Rho-dependent transcription termination proceeds via three ...

Rho is a general transcription termination factor in bacteria, but many aspects of its mechanism of action are unclear. Skiptomaincontent Thankyouforvisitingnature.com.YouareusingabrowserversionwithlimitedsupportforCSS.Toobtain thebestexperience,werecommendyouuseamoreuptodatebrowser(orturnoffcompatibilitymodein InternetExplorer).Inthemeantime,toensurecontinuedsupport,wearedisplayingthesitewithoutstyles andJavaScript. Advertisement nature naturecommunications articles article Rho-dependenttranscriptionterminationproceedsviathreeroutes DownloadPDF DownloadPDF Subjects BacterialtranscriptionTranscription AbstractRhoisageneraltranscriptionterminationfactorinbacteria,butmanyaspectsofitsmechanismofactionareunclear.DiversemodelshavebeenproposedfortheinitialinteractionbetweentheRNApolymerase(RNAP)andRho(catch-upandstand-bypre-terminationalmodels);fortheterminationalreleaseoftheRNAtranscript(RNAshearing,RNAPhyper-translocationordisplacing,andallostericmodels);andforthepost-terminationaloutcome(whethertheRNAPdissociatesorremainsboundtotheDNA).Here,weusesingle-moleculefluorescenceassaystostudythosethreestepsintranscriptionterminationmediatedbyE.coliRho.Wefindthatdifferentmechanismspreviouslyproposedforeachstepco-exist,butapparentlyoccuronvarioustimescalesandtendtoleadtospecificoutcomes.Ourresultsindicatethatthreekineticallydistinctroutestakeplace:(1)thecatch-upmodeleadsfirsttoRNAshearingforRNAPrecyclingonDNA,and(2)latertoRNAPdisplacementfordecompositionofthetranscriptionalcomplex;(3)thelastterminationusuallyfollowsthestand-bymodewithdisplacingfordecomposing.Thisthree-routemodelwouldhelpreconcilecurrentcontroversiesonthemechanisms. IntroductionTranscriptionterminationisessentialforpreciseexpressionandproperregulationofgenes.Inbacteria,intrinsicterminationrequiresonlycis-actingelementsonnascentRNAs,whilefactor-dependentterminationreliesonbothcis-actingRNAelementsandatrans-actingterminationfactor,Rho(ρ).Bacterialρ,participatingintranscriptiontermination1andmanyothers2,3,isanATP-dependentRNAtranslocase4.ρbindsnascentRNAatrut(Rhoutilization)sitesusingitsprimaryRNA-bindingsurfaceinanopenstateandwrapstheRNAwithitssecondaryRNA-bindingsurfacetotransformintotheactivestate4.Withrespecttothepre-terminationalmodesregardinghowρencountersRNApolymerase(RNAP)orelongationcomplex(EC)inordertomediatetermination,widelyacceptedisthecatch-upmechanism,inwhichρfirstbindsnascentRNAatarutsiteandcatchesupwithECthatispausingataterminationsiteaccordingtotheso-calledkinetic-coupling5,RNA-dependent6,RNA-centric7,andtracking8models.Thisclassicalmechanismforthepre-terminationalstepischallengedbythestand-bymechanism,inwhichρfirstbindsRNAPofstableECandstandsbyforbindinganincipientRNArutsiteemergingoutofRNAPaccordingtotheso-calledRNAP-dependent6,EC-centric7,andallosteric9models.Thisalternativepre-terminationalmechanismhasbeensupportedbyseveralbiochemical9,10,11andstructuraldata7,12,althoughquestionedbyotherbiochemicaldata13.FortheRNArelease,whichdefinesterminationoftranscription,thecatch-upmodegenerallyassumesacriticalroleinρ′smotoractivity.ThemechanicalforceexertedbyρisexpectedtoshearRNAofffromRNA·DNAhybridintheRNAshearingmodelordisplaceRNAPforwardoutoftheribonucleotide(NTP)incorporationsiteintheRNAPhyper-translocationmodel14.Thestand-bymode,ontheotherhand,generallyassumesthatRNAisreleasedorECisdisassemblednotbythemotoractionofρbutbyallostericchangesinECconformationtriggeredbyρ7,9,12.Astothepost-terminationaloutcomes,recentsingle-moleculestudiesrevealedthatRNAPsmostlyremainonDNAafterRNAreleaseatρ-independent,intrinsicterminationandaredirectlyreusedforreinitiationoftranscriptiononthesameDNAmolecule15,16,17.AfterreleasingRNAattermination,RNAPsdiffuseonDNAinupstreamanddownstreamdirections,changethemovingdirection,andevenflipthemselvesonDNA,untilreinitiatingtranscriptiononpromotersorfallingoffDNA15,16,17.Thispost-terminationalone-dimensional(1D)diffusionofRNA-freeRNAPonDNAisnamed1Drecycling,shortenedhereasrecycling,andthereinitiationbysuch1DrecyclingRNAPisnamed1Dreinitiation.Muchlessfrequentlyatintrinsictermination,ECdecomposesintothethreeessentialcomponentsallatoncewithsimultaneousornear-simultaneousdissociationsofRNAandRNAPfromDNA.Thisone-stepdisassemblyofECiscalledheredecomposingforshort.Whilerecyclingismuchmorefrequentthandecomposingfromintrinsictermination15,16,17,itisunknownyetwhetherrecycling,decomposing,orbothresultfromρ-dependenttermination.Inthisstudyusingsingle-moleculefluorescencemeasurements,wediscoverthatρ-dependentterminationofasingleterminatorfollowsthreedifferentroutesratherthanasinglepathwhilecharacterizingtheterminationeventsofindividualtranscriptioncomplexesinthreesteps;pre-terminationalsetup,terminationalrelease,andpost-terminationaloutcome.Thethreerouteseachleadtoaspecificdestinationonadistincttimescale.ResultsSingle-moleculemonitoringoftranscriptionterminationWeconstructedaDNAtemplatecontainingtheT7A1promoterandtheSalmonellamgtAleaderwithaρ-dependentterminator18,19(Fig. 1a,SupplementaryTable 1).ItsupstreamendislabeledwithbiotinforimmobilizationanditsdownstreamendwithfluorescentCy5.Togeneratestalledtranscriptioncomplexes,weincubateaDNAtemplatewithfluorescentCy3-5′-labeledApUdinucleotide,unlabeledATP,CTP,andEscherichiacoliRNAPwithσ70inatranscriptionstallingbuffer.TranscriptioninitiatespreferentiallywithCy3-ApUbutstallswithashortRNA(5′-Cy3-AUACC-3′)duetomissingGTP,soactivetranscriptioncomplexesarelabeledwithtwofluorescentdyes,Cy3onRNAandCy5onDNA(Fig. 1a).Fig.1:Single-moleculeassayofρ-dependenttermination.aTranscriptionofafluorescentCy5-labeledDNAtemplatewithaρ-dependentmgtAterminatorproducesRNAtranscriptscontainingarutsiteandlabeledwithfluorescentCy3atthe5′end.bSchemeofsingle-moleculeholisticassayoftranscriptiontermination.TheoccurrenceofCy3andCy5PIFEsareindicatedbyenlargedgreenandmagentacircles,respectively.cThreedistinctpatternsoffluorescencechangesinactivetranscription.Readthrough(left,n = 251complexes),recyclingtermination(center,n = 51),anddecomposingtermination(right,n = 289)aredistinguishedbytheircharacteristicpatternsinfluorescencetimetracesofCy3(green)andCy5(red)atCy3excitation(top)andCy5excitation(bottom).NTP + ρwereinjectedat30 s(graylines)afterfluorescencemonitoringstarts.Aschematicdiagramisshownbeloweachrepresentativetrace,andadditionaldetailsaredescribedinSupplementaryFig. 