Isolation and molecular characterization of dengue virus clinical isolates from pediatric patients in New Delhi
Meenakshi Kar
a,1, Amul Nisheetha
c,1, Anuj Kumar
c,1, Suraj Jagtap
d,1, Jitendra Shinde
a, Mohit Singla
b, Saranya M
e, Awadhesh Pandit
c, Anmol Chandele
f, Sushil K. Kabra
b, Sudhir Krishna
c, Rahul Roy
d,e,g, Rakesh Lodha
b, Chitra Pattabiraman
h,
Guruprasad R. Medigeshi
a,*
aTranslationalHealthScienceandTechnologyInstitute,Faridabad,Haryana,India
bDepartmentofPediatrics,AllIndiaInstituteofMedicalSciences,NewDelhi,India
cNationalCentreforBiologicalSciences,TIFR,Bengaluru,India
dDepartmentofChemicalEngineering,IndianInstituteofScience,Bengaluru,India
eMolecularBiophysicsUnit,IndianInstituteofScience,Bengaluru,India
fICGEB-EmoryVaccineCenter,ICGEBCampus,NewDelhi,India
gCenterforBiosystemsScienceandEngineering,IndianInstituteofScience,Bengaluru,India
hNationalInstituteofMentalHealthandNeurosciences,Bengaluru,India
ARTICLE INFO
Articlehistory:
Received26October2018
Receivedinrevisedform1December2018 Accepted4December2018
Keywords:
Denguevirus Growthcurve Phylogenetics Genotype Mutations
ABSTRACT
Objective:Tocharacterizetheinvitroreplicationfitness,viraldiversity,andphylogenyofdengueviruses (DENV)isolatedfromIndianpatients.
Methods:DENVwasisolatedfromwholebloodcollectedfrompatientsbypassagingincellculture.
Passage3viruseswereusedforgrowthkineticsinC6/36mosquitocells.Paralleleffortsalsofocusedon theisolationofDENVRNAfromplasmasamplesofthesamepatients,whichwereprocessedfornext- generationsequencing.
Results:Itwaspossibletoisolate64clinicalisolatesofDENV,mostlyDENV-2.Twenty-fiveofthesewere furtherusedforgrowthcurveanalysisinvitro,whichshowedawiderangeofreplicationkinetics.The highestviraltiterswereassociatedwithisolatesfrompatientswithdenguewithwarningsignsand severe denguecases.Fullgenome sequencesof21 DENVisolateswereobtained.Genome analysis mappedthecirculatingDENV-2strainstotheCosmopolitangenotype.
Conclusions:Thereplicationkineticsofisolatesfrompatientswithmildorsevereinfectiondidnotdiffer significantly,buttheviraltitersvariedbytwoordersofmagnitudebetweentheisolates,suggesting differencesinreplicationfitnessamongthecirculatingDENV-2.
©2018TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by- nc-nd/4.0/).
Introduction
Denguevirus(DENV)isacommonarbovirusthatcausesabout60 millionapparentinfectionseveryyear,leadingtoover10000deaths.
Thenumberofdenguecasesdoubledevery10yearsbetween1990 and2013(Stanawayetal.,2016).Antibody-dependentenhancement ofviralinfectionisamajordriverofseveredengueinfections,but
pre-existingantibodiesdonotalwaysleadtosevereinfectionforall serotypes (Endyetal.,2004).The antibodyresponseinprimary infections,whichdoesprovidecross-protectionalbeittransiently, hasalsobeenshowntodifferbetweenserotypes,suggestingthatthe genotypeandserotypeofDENVisolateselicitdifferentialimmune responses(Chaudhuryetal.,2017;Claphametal.,2016;VanBlargan et al., 2013). Serotype-specific differences in efficacy havebeen reportedfortheonly licensedlive-attenuateddenguevaccine,which isconstructedonayellowfevervirusbackbone(Capedingetal., 2014).Someofthereasonsforthesedifferencescouldbethelackof T-cell epitopes of DENV non-structural regions and possible divergencein sequencebetween thestrains usedinthevaccine andthecirculatingserotypesintheendemiccountrieswherethe vaccinetrialswereconducted(Guyetal.,2015;Juraskaetal.,2018).
* Corresponding author at: Translational Health Science and Technology Institute,NCR-BiotechScienceCluster,3rdMilestone,Faridabad-GurgaonHighway, POBox#4,Faridabad121001,Haryana,India.
E-mailaddress:gmedigeshi@thsti.res.in(G.R. Medigeshi).
1 MeenakshiKar,AmulNisheetha,AnujKumar,andSurajJagtapcontributed equally.
https://doi.org/10.1016/j.ijid.2018.12.003
1201-9712/©2018TheAuthor(s).PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).
ContentslistsavailableatScienceDirect
International Journal of Infectious Diseases
j o u r n a l h o m ep a g e : w w w . e l s e v i e r . c o m / l o c a te / i j i d
Weandothers havereportedthatasignificantproportionof primary dengue cases also result in severe dengue disease in endemiccountries,suggestingthatthecirculatingviralstrainsand intrinsic host responses play an important role in disease outcomes(Balmasedaetal.,2006;NgweTunetal.,2013;Nunes etal.,2018;Singlaetal., 2016).Cohortstudiestrackingdengue epidemicsovermanyyearshaveidentifiedviralevolutionasakey playerthatimpactsserotype-specificresponses.Differentcladesof DENV-2 were found to lead to different disease outcomes in childrenwhohadpriorimmunitytoDENV-1orDENV-3(OhAinle etal.,2011).GeneticvariationinDENVsequencesisofsignificance forbothepidemiologyandvaccinedevelopment.Ithasbeenwell documentedin manyearlierreportsthatthereplicationfitness, infectiondynamics,andpathogenicityvaryamongDENVisolates (Balmasedaetal.,2006;Borgesetal.,2018;Fontaineetal.,2018;
Leitmeyeretal.,1999;Vaughnetal.,2000;Wattsetal.,1999).
