in between β-3 and β-4 strands and helix α-2 is presentbetween β-5 and loop region connecting the two subdomains. Thesubdomain II consists of seven β-strands and devoid of any helixlike other CP's. The two subdomains are connected by a linker loopregion of 10 residues (Ala175-Glu184). The antiparallel strands ofβ-barrel subdomains [β1 (114–119), β2 (122–130), β3 (133–137), β4(155–157), β5 (161–166) of sub-domain I and β6 (185–188), β7(193–197), β8 (200–204), β9 (214–218), β10 (225–233), β11 (237–245), β12 (250–253) of sub-domain II] are folded in to two Greekkeys, characteristic of chymotrypsin like serine protease. The al-phavirus CP possesses cis-autoproteolytic activity and it cleavesitself from the rest of the structural polyprotein only once, in thevirus life cycle. The further protease activity of CP was inhibited byC-terminal Trp residue as it blocks the active site of the protein(Tong et al., 1993). In between the two β-barrel subdomains theconserved catalytic triad, consisting of Ser213, His139 and Asp161

residues was present (Fig. 2). These catalytic residues architectureare conserved among the alphavirus CP's as well as other serineproteases. This protease domain of capsid protein cleaves theconserved scissile peptide bond between Trp-Ser residues. Afterthe truncation of C-terminal Trp residue, CP protease activity intrans can be restored. Recent studies on FRET based assays haveshown the trans-activity of C-terminal capsid protein (AVCP andCVCP) by using synthetic fluorogenic peptide containing cleavagesite (Aggarwal et al., 2014; Aggarwal et al., 2015). Similar to theother chymotrypsin like serine proteases including alphaviruses,the GDSG motif containing the conserved active site nucleophilicserine residue is structurally conserved in CVCP (211GDSG214) (Choiet al., 1996; Forsell et al., 1995) ( Fig. 2).

3.2. Molecular docking

Docking studies were performed using dioxane, proline andPCA as ligands and their binding in to the hydrophobic pocket ofCVCP was analyzed. Previous studies suggested the role of dioxaneand its derivatives in disrupting the capsid-cdE2 interactions andthus alphaviral replication (Kim et al., 2007, 2005). Recent pub-lished data suggested the proline residue (Pro 405 in AVCP) ofhelix-loop-helix of cdE2 fits into the hydrophobic pocket of CP andis involved in CP-cdE2 PPIs (Aggarwal et al., 2012). The crystalstructures of CP-dioxane complex demonstrate dioxane moleculebinding in the CP hydrophobic pocket that mimics the pyrollidinering of cdE2 proline. Thus, docking of proline and its derivativePCA into the hydrophobic pocket of CVCP was done to analyzeligand binding. The binding energy of proline (�3.63 kcal/mol)and picolinic acid (�3.23 kcal/mol) into the CVCP hydrophobicpocket was found to be more than that of dioxane (�2.86 kcal/mol). Moreover, the inhibition constant obtained in silico were8 mM, 4.31 mM and 2.17 mM for dioxane, PCA and proline, re-spectively (Table 3). The interactions of dioxane, proline and pi-colinic acid with the hydrophobic residues of the CVCP wereanalyzed and shown in Fig. 3. Picolinic acid was found to interactwith CVCP hydrophobic pocket residues through 4 hydrogenbonds. While, dioxane and proline bound to hydrophobic pocketthrough 2 Hydrogen bonds each. This is suggestive of that PCA andits derivatives might be more effective in preventing CP-cdE2 PPIand thus inhibit the alphavirus replication.

3.3. Secondary structure analysis

The secondary structure of the CVCP was determined usingcircular dichroism analysis. The online web server Dichroweb wasused for the data analysis. The deconvoluted data reveal the sec-ondary structure content of CVCP was 5% α-helix, 36% β-sheet, 25%β-turn and 32% random coil. CD results show that purified CVCPused in this study has well defined secondary structures similar toother alphavirus capsid protease and is natively folded.

3.4. Isothermal studies of picolinic acid binding

The binding of dioxane and picolinic acid to the purified CVCPwas studied using MicroCal iTC200 at 25 °C by following the heateffect after injection of these ligands into cell containing protein.The ITC binding isotherms were fitted as single set of site model byusing Origin7 Data Analysis software (GE Healthcare) and shownin Fig. 4. The binding constant of dioxane and PCA were calculated.The thermodynamic analysis of dioxane and picolinic acid bindingto CVCP is shown in Table 4. There was stronger affinity of CVCPprotein for picolinic acid with a KD of 61 mM than dioxane with aKD of 80 mM.

