considered123408.com

Receptorsequestrationinresponsetoβ-arrestin-2phosphorylationbyERK1/2governssteady-statelevelsofGPCRcell-surfaceexpressionJustineS.
Paradisa,StevensonLya,b,lodieBlondel-Tepaza,JacobA.
Galana,c,1,AlexandreBeautraita,MarkG.
H.
Scottd,HervéEnslend,StefanoMarullod,PhilippeP.
Rouxa,c,2,andMichelBouviera,b,2aInstituteforResearchinImmunologyandCancer,UniversitédeMontréal,Montréal,QCH3C1J4,Canada;bDepartmentofBiochemistryandMolecularMedicine,UniversitédeMontréal,Montréal,QCH3CIJ4,Canada;cDepartmentofPathologyandCellBiology,UniversitédeMontréal,Montréal,QCH3CIJ4,Canada;anddInstitutCochin,InsermU1016,CNRSUMR8104,UniversitéParisDescartes,SorbonneParisCité,75014Paris,FranceEditedbySolomonH.
Snyder,JohnsHopkinsUniversitySchoolofMedicine,Baltimore,MD,andapprovedAugust13,2015(receivedforreviewMay5,2015)MAPKsareactivatedinresponsetoGprotein-coupledreceptor(GPCR)stimulationandplayessentialrolesinregulatingcellularprocessesdownstreamofthesereceptors.
However,verylittleisknownaboutthereciprocaleffectofMAPKactivationonGPCRs.
ToinvestigatepossiblecrosstalkbetweentheMAPKandGPCRs,weassessedtheeffectofERK1/2ontheactivityofseveralGPCRfamilymembers.
WefoundthatERK1/2activationleadstoareductioninthesteady-statecell-surfaceexpressionofmanyGPCRsbecauseoftheirintracellularsequestration.
Thissubcellularredistributionresultedinaglobaldampeningofcellresponsive-ness,asillustratedbyreducedligand-mediatedG-proteinactiva-tionandsecond-messengergenerationaswellasbluntedGPCRkinasesandβ-arrestinrecruitment.
ThisERK1/2-mediatedregula-toryprocesswasobservedforGPCRsthatcaninteractwithβ-arrestins,suchastype-2vasopressin,type-1angiotensin,andCXCtype-4chemokinereceptors,butnotfortheprostaglandinFreceptorthatcannotinteractwithβ-arrestin,implicatingthisscaf-foldingproteininthereceptor'ssubcellularredistribution.
Com-plementationexperimentsinmouseembryonicfibroblastslackingβ-arrestinscombinedwithinvitrokinaseassaysrevealedthatβ-arrestin-2phosphorylationonSer14andThr276isessentialfortheERK1/2-promotedGPCRsequestration.
Thispreviouslyuniden-tifiedregulatorymechanismwasobservedafterconstitutiveacti-vationaswellasafterreceptortyrosinekinase-orGPCR-mediatedactivationofERK1/2,suggestingthatitisacentralnodeinthetonicregulationofcellresponsivenesstoGPCRstimulation,actingbothasaneffectorandanegativeregulator.
β-arrestin|MAPK|Gprotein-coupledreceptor|internalization|cellsignalingTheERK/MAPKpathway(1,2)traditionallyhasbeenlinkedtotheactivationofreceptortyrosinekinases(RTKs)(3,4),butinrecentyears,Gprotein-coupledreceptor(GPCR)-medi-atedERK/MAPKactivationalsohasbeenshowntoplayimpor-tantroles(5).
SuchconvergenceontheERK/MAPKsignalingmodulegeneratesmultipleopportunitiesforregulatingcross-talkbetweenGPCRandRTKsignaling,andmanydistinctmo-lecularmechanismshavebeendescribed(6).
GPCRsactivatetheERK/MAPKpathwayviabothGprotein-dependent(canonical)and-independent(noncanonical)pathways.
Thecanonicalpath-waycaninvolvediverseGαand/orGβγsubunitsthatcauseERK1/2activationthroughvariousdownstreameffectors,suchasPI3KandexchangeproteindirectlyactivatedbycAMP(EPAC),aswellasPKAandPKC(7–11).
Amongthenoncanonicalmech-anisms,engagementofβ-arrestins(βarrs)scaffoldingtheERK/MAPKmodulehasattractedconsiderableattention(12–14).
Forbothcanonical(15)andnoncanonical(16,17)pathways,transactivationofRTKsdownstreamofsomeGPCRsalsohasbeenshowntocontributetoERK1/2regulation.
Particularly,di-rectrecruitmentofβarrstoseveralRTKs(18),includinginsulin-likegrowthfactorreceptor(IGFR)(19,20),insulinreceptor(IR)(21),andepidermalgrowthfactorreceptor(EGFR)(22–24),followingactivationofeitherRTKorGPCRalsoresultsinERK1/2activation,emphasizingthecentralroleofβarrsincon-trollingERK1/2activity.
ForclassicalGprotein-mediatedsignaling,severalprocessescontrollingthedurationofthesignalhavebeenelucidated,in-cludingbothhomologousandheterologousdesensitizations.
TheformerinvolvesphosphorylationoftheactivatedreceptorbyGPCRkinases(GRKs)leadingtothehigh-affinitybindingofβarrstotheactivatedreceptor,uncouplingofthereceptorfromGproteins,andsubsequentinternalization(25–27).
Heterolo-gousdesensitizationoccursuponactivationofsecondmessen-ger-dependentkinasessuchasPKAorPKCandmayormaynotbefollowedbyreceptorinternalization(28,29).
Although,asmentionedabove,ERK1/2activationnowisrecognizedasamainsignalingmoduleinbothGprotein-andβarr-dependentpathways,verylittleisknownaboutthepossibleadaptivemech-anismsresultingfromtheactivationofthesekinases.
Inonestudy(30),phosphorylationofβarr1,butnotβarr2,byERK1/2wasproposedtoinhibittheagonist-promotedtranslocationofβarr1totheβ2-adrenergicreceptor,preventingitsendocytosis.
Morerecently,ERK1/2-promotedphosphorylationofratβarr2wasshowntostabilizethecomplexbetweentheB2-bradykininSignificanceERK1/2areimportantGprotein-coupledreceptor(GPCR)sig-nalingeffectors,buttheirroleaspossibleGPCRregulatorsre-mainslargelyuncharted.
WereportthatERK1/2activationleadstothephosphorylationofβ-arrestin-2onSer14andThr276,promotingtheintracellularsequestrationofunligandedGPCRs.
ThissubcellularredistributionresultsinthedampeningofcellresponsivenesstoGPCRs'ligand-mediatedactivation,positioningERK1/2asbothadownstreameffectorandanega-tiveregulatorofGPCRs.
BecauseERK1/2alsoisstimulatedbyreceptortyrosinekinasesandisderegulatedinmanydiseases,andbecauseGPCRsrespondtoalargenumberofhormonesandneurotransmitters,thisnewlyuncoveredregulatoryprocessispoisedtoplayacentralroleincontrollingcellresponsivenessinhealthanddisease.
Authorcontributions:J.
S.
P.
,S.
L.
,.
B.
-T.
,J.
A.
G.
,A.
B.
,M.
G.
H.
S.
,H.
E.
,S.
M.
,P.
P.
R.
,andM.
B.
designedresearch;J.
S.
P.
,S.
L.
,.
B.
-T.
,J.
A.
G.
,A.
B.
,andP.
P.
R.
performedresearch;J.
S.
P.
,S.
L.
,.
B.
-T.
,J.
A.
G.
,A.
B.
,M.
G.
H.
S.
,H.
E.
,S.
M.
,P.
P.
R.
,andM.
B.
analyzeddata;andJ.
S.
P.
,P.
P.
R.
,andM.
B.
wrotethepaper.
Theauthorsdeclarenoconflictofinterest.
ThisarticleisaPNASDirectSubmission.
1Presentaddress:DepartmentofMolecularBiology,MassachusettsGeneralHospital,Boston,MA02114.
2Towhomcorrespondencemaybeaddressed.
Email:philippe.
roux@umontreal.
caormichel.
bouvier@umontreal.
ca.
Thisarticlecontainssupportinginformationonlineatwww.
pnas.
org/lookup/suppl/doi:10.
1073/pnas.
1508836112/-/DCSupplemental.
E5160–E5168|PNAS|PublishedonlineAugust31,2015www.
pnas.
org/cgi/doi/10.
1073/pnas.
1508836112DownloadedbyguestonJanuary22,2021receptorandβarr2,therebyslowingtherecyclingofthereceptortotheplasmamembraneafteragonist-promotedendocytosis(31).
AlthoughthesedatapointtopossiblefeedbackrolesfortheERK/MAPKpathwayinreceptortrafficking,theyalsoraiseimportantquestionsaboutthegeneralrolethatERK1/2acti-vationmayplayinGPCRactivity.
HereweinvestigatedtheinfluenceofERK1/2activationonGPCRresponsivenessusingthechemokineCXCtype-4receptor(CXCR4)asamodel.
WefoundthatactivationoftheERK/MAPKpathwayresultsindecreasedCXCR4signalinguponstimulationbyitsnativeagonist,CXCtype-12(CXCL12).
Thisreducedresponsivenessresultedfromthedirectphosphorylationofβarr2byERK1/2ontwoSer/Thrresidues,ultimatelyleadingtotheconstitutiveintracellularredistributionofthereceptor.
Thispreviouslyunidentifiedregulatorymechanismisnotre-strictedtoCXCR4andissharedbyotherGPCRsthatareabletointeractwithβarrs.
Ourstudythereforeunravelsanadaptivemechanismleadingtoageneraldampeningofcellresponsive-nesstoGPCRstimulationinresponsetotheactivationoftheERK/MAPKbyeitherRTKsorGPCRs.
ResultsActivationoftheERK/MAPKPathwayDecreasesCXCR4Agonist-InducedGiActivation.
Giproteinactivationdownstreamofago-nist(CXCL12)stimulationofCXCR4wasexaminedinthepresenceorabsenceofupstreammodulatorsoftheERK/MAPKpathway.
Gproteinactivationwasmonitoredusinganassaybasedonbioluminescenceresonanceenergytransfer(BRET)Fig.
1.
ERK/MAPKpathwayactivationreducesGiactivationandsecond-messengerproductionaswellasGRK2andβarr2translocationinresponsetoCXCR4activation.
(A)SchematicrepresentationoftheBRET-basedligand-inducedGiactivationassay.
(BandC)CXCL12-promotedGiactivationmeasuredbyBRETinHEK293TcellscotransfectedwithHA-CXCR4,Gαi1-RLucII,Gβ1,andGγ2-GFP10,without(Control)orwitheitherFlag-RasCAorFlag-RasDN(B)orFlag-MEKCAorFlag-MEKDN(C).
BRET400-GFP10betweenGαi1-RLucIIandGγ2-GFP10wasmeasuredaftertheadditionofcoel-400a,3minfollowingtheadditionofCXCL12.
Dataareexpressedasagonist-promotedBRET(ΔBRET;seeFig.
S1A).
(D)SchematicrepresentationoftheObelin-basedCa2+mobilizationassay.
(EandF)TheincreaseinCXCL12-promotedintracellularcalciummeasuredinHEK293TcellstransfectedwithObelin-Cherryintheabsence(Control)orpresenceoftheindicatedRasandMEKmutants.
Luminescencewasmeasuredeverysecondfor60saftertheinjectionofCXCL12orcalciumionophoreA23187.
Barsrepresenttheareaunderthecurvecalculatedfromthekineticscurves(Fig.
S1B).
(GandJ)SchematicrepresentationsoftheBRET-basedligand-inducedGRK2(G)andβarr2(J)translocation.
(H,I,K,andL)CXCL12-promotedBRETwasmeasuredinHEK293TcellstransfectedwithCXCR4-RLucIIandGRK2-GFP10(HandI)orβarr2-GFP10(KandL),intheabsence(Control)orpresenceoftheindicatedRasorMEKmutants.
BRET400-GFP10betweenCXCR4-RLucIIandGRK2-GFP10orβarr2-GFP10wasmeasuredaftertheadditionofcoel-400a,15minaftertheadditionofCXCL12.
Dataareexpressedasagonist-promotedBRET(ΔBRET;seeFig.
S1CandD).
Inallcases,datashownrepresentthemean±SEMofthreeindependentexperimentsandwerenormalizedto100%ofthecontrolcondition.
ExpressionofCAorDNformsofFlag-RasorFlag-MEKandthetotalERK1/2andERK1/2activationstatus[(p)ERK1/2],wasassessedbyimmunoblotting.
Representativeexperimentsareshownbelowthegraphs.
*P<0.
05;**P<0.
01;***P<0.
001;NS,notsignificant.
Paradisetal.
PNAS|PublishedonlineAugust31,2015|E5161CELLBIOLOGYPNASPLUSDownloadedbyguestonJanuary22,2021thatdetectstheagonist-promotedseparationofGαi1andGγ2inintactcells(Fig.
1A)(32).
CXCL12inducesarapiddecreaseinBRETsignalbetweenGαi1-91RLucIIandGFP10-Gγ2incellscoexpressingtheunmodifiedCXCR4andGβ1(Fig.
S1A).
Constitutivelyactive(CA)formsofH-Ras[Ras-CA(Ras-G12V)]andMAPK/ERKkinase1(MEK1)[MEK-CA(MEK-S218/222D)],whichpromoteERK1/2activation(33),diminishedCXCL12-inducedactivationofGicomparedwithcellsexpressinganemptyvectororthedominant-negative(DN)formsofH-Ras[Ras-DN(Ras-S17N)]andMEK1[MEK-DN(MEK-K97A)](Fig.
1BandC).
SubsequentlytheeffectofmodulatingtheERK/MAPKpathwayontheCXCL12-evokedcalciumresponsewasexaminedusingtheObelinbiosensor(34)asacalciumreporterinHEK293TcellsthatendogenouslyexpresslowlevelsofCXCR4(Fig.
1DandFig.
S1B).
TheexpressionofbothRas-CAandMEK-CAreducedCXCL12-inducedcalciummobilization,butRas-DNdidnot(Fig.
1EandF).
Incontrast,calciummobilizationtriggeredbythecalciumionophoreA23187wasunaffectedbymodulationoftheERK/MAPKpathway(Fig.
1EandF),indicatingthatCXCR4-dependentsignalingisselectivelyreducedbyactivatedERK1/2.
BecauseCXCR4-mediatedcalciummobilizationisaGi-dependentresponse(35,36),thesefindingsdemonstratethatactivationoftheERK/MAPKpathwaysignificantlyreducesGiactivationanddownstreamcalciumsignalinguponCXCR4stimulation.
Thelargerreductioninthecalciumresponsecom-paredwiththeGiactivationsignalmostlikelyreflectsthedif-ferentassaymodes(i.
e.
,thegreaternumberofsparereceptorsintheBRET-basedGiactivationassaythatrequiresexogenousCXCR4coexpression).
ActivationoftheERK/MAPKPathwayDecreasesAgonist-InducedGRK2andβarr2Translocation.
ToinvestigatefurthertheeffectofERK1/2activationonCXCR4signaling,weassessedCXCL12-promotedGRK2andβarr2translocationbyBRETusingCXCR4-RLucIIandeitherGRK2-GFP10orβarr2-GFP10constructs(Fig.
1GandJandFig.
S1CandD).
CoexpressionofRas-CAandMEK-CA,butnotRas-DNandMEK-DN,decreasedthetrans-locationofGRK2andβarr2followingCXCR4stimulation(Fig.
1H,I,K,andL).
Thus,inadditiontobluntingGiresponses,acti-vationoftheERK/MAPKpathwaygreatlyreducesGRK2andβarr2translocation.
ERK1/2ActivationDecreasesCXCR4LocalizationattheCellSurface.
ThedecreaseinGiactivationandthereducedtranslocationofGRK2andβarr2couldresulteitherfromagenerallossofthereceptor'sabilitytorespondproductivelytoCXCL12orfromareducednumberofreceptorsatthecellsurface.
Totestthelatterpossibilitydirectly,weassessedtheeffectofstimulatingtheERK/MAPKpathwayonCXCR4cell-surfaceexpressionbydual-flowcytometry(dualFACS).
Thisapproachallowsthemeasurementofcell-surfacereceptordensity(reflectedbyHAimmunoreactivity)asafunctionoftotalreceptornumber[reflectedbythevenusYFP(vYFP)fluorescence]inunpermeabilizedHEK293TcellsstablyexpressingHA-CXCR4-vYFP(Fig.
2A).
Usingthissystem,wefoundthatRas-CAandMEK-CA,butnotRas-DNandMEK-DN,ledtoa70%reductioninthenumberofCXCR4atthecellsurface(Fig.
2BandC).
Tovalidatethesefindingsfurther,weimagedthedistributionofCXCR4-YFPbyconfocalfluorescencemicroscopyinHeLacells.
AsshowninFig.
3A,intheabsenceofERK/MAPKacti-vation,CXCR4islocalizedmainlyatthecellsurface.
ExpressionofeitherRas-CA(Fig.
3B)orMEK-CA(Fig.
3D)resultedinasubstantiallossofcell-surfacereceptorsandtheappearanceofapunctateintracellularfluorescencesignalindicatingthatactiva-tionoftheERK/MAPKpathwaytriggerstheredistributionofCXCR4tointracellularvesicles.
TheDNformsofRasandMEKwerewithouteffectonCXCR4distribution(Fig.
3CandE).
SemiquantificationofthereceptorredistributionusingImageJsoftware(37)wasperformedbyselectingtransversalsectionsofthecellsthatallowdefinitionoftheplasmamembraneandintracellularcompartment(Fig.
3,Right).
Thesedata(Fig.
S2)supporttheFACSresultsdemonstratingthatactivationoftheERK/MAPKpathwaypromotesasignificantsubcellularre-distributionofCXCR4fromthecellsurfacetotheintracellularcompartment.
AcuteStimulationoftheERK/MAPKPathwayResultsinCXCR4IntracellularSequestrationandBluntedSignaling.
Next,weevalu-atedtheeffectsofpharmacologicalandendogenousagonistsoftheERK/MAPKpathway(38,39)onthesubcellulardistributionandsignalingcapacityofCXCR4.
Acutestimulationwithphor-bol12-myristate13-acetate(PMA)orEGFresultedinaweakerCXCL12-promotedrecruitmentofβarr2(Fig.
4A)andan20%reductionintheproportionofcell-surfacereceptorsintheab-senceofagoniststimulation(Fig.
4B).
Interestingly,thekineticsofsurface-receptorlossparalleledthatofERK1/2activation,beingtransientwithEGFstimulationandsustainedwithPMAstimulation,peakingat15or30min,respectively(Fig.
4C).
Stimulationofthevasopressintype-2receptor(V2R),aGPCRknowntoactivateERK1/2strongly(40),alsowasfoundtoin-duceCXCR4sequestrationtothesameextentasEGFandPMAstimulation(Fig.
4D).
Inallcases,inhibitionofERK1/2activa-tionwiththeselectiveMEK1/2inhibitorPD184352(41)restoredtheinitialproportionofcell-surfaceCXCR4confirmingthatERK1/2activationisfunctionallycorrelatedwiththelossofcell-surfacereceptors.
TheeffectofEGFstimulationoncell-surfacereceptorsalsowastestedintheT-lymphoidcelllineSUP-T1thatendogenouslyexpressesCXCR4(Fig.
4E).
Indeed,FACSanalysisusingamonoclonalantibodydirectedagainsttheextracellulardomainofthehumanCXCR4showedasignificantEGF-evokedlossofcell-surfaceCXCR4,indicatingthatthismodulationalsoisobservedwithendogenousreceptors.
Takentogether,thesedataindicatethatbothacuteandsustainedstimulationoftheERK1/2Fig.
2.
ActivationoftheERK/MAPKpathwayreducescell-surfaceexpressionofCXCR4.
(A)Schematicrepresentationofthereceptorcell-surfaceexpres-sionassay.
PM,plasmamembrane.
(BandC)Cell-surfaceCXCR4levelsweredetectedinHEK293TcellsstablyexpressingHA-CXCR4-vYFPtransfectedwithout(Control)orwiththeindicatedRasandMEKmutants.
Cellswerelabeledwithaprimaryanti-HAandsecondaryAlexa647-coupledchickenanti-mouseIgGantibodies,withoutpermeabilization.
Cell-surfaceexpres-sionofHA-CXCR4-vYFPintheabsenceofCXCL12stimulationwasmeasuredbydual-flowcytometry.
vYFPemissionrepresentsCXCR4totalexpression;AlexaFluor647emissionrepresentsCXCR4plasmamembraneexpression.
Therelativecell-surfaceexpression(theratioofAlexa:vYFPmeanemissions)iscalculatedfromtheYFP+cellpopulationandisexpressedasapercentageofthecontrolcondition.
Datashownrepresentthemean±SEMofthreeindependentexperiments.
ExpressionofCAorDNformsofFlag-RasorFlag-MEKandtotalERKandERK1/2activationstatus[(p)ERK1/2]wasassessedbyimmunoblotting.
Representativeexperimentsareshownbelowthegraphs.
**P<0.
01;NS,notsignificant.
E5162|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1508836112Paradisetal.
DownloadedbyguestonJanuary22,2021signalingpathwaypromoteCXCR4redistribution,but,aswouldbeexpected,long-lastingactivationresultsinmoredramaticef-fects,reachingasmuchas70%inthe48hfollowingtheexpres-sionofconstitutivelyactiveRasandMEKmutants.
TheERK/MAPKPathwayNegativelyModulatesSeveralGPCRs.
Ourresultsindicatethatcell-surfaceCXCR4densityisnegativelyregulateduponactivationofERK1/2byseveralagonistsandindifferentcelltypes.
TodeterminewhetherthisregulatorymechanismcanbegeneralizedtootherGPCRs,weassessedtheeffectofMEK-CAonthreeadditionalGPCRs,thevