英文名
Huntingtin
effector
MaxPlanckInstituteofMolecularCellBiologyandGenetics,01307Dresden,Germany
ThemolecularmechanismsunderlyingthetargetingofHuntingtin(Htt)toendosomesanditsmultifacetedroleinendocytosisarepoorlyunderstood.Inthisstudy,wehaveidentifiedHtt-associatedprotein40(HAP40)asanoveleffectorofthesmallguanosinetriphosphataseRab5,akeyregulatorofendocytosis.HAP40mediatestherecruitmentofHttbyRab5ontoearlyendosomes.HAP40overexpressioncausedadrasticreductionofearlyendosomalmotilitythroughtheirdisplacementfrommicrotubulesandpreferentialassociationwithactinfilaments.Remarkably,endogenousHAP40wasup-regulatedinfibroblastsandbraintissuefromhumanpatientsaffectedbyHuntington'sdisease(HD)aswellasinSTHdhQ111striatalcellsestablishedfromaHDmousemodel.Thesecellsconsistentlydisplayedalteredendosomemotilityandendocyticactivity,whichwasrestoredbytheablationofHAP40.InrevealinganunexpectedlinkbetweenRab5,HAP40,andHtt,weuncoveredanewmechanismregulatingcytoskeleton-dependentendosomedynamicsanditsdysfunctionunderpathologicalconditions.
F.Severin'spresentaddressisBiotechnologyCentre,UniversityofTechnologyDresden,CellularMachines,01307Dresden,Germany.
Abbreviationsusedinthispaper:EEA1,earlyendosomeantigen1;F-actin,filamentousactin;GDI,GDPdissociationinhibitor;GDP,guanosinediphosphate;HAP,Htt-associatedprotein;HD,Huntington'sdisease;HIP;Htt-interactingprotein;Htt,Huntingtin;LAMP,lysosome-associatedmembraneprotein;polyQ,polyglutamine;RNAi,RNAinterference;siRNA,shortinterferingRNA;Tfr,transferrinreceptor.
Introduction
Huntington'sdisease(HD)isaneurodegenerativedisordercausedbyexpansionoftheCAGrepeatinthegeneencodingHuntingtin(Htt),whichconferstotheproteinanexpandedNH2-terminalpolyglutamine(polyQ)stretchof>35residues(forreviewseeHarjesandWanker,2003).ThefunctionofHttislargelyunclear.Ithasbeenshowntointeractwithmicrotubules(Hoffneretal.,2002)andtodisplayanti-apoptoticactivity(Rigamontietal.,2000,2001).InsightsintoitsfunctioncamefromstudiesofHtt-interactingproteins(HIPs)andHtt-associatedproteins(HAPS).Forexample,interactionswithHIP1,HIP1R,PACSIN1,SH3GL3,andHIP14haveimplicatedHttinclathrin-mediatedendocytosis.StudiesofHAP1havesuggestedaroleforHttinaxonaltransportinneuronsbylinkingvesiclestothedyneindynactinmotorcomplex(Block-Galarzaetal.,1997;Engelenderetal.,1997).ThepolyQexpansionconferstheadjacentproline-richsequenceinHttalterationsinbindingaffinitytoHIPs/HAPs.Thus,releasefromorsequestrationofthesemoleculesbymutantHtthasbeenimplicatedinthepathogeneticmechanisms.
Forexample,thetighterbindingofHAP1tomutantHttisthoughttoimpairthecorrectdynactindyneinmotorcomplexassemblyandcauseatraffickingdefect,leadingtoneuronaldegeneration(Block-Galarzaetal.,1997;Engelenderetal.,1997;Gauthieretal.,2004).Consistently,mutantHttwasrecentlyshowntoreleasedyneinfrommicrotubulesandreducethemotilityofEGFPbrain-derivedneurotrophicfactorcontainingvesiclesinvivo(Gauthieretal.,2004).However,theupstreameventsthattargetHttanditspartnerstotheirvarioussitesoffunctionandthemechanismswherebytheyregulateintracellulartraffickingremainelusive.
Inthisstudy,wereportanunexpectedlinkbetweenHttandthesmallGTPaseRab5viatheadaptorproteinHAP40.Rab5isakeyregulatorofendocytosisthatorchestratestherecruitmentofmultipleeffectorproteinsontheearlyendosomemembranetoregulateorganelletethering,fusion,andmicrotubule-dependentmotility(ZerialandMcBride,2001).OurdataextendtheanalysisoftheRab5effectormachinerybyfunctionallyimplicatingtheinteractionbetweenRab5andHttintheregulationofthedifferentialassociationofearlyendosomeswiththeactinandtubulincytoskeleton.
Results
TheHttHAP40complexisanovelRab5effector
AffinitychromatographyrevealedseveraldownstreameffectorsofthesmallGTPaseRab5(Christoforidisetal.1999).Surprisingly,amongtheproteinsspecificallyelutedfromtheGST-Rab5GTPSbutnotfromtheRab5guanosinediphosphate(GDP)norGST-Rab4affinitycolumn,weidentifiedHttandHAP40(PetersandRoss,2001)bymassspectrometryandimmunoblotting(Fig.1,AandB).therefore,weinvestigatedthefunctionofHAP40andHttwithrespecttoRab5.
WefirsttestedwhetherHAP40,Htt,orbothinteractdirectlyandspecificallywithRab5.Tothisend,full-lengthHAP40wasclonedfromaratbraincDNAlibraryandinvitrotranslated.Becauseofitslargemolecularmass(348kD;Fig.2A),fragmentsofwild-typeHttweretranslatedtofacilitatetheanalysis.Invitrotranslationyieldedmajorproductsofpredictedsizeaswellaslowermolecularmassbandspresumablyasaresultofinternalinitiation(Fig.2B).ImmobilizedRab5-GSTfusionproteinpreloadedwitheitherGDPorGTPSwereincubatedwiththetranslationproductsandwashed,andboundproteinswereelutedwithglutathioneanalyzedbySDS-PAGEandautoradiography(Christoforidisetal.,1999).Similartoearlyendosomeantigen1(EEA1),whichservedasapositivecontrol,HAP40displayedspecificbindingtoGTPS-versusGDP-boundRab5(Fig.2C,comparelanes1and11withlanes2and12).Incontrast,noneoftheHttfragmentsexhibitedsignificantbindingtoRab5(Fig.2C,lanes310).BecauseHttwaspurifiedontheGST-Rab5affinitycolumn,wenexttestedwhetherbindingofHtttoRab5occursindirectlyandrequiresHAP40asabridge.HAP40andHttfragmentswerecotranslatedinvitro(Fig.2B,lanes69)andappliedontheRab5columns.indeed,theCOOH-terminalpartofHtt(Fig.2A,Htt4)waselutedtogetherwithHAP40fromGTPS-butnotGDP-boundGST-Rab5(Fig.2D,lanes7and8).NoneoftheotherHttfragmentsdisplayedsuchbinding(Fig.2D,lanes16),whichisconsistentwiththereportedinteractionmapbetweenHAP40andHtt(PetersandRoss,2001).NeithertheHttfragmentsnorHAP40displayedbindingtoRab4,Rab6,Rab7,orRab11(Fig.2D),suggestingtheinteractionwithRab5isspecific.Thus,weconcludethatHAP40bindstotheCOOH-terminalpartofHttandlinksthecomplextoactiveRab5.Byapplyinga10-foldexcessoftheCOOH-terminalfragmentofHttontotheRab5columntoreducebindingoffreeHAP40toRab5(Fig.2E),weestimatedthestoichiometryoftheRab5/HAP40/Httinteractioninthisassaytobe1:1:1(seeRab5affinity...cloning).
