agarose44pzpz.com

1Thecommensalinfantgutmeta-mobilomeasapotentialreservoirforpersistentmultidrugresistanceintegronsAnuradhaRavi1,*,EkaterinaAvershina1,*,StevenL.
Foley2,JaneLudvigsen1,OlaStorr3,Torbjrnien3,RoarJohnsen3,AnneL.
McCartney4,TrineM.
L'Abée-Lund5&KnutRudi1Despitetheaccumulatingknowledgeonthedevelopmentandestablishmentofthegutmicrobiota,itsroleasareservoirformultidrugresistanceisnotwellunderstood.
Thisstudyinvestigatedtheprevalenceandpersistencepatternsofanintegrasegene(int1),usedasaproxyforintegrons(whichoftencarrymultipleantimicrobialresistancegenes),inthefecalmicrobiotaof147mothersandtheirchildrensampledlongitudinallyfrombirthto2years.
Thestudyshowedtheint1genewasdetectedin15%ofthestudypopulation,andapparentlymorepersistentthanthemicrobialcommunitystructureitself.
Wefoundint1tobepersistentthroughoutthefirsttwoyearsoflife,aswellasbetweenmothersandtheir2-year-oldchildren.
Metagenomesequencingrevealedintegronsinthegutmeta-mobilomethatwereassociatedwithplasmidsandmultidrugresistance.
Inconclusion,thepersistentnatureofintegronsintheinfantgutmicrobiotamakesitapotentialreservoirofmobilemultidrugresistance.
Thespreadofantibioticresistance(AR)genesanddevelopmentofmultidrugresistancerepresentmajorthreatstopublichealth1.
Untilrecently,pathogenshavebeentheprimefocuswithrespecttounderstand-ingthespreadofmultidrugresistance,withthecommensalmicrobiotareceivingmuchlessattention.
However,recentstudieshaveshowntheprevalenceofARgenesinthecommensalgutmicrobiota2–5.
Furthermore,thegutmicrobiotashowsahighrateofhorizontalgenetransfer(HGT),whichwasindi-catedtobeupto25-foldgreaterthanthatofbacteriainotherenvironments6.
Hence,thecollectivemobilegeneticelements(MGEs)inthegutmicrobiota(i.
e.
thegutmeta-mobilome)representanimpor-tanttargetforbothunderstandingandcombatingthespreadofmultidrugresistance5,7.
Thegutmicrobiotaformsacomplexecosystem.
Thegutisassumedsterileatbirth8,9whereasjustafterbirth,itgoesthroughmajorshiftsstartingwithfacultativeanaerobicbacteria(EnterococcaceaeandStreptococcaceae)10,11.
Asoxygenlevelsdeplete,strictlyanaerobicbacteria(BifidobacterialesandBacteroidetes)takeoveranddominateinthegut12.
Thisprogressionslowsdownasthemicrobiotareachestheadult-likestatewhereanestimated100–200speciesco-existincloseproximity13.
Althoughscientistshavestartedtounderstandtheshiftsinthetaxonomiccompositionofthedevelopingmicrobiotafrominfancytoadulthood,theknowledgeofthemeta-mobilome,includingthetransmissionandpersistenceofmultipleantimicrobialresistancegenes,islimited.
1NorwegianUniversityofLifeSciences,Chemistry,BiotechnologyandFoodsciencedepartment(IKBM),Campuss,s1432,Norway.
2NationalCenterforToxicologicalResearch,U.
S.
FoodandDrugAdministration,DivisionofMicrobiology,Jefferson,AR72079.
3DepartmentofPublicHealthandGeneralPractice,NorwegianUniversityofScienceandTechnology,9491Trondheim,Norway.
4MicrobialEcology&HealthGroup,DepartmentofFoodandNutritionalSciences,UniversityofReading,Reading,UK.
5NorwegianUniversityofLifeSciences,DepartmentofFoodsafetyandInfectionBiology,CampusAdamstuen,Oslo0454,Norway.
*Theseauthorscontributedequallytothiswork.
CorrespondenceandrequestsformaterialsshouldbeaddressedtoK.
R.
(email:knut.
rudi@nmbu.
no)received:03March2015accepted:21September2015Published:28October2015OPEN2Antimicrobialresistancegenescanbecarriedinintegrons,whicharenon-mobileelementsthem-selves,butareoftenfoundwithinMGEsliketransposonsandplasmids14,15.
Integronsareplatformsforintegration,assemblyandexpressionofspecificgenecassetteswithintheMGEsthatoftenencodeantimicrobialresistance16.
Theindividualgeneticcassettestypicallylacktheirownpromoters,butareexpressedbyacommonpromoterforallthecassetteswithintheintegron(Fig.
1).
Therehavebeen5classesofintegrons(classI–V)classifiedtodate5.
TheclassIintegronsarethemostwidelystudiedandarefoundinabroadhostrangeofcommensalandpathogenicbacteria17.
ClassIintegronsarefoundextensivelyinclinicalisolatescontainingseveraldifferentARgenecassettesconferringresistancetoanti-bioticscommonlyusedagainstbacterialinfections16,18.
Upto8genecassetteshavebeenfoundinasingleclassIintegron16,howeverhundredsofgenecassetteshavebeendetectedinso-calledsuper-integrons19.
TheaimofthecurrentstudywastoinvestigatetheprevalenceandpersistenceofclassIintegronsinalargeunselectedlongitudinalcohortofmothersandtheirchildren.
WeusedquantitativePCRtoidentifyandstudythepersistencepatternsofintegrons.
16SrRNAandmetagenomedeepsequencingwereusedtoanalyzethephylogenyandgeneticbackgroundoftheintegronsinthesamplesandtotracetheseelementslongitudinally.
MaterialsandMethodsTheschematicoverviewoftheworkflowisdisplayedinFig.
2.
Themethodswereperformedinaccord-ancetotheapprovedguidelinesandallexperimentalprotocolswereapprovedbyNorwegianUniversityofLifeSciences.
Cohortdescription.
IMPACT(ImmunologyandMicrobiologyinPreventionofAllergyamongChildreninTrondheim)studyisacontrollednon-randomizedlongitudinalstudy,whichbeganin2000.
TheregionalcommitteeforMedicalResearchEthicsforCentralNorwayhasapprovedtheIMPACTstudy(ref.
120–2000).
ThisstudywasgrantedalicensebytheNorwegianDataInspectoratetopro-cesspersonalhealthdataandoneoftheparentsofeachchildsignedawritteninformedconsentform(r.