Indiacontributesabout34%oftheglobalDENVcasesandthere islimitedinformationontheevolutionofDENVstrainscirculating inIndia(Bhattetal.,2013;Diasetal.,2018).Consideringtherecent progress in dengue vaccine development, it is important to understandthedriversofpositiveselectionandtheevolutionof DENV strains in India to better prepare for vaccine trials (Wichmannet al.,2017).In thisstudy,DENV wasisolatedfrom patientbloodandviralgrowthkineticswerecharacterizedinvitro;
furthermore, complete DENV genomes were recovered from plasmaandcompared, inorder tocreateamolecularprofile of theDENVstrainscirculatinginIndia.
Materialsandmethods
Thestudypopulation,isolationofviralRNAfromwholeblood or plasma, and estimation of viremia have been described previously; a description is provided in the Supplementary Material.
Virusisolationfromwholeblood
Bloodsampleswerediluted1:10inminimalessentialmedium (MEM)containing2%fetalbovineserum(FBS)(Gibco)andadded toaculturedmonolayerofC6/36cellsfor1hwithgentlerockingin a28CCO2incubator.After1h,theinoculumwas removedand cellswereculturedfor7daysinMEMwith2%FBSandantibiotics.
Theculturesupernatantwascollectedonday7andusedtoinfecta newbatchofC6/36cells,asabove.Thiswasrepeatedthreetimesto achievepassage3(P3)isolates.Viraltitersinthesupernatantwere measured by plaque assays on BHK-21 cells, as described previously(Agrawaletal.,2013).
Wholegenomesequencing
SpecificcDNAsynthesiswasperformedusingaprimerspecific for the 30 untranslated region (UTR) of flaviviruses (50 GGGTCTCCWCTAACCTCTAGTCCT30)usingtheThermofisherSci- entific Maxima H Minus Reverse Transcriptase. Second strand synthesis was performed using the NEBNext Second Strand Synthesis module. Resulting DNA was purified using magnetic beads.Sequencinglibrarypreparationwas performedusingthe NexteraXTIlluminasequencingkitwiththe96barcodes.About10 pMDNAfrompooledlibraries wastaken forsequencingofthe MiSeq platform. Fastq reads were extracted from the run, demultiplexed, and used for de novo assembly using SPAdes (ver.3.12.0)(Nurketal.,2013).
Reference recruitment tothe closestBlastn hitand genome editingwasperformedusingGeneioussoftware(ver.11.15),with coverageofaleast9andmajorityruleforcallingtheconsensus.
Resulting consensus sequences were annotated by transferring
annotationsfromthereferencesequence(RefSeq)andsubmitted toGenBank.TheGenBankaccessionnumbersareprovidedinthe SupplementaryMaterial.
Molecularphylogeneticanalysis
Multiplesequencealignmentwasperformedforrepresentative sequences from known genotypes (obtained from the Virus Pathogen Database and Analysis Resource, ViPR), for which complete genomes areavailable (the accessionnumbers of the sequences are provided in Table T4 of the Supplementary Material),usingMUSCLE.Thephylogenetictreesforwholegenome sequences were constructed using the maximum likelihood methodinMEGAX software(Kumar etal.,2018).Initialtree(s) fortheheuristicsearchwereobtainedautomaticallybyapplying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the maximum composite likelihood (MCL)approachandthenselectingthetopologywithsuperiorlog- likelihoodvalue.Bootstrappingwith1000replicateswasusedto improvethestatisticalsignificanceofeachnode.Around91%ofthe nodesinalltreeshadbootstrapvalues>90%.Thephylogenetictree for theenvelope proteinof DENV-2 was constructedusing the neighbor-joiningmethodwith100bootstraps.FigTree(v1.4.3)was usedforgraphicalpresentationofthephylogenetictrees.
Statisticalmethods
Appropriatestatisticaltestswereperformed,asindicatedinthe figure legends and tables. Prism 7 was used for graphical representationofthedata.
Results Denguecohort
Thecharacteristicsandclinicalparametersofthe119patients enrolled in the study are presented in Table T1 of the SupplementaryMaterial.Asexpected,mostpatientswithsevere dengue disease came to the hospital at later days of fever as compared to patientswith mild dengue (DI) and dengue with warningsigns(DW).Patientswithseveredenguehadsignificantly lowerplateletcountsatthetimeofenrolmentascomparedtoDI andDWcases.BasedontheIgMandIgGELISA,56(47%)patients wereclassifiedaseitherseronegativeorhavingprimaryinfections andtheother63patients(53%)ashavingsecondaryinfections.
IsolationofDENVRNAismoreefficientfromwholeblood
Patients were enrolled into the study based on a positive dengueNS1testusingapoint-of-carekit.TofurtherconfirmDENV positivity,determinetheinfectingDENVserotype,andtoestimate viremia, RNA was isolated from whole blood sampled from patients at the time of enrolment. Furthermore, to determine whetherRNAisolationfromtheplasmapreparationofthesame samplesisequallyefficient, RNAisolationwas performedusing plasmasamplesfrommostof thesepatients.TotalRNAisolated frombloodorplasmawasusedforthedetectionandquantitation ofviralRNAasdescribedpreviously(Agrawaletal.,2013).
Viral RNAwas detected in a significantly higher number of wholebloodsamplesascomparedtoplasmasamples.Overall,92%
of the RNA samples isolatedfrom wholeblood had detectable levelsofviralRNA,whereasthesamewas39%forplasmasamples (Supplementary Material, Figure S1A). This suggests that the efficiency of RNAisolation is higher for wholeblood samples, whichmaybeduetothepresenceofcellularRNA.Althoughcarrier
RNAwasaddedasperthemanufacturer’sinstructionsduringRNA isolation from plasma, the process was not as optimal for the detectionofviralRNAunderourconditions.
TheDENVgenomecopynumbersinwholeblooddidnotdiffer significantlybetweenthedifferentdiseaseseverities(Supplemen- taryMaterial,TableT1),asreportedpreviously(Singlaetal.,2016).
However,viralgenomecopynumbersweresignificantlyhigherin plasmasamplesascomparedtowholebloodsamples,asthetotal RNAobtainedfromplasmaislikelytobepredominantlyofviral originascomparedtowholebloodRNAwhichwillbedominated bycellularRNA(SupplementaryMaterial,FigureS1B).
Eightypercentofthepatientswereinfected withDENV-2as determinedbyserotypingPCR(SupplementaryMaterial,TableT2).
About11%ofthesampleswereeithernegativeorundeterminedby serotypingPCR,buttheywereconfirmedtobedengue-positiveby antibodyELISAorviralRNAdetectionor virusisolationexperi- ments.
DENVisolationfrompatientsamples
LikeallotherRNAviruses,denguevirusesevolverapidly,and changes in the genome regions encoding theT celland Bcell epitopeshaveanimpactonvaccinedevelopmentandtheimmune response. Therefore, focus was next placed on the isolation of DENVfrompatientsamplescollectedinthisstudy.Aflowdiagram oftheprocessthatwasfollowedisshowninFigure1.
Wholebloodwasusedfortheisolationofvirusesbythreeblind passagesinC6/36cells.Sixty-fourisolateswithmeasurableviral titers in P3 supernatants were obtained. Segregation of these isolatesbased onthedayoffever (DOF)of theoriginalsample showedthat72.5%ofthesamplescollectedbetweenDOF1and DOF3yieldedisolatesand44%ofthesamplescollectedbetween DOF 4 and DOF 6 yielded isolates, clearly demonstrating that samplescollectedbetweenDOF1andDOF6wouldbeidealfor virusisolation(SupplementaryMaterial,FigureS2A).Thenumber ofvirusisolatesobtainedandtheviraltitersatP3didnotdiffer
significantly betweensamplesobtainedfromDI,DW, orsevere denguepatients(SupplementaryMaterial,FigureS2B,S2C).
Forty-six of the isolates were confirmed to be of DENV-2 serotypebyRT-PCRand25oftheseisolateshadplaqueassaytiters highenoughforgrowthcurveassaysinC6/36cells(105–108pfu/
ml).The 3-daygrowth kineticsofP3 isolateswereassessed by infectingC6/36cellswith0.1MOI(multiplicityofinfection)ofP3 supernatants.At24h,48h,and72hpost-infection,supernatants werecollectedandviraltitersweremeasuredbyplaqueassay.Of the25isolates,itwaspossibletostudy22isolatesinC6/36cells.
Thegrowthratesoftheisolatesdiffereddrasticallyandtheviral titersonday3post-infectionfromtheclinicalisolatesrangedfrom 103pfu/mlto106pfu/ml(Figure2A).Thehighesttiters(>106pfu/
ml)wereassociatedwiththesamplesfromapatientwithsevere dengue who had died of the disease and from a DW case.
Nevertheless, viraltiters on day3 post-infectiondid not differ significantlybetweenvirusesisolatedfrompatientswithdifferent diseaseseverities(Figure2BandC).Ofthethreeisolatesthatfailed toamplify,onewasfromaDIpatientandtwowerefrompatients withseveredengue,andallthreesampleswereobtainedwithin 3days of fever. These in vitro data suggest that the isolation efficiency and growth kinetics of clinical isolates from DENV patients with diseaseof differentseverity are not significantly differentincellcultureconditions.
WholegenomesequencingofclinicalDENVisolates
Next,viralRNAwasisolatedfrom84plasmasamplesandthe amount of DENV RNAwas determined by quantitative RT-PCR.
Thirty-five samples with copy numbers exceeding 105 DENV genome equivalents per milliliter were processed for whole genome sequencing. A modified RNA sequencingprotocol was used to generate thewhole genome sequences (see Methods).
Using de novo assembly followed by reference recruitment (mapping),completegenomeswererecoveredwithgreaterthan 9 coverage across the genome from 21 RNA samples. These
Figure1.Workflowofthestudy.Flowchartofactivities,depictingvirusisolationandgrowthcurveanalysisfromwholeblood/plasmainC6/36cells,viralRNAisolationfrom plasma,andwholegenomesequencing.
Figure2.Growthkineticsofpatientisolates(passage3).(A)C6/36cellswereinfectedatamultiplicityofinfection(MOI)of0.1andculturesupernatantswerecollectedon days1–3.Viraltiterswereestimatedbyplaqueassays.(B)Comparisonofday3viraltitersfromtheaboveisolatesaccordingtothediseaseclassificationofthesamplesat origin.(C)Samplesindicatingindividualviraltitersonday3.ND,notdetected.Theasterisk(*)indicatesthoseisolatesforwhichwholegenomesequenceswereobtained.
included 18 DENV-2, two DENV-1, and one DENV-3 genomes (SupplementaryMaterial,TableT3).GenBankaccessionnumbers areprovidedinTableT4oftheSupplementaryMaterial.
Phylogeneticanalysis
MEGA X (Kumar et al., 2018) was used to reconstruct the evolutionary tree(s) for the 21 near-complete DENV genome sequences.TheDENV-3samplesmappedtogenotypeIIIandthe DENV-1samplemappedtogenotypeV(SupplementaryMaterial, Figure S3A, S3B). Since most of the whole genome sequences recoveredbelongedtotheDENV-2serotype,focuswasplacedon furtheranalysisofthesesamples.
The DENV-2 serotype has six genotypes that are largely spatiallyrelated,namelyAmerican,American-Asian,Cosmopoli- tan,Asian-I,Asian-II,andSylvatic(Wamanetal.,2016).Basedon theneighbor-joining(NJ)treeoftheenvelopeprotein(E),itwas observedthattheIndianDENVstrainsfrompre-2000wereofthe Americangenotype,whiletherecentstrains(post-2000)mostly seemed to be dominated by the Cosmopolitan genotype (Figure 3A). Even within the Cosmopolitan genotype, recent studieshavereportedtheprevalenceofthreelineageclustersin India(Afreenetal.,2016).ClusteringoftheDENV-2genomeswith twodistinctlineageswithintheCosmopolitangenotypewasalso observed. Lineage III sequences were largely prevalent in the secondhalf ofthe epidemic(2014–15) and lineageI sequences dominated the first half of the epidemic (2012–13) while co- circulatingatthesametime intheDelhiregion.Themaximum likelihoodphylogenetictreewasthenconstructedusingthewhole genome sequences of the DENV-2 from GenBank (Figure 3B).
DENV-2 genomes from the present study largely followed the phylogenetic analysis of the E protein sequences, suggesting insignificantrecombinationinthelineages.
SequencediversityfromglobalDENV-2strains
Next, the level of diversity between the present DENV-2 sequencesandtheDENV-2Cosmopolitangenotypewasevaluated.
AconsensusDENV-2Cosmopolitangenomewas firstgenerated usingthe115wholegenomesintheGenBankdatabase(Supple- mentary Material, Table T5). Synonymous single nucleotide polymorphisms (SNPs) in the polyproteincoding regionof the virus genome were high (1051 mutations) and uniformly distributedacrossthewholegenome,asexpectedforRNAviruses.
The non-synonymousmutations wererare but significant (138 mutations) and spread non-uniformly across the polyprotein (Supplementary Material, FigureS4; SupplementaryMaterial, TableT6).MostofthechangesintheEproteinwereintheBcell epitopes,andtwoofthechangesintheNS3regionwerepartofthe CD8Tcellepitope(SupplementaryMaterial,TableT6)(Vaughan etal., 2010).We nextlocated theamino acidchangesin theE proteinfromtheconsensusandthe16881DENV-2sequencethat hasbeenadaptedtogeneratetheDNAinfectious clone(Kinney etal.,1997).AlmostalloftheEproteinmutationsmappedtothe exposedfaceoftheEproteininitsdimericconformation,possibly suggesting immunological selection being the driving force (Figure 4A and 4B). For example, a large proportion of the mutationsspannedtheEDI/IIregion,whichisthebindingepitope for human monoclonal antibody (hMAb) 3F9, whereas the mutationH346YmappedtotheEDIIIbindingepitoperecognized by hMAbs 2D22 and 1L12 (Gallichotte et al., 2018). We also comparedtheUTRsofthegenomessincetheyarereportedtobe criticalforvirusreplication(Gebhardetal.,2011;Ngetal.,2017).
Thelargelyconserved50UTRamongtheCosmopolitangenotype carriedaC–UmutationthatliesintheRNApolymerase binding SLAloop(SupplementaryMaterial,FigureS5).The30UTRonthe
other hand displayed several mutations that fell into two categories:onesetthatwascommonamongalltheCosmopolitan genotypeandanothersetofmutationsthatsetthetwolineages apartwithinthegenotype(SupplementaryMaterial,FigureS6).
Discussion
PreviousreportshavecomparedDENVdetectionusingdifferent RNA isolation methods from serum, plasma, or whole blood (Anwaretal.,2009;DettogniandLouro,2011).Thisstudyconfirms theresultsofpreviousstudiesthathavereportedviraldetectionto be more efficient with whole blood as compared to plasma (Klungthong et al., 2007). Dengue serotyping using rapid diagnostictestshasalsodemonstratedtheconvenienceandutility ofusingwholebloodfortheisolationofviralRNA(Vongsouvath etal.,2016).Anotherstudythatusedcellularcomponentsofwhole blood,serum,clotspecimens,andplasmafrombloodbankdonors who were DENV RNA-positive also demonstrated that the detection of viral RNA by RT-PCR was more consistent with cellularcomponentsofblood(Añezetal.,2016).In thepresent study, it was foundthat viralRNA isolationwas more efficient usingwholebloodandinsamplescollectedwithin6daysoffever.
Blood viremia did not differ significantly between samples obtainedfrompatientswithdifferentdiseaseseverities;however, thestudyhasacaveatin thatmostofthepatientswithsevere dengue sought clinical care at later stages of infection when viremiaisonthedecline.Itispossiblethatthehighviralloadinthe earlystagesofinfectionleadstohigherinflammation,resultingin severedisease.Therefore,estimationofviremiaasafunctionof timestartingfromtheearliestpossibletimeuponinfectionuntil hospitalization/recoverywouldprovideaclearerpictureoftherole ofviralloadindiseaseprogression.
Abouthalfofthepatientsenrolledinthisstudyhadaprimary/
seronegative status based on IgG and IgM ELISA. Thirty-two percentoftheseveredenguecaseswerealsoprimaryinfections (nodengueIgGatthetimeenrolment),suggestingthatinasubset of patients, severe disease was a consequence independent of infection-enhancingantibodies.
VeryfewstudieshavecharacterizedclinicalisolatesfromIndia anditisnecessarytodeterminewhethertheantibodiesgenerated frompreviousnatural infectionsorvaccinetrialsarecapableof neutralizing circulating viruses in the country. In addition, sequencing information and corresponding virusneutralization datawillalsoprovidecluestotheemergenceofescapevariants.
Clinical isolates minimallypassagedin vitrohave beenusedto assessreplicationfitnessincellculturemodelsofinfectionandin immunocompetentandimmunocompromisedmousemodelsor rhesusmacaques(Borgesetal.,2018;Ferreiraetal.,2010;Sarathy etal.,2018;Tuiskunenetal.,2011a,b).Ithasbeendemonstrated that isolates frompatients withsevereor fatal dengue disease demonstrateenhancedreplicationandinduceinflammationata muchhigherlevel(Murgueetal.,1997;Silveiraetal.,2011).The present study revealed that the low passage clinical DENV-2 isolatesshoweddifferentgrowthkineticsincellculturebutthat thedifferenceswerenotassociatedwiththediseaseseverityinthe patients.Althoughtheseisolateswerepassagedonlythreetimesin cellculturetolimit adaptivemutations,thepossibility thatthe sequence of thevirusobtainedafterthethird passagewas not identicaltothatinfectingthepatientcannotberuledout.Further sequencingofP3isolateswillclarifythisissue.
Fullgenomesequencingdatafromrecentreportssuggestthat inmostpartsofSouthAmericaandLatinAmerica,thecirculating DENV-2strainsbelongtotheAmerican/Asiangenotypeandthat phylogenetically close viruses are circulating between these countries (Lizarazo et al., 2018; Stewart-Ibarra et al., 2018;
Williams et al.,2014).Similarly, thefull genomesequence ofa
Figure3. PhylogeneticanalysisofDENV-2clinicalisolates.(A)Theoptimaltreewiththesumofbranchlength=0.43786479isshown.Thetreeisdrawntoscale,withbranch lengthsinthesameunitsasthoseoftheevolutionarydistancesusedtoinferthephylogenetictree.Theevolutionarydistancesareintheunitsofthenumberofaminoacid substitutionspersite.Theanalysisinvolved121aminoacidsequencesoftheEprotein.Therewereatotalof473positionsinthefinaldataset.(B)Phylogenetictreeforwhole genomesofDENV-2usingthemaximumlikelihoodmethodbasedonthegeneraltimereversiblemodel.Thetreewiththehighestloglikelihood(47856.75)isshown.The treeisdrawntoscale,withbranchlengthsmeasuredinthenumberofsubstitutionspersite.Theanalysisinvolved53nucleotidesequences.Therewereatotalof10743 positionsinthefinaldataset.SamplesfromthisstudyareindicatedasinTableT3oftheSupplementaryMaterial.
DENV-2 isolate from Myanmar identified the Asian genotype closelyrelatedtotheisolatescirculatinginVietnam,Thailand,and Cambodia (Zeng et al., 2018). The Asian genotype of DENV-2 dominatedinChinapriorto2010andhasbeenreplacedbythe Cosmopolitangenotypethatissimilartotheisolatescirculatingin Singapore(Jiangetal.,2018;Zhaoetal.,2012).Allfourserotypes arecirculatinginIndia,althoughDENV-2hasdominatedinrecent yearsinthestudyregionandalsoinotherpartsofthecountry (Mishraetal.,2015;Mukherjeeetal.,2017;Shrivastavaetal.,2018;
Singlaet al., 2016).Phylogenetic analysisof DENV-2 sequences from Pakistan suggests that the Cosmopolitan genotype was introducedfromIndiaandSriLankaabout30yearsago,andasub- lineage within the Cosmopolitan genotype has emerged in PakistanandintheIndiansub-continent,causingmajorepidemics (Akrametal.,2015;Khanetal.,2013).Thisisfurtherconfirmedby DENV-2sequences (completegenome or thestructural region) fromIndiafortheperiod1960to2012,whichindicateashiftfrom American genotype to a unique South Asian clade within the Cosmopolitangenotype(Dashetal.,2013;Mishraetal.,2015).The envelopesequenceofDENV-2isolatesfromNepalhasshownco- circulationoftwoDENV-2genotypes—AsianIIandCosmopolitan IVaandIVb. AmongtheDENVserotypes,DENV-2hasbeenthe mostgeneticallydiversepopulation.Itisthemostfrequentcause ofdengueepidemicsworldwideandhasalsobeenassociatedwith severedenguecases(WeiandLi,2017).
In this study, complete genome sequences of 21 DENV-2 isolates from India isolated between 2012 and 2015 were obtained, which further confirmed a positive selection of Cosmopolitan genotype in India. Fifteen of the 21 sequences wereobtainedfromseronegativepatientsorthosewithprimary infections, and these were isolated in the early days of fever (medianDOF2.5).Thisisconsideredtobeauniquefeatureofthe presentstudyandwillserveasusefulinformation.Theanalysisof DENV-2 sequences from primary and secondary infections in Thailand showed that the sequences were homogeneous in secondary infections whilethe viral diversitywasmuch more heterogeneous in primary infections, suggesting that most primary infections could drive the generation of quasispecies leadingtotheevolutionandselectionofdominantisolatesduring subsequentexposure(Kurosuetal., 2014).Itis speculatedthat non-synonymousmutationsinthecapsid,NS2A,andNS5could
beactingasastrongdriverofpositiveselection.Inaddition,some of the mutations in the E protein mapped to B cell epitopes.
Further studies should be performedtoverify whetherany of theseisolatesareneutralizationescapemutants.Futurestudies shouldbedesignedwithafocusoncapturingthekineticsofviral replication by isolation of viral RNA and viruses from serial samplesfrombothprimaryandsecondaryinfections,inorderto determine the inter-host diversity and evolution of the viral genomewithdiseaseprogressionorrecovery.
Acknowledgements
WethankAmreshKumarSinghfortechnicalsupport.