R. Sharma et al. / Virology 498 (2016) 265–276270

3.5. Surface plasmon resonance

To get the quality sensor data the cleaning of instrument withDESORB followed by preconditioning of sensor chip was done.Standard cleaning procedure was performed using 0.5% (w/v) SDSand 50 mM Glycine pH¼9.5 as the running buffer and thenprimed for five to six times using distilled deionized filtered /de-gassed water. The SPR instrument was then kept in the standbymode overnight. Next day, a fresh CM5 sensor chip (research-grade) was docked into the clean instrument followed by pre-conditioning with passage of short pulses of acid, base, and de-naturant over the chip surface that hydrates and cleans the CM5layer (carboxymethyl dextran).

3.5.1. CVCP protein immobilizationWe used standard amine-coupling method for the im-

mobilization of CVCP protein. The final immobilization levelachieved was 4500 RU on the surface of Fow cell 2 (FC-2). Fow cell

1 (FC-1) was selected as a reference to minimize the unwanteddrift and systematic noise. The desired immobilization level wasachieved by using the “aim for immobilization” method within thecontrol software of Biacore. The specified “aim for immobilization”for CVCP protein was 5000 RU and the final immobilized level was4500 RU. In this method a series of short pulses of the CVCPprotein were injected over the activated surface until the requiredimmobilization level was attained. After the immobilization step,the surfaces of both the reference (FC-1) and activated (FC-2) flowchannels of the sensor chip was blocked by injecting 1.0 M etha-nolamine (pH 8.5) with the flow rate of 20 ml/min followed by“postconditioned” with three pulses of 100 mM HCl. These stepswere carried out to wash away any non-covalently bound materialto chip surfaces and to stabilize the immobilized ligand. PCA wasinjected over the immobilized CVCP, at concentrations rangingfrom 0.0625 mM to 2 mM, for binding studies. The responses ob-tained upon binding, from these varied analyte concentrationrange was used to measure the kinetic parameters of the inter-action. Each PCA concentration was tested in duplicate that definesthe activity of immobilized CVCP is retained throughout the ex-periment. Reproducibility of the results rules out any mishandlingand further supports the regeneration buffer used was appropriatefor disrupting PCA/CVCP complex after each cycle without affect-ing the CVCP activity. For kinetic studies the contact time of PCAbinding to the CVCP protein was monitored for 120 s with the flowrate of 30 μl/min which was sufficient to get a nice binding re-sponse. To demonstrate the stability of the base line during the

Fig. 2. Multiple sequence alignment of CVCP with CP's from other alphaviruses: the conserved residues are highlighted in red background. The star under the amino acidsrepresented the residues of the hydrophobic pocket involved in interaction with Dioxane, Proline and Picolinic acid. The highly conserved catalytic residues (His139,Asp161and Ser213 in CVCP) of CP's are represented by solid circles below the residues. The GDSG motif (211GDSG214) containing the conserved active site nucleophilic serineresidue is shown in the box.

Table 3The values of binding energy and inhibition constant as obtained from docking ofdioxane, proline and picolinic acid into the CVCP hydrophobic pocket.

CVCP modeled structure Dioxane Proline Picolinic acid

Binding Energy (kcal/mol) �2.86 �3.63 �3.23Inhibition constant (mM) 8 2.17 4.31

R. Sharma et al. / Virology 498 (2016) 265–276 271

experiment, blank injections (only running buffer) were injectedalong with the PCA injections. The responses obtained for thebuffer injections were flat, indicating the stability of the baseline.

3.5.2. Kinetic analysisKinetic parameters for binding of PCA to the immobilized CVCP

protein were measured. The data obtained was evaluated by theBiacore Evaluation software (Biacore T200 Evaluation Software,Version: 2.0). The data was fitted as per the global fitting of thekinetic model with one state model and stoichiometry 1:1(Fig. 5A). Final experiment reports showed no bulk factor whichensures the proper sample preparations and handling. The valuesof the rate constants obtained from SPR for binding of PCA to CVCPwere comparable to the values obtained from ITC studies. The KD

value obtained for binding of PCA to CVCP from ITC and SPR was61 μM and 0.21 μM, respectively. This difference in rate constantscould be due to steric hindrance caused by immobilization strat-egy, kind of buffers used in performing the experiments, setparameters and principle behind these two biophysical techniques(ITC and SPR).

3.6. Effect of PCA on CVCP intrinsic fluorescence

The intrinsic fluorescence of CVCP was measured in its nativeform (residue range: 106–261) and upon binding to PCA. As sug-gested from docking studies of PCA with CVCP, the PCA binds tothe hydrophobic pocket of CVCP, rather more strongly than pre-viously reported dioxane. The binding of the PCA to CVCP

hydrophobic pocket results in decrease of fluorescence intensitywith the increase in concentration of PCA as shown in Fig. 5B. Thebinding constant for PCA binding to CVCP estimated from ITC andSPR was comparable. Fluorescence studies further validated theability of PCA to bind to CVCP hydrophobic pocket that is expectedto disrupt the CP-cdE2 PPIs.