3.dTerminationefficiencies(TEs)andbackgroundtermination.RawTEsareestimatedastheratioofdecomposingorrecyclingeventstothesumofallterminationandreadthroughevents.Negativecontrolswereperformedwithoutρorwithρplusρ-inhibitorbicyclomycin,andthreeρmutantsexhibitonlybackgroundlevelsofterminationactivity.eρ-dependencyoftermination.Theρ-dependentTEsofrecycling(solid)anddecomposing(open)alongwiththeirsums(total)inthey-axisareplottedagainstinputρconcentrationsonabase-10logscaleinthex-axis.ρ-dependentTEsarecalculatedbysubtractionoftheρ-independentbackgroundTE(6.3%)fromrawTEs.fRelativefrequenciesofreadthrough,recyclingtermination,anddecomposingterminationtimings.TerminationwastimedasthedelayfromCy3PIFEdiminishingtoCy3vanishing,whereasreadthroughwastimedasthedelayfromCy3PIFEdiminishingtoCy5PIFEstarting.Thesetimingswereestimatedwiththedatafittingtosingleexponentialfunctions.Errorbarrepresentsthestandarddeviationofthemeanfromn ≥ 3independentexperiments.ThenumbersofanalyzedmoleculesfordandeareinSupplementaryTable 2.SourceDatafileincludesthedatafor(d–f).FullsizeimageThefluorescentinitiallystalledcomplexes(ISCs)areimmobilizedonpolymer-coatedquartzslidesusingbiotin–streptavidinconjugationandextensivelywashedtoremoveallunimmobilizedcomplexes.Insingle-moleculeterminationassays(Fig. 1b),theimmobilizedISCsareelongatedbytheadditionofNTPsandρ(100 nMunlessvarying),whilethefluorescenceofCy3-RNAandCy5-DNAinindividualtranscriptioncomplexesismonitoredinreal-timeusingtotal-internal-reflectionfluorescencemicroscopy.VanishingofCy3spotsafterNTP + ρinjectionindicatesreleaseoftranscriptRNAoutofimmobilizedcomplexesresultingfromtranscriptionterminationatthetermination-associatedpausingsite(terminationsite)orfromrunofftranscriptionattheDNAdownstreamendfollowingtranscriptionreadthroughattheterminationsite(SupplementaryFig. 1).ThisCy3vanishingisdistinguishedfromphotobleaching,whichtakesmuchlongerthanterminationorrunoff(SupplementaryFig. 2)andlittleaffectsthemeasurementsofterminationandrunoff.Co-occurringfastandslowρ-dependentterminationsAnothernoticeablefluorescencechangewasprotein-inducedfluorescenceenhancement(PIFE).PIFEoccurstocyanine(Cy)dyeswhentheirlocalenvironmentbecomesmoreviscousbybindingproteinstohinderthedye’sphotoisomerizationfromfluorescenttrans-isoformtonon-fluorescentcis-isoform20,21,22,asithasbeenseenwithmanyDNA-bindingandmotorproteins23,24.Inthisstudy,Cy3PIFEoccurswhentheCy3labeledatthe5′endofRNAremainsproximaltoRNAPinISCs,andthePIFEdecreaseswhenthe5′endmovesawayfromRNAPastheRNAgrowsinlengthuponresumptionoftranscriptionelongationfromISCs.IndividualcomplexesshowingthispatternofCy3fluorescencechangeshavegonethroughactivetranscriptionandthereforearesolelyincludedinourdataanalysis.Bycontrast,Cy5PIFEariseswhentheCy5labeledatthedownstreamendofDNAisapproachedbyRNAP.EithertranscribingRNAPwithRNAafterreadthroughattheterminationsiteorrecyclingRNAPwithoutRNAafterterminationattheterminationsitecanreachthedownstreamendtocausethePIFE.Ontheotherhand,onceRNAPfallsoffDNA,itreadilydiffusesawayandseldomrebindsDNAevenattheendsinthissingle-moleculeassay.Real-timemonitoringofCy3-RNAandCy5-DNAsignalsrevealedthreedistinctpatternsoftheirfluorescencechanges(Fig. 1c,SupplementaryFig. 3).ThefirstpatternarisesfromthereadthroughfollowedbyrunofffromDNA,asCy5PIFEstartsbeforetheCy3signalvanishes(Fig. 1cleft,SupplementaryFig. 3a).TranscribingRNAPignorestheterminationsignal,passesbeyondtheterminationsite,reachestheDNAdownstreamendtobringaboutCy5PIFE,andfinallydissociatesfromtheDNAend,releasingaCy3-labeledreadthrough-productrunoff-transcriptRNAtomaketheCy3signalvanish.ThesecondpatternrepresentstheterminationresultinginRNAPrecyclingonDNA(calledrecyclingtermination),asCy5PIFEstartsassoonasCy3signalvanishes(Fig. 1ccenter,SupplementaryFig. 3b).OnlyRNAisreleasedattheterminationsite,vanishingtheCy3signal,andthenpost-terminationalrecyclingRNAPdiffusesonDNAtoreachthedownstreamendandsticksthereforawhile,generatingCy5PIFE.ThethirdpatternreflectstheterminationresultinginECdecomposing(calleddecomposingtermination),asCy3signaldisappearsbutCy5PIFEneverarises(Fig. 1cright,SupplementaryFig. 3c).RNAPreleasesterminatedRNAtovanishtheCy3signalbutdoesnotreachtheDNAdownstreamendtoproduceCy5PIFE.Inthisevent,thethreeessentialcomponentsofthetranscriptioncomplexfallaparttogetherfordecomposing.Alternatively,RNA-freeRNAPonlybrieflyremainsonDNAuponterminationandfallsoffDNAtodiffuseawaybeforereachingthedownstreamend.The1DreinitiationcouldoccurbyrecyclingRNAPafterterminationandsubsequentbackwarddiffusiontothepromoteronthesameDNAmolecule15,16,17.Itcanbefollowedbythesecond-roundtranscriptionforterminationorreadthroughandsubsequentsuchevents,butthesecycliceventsarenotmonitoredinthisassayasthereinitiatedtranscriptsarenotlabeledwithCy3aftertheISC-washingstep.Thisbasicsetupofsingle-moleculeterminationassayiscalledholisticassaytobedistinguishedfromthetwobelow-describedassaysthateachmeasureonlyacertaindiscretemodeoftermination.Inanegativecontrolexperimentwithoutρ,terminationefficiency(TE)dropstoabackgroundlevel,6.3 ± 0.7%(Fig. 1d).Thisρ-independentTEissubtractedfromtherawTEstoyieldρ-dependentTEsinallthefollowingestimations.Additionally,withρplusρ-inhibitorbicyclomycin,TEdropstoasimilarbackgroundlevel(Fig. 1d).Ineachassay,everytranscriptioncomplexiscountedasreadthrough,recyclingtermination,ordecomposingtermination.Theirrelativefrequenciesare42.6 ± 3.4%,8.6 ± 2.0%,and48.8 ± 3.9%,respectivelyasmeasuredinthreeindependentexperimentswith10 mMMg2+.Thus,rawTEis57.4 ± 3.4%(sumoftherecyclinganddecomposingterminations,Fig. 1d)orρ-dependentTEis51%,whichiscomparabletoapreviousmeasurement18at3.5 mMMg2+.Interestingly,decomposingoutcomedominatesoverrecyclingoutcomefromρ-dependentterminationandonlydecomposingisobservedinbackgroundterminations(Fig. 1d),incontrasttointrinsictermination,whererecyclingoutcomeispredominant15,16,17.Intitrationexperiments,decomposingandrecyclingTEsbothsteeplyincreaseonrisingρconcentration(Fig. 1e)demonstratingtheρdependence,buttheratioofdecomposingandrecyclingTEslittlechanges(SupplementaryFig. 4)implyingthatthetwoarebothintrinsicoutcomesofρ-dependenttermination.Interestingly,recyclingterminationoccursmuchearlier(27 s)thandecomposingtermination(199 s)orreadthrough(198 s)asshowninFig. 1f.TerminationwastimedastheintervalfromtheCy3PIFEdiminishingtimepointtotheCy3signalvanishingtimepoint,whereasreadthroughwastimedasthedelayfromCy3PIFEdiminishingtoCy5PIFEstarting.Becausereadthroughismuchmoresynchronouswithdecomposingterminationthanrecyclingtermination,recyclingappearstobedecidedoverreadthroughmuchearlierthandecomposing.Co-presentcatch-upandstand-bypre-terminationalmodesTheholisticassaydoesnotdifferentiatethecatch-upandstand-bypre-terminationalpathways.Inordertosolelymonitortheterminationthatismediatedbyρinthestand-bymode(calledstand-bytermination)ratherthanbyρinthecatch-upmode(calledcatch-uptermination),wepre-incubatedISCswithρandwashedawaytheunboundbeforeresumingtranscriptionwithNTPsbutwithoutaddingρ(Fig. 2a).AlthoughitisnotknownwhatfractionofISCsisboundtoρundertheconditionsofthisstand-bysingle-moleculeassay,onlypre-boundρinthestand-bymodecanmediateterminationbecauseunboundρforcatch-upmodeiswashedaway.TEincreasesasmoreρisinputinthepre-incubation(Fig. 2b)demonstratingthestand-byρ-dependenttermination9.Fig.2:Stand-byρmediatingonlydecomposingbutnotrecyclingtermination.aSchemeforthestand-bysingle-moleculeassay.Terminationismediatedonlybythestand-byρpre-boundonRNAPintheabsenceofcatch-upρ.bρ-dependencyofstand-bydecomposingtermination.Theρ-dependentTEsofdecomposing(open)andrecycling(solid)inthey-axisareplottedagainstρconcentrationsonabase-10logscaleinthex-axis.Recyclingisvirtuallynullovertheentirerangeofρconcentrationsused,sothedecomposingdataoverlapwiththeirsumdata(notshown).cRelativefrequenciesofreadthrough(n = 326complexes)anddecomposing(n = 93)timingsmeasuredinstand-byassays.Thedataarefittedtosingleexponentialfunctions.Errorbarrepresentsthestandarddeviationofthemeanfromn ≥ 3independentexperiments.Thenumbersofanalyzedmoleculesfor(b)areinSupplementaryTable 2.SourceDatafileincludesthedatafor(b,c).FullsizeimageInterestingly,almostallstand-byterminationeventsarefollowedbydecomposingoutcome(SupplementaryFig. 5a)andveryfewbyrecyclingoutcome(Fig. 2b).Becauserecyclingisvirtuallynullthroughoutawiderangeofρconcentration,thedecomposingdataoverlapwiththeirsumdata(notshown)inFig. 2b.TheseresultsindicatethatRNAPdissociatesoffDNAtogetherwithRNAreleaseatstand-bytermination,andfewremainonDNAafterstand-bytermination.Withstand-byρ,decomposingterminationtakesplacemuchlater(485 s)thanreadthrough(207 s)asshowninFig. 2c.AsthetimeintervalbetweenISCwashingandNTPinjectionprolongs,thestand-byTEgraduallydecreaseswithatimeconstantof693 s(SupplementaryFig. 6),indicatingthattheactivitylifetimeofρ·RNAPcomplexisfinitebutlongenoughtorevealstand-bytermination.Next,inordertosolelymeasurethecatch-uptermination,wepre-incubatedISCswithinactiveρP297Smutant25,washedawaytheunbound,andresumedtranscriptionbyaddingNTPsandwild-typeρ(Fig. 3a).Inthiscatch-upsingle-moleculeassay,thestand-bysiteonRNAPshouldbestablypreoccupiedandmaskedbyafullyinactivemutantsothatonlycatch-upwild-typeρcanmediatetermination.Fig.3:Catch-upρmediatingbothdecomposingandrecyclingterminations.aSchemeforthecatch-upsingle-moleculeassay.Terminationismediatedonlybycatch-upρ,whileaninactiveρpre-occupiesthestand-bysiteonRNAPandisnotreplacedbysubsequentlyaddedwild-typeρ.bρ-dependencyofcatch-uprecyclinganddecomposingterminations.Theρ-dependentTEsofrecycling(solid)anddecomposing(open)inthey-axisareplottedagainstρconcentrationsonabase-10logscaleinthex-axis.cRelativefrequenciesofreadthrough(n = 251complexes),recycling(n = 14),anddecomposing(n = 70)timingsmeasuredincatch-upassays.Thedataarefittedtosingleexponentialfunctions.Errorbarrepresentsthestandarddeviationofthemeanfromn ≥ 3independentexperiments.ThenumbersofanalyzedmoleculesforbareinSupplementaryTable 2.SourceDatafileincludesthedatafor(b,c).FullsizeimageForthispurpose,threepreviouslycharacterizedρ-mutants25,ρP279S,ρG51V,andρY80C,wereconfirmedtobecompletelyinactiveexhibitingonlybackgroundtermination(Fig. 1d)andtestedfortheirmaskingactivityinamodifiedstand-byassay.Wepre-incubatedISCswithaninactiveρmutant,washedawaytheunbound,addedwild-typeρ,washedawaytheunboundagain,andresumedtranscriptionwithNTPsalone(SupplementaryFig. 7).WhenρP297Sispre-bound,itishardlyreplacedbysubsequentlyaddedwild-typeρ,exhibitingfullmasking,whereasρG51VandρY80Cmutantsachieveonlypartialmasking.Whilethecatch-upassay’sexperimentalsetupwithpre-boundρP297Sisvalidated,bothdecomposingandrecyclingterminationsareobservedwithcatch-upρinoursingle-moleculeassay(Fig. 3b,SupplementaryFig. 5b)inwelcomecontrasttoapreviousreportthatpre-bindingofthemutantρcompletelyblockstheterminationbysubsequentwild-typeρinabulkassay9.Withcatch-upρ,recyclingtermination(21 s)happensmuchearlierthandecomposingtermination(189 s)orreadthrough(177 s)asshowninFig. 3c,whichisconsistentwiththeresultsofholisticassays(Fig. 1f).ThegeneralityofterminationheterogeneitywithfivedifferentterminatorsInordertotestthegeneralityofthemolecularheterogeneityofE.coliρ-dependentterminationobservedwithmgtAterminator,weexaminedfouradditionalρ-dependentterminators,rho,ribB,trpt’,andλtR1(SupplementaryTable 1).Theirρ-dependentterminationproficienciesareconfirmedinthisstudybyperforminginvitrobulktranscriptionassays(SupplementaryFig. 8),allconsistentwiththepreviousreports7,9,12,18,19,26,27,28,29,30.Theρ-dependentTEsofthefiveterminatorsmeasuredbythesingle-moleculeholisticassaysrangefrom12to51%asbothcatch-upandstand-byTEsvaryonterminators(Fig. 4a,SupplementaryFig. 9).TheseresultsareconsistentwiththatρineithermodeinteractsspecificallywithRNArutsequenceaswellaswithRNAP.Furthermore,bothcatch-up(cyaninFig. 4a)andstand-by(red)pre-terminationalmodesareco-presentintheoperationofeveryterminatorexceptλtR1,inwhichstand-bymodeisnegligible.Accordingly,thepre-terminationalmechanismisgenerallyheterogeneousintheoperationofasingleterminator.Fig.4:Generalityofρ-dependentterminationrouteheterogeneity.Recycling(solid)anddecomposing(open)terminationswerecountedandtimedwitheachofthefiveterminatorsinholistic(black),stand-by(red),andcatch-up(cyan)assays.aρ-dependentTEsmeasuredinthethreedifferentassays.Theρ-dependentTEsincatch-upandstand-byassaysaresummedintherightmostbarsforcomparisonwiththeholisticmeasurementsintheleftmostbars.bThesameasinabutformgtAterminatorwithNusA/G.cRelativefrequenciesofthethreeterminationroutes.Witheachterminator,thefrequencyofeachterminationrouteintotalterminationeventswasestimatedusingthedatain(a)and(b).Thecatch-up→decomposingrouteisthemostfrequentwitheveryterminator.dRelativetimingsofreadthroughandtermination.Foreachtranscriptionroutewithρ,aboxplotisdrawnwithreadthroughorterminationtimingsobservedinallterminatorsin(a)and(b)(n = 6),theirmedian(centerline),andthefirstandthirdquartiles(boundaries).Whenthetimingdistributionsarecomparedamongthefivetranscriptionroutes,thecatch-upρ’srecyclingterminationistheearliest.eComparisonbetweenterminationandreadthroughtimings.Theaverageterminationtimingsofallterminatorsinthey-axisareplottedagainsttheaveragereadthroughtimingsinthex-axis.ThedataofmgtAwithNusA/Garecoloredgreen.Allopensquares(decomposingtermination)arelocatedabovethediagonaleye-guidelinewithaunitslope(dotted),andallsolidsquares(recyclingtermination)belowtheguideline.fComparisonofdecomposingterminationtimingsbetweencatch-upandstand-bymodes.Theaveragetimingsofdecomposingterminationbycatch-upρinthey-axisareplottedagainstthosebystand-byρinthex-axis.Allopensquaresarelocatedbelowtheguideline,exceptforrhoterminatorslightlyabovetheline.Errorbarrepresentsthestandarddeviationofthe mean fromn ≥ 3independentexperiments.Thenumbersofanalyzedmoleculesfor(a)and(b)areinSupplementaryTable 2.Valuesandstatisticsoftimingsford–fareinSupplementaryTables 3and4.SourceDatafileincludesthedatafor(a–f).FullsizeimageFromallthefiveterminators,theonlydecomposingoutcomeisobservedinstand-byassays,whilebothdecomposingandrecyclingoutcomesaredetectedinbothcatch-upandholisticassays(Fig. 4a).Ontheotherhand,sumsofcatch-upandstand-byTEsarenotalwaysthesameastheTEmeasurementsinholisticassays:thesumissmallerthantheholisticmeasurementformgtAandlargerforrhoandribB,andtheyaresimilartoeachotherfortrpt’andλtR1.E.coliNusAandNusGfactors(NusA/G)areknowntoparticipateinstableECalongwithρ7,12.Inordertoexaminehowtheyinfluencetheterminationroutes,weaddedthemtoISCsinallthreeassayswithmgtAterminator.Expectedly,theactivitylifetimeofstand-byρislongerwithNusA/Gthanwithoutthem(SupplementaryFig. 6),andthebackgroundterminationlittlechanges.WithNusA/G,thecatch-upandstand-bymodesstillcoexistandtheterminationroutecompositionlittlechanges(Fig. 4b).AlltheresultswithandwithoutNusA/Gindicatethattranscriptionwithρproceedsinfivedifferentroutes;(1)catch-up→readthrough→runoff,(2)catch-up→termination→decomposing,(3)catch-up→termination→recycling,(4)stand-by→readthrough→runoff,and(5)stand-by→termination→decomposing.Whilethestand-by→termination→recyclingpathisnotoperatinginanyterminator,theoperationofasingleρ-dependentterminatorrunsviathethreeroutesstartingwithcatch-uporstand-byρ.Whenportionsofthethreeterminationroutesinallterminationeventsfromeachterminatorwereseparatelycalculated(Fig. 4c),thecatch-up→decomposingroute(blankbluebox)isthemostfrequentterminationwithallterminators,whilerelativefrequenciesoftheothertwoterminationsdependonterminators.Thestand-by→decomposingroute(blankredbox)ismorefrequentthanthecatch-up→recyclingroute(solidbluebox)inthreeterminatorsbutlessintheothertwo.Next,themeasuredtimingsofrecyclingtermination,decomposingtermination,andreadthrougheventswerecomparedamongeachother,whilethefiveterminatorswithoutNusA/GandmgtAterminatorwithNusA/Gweregroupedtogetherinaboxplotwithmedianandquartilevalues(Fig. 4d).Accordingtostudentt-tests,thecatch-uprecyclingtermination(route3)istheearliestandespeciallyearlierthanthecatch-upreadthrough(route1)(P = 0.001).Decomposingterminationisobservedlaterthanreadthroughbycatch-up(P = 0.03)orstand-byρ(P = 0.01).Thesechronologicalordersareinfactmoreevidentwhencomparisonsaremadedirectlybetweenterminationandreadthroughtimingsforeachindividualterminator(Fig. 4e).Inthegraphsofholistic(left),stand-by(center),andcatch-up(right)assays,allopensquaresarelocatedabovethediagonalguideline,indicatingthatdecomposingterminationisalwayslaterthanreadthroughinallterminatorswithcatch-uporstand-byρ.Instarkcontrast,allsolidsquaresarelocatedbelowtheguidelineinthecatch-upandholisticassaygraphs,indicatingthatrecyclingterminationisalwaysearlierthanreadthroughwithcatch-upρineveryterminator.Furthermore,decomposingresultslaterwithstand-byρthancatch-upρoralmostatthesametimeinallterminators,whenthedecomposingterminationtimingsbythetwoρmodesaredirectlycomparedfromeachotherinindividualterminatorsasshowninFig. 4f,whileitisnotrevealedinbox-plotcomparisonsofFig. 4d.Theseresultssuggestthatstand-byρtendstotakemoretimeorenergytodecomposeECthancatch-upρ.Itisinterestingthatstand-byρwouldencounterRNAPearlierthancatch-upρbuttendtoachieveterminationnotsomuchearlierthancatch-upρ.RNAshearingandRNAPdisplacingforterminationalreleaseofRNAFinally,weexaminedwhetherRNAshearingorRNAPhyper-translocationisutilizedinρ-dependenttermination.IntheRNAshearingmodel14,ρpullsRNAtoresolveRNA·DNAhybridforterminationalreleaseofRNAandimmediateordelayeddissociationofRNAPfromDNA.IntheRNAPhyper-translocationmodel14,bycontrast,ρpushesRNAPforwardonDNAforthecollapseoftranscriptionbubbleleadingtoRNAreleasewithsimultaneousorsubsequentRNAPdissociation.FortestingtheRNAshearingmodel,wedesignedaDNAscaffoldwiththemgtAterminatormutantAU100whereAUcontentoftheRNA·DNAhybridformedattheterminationsiteisincreasedto100%fromthewild-type44%(Fig. 5a,SupplementaryTable 1).Thus,AU100wouldformaweakerhybridthanthewild-typewhilepausingattheterminationsite.TheweakerhybridinAU100wouldfacilitatetheterminationbyRNAshearing,whileitwouldnotbringmuchenergeticadvantagetoRNAPhyper-translocation.InthistestusedwasmgtAterminatorexhibitingasinglemajorterminationsite18,19(SupplementaryFig. 8).Fig.5:RNAshearingresultingonlyinRNAPrecycling onDNA.ForRNAshearingtests,comparisonsweremadeamongsixtemplatesproducingRNA·DNAhybridswithvaryingAUcontents.ρ-dependentTEs(=rawTE—backgroundTE)ofrecycling(solid)anddecomposing(open)weremeasuredinholistic(black),stand-by(red),andcatch-up(cyan)assaysandnormalizedtothemaximumof100%bydividingwith(1—backgroundTE).aPutativeRNA·DNAhybridsformedattheterminationsite.TheAUcontentis44%inthewild-typemgtAand100%inAU100mutant.bAugmentofrecyclingterminationbyhybridweakeninginAU100.Normalizedρ-dependentTEsarecomparedbetweenthewild-type(W)andAU100(M).cPositivecorrelationofrecyclingterminationwithhybridinstability.NormalizedrecyclingTEsinthey-axisareplottedagainsttheAUcontentinthex-axis.dNocorrelationofdecomposingTEwithhybridinstability.NormalizeddecomposingTEsinthey-axisareplottedagainsttheAUcontentinthex-axis.Errorbarrepresentsthestandarddeviationofthemeanfromn ≥ 3independentexperiments.Thenumbersofanalyzedmoleculesfor(c)and(d)areinSupplementaryTable 2.SourceDatafileincludesthedatafor(c)and(d).FullsizeimageRecyclingTEisroughlydoubledwithAU100inFig. 5b,asρ-dependentTEs(=rawTE—backgroundTE)arenormalizedtothemaximumof100%bydividingthemwith(100%—backgroundTE),whenblacksolidpartsarecomparedbetweenthewild-typeandAU100inholisticassaysorcyansolidpartsincatch-upassays.Includingfourmoreterminatormutants(SupplementaryTable 1),atotalofsixtemplateswithvaryingAUcontentstogetherexhibitthatrecyclingTEincreaseswithrisingAUcontent(Fig. 5c),whiledecomposingTEislittlecorrelatedwithit(Fig. 5d).TheseresultssupportthatRNAshearingresultsinrecyclingonDNAmuchmorelikelythandecomposingoutcome,i.e.,RNAreleaseisfollowedbymuch-delayeddissociationofRNAPfromDNA.TheuncorrelatedvariationsinFig. 5daresolargethatsomemechanism(s)differentfromRNAshearingcanbepresumedfordecomposingtermination,sothehyper-translocationmodelwastested(Fig. 6a)aspreviouslydemonstratedusingmodifiedDNAscaffoldswhere3-bpmismatch-causingmutationsareintroducedinthenontemplatestrandoftemplateDNAatvaryingpositions14.Ifamismatchoverlapswithatranscriptionbubbleregionthatisnotbase-pairedbeforeterminationbutbecomesbase-pairedafterhyper-translocationfortermination,themismatchwouldinhibitrewindingtodiminishhyper-translocationproficiencybutlittleaffectRNAshearing,whichdoesnotneedrewindingoftheregion14,31.Fig.6:RNAPdisplacingresultingonlyinECdecomposing.ForRNAPdisplacingtests,comparisonsweremadeamong15DNAtemplateswitha3-bpmismatch.ρ-dependentTEs(=rawTE—backgroundTE)ofrecycling(solid)anddecomposing(open)weremeasuredinstand-by(red)andcatch-up(cyan)assaysandnormalizedtothemaximumof100%bydividingwith(1—backgroundTE).aPutativeelongation-coupledprogressionoftranscriptionbubbleinthemgtAwild-typetemplateandamutantwitha3-bpmismatch(blue).bApeakedeffectofthe3-bpmismatchondecomposingtermination.NormalizeddecomposingTEsofthe15mutantsinthey-axisareplottedagainstthecenterpositionofthe3-bpmismatchinthex-axis.Horizontallinesindicatethevaluesofthewild-typetemplate(16.7%forstand-byassayand24.9%forcatch-upassay).cNoeffectofthemismatchonrecyclingtermination.NormalizedrecyclingTEsinthey-axisareplottedagainstthemismatchpositioninthex-axis.Errorbarrepresentsthestandarddeviationofthemeanfromn ≥ 3independentexperiments.Thenumbersofanalyzedmoleculesfor(b)and(c)areinSupplementaryTable 2.SourceDatafileincludesthedatafor(b)and(c).FullsizeimageUsingaseriesof15templatemutantswitha3-bpmismatchatvaryingpositions(SupplementaryTable 1),weobservedthatthemismatchesatcertainpositionsupstreamoftheterminationsite(denotedasthe−1position)substantiallydecreasedecomposingtermination(Fig. 6b)butnotrecyclingtermination(Fig. 6c).Unexpectedly,however,themismatcheffectpeaksatthe−12position,whichisjustupstreamofthetranscriptionbubbleratherthanwithinitaccordingtotheECstructuresthatwerereportedforρ-dependentterminators7,12,32,33,34,35,36,37althoughnotyetformgtAterminator.Speculatively,thetranscriptionbubblepositioncouldbedifferentinmgtAterminator,orRNAPbacktrackingcouldbeinvolvedinρ-dependenttermination,suchthatthe−13to−11positionsarewhollyorpartlylocatedwithinthebubblejustbeforetermination.Alternatively,themismatcheffectsmaynotberelatedtohyper-translocationbutwithsomethingelse.Forexample,aDNAregionaroundthe−12positioninteractswithRNAPintherecentX-raycrystalstructures7,16,andsuchinteractionmaybehypotheticallyimportantfortermination.Ourresults,however,atleastsuggestthatduplexformationbytemplaterewindingaroundthe−12positionratherthanothersiscriticalfordecomposingterminationregardlessofwhetherthehyper-translocationisinvolvedornotinρ-dependenttermination,soitisdescribedherelooselyasRNAPdisplacinginsteadofhyper-translocation.TheseresultstogethersupportthatRNAPdisplacingresultsindecomposingoutcomesmuchmorereadilythanrecyclingonDNA,i.e.,RNAreleaseisaccompaniedbyfairlysimultaneousdissociationofRNAPfromDNA.DiscussionMechanismsofρ-dependentterminationhavelongbeenstudiedbutstillremainelusive.AprevailingdebatehasbeenconcernedwithwhetherρfirstbindsRNAandcatchesupwithRNAPlater13,38(catch-upmode)orρfirstbindsRNAPandstandsbyforRNAemergence9(stand-bymode).Contrarytothepreviousreportsthatonlyonepre-terminationalmodeispresent,wediscoverthatbothcatch-upandstand-bymodesareusedbyasingleterminator.Theirproportions,however,varyindifferentterminators,raisinginterestingquestionsofhowtheirproportionsaredeterminedandinfluencegeneregulation.Weadditionallydiscoverthatdecomposingoutcomeismuchmorefrequentthanrecyclingfromeachofthefivetestedρ-dependentterminators,asopposedtothatrecyclingoutcomeismuchmorefrequentthandecomposingfromintrinsicterminators15,16,17.Frequentlywithcatch-uporstand-byρ,RNAreleaseandRNAPdissociationfromDNAoccurtogether,asRNAPdisplacingresultsonlyinECdecomposing.Bycontrast,lessfrequentlywithcatch-upρ,onlyRNAisreleasedbutRNAPisnotconcomitantlydissociatedfromDNAwhenRNAshearingresultsonlyinRNAPrecyclingonDNA,whilethisismuchmorefrequentinintrinsicterminationwithoutρ.Ourkineticanalysesadditionallyrevealhowtheseco-occurringmechanismsarechronologicallyarrangedinthecourseofρ-dependenttermination(Fig. 7).Thechoiceforrecyclingterminationisalwaysmadeearlierthandecomposingterminationinalltestedterminators.RecyclingterminationresultsfromRNAshearingandobservedonlyinourcatch-upsingle-moleculeassaysbutnotinstand-byassays.Therefore,wesuspectthatcatch-upρmovingonRNAcanshearoffRNAmuchmorehandilythanstand-byρresidingonRNAP,andpullingRNAalonefromECmayneedlessforcethanpushingRNAPorECtobedisplaced.Fig.7:Thethree-routemodelofρ-dependenttermination.Threekineticallydifferentroutesoftranscriptionterminationareoperatingwithasingleterminatorinourmodelforρ-dependenttermination,whereasallpreviousmodelsassumeasinglepath.(1)Catch-uprecyclingtermination. Thefastestterminationrouteistakenbycatch-upρ,whichbindsnascentRNAatarutsite,translocatesdownonRNA,andcatchesupwithRNAPthatispausingataterminationsite.OnlyRNA,notRNAP,isreleased,andpost-terminationalRNAPrecyclesonDNAfor1Dreinitiation(likeinmostintrinsictermination15,16,17,whichisρ-independentbutRNAhairpin-dependent).(2) Catch-updecomposingtermination.Ifthefirst,fastestterminationfails,thesecondterminationroutecanbefollowedbycatch-upρ.TranscriptioncomplexdecomposesinastepasbothRNAandRNAPdepartDNA.Thissecond-routeterminationbycatch-upρfordecomposingalwaysoccursmuchlaterthanthefirst-routeterminationforrecycling.