HttisrecruitedontoearlyendosomeinaHAP40-andRab5-dependentmanner
WebegantestingthefunctionalrelevanceofthisinteractioninvivobyimmunofluorescencemicroscopyanalysisofHeLacells.First,weverifiedthattheanti-HAP40antibodyresultedinspecificstainingabovebackgroundlevelsfordetectionoftheendogenousantigen(Fig.3A).EndogenousHAP40displayedadiffusestaininginthecytoplasmandaccumulatedinthenucleus,whereasendogenousHttlocalizedtodiscretecytoplasmicstructures(Fig.3B)asreportedpreviously(PetersandRoss,2001;TaoandTartakoff,2001).ColocalizationofendogenousHAP40andHttwashardlydetectable(Fig.3B).EarlyendosomeslabeledwithEGFP-Rab5displayedlittlecolocalizationwithendogenousHtt(7±5%overlap;n=10;Fig.3C).However,theassociationofHAP40withearlyendosomesdramaticallyincreaseduponoverexpression.IncellsoverexpressingHAP40(seeFig.7D)butnotEGFP-Rab5alone,HAP40significantlycolocalizedwithendogenousHttonEGFP-Rab5positiveearlyendosomes(Fig.3,compareBwithD;endogenousHttandEGFP-Rab5:43±6%overlap,n=10;HAP40andEGFP-Rab5:31±7%overlap,n=10;Fig.3,compareCwithD).WeexperimentallyverifiedthattheHAP40fluorescencesignals(Fig.3D)werenotcausedbybleed-throughoftheHttsignals(becauseofextendedAlexaFluor568emissionintheCy5channel)andthatswappingthesefluorescentdyesonthesecondaryantibodiesresultedinsimilardistributionpatternsofHttandHAP40(seeCellcultureprocedures).TheendosomalcolocalizationofendogenousHAP40andHttwasevenmorestrikinguponexpressionoftheactivatedEGFP-Rab5Q79Lmutant(Htt:52±7%overlap,n=8),whichcausedthecharacteristicswellingofearlyendosomes(Fig.4A;Stenmarketal.,1994).
GiventhedifficultiesdetectingendogenousHAP40andHttonearlyendosomesinuntreatedcells(Fig.3,BandC)wesoughttoverifythechangesinthelocalizationofbothproteinsupontheoverexpressionofHAP40biochemically.Tothisend,wepreparedearlyendosomesfromHeLacellsforWesternblotanalysis.Indeed,wefoundthatthelevelsofHAP40andHttincreasedonearlyendosomesfromtheHAP40overexpressorcomparedwithuntreatedcells(Fig.3E).Earlyendosome(EEA1andtransferrinreceptor[Tfr])aswellaslysosomal(lysosome-associatedmembraneprotein(LAMP);Eskelinenetal.,2003)andGolgi(GM130;Nakamuraetal.,1995)markersremainedunchanged,confirmingequalloadingandthespecificityofchangesthroughelevatedHAP40onearlyendosomes.
ToconfirmtherequirementofHAP40fortherecruitmentofHttontoearlyendosomes,wetransfectedHeLacellswithshortinterferingRNA(siRNA)duplexesagainstHAP40andunrelatedsiRNA(againstGFP)ascontrol.TheHAP40siRNAspecificallyandefficientlyreducedtheproteinlevelsby90%(Fig.4C),whereasthelevelofEEA1remainedunchanged.WhencellswerecotransfectedwiththeexpressionvectorforEGFP-Rab5Q79LandHAP40siRNA,Httwasnolongerdetectableontheenlargedendosomes(8±6%overlap,n=9;Fig.4B),confirmingHAP40asaprerequisitetobridgeHtttoactiveRab5.Collectively,thesedatasuggestthatactiveRab5andHAP40areratelimitingfortherecruitmentofHttontoearlyendosomes.
TheHttHAP40complexinhibitsthebindingandmotilityofearlyendosomesonmicrotubules
Htthaspreviouslybeenshowntobindmicrotubulesandregulatemicrotubulemotorinteractions(Block-Galarzaetal.,1997;Engelenderetal.,1997;Hoffneretal.,2002;Gauthieretal.,2004).BecauseRab5regulatesendosomemotilityalongmicrotubules(Nielsenetal.,1999;Hoepfneretal.,2005),weexploredtheroleoftheHttHAP40complexinthisprocess.First,byusingacell-andcytosol-freeassaythatrecapitulatestheRab5-dependentmovementofearlyendosomesalongmicrotubules(Hoepfneretal.,2005),wefoundthatHttHAP40inhibitedmicrotubule-dependentearlyendosomemotility.AdditionoftheGTPS-(containingHttandHAP40;Fig.1B)butnottheGDP-loadedRab5columneluatereducedthemotilitycomparedwithcontrolconditions(Fig.5A).Thisinhibitionwasspecificallyrescuedwithanti-Httbutnotunrelatedantibodies.Theadditionof1μMGST-HAP40fusionproteincompletelyblockedtheinvitromotility(Fig.5,AandB).Second,weperformedabiochemicalearlyendosomemicrotubule-bindingassayasdescribedpreviously(Nielsenetal.,1999)withsomemodificationstoimprovethequantitativeassessment.AnearlyendosomeenrichedfractionwaspreparedfromHeLacellspulsedwithrhodaminetransferrin,incubatedwithtaxol-stabilizedmicrotubules,ATP,andfactorstobetested,andcentrifugedthroughasucrosecushion.Theresultingpelletofmicrotubule-associatedmaterialwasanalyzedbyimmunoblotting(Fig.6A)andfluorimetricallyfortherhodaminetransferrincontent(Fig.6D).Inadilutionseriesforcalibration,weverifiedthattheamountofendosomesandfluorescenceintensitycorrelatedlinearly(seeMicrotubuleandactinspin-downassays).Westernblottingrevealedthatthe-tubulincontentwassimilarbetweensamples,rulingoutsecondaryeffectsonmicrotubulestability(Fig.6A,comparelanes25withlane1).Theadditionof1μMGST-HAP40proteindecreasedtheamountofearlyendosomesinthepelletasrevealedbyEEA1andTfr(Fig.6A,comparelane3withlane2).TheGST-Rab5GTPScolumneluateadding0.3μMHAP40(Fig.6C)causedasimilarinhibition(Fig.6A,comparelane4withlane2)thatwasrescuedthroughtheadditionofantiserumagainstHtt(Fig.6A,comparelane5withlanes4and2).Toruleouttheideathattheobserveddifferencesresultfromthebundlingofmicrotubulescausingunspecificcosedimentationofanymembranousstructure,weprobedthepelletsfornonendosomalcontaminantsinthefraction.Boththelysosomal(LAMP1)andGolgimarker(GM130)pelletedwithsimilarefficiencyinallsamples.Collectively,thesedataindicatethatHttHAP40specificallylowersthebindingofearlyendosomestomicrotubules.