3.7. Cytotoxicity testing

In order to determine the cytotoxic effect of this compoundbefore subsequent assays, the maximal nontoxic dose (MNTD) wasdetermined on Vero cells by the conventional neutral red dyeuptake assay. The MNTD value of the compound in this study wasfound to be approximately 2 mM that showed more than 90% cellviability. Both TUNEL and Caspase 3/FITC assay followed by DAPIand Hoechst 33342 counterstaining, respectively of PCA treatedVero cells did not show apoptosis at 2 mM PCA. Thus, 2 mM PCAconcentration was used to carry out all the subsequent experi-ments (Fig. S1).

3.8. Antiviral assay

The antiviral activity of the compound against CHIKV was de-termined by using qRT-PCR. Viral RNA was isolated from the in-fected culture supernatant of compound treated and untreatedcells, and was quantitated. The inhibition of CHIKV RNA load asdepicted in Fig. 6A was significantly reduced by up to 2 log in si-multaneous treatment (po0.01). The treatment of infected cells

Fig. 3. Binding analysis of docked compound into CVCP hydrophobic pocket: The diagram showing the binding of (A) dioxane, (B) proline and (C) picolinic acid in thehydrophobic pocket of CVCP. The residues of the hydrophobic pocket involved in hydrophobic interactions are shown in green color and dioxane, proline and picolinic acidare shown in cyan, yellow and magenta color, respectively. The polar interactions of dioxane (2 hydrogen bonds), proline (2 hydrogen bonds) and picolinic acid (4 hydrogenbonds) with the hydrophobic pocket residues are shown in red dashes. Picolinic acid binds to CVCP hydrophobic pocket more strongly than dioxane and proline.

R. Sharma et al. / Virology 498 (2016) 265–276272

with compound showed highly significant inhibition in simulta-neous-treatment up to 48 h compared to virus control. Marginalinhibition was observed in pre-treatment up to 24 h and no in-hibition was observed in post-treatment.

The antiviral activity of the compound against CHIKV was alsoconfirmed by conventional plaque assay. Our results showed thatthe CHIKV load was significantly reduced following compoundtreatment by 60.63% in simultaneous treatment at 48 hpi(po0.01). Approximately 45% reduction was observed in pre-treatment only up to 24 hpi. However, no significant reduction wasseen in post-treatment study (Fig. 6B).

4. Discussion

The genus Alphavirus includes notable human and animal pa-thogens. The members of this genus are divided into two maingroups: Old and New World alphaviruses. They have differentdisease presentation and different mechanism to shut the hosttranscription off. Old World viruses such as Sindbis virus, SemlikiForest virus (SFV), Chikungunya virus (CHIKV) etc. are arthralgic indisease presentation and employ nsp2 protein for shutting the

host transcription off. New World viruses such as Venezuelanequine encephalitis virus (VEEV) and Eastern equine encephalitisvirus (EEEV) etc. are encephalitic in disease presentation andemploy capsid protein for inhibition of host transcription (Gar-mashova et al., 2007; Hahn et al., 1988). The N-terminal domain orRNA binding domain of capsid protein involve in this mechanismof transcription inhibition and not the protease domain. However,over the last few decades, chikungunya disease associated withneurological and haematological complications have been re-ported. Recent outbreaks of chikungunya, from 2005 onwardshave been reported to be associated with encephalitis and me-ningoencephalitis (Chandak et al., 2009; Chusri et al., 2011). There-emergence of chikungunya as an epidemic, evolution of newmosquito vector of the virus (due to mutation in E1 glycoprotein)(Schuffenecker et al., 2006) and the pathogenicity of virus espe-cially encephalitis and meningoencephalitis apart from arthralgiamakes it a significant pathogen which needs proper and im-mediate attention. Till date, there is no vaccine to prevent chi-kungunya disease and no specific medicine to treat it. This ne-cessitates, the study of alphaviral biology for specific anti-CHIKVdrug development.

In the current study, we focused on finding inhibitors that bind

Fig. 4. ITC raw and integrated data: (A) Isothermal Titration Calorimetry (ITC) raw and integrated data for CVCP-Dioxane titrations. The titrations were performed at 25 °C inbuffer containing 50 mM Tris (pH7.6), 20 mM NaCl. Dioxane at the concentration of 1 mM (Syringe) was titrated in to CVCP protein 50 mM (Cell). (B) ITC raw and integrateddata for CVCP-picolinic acid titrations. The temperature and the buffer components were similar to that of CVCP- Dioxane titrations. Picolinic acid at the concentration of12 mM (Syringe) was titrated in to CVCP protein 50 mM (Cell). The experimental data points for the titrations were represented in solid squares and best fit of these datapoints were shown as solid curve by taking one site binding model.

Table 4The thermodynamic analysis of dioxane and picolinic acid binding to CVCP as obtained from ITC.