(3)Stand-bydecomposingtermination. Theslowestterminationrouteisusuallypursuedbystand-byρ,whichpre-bindsRNAP,movesalongDNAwithRNAP,andstandsbyforbindingarutsiteemergingoutofRNAP.Thisterminationresultsinone-stepdisassemblyofthetranscriptioncomplex.Interestingly,stand-byρpresumablybindsRNAPearlierthancatch-upρbuttendstoachieveterminationnotsomuchearlier.Whilethecatch-updecomposingtermination route(2)isthemostfrequentinallterminators,relativefrequenciesoftheotherroutes(1)and(3)dependonterminators.FullsizeimageTranscriptionalreadthroughnaturallysynchronizeswiththelatertermination,whichisdecomposingterminationmuchmorelikelythanrecyclingtermination.However,decomposingterminationcompletesevenlaterthanreadthrough,regardlessofwhetheritismediatedbycatch-uporstand-byρ(Fig. 4e).ItispossiblybecauseECdecomposingrequiresmoremassiveorslowerconformationalchangesthanreadthrough,orlesslikelyRNAtemporarilyremainsonDNAafterRNAPdisplacing.Asimplespeculativecourseofρ-dependenttermination(Fig. 7)couldstartwiththecatch-upρshearsoffRNAforrecyclingterminationusinglessenergy.Ifitfails,thecatch-upρ·ECcomplexundergoesconformationalchangesusingmoreaccumulatedenergyfordecomposingtermination.Lastly,stand-byρcanadditionallymediatethesamedecomposingterminationoftenmoreslowlythancatch-upρ.Interestingly,stand-byρcouldbindRNAPearlierthancatch-upρbutdoesnotachievethesameterminationmorereadily.Thecatch-upandstand-bypre-terminationalmodeswouldfollowseparatepathsondifferenttimescalesandcouldbuilddisparatestructuresoftermination-proficientcomplexfavoringspecificoutcomessuchasRNAPrecyclingonDNAandECdecomposingatonce.Wespeculatethatevenforthesamedecomposingtermination,thecatch-upandstand-byρ′scouldinducedissimilarchangesinaterminationcomplexstructurebecausetheirterminationtimingsoftendiffermuchfromeachother(Fig. 4f).Most,ifnotall,previousmodelsofρ-dependentterminationsimulatewithasingleρmolecule.However,itshouldnotberuledoutthattwoρmoleculesparticipatesidebysideinterminationandcompeteorcollaboratewitheachother.Becausethesumsofstand-byandcatch-upTEsarelargerorsmallerthantheholisticassaymeasurements(Fig. 4a,b),onecanspeculateapossibilitythattheinteractionbetweenthetwoρmoleculesindifferentmodesissynergistic,antagonistic,ornullinaterminator-dependentfashion.Furthermore,thestand-byρactinglatercouldofferastandbybackupincasetheearlier-actingcatch-upρfailstomediateterminationinthesamecomplex.Incidentally,onecanconsideritoddthatthebackgroundterminationoccurringwithoutρresultssolelyinthedecomposingoutcome,whileρ-independent,intrinsicterminationresultsmostlyinrecycling15,16,17.However,thedecomposingoutcomeisnotabsentinintrinsicterminationrequiringRNAhairpinformation,soonecanhypothesizethatthebackgroundterminationresultsfrominfrequentdecomposingofECunstabilizedwithoutρorRNAhairpin,andthatshearingRNAaloneforrecyclingoccursactivelyratherthanpassivelyinbothintrinsicandρ-dependentterminations.Apparently,ρplaysthesameroleastheterminatorRNAhairpintomediaterecyclingterminationandadifferentroletomediatedecomposingtermination.Inthisstudy,weshowthatρfollowscatch-uporstand-bypre-terminationalsetting,executesRNA-shearingorRNAP-displacingterminationalrelease,andthenresultsinRNAP-recyclingorEC-decomposingpost-terminationaloutcome.Thesemechanismsinthethreesequentialstepsarerathertightlycoupledsothatonlythreeroutesareco-presentoutofeight(23)possiblestepwisecombinationsintheperformanceofasingleterminator.Thethreeroutesapparentlyoperateonindividualtimescalesandleadtotheirspecificoutcomes.Thisthree-routemodel(Fig. 7)couldprovideaplatformtoreviewthepriorstudiesonρ-dependentterminationandtoinvestigatefurthercompellingaspectsoftranscriptionregulation.MethodsSingle-moleculeexperimentsoftranscriptionterminationIndividualtranscriptioncomplexeswereimmobilizedtoquartzmicroscopeslidesviabiotin-streptavidinconjugationontheslidesurfacesthatwerecoatedwithpolyethyleneglycol(PEG)andbiotin-PEG5000(LaysanBio)in1:40ratio39andthenwithstreptavidin(0.2 g/l,Invitrogen)for5 min.Thecomplexeswerethenimagedunderahomemadewide-fieldtotal-internal-reflectionfluorescencemicroscopeequippedwith532-nmgreen(Excelsior-532-50-CDRH)and640-nmred(Excelsior-640c-35)lasers(Spectra-Physics)usinganelectron-multiplyingcharge-coupledcamera(iXonDU-897,AndorTechnology).Allexperimentswereperformedat37 °Cwith1-sexposuretimeinalternatinglaserexcitationmode40andtimeresolutionwas2 s.SoftwaresIDL7.0(ITT),MATLABR2018a(MathWorks),andOrigin8.5(OriginLab)wereusedfordataanalysis.ISCswerepreparedbyincubationofDNA,Cy3-labeledApU(250 μM,TriLinkBioTechnologies),ATP,CTP(10 μMeach,GEHealthcare),andRNAPholoenzyme(340 nM)inatranscriptionstallingbuffer(20 mMTris-HCl,pH8.0,20 mMMgCl2,20 mMNaCl,and1 mMdithiothreitol)for20 min.TranscriptionwasresumedwithallfourNTPs(200 μMeach)andρ(100 nMorvarying)inatranscriptionresumingbuffer(40 mMTris-HCl,pH8.0,10 mMMgCl2,150 mMKCl,1 mMdithiothreitol,5 mMprotocatechuateacid,100 mMprotocatechuate-3,4-dioxygenase,andsaturatedTrolox).Whenneeded,NusA(500 nM)plusNusG(500 nM)oraρmutant(100 nM)wasaddedtothestallingbuffer.PreparationofE.coliproteinsRNAPholoenzymewaseitherpurchasedfromNewEnglandBiolab(M0551S)orcustom-purifiedaspreviouslydescribed41,42.ρfactorwaspurchasedfromBioprogen.Threeρmutantswereexpressedandpurifiedaspreviouslydescribed25(SupplementaryFig. 10)usingtheirexpressionplasmidsthatwereprovidedasagiftbyRanjanSeninHyderabad,India.NusAwasexpressedwithanN-terminalHis6tagfromaclonedplasmidpNG5inE.coliBL21(DE3)andpurifiedbyusingHiTrapIMACHPaffinityandSuperdex200gelfiltration35.NusGwasexpressedwithaC-terminalHis6tagfromaclonedplasmidpRM1160inE.coliBL21(DE3)andpurifiedbyusingHiTrapIMAC,HiTrapQ,andHiLoadSuperdex75columns32.InvitrobulktranscriptionassayE.coliRNAPholoenzyme(30 nM)andatemplateDNA(15 nM)withoneofthefiveterminatorswereincubatedinatranscriptionbuffer(20 mMTris-HCl,pH8.0,10 mMMgCl2,100 mMKCl,and1 mMdithiothreitol)at37 °Cfor10 min.Theincubationwascontinuedwithρat0,15,30,or60 nMandNusA/Geachat0or50 nMfor10 min,andthenwith200 μMATP,200 μMGTP,200 μMUTP,0.025 μMCTP,and0.05 μM[ɑ-32P]CTPforanother10 min.Thetranscriptionreactionwasquenchedbytheadditionofa2×urea-denaturinggelloadingbufferandanalyzedona6%-polyacrylamideureagel.WhilerawTEistheratioofterminationbandintensitytothesumofterminationandreadthroughbandintensities,ρ-dependentTEsareestimatedbysubtractionofthebackgroundTEmeasuredwithoutρfromtherawTEsmeasuredwithρ.PreparationoftranscriptiontemplatesUsingDNAoligomers(SupplementaryTable 1)purchasedfromIntegratedDNATechnologies,nontemplateDNAstrandswerepreparedbyannealingtheirtwopartswithaDNAsplintbyslowlycoolingfrom90to16 °Cin50 mMTris-HCl,pH8.0,10 mMMgCl2,10 mMdithiothreitol,and1 mMATP,andbyligatingthemusingT4DNAligase2(NewEnglandBiolabs).Homo-duplexDNAsweremadebypolymerasechainreactions(PCRs)usingbiotin-labeledforwardandCy5-labeledbackwardprimers.Hetero-duplexDNAsweremadebyPCRsusingbiotin-labeledforwardandunlabeledbackwardprimersforwild-typeDNAs,andusingunlabeledforwardandCy5-labeledbackwardprimersformutantDNAs.Then,PCRproductswereannealed,andproductswithbothbiotinandCy5wereusedinexperiments.ReportingsummaryFurtherinformationonresearchdesignisavailableinthe NatureResearchReportingSummarylinkedtothisarticle. Dataavailability SourcedataareprovidedinaSourceDatafile.Allotherrelevantdataareavailablefromthecorrespondingauthorsuponrequest.AreportingsummaryforthisArticleisavailableasa SupplementaryInformationfile. Sourcedataareprovidedwiththispaper. ReferencesRoberts,J.W.TerminationfactorforRNAsynthesis.Nature224,1168–1174(1969).ADS  CAS  PubMed  GoogleScholar  Leela,J.K.,Syeda,A.H.,Anupama,K.