Havingvalidatedtheassay,wenextquantifiedtheamountofendosomesboundtomicrotubulesfluorimetrically.Omissionofeithermicrotubulesorearlyendosomesreducedthefluorescencesignaltobackgroundlevels(Fig.6D).Asreportedpreviously(Nielsenetal.,1999),theassociationofearlyendosomeswithmicrotubuleswasenergydependentandrequiredactiveRab5.OmittingATPorsubstitutingitwiththenonhydrolyzableadenylyl-imidodiphosphateanalogueresultedinan50%reductioninmicrotubulebinding.TherequirementforATPcanbeexplainedbytheroleofPI3-KintheassemblyandmaintenanceofafunctionalRab5domainonendosomes(ZerialandMcBride,2001)andinrecruitmentoftheendosomalkinesinKIF16B(Hoepfneretal.,2005).ExtractionofRabproteinsfrommembranesbytheadditionof1μMofrecombinantRabGDPdissociationinhibitor(GDI;Ullrichetal.,1994)ortreatmentwithRN-tre,aGTPase-activatingproteinforRab5(Lanzettietal.,2000),causedan40%reductionofendosomesinthepellet(Fig.6D).Asforendosomemotility(Fig.5A),HttandHAP40inhibitedendosomemicrotubulebinding.TheGST-Rab5GTPScolumneluateledtoan25%reductioninbinding(Fig.6D),butsupplementingthereactionwithantibodiesagainstHttrestoredbindingtocontrollevels.Theadditionof1μMHAP40-GSTfusionproteindecreasedthebindingby60%,whereasGSTalonedidnothaveanyeffect.Collectively,thesedataunderpintheabilityofHAP40andHtttodestabilizeendosomemicrotubuleassociation.
Wenextperformedtime-lapsevideomicroscopystudiestocorrelatethesebiochemicalfindingsinvitrowiththeregulationofearlyendosomedynamicsinvivo.WedetectedadrasticreductioninmotilityofEGFP-Rab5positiveendosomesinHeLacellsoverexpressingHAP40(Fig.7A)comparedwithcellstransfectedwithEGFP-Rab5alone(Fig.7AandVideos1and2,availableathttp://www.jcb.org/cgi/content/full/jcb.200509091/DC1).Whereassomeresidualmotilityactivitywasobservedinthecellperiphery,earlyendosomesintheperinuclearregionappearedstatic,withfrequentshort-rangemovementsalmostcompletelyimpairedinlong-rangemotility(Nielsenetal.,1999,rinketal.2005).Collectively,thesedatasuggestthattheRab5-dependentrecruitmentofHttontoendosomesbyHAP40disruptsearlyendosomemicrotubuleinteractions,thusleadingtoareductioninorganellemotility.
HDcellsdisplayincreasedlevelsofHAP40andareimpairedinearlyendosomemotility
WenextaskedwhetheralterationsofendosomemotilitycouldoccurincellsbearingtheHDmutation.PrimaryfibroblastsfromfivehealthyindividualsandfiveunrelatedHDpatientsweretransfectedwithEGFP-Rab5tocomparethemotilityofearlyendosomes.TheidentityoftheHDcelllineswasconfirmedbyWesternblottingtodetectthepolyQ-expandedHtt.BecausethecellswerederivedfrompatientsheterozygousfortheHDgene,theyexpressbothnormalandmutantHtt.ThelatterisknowntodisplayalowermobilityinSDS-PAGE(Trottieretal.,1995),thuscausingadoubletontheblot(Fig.7D).Strikingly,weobservedaseverereductioninearlyendosomemotilityinallHDcelllinescomparedwithfibroblastsfromhealthyindividuals(Fig.7BandVideos3and4,availableathttp://www.jcb.org/cgi/content/full/jcb.200509091/DC1).ThesimilaritybetweenthisphenotypeandthealterationsinduceduponHAP40overexpressioninHeLacells(Fig.7AandVideos1and2)hintedtoacommonmolecularbasis.Interestingly,wediscoveredan10-foldup-regulationofendogenousHAP40levelsbyWesternblottinginallHDcelllinescomparedwithnormalfibroblasts(Fig.7D).Asacontrol,thelevelsofEEA1remainedunchanged.moreover,wefoundthatHAP40proteinlevelsweresignificantlyelevatedinstriataltissue(caudate,putamen,accumbens,andglobuspallidus)fromhumanpostmortembrainsaffectedbyHD(adultonsetgrade)comparedwithcontrolbrains(Fig.8B).OurdatasuggestthatthemotilityblockinHDcellsmaybecausedbyelevatedHAP40levels,asmimickedbyoverexpressionofHAP40inHeLacells.Consistently,endogenousHttlocalizedtoEGFP-Rab5labeledendosomesinfibroblastsfromhumanHDpatientsaswellasstriatalSTHdhQcellsfromaHDmousemodel(Tretteletal.,2000)butnotincellslackingthemutantHtt(Fig.9).ThisphenotypeisalsocausedbyoverexpressedHAP40inHeLacells(Fig.3D).
Ourdatadonotexcludethepossibilitythattheobservedinhibitionofearlyendosomemotilitymayhaveotherunderlyingcauses.Asatesttoourhypothesis,weattemptedtorescuetheinhibitionofearlyendosomemotilitybyspecificallyablatingHAP40fromtheHDfibroblastsbyRNAinterference(RNAi).TransfectionofHAP40siRNAefficientlyreducedtheHAP40proteinlevelsinthesecells(Fig.7E)asinHeLacells(Fig.4C).Indeed,endosomemotilitywasrestoredbyRNAiagainstHAP40(Fig.7CandVideo5,availableathttp://www.jcb.org/cgi/content/full/jcb.200509091/DC1)butnotagainstGFP,suggestingthattheup-regulationofHAP40isindeedtheunderlyingmechanismofthemotilitydefectinHDfibroblasts.