Protein Ligand N KD (mM) KA (M�1) ΔH (kcal/mol) ΔS (cal/mol/degree)

CVCP Dioxane 1 80 (12.670.9)103 �10407 0.8 �3.4�103

CVCP Picolinic acid 1 61 (16.370.9)103 �44.670.6 �130

R. Sharma et al. / Virology 498 (2016) 265–276 273

the hydrophobic pocket of CP, and obstruct the CP and cdE2 gly-coprotein interactions. The interaction of cdE2 with nucleocapsidcore is critical for the virus life cycle. Although the crystal struc-tures of CP from other alphavirus members like Semliki Forestvirus (SFV), Sindbis virus (SINV), Aura virus etc. have been re-ported, the crystal structure of Chikungunya virus capsid protein isyet to be determined. The cryo-electron microscopy (cryo-EM)

studies of capsid protein (CP) and trans-membrane glycoprotein(cdE2) interactions provide important breakthrough in under-standing this key interaction. Previous studies have found a sol-vent-derived dioxane molecule bound to hydrophobic pocket inthe crystal structure of SCP, AVCP (Aggarwal et al., 2012; Choi et al.,1991) suggesting that dioxane and its derivatives may prevent thecapsid-cdE2 interactions. Many Dioxane derivatives were ex-tensively studied in Sindbis virus and found to have antiviralproperty (Kim et al., 2007; Kim et al., 2005). Alphavirus capsidprotein is a serine protease and possesses a chymotrypsin-like foldwith a hydrophobic pocket at the surface responsible for interac-tion with transmembrane glycoproteins. The crystal structure ofalphavirus capsid protein (AVCP, SCP, and SFCP) has structurallyconserved hydrophobic pocket on the CP surface. Therefore, li-gands that bind the CP hydrophobic pocket more strongly thandioxane are expected to inhibit capsid–cdE2 interaction in variousalphaviruses.

The CP possesses cis-autoproteolytic activity and it cleaves it-self from the rest of the structural polyprotein only once, in thevirus life cycle. The subsequent protease activity of CP was in-hibited by C-terminal Trp residue (AVCP, W-267) as it blocks theP1 position near the active site. After the truncation of C-terminalTrp residue, CP protease activity in trans can be restored (Aggarwalet al., 2014). Like other serine protease, CHIKV CP contains a cat-alytic triad formed of His139, Ser213 and Asp161 residues. Thetriad is located in the cleft between the β-barrel sub-domains. Dueto the absence of crystal structure of Chikungunya CP (CVCP),homology models were generated by modeler 9.14. After homol-ogy modeling the 3D structure obtained was validated and energyminimized. AutoDock 4.2, was used for studying the docking ofdioxane and PCA into the hydrophobic pocket of CVCP. The bindingenergy of picolinic acid (�3.23) into the CVCP hydrophobic pocketwas found to be more than that of dioxane (�2.86). Moreover, theinhibition constant of these ligands for CVCP hydrophobic pocketinteractions were calculated by AutoDock 4.2. The values for theinhibition constant obtained were 8 mM and 4.31 mM for dioxaneand PCA, respectively. It clearly reflects that PCA binds to the CPhydrophobic pocket more strongly and efficiently than the pre-viously reported dioxane. The thermodynamic parameters of thebinding were calculated based on the isothermal titration

Fig. 5. (A) SPR binding and kinetic analysis: The relative responses obtained for the binding of analyte (PCA) to the immobilized CVCP protein. The data was fitted by 1:1binding state model. The KD value obtained for the binding of PCA to immobilized CVCP was 0.21 mM. (B) Fluorescence spectroscopy: The intrinsic fluorescence of CVCPnative protein and upon binding to PCA was determined. The emission spectra was recorded upon addition of PCA at different concentrations up to 5 mM (0.5 mM, 1.0 mM,1.5 mM, 2.0 mM and 5.0 mM) to the native CVCP (25 mM), after incubation for 15 min. The PCA was supposed to bind to CVCP hydrophobic pocket and thereby decreased thefluorescence intensity with the increase in concentration of PCA.

Fig. 6. Antiviral activity of PCA in Vero cells. (A) Viral RNA copy number by qRT-PCR. Vero cells were treated with PCA in pre-treatment, simultaneous treatmentand post-treatment modes (infected with 0.01 MOI of CHIKV). Cell supernatantswere harvested at 24 and 48 hpi and RNA was isolated. The qRT-PCR was per-formed with specific primers for CHIKV E1 region. Titer of CHIKV RNA was de-termined from a standard curve drawn using 10-fold serially diluted in vitrotranscribed CHIKV RNA. (B) Relative CHIKV titer as assessed by plaque assay at 24and 48 hpi. Percent inhibition with respect to virus control was calculated. Datarepresents the mean 7 standard deviation of three independent experiments. Theasterisk indicates statistical significance (*po0.05,**po0.01).

R. Sharma et al. / Virology 498 (2016) 265–276274

calorimetry and surface plasmon resonance. The binding constant(KD) calculated for the binding of dioxane and PCA (picolinic acid)to CVCP hydrophobic pocket from ITC was 80 mM and 61 mM re-spectively. This result was further validated by SPR. The KD value ofPCA binding to CVCP hydrophobic pocket obtained from SPR ki-netic experiment was 0.21 mM. These results of PCA binding in-dicate that other small molecules related to PCA and its derivativesthat bind to the hydrophobic binding pocket of the alphavirusescould effectively inhibit the viral replication cycle by blockingcapsid assembly as well as capsid-cdE2 interactions.

Prior to the estimation of antiviral properties of PCA, its cyto-toxic effect has been determined. Cell viability assay revealed thatMNTD value of the PCA, was approximately 2 mM in Vero cellsthat showed more than 90% cell viability at 72 h. Further, thisconcentration (2 mM) of PCA did not show apoptosis as de-termined using TUNEL and Caspase 3 apoptosis assay. In the pre-sent investigation, we determined the inhibitory effects of PCAagainst CHIKV in the pre, simultaneous and post-treatment in Verocells. The computational modeling studies indicated that PCAbinds to the hydrophobic region of CHIKV capsid protein, whichwas also confirmed by fluorescence spectroscopy and SPR results.Thus, hypothesized that PCA should have antiviral activity againstCHIKV via inhibiting the disassembly and assembly functions ofthe CP or the virus budding process. Our results showed that pre-treatment of cells with PCA had partial effect on CHIKV replicationthat supports the hypothesis that PCA is inhibiting the initialdisassembly. Whereas, in the simultaneous treatment, PCA in-hibited CHIKV up to 48 hpi by affecting virus fusion and assemblyprocess. Interestingly, no antiviral effects were observed in thepost-treatment, indicating that PCA does not interfere with theprocess of virus release, confirming that the main inhibitory effectof PCA is at the step of disassembly, replication and/or assembly ofnucleocapsids. However, it is observed that the post-treatmentafter 24 hpi, but not after 48 hpi, inhibited viral RNA abundance.This may be representing the inhibition of re-infection by newlyreleased virions at the early disassembly step. Interestingly, at 48hpi no inhibition at RNA level is observed and this could be due tooverabundance of capsid protein. Therefore, at this stage enoughamount of PCA is not available to saturate the hydrophobic pocketsof all the CP molecules and thus, not able to inhibit CHIKV. Theseresults clearly demonstrate the antiviral activity of PCA againstCHIKV, supporting the postulate that PCA binds to the hydro-phobic pocket of CVCP and inhibits its function. However, theexact mechanism by which PCA inhibits CHIKV replication will beelucidated by performing additional in vitro and in vivo capsidassembly and disassembly experiments. In nut shell PCA at con-centration 2 mM significantly inhibited CHIKV replication in si-multaneous-treatment mode, decreasing viral mRNA and viralload as assessed by quantitative real-time RT-PCR and plaque re-duction assay respectively, in infected Vero cells. Hence, this studypaves the way for developing more precise and efficient antiviralsagainst re-emerging CHIKV and other pathogenic alphaviruses. Ofparticular note, outstanding inhibitory effects against alphavirusesmay be observed following treatment with PCA in combinationwith other antiviral drugs. However, further in vivo experimentalstudies are required to investigate their potential application forclinical intervention of alphaviruses.

5. Conclusions

Molecular docking results showed stronger binding of the PCA(�3.23 kcal/mol) to the hydrophobic pocket than that of the di-oxane (�2.86 kcal/mol). The binding constant (KD) calculatedusing isothermal titration calorimetry (ITC) for dioxane and PCAwere 80 mM and 61 mM respectively confirming the results of

molecular docking. The binding kinetics of PCA with CVCP werefurther studied using surface plasmon resonance (SPR). Thebinding constant KD (M) obtained for PCA was 2.1�10�7 M. Theintrinsic fluorescence spectroscopy confirms the binding of PCA inthe hydrophobic pocket. Additionally, PCA was found to have in-hibitory activity against CHIKV replication in Vero cells, decreasingviral mRNA and viral load as assessed by qRT-PCR and plaque re-duction assay, respectively. This study suggests antiviral potentialof heterocyclic ring compounds similar to PCA and its derivativesagainst CHIKV and alphaviruses.

Acknowledgments

ST acknowledges the financially support from Defence Researchand Development Establishment (DRDE) Proj. No. DRDE/TC/05414/sub-Project/DRD-203/01/16. Department of Biotechnology (DBT) isacknowledged by RS for his Senior Research Fellowship (SRF). Wethank the Macromolecular Crystallographic Facility (MCU) at IIC,Indian Institute of Technology Roorkee.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.virol.2016.08.029.

References

Agarwal, A., Singh, A.K., Sharma, S., Soni, M., Thakur, A.K., Gopalan, N., Parida, M.M.,Rao, P.V.L., Dash, P.K., 2013. Application of Real-time RT-PCR in vector surveil-lance and assessment of replication kinetics of an emerging novel ECSA gen-otype of Chikungunya virus in Aedes aegypti. J. Virol. Methods 193, 419–425.

Aggarwal, M., Dhindwal, S., Kumar, P., Kuhn, R.J., Tomar, S., 2014. trans-Proteaseactivity and structural insights into the active form of the alphavirus capsidprotease. J. Virol. 88, 12242–12253.

Aggarwal, M., Sharma, R., Kumar, P., Parida, M., Tomar, S., 2015. Kinetic character-ization of trans-proteolytic activity of Chikungunya virus capsid protease anddevelopment of a FRET-based HTS assay. Sci. Rep., 5.

Aggarwal, M., Tapas, S., Siwach, A., Kumar, P., Kuhn, R.J., Tomar, S., 2012. Crystalstructure of aura virus capsid protease and its complex with dioxane: newinsights into capsid-glycoprotein molecular contacts. PloS One 7, e51288.

Baron, S., Schmaljohn, A.L., McClain, D., 1996. Alphaviruses (Togaviridae) and fla-viviruses (flaviviridae).

Cavrini, F., Gaibani, P., Pierro, A.M., Rossini, G., Landini, M.P., Sambri, V., 2009.Chikungunya: an emerging and spreading arthropod-borne viral disease. J.Infect. Dev. Ctries 3, 744–752.

Chandak, N.H., Kashyap, R.S., Kabra, D., Karandikar, P., Saha, S.S., Morey, S.H., Pur-ohit, H.J., Taori, G.M., Daginawala, H.F., 2009. Neurological complications ofChikungunya virus infection. Neurol. India 57, 177.

Cheng, R.H., Kuhn, R.J., Olson, N.H., Rossmann, M.G., Choi, H.-K., Smith, T.J., Baker, T.S., 1995. Nucleocapsid and glycoprotein organization in an enveloped virus. Cell80, 621–630.

Choi, H.-K., Lee, S., Zhang, Y.-P., McKinney, B.R., Wengler, G., Rossmann, M.G., Kuhn,R.J., 1996. Structural analysis of Sindbis virus capsid mutants involving as-sembly and catalysis. J. Mol. Biol. 262, 151–167.

Choi, H.-K., Lu, G., Lee, S., Wengler, G., Rossmann, M.G., 1997. Structure of SemlikiForest virus core protein. Proteins 27, 345–359.

Choi, H.-K., Tong, L., Minor, W., Dumas, P., Boege, U., Rossmann, M.G., Wengler, G.,1991. Structure of Sindbis virus core protein reveals a chymotrypsin-like serineproteinase and the organization of the virion. Nature 354, 37–43.

Chusri, S., Siripaitoon, P., Hirunpat, S., Silpapojakul, K., 2011. Case reports of neuro-Chikungunya in southern Thailand. Am. J. Trop. Med. Hyg. 85, 386–389.

Corrêa, P.R.C., da Silveira Pinto, A., de Fátima Silva, G.D., Vieira Filho, S.A., Duarte, L.P., Yamada-Ogatta, S.F., 2011. In vitro Activity of 2-pyridinecarboxylic Acidagainst Trypanosomes of the Subgenus Schizotrypanum Isolated from the BatPhyllostomus hastatus. Acta Sci. Biol. Sci. 33, 437–443.

Fernandez-Pol, J.A., Fernandez-Pol, S., 2008. Method to control dengue viruses inhumans by picolinic acid and derivatives thereof. Google Pat.

Fernandez-Pol, J.A., Klos, D.J., Hamilton, P.D., 2000. Antiviral, cytotoxic and apop-totic activities of picolinic acid on human immunodeficiency virus-1 and hu-man herpes simplex virus-2 infected cells. Anticancer Res. 21, 3773–3776.

Forsell, K., Suomalainen, M., Garoff, H., 1995. Structure-function relation of theNH2-terminal domain of the Semliki Forest virus capsid protein. J. Virol. 69,1556–1563.

R. Sharma et al. / Virology 498 (2016) 265–276 275

Garmashova, N., Gorchakov, R., Volkova, E., Paessler, S., Frolova, E., Frolov, I., 2007.The Old World and New World alphaviruses use different virus-specific pro-teins for induction of transcriptional shutoff. J. Virol. 81, 2472–2484.

Hahn, C.S., Lustig, S., Strauss, E.G., Strauss, J.H., 1988. Western equine encephalitisvirus is a recombinant virus. Proc. Natl. Acad. Sci. USA 85, 5997–6001.

Hong, E.M., Perera, R., Kuhn, R.J., 2006. Alphavirus capsid protein helix I controls acheckpoint in nucleocapsid core assembly. J. Virol. 80, 8848–8855.

Jadav, S.S., Sinha, B.N., Hilgenfeld, R., Pastorino, B., de Lamballerie, X., Jayaprakash,V., 2015. Thiazolidone derivatives as inhibitors of chikungunya virus. Eur. J.Med. Chem. 89, 172–178.