2003/953-3KBE/-).
Currentcontrolledtrialsregistrationnumber:ISRCTN28090297.
Thestudyinvolved720pairsofpregnantwomenandtheirchildren(uptotwoyearsofage).
Ninetypercentofthechildrenwerevaginallydeliveredandatterm.
Ninety-sevenpercentoftheinfantswerebreast-fedexclusivelyforthefirstsixweeksoflife.
Fecalsampleswerecollectedfromthepregnantwomenduringthefirst/second(7–20weeks)trimesterandthethird(32–40weeks)trimester,andfromthechildrenat3–10days,4months,1and2yearsofage.
Inthecurrentstudy,samplesfromarandomlyselectedsubgroupof147mother-childpairsfromtheIMPACTcohortwereanalyzed.
Informationonallergyrelatedhereditarydiseases,atopyandantibioticusage;healthandexposurefactorsfortheparentandchildissummarizedinSupplementaryTableS1.
Figure1.
StructureofclassIintegron.
AgeneralrepresentationofaclassIintegronwithresistancegenecassettesattheattachmentsites(attC)andacommonpromoterforthecassettesasPcandfortheintegraseasPint.
Thefollowingcassettesareapartofthe3′conservedregionandnotmobile:sul1geneencodingresistancetosulfonamidesandqacEencodingresistancetoquaternaryammoniumcompounds.
IMPACTcohortn=663;id=147Intsamplesn=99;id=5816SrRNAsequencing&analysis16SrRNAseqn=465;id=145Overlapdatasetn=56;id=434000seq/sample16SrRNAsubsetn=378;id=141Int1qPCRLong-rangePCRLong-rangePCRsetn=6;id=5Fullmetagenomedatasetn=15;id=9Metagenomeanalysishighint1copy/samplesn=647;id=147Figure2.
Workflowofexperimentalsetup.
3Samplecollection.
FecalsamplesfromthesubjectsoftheIMPACTcohortwerecollectedinCary-Blairetransportandholdingmedium(BDDiagnostics,Sparks,MD).
Thesampleswerefrozenat-20°Cwithin2hfromcollection.
Thesampleswerethenstoredat80°Cwithinonemonthforchildrenandmothers.
DNApurification.
FecalDNAwaspurifiedwithanautomatedprotocolusingDNAextractionkitbasedonparamagneticparticles(LGCGenomics,UK).
Inbrief,thesamplesweresubjectedtomechan-icallysisusingglassbeadsandtheDNAwaspurifiedbyelutingfromtheparamagneticparticlesbydownstreamprocessesasdescribedbymanufacturer.
TheDNAwasstoredat40°C.
Genequantification.
Theabundanceofintegrons(usingtheintegrase(int1)gene20asaproxy)inthesampleswascalculatedrelativetothe16SrRNAgenebyquantitativePCR.
EachPCRreaction(25μl)contained1*HOTFIREPolPCRmix(SolisBioDyne,Estonia);200nMforwardandreverseprimers;oneμlofsampleDNAandwater.
ThereactionmixwasrunonLightCycler480(Roche,Germany).
FollowingthethermalcyclingtherawfluorescencedatawasexportedintoLinRegPCRprogram.
ThesoftwareperformedbaselinecorrectionsandcalculatedthemeanPCRefficiency.
Fortheint1amplicon,wealsousedHighResolutionMelting(HRM)curveanalysis,inadditiontoSangersequencingusingtheBigDyeTerminatorv.
1.
1chemistry(AppliedBiosystems)forverification.
Thethermalcyclingforthe16SrRNAprimer-pair(5′-TCCTACGGGAGGCAGCAGT-3′;5′-GGACTACCAGGGTATCTAATCCTGTT-3′)wasaninitialdenaturationof95°Cfor15minfol-lowedby40cyclesof95°Cfor30secand60°Cfor30sec.
Thisprimer-pairtargetsconservedregionsofthe16SrRNAgene21.
Theprimersflankingtheint1gene(5′-ACGAGCGCAAGGTTTCGGT-3′;5′-GAAAGGTCTGGTCATACATG-3′)fromSrumetal.
11wereusedwiththermalcyclingconditions95°Cfor15minand40cyclesof95°Cfor30sec,53°Cfor30secand72°Cfor30sec.
Microbialcommunityanalyses.
MicrobialcommunitieswereassessedusingIlluminasequencingof16SrRNAgeneamplicons(n=465),withsubsetssubjectedtofullmetagenome(n=15)andlong-rangePCRamplicon(n=6)analyses.
Forfullmetagenomicsanalysis,sampleswereselectedbasedonthehighrelativequantitiesofint1geneinthesamples.
Foralong-rangePCR,sixint1-positivesampleswererandomlychosenforamplification.
Long-rangeprimerswereusedtoamplifythesequenceflankingtheregionfromattItothe3′con-sensusregionincludingthegenecassettes(5′-GGCATCCAAGCAGCAAG-3′;5′-AAGCAGACTTGACCTGA-3′)11withtheTaKaRaLAPCRkitVer.
2.
1.
Thethermalcyclingconditionsof94°Cfor5minfollowedby35cyclesof98°Cfor10min,54°Cfor30secand72°Cfor1min,withthefinalextensionstepat72°Cfor5min.
TheresultantPCRproductswereanalyzedbyagarosegelelectrophoresisandIlluminasequencing.
Forfullmetagenome,long-rangeampliconandmetagenomeanalyses,gDNAwasrandomlyfrag-mented,tagged,amplifiedandpreparedforsequencingusingNexteraXTkit(Illumina,USA).
Portionsofthe16SrRNAgeneswereamplifiedusingPRK341F/PRK806RprimerstargetingV3-V4regions22,modifiedbyadditionofIllumina-specificadapters.
EachPCRreaction(25μl)contained1*HOTFIREPolPCRmix(SolisBioDyne,Estonia);200nMuniquelytaggedforwardandreverseprimers;1μlofsampleDNAandwater.
Thethermalcyclingconditionswere95°Cfor15minand30cyclesof95°Cfor30sec,50°Cfor1minand72°Cfor45sec.
PCRproductswerethenpooled,basedontheirconcentrationsmeasuredusingQuant-iTTMPicoGreendsDNAassaykit(LifeTechnologies,USA),column-purifiedusingE.
Z.
N.
A.
CyclePurekit(OmegaBio-tek,USA)andsubmittedforsequencing.
SequencingwasperformedonMiSeqplatform(Illumina,USA)usingV3sequencingchemistrywith300bppaired-endreads.
16SrRNAgeneampliconsampleswereprocessedatNorwegianSequencingCentre(Oslo,Norway),whereasfullmetagenomesamplesweresequencedin-house.
Bacterialculturing.
ForisolationofBifidobacteriumspecies,10-folddilutionsoffecalsamplesin1%peptonewaterwereanaerobicallyculturedonBeerensagarat37°C.
Isolatedcolonieswerethensubculturedtopurityusingthesameconditions.
DNAwasextractedforsequencingof16SrRNAgeneasdescribedabovetoconfirmisolatesbelongingtoBifidobacteriumgenus.
Three-foldserialdilutionsoffecalsamplesfromthecohortwerepreparedindistilledwater,cul-turedonlactoseagarandintrypticsoyabrothwith5%horseblood,incubatedat37°Cfor24h.
Thebrothwassupplementedwith0.
1%ofbothTween80andmagnesiumchloridetorecoverdamagedEnterobacteriaceaecells.
Dataanalyses.
16SrRNAgeneamplicondatawereanalyzedusingQIIMEpipeline23.
Sequenceswerefirstqualityfiltered(split_libraries.
py;sequencelengthbetween200bpand1000bp;minimumaveragequalityscore25;notmorethan6ambiguousbases;andnoprimermismatchallowed)andthenclus-teredat99%homologylevelusingclosed-referenceuclustsearchagainstGreengenesdatabase24(pick_closed_reference_otus.
py).
Persistenceofoperationaltaxonomicunits(OTUs)overtimeinindividualswasassessedusingmulti-waydecompositionPARAFACanalysisofmean-centeredabundancedata25.
ThisanalysisallowsdetectionoftheOTUsthatbringmostofthevariationintothesystem,simultane-ouslywithdetectingthetimepointsatwhichtheseOTUsaremostpronounced.
Simpson'sreciprocal4diversityindexandBray-Curtisdissimilarityindexwereusedforalpha-andbeta-diversityassessment,respectively.
MetagenomedatamappingandassemblywasperformedusingGeneiouspipelinefollowingauthors'recommendations26.
MG-RASTmetagenomeanalyzerwasusedtoanalyzethetaxonomyandfunctionalclassificationofthesamples27.
PATRICdatabase28inMG-RASTwasusedtochecktheintegronabun-danceinthesamples.
E-value<105wasusedasthecut-offtoselectintegronhits.
Int1genepersistencewascalculatedastheratioofthenumberofmother-childpairsinwhoint1wasdetectedatbothtimepointstothetotalnumberofmother-childpairsforwhoinformationforbothtimepointswasavailable.
Theoddsratioforint1genedetectionwascalculatedbytheratioofint1persistencetotheprevalenceofint1atalatertimepoint.
Fisherexacttest,PearsoncorrelationcoefficientandSpearmancorrelationcoefficientwereusedforpairwisecomparisonsofint1and16SrRNAdata(includingdiversity,OTUabundanceandbacterialclassabundancedata).
ThesignificanceofthechangeovertimewastestedwithFriedman'stest-anon-parametricversionofANOVAtestwhichtakesintoaccountrepeatedmeasurements.
Thechangeinint1generelativeabundancewasalsocomparedtothechangeinlog-transformedOTUrelativeabun-dancesovertimeinanattempttoidentifyOTUsthatcorrelatedtoint1.
Regressionandclassificationdecisiontreeswerealsobuiltinanattempttoidentifybacterialclassesthatcorrelatedtoint1.
DataanalyseswereperformedusingMATLABR2014asoftware(TheMathWorksInc.
,NatickMA,USA).
ResultsMicrobiotacompositionanddevelopment.
Thephylogeneticcompositionofthemicrobiotawasassessedusingdeep16SrRNAgenesequencing.
Allsamplesthatwereamplifiedwith16SrRNAgene-tar-getingprimersandfurtheramplifiedwithIllumina-adaptedprimersetwereincludedintheanalysis.
Intotal,sequencingdatawereavailablefor451samples.
Inaddition,sevenofthesampleswereana-lyzedintriplicatetodeterminetechnicalvariation,whichwasfoundtobelow(SupplementaryFig.
S1).
Theaveragequalityscoreforthesequencerangeof250–299bpwas25.
Onaverage,21,277sequencespersampleweregeneratedafterqualityfilteringandassembly.
Toensureevenamountofsequencinginformation,4,000readspersamplewererandomlypickedfromthefulldatasetbasedontherecommendationsbySrensenetal.
29.
Thefinaldatasetafterqualityfilteringandunificationofthesequencinginformationpersamplecomprised378samples,withatotalof8,288OTUsbelongingto27classes.
The10mostabundantclassescomprisednearly100%ofthemicrobiotaatallages(Fig.
3).
Stoolsamplesfromnewbornsand4-month-oldinfantsweresignificantlylowerinalpha-diversityandsignificantlyhigherinbeta-diversitythanstoolsamplesfrom2-year-oldsandtheirmothers(Fig.
4).
At1yearofage,bothalpha-andbeta-diversityestimatesweresignificantlyhigherthanthatof4month-olds.
TherewasahighdominanceofClostridiainstoolsamplesfrommothers,aswellasfrom1-and2-year-olds.
Fivebacterialclasseswererelativelyequalinabundanceinneonatalstoolsamplescollectedsoonafterbirth(3days),whereasActinobacteriabecamedominantthereafter(4–10days)andremainedsothroughatleastfirst4monthsofage.
By1yearofage,theaverageprofileofstoolsamplesfromchildrenhadstartedconvergingtowardstheadultprofile.
However,pronounceddifferencesintheabundanceofActinobacteriaandBacteroidiawereseenbetweenadultsand2-year-oldchildren,suggestingclimaxadultcommunitywasstillnotreachedby2yearsofage.
Figure3.
Bacterialclasscompositionofstoolsamplesofinfants(from3daysto2yearsofage)andtheirmothersduringearly(1/2trimester;EarlyPreg)andlate(finaltrimester;LatePreg)pregnancybasedonthedeepsequencingof16SrRNAgeneamplicons.
s,Numberofsamplesateachtimeperiod.
5Microbiotapersistenceandstability.
Thepersistenceof599mostabundantOTUsinthedataset(withanabundancelevel≥0.
5%inatleastonesample)wereanalyzedusingPARAFAC.
NosignificantassociationsofOTUstoagewereidentifiedwhenonlyconsideringthedetected/non-detectedinforma-tion.
Whenabundancelevelswereconsidered,twoOTUsbelongingtoBifidobacteriumspecies(B.
longumOTU594044andB.
breveOTU484303),andoneassignedtoEnterobacteriaceaefamily(OTU1109087),showedhigheststabilityovertimeinthecohort(Fig.
5).
Spearmancorrelationtestidentifiedtheper-sistenceoftheB.
longum-assignedOTU,whichhadahighestloadinginPARAFAC,from3–10daysto4monthsofage(correlationcoefficient=0.
49;p=0.
007).
ThetwootherOTUs,however,didnotshowanysignificantcorrelationsbetweentheagegroups.
Integrondistributionandpersistence.
ThedistributionofintegronswasanalyzedbyquantitativePCRoftheint1gene.
Allsampleswereincludedandamplificationwascontrolledby16SrRNAgeneamplification.
Outofinitial663IMPACTsamples,16failedtoamplifyPCRproductsusing16SrRNAgene-targetingprimersandthuswereexcludedfromtheanalysis.
Intotal,99ofthe647samplesanalyzedshowedthepresenceofintegrons.
Theprevalenceoftheintegron-positivesampleswashighestfrom4-month-oldchildrencomparedtoanyotherage(Fig.
6a).
Thehighestpersistencepatternsforintegronswereseeninchildrenbetween3–10daysand4months,and4monthsto1year(Fig.
6b).
Persistencebetweensomemother-childpairswasalsodetected.
Theint1genecopynumbersofthepositivesamples,correctedfortheestimatedgenomeequivalents30,weresignificantlyhigherinsamplesfrominfants(3–10daysand4months)comparedtobothpregnantmothersand2-year-oldchildren(Fig.
6c).
Forthechildrenwithpersistentint1genes,17%(1of6childrenwithantibioticusageinformation)receivedantibioticsduringthefirstyearoflife.
Inaddition,31%(46outof147)ofthechildreninthewholecohorthadantibioticusageinformationdocumented.
Correlationofint1geneto16SrRNAgene.
Detectionofint1genedidnotcorrelatetoalpha-diversity(Simpson'sreciprocalindex1/D=12.
3±1.
74[mean±SEM]and1/D=13.
7±0.
66forint1-positiveandint1-negativesubgroups,respectively)ortobeta-diversity(Bray-CurtisDissimilarityFigure4.
Diversitycharacteristicsofstoolsamplesofinfants(from3daysto2yearsofage)andtheirmothersduringearly(1/2trimester;EarlyPreg)andlate(finaltrimester;LatePreg)pregnancybasedonthedeepsequencingof16SrRNAgeneamplicons.
(a)Simpson'sreciprocalindexofalpha-diversity.
(b)Bray-Curtisdissimilarityindexofbeta-diversity.
*pvalue<0.
05;and***pvalue<0.
001.
6indexBC=0.
85±0.
03andBC=0.
86±0.
04forint1-positiveandint1-negativesubgroups,respectively).
Therewasalsonosignificantcorrelationdetectedbetweenalpha-diversityandint1generelativeabun-dance(correlationcoefficient=0.
389,p=0.
45).
WithrespecttoOTUquantity,themostpersistentOTU(B.
longumOTU594044)showedapositivecorrelationwiththeint1gene(p=0.
03)at3–10days.
Noothersignificantcorrelations,however,werefound(SupplementaryFig.
S2).
Additionally,itwasinvestigatedwhetherachangeinOTUrelativeabun-dancecouldbeassociatedtothechangeinint1generelativeabundanceovertime,buttherewasnotanOTUidentifiedthatwassignificantlyassociatedtotheint1gene.
Finally,theanalysesconcentratedontheOTUsthatweredetectedinallsamplesforwhichint1weredetected(SupplementaryTableS2);however,theseOTUsdidnotshowanyquantitativecorrelationswithint1either.
Therewereadditionalattemptstofindbacterialclassesthatmightcorrelatetoint1detectionorint1geneabundance;however,nosignificantpairwisecorrelationsbetweenbacterialclassesandint1geneabundanceweredetected(SupplementaryFig.
3).
Regressionandclassificationdecisiontreeswerethenbuilttotestforthecumulativeeffectsofbacterialclasses,buttheseanalysesalsosuggestedweakcorre-lationsbetween16SrRNAgeneandint1genedata(SupplementaryFig.
S4andSupplementaryFig.
S5forregressionandclassification,respectively).
Searchforint1geneinbacterialisolates.
Thedetectionoftheint1geneinthegenomesofsequencedrepresentativesofpersistent/stableOTUsidentifiedbyPARAFACwascarriedoutbyBLASTsearchingwhole-genomesequencingdatafrom16B.
longum,2B.
breveand10E.
colistrains,isolatedfrompreviouslypublishedsubsetoftheIMPACTdataset31,32.
Despitehighnumbersofhitsto16SrRNAgeneperisolate(451.
5±37.
8),whichwasusedasaproxyforthegenomecoverage,theanalysesfailedtoidentifyreadsthatshowedhomologytotheint1genesequenceidentifiedinourwork.
Long-rangePCRandampliconsequencing.
Sixintegron-positivesampleswererandomlyselectedforlong-rangePCRandsequencing.
Thereadswereassembledintocontigsandtwocontigsoflengths1541bpand1019bpshowedBLASThitstoE.
colistrainDK510(GQ906578.
1)containingdihydro-folatereductase(dfrA17)andaminoglycosideadenylyltransferase(aaDA5)genes(E-value=0)with100%identity(Fig.
7a)andE.
colistrainA30(KF921570.
1)containingdihydrofolatereductase(dfrA12),hypotheticalprotein(orfF)andaminoglycosideadenylyltransferase(aadA2)genecassettes(E-value=0)with100%identity(Fig.
7b),respectively.
Integronpresenceinshotgunmetagenomedata.
Fifteensamples(latepregnancy,n=1;3–10days,n=6;4months,n=5;and2years,n=3)havingmicrobiotaprofileinformationandhighestrel-ativeabundanceofint1wereselectedforshotgunmetagenomesequencing.
Onaverage,837,048readswithasizerangefrom35bpto301bpwereobtainedforeachsample.
ByNCBIBLASTsearches699Figure5.
SummaryofPARAFACanalysis(multi-waydecompositionwhichdefinesmostinfluentialOTUsovertime)ontherelativeabundancesof599mostabundantbacterialOTUs.