GiventheabilityofHAP40toreduceendosomemicrotubulebindinginvitro(Fig.6,AandD),weinvestigatedwhethertheobservedmotilityblockinHDcelllineswascausedbyareleaseofendosomesfrommicrotubulesinvivo.ImmunofluorescenceanalysisoncellstransfectedwithEGFP-Rab5showedaconsiderablealignmentofearlyendosomestomicrotubulesinhealthyfibroblasts(82±9%overlap,n=10)buttoamuchlesserextentinHDcelllines(15±6%overlap,n=9;Fig.7,compareFwithG).Incontrast,earlyendosomeswerestrikinglyalignedwithfilamentousactin(F-actin)inallfiveHDcelllines(44±8%overlap,n=10)butnotinhealthyfibroblasts(2±1%overlap,n=10;Fig.7,compareGwithF).AsimilarphenotypewasobtainedbytheoverexpressionofHAP40inHeLacells(notdepicted).
TheHttHAP40complexenhancesthebindingofearlyendosomestoactin
ThealignmentofearlyendosomesonF-actininHDcelllinescouldbeasecondaryeffectfromtheinhibitionofendosomemicrotubuleinteractions.TodirectlytesttheroleofHAP40intheassociationbetweenearlyendosomesandactinfilaments(Fig.6,BandE),wemodifiedthebiochemicalsedimentationassayusedtostudyendosomemicrotubuleinteractions(Fig.6,AandD)byreplacingtaxol-stabilizedmicrotubuleswithfreshlyinvitropolymerizedF-actin.Unlikeforendosomemicrotubulebinding,depletionofactiveRab5fromendosomalmembranesbytreatmentwitheitherRab-GDIorRN-tredidnotdecreaseendosomeactininteractions(Fig.6E).Thus,abasallevelofendosomeF-actinbindingactivitywasindependentofRab5.However,theadditionofHAP40-GSTfusionproteinstimulatedbinding(260%)overcontrollevels,whereasGSTalonewasineffective.Evidently,endosomebindingtomicrotubulesandF-actinisreciprocallyregulatedthroughHAP40becauseconcentrations(1μM)inhibitingbindingtomicrotubules(Fig.6D)clearlystimulatedbindingtoF-actin(Fig.6E).ThiseffectwasRab5dependentbecausewhenRab-GDIorRN-trewereaddedtogetherwithHAP40,thestimulationwasnearlyabolished(Fig.6E).Again,immunoblottingconfirmedthattheamountofactininthepelletswasunaffectedbyanyaddedproteinandthatallchangesinpelletablefluorescencecorrespondedconsistentlytoalteredbandintensitiesofearlyendosomalmarkers(EEA1andTFr)butnottoothers(LAMP1andGM130),indicatingspecificeffectsofHttHAP40onendosomeactinbinding(Fig.6B).
AlterationsinearlyendosomemotilityinaHDmousemodelofstriatalcells
Next,weinvestigatedtheroleofHAP40inRab5dynamicsinanexperimentalsystemthatismorerelevantforHDusingimmortalizedSTHdhQstriatalcells(Tretteletal.,2000).ThesecellswereestablishedfromembryonicnormalorHDknock-inmiceandeitherexpressnormal(STHdhQ7/7)ormutantHttasaresultofaCAGexpansioninsertedintotheendogenousHttgene(heterozygousSTHdhQ7/111andhomozygousSTHdhQ111/111).Thus,theyreflecttheclosestsituationtoHDpatientsasnormal,andmutantHttareexpressedatendogenouslevels.Remarkably,weagainfoundendogenousHAP40proteinlevelselevatedinSTHdhQ7/111andSTHdhQ111/111comparedwithSTHdhQ7/7cells(Fig.8B),whichisconsistentwiththedataonHDfibroblasts(Fig.7D)andbraintissuefromHDpatients(Fig.8B).LivecellimagingrevealedEGFP-Rab5positiveorganellesmovingbidirectionallyinneuronlikeoutgrowthsaswellasinthecellbodyofnormalSTHdhQ7/7cells(Fig.8AandVideo6,availableathttp://www.jcb.org/cgi/content/full/jcb.200509091/DC1).conversely,STHdhQ111/111andSTHdhQ7/111cellsclearlydisplayedadrasticreductionofendosomedynamics(Fig.8AandVideos7and8),whichisconsistentwiththeobservationsonHDfibroblasts(Fig.7BandVideos3and4).Collectively,defectsofRab5dynamicscausedbypathogenicexcessofHAP40apparentlyoccurinperipheraltissuessuchasfibroblastsaswellasinneuronalsystemsand,therefore,areofpotentialrelevanceforHD.
IncreasedHAP40levelscausealterationsinendocytictrafficking
TogainsomeinsightsintopossiblealterationsofendocytictransportcausedbyincreasedlevelsofHAP40,wetestedtheuptakeoftransferrininHeLacellsoverexpressingHAP40andinfibroblastsfromhealthyandHDpatients.Fig.10Ashowsthattheuptakeoftransferrinwasreducedby30%inHeLacellsoverexpressingHAP40comparedwithmock-transfectedorcontrolcells.Consistentwiththisresult,fibroblastsfromHDpatientsdisplayinghigherlevelsofendogenousHAP40(Fig.7D)displayedasimilarlyreduceduptakeoftransferrincomparedwithnormalfibroblasts(Fig.10B).Asfortheblockofendosomemotility,theinhibitoryeffectontransferrinuptakewasrescuedbythedepletionofHAP40byRNAi.Thesedatasuggestthattheinhibitoryeffectsonendosomemotilitycausedbyup-regulationofHAP40alsoresultindefectsincargotransportthroughtheendocyticpathway.
Discussion
ThekeyfindingofthisstudyisthatthecomplexbetweenHAP40andHttisadirectdownstreameffectorofRab5thatregulatesthedynamicsofearlyendosomesthroughaswitchfrommicrotubulestoF-actin.Thesefindingsprovideimportantnewinsightsintohowthemotilityofearlyendosomesisregulatedunderphysiologicalandpathologicalconditions.Htthasbeenimplicatedinclathrin-mediatedendocytosis,regulationoftheactincytoskeleton,andmicrotubule-dependenttransportalongtheendocyticpathwayviainteractionswithitsnumerousbindingpartners(HarjesandWanker,2003).SuchmultiplicityofrolesimpliesthattheactivitiesofHttinendocyticmembranetraffickingneedtobespatiallyandtemporallycoordinated.InshowingforthefirsttimethatacomplexbetweenHttandoneofitsbindingpartners,HAP40,canberecruitedontoendosomesbyinteractingdirectlywithRab5,ourdataprovidenovelinsightsintothemechanismsgoverningthetargetingofHtttoearlyendosomesanditsregulatoryactivityoncytoskeleton-dependentdynamics.
Undernormalphysiologicalconditions,earlyendosomesundergofrequentshort-rangemovementsonactinbutalsolong-rangebidirectionalmovementsalongmicrotubules(Nielsenetal.,1999;gasmanetal.,2003;Rinketal.2005).PlusendmovementofearlyendosomesalongmicrotubulesispropelledbyKIF16B(Hoepfneretal.2005).OverexpressionofHAP40,whichisratelimitingfortherecruitmentofHttonthemembrane,causedthedetachmentofearlyendosomesfrommicrotubulesandtheirpreferentialassociationwithactinfilaments,thuslimitingboththeirvelocityandrangeofmovements.Therefore,theHttHAP40Rab5complexisakeyregulatoroftheswitchfromonetypeoffilamentstoanother.Ourdataareconsistentwithpreviousstudiesdocumentingalterationsinmicrotubule-dependentmotilityinHDmodelsystems(Block-Galarzaetal.,1997;Engelenderetal.,1997;Gauthieretal.,2004).However,thesestudieshaveexclusivelyimplicatedalterationsbetweenmutantHttandHIP/HAPeffectors.Forexample,wild-typeHtthasbeenshowntointeractwiththedynactinsubunitp150GluedviaHAP1andmutantHtttodisruptthedyneinmotorcomplexinaxonaltransport(Gauthieretal.,2004).Inthisstudy,wehaveuncoveredadifferentmechanismbasedontheup-regulationofanHttadaptor,HAP40.CompellingevidenceinsupportofthismechanismwasprovidedbytherescueofthemotilityblockupondepletionofHAP40byRNAibothinHDfibroblastsandinstriatalcells.BecausetheCOOH-terminalpartofHttisresponsiblefortheunderlyinginteractionswithHAP40andRab5,thisendosomalrecruitmentaffectsnormalaswellasmutantHtt.Consistently,wild-type(overexpressionofHAP40inHeLacells),heterozygous(STHdhQ7/111),andhomozygous(STHdhQ111/111)HDcellsdisplayverysimilarphenotypes.Inthisway,functionalcompetitionwiththeRab5-dependentendosomalkinesinKIF16Banddisruptionofthedyneindynactincomplex(Gauthieretal.,2004)couldaccountforthecompromisedbidirectionalmicrotubule-basedmotility.Therefore,ourdatasuggestthatmultipledefectsmaycontributetotheblockofvesiculartransportinHDcells,whichimplicatesvariousHIP/HAPsinteractingwithdifferentregionsofHtt.
PreviousfindingshaveshownthatHttisalsolinkedtoRab8throughFIP-2(HattulaandPeranen,2000)andthatthisinteractionregulatescellpolarizationandmorphogenesis.Whereasthemechanismunderlyingthelatterprocessisunclear,inlightofourobservation,weproposethatHttisamultifunctionalproteinthatmayberecruitedbydifferentRabGTPasesthroughdifferentadaptorsatdifferentintracellularlocationstoregulateorganelledynamics.
TheprimarypathologicalcauseofHDisattributedtotheabnormalactivityofHttanditsinteractingpartnersinthenucleus(LandlesandBates,2004).ThemajorityofHAP40isalsonuclearundernormalconditions(Fig.3;PetersandRoss,2001),althoughthefunctionalsignificanceofthislocalizationisunclearatpresent.OurdataindicatethatHAP40fulfillsafunctioninthecytoplasm.ItisinterestinginthisrespecttonotethatHAP40addstotheincreasinglistofproteinsimplicatedinadualroleinendocytictraffickingandnuclearsignaling(Miaczynskaetal.,2004).Themolecularmechanismunderlyingtheup-regulationofHAP40unexpectedlyobservedincellsandbraintissuefromHDpatientsremainstobedetermined.ThemostlikelyexplanationisthatitarisesasaconsequenceofalterationsingeneexpressioncausedbymutatedHttinthenucleus.MessengerRNAmicroarraystudieshaverevealedmanytranscriptionalabnormalitiesinHD(Chanetal.,2002;Sipioneetal.,2002;SugarsandRubinsztein,2003),althoughnochangesforHAP40oranyotherHtt-interactingpartnershavebeenreportedsofar.Up-regulationofHAP40attheproteinlevelmightthusserveasanewdiagnosticindicatorforthedisease.However,itisprematuretojudgetowhatextentthisabnormalityandtheconsequentalterationsinmembranedynamicscontributetothepathogeneticmechanismofHDascomparedwithnuclearactivitiesofHttandHAP40.AlthoughendosomemotilityisnotperturbedinHDcellstotheextentofcompromisingcellviability,weneverthelessuncovereddefectsinendocytictransport,specificallyadecreaseintransferrinuptake.Theseobservationsmaybeofparticularfunctionalrelevanceinneuronsgiventheimportanceoflong-rangetransport(Block-Galarzaetal.,1997;Engelenderetal.,1997;Gauthieretal.,2004).Therefore,itispossiblethatalterationsinendosomemotilityandtransportmayfurthercompromisethepathologicalstateinducedbymutatedHttonthesurvivalandfunctionofneuronsorsomeselectedpopulationsofneuronswhereendosomaltransportisparticularlyratelimitinginvivo.AddressingthishypothesisrequiresamorethoroughinvestigationofHAP40andHttfunctioninthetransportofdifferenttypesofcargoalongtheendocytic/recyclingaswellasdegradativepathwayinneuronalcellsratherthanfibroblasts.AdeeperanalysisoftheroleofHAP40onthemolecularlevelwillhopefullyimproveourunderstandingofbothHDpathogenesisandthemechanismscoordinatingorganellecytoskeletoninteractionsduringmembranetrafficking.
Materialsandmethods
Rab5affinitychromatographyandHAP40andHttcloning
GST-Rab5affinitychromatographywithbovinebraincytosolorinvitrotranslatedproteinswasperformedasdescribedpreviously(Christoforidisetal.,1999).HAP40wasclonedfromratbraincDNAbyPCR,HttfragmentswereclonedbyPCRonfull-lengthHttcDNAinpBlueScript(agiftfromM.Sherman,BostonUniversitySchoolofMedicine,Boston,MA),andallPCRproductswereclonedintopcDNA3.1(Invitrogen)andsequencedtoconfirmtheiridentity.Toestimatethestoichiometryoftheboundcomplex,GST-Rab5wasincubatedinthepresenceofa10-foldexcessoftheCOOH-terminalfragmentofHttoverHAP40.BandintensitiesobtainedundertheseconditionswerequantifiedinImageJsoftware(NationalInstitutesofHealth;Htt4/HAP40=6:1)andcorrectedforthenumberof[35S]methionineresiduesincorporated(Htt4/HAP40=22:4).Theresultingratiowas1:1.Assuminga1:1ratioofRabproteintoeffector,thestoichiometryoftheRab5/HAP40/Httinteractionwasestimatedat1:1:1.