Authorcontributions
MK,AN,AK,SJ,JS,MS,SM,andAPperformedexperimentsand analyzed data. AC, SKK, SK, and RL contributed reagents and analyzeddata.RR,CP,andGMconceivedthestudy,designedand performed the experiments, analyzed the data, and wrote the manuscript. All authors havereviewed the final version of the manuscript.
Funding
ThisworkwassupportedbyfundingtoGRM,SKK,RL,andAC underthe“VaccineGrandChallengeProgram”oftheDepartmentof Biotechnology,GovernmentofIndia(SanctionNo.BT/PR5132/MED/
15/85/2012)andbyWellcometrust-DBTIndiaAllianceIntermediate fellowshipstoGM(IA/S/14/1/501291)andRR(IA/I/13/1/500901).
MKreceivedafellowshipfromtheDepartmentofBiotechnology, India. This study was also supported by the Department of Biotechnology,GlueGrant,andNCBScorefundstoSK,aswellasa grant for dengue virus sequencing from Mr Narayana Murthy (Infosys). CP was supported bya Department of Biotechnology ResearchAssociateFellowship,RoyalSociety-SERBNewtonInterna- tionalFellowship,andtheIndiaAlliance(DBT-WellcomeTrust)Early CareerFellowship(IA/E/15/1/502336).Thefundershadnoroleinthe studydesign,collectionandinterpretation ofthedata,orinthe decisiontosubmittheworkforpublication.Thisarticleispartofa supplemententitled‘DengueFeverinIndia’whichissponsoredby Figure4.Non-synonymousmutations(yellow)mappedontheenvelopeprotein(ProteinDataBank(PDB)accessioncode1OAN)mutations,comparedto(A)theconsensus generatedfrom115CosmopolitanDENV-2sequencesfromallovertheworldand(B)strain16881(NC_001474),areshown.
theDepartmentofBiotechnology,GovernmentofIndia,andcollated bytheTranslationalHealthScienceandTechnologyInstitute.
Ethicsstatement
Thestudywasapprovedbytheinstitutionalethicscommit- teesof allthreeparticipatinginstitutes(Ethics/THSTI/2011/2.1 dated 16 November 2011; AIIMS: IEC/NP-338/2011 dated 17 November2011;ICGEB/IEC/2011/01dated12November2011).
Writteninformedconsentforthestudywasobtainedfromthe parents/guardians to collect blood samples at the time of admission.
Conflictofinterest
The authorsdeclarethatthey havenoconflict ofinterest to disclose.
AppendixA.Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttps://doi.org/10.1016/j.ijid.2018.12.003.
References
AfreenN,NaqviIH,BroorS,AhmedA,KazimSN,DohareR,etal.Evolutionary analysisofdengueserotype2virusesusingphylogeneticandBayesianmethods fromNewDelhi,India.PLoSNeglTropDis2016;10(3)e0004511.
AgrawalT,SchuP,MedigeshiGR.Adaptorproteincomplexes-1and3areinvolvedat distinctstagesofflaviviruslife-cycle.SciRep2013;3:1813.
AkramM,FatimaZ,PurdyMA,SueA,SaleemS,AminI,etal.Introductionand evolutionofdenguevirustype2inPakistan:aphylogeographicanalysis.VirolJ 2015;12:148.
AñezG,HeiseyDAR,ChanceyC,FaresRCG,EspinaLM,SouzaKPR,etal.Distribution ofdenguevirustypes1and4inbloodcomponentsfrominfectedblooddonors fromPuertoRico.PLoSNeglTropDis2016;10(2)e0004445.
AnwarA,WanG,ChuaK-B,AugustJT,TooH-P.Evaluationofpre-analyticalvariables inthequantificationofdenguevirusbyreal-timepolymerasechainreaction.J MolDiagn2009;11(6):537–42.
BalmasedaA,HammondSN,PerezL,TellezY,SaborioSI,MercadoJC,etal.Serotype- specificdifferencesinclinicalmanifestationsofdengue.AmJTropMedHyg 2006;74(3):449–56.
BhattS,GethingPW,BradyOJ,MessinaJP,FarlowAW,MoyesCL,etal.Theglobal distributionandburdenofdengue.Nature2013;496(7446):504–7.
BorgesMB,MarchevskyRS,MendesYS,MendesLG,DuarteAC,CruzM,etal.
CharacterizationofrecentandminimallypassagedBraziliandengueviruses inducingrobustinfectioninrhesusmacaques.PLoSOne2018;13(4)e0196311.
CapedingMR,TranNH,HadinegoroSR,IsmailHI,ChotpitayasunondhT,ChuaMN, etal.Clinicalefficacyandsafetyofanoveltetravalentdenguevaccineinhealthy childreninAsia:aphase3,randomised,observer-masked,placebo-controlled trial.Lancet2014;384(9951):1358–65.
ChaudhuryS,GromowskiGD,RipollDR,KhavrutskiiIV, DesaiV,WallqvistA.
Denguevirusantibodydatabase:systematicallylinkingserotype-specificity withepitopemappingindenguevirus.PLoSNeglTropDis2017;11(2)e0005395.
ClaphamHE,Rodriguez-BarraquerI,AzmanAS,AlthouseBM,SaljeH,GibbonsRV, etal.Denguevirus(DENV)neutralizing antibodykineticsinchildrenafter symptomaticprimaryandpostprimaryDENVinfection.JInfectDis2016;213 (9):1428–35.
DashPK,SharmaS,SoniM,AgarwalA,ParidaM,RaoPVL.Completegenome sequencingandevolutionaryanalysisofIndianisolatesofdenguevirustype2.
BiochemBiophysResCommun2013;436(3):478–85.
DettogniRS,Louro ID.DenguevirusRNApurification fromhumanplasma: a comparisonoftwotechniques.MolBiolRep2011;38(8):4979–83.
DiasM,PattabiramanC,SiddappaS,GowdaM,ShetA,SmithD,etal.Complete assemblyofadenguevirustype3genomefromarecentgenotypeIIIcladeby metagenomicsequencingofserum[version1;referees:2approved].Wellcome OpenRes2018;3(44).
Endy TP,Nisalak A, ChunsuttitwatS, Vaughn DW, Green S,Ennis FA, et al.
Relationshipofpreexistingdenguevirus(DV)neutralizingantibodylevelsto viremiaandseverityofdiseaseinaprospectivecohortstudyofDVinfectionin Thailand.JInfectDis2004;189(6):990–1000.
FerreiraGP,FigueiredoLB,CoelhoLF,SPA,Jr,CecilioAB,FerreiraPC,etal.Dengue virus3clinicalisolatesshowdifferentpatternsofvirulenceinexperimental miceinfection.MicrobesInfect2010;12(7):546–54.
FontaineA,LequimeS,Moltini-ConcloisI,JiolleD,Leparc-GoffartI,ReinerJr.RC, etal.Epidemiologicalsignificanceofdenguevirusgeneticvariationinmosquito infectiondynamics.PLoSPathog2018;14(7)e1007187.
GallichotteEN,BaricTJ,YountJr.BL,WidmanDG,DurbinA,WhiteheadS,etal.
Humandenguevirusserotype2neutralizingantibodiestargettwodistinct quaternaryepitopes.PLoSPathog2018;14(2)e1006934.
GebhardLG,FilomatoriCV,GamarnikAV.FunctionalRNAelementsinthedengue virusgenome.Viruses2011;3(9):1739–56.
GuyB,BriandO,LangJ,SavilleM,JacksonN.DevelopmentoftheSanofiPasteur tetravalentdenguevaccine:onemorestepforward.Vaccine2015;33(50):7100–11.
JiangL,MaD,YeC,LiL,LiX,YangJ,etal.Molecularcharacterizationofdenguevirus serotype2cosmospolitangenotypefrom2015dengueoutbreakinYunnan, China.FrontCellInfectMicrobiol2018;8:219.
JuraskaM,MagaretCA,ShaoJ,CarppLN,Fiore-GartlandAJ,BenkeserD,etal.Viral geneticdiversityandprotectiveefficacyofatetravalentdenguevaccineintwo phase3trials.ProcNatlAcadSciUSA2018;115(36):E8378–87.
KhanMA,EllisEM,TisseraHA,AlviMY,RahmanFF,MasudF,etal.Emergenceand diversificationofdengue2cosmopolitangenotypeinPakistan,2011.PLoSOne 2013;8(3)e56391.
KinneyRM,ButrapetS,ChangGJ,TsuchiyaKR,RoehrigJT,BhamarapravatiN,etal.
ConstructionofinfectiouscDNAclonesfordengue2virus:strain16681andits attenuatedvaccinederivative,strainPDK-53.Virology1997;230(2):300–8.
KlungthongC,GibbonsRV,ThaisomboonsukB,NisalakA,KalayanaroojS,Thirawuth V,etal.DenguevirusdetectionusingwholebloodforreversetranscriptasePCR andvirusisolation.JClinMicrobiol2007;45(8):2480–5.
KumarS,StecherG,LiM,KnyazC,TamuraK.MEGAX:molecularevolutionary geneticsanalysisacrosscomputingplatforms.MolBiolEvol2018;35(6):1547–9.
KurosuT,ChaichanaP,PhanthanawiboonS,KhamlertC,YamashitaA,AnA,etal.
Sequence variationofdenguetype 2virusisolated fromclinicalcases in Thailand.JpnJInfectDis2014;67(2):132–4.
LeitmeyerKC,VaughnDW,WattsDM,SalasR,VillalobosI,DeC,etal.Denguevirus structural differences that correlate with pathogenesis. J Virol 1999;73 (6):4738–47.
LizarazoE,CoutoN,Vincenti-GonzalezM,RaangsEC,JaenischT,FriedrichAW,etal.
Complete coding sequences of five dengue virus type 2 clinical isolates fromvenezuelaobtainedthroughshotgunmetagenomics.GenomeAnnounc 2018;6(25).
Mishra G, JainA, Prakash O, PrakashS,Kumar R, GargRK, etal. Molecular characterization of denguevirusescirculating during 2009–2012 in Uttar Pradesh,India.JMedVirol2015;87(1):68–75.
MukherjeeS,DuttaSK,SenguptaS,TripathiA.Evidenceofdengueandchikungunya virusco-infectionandcirculationofmultipledengueserotypesinarecent Indianoutbreak.EurJClinMicrobiolInfectDis2017;36(11):2273–9.
Murgue B, Cassar O, Guigon M, Chungue E. Dengue virus inhibits human hematopoieticprogenitorgrowthinvitro.JInfectDis1997;175(6):1497–501.
NgWC, Soto-Acosta R,Bradrick SS,Garcia-BlancoMA, OoiEE.The50 and30 untranslatedregionsoftheflaviviralgenome.Viruses2017;9(6).
NgweTunMM,ThantKZ,InoueS,KurosawaY,LwinYY,LinS,etal.Serological characterizationofdenguevirusinfectionsobservedamongdenguehemor- rhagicfever/dengueshocksyndromecasesinupperMyanmar.JMedVirol 2013;85(7):1258–66.
NunesPCG,deFilippisAMB,LimaM,FariaN,deBruycker-NogueiraF,SantosJB,etal.
30yearsofdenguefatalcasesinBrazil:alaboratorial-basedinvestigationof 1047cases.BMCInfectDis2018;18(1):346.
NurkS,BankevichA,AntipovD,GurevichAA,KorobeynikovA,LapidusA,etal.
Assemblingsingle-cellgenomesandmini-metagenomesfromchimericMDA products.JComputBiol2013;20(10):714–37.
OhAinleM,BalmasedaA,MacalaladAR,TellezY,ZodyMC,SaborioS,etal.Dynamics ofdenguediseaseseveritydeterminedbytheinterplaybetweenviralgenetics andserotype-specificimmunity.SciTranslMed2011;3(114)114ra28.
Sarathy VV, White M, Li L, Kaiser JA, Campbell GA, Milligan GN, et al.
Characterizationofamurinemodelofnon-lethal,symptomaticdenguevirus infection.SciRep2018;8(1)4900.
ShrivastavaS,TirakiD, DiwanA,LalwaniSK,ModakM,MishraAC,etal.Co- circulationofallthefourdenguevirusserotypesanddetectionofanovelclade ofDENV-4(genotypeI)virusinPune,Indiaduring2016season.PLoSOne 2018;13(2)e0192672.
SilveiraGF,MeyerF,DelfraroA,MosimannALP,ColuchiN,VasquezC,etal.Dengue virustype3isolatedfromafatalcasewithvisceralcomplicationsinduces enhancedproinflammatoryresponsesandapoptosisofhumandendriticcells.J Virol2011;85(11):5374.
SinglaM,KarM,SethiT,KabraSK,LodhaR,ChandeleA,etal.Immuneresponseto denguevirusinfectioninpediatricpatientsinNewDelhi,India—associationof viremia,inflammatorymediatorsandmonocyteswithdiseaseseverity.PLoS NeglTropDis2016;10(3)e0004497.
StanawayJD,ShepardDS,UndurragaEA,HalasaYA,CoffengLE,BradyOJ,etal.The globalburdenofdengue:ananalysisfromtheGlobalBurdenofDiseaseStudy 2013.LancetInfectDis2016;16(6):712–23.
Stewart-IbarraAM,RyanSJ,KennesonA,KingCA,AbbottM,Barbachano-Guerrero A,etal. Theburdenofdenguefeverandchikungunyainsoutherncoastal Ecuador:epidemiology,clinicalpresentation,andphylogeneticsfromthefirst twoyearsofaprospectivestudy.AmJTropMedHyg2018;98(5):1444–59.
Tuiskunen A,Monteil V, PlumetS, BoubisL, Wahlström M, Duong V,et al.
Phenotypic andgenotypiccharacterization ofdenguevirusisolatesdiffer- entiates denguefever and denguehemorrhagic fever from dengueshock syndrome.ArchVirol2011a;156(11):2023.
TuiskunenA,WahlströmM,BergströmJ,BuchyP,Leparc-GoffartI,LundkvistÅ.
PhenotypiccharacterizationofpatientdenguevirusisolatesinBALB/cmice
differentiatesdenguefeveranddenguehemorrhagicfeverfromdengueshock syndrome.VirolJ2011b;8(1):398.
VanBlarganLA,MukherjeeS,DowdKA,DurbinAP,WhiteheadSS,PiersonTC.The type-specificneutralizing antibody responseelicitedby a denguevaccine candidateisfocusedontwoaminoacidsoftheenvelopeprotein.PLoSPathog 2013;9(12)e1003761.
VaughanK,GreenbaumJ,BlytheM,PetersB,SetteA.Meta-analysisofallimmune epitopedataintheFlavivirusgenus:inventoryofcurrentimmuneepitopedata statusinthecontextofvirusimmunityandimmunopathology.ViralImmunol 2010;23(3):259–84.
VaughnDW,GreenS,KalayanaroojS,InnisBL,NimmannityaS,SuntayakornS,etal.
Dengueviremiatiter,antibodyresponsepattern,andvirusserotypecorrelate withdiseaseseverity.JInfectDis2000;181(1):2–9.
VongsouvathM,PhommasoneK,SengvilaipaseuthO, KosoltanapiwatN,Chan- tratitaN,BlacksellSD,etal.UsingrapiddiagnostictestsasasourceofviralRNA fordengueserotypingbyRT-PCR—anovelepidemiologicaltool.PLoSNeglTrop Dis2016;10(5)e0004704.
WamanVP,KolekarP,RamtirthkarMR,KaleMM,Kulkarni-KaleU.Analysisof genotypediversityandevolutionofdenguevirusserotype2usingcomplete genomes.PeerJ2016;4:e2326.
WattsDM,PorterKR,PutvatanaP,VasquezB,CalampaC,HayesCG,etal.Failureof secondary infection with American genotype dengue 2 to cause dengue haemorrhagicfever.Lancet1999;354(9188):1431–4.
WeiK,LiY.Globalevolutionaryhistoryandspatio-temporaldynamicsofdengue virustype2.SciRep2017;7:45505.
WichmannO,VanniceK,AsturiasEJ,deAlbuquerqueLunaEJ,LonginiI,LopezAL, etal.Live-attenuatedtetravalentdenguevaccines:theneedsandchallengesof post-licensureevaluationofvaccinesafetyandeffectiveness.Vaccine2017;35 (42):5535–42.
WilliamsM,MayerSV,JohnsonWL,ChenR,VolkovaE,VilcarromeroS,etal.Lineage IIofSoutheastAsian/AmericanDENV-2isassociatedwithaseveredengue outbreakinthePeruvianAmazon.AmJTropMedHyg2014;91(3):611–20.
ZengZ,ShiJ,GuoX,MoL,HuN,SunJ,etal.Full-lengthgenomeandmolecular characterizationofdenguevirusserotype2isolatedfromanimportedpatient fromMyanmar.VirolJ2018;15(1):131.
ZhaoH,DengY-Q,HongW-X,YuX-D,JiangT,YuM,etal.Completegenome sequence of dengue virus serotype 2 cosmopolitan genotype strain in Guangdong,China.JVirol2012;86(24):13808.