Jose, J., Przybyla, L., Edwards, T.J., Perera, R., Burgner, J.W., Kuhn, R.J., 2012. Inter-actions of the cytoplasmic domain of Sindbis virus E2 with nucleocapsid corespromote alphavirus budding. J. Virol. 86, 2585–2599.

Jose, J., Snyder, J.E., Kuhn, R.J., 2009. A structural and functional perspective of al-phavirus replication and assembly. Future Microbiol. 4, 837–856.

Kim, H.Y., Kuhn, R.J., Patkar, C., Warrier, R., Cushman, M., 2007. Synthesis of diox-ane-based antiviral agents and evaluation of their biological activities as in-hibitors of Sindbis virus replication. Bioorg. Med. Chem. 15, 2667–2679.

Kim, H.Y., Patkar, C., Warrier, R., Kuhn, R., Cushman, M., 2005. Design, synthesis,and evaluation of dioxane-based antiviral agents targeted against the Sindbisvirus capsid protein. Bioorg. Med. Chem. Lett. 15, 3207–3211.

Kostyuchenko, V.A., Jakana, J., Liu, X., Haddow, A.D., Aung, M., Weaver, S.C., Chiu,W., Lok, S.-M., 2011. The structure of barmah forest virus as revealed by cryo-electron microscopy at a 6-angstrom resolution has detailed transmembraneprotein architecture and interactions. J. Virol. 85, 9327–9333.

Larrieu, S., Pouderoux, N., Pistone, T., Filleul, L., Receveur, M.-C., Sissoko, D., Ezze-dine, K., Malvy, D., 2010. Factors associated with persistence of arthralgiaamong Chikungunya virus-infected travellers: report of 42 French cases. J. Clin.Virol. 47, 85–88.

Lopez, S., Yao, J.-S., Kuhn, R.J., Strauss, E.G., Strauss, J.H., 1994. Nucleocapsid-gly-coprotein interactions required for assembly of alphaviruses. J. Virol. 68,1316–1323.

Mancini, E.J., Clarke, M., Gowen, B.E., Rutten, T., Fuller, S.D., 2000. Cryo-electronmicroscopy reveals the functional organization of an enveloped virus, SemlikiForest virus. Mol. Cell 5, 255–266.

Mavalankar, D., Shastri, P., Bandyopadhyay, T., Parmar, J., Ramani, K.V., 2008. In-creased mortality rate associated with chikungunya epidemic, Ahmedabad,India. Emerg. Infect. Dis. 14, 412–415.

Morrison, T.E., 2014. Reemergence of chikungunya virus. J. Virol. 88, 11644–11647.Mukhopadhyay, S., Zhang, W., Gabler, S., Chipman, P.R., Strauss, E.G., Strauss, J.H.,

Baker, T.S., Kuhn, R.J., Rossmann, M.G., 2006. Mapping the structure and func-tion of the E1 and E2 glycoproteins in alphaviruses. Structure 14, 63–73.

Nimmannitya, S., Halstead, S.B., Cohen, S.N., Margiotta, M.R., 1969. Dengue andchikungunya virus infection in man in Thailand, 1962-64. I. Observations onhospitalized patients with hemorrhagic fever. Am. J. Trop. Med. Hyg. 18,954–971.

Owen, K.E., Kuhn, R.J., 1996. Identification of a region in the Sindbis virus nucleo-capsid protein that is involved in specificity of RNA encapsidation. J. Virol. 70,2757–2763.

Owen, K.E., Kuhn, R.J., 1997. Alphavirus budding is dependent on the interactionbetween the nucleocapsid and hydrophobic amino acids on the cytoplasmicdomain of the E2 envelope glycoprotein. Virology 230, 187–196.

Padbidri, V.S., Gnaneswar, T.T., 1978. Epidemiological investigations of chikungunyaepidemic at Barsi, Maharashtra state, India. J. Hyg. Epidemiol. Microbiol. Im-munol. 23, 445–451.

Perera, R., Owen, K.E., Tellinghuisen, T.L., Gorbalenya, A.E., Kuhn, R.J., 2001. Al-phavirus nucleocapsid protein contains a putative coiled coil α-helix importantfor core assembly. J. Virol. 75, 1–10.

Powers, A.M., Brault, A.C., Shirako, Y., Strauss, E.G., Kang, W., Strauss, J.H., Weaver, S.C., 2001. Evolutionary relationships and systematics of the alphaviruses. J. Virol.

75, 10118–10131.Preziosi, P., 2007. Isoniazid: metabolic aspects and toxicological correlates. Curr.

Drug Metab. 8, 839–851.Ross, R.W., 1956. The Newala epidemic: III. The virus: isolation, pathogenic prop-

erties and relationship to the epidemic. J. Hyg. 54, 177–191.Rupp, J.C., Jundt, N., Hardy, R.W., 2011. Requirement for the amino-terminal domain

of Sindbis virus nsP4 during virus infection. J. Virol. 85, 3449–3460.Schuffenecker, I., Iteman, I., Michault, A., Murri, S., Frangeul, L., Vaney, M.-C., La-

venir, R., Pardigon, N., Reynes, J.-M., Pettinelli, F., 2006. Genome microevolutionof chikungunya viruses causing the Indian Ocean outbreak. PLoS Med. 3, e263.

Sergon, K., Njuguna, C., Kalani, R., Ofula, V., Onyango, C., Konongoi, L.S., Bedno, S.,Burke, H., Dumilla, A.M., Konde, J., 2008. . Seroprevalence of chikungunya virus(CHIKV) infection on Lamu Island, Kenya, October 2004. Am. J. Trop. Med. Hyg.78, 333–337.

Sergon, K., Yahaya, A.A., Brown, J., Bedja, S.A., Mlindasse, M., Agata, N., Allaranger, Y.,Ball, M.D., Powers, A.M., Ofula, V., 2007. Seroprevalence of Chikungunya virusinfection on Grande Comore Island, union of the Comoros, 2005. Am. J. Trop.Med. Hyg. 76, 1189–1193.

Shah, K.V., Gibbs Jr, C.J., Banerjee, G., 1964. Virological Investigation of the Epidemicof Haemorrhagic Fever in Calcutta: Isolation of Three Strains of ChikungunyaVirus. Indian J. Med. Res. 52, 676–683.

Soonsawad, P., Xing, L., Milla, E., Espinoza, J.M., Kawano, M., Marko, M., Hsieh, C.,Furukawa, H., Kawasaki, M., Weerachatyanukul, W., 2010. Structural evidence ofglycoprotein assembly in cellular membrane compartments prior to Alphavirusbudding. J. Virol. 84, 11145–11151.

Sourisseau, M., Schilte, C., Casartelli, N., Trouillet, C., Guivel-Benhassine, F., Rud-nicka, D., Sol-Foulon, N., Le Roux, K., Prevost, M.-C., Fsihi, H., 2007. Character-ization of reemerging chikungunya virus. PLoS Pathog. 3, e89.

Sreerama, N., Woody, R.W., 2000. Estimation of protein secondary structure fromcircular dichroism spectra: comparison of CONTIN, SELCON, and CDSSTRmethods with an expanded reference set. Anal. Biochem. 287, 252–260.

Strauss, J.H., Strauss, E.G., 1994. The alphaviruses: gene expression, replication, andevolution. Microbiol. Rev. 58, 491–562.

Tang, J., Jose, J., Chipman, P., Zhang, W., Kuhn, R.J., Baker, T.S., 2011. Molecular linksbetween the E2 envelope glycoprotein and nucleocapsid core in Sindbis virus. J.Mol. Biol. 414, 442–459.

Thiboutot, M.M., Kannan, S., Kawalekar, O.U., Shedlock, D.J., Khan, A.S., Sarangan, G.,Srikanth, P., Weiner, D.B., Muthumani, K., 2010. Chikungunya: a potentiallyemerging epidemic? PLoS Negl. Trop. Dis. 4, e623.

Tomar, S., Hardy, R.W., Smith, J.L., Kuhn, R.J., 2006. Catalytic core of alphavirusnonstructural protein nsP4 possesses terminal adenylyltransferase activity. J.Virol. 80, 9962–9969.

Tong, L., Wengler, G., Rossmann, M.G., 1993. Refined structure of Sindbis virus coreprotein and comparison with other chymotrypsin-like serine proteinasestructures. J. Mol. Biol. 230, 228–247.

Vasiljeva, L., Merits, A., Auvinen, P., Kääriäinen, L., 2000. Identification of a NovelFunction of the AlphavirusCapping Apparatus RNA 5′-TRIPHOSPHATASE AC-TIVITY OF Nsp2. J. Biol. Chem. 275, 17281–17287.

Wang, E., Volkova, E., Adams, A.P., Forrester, N., Xiao, S.-Y., Frolov, I., Weaver, S.C.,2008. Chimeric alphavirus vaccine candidates for chikungunya. Vaccine 26,5030–5039.

Weaver, S.C., Winegar, R., Manger, I.D., Forrester, N.L., 2012. Alphaviruses: popu-lation genetics and determinants of emergence. Antivir. Res 94, 242–257.

Zhang, R., Hryc, C.F., Cong, Y., Liu, X., Jakana, J., Gorchakov, R., Baker, M.L., Weaver, S.C., Chiu, W., 2011. 4.4 Å cryo-EM structure of an enveloped alphavirus Vene-zuelan equine encephalitis virus. EMBO J. 30, 3854–3863.

Zhang, W., Mukhopadhyay, S., Pletnev, S.V., Baker, T.S., Kuhn, R.J., Rossmann, M.G.,2002. Placement of the structural proteins in Sindbis virus. J. Virol. 76,11645–11658.

R. Sharma et al. / Virology 498 (2016) 265–276276