Antibodiesandrecombinantproteins
MonoclonalmouseantibodiesagainstHttwereobtainedfromChemiconandY.Trottier(InstitutdeGenetiqueetdeBiologieMoleculaireetCellulaire,UniversiteLouisPasteur,Illkirch,France),andpolyclonalrabbitantibodyagainstHAP40wasobtainedfromChemicon.Microtubuleswererevealedwithmousemonoclonalanti-tubulinantibody(BDBiosciences),andF-actinwasrevealedwithAlexaFluor568-conjugatedphalloidin(Invitrogen).SecondaryantimouseandrabbitIgGforimmunofluorescencemicroscopywereconjugatedtoAlexaFluor568andCy5(Invitrogen).Rabbitpolyclonalanti-EEA1antibodywasdescribedpreviously(Christoforidisetal.1999),mousemonoclonalantibodyagainstthecytoplasmictailofTfrwaspurchasedfromInvitrogen,mousemonoclonalanti-LAMP1antibodywasobtainedfromBectonDickinson,andmousemonoclonalanti-GM130antibodywasobtainedfromAbcam.Rab-GDIandRN-trewerepreparedasdescribedpreviously(Ullrichetal.,1994;Lanzettietal.,2000).GSTandHAP40-GSTfusionproteinswereaffinitypurifiedfromEscherichiacolilysatesaccordingtostandardprocedures.
Cellcultureprocedures
HeLaandhumanprimaryfibroblastcellsweregrownaccordingtostandardprocedures.ImmortalizedSTHdhQstriatalcelllinesfromcontrolandHDknock-inmicewereculturedanddifferentiatedasdescribedpreviously(Tretteletal.,2000).Fortransientexpressionstudies,cellsweretransfectedusingLipofectAMINE2000(Invitrogen)andused12haftertransfectionforimmunoblotanalysisorintracellularlocalizationstudiesaccordingtostandardprotocols.Fixedcellswereanalyzedusinga100x/NA1.40plan-Apochromatoilimmersionlens(CarlZeissMicroImaging,Inc.)onamicroscope(AxiovertS100TV;CarlZeissMicroImaging,Inc.).Illuminationwasperformedwitha100-WmercurylampwithfiltersetsforGFP,AlexaFluor568,DAPI,andCy5fluorescence.Imageswereacquiredat1,300x1,000pixelswithadigitalcamera(SP1.4.4;DiagnosticInstruments)controlledbytheMetaVue6.1softwarepackage(UniversalImagingCorp.).RawimagesfromvariouscolorchannelswereassembledandcolorizedinAdobePhotoshop7.0.Brightnessandcontrastwereadjustedforvisualclarity.TovalidateimmunofluorescencemicroscopystudiesusingAlexaFluor568-andCy5-conjugatedsecondaryantibodies,HAP40wasoverexpressedinHeLacellsandcellsstainedwithamixtureofanti-HAP40and-Httantisera.SingleadditionofsecondaryantibodyagainstmouseIgG(forHtt)conjugatedwithAlexaFluor568resultedinnodetectablebleed-throughintotheCy5channelattheexposuretimeandfiltersettingsusedforHAP40detectionwithCy5.Moreover,signalsforHAP40thatwereobtainedwithCy5lookedverysimilarinthepresenceorabsenceofAlexaFluor568.Finally,whenthesedyeswereswappedonthesecondaryantibodies,verysimilardistributionpatternsforHAP40andHttwereobtained.Collectively,thecolocalizationbetweenHAP40andHttrevealedinFigs.3and9isnotcausedbybleed-through.Thepercentageofcolocalizationbetweenimmunofluorescencesignalswasdeterminedasfollows:rawsignalsfromdiscretemembranestructuresweremanuallycountedandinspectedforoverlappingwithsignalsinotherchannelstocalculatethepercentageofcolocalization.Onlysignalsabovevalue70ontheeight-bittonalscalewereconsideredspecificandusedfortheanalysis.Typically,10cells(n=10)werecountedtocalculatethemeancolocalizationandSD.ForoverlappingstudiesofEGFP-labeledearlyendosomeswithcytoskeletalfilaments,theseorganelleswereconsideredcolocalizedtoeithermicrotubulesorF-actinifatleastthreeearlyendosomesinarowwereclearlyalignedtoafilament.InareaswheremicrotubulesandF-actinconverged(e.g.,closetothecelledge),earlyendosomescouldoftennotbeassignedtoeithertypeoffilament.Thesumofearlyendosomesalignedtomicrotubules,F-actin,andunassignableearlyendosomes(100%)wasusedtocalculatepercentagesofcolocalization.Internalizedbiotinylatedtransferrinwasdetectedwithruthenium-labeledantitransferrinantibodiesandsubsequentlyquantifiedbyECLanalysisasdescribedpreviously(Hoepfneretal.2005).
Celllinesandtissuesamples
TheprimaryhumanfibroblastcelllinesGM00023,GM00024,GM00037,GM00038,andGM00041(ApparentlyHealthycellcollection)andGM04281,GM04723,GM03621,GM04869,andGM04847(HuntingtonDiseasecellcollection)wereobtainedfromCoriellCellRepositories.Celllineswereestablishedfrombothgendersfromhealthy,unrelatedindividualsbytheageof331yrorfromclinicallyaffectedonsetHDpatientsbytheageof1932yr.Immortalizedstriatalcelllinesestablishedfrommouseembryos(STHdhQ)weregiftsfromM.MacDonald(RichardB.SimchesResearchCenter,Boston,MA;Tretteletal.,2000).Humanpostmortembrainsamples(CAP;GlobusPallidus)fromhealthyandHDgrades3and4donorswereprovidedbytheHarvardBrainTissueResourceCenter.
Proteinidentificationbymassspectrometry
Gel-separatedCoomassie-stainedproteinswereexcisedfromthegelslabandin-geldigestedwithtrypsinasdescribedpreviously(Shevchenkoetal.,1996).Trypticpeptidesweresequencedbynanoelectrospraytandemmassspectrometryonhybridquadrupoletime-of-flightmassspectrometers(Q-TOFI;Micromass,Ltd.andQSTARPulsarI;MDSSciex)asdescribedpreviously(Shevchenkoetal.,1997).DatabasesearchingwasperformedbyMascotsoftware(MatrixScience,Ltd.).
Theinvitromotilityassaywasessentiallyperformedasdescribedpreviously(Nielsenetal.,1999)withthefollowingmodifications(Hoepfneretal.,2005).KHMG(110mMKCl,50mMHepes-KOH,pH7.4,2mMMgCl2,and10%glycerol)wasusedasanassaybuffer.Forpreparationoftheantifadesolution,BRB80bufferwassubstitutedwithKHMG,andthesolutionwasfurthersupplementedwith10%serum.Theenergymixconsistedof75mMcreatinephosphate,10mMATP,10mMGTP,and20mMMgCl2inBRB80.Fortheassay,taxol-stabilizedmicrotubuleswereperfusedinamicroscopychamberandallowedtobindtothecoverslip.Next,10μlof10%nonspecificrabbitseruminantifadesolutionwasperfusedinthechamberfollowedby10μloftheassaymixture(2μlof5mg/mlfluorescentlylabeledearlyendosome,1μlofenergymix,6μloftheantifadesolution,and1μlofsaturatedhemoglobinsolutioninKHMG)andwasincubatedfor5minatRT.Atleastthreevideospersamplewererecorded(60frameswith2-sintervals)andanalyzedasdescribedpreviously(Nielsenetal.,1999).