(A)MostinfluentialOTUs.
(B)TimepointsassociatedwithmostinfluentialOTUs.
7Figure6.
Prevalence,persistenceofintegron-positivesamplesandrelativequantityofintegronsinthepositivesamplesbetweentimepoints.
(a)Relativeprevalenceofintegron-positivesamplesinthedataset.
*Binomialtestingbetweenthehighestabundance(4months)andtherest(pvalue=0.
005).
(b)Persistenceofintegronsateachtimepoint.
Thenumbersrepresenttheodds-ratio;thecolorgradientrepresentsthepercentageofpersistencebetweentimepoints.
SignificantpvaluesbyFisherexacttestarealsoindicated(*pvalue<0.
05;**pvalue<0.
01;***pvalue<0.
001).
(c)Relativeintegronquantificationateachtimepoint(log(int1copies/genomeequivalent1)forintegron-positivesamples.
Errorbarsrepresentstandarderrorofthemean(SEM).
ThesignificantdifferencebetweensamplegroupswascalculatedbyKruskal-Wallistest;pvalue<0.
05isindicatedbybracketing.
Early_latePreg,samplescollectedfrommothersduringearly(7–20weeks)andlate(32–40weeks)pregnancy;3–10days,samplesfrom3-to10-day-oldinfants;4months,1yearand2years;samplesfrom4-month-oldinfants,1-year-oldand2-year-oldchildren,respectively.
116SrRNAcopiesofallsamplesfromdifferentagegroupswerenormalizedtoreflectgenomeequivalentstakingintoaccountcopynumberinformationgivenbyVetrovskyetal.
30.
8shotgunmetagenomicreadsfrom12sampleswereidentifiedthatshowedhighhomologytotheint1gene(E-value<105;averageidentity[range]97.
5%[85.
1%;100%];averagequerycoverage99.
7%[98.
4%;100.
0%]).
UsingtheMG-RASTmetagenomeanalyzer27,itwasfoundthatallthesamplesshowedthepresenceofintegronsandintegron-relatedgenes.
TheidentityoftheintegronhitsofthesampleswereobtainedfromPATRICdatabase(SupplementaryTableS3).
Metagenomeassemblyandidentificationofcompleteintegrons.
Thereadswereextractedthatshowedint1homologyinonlyonedirectionofthepaired-endreads(n=71)toinvestigatethegeneticbackgroundoftheirpairedmates.
ByBLASTsearchingofthesesequencesagainstNCBIdatabase,can-didateplasmidpSH1148_107(GenBankJN983049)wasidentifiedthatwasmostprevalentamongthehits(SupplementaryTableS4).
Themetagenomicreadswerethenmappedontothecompleteplasmidsequenceandapproximately60%oftheplasmidwasencompassedbythemetagenomicreads.
Seventeenofthe25conjugationproteinsoftheplasmidmappedtoourreads,includingtheInc1conjugativetransferproteins,DNAprimaseandpilusbiogene(SupplementaryFig.
S6).
Thereadspartiallycoveredtheoriginofreplication.
TherewasonechildwhoshowedhighprevalenceofaplasmidrelatedtopSH1148_107(morethan1%ofallreads)instoolsamplesfromboth3–10daysand4months(20*and34*meancoveragefor3–10daysand4months,respectively).
The3–10daysand4monthsreadsmappedsimilarlytotheplasmid.
Thedenovoassemblyofthereadsmappedtoatransposoncontainingintegronwiththesul1geneandaadAgenecassette,whichwassimilartotheresistancegenesinpSH1148_107,andanadditionaldfrA17genecassette(Fig.
8).
Thegenecassettesencoderesistancetosulphonamides,spectin-omycinandstreptomycin,andtrimethoprimrespectively.
Thelong-rangePCRampliconcontigswerealsomappedtotheintegronassembledfromourmetage-nome.
The1541bp-longcontigshowed97%coverage,suggestingbothassembliescamefromthesameintegron.
Theothercontigof1019bplengthhaddifferentgenecassettesandthusshowedonlypartialcoverage.
Taxonomicrangeoftheintegronsidentifiedbylong-rangePCR.
BLASTsearchingoftheNCBIdatabasewiththeint1-containingcontigsidentifiedbylong-rangePCRrevealedhighhomology(100%pairwiseidentitywith100%querycoverage)towardsplasmidsisolatedfromE.
coli,Kluyverageorgiana,SalmonellaentericaandShigellaflexneri(SupplementaryTableS5),allbelongingtoEnterobacteriaceaefamily.
Searchforintegronsinothermetagenomes.
Tosearchforthesameintegroninotherpublicallyavailablemetagenomes,datawasextractedfrom60metagenomesamplesfromthecohortprovidedbyYatsunenkoetal.
33.
TheavailablecohortcontainedfecalsamplesfromhealthychildrenandadultsinMalawi,UnitedStatesandVenezuela;and20metagenomesfromeachoftherespectivecountrieswasanalyzed.
Eleven(18.
3%)ofthemetagenomesshowedthepresenceofint1gene.
Sevenoftheint1-positivemetagenomesamplesalsocontainedreadsmappingtothetransposonflankingregions.
However,theintegron-associatedgenecassetteswerenotsimilartothosedetectedinourdataset(SupplementaryTableS6).
DiscussionSeveralstudieshaveshownahighprevalenceofARgenesininfantswiththeabsenceofantibiotictreat-ments3,34,35whichisinlinewithourfindings.
However,toourknowledge,thisstudyisthefirstonetoobservehighint1geneprevalenceandpersistence.
Ahighprevalenceofintegronswasfoundat3–10daysand4monthsofage.
Intheearlyperiodsoflifetheresistanceagainstcolonizationbyexogenousbacteriaislow35,thereforeopeningforthepossibilityofestablishmentofbacteriafromtheenvironment.
Figure7.
Integronsdetectedbylong-rangePCR.
(a)1.
5kbpartiallysequencedintegronbylong-rangePCRproduct.
(b)1.
1kbpartiallysequencedintegronbylong-rangePCR.
Cylindricalboxesshowindividualgenesthataresizedependent,i.
e.
largerboxislongergene;dottedarrowsindicatethedirectionoftranscription;andgradientbluecolortheendoftheacquiredsequence.
Geneandstructuralfeatures:attI,primaryrecombinationsite;dfrA17anddfrA12,dihydrofolatereductase;attC,recombinationsite;aadA5andaadA2,aminoglycosideadenylyltransferase;andorfF,hypotheticalprotein.
9Aplausibleexplanationforthehighintegronprevalenceatearlyagecouldbethehospitalenvironment,sincechildrenarefirstexposedtothisatmosphere36.
Therewasalsopersistenceofint1genethrough-outthefirsttwoyearsoflifeandbetweenmothersandtheir2-year-oldchildren,pointingtowardsmaternalsourceasanotherpotentialroutefortransmission.
Similarpatternshavealsobeendetectedintransposon-associatedgenesinmother-infantpairs3,31.
Analternativeexplanation,though,couldbethecolonizationbyvariousintegronsatdifferentages.
However,takingintoaccounttheincreasedlikelihoodofint1detectionatonetimeperiodgivenitwasdetectedpreviously,themoreprobableexplanationwouldbethepersistenceofthesameintegronratherthanthedetectionofindependentmultiplecolo-nizationevents.
ThepersistenceofintegronsinthegutmicrobiotaindicatestheversatilityofMGEstoendurethedrasticchangesthatoccurduringfirstyearsoflife10,33.
However,itisunlikelythatantibiotictreatmentinfluencesthepresenceofmultidrugresistanceintegronssincewedidnotfindanyalterationofpersis-tencepatternsinourdatasetassociatedwithantibioticusage.
Diversityestimatesofthecohortcorrespondedwellwithpreviouslypublishedobservationsofincreaseinalpha-anddecreaseinbeta-diversitywithage11,33.
Interestingly,whenint1geneabundancewashighestatearlydaysoflife,themicrobialdiversitywaslowest,suggestingthatint1geneshouldbeassociatedtothosefewbacteriathatareestablishedbythen.
However,therewasnocorrelationbetweenint1geneanddiversityestimatesorbacterialclasses.
Moreover,despitenumerousattempts,wecouldnotassociateint1genetoanyparticularphylotypeacrossindividualswithinourcohort.
Hence,itisunlikelythattheintegronshaveastrictphylotypeassociation.
Inaddition,whenwetriedtosearchforint1geneinBifidobacteriumisolatesthatrepresentthemostabundantbacterialgroupininfancyandthatwastheonlybacterialgenuscorrelatingtoint1geneabundanceatearlyinfancy,wefailedtofindanyindicationofintegronpresenceinitsgenomes.
Lackofassociationbetweenintegronsandphylotypesacrosslargephylogeneticdistanceshaspreviouslybeenobserved37.
Statisticalinconsistencieshavebeenreportedwhenphylogenetictreeswereobtainedforint1geneandmolecularmarkerforphylogenysuchasRNApolymerasesubunitB(rpoB)37.
Therefore,giventhebroadhostrangeforintegrons38,themostplausibleexplanationforthelackofphylotypeassociationishighratesofHGT.
InconcordancewithpotentiallyhighHGTrates,apossibletransposoncarryinganintegronwasidentifiedinoursamples,suggestingMGEsasthelikelyvehicleformobilityofintegrons.
Themobilenatureofintegrons-associatedMGEshasbeenpreviouslyobservedinpathogenicbacteria39,40,environmentalsamples41andinhospitalenvironments42.
Wealsoobservedthepersistenceofatransposon-containingintegrononapotentialconjugativeplasmidinoneinfantattwotimeperiods.
ThisintegroncontainedgenesassociatedwithaminoglycosidesandsulfonamideresistancesimilartotheconjugativeplasmidpSH1148_107,alongwithandadditionaltrimethoprimresistancegene.
Weexpandedoursearchforintegronsfromdifferentsamplesinourdatasetbyinvolvinglong-rangePCRsthatcouldamplifythewholeintegron.
TwoclassIintegronswereidentifiedwithpotentialasso-ciationtoamobileelementhavingresistancegenestotrimethoprim,streptomycinandspectinomycin.
Interestingly,astudybyShahcheraghietal.
alsofoundasimilarintegroncontainingresistancegenesinenteropathogenicE.
colistrains(JX442969.
1)isolatedfromfecalsamplesofchildrenlessthan5yearsofage43.
TheseevidencesgivefurthersupportthatintegronscanbereservoirsforARgenesininfants,withthepotentialfortransmissiontopathogens4,35.
Additionally,wealsodetectedintegronswithdifferentgenecassettesinpubliclyavailablemetagenomes,suggestingthediversityofintegronsinglobalhumanpopulations.
Ourobservationofintegron-containingelementsregardlessofantibioticsintakesuggeststhattheycanpersistwithoutouterselectionpressure.
ArecentstudyonthegutmicrobiotaofanisolatedgroupofYanomaniAmerindiantribeshowedasimilarpatternofthecarriageofapoolofmobilizablenext-generationantibioticresistancegeneswithoutanypriorantibioticpressure44.
Moreover,Sternandcolleaguesfoundover10,000contigscontainingpotentialmobileelementsintheMetaHITdataset45,whichwerelikelytobequitecommonconstituentsofthegutmicrobiotasinceallwereidentifiedastargetsforCRISPRelements.
Interestingly,onlyaround10%ofthesecontigswereofviralnature,leavingtheresttoplasmidsandMGEs,suggestingthatthehostactuallycounterselectsthesemobilesul1dfrA17IntegraseaadA5qacEdeltaTn21proteins5'3'MobileelementproteinsChromatetransportproteinFigure8.
Graphicalrepresentationofatransposon-containingintegronbydenovoassembly.
Theboxesillustratethecodingregionofgenes,darkbluerepresentsgenesofthetransposonandthelightblueindicatesgenesoftheintegron.
Geneticfeatures:dfrA17,trimethoprimresistanceprotein;aadA5,streptomycinandspectinomycinresistanceprotein;qacEdelta,quaternaryammoniumcompoundresistanceprotein;sul1,sulphonamideresistanceprotein.
10elements.
Thisfindingsupportstheselfishparasitic-likespreadofconjugativeplasmidsassociatedintegronsinthegut.
Theoverallresultsofthestudyprovideevidenceforhighprevalenceofintegronsinthefecalmicrobi-otaatearlystagesoflifeandfurthersuggestthatthecommensalgutmicrobiotacanserveasareservoirformultidrugresistance,potentiallycontributingtoitsrapidspread.
References1.
McKenna,M.
Antibioticresistance:thelastresort.
Nature499,394–396,doi:10.
1038/499394a(2013).
2.
Hu,Y.
etal.
Metagenome-wideanalysisofantibioticresistancegenesinalargecohortofhumangutmicrobiota.
