DNA extraction protocol and a novel multiplex real-time PCR

Copyright©2011,AmericanSocietyforMicrobiology.AllRightsReserved.

Vol.49,No.7

RapidandSimultaneousDetectionofMycobacteriumtuberculosisComplexandBeijing/WGenotypeinSputumbyanOptimizedDNAExtraction

ProtocolandaNovelMultiplexReal-TimePCRᰔEricT.Y.Leung,1*L.Zheng,1RityY.K.Wong,2EdwardW.C.Chan,1T.K.Au,1RaphaelC.Y.Chan,1GraceLui,2NelsonLee,2andMargaretIp1DepartmentofMicrobiology1andDepartmentofMedicineandTherapeutics,2TheChineseUniversityof,,China

Received19January2011/Returnedformodification22February2011/Accepted2May2011

RapiddiagnosisandgenotypingofMycobacteriumtuberculosisbymolecularmethodsareoftenlimitedbytheamountandpurityofDNAextractedfrombodyfluids.Inthisstudy,weevaluated12DNAextractionmethodsanddevelopedahighlysensitiveprotocolformycobacterialDNAextractiondirectlyfromsputausingsurface-coatedmagneticparticles.Wehavealsodevelopedanovelmultiplexreal-timePCRforsimultaneousidenti-ficationofM.tuberculosiscomplexandtheBeijing/Wgenotype(ahypervirulentsublineageofM.tuberculosis)byusingmultiplefluorogenicprobestargetingboththeM.tuberculosisIS6110andtheRv0927c-pstS3intergenicregion.Withreferencestrainsandclinicalisolates,ourreal-timePCRaccuratelyidentified20non-Beijing/Wand20Beijing/WM.tuberculosisstrainsfrom17differentspeciesofnontuberculosisMycobacterium(NTM).FurtherassessmentofourDNAextractionprotocolandreal-timePCRwith335nonduplicatesputumspec-imenscorrectlyidentifiedall74M.tuberculosisculture-positivespecimens.Inaddition,15culture-negativespecimensfrompatientswithconfirmedtuberculosiswerealsoidentified.Nocross-reactivitywasdetectedwithNTMspecimens(n؍31).Thedetectionlimitoftheassayis10M.tuberculosisbacilli,asdeterminedbyendpointdilutionanalysis.Inconclusion,anoptimizedDNAextractionprotocolcoupledwithanovelmulti-probemultiplexreal-timePCRforthedirectdetectionofM.tuberculosis,includingBeijing/WM.tuberculosis,wasfoundtoconferhighsensitivityandspecificity.Thecombinedprocedurehasthepotentialtocompensateforthedrawbacksofconventionalmycobacterialcultureinroutineclinicallaboratorysetting,suchasthelengthyincubationperiodandthelimitationtoviableorganisms.

Fightingtuberculosis(TB)remainsaglobalpublichealthchallenge.TheWorldHealthOrganizationhasestimatedthatover2billionpeopleareinfectedwithMycobacteriumtubercu-losis,thecausativeagentofTB(21,22),andthat10%ofthemwilleventuallydevelopactivedisease.Despitethefactthatnumerouscontrolprogramshavebeenattemptedworldwide,therewerestill9.4millionnewcasesand1.7millionreporteddeathsintheyear2009(22).ThespreadofthediseaseisalsocomplicatedbytheemergenceofhypervirulentstrainsfromtheBeijing/Wlineage(9,19).Incontrasttonon-Beijingstrains,recentclinicalandepidemiologicalstudieshaveshownthatBeijing/Wstrainshaveevolveduniqueproperties,includ-ingtheabilitiestoacquiredrugresistancemorefrequently(9,13,15),todisseminatemoreefficiently(3,6),andtoevadehostimmunedefenseandtheprotectiveeffectofBCGvac-cination(12,14,15,17,23).Insomereports,Beijing/Wstrainshavebeencloselyassociatedwithtreatmentrelapse(5,19).

CentraltoeffectivecontrolofthediseaseisthedevelopmentofrapidandreliablemethodsfordiagnosisofM.tuberculosisinfection.Conventionalidentificationrequiresacultureperiodfrom6to8weeks,whileBeijinggenotypingbytraditional

IS6110fingerprintingmethodtakesanother3daysaftersuc-cessfulculturehasbeenobtained.Thelatteralsorequirestechnicalskillsandequipmentthatareoftenunavailableinroutineclinicalmicrobiologylaboratories.Despitetheavail-abilityofanumberofcommercialdiagnosticsystemsforrapiddetectionoftheM.tuberculosiscomplex,theuseofthesesystemsinroutineclinicallaboratoriesislimitedbytheirpro-hibitorycostsand/ortheirrequirementofspecificequipments.Moreover,clinicaldetectionandreportingofBeijing/Wclonesareworthwhilebutnotcommonduetothelackofconvenientandreliableprocedures.Wethereforeseetheadvantagesofdevelopinganin-housemultiprobemultiplexreal-timePCRprotocolforsimultaneousTBdiagnosisandthedetectionofBeijing/Wgenotype.

CurrentunderstandingoftheunderlyinggeneticsofM.tu-berculosishasfacilitateddevelopmentofsuchprotocol.RecentstudiesbyJiangetal.comparedtheDNAsequencesofBei-jing/Wandnon-Beijing/WM.tuberculosisisolatesandfoundacharacteristicpointmutation,G3A(bpϪ127)intheRv0927c-pstS3intergenicregion,thatisspecifictoBeijing/Wisolates(10).Thismutationdisruptsthepromoterregionanddown-regulatestheexpressionofRv0927c,whichencodesaputativeshort-chaindehydrogenase.SincethemutationisamolecularmarkerforBeijing/W,itcouldbeincorporatedintoamultiplexreal-timePCRforsimultaneousdetection.

ToallowhighsensitivityofdetectionofM.tuberculosisnu-cleicacidsdirectlyfromsputumspecimens,thechoiceofmethodforbacterialnucleicacidextractiontakesthecenterstage.Twostepsintheextractionprocedure—celllysisand

2509

*Correspondingauthor.Mailingaddress:DepartmentofMicrobi-ology,TheChineseUniversityof,PrinceofWalesHospi-tal,30-32NganShingSt.,Shatin,NewTerritories,SAR,China.Phone:(852)26322573.Fax:(852)273227.E-mail:ericleung@cuhk.edu.hk.ᰔPublishedaheadofprinton18May2011.

2510LEUNGETAL.DNApurification—aredeterminantsofdetectionsensitivity.Itisparticularlychallengingforthesetwostepstobeaccom-plishedeffectivelyinthesituationofM.tuberculosisinsputum,consideringthethickdurablebacterialcellwall,theunevendistributionoforganismsinthefluidduetothehightendencyofclumping,theabundanceofPCRinhibitorsinsputum,andtheminuteamountofM.tuberculosisDNAthatisoftenquenchedbyarichhumanDNAbackground.Previousstudieshavemadevaluableattemptstoovercometheseobstacles.UsingM.tuberculosis-spikedpleuralfluids,Santosetal.ana-lyzedtheyieldandthepresenceofPCRinhibitorsinDNAextractedbydifferentmethods(16),includingphenol-chloro-formbasedmethods,freeze-thaw-boiling,detergent-basedprocedures,Chelexresin,aQIAmpDNAMinikit,RocheCo-basTNAI-AMPLI-Prep,andfinallythePromegaDNAIQCaseworksamplekitwithMaxwell16robot.Amongthesecommercialandin-housemethods,itwasfoundthat,surpris-ingly,onlythreemethods—theQIAmpDNAMinikit,Chelex100,andCobasTNAI-AMPLI-Prep—couldefficientlyremovePCRinhibitorsandyieldPCR-detectableamountofM.tuber-culosisDNAatlow-levelspiking(Ͻ100bacilli).ItisnotedthatCobasTNAI-AMPLI-Prepisintendedtobeusedincombina-tionwithspecificmachinewhichmaynotbeaccessiblebysomelaboratories.AnotherreportbyAldousetal.(1)comparedthesensitivitiesofDNAextractionmethodsforM.tuberculosis-spikedsputumandfoundthatIDIlysistubes(InfectioDiag-nostics,Inc.),whichfunctionbyactualmechanicaldisruptionwithglassbeads,releasedthegreatestamountofmycobacte-rialDNA.Detergent-basedmethods,however,couldnotcom-pletelyremovePCRinhibitors.Althoughthesefindingsweregeneratedbyspikeanalysisonly,theyprovideimportanthintsforthesearchofthepotentiallymostsensitiveextractionpro-tocoltailoredforroutineclinicalsputumspecimens.

Amongthelimitedoptionsofreagentsforbodyfluidspec-imens,acouplemeritconsideration.DynabeadsMyOneSIL-ANEmagneticbeads(Invitrogen,Carlsbad,CA)aresmallmagneticparticles(1␮mindiameter)withasilica-likesurfacetocapturenucleicacids.Theproductisreportedtohaveahighaffinityfornucleicacidsandiscompatiblewithbiologicalsam-ples.Thebeadsareavailablealoneaswellasinkitsthatprovideextractionbuffers.Itcouldbeusedmanuallyorinconjunctionwithautomatedplatformsforlargethroughput.TheQIAmpMinElutevirusspinkit(Qiagen)utilizescarrierRNAtoenhancebindingofnucleicacidstosilicamembrane,whichmaybeparticularlyusefuliftherearefewtargetmole-culesinthesample.Itisalsooptimizedforpurificationfromhumanbodyfluids.Neithermethodhasbeenevaluatedforextractionfromsputum.Thepresentstudyaimsto(i)developanextractionprotocolwithhighsensitivityforM.tuberculosisdetectiondirectlyfromsputumspecimens;(ii)developamul-tiplexreal-timePCRforTBdiagnosisandsimultaneousde-tectionoftheBeijing/Wgenotype;and(iii)evaluatethede-velopedassay,includingbothextractionandreal-timePCRprotocols,withroutineclinicalspecimens.

MATERIALSANDMETHODS

Bacterialstrains.M.tuberculosisH37Rv(ATCC27294)wasusedforthespikecomparisonofDNAextractionmethods.FortheevaluationofBeijing/Wdetec-tionbyreal-timePCR,40clinicalisolatesofM.tuberculosis(20Beijing/Wand20non-Beijing/W)previouslycollectedfromPrinceofWalesHospital,,

J.CLIN.MICROBIOL.

TABLE1.ATCCreferencestrainsusedinthisstudy

Mycobacteriumstrain

ATCCno.

MTCaM.tuberculosisH37Rv................................................................27294M.bovis........................................................................................35720NTM

M.asiaticum.................................................................................25276M.aurum......................................................................................23366M.avium......................................................................................35718M.chelonae..................................................................................14472M.chitae.......................................................................................19627M.diernhoferi...............................................................................19340M.flavescens................................................................................14474M.fortuitum.................................................................................6841M.gastri........................................................................................157M.gordonae.................................................................................14470M.intracellulare...........................................................................13950M.kansasii...................................................................................12478M.marinum.................................................................................927M.neoaurum................................................................................25795M.scrofulaceum...........................................................................19981M.smegmatis...............................................................................19420M.szulgai......................................................................................35799

aMTC,M.tuberculosiscomplex.

duringtheperiodfrom2004to2009wereused.Thegenotypeswereconfirmedbythedeletion-targetedmultiplexPCR(DTM-PCR)methodasdescribedbyChenetal.(7).Inaddition,M.bovisand17referencestrainsofnontuberculosismycobacteria(NTM)wereusedforprimerspecificitytesting(Table1).

Specimenscollectionandprocessing.TheuseofclinicalmaterialsinthepresentstudywasapprovedbytheJointCUHK-NTECClinicalResearchEthicsCommittee(CRE-2008.500).Atotalof352nonduplicatesputumspecimenswerecollectedduringtheperiodJanuary2009toApril2010.Amongthe352specimens,17wereusedforthecomparisonofDNAextractionmethods,and335wereusedforevaluationofthefinaldevelopedassay,whichconsistedofbothoptimizedextractionandreal-timePCRprotocol.Thesespecimenswerese-lectedfromover3,000clinicalrespiratoryspecimensbasedontwocriteria:(i)theywereacid-fastbacillus(AFB)smearpositive,and(ii)theywereobtainedfrompatientswithastrongclinicalpresentationofTBorwithseverepulmonaryinfection.Thetwocriteriarepresentsituationsinwhichrapidmoleculardiag-nosisofTBismostneeded.Sputumspecimenswereinitiallydecontaminatedbysodiumhydroxidedigestionandsubjectedtoconventionalmycobacterialcultureandreal-timePCRinparallel.Forconventionalmycobacterialculture,250␮lofeachdecontaminatedsputumwasinoculatedinduplicateontotwoLowenstein-Jensenmediumslants.Theslantswereincubatedat37°Candmonitoredweeklyforeightconsecutiveweeks.Positivemycobacterialcultureswerespeciesdiffer-entiatedbypreviouslydescribed16Sreal-timePCR(11),hsp65PCR-RFLP(20)and,ifthesetestswereinconclusive,16SDNAsequencing(8)toidentifyM.tuberculosisandNTMspecies.

Designofprimersandprobes.Gene-specificprimersandprobeswerede-signedbyusingPrimerExpresssoftwareversion3.0(AppliedBiosystems,FosterCity,CA)(Table2).PrimersandTaqManMGBprobestargetinginsertionsequenceIS6110andRv0927c-pstS3intergenicregionwerebasedontheM.tuberculosisH37Rvgenomesequence(GenBankaccessionnumbersBX842579andBX842575).ThespecificityoftheprimerswasfirstverifiedbyusingtheNCBIBLASTalgorithm,followedbyreal-timePCRspecificitytestingwithDNAextractedfromthereferencestrainslistedinTable1.

Evaluationof12DNAextractionprotocolswithspikedspecimensandclinicalspecimens.Forthespikedspecimens,M.tuberculosisH37Rvstandardstrainwasgrownat37°Cwithorbitalshaking(250rpm)inMiddlebrook7H9brothsup-plementedwitholeicacid-albumin-dextrose-citricacid(OADC)to1McFarlandstandard.Then,2-mlportionsof1:10dilutedculturewereaddedto50mlofpooledculture-negativesputumspecimens.Aliquots(2ml)ofspikedsputumwerecentrifugedat5,000ϫgfor5min.Eachpelletwasresuspendedwith1mlofTEbuffer(pH8.0)forsubsequentDNAextraction.Fortheclinicalspecimens,17M.tuberculosisculture-positivesputumspecimens—9AFBsmear-positiveand8AFBsmear-negativespecimens—wereused.Thesespecimensrepresentedarangeofdifferentbacterialloadsinaspectrumofdifferentsputumviscosities.

VOL.49,2011DETECTIONOFM.TUBERCULOSISCOMPLEXINSPUTUM

TABLE2.Primersandprobesusedforreal-timePCR

2511

NameSequence(5Ј–3Ј)PositionaGenetarget(GenBankaccessionno.)Ampliconsize(bp)

MTB-FMTB-RMTB-PBJW-FBJW-RBJW-PHB-FHB-RHB-P

aGCCGGATCAGCGATCGT

GCAAAGTGTGGCTAACCCTGAAFAM-TTCGACGGTGCATCTG-MGBATGCACGGCATACGGACATGGTTGACCCCTGATGATGGACNED-TGAGATCCGCGGTCG-MGBTTCTGACACAACTGTGTTCACTAGCCAACTTCATCCACGTTCACC

VIC-CTCCTGAGGAGAAGTC-MGB–59106005–60265968–5983939–957993–1013966–9809–33104–123–79

IS6110(BX842579)

Rv0927c-pstS3intergenicregion(BX842575)Human␤-globin(NM_000518)

13375115

Thatis,thebasepairpositionsofthegenesequencesinthecorrespondingGenBankaccession.

Aliquots(2ml)ofsputumwerecentrifugedandresuspendedthesamewayasthespikedsputumaboveforsubsequentDNAextraction.

DNAextraction.ThreedifferenttypesofbeadsusedformechanicaldisruptioninthebacterialcelllysisstepandfourdifferentdownstreamDNApurificationmethodswereevaluatedforatotalof12protocols.Thethreedifferentbeadtypesincluded0.1-mmzirconiabeads(BioSpecProducts,Inc.,Bartlesville,OK),0.2-mmglassbeads(Sigma-Aldrich,St.Louis,MO),and1-mmglassbeads(SigmaAldrich,St.Louis,MO).Aportion(1ml)oftheTE-resuspendedpelletwasaddedtoa2-mlscrew-capmicrocentrifugetubecontainingϳ200␮lofoneofthethreetypesofbeads.ThetubeswerevortexedinaDisruptorGenie(ScientificIndustries,Inc.,Bohemia,NY)for5minandincubatedat95°Cfor10minforcompleteinactivationofthebacteria.Afterabriefcentrifugation,100-␮laliquotsofthelysatecontainingnodebriswereusedforeachDNApurificationmethod.

FourdifferentDNApurificationmethodswereevaluated.MethodCutilizedChelex100resins(Bio-RadLaboratories,Hercules,CA).Thelysatewasaddedtoanequalvolumeof10%Chelex100resins,followedbyincubationatroomtemperaturefor5minwithoccasionalmixing.Themixturewascentrifugedat1,000ϫgfor2mintocollectthesupernatant.ThesupernatantsweredirectlyusedasDNAtemplatesforreal-timePCR.MethodDutilizedaQIAmpDNAminikit(Qiagen,Hilden,Germany).Thelysatewasprocessedaccordingtothebodyfluidspinprotocolsuppliedbythemanufacturer.ThepurifiedDNAwaselutedin30␮lofbufferAE.MethodVutilizedaQIAmpMinElutevirusspinkit(Qiagen).Thelysatewasprocessedaccordingtothemanufacturer’srecom-mendedprotocol(2010-Aprversion).ThepurifiedDNAwaselutedin30␮lofBufferAVE.MethodHM:in-housebufferswithDynabeadsMyOneSILANEmagneticbeads(Invitrogen).Theformulationofbuffersisbasedonthebuffersystempreviouslydescribedwithslightmodification(4).Briefly,bindingbufferwaspreparedbydissolving120gofguanidiumthiocyanatein100mlof0.1MTris-HCl(pH6.4).Subsequently,22mlof0.2MEDTA(pH8.0)and2.6gofTritonX-100wereaddedtothesolution.Awashbufferwaspreparedbycom-bining55mlofethanoland45mlofsolutioncontaining3Mguanidiumthio-cyanate,10mMTris-HCl,and10mMNaCl(pH8.0).Toeach100-␮lportionoflysate,200␮lofbindingbufferwasadded,andthemixturewaspulsevortexed.Then,50␮lofDynabeadsand200␮lofisopropanolwereadded,respectively,withbriefvortexinginbetween.Thetubeswereincubatedatroomtemperaturefor3min.Subsequently,thebeadswereimmobilizedbyamagnet,andthesupernatantwasremoved.Thebeadswerewashedtwicewith800␮lofwashbufferandtwicewith800␮lof70%ethanol.Theethanolwasremovedbyairdryingthebeadsatroomtemperaturefor10min.DNAwaselutedwith30␮lofTEbuffer(pH9.0).Theextractionproceduresforspikedspecimenswerere-peatedninetimesforeachmethodforstatisticalanalysis.Forclinicalspecimens,eachofthe17specimenswasextractedbyall12evaluationmethods.

Cloningofstandardsforreal-timePCR.Tomonitortheefficiencyofeachrunofreal-timePCR,dilutionsofstandardwithknownquantitieswereincluded.Forcloning,PCRwasperformedina50-␮lreactionbyusinganAmpliTaqGoldPCRreagentkit(AppliedBiosystems),theprimersMTB-FandMTB-R,andH37RvgenomicDNA.Thereactionmixturewassubjectedtoaninitialpoly-meraseactivationof95°Cfor10min,followedby30thermalcyclesof94°Cfor30s,60°Cfor30s,and72°Cfor30s,withafinalextensionstepat72°Cfor7mininanAppliedBiosystems9700thermalcycler.Theampliconappearedasasinglebandafterelectrophoresis.TheampliconwasligatedtopCR2.1-TOPOvectorbyusingaTOPOTAcloningkit(Invitrogen),andtherecombinantvectorwastransformedintoTop10Escherichiacolicompetentcellsaccordingtotheman-ufacturer’srecommendations.TransformantswereisolatedandsubculturedforplasmidpreparationbyusingaQIAprepspinminiprepkit(Qiagen).TheDNA

sequenceoftheconstructwasconfirmedbyautomaticDNAsequencing.Quan-tificationoftheplasmidwasperformedbymeasuringabsorbanceat260nmwithaNanodropUVspectrophotometer(Invitrogen).

Multiprobemultiplexreal-timePCR.Amplificationwasdoneintriplicatesin96-wellopticalreactionplateswithanAppliedBiosystems7700real-timePCRinstrument.Amplificationofhuman␤-globingenebytheprimersHB-FandHB-RactsasinternalamplificationcontroltomonitorPCRinhibition.Theconcentrationsoftheprimersandprobeswereoptimizedoneatatimebytestingarangeofconcentrationsforonecomponentwhilefixingthatoftheothers.Theconcentrationsofthehuman␤-globinprimersandprobewereoptimizedtogeneratedetectablesignalwhentheotheramplificationisnegative,sothattheassociatedinterferenceandresourcecompetitionisminimized.ForcomparisonoftheDNAextractionmethods,eachwellcontained25␮lofreactionvolume,including12.5␮lofTaqManUniversalMasterMix,500nMMTB-F,300nMMTB-R,50nMMTB-P,and2.5␮lofDNAextracts.Forevaluationofthefinaldevelopedassaywith335clinicalspecimens,the25-␮lreactionvolumealsocontainsanadditionof600nMBJW-F,900nMBJW-R,80nMBJW-P,50nMHB-F,30nMHB-R,and50nMHB-P.Theinstrumentwasprogrammedto95°Cfor10min,followedby40cyclesof95°Cfor15sand60°Cfor1min.Quantifiedplasmidstandardsin10-folddilutionsequivalenttoarangeof20to2,000,000IS6110copieswereincludedineachrun.Cyclethreshold(CT)valuesofthedilutionsofstandardswereplottedagainstthenumberofcopiesoftemplatetogeneratethestandardcurves.Thevaluesofslopes,yintercepts,andthecorre-lationcoefficients(r2)ofthelinearregressionlineswereusedasindicatorsofefficiency.Amplificationthatgeneratedstandardcurveswithslopesbetween3.3and3.6,yinterceptsbetween39and40,andr2valuesϾ0.995wasconsideredvalid.

Statisticalanalysis.TheefficiencyofeachDNAextractionmethodwaseval-uatedbycomparingtheCTvaluesgeneratedfromreal-timePCR,withtheuseofthedataanalysispackageincludedinMicrosoftExcelsoftware(MicrosoftCorp.,Redmond,WA).Statisticalsignificancefordifferentgroupsofspikedspecimensextractedbydifferentmethodswasanalyzedbysingle-varianceanalysisofvari-ance(ANOVA).Detailedstatisticalsignificancefindingsforeachmethodagainsteachofanytheothermethodsforspikedspecimensandclinicalspeci-mensweretabulatedbyusingthetwo-tailedStudentttestandpairedttest,respectively.Pvaluessmallerthanorequalto0.05wereconsideredsignificant.

RESULTS

Primerspecificities.Ourreal-timePCRcorrectlyidenti-fiedall40M.tuberculosisisolatesandM.bovis(membersoftheM.tuberculosiscomplex).Nocross-reactionwiththeNTMstrainswasobserved.The20isolateswithaBeijing/Wgenotypewerealsoaccuratelydifferentiatedfrom20non-Beijing/Wstrains,asverifiedbyDTM-PCRresults,provid-ing100%specificityandsensitivity.

ComparisonofDNAextractionmethods.TheefficienciesofDNAextractionprotocolswerecomparedbytheirCTvaluesgeneratedfromreal-timePCRanalysisofbothspikedspeci-mensandclinicalspecimens(Table3).ThePvaluefromANOVAanalysisofthedifferentprotocolsofextractiononspikedspecimenswas2.57ϫ10Ϫ17(PՅ0.05),whichindicates

2512LEUNGETAL.J.CLIN.MICROBIOL.

TABLE3.MeanCTvaluesof12differentextractionmethods

CTbMethodaSpikedspecimens

(nϭ9)Mean

SD

Clinicalspecimens

(nϭ17)Mean

SD

L-CL-DL-VL-HMS-CS-DS-VS-HMZ-CZ-DZ-VZ-HM

a30.4328.1528.9528.9627.7926.0226.3125.3027.0525.9926.0925.311.120.980.901.111.191.551.261.411.411.181.331.3627.2426.5125.4727.7725.5125.1624.6024.6725.1424.2223.7923.705.555.845.396.035.155.425.455.465.385.785.275.57

L,large(1-mm)glassbeads;S,small(0.2-mm)glassbeads;Z,zirconiabeads(0.1mm);C,Chelexpurification;D,DNAminikit(Qiagen);V,QIAmpMin-Elutevirusspinkit;HM,in-housebufferwithInvitrogenDynabeadsMyOneSILANEmagneticbeads.bComparisonsofCTvaluesconsideredstatisticallysignificant(PՅ0.05)areindicatedinboldface.

existenceofsignificantdifferenceamongtheefficienciesofdifferentextractionprotocols.Ourresultsshowedthatme-chanicalcelldisruptionwith0.2-mmglassbeads(S)wasmoreefficientthanwith1-mmglassbeads(L).ContrastingtheCTvaluesofprotocolsS-CversusL-C,S-DversusL-D,S-VversusL-V,andS-HMversusL-HMrevealedcycledifferences(⌬CTs)from2.1to3.7withstatisticalsignificance(PϽ0.005)(Table4),suggesting4.5-to12.5-folddifferencesintheDNAyields.Zirconiabeads(Z)havesimilarefficienciesasS(Z-CversusS-C,Z-DversusS-D,Z-VversusS-V,andZ-HMversusS-HM;⌬CTsՅ0.74;PՆ0.24).ComparisonofthedifferentDNApurificationmethodsusingthesametypeofbeadsforcelllysis(i.e.,L-C/L-D/L-V/L-HM,S-C/S-D/S-V/S-HM,andZ-C/Z-D/Z-V/Z-HM)showedthatChelexresins(C)hadthepoorestefficiencyamongthefour(0.96Յ⌬CTsՅ2.49,1.9-to5.6-folddifference),whiletheotherthreepurificationmethodshadsimilarefficiencies(⌬CTsՅ0.81,PՆ0.09).ThethreemostsensitiveprotocolswithspikedspecimensareS-HM,

Z-HM,andZ-DinincreasingorderofCTvalues(decreasingsensitivity).

Comparisonusingclinicalspecimens(including9AFBsmearpositiveand8AFBsmearnegativesputumspecimens)alsosuggestedthatZandSweremoreefficientthanLindisruptingM.tuberculosiscells,whichisinagreementwiththeresultfromspikeanalysis.Zirconiabeadsincombinationwithmagneticbeads(Z-HM)wasfoundtobemoreefficientthanwith0.2-mmglassbeads(⌬CTϭ0.97,Pϭ0.01).ThethreemostsensitiveprotocolswithclinicalspecimensareZ-HM,Z-V,andZ-DinincreasingorderofCTvalues.

Tosummarizetheresultsfrombothspikedandclinicalspec-imens,celllysisbyzirconiabeadsincombinationwithDNApurificationbymagneticbeadsorspincolumnsconsistentlyproducedthelowestCTvalues(highestyieldofDNA)incon-trasttotheotherprotocols.CombinationsinvolvinglargeglassbeadsorChelexresinpurificationyieldedsignificantlylessM.tuberculosisDNA.Takingintoconsiderationthesensitivity,technicalsimplicityofprotocol,andpotentialforautomation,thecombinationofzirconiabeadsandmagneticbeads(Z-HM)wasselected.

Evaluationofthedevelopedextractionandreal-timePCRprotocol.Atotalof335sputumspecimenswereextractedbytheZ-HMprotocol,including74M.tuberculosisculture-posi-tive,31NTMculture-positive,and230mycobacterialculture-negativespecimens(Table5).Ofthe335specimens,(26.6%)werepositiveforM.tuberculosisbyreal-timePCR.All74M.tuberculosisculture-positivespecimenswerepositivelyidentifiedbyreal-timePCR.Fifteenculture-negativespeci-mens,includingfivethatwereAFBsmearpositiveandtenthatwereAFBsmearnegative,werealsodeterminedtobepositivebyreal-timePCR.Ofthese15specimens,12wereobtainedfrompatientsdiagnosedwithTB(asdeterminedeitherbyhistology/radiologyorbyconventionalmycobacterialcultureofotherrespiratoryspecimensfromthesameindividual)andwereundergoingTBchemotherapies(Table6).Theremainingthreespecimensoriginatedfrompatientswithrecordsofre-centlytreatedTB.OneofthethreepatientsiscurrentlyontreatmentforTBrelapse.NoneoftheNTMculture-positivespecimenswaspositiveforM.tuberculosisbyreal-timePCR(Table5).Takingconventionalmycobacterialcultureasthe

TABLE4.Pvaluesforspikedandclinicalspecimensderivedbystatisticalsignificanceanalysisbetweenmethods

Method

aP(spiked/clinical)bL-D

L-V

L-HM

S-C

S-D

S-V

S-HM

Z-C

Z-D

Z-V

Z-HM

L-CL-DL-VL-HMS-CS-DS-VS-HMZ-CZ-DZ-V

ab0.00/0.00

0.01/0.000.09/0.00

0.00/0.150.16/0.000.58/0.00

0.00/0.000.50/0.030.03/0.920.05/0.00

0.00/0.000.00/0.000.00/0.360.00/0.000.02/0.23

0.00/0.000.01/0.000.00/0.050.00/0.000.03/0.080.65/0.15

0.00/0.000.00/0.000.00/0.010.00/0.000.00/0.010.32/0.040.14/0.

0.00/0.000.07/0.000.00/0.360.00/0.000.24/0.130.16/0.960.31/0.330.02/0.07

0.00/0.000.00/0.000.00/0.020.00/0.000.01/0.010.96/0.070.57/0.0.28/0.240.10/0.00

0.00/0.000.00/0.000.00/0.000.00/0.000.01/0.000.92/0.000.70/0.130.24/0.010.16/0.000.86/0.21

0.00/0.000.00/0.000.00/0.000.00/0.000.00/0.000.32/0.010.14/0.200.99/0.010.02/0.000.27/0.060.23/0.82

ThemethodsareasdefinedinTable3,footnotea.

Statisticallysignificantresults(PՅ0.05)areindicatedinboldface.

VOL.49,2011DETECTIONOFM.TUBERCULOSISCOMPLEXINSPUTUM2513

TABLE5.Comparativesensitivitiesandspecificitiesofreal-time

PCRwithreferencetoconventionalmycobacterialculture

Sputumspecimens(nϭ335)

aNo.ofspecimensdetectedby

real-timePCRbNon-Beijing/W

Beijing/W

M.tuberculosisculturepositive

(nϭ74)

AFBsmearpositive(nϭ47)AFBsmearnegative(nϭ27)NTMculturepositive(nϭ31)cAFBsmearpositive(nϭ7)AFBsmearnegative(nϭ24)Culturenegativeformycobacterium

(nϭ230)

AFBsmearpositive(nϭ6)AFBsmearnegative(nϭ224)

abdetectionlimitoftheassayis10bacilli,asdeterminedbyendpointdilutionanalysis.Aliquotsof100␮lofTEbufferwereincludedineachextractionandreal-timePCRexperimentasnegativecontrols.Cross-contaminationbetweenspecimenswasnotdetected.InternalcontrolsremainpositiveforallM.tuberculosis-negativespecimens,indicatingthatinhibitionofamplificationwasnotencounteredinthisevaluation.

472700

286

DISCUSSION

00

5*10*32

AFBsmear,acid-fastbacillusdirectsmearmicroscopy.

*,ClinicalinformationregardingthecorrespondingpatientsisgiveninTable6.cThisnumberincludes7M.avium,6M.chelonae,6M.fortuitum,1M.gordonae,4M.kansasii,2M.neoaurum,1M.simiae,1M.terrae,and3RunyongroupIIImycobacterialspecimens.

goldstandard,theoverallsensitivityandspecificityare100%(74/74)and94.3%(246/261),respectively.Takingintoaccountthefinalclinicaldiagnosisofthepatients,thespecificityis100%(246/246).AssociationofBeijing/WgenotypeswithAFBsmear-positiveandAFBsmear-negativespecimenswere59.6%(31/52)and21.6%(8/37),respectively.Theoverallprev-alenceofBeijing/Wis43.8%(39/).TheresultwasfurtherconfirmedbyDTM-PCR(7)withcompleteconcordance.The

SensitivedetectionofM.tuberculosiscomplexfrombodyfluidsrequiresoptimalcelllysisandefficientDNApurificationtoremoveassociatedPCRinhibitors.Thechoiceofreagentsandmethodsisthuscrucial.ChelexandtheQIAmpDNAextractionkitarecommonlyusedinDNApreparations,al-thoughtheQIAmpDNAextractionkitisincompatiblewithmainstreamautomationforhighthroughput.

Accordingtotheresultsofthepresentstudy,thereagentsandmethodsusedinDNAextractionhaveagreatimpactontheDNAyield.ComparisonusingaspikedspecimenpoolsuggeststhatS-HMisthemostsensitivesinceitrequiredthefewestcyclestoreachthethreshold(CT)inreal-timePCRassay,followedbyZ-HM,Z-D,Z-V,S-D,andS-V.Intermsofsensitivity,thesesixprotocolsarebetterthantheotherstested.Theevaluationwithclinicalspecimensyieldedsimilarbutnotidenticalorder.Z-HM,Z-V,Z-D,andS-Vweresignificantlymoresensitivethanotherprotocols,whiletheperformanceofS-HMandS-Dwereinferior.Apossibleexplanationisthattheefficienciesofmechanicaldisruptionbybeadscouldbeaffectedbytheviscosityandtheactualbacterialloadofsputumspeci-

TABLE6.Clinicalinformationfor15culture-negativecasesdeterminedtobepositivebyreal-timePCR

Case

Gender/age(yr)

AFBsmear

Clinicalinformationa1234567101112131415

aM/70F/36F/70M/69M/73M/72M/75F/20M/79M/58M/52M/M/77M/50M/59

PositivePositivePositivePositivePositiveNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegative

M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;onantituberculosistreatment.

BothsputumandcervicallymphnodespuspositiveinAFBsmear;apicalnodules;possiblygranulomas;onantituberculosistreatment.

M.tuberculosisculturepositiveofanothersputumspecimenandBALduringthesameadmission,tissuepositiveinAFBsmear;onantituberculosistreatment.

M.tuberculosisculturepositiveforanothersputumspecimenfromlastadmission;onantituberculosistreatment.

M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;onantituberculosistreatment.

CTscanofrightupperzoneshowedtree-in-budappearance;onantituberculosistreatmentpriortoadmission.

M.tuberculosisculturepositiveforanothersputumspecimenduringasubsequentadmission1monthlater;pleuraleffusion;onantituberculosistreatment.M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;onantituberculosistreatment.

M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;startedantituberculosistreatment;deceased.

M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;onantituberculosistreatment.

M.tuberculosisculturepositiveforanothersputumspecimenduringthesameadmission;onantituberculosistreatment.

AFBsmearpositiveinbiopsyofupperrightlobeoflung;presenceofgranulomas;onantituberculosistreatment.

TBhistory;recentsputumpositiveinAFBsmear;onantituberculosistreatment.

TBhistory;completedantituberculosistreatment;recurrenthemoptysisandbronchiectasis.TBhistory;completedantituberculosistreatment.

BAL,bronchoalveolarlavage.

2514LEUNGETAL.mens.Theresultspresentedheresuggestthatzirconiabeadsaremorecompatiblewiththewiderangeofsputumviscosities.Inbothexperimentsusingspikedspecimensandclinicalspecimens,theperformanceofChelexresinsandlarge(1-mm)glassbeadswasrelativelypoor.Forinstance,theCTofthecombinationusinglargeglassbeadsplusChelex(L-C)withspikedspecimenswas30.43Ϯ1.12,whichsuggeststheDNAyieldwasϳ32-foldlowerthanthatofS-HM.Fromourobser-vation,1-mmglassbeadswererelativelysluggishcomparedtotheothertwotypesofbeadsduringvortex.Themildbombard-mentduetoitsheavyweightmightleadtoincompletecelllysisandlowDNAyield.Ontheotherhand,Chelexpurificationsuffersfromamajordisadvantageofdilutioneffect.WhilecolumnandmagneticpurificationsignificantlyconcentratedtheDNAbysmall-volumeelution,Chelexpurificationinevita-blycauseda2-folddilution,whichmayaccountforitspoorperformance.Takingintoconsiderationtheresultsfromthecomparisonanalysiswithboththespikedandtheclinicalspec-imens,thecombinationZ-HMwasselectedforitssimplicity,highsensitivity,lowcostrelativetocommercialspincolumns,andpotentialforhigh-throughputautomation.

AssaysthatdetectinsertionsequenceIS6110formoleculardiagnosisofM.tuberculosis,suchastheonedescribedhere,benefitsfromthefactthatIS6110oftenpresentsinhighcopynumbers(Ͼ10copies)inthegenomeofM.tuberculosis.AcommonlimitationofIS6110detectionsistheinabilitytodis-criminatemembersoftheM.tuberculosiscomplexsuchasM.bovisandM.africanum,althoughtheprevalenceoftheseagentsinhumanisconsiderablylessthanthatofM.tubercu-losis.Inaddition,asmallportion(Ͻ2%)ofM.tuberculosisisolatesdoesnotcontainIS6110repeatedsegmentsintheirgenome,whichwouldgiverisetofalse-negativeresults.Ournovelmultiprobemultiplexreal-timePCRsuccessfullydetectsall74cultivableM.tuberculosisisolateswithnocross-reactionwithNTM.Fifteenspecimensdeterminedtobepositivebyreal-timePCRwereculturenegative.SincecultureisolationisthegoldstandardforTBdiagnosis,thepossibilityoffalsepositivityamongthesespecimensshouldnotbeexcluded.However,whentheclinicalpresentationofthecorrespondingpatientsandtheresultsofrepeatedsamplingaretakenintoconsideration,thechancethatthesearerealfalsepositivesislow(Table6).Discrepanciesbetweenconventionalmycobac-terialcultureandPCR-baseddetectionhavebeenpreviouslyassociatedwithTBpatientswhounderwentantitubercularche-motherapies(2,18).SincePCRcoulddetectDNAfromnon-viableM.tuberculosisasaresultofantituberculartreatmentandalsoviableM.tuberculosisininsufficientquantityforsuc-cessfulculture,thediscrepancyfoundinthepresentstudy,togetherwiththeavailableclinicalinformation,indicatesanadvantageinsensitivityofPCR-baseddetectionoverconven-tionalmycobacterialculture.

TheworldwideprevalenceofBeijinggenotypevariesfrom0toϾ80%dependingongeographiclocations(19).Ourfindingof43.8%concurswiththecharacteristichighprevalenceintheAsia-Pacificregion.However,itshouldbenotedthatthespec-imenscollectedforthisassaydevelopmentweresubjectedtopriorselectionbasedonthetwocriteriastated;hence,thepercentageisnotnecessarilyarepresentativereflectionoftheregionalstatus.

Contrastingconventionalmycobacterialculture,themolec-

J.CLIN.MICROBIOL.

ularprotocoldescribedinthepresentstudygreatlyreducestheturnaroundtime,whichwouldbebeneficialtobothTBcontrolandpatienttreatmentoutcome.ItalsoavoidsthehazardsofmaintainingaTBcultureroom.However,theprotocolismorelabor-intensiveunlessautomationisavailable.AlthoughthecostofreagentsinvolvedissignificantlylowerthanthatofcommerciallyavailablerapidTBdiagnosissystems,itisstillhigherthanconventionalmycobacterialculture,andthiscouldbeabarrieratthemoment.Asthecostofmolecularreagentsgraduallydrops,onemayfindthisapproachjustifiableforitspotentialhealthcarebenefitsintheforeseeablefuture.Incon-clusion,themethodusedtoextractDNAfromsputumspeci-mensgreatlyaffectstheoutcomefrommoleculardiagnosis.WedescribehereanoptimizedprotocolforDNAextractionfromsputumspecimensandamultiprobemultiplexreal-timePCRforthesimultaneousdetectionofM.tuberculosisandBeijing/Wgenotypewithhighsensitivityandspecificity.

ACKNOWLEDGMENTS

WesincerelythankRaymondLai,DepartmentofMicrobiology,PrinceofWalesHospital,forhisgeneroussupportandW.Y.Laufromtheclinicalmicrobiologylaboratoryforhisassistanceinspecimencollection.

ThisstudywassolelysupportedbytheResearchFundfortheCon-trolofInfectiousDiseases(RFCID08070212),FoodandHealthBu-reau,SARgovernment.

REFERENCES

1.Aldous,W.K.,J.I.Pounder,J.L.Cloud,andG.L.Woods.2005.Compar-isonofsixmethodsofextractingMycobacteriumtuberculosisDNAfromprocessedsputumfortestingbyquantitativereal-timePCR.J.Clin.Micro-biol.43:2471–2473.

2.Bennedsen,J.,etal.1996.UtilityofPCRindiagnosingpulmonarytubercu-losis.J.Clin.Microbiol.34:1407–1411.

3.Bifani,P.J.,B.Mathema,N.E.Kurepina,andB.N.Kreiswirth.2002.GlobaldisseminationoftheMycobacteriumtuberculosisW-Beijingfamilystrains.TrendsMicrobiol.10:45–52.

4.Boom,R.,etal.2005.Rapidandsimplemethodforpurificationofnucleicacids.J.Clin.Microbiol.28:495–503.

5.Burman,W.J.,etal.2009.RelapseassociatedwithactivediseasecausedbyBeijingstrainofMycobacteriumtuberculosis.Emerg.Infect.Dis.15:1061–1067.

6.Chang,J.R.,etal.GenotypicanalysisofgenesassociatedwithtransmissionanddrugresistanceintheBeijinglineageofMycobacteriumtuberculosis.Clin.Microbiol.Infect.,inpress.

7.Chen,J.,etal.2007.Deletion-targetedmultiplexPCR(DTM-PCR)foridentificationofBeijing/WgenotypesofMycobacteriumtuberculosis.Tuber-culosis(Edinb.)87:446–449.8.Edwards,U.,T.Rogall,H.Blo¨cker,M.Emde,andE.C.Bo¨ttger.19.Isolationanddirectcompletenucleotidedeterminationofentiregenes:characterizationofagenecodingfor16SrRNA.NucleicAcidsRes.17:7843–7853.

9.EuropeanConcertedActiononNewGenerationGeneticMarkersandTech-niquesfortheEpidemiologyandControlofTuberculosis.2006.Beijing/WgenotypeMycobacteriumtuberculosisanddrugresistance.Emerg.Infect.Dis.12:736–743.

10.Jiang,X.,etal.2007.IdentificationofuniquegeneticmarkersinRv0927c

amongMycobacteriumtuberculosisW-Beijingstrains.MicrobesInfect.9:241–246.

11.Leung,K.L.,etal.2009.Developmentofasimpleandlow-costreal-time

PCRmethodfortheidentificationofcommonlyencounteredmycobacteriainahighthroughputlaboratory.J.Appl.Microbiol.107:1433–1439.

12.Manca,C.,etal.2005.HypervirulentMycobacteriumtuberculosisW/Beijing

strainsupregulatetypeIIFNsandincreaseexpressionofnegativeregulatorsoftheJak-Statpathway.J.InterferonCytokineRes.25:694–701.

13.Rad,M.E.,etal.2003.MutationsinputativemutatorgenesofMycobacte-riumtuberculosisstrainsoftheW-Beijingfamily.Emerg.Infect.Dis.9:838–845.

14.Reed,M.B.,etal.2004.Aglycolipidofhypervirulenttuberculosisstrains

thatinhibitstheinnateimmuneresponse.Nature431:84–87.

15.Reed,M.B.,S.Gagneux,K.Deriemer,P.M.Small,andC.E.Barry3rd.

2007.TheW-BeijinglineageofMycobacteriumtuberculosisoverproducestriglyceridesandhastheDosRdormancyregulonconstitutivelyupregulated.J.Bacteriol.1:2583–25.

VOL.49,2011DETECTIONOFM.TUBERCULOSISCOMPLEXINSPUTUM2515

16.Santos,A.,etal.2010.ComparisonofmethodsofDNAextractionfor

real-timePCRinamodelofpleuraltuberculosis.APMIS118:60–65.

17.Sun,Y.J.,etal.2006.TuberculosisassociatedwithMycobacteriumtubercu-losisBeijingandnon-Beijinggenotypes:aclinicalandimmunologicalcom-parison.BMCInfect.Dis.6:105.

18.Traore,H.,A.vanDeun,I.C.Shamputa,L.Rigouts,andF.Portaels.2006.

DirectdetectionofMycobacteriumtuberculosiscomplexDNAandrifampinresistanceinclinicalspecimensfromtuberculosispatientsbylineprobeassay.J.Clin.Microbiol.12:4384–4388.

19.vanSoolingen,D.,etal.1995.PredominanceofasinglegenotypeofMyco-bacteriumtuberculosisincountriesofeastAsia.J.Clin.Microbiol.33:3234–3238.

20.Wong,D.A.,P.C.Yip,D.T.Cheung,andK.M.Kam.2001.SimpleandrationalapproachtotheidentificationofMycobacteriumtuberculosis,Myco-bacteriumaviumcomplexspecies,andothercommonlyisolatedmycobacte-ria.J.Clin.Microbiol.39:3768–3771.

21.WorldHealthOrganization.2006.ThestopTBstrategy:buildingonand

enhancingDOTStomeettheTB-relatedmillenniumdevelopmentgoals.ReportWHO/HTM/TB/2006.368.WorldHealthOrganization,Geneva,Switzerland.

22.WorldHealthOrganization.2010.Globaltuberculosiscontrol.Report

WHO/HTM/TB/2010.7.WorldHealthOrganization,Geneva,Switzerland.23.Wu,S.,etal.2009.ActivationoftheeisgeneinaW-Beijingstrainof

MycobacteriumtuberculosiscorrelateswithincreasedSigAlevelsanden-hancedintracellulargrowth.Microbiology155:1272–1281.