&Gowrishankar,J.Rho-dependenttranscriptionterminationisessentialtopreventexcessivegenome-wideR-loopsinEscherichiacoli.Proc.NatlAcad.Sci.USA110,258–263(2013).ADS  CAS  PubMed  GoogleScholar  Jain,S.,Gupta,R.&Sen,R.Rho-dependenttranscriptionterminationinbacteriarecyclesRNApolymerasesstalledatDNAlesions.Nat.Commun.10,1207(2019).ADS  PubMed  PubMedCentral  GoogleScholar  Thomsen,N.D.,Lawson,M.R.,Witkowsky,L.B.,Qu,S.&Berger,J.M.Molecularmechanismsofsubstrate-controlledringdynamicsandsubsteppinginanucleicacid-dependenthexamericmotor.Proc.NatlAcad.Sci.USA113,E7691–E7700(2016).CAS  PubMed  PubMedCentral  GoogleScholar  Jin,D.J.,Burgess,R.R.,Richardson,J.P.&Gross,C.A.Terminationefficiencyatrho-dependentterminatorsdependsonkineticcouplingbetweenRNApolymeraseandrho.Proc.NatlAcad.Sci.USA89,1453–1457(1992).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Mitra,P.,Ghosh,G.,Hafeezunnisa,M.&Sen,R.Rhoprotein:rolesandmechanisms.Annu.Rev.Microbiol.71,687–709(2017).CAS  PubMed  GoogleScholar  Said,N.etal.Stepstowardtranslocation-independentRNApolymeraseinactivationbyterminatorATPaseρ.Science371,eabd1673(2021).CAS  PubMed  GoogleScholar  Steinmetz,E.J.,Brennan,C.A.&Platt,T.Ashortinterveningstructurecanblockrhofactorhelicaseactionatadistance.J.Biol.Chem.265,18408–18413(1990).CAS  PubMed  GoogleScholar  Epshtein,V.,Dutta,D.,Wade,J.&Nudler,E.AnallostericmechanismofRho-dependenttranscriptiontermination.Nature463,245–249(2010).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Darliz,J.L.,Sentenac,A.&Frogmageot,P.BindingofterminationfactorRhotoRNApolymeraseandDNA.FEBSLett.13,165–168(1971). GoogleScholar  Mooney,R.A.etal.Regulatortraffickingonbacterialtranscriptionunitsinvivo.Mol.Cell33,97–108(2009).CAS  PubMed  PubMedCentral  GoogleScholar  Hao,Z.etal.Pre-terminationtranscriptioncomplex:structureandfunction.Mol.Cell81,281–292(2021).CAS  PubMed  GoogleScholar  Kalyani,B.S.,Muteeb,G.,Qayyum,M.Z.&Sen,R.InteractionwiththenascentRNAisaprerequisitefortherecruitmentofRhotothetranscriptionelongationcomplexinvitro.J.Mol.Biol.413,548–560(2011).CAS  PubMed  GoogleScholar  Park,J.&Roberts,J.W.RoleofDNAbubblerewindinginenzymatictranscriptiontermination.Proc.NatlAcad.Sci.USA103,4870–4875(2006).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Kang,W.etal.TranscriptionreinitiationbyrecyclingRNApolymerasethatdiffusesonDNAafterreleasingterminatedRNA.Nat.Commun.11,450(2020).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Kang,W.,Hwang,S.,Kang,J.Y.,Kang,C.&Hohng,S.HoppingandflippingofRNApolymeraseonDNAduringrecyclingforreinitiationafterintrinsicterminationinbacterialtranscription.Int.J.Mol.Sci.22,2398(2021).CAS  PubMed  PubMedCentral  GoogleScholar  Harden,T.T.etal.AlternativetranscriptioncycleforbacterialRNApolymerase.Nat.Commun.11,448(2020).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Hollands,K.etal.RiboswitchcontrolofRho-dependenttranscriptiontermination.Proc.NatlAcad.Sci.USA109,5376–5381(2012).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Hollands,K.,Sevostiyanova,A.&Groisman,E.A.Unusuallylong-livedpauserequiredforregulationofaRho-dependenttranscriptionterminator.Proc.NatlAcad.Sci.USA111,E1999–E2007(2014).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Rashid,F.etal.InitialstateofDNA-dyecomplexsetsthestageforproteininducedfluorescencemodulation.Nat.Commun.10,2104(2019).ADS  PubMed  PubMedCentral  GoogleScholar  Hwang,H.&Myong,S.Proteininducedfluorescenceenhancement(PIFE)forprobingprotein-nucleicacidinteractions.Chem.Soc.Rev.43,1221–1229(2014).CAS  PubMed  PubMedCentral  GoogleScholar  Hwang,H.,Kim,H.&Myong,S.Proteininducedfluorescenceenhancementasasinglemoleculeassaywithshortdistancesensitivity.Proc.NatlAcad.Sci.USA108,7414–7418(2011).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Myong,S.etal.CytosolicviralsensorRIG-Iisa5′-triphosphate–dependenttranslocaseondouble-strandedRNA.Science323,1070–1074(2009).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  Markiewicz,R.P.,Vrits,K.B.,Rueda,D.&Romano,L.J.Single-moleculemicroscopyrevealsnewinsightsintonucleotideselectionbyDNApolymeraseI.NucleicAcidsRes.40,7975–7984(2012).CAS  PubMed  PubMedCentral  GoogleScholar  Chalissery,J.,Banerjee,S.,Bandey,I.&Sen,R.TranscriptionterminationdefectivemutantsofRho:roleofdifferentfunctionsofRhoinreleasingRNAfromtheelongationcomplex.J.Mol.Biol.371,855–872(2007).CAS  PubMed  PubMedCentral  GoogleScholar  Bastet,L.etal.TranslationalcontrolandRho-dependenttranscriptionterminationareintimatelylinkedinriboswitchregulation.NucleicAcidsRes.45,7474–7486(2017).CAS  PubMed  PubMedCentral  GoogleScholar  Silva,I.J.etal.SraLsRNAinteractionregulatestheterminatorbypreventingprematuretranscriptionterminationofrhomRNA.Proc.NatlAcad.Sci.USA116,3042–3051(2019).CAS  PubMed  PubMedCentral  GoogleScholar  Matsumoto,Y.,Shigesada,K.,Hirano,M.&Imai,M.AutogenousregulationofthegenefortranscriptionterminationfactorrhoinEscherichiacoli:localizationandfunctionofitsattenuators.J.Bacteriol.166,945–958(1986).CAS  PubMed  PubMedCentral  GoogleScholar  Barik,S.,Bhattacharya,P.&Das,A.AutogenousregulationoftranscriptionterminationfactorRho.J.Mol.Biol.182,495–508(1985).CAS  PubMed  GoogleScholar  Brown,S.,Albrechtsen,B.,Pedersen,S.&Klemm,P.Localizationandregulationofthestructuralgenefortranscription-terminationfactorrhoofEscherichiacoli.J.Mol.Biol.162,283–298(1982).CAS  PubMed  GoogleScholar  Santangelo,T.J.