Naturecommunications4,2151,doi:10.
1038/ncomms3151(2013).
3.
deVries,L.
E.
etal.
TheGutasReservoirofAntibioticResistance:MicrobialDiversityofTetracyclineResistanceinMotherandInfant.
PloSone6,e21644,doi:10.
1371/journal.
pone.
0021644(2011).
4.
Sommer,M.
O.
A.
,Dantas,G.
&Church,G.
M.
FunctionalCharacterizationoftheAntibioticResistanceReservoirintheHumanMicroflora.
Science325,1128–1131,doi:10.
1126/science.
1176950(2009).
5.
Ravi,A.
etal.
IntegronsintheIntestinalMicrobiotaasReservoirsforTransmissionofAntibioticResistanceGenes.
Pathogens3,238–248(2014).
6.
Smillie,C.
S.
etal.
Ecologydrivesaglobalnetworkofgeneexchangeconnectingthehumanmicrobiome.
Nature480,241–244,doi:10.
1038/nature10571(2011).
7.
Modi,S.
R.
,Lee,H.
H.
,Spina,C.
S.
&Collins,J.
J.
Antibiotictreatmentexpandstheresistancereservoirandecologicalnetworkofthephagemetagenome.
Nature499,219–222,doi:10.
1038/nature122128.
Sekirov,I.
,Russell,S.
L.
,Antunes,L.
C.
M.
&Finlay,B.
B.
GutMicrobiotainHealthandDisease.
Physiol.
Rev.
90,859–904,doi:10.
1152/physrev.
00045.
2009(2010).
9.
Jiménez,E.
etal.
IsolationofCommensalBacteriafromUmbilicalCordBloodofHealthyNeonatesBornbyCesareanSection.
Curr.
Microbiol.
51,270–274,doi:10.
1007/s00284-005-0020-3(2005).
10.
Palmer,C.
,Bik,E.
M.
,DiGiulio,D.
B.
,Relman,D.
A.
&Brown,P.
O.
DevelopmentoftheHumanInfantIntestinalMicrobiota.
PLoSBiol.
5,e177,doi:10.
1371/journal.
pbio.
0050177(2007).
11.
Avershina,E.
etal.
Majorfaecalmicrobiotashiftsincompositionanddiversitywithageinageographicallyrestrictedcohortofmothersandtheirchildren.
FemsMicrobiol.
Lett.
n/a-n/a,doi:10.
1111/1574-6941.
12223(2013).
12.
Avershina,E.
&Rudi,K.
IsitwhoyouareorwhatyoudothatisimportantinthehumangutBenef.
Microbes.
inpress(2013).
13.
Faith,J.
J.
etal.
TheLong-TermStabilityoftheHumanGutMicrobiota.
Science341,doi:10.
1126/science.
1237439(2013).
14.
Kovalevskaya,N.
P.
MobileGeneCassettesandIntegrons.
Mol.
Biol.
36,196–201,doi:10.
1023/A:1015361704475(2002).
15.
Fluit,A.
C.
&Schmitz,F.
J.
Resistanceintegronsandsuper-integrons.
Clin.
Microbiol.
Infect.
10,272–288,doi:10.
1111/j.
1198-743X.
2004.
00858.
x(2004).
16.
Mazel,D.
Integrons:agentsofbacterialevolution.
Nat.
Rev.
Microbiol.
4,608–620,doi:10.
1038/nrmicro1462(2006).
17.
Gillings,M.
R.
etal.
Usingtheclass1integron-integrasegeneasaproxyforanthropogenicpollution.
ISMEJ.
doi:10.
1038/ismej.
2014.
226(2014).
18.
Gillings,M.
R.
,Xuejun,D.
,Hardwick,S.
A.
,Holley,M.
P.
&Stokes,H.
W.
Genecassettesencodingresistancetoquaternaryammoniumcompounds:aroleintheoriginofclinicalclass1integronsISMEJ.
3,209–215,doi:10.
1038/ismej.
2008.
98(2009).
19.
Rowe-Magnus,D.
A.
etal.
Theevolutionaryhistoryofchromosomalsuper-integronsprovidesanancestryformultiresistantintegrons.
Proc.
Natl.
Acad.
Sci.
USA98,652–657,doi:10.
1073/pnas.
98.
2.
652(2001).
20.
Sorum,H.
,L'Abee-Lund,T.
M.
,Solberg,A.
&Wold,A.
Integron-containingIncURplasmidspRAS1andpAr-32fromthefishpathogenAeromonassalmonicida.
Antimicrob.
AgentsChemother.
47,1285–1290(2003).
21.
Nadkarni,M.
A.
,Martin,F.
E.
,Jacques,N.
A.
&Hunter,N.
Determinationofbacterialloadbyreal-timePCRusingabroad-range(universal)probeandprimersset.
Microbiology148,257–266(2002).
22.
Yu,Y.
,Lee,C.
,Kim,J.
&Hwang,S.
Group-specificprimerandprobesetstodetectmethanogeniccommunitiesusingquantitativereal-timepolymerasechainreaction.
Biotechnol.
Bioeng.
89,670–679,doi:10.
1002/bit.
20347(2005).
23.
Caporaso,J.
G.
etal.
QIIMEallowsanalysisofhigh-throughputcommunitysequencingdata.
Nat.
Methods7,335–336,doi:10.
1038/nmeth.
f.
303(2010).
24.
DeSantis,T.
Z.
etal.
Greengenes,achimera-checked16SrRNAgenedatabaseandworkbenchcompatiblewithARB.
Appl.
Environ.
Microbiol.
72,5069–5072,doi:10.
1128/AEM.
03006-05(2006).
25.
Bro,R.
PARAFAC.
Tutorialandapplications.
ChemometricsandIntelligentLaboratorySystems38,149–171,doi:10.
1016/s0169-7439(97)00032-4(1997).
26.
Kearse,M.
etal.
GeneiousBasic:anintegratedandextendabledesktopsoftwareplatformfortheorganizationandanalysisofsequencedata.
Bioinformatics28,1647–1649,doi:10.
1093/bioinformatics/bts199(2012).
27.
Meyer,F.
etal.
ThemetagenomicsRASTserver-apublicresourcefortheautomaticphylogeneticandfunctionalanalysisofmetagenomes.
BMCBioinformatics9,386,doi:10.
1186/1471-2105-9-386(2008).
28.
Wattam,A.
R.
etal.