Cellsweregrownonglasscoverslips,transfectedwithEGFP-Rab5aloneortogetherwithsiRNAagainstHAP40,andtransferredtocustom-builtaluminummicroscopeslidechambers(Nielsenetal.,1999)justbeforeobservation.CellswereanalyzedinCO2-independentmedium/10%FBS(Invitrogen)at37°Cusinga100x/NA1.40plan-Apochromatoilimmersionlens(CarlZeissMicroImaging,Inc.)onamicroscope(IX70;Olympus)placedinatemperature-controlledchamber.Illuminationwasperformedwitha100-WXenonlampfittedwithamonochromatortoexciteGFPfluorescence.Imageswereacquiredat512x512pixelswithacamera(Cascade512B;RoperScientific)instreamacquisitionmodeover2minin300-msintervalsusingtheMetaVue6.1softwarepackage.ImagestackswereeitherconvertedtoQuickTimevideosormotilityeventshighlightedbygeneratingZprojectionsoftheentirestackusingtheImageJsoftwarepackage.
Microtubuleandactinspin-downassays
Bindingofearlyendosomestomicrotubuleswasperformedasdescribedpreviously(Nielsenetal.,1999)butwithsomemodifications.15μgofpreparedearlyendosomespulsedwithrhodaminetransferrin(Sigma-Aldrich)wereincubatedatRTwith5μlofenergymix(seeprevioussection),1μlofnonspecificrabbitserum,and1μMofrecombinantcandidatefactors(HAP40,Rab-GDI,orRN-tre)inKHMGbuffersupplementedwith1mg/mlHeLacellcytosol.After20min,16μgoftaxol-stabilizedmicrotubuleswereaddedtoatotalvolumeof50μl.After10minmoreatRT,thereactionmixturewaslaidover100μlofa35%(wt/vol)sucrosecushion.Aftersedimentationat100,000gfor20minat22°Cinarotor(TLA100;BeckmanCoulter),thesupernatantwasremoved,andtheresultingpelletwaswashedtwiceincytosol-freeKHMGbuffer.Thepelletwaslysedin150μlKHMGsupplementedwith1%sodiumdeoxycholate,andthereleasedrhodaminetransferrinwasquantifiedinafluorimeter.Excitationwasat550nm,andemissionwasat582nm.Acalibrationcurverevealedalinearcorrelationbetweenemissionandearlyendosomeconcentration.Theequationofthelinearregressionanalysiswasy=4,548.4x2,653.1withR2=0.9879,whereyistheemissioninarbitraryunitsandxistheconcentrationofearlyendosomesinproteinmassunits.ForbindingofearlyendosomestoF-actin,globular-actin(Cytoskeleton,Inc.)waspolymerizedinKHMGbufferatRTfor1handcentrifugedfor30minat100,000g.TheresultingpelletwaswashedwithKHMGbufferandresuspendedat10mg/ml.Theassaywasthenperformedwith10μgofF-actinasdescribedformicrotubules.
Onlinesupplementalmaterial
AllvideosshowEGFP-Rab5dynamics,andimageswereacquiredasdescribedinInvivomotilityofearlyendosomes.Video1showsawild-typeHeLacelltransfectedwithplasmidforEGFP-Rab5expressiononly.Video2showsaHeLacellcotransfectedwithplasmidsforEGFP-Rab5andHAP40.Video3showsaprimaryhumanfibroblastfromahealthypersontransfectedwithplasmidforEFGP-Rab5expressiononly,whereasVideo4showsthefibroblastfromaHDpatient.Video5showsaprimaryhumanfibroblastfromaHDpatientcotransfectedwithplasmidforEGFP-Rab5transfectionandsiRNAduplexesagainstHAP40.Video6showsastriatalSTHdhQ7/7celltransfectedwithplasmidforEGFP-Rab5expressiononly,Video7showsthesameforanSTHdhQ7/111cell,andVideo8showsthisforanSTHdhQ111/111cell.Onlinesupplementalmaterialisavailableathttp://www.jcb.org/cgi/content/full/jcb.200509091/DC1.
Acknowledgments
WethankM.Shermanforfull-lengthHttcDNA,M.MatyashfortheratbraincDNAlibrary,M.MacDonaldforSTHdhQstriatalcelllines,andY.Trottierforinvaluableadvice,discussions,andforprovidinganti-Httantibodies.WethankK.Simons,J.Howard,L.Pelkman,andJ.Rinkforcriticalreadingofthemanuscript.TheCoriellCellRepositoriesisacknowledgedforprovidingthehumanprimaryfibroblastcelllines,andtheHarvardBrainTissueResourceCenterisacknowledgedforprovidinghumanpostmortembrainsamplesfromhealthyandHDgrades3and4donors.
ThisworkwassupportedbytheMaxPlanckSocietyandgrantsfromtheHumanFrontierScienceProgram(RG-0260/1999-M)andtheEuropeanUnion(HPRN-CT-2000-00081).A.Pal,F.Severin,andB.Lommerarelong-termMaxPlanckfellows.
References
Block-Galarza,J.,K.O.Chase,E.Sapp,K.T.Vaughn,R.B.Vallee,M.DiFiglia,andN.Aronin.1997.FasttransportandretrogrademovementofhuntingtinandHAP1inaxons.Neuroreport.8:22472251.
Chan,E.Y.,R.Luthi-Carter,A.Strand,S.M.Solano,S.A.Hanson,M.M.DeJohn,C.Kooperberg,K.O.Chase,M.DiFiglia,A.B.Young,etal.2002.Increasedhuntingtinproteinlengthreducesthenumberofpolyglutamine-inducedgeneexpressionchangesinmousemodelsofHuntington'sdisease.Hum.Mol.Genet.11:19391951.
Christoforidis,S.,H.M.McBride,R.D.Burgoyne,andM.Zerial.1999.TheRab5effectorEEA1isacorecomponentofendosomedocking.Nature.397:621625.
Cuif,M.H.,F.Possmayer,H.Zander,N.Bordes,F.Jollivet,A.Couedel-Courteille,I.Janoueix-Lerosey,G.Langsley,M.Bornens,andB.Goud.1999.CharacterizationofGAPCenA,aGTPaseactivatingproteinforRab6,partofwhichassociateswiththecentrosome.EMBOJ.18:17721782.
deRenzis,S.,B.Sonnichsen,andM.Zerial.2002.DivalentRabeffectorsregulatethesub-compartmentalorganizationandsortingofearlyendosomes.Nat.CellBiol.4:124133.
Engelender,S.,A.H.Sharp,V.Colomer,M.K.Tokito,A.Lanahan,P.Worley,E.L.Holzbaur,andC.A.Ross.1997.Huntingtin-associatedprotein1(HAP1)interactswiththep150Gluedsubunitofdynactin.Hum.Mol.Genet.6:22052212.
Eskelinen,E.L.,Y.Tanaka,andP.Saftig.2003.Attheacidicedge:emergingfunctionsforlysosomalmembraneproteins.TrendsCellBiol.13:137145.
Gasman,S.,Y.Kalaidzidis,andM.Zerial.2003.RhoDregulatesendosomedynamicsthroughDiaphanous-relatedForminandSrctyrosinekinase.Nat.CellBiol.5:195204.
Gauthier,L.R.,B.C.Charrin,M.Borrell-Pages,J.P.Dompierre,H.Rangone,F.P.Cordelieres,J.DeMey,M.E.MacDonald,V.Lessmann,S.Humbert,andF.Saudou.2004.HuntingtincontrolsneurotrophicsupportandsurvivalofneuronsbyenhancingBDNFvesiculartransportalongmicrotubules.Cell.118:127138.
Harjes,P.,andE.E.Wanker.2003.Thehuntforhuntingtinfunction:interactionpartnerstellmanydifferentstories.TrendsBiochem.Sci.28:425433.
Hattula,K.,andJ.Peranen.2000.FIP-2,acoiled-coilprotein,linksHuntingtintoRab8andmodulatescellularmorphogenesis.Curr.Biol.10:16031606.
Hoepfner,S.,F.Severin,A.Cabezas,B.Habermann,A.Runge,D.Gillooly,H.Stenmark,andM.Zerial.2005.ModulationofreceptorrecyclinganddegradationbytheendosomalkinesinKIF16B.Cell.121:437450.
Hoffner,G.,P.Kahlem,andP.J.Djian.2002.Perinuclearlocalizationofhuntingtinasaconsequenceofitsbindingtomicrotubulesthroughaninteractionwithbeta-tubulin:relevancetoHuntington'sdisease.J.CellSci.115:941948.
Landles,C.,andG.P.Bates.2004.HuntingtinandthemolecularpathogenesisofHuntington'sdisease.EMBORep.5:958963.
Lanzetti,L.,V.Rybin,M.G.Malabarba,S.Christoforidis,G.Scita,M.Zerial,andP.P.DiFiore.2000.TheEps8proteincoordinatesEGFreceptorsignallingthroughRacandtraffickingthroughRab5.Nature.408:374377.
Miaczynska,M.,L.Pelkmans,andM.Zerial.2004.Notjustasink:endosomesincontrolofsignaltransduction.Curr.Opin.CellBiol.16:400406.
Nakamura,N.,C.Rabouille,R.Watson,T.Nilsson,N.Hui,P.Slusarewicz,T.E.Kreis,andG.Warren.1995.Characterizationofacis-Golgimatrixprotein,GM130.J.CellBiol.131:17151726.
Nielsen,E.,F.Severin,J.M.Backert,A.A.Hyman,andM.Zerial.1999.Rab5regulatesmotilityofearlyendosomesonmicrotubules.Nat.CellBiol.1:376382.
Peters,M.F.,andC.A.Ross.2001.Isolationofa40-kDaHuntingtin-associatedprotein.J.Biol.Chem.276:31883194.
Rigamonti,D.,J.H.Bauer,C.De-Fraja,L.Conti,S.Sipione,C.Sciorati,E.Clementi,A.Hackam,M.R.Hayden,Y.Li,etal.2000.Wild-typehuntingtinprotectsfromapoptosisupstreamofcaspase-3.J.Neurosci.20:37053713.
Rigamonti,D.,S.Sipione,D.Goffredo,C.Zuccato,E.Fossale,andE.Cattaneo.2001.Huntingtin'sneuroprotectiveactivityoccursviainhibitionofprocaspase-9processing.J.Biol.Chem.276:1454514548.
Rink,J.,E.Ghigo,Y.Kalaidzidis,andM.Zerial.2005.Rabconversionasamechanismofprogressionfromearlytolateendosomes.Cell.122:735749.
Shevchenko,A.,M.Wilm,O.Vorm,andM.Mann.1996.Massspectrometricsequencingofproteinssilver-stainedpolyacrylamidegels.Anal.Chem.68:850858.
Shevchenko,A.,I.Chernushevich,W.Ens,K.G.Standing,B.Thomson,M.Wilm,andM.Mann.1997.Rapid‘denovo’peptidesequencingbyacombinationofnanoelectrospray,isotopiclabelingandaquadrupole/time-of-flightmassspectrometer.RapidCommun.MassSpectrom.11:10151024.
Sipione,S.,D.Rigamonti,M.Valenza,C.Zuccato,L.Conti,J.Pritchard,C.Kooperberg,J.M.Olson,andE.Cattaneo.2002.Earlytranscriptionalprofilesinhuntingtin-induciblestriatalcellsbymicroarrayanalyses.Hum.Mol.Genet.11:19531965.
Stenmark,H.,R.G.Parton,O.Steele-Mortimer,A.Lutcke,J.Gruenberg,andM.Zerial.1994.Inhibitionofrab5GTPaseactivitystimulatesmembranefusioninendocytosis.EMBOJ.13:12871296.
Sugars,K.L.,andD.C.Rubinsztein.2003.TranscriptionalabnormalitiesinHuntingtondisease.TrendsGenet.19:233238.
Tao,T.,andA.M.Tartakoff.2001.Nuclearrelocationofnormalhuntingtin.Traffic.2:385394.
Trettel,F.,D.Rigamonti,P.Hilditch-Maguire,V.C.Wheeler,A.H.Sharp,F.Persichetti,E.Cattaneo,andM.E.MacDonald.2000.DominantphenotypesproducedbytheHDmutationinSTHdhQ111striatalcells.Hum.Mol.Genet.9:27992809.
Trottier,Y.,D.Devys,G.Imbert,F.Saudou,I.An,Y.Lutz,C.Weber,Y.Agid,E.C.Hirsch,andJ.L.Mandel.1995.CellularlocalizationoftheHuntington'sdiseaseproteinanddiscriminationofthenormalandmutatedform.Nat.Genet.10:104110.
Ullrich,O.,H.Horiuchi,C.Bucci,andM.Zerial.1994.MembraneassociationofRab5mediatedbyGDP-dissociationinhibitorandaccompaniedbyGDP/GTPexchange.Nature.368:157160.
Zerial,M.,andH.M.McBride.2001.Rabproteinsasmembraneorganizers.Nat.Rev.Mol.CellBiol.2:107117.