&Roberts,J.W.ForwardtranslocationisthenaturalpathwayofRNAreleaseatanintrinsicterminator.Mol.Cell14,117–126(2004).CAS  PubMed  GoogleScholar  Kang,J.Y.etal.StructuralbasisfortranscriptelongationcontrolbyNusGfamilyuniversalregulators.Cell173,1650–1662(2018).CAS  PubMed  PubMedCentral  GoogleScholar  Kang,J.Y.etal.StructuralbasisoftranscriptionarrestbycoliphageHK022nuninanEscherichiacoliRNApolymeraseelongationcomplex.eLife6,e25478(2017).PubMed  PubMedCentral  GoogleScholar  Kang,J.Y.etal.StructuralbasisfortranscriptioncomplexdisruptionbytheMfdtranslocase.eLife10,1–86(2021).CAS  GoogleScholar  Guo,X.etal.StructuralbasisforNusAstabilizedtranscriptionalpausing.Mol.Cell69,816–827.e4(2018).CAS  PubMed  PubMedCentral  GoogleScholar  Kang,J.Y.etal.RNApolymeraseaccommodatesapauseRNAhairpinbyglobalconformationalrearrangementsthatprolongpausing.Mol.Cell69,802–815(2018).CAS  PubMed  PubMedCentral  GoogleScholar  Krupp,F.etal.Structuralbasisfortheactionofanall-purposetranscriptionanti-terminationfactor.Mol.Cell74,143–157(2019).CAS  PubMed  GoogleScholar  Koslover,D.J.,Fazal,F.M.,Mooney,R.A.,Landick,R.&Block,S.M.BindingandtranslocationofterminationfactorRhostudiedatthesingle-moleculelevel.J.Mol.Biol.423,664–676(2012).CAS  PubMed  PubMedCentral  GoogleScholar  Roy,R.,Hohng,S.&Ha,T.Apracticalguidetosingle-moleculeFRET.Nat.Methods5,507–516(2008).CAS  PubMed  PubMedCentral  GoogleScholar  Lee,S.,Lee,J.&Hohng,S.Single-moleculethree-colorFRETwithbothnegligiblespectraloverlapandlongobservationtime.PLoSONE5,e12270(2010).ADS  PubMed  PubMedCentral  GoogleScholar  Ingham,C.J.,Dennis,J.&Furneaux,P.A.AutogenousregulationoftranscriptionterminationfactorRhoandtherequirementforNusfactorsinBacillussubtilis.Mol.Microbiol.31,651–663(1999).CAS  PubMed  GoogleScholar  Bae,B.etal.PhageT7Gp2inhibitionofEscherichiacoliRNApolymeraseinvolvesmisappropriationofσ70domain1.1.Proc.NatlAcad.Sci.USA110,19772–19777(2013).ADS  CAS  PubMed  PubMedCentral  GoogleScholar  DownloadreferencesAcknowledgementsWethankProf.KookSunHafortheinitialdesignandconstructionoffluorescenttranscriptioncomplexes,andDr.RanjanSenforprovidingexpressionplasmidsofthethreeρmutants.ThisworkwassupportedbygrantsfromtheNationalResearchFoundationofKorea(2019R1A2C2005209toS.Hohng,2019M3E5D6066058toJ.Y.K.,and2018R1A2B2004602toY.S.)andfromtheHigh-RiskHigh-ReturnProjectofKAIST(N10110078toC.K.).AuthorinformationAuthornotesHeesooUhmPresentaddress:DepartmentofPhysics,UniversityofOxford,Oxford,OX13PU,UKAuthorsandAffiliationsDepartmentofPhysicsandAstronomy,andInstituteofAppliedPhysics,SeoulNationalUniversity,Seoul,08826,RepublicofKoreaEunhoSong, HeesooUhm & SungchulHohngDepartmentofBiologicalSciences,KoreaAdvancedInstituteofScienceandTechnology,Daejeon,34141,RepublicofKoreaPalindaRuvanMunasingha, Yeon-SooSeo & ChangwonKangDepartmentofChemistry,KoreaAdvancedInstituteofScienceandTechnology,Daejeon,34141,RepublicofKoreaSeunghaHwang & JinYoungKangAuthorsEunhoSongViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarHeesooUhmViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarPalindaRuvanMunasinghaViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSeunghaHwangViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarYeon-SooSeoViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarJinYoungKangViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarChangwonKangViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarSungchulHohngViewauthorpublicationsYoucanalsosearchforthisauthorin PubMed GoogleScholarContributionsS.Hohng,E.S.,H.U.,C.K.,andJ.Y.K.conceivedthestudy.E.S.andH.U.designedandperformedallbiophysicalexperimentsbymeasuringsingle-moleculefluorescenceimages.P.R.M.andY.S.preparedρmutantsandperformedallbiochemicalexperiments.J.Y.K.andS.HwangpreparedE.coliRNAP,NusA,andNusGandperformedbulktranscriptionassays.C.K.,J.Y.K.,andY.S.supervisedthemolecularbiologyandbiochemistryparts.S.Hohngsupervisedthebiophysicspart.E.S.,S.Hohng,C.K.,andJ.Y.K.collectivelyanalyzedallthedata.Allauthorscontributedtowritingandrevisingthepaper.CorrespondingauthorsCorrespondenceto JinYoungKang,ChangwonKangorSungchulHohng.Ethicsdeclarations Competinginterests Theauthorsdeclarenocompetinginterests. Peerreview Peerreviewinformation NatureCommunicationsthankstheanonymousreviewersfortheircontributiontothepeerreviewofthiswork.Peerreviewerreportsareavailable. AdditionalinformationPublisher’snoteSpringerNatureremainsneutralwithregardtojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.Supplementaryinformation SupplementaryInformationReportingSummaryPeerReviewFileSourcedata SourceDataRightsandpermissions OpenAccessThisarticleislicensedunderaCreativeCommonsAttribution4.0InternationalLicense,whichpermitsuse,sharing,adaptation,distributionandreproductioninanymediumorformat,aslongasyougiveappropriatecredittotheoriginalauthor(s)andthesource,providealinktotheCreativeCommonslicense,andindicateifchangesweremade.Theimagesorotherthirdpartymaterialinthisarticleareincludedinthearticle’sCreativeCommonslicense,unlessindicatedotherwiseinacreditlinetothematerial.Ifmaterialisnotincludedinthearticle’sCreativeCommonslicenseandyourintendeduseisnotpermittedbystatutoryregulationorexceedsthepermitteduse,youwillneedtoobtainpermissiondirectlyfromthecopyrightholder.Toviewacopyofthislicense,visithttp://creativecommons.org/licenses/by/4.0/. ReprintsandPermissionsAboutthisarticleCitethisarticleSong,E.,Uhm,H.,Munasingha,P.R.etal.Rho-dependenttranscriptionterminationproceedsviathreeroutes. NatCommun13,1663(2022).https://doi.org/10.1038/s41467-022-29321-5DownloadcitationReceived:20July2021Accepted:09March2022Published:29March2022DOI:https://doi.org/10.1038/s41467-022-29321-5SharethisarticleAnyoneyousharethefollowinglinkwithwillbeabletoreadthiscontent:GetshareablelinkSorry,ashareablelinkisnotcurrentlyavailableforthisarticle.Copytoclipboard ProvidedbytheSpringerNatureSharedItcontent-sharinginitiative CommentsBysubmittingacommentyouagreetoabidebyourTermsandCommunityGuidelines.Ifyoufindsomethingabusiveorthatdoesnotcomplywithourtermsorguidelinespleaseflagitasinappropriate. DownloadPDF Advertisement Explorecontent Researcharticles Reviews&Analysis News&Comment Videos Collections Subjects FollowusonFacebook FollowusonTwitter Signupforalerts RSSfeed Aboutthejournal Aims&Scope Editors JournalInformation Openaccess EditorialValuesStatement JournalMetrics Editors'Highlights Contact Editorialpolicies TopArticles Publishwithus Forauthors ForReviewers Submitmanuscript Search Searcharticlesbysubject,keywordorauthor Showresultsfrom Alljournals Thisjournal Search Advancedsearch Quicklinks Explorearticlesbysubject Findajob Guidetoauthors Editorialpolicies Closebanner Close SignupfortheNatureBriefingnewsletter—whatmattersinscience,freetoyourinboxdaily. Emailaddress Signup IagreemyinformationwillbeprocessedinaccordancewiththeNatureandSpringerNatureLimitedPrivacyPolicy. Closebanner Close Getthemostimportantsciencestoriesoftheday,freeinyourinbox. SignupforNatureBriefing