PATRIC,thebacterialbioinformaticsdatabaseandanalysisresource.
NucleicAcidsRes.
42,D581–591,doi:10.
1093/nar/gkt1099(2014).
29.
Srensen,L.
L.
,Coddington,J.
A.
&Scharff,N.
InventoryingandEstimatingSubcanopySpiderDiversityUsingSemiquantitativeSamplingMethodsinanAfromontaneForest.
Environ.
Entomol.
31,319–330,doi:10.
1603/0046-225X-31.
2.
319(2002).
30.
Vetrovsky,T.
&Baldrian,P.
Thevariabilityofthe16SrRNAgeneinbacterialgenomesanditsconsequencesforbacterialcommunityanalyses.
PloSone8,e57923,doi:10.
1371/journal.
pone.
0057923(2013).
31.
Avershina,E.
etal.
Bifidobacterialsuccessionandcorrelationnetworksinalargeunselectedcohortofmothersandtheirchildren.
Appl.
Environ.
Microbiol.
79,497–507,doi:10.
1128/AEM.
02359-12(2013).
32.
deMuinck,E.
J.
etal.
Diversity,transmissionandpersistenceofEscherichiacoliinacohortofmothersandtheirinfants.
Environ.
Microbiol.
Rep.
3,352–359,doi:10.
1111/j.
1758-2229.
2010.
00231.
x(2011).
33.
Yatsunenko,T.
etal.
Humangutmicrobiomeviewedacrossageandgeography.
Nature486,222–227,doi:10.
1038/nature11053(2012).
34.
Moore,A.
M.
etal.
Pediatricfecalmicrobiotaharbordiverseandnovelantibioticresistancegenes.
PloSone8,e78822,doi:10.
1371/journal.
pone.
0078822(2013).
35.
Sepp,E.
etal.
Theoccurrenceofantimicrobialresistanceandclass1integronsamongcommensalEscherichiacoliisolatesfrominfantsandelderlypersons.
Ann.
Clin.
Microbiol.
Antimicrob.
8,34(2009).
36.
Davies,J.
&Davies,D.
Originsandevolutionofantibioticresistance.
Microbiol.
Mol.
Biol.
Rev.
74,417–433,doi:10.
1128/MMBR.
00016-10(2010).
37.
Boucher,Y.
,Labbate,M.
,Koenig,J.
E.
&Stokes,H.
W.
Integrons:mobilizableplatformsthatpromotegeneticdiversityinbacteria.
TrendsMicrobiol.
15,301-309,doi:10.
1016/j.
tim.
2007.
05.
004(2007).
1138.
Nandi,S.
,Maurer,J.
J.
,Hofacre,C.
&Summers,A.
O.
Gram-positivebacteriaareamajorreservoirofClass1antibioticresistanceintegronsinpoultrylitter.
Proc.
Natl.
Acad.
Sci.
USA101,7118–7122,doi:10.
1073/pnas.
0306466101(2004).
39.
Meng,H.
etal.
Characterizationandhorizontaltransferofclass1integronsinSalmonellastrainsisolatedfromfoodproductsofanimalorigin.
Int.
J.
FoodMicrobiol.
149,274–277,doi:10.
1016/j.
ijfoodmicro.
2011.
07.
006(2011).
40.
Dalsgaard,A.
,Forslund,A.
,Serichantalergs,O.
&Sandvang,D.
Distributionandcontentofclass1integronsindifferentVibriocholeraeO-serotypestrainsisolatedinThailand.
Antimicrob.
AgentsChemother44,1315–1321(2000).
41.
Vaisvila,R.
,Morgan,R.
D.
,Posfai,J.
&Raleigh,E.
A.
Discoveryanddistributionofsuper-integronsamongpseudomonads.
Mol.
Microbiol.
42,587–601(2001).
42.
Gonzalez,G.
etal.
PresenceofintegronsinisolatesofdifferentbiotypesofAcinetobacterbaumanniifromChileanhospitals.
FEMSMicrobiol.
Lett.
161,125–128(1998).
43.
Shahcheraghi,F.
etal.
Identificationandcharacterizationofclass1integronsamongatypicalenteropathogenicEscherichiacoliisolatedfromchildrenunder5yearsofage.
Iran.
J.
Microbiol.
6,156–162(2014).
44.
Yeagle,P.
MicrobiomeofuncontactedAmerindians.
Science348,298,doi:10.
1126/science.
348.
6232.
298-a(2015).
45.
Stern,A.
,Mick,E.
,Tirosh,I.
,Sagy,O.
&Sorek,R.
CRISPRtargetingrevealsareservoirofcommonphagesassociatedwiththehumangutmicrobiome.
GenomeRes.
22,1985–1994,doi:10.
1101/gr.
138297.
112(2012).
AcknowledgementsThisworkwassupportedbyQuotascholarshipandfundingfromNorwegianUniversityofLifeSciences,s,NorwayandNorwegianUniversityofScienceandTechnology,Trondheim,Norway.
AuthorContributionsA.
R.
-mainauthorofthearticle.
E.
A.
-mainauthorofthearticle.
S.
F.
-Providedplasmidsandintegron-relatedexperiments.
J.
L.
-Culturingoffecalsamples.
O.
S.
-SamplecollectionforIMPACTstudy.
T.
.
-SamplecollectionforIMPACTstudy.
R.
J.
-SamplecollectionforIMPACTstudy.
A.
L.
M.
-IsolationofBifidobacterium.
T.
M.
L.
-reviewofscientificcontentinthearticle.
KR-mainleadofthearticleAdditionalInformationSupplementaryinformationaccompaniesthispaperathttp://www.
nature.
com/srepCompetingfinancialinterests:Theauthorsdeclarenocompetingfinancialinterests.
Howtocitethisarticle:Ravi,A.
etal.
Thecommensalinfantgutmeta-mobilomeasapotentialreservoirforpersistentmultidrugresistanceintegrons.
Sci.
Rep.
5,15317;doi:10.
1038/srep15317(2015).
ThisworkislicensedunderaCreativeCommonsAttribution4.
0InternationalLicense.
Theimagesorotherthirdpartymaterialinthisarticleareincludedinthearticle'sCreativeCommonslicense,unlessindicatedotherwiseinthecreditline;ifthematerialisnotincludedundertheCreativeCommonslicense,userswillneedtoobtainpermissionfromthelicenseholdertoreproducethematerial.
Toviewacopyofthislicense,visithttp://creativecommons.
org/licenses/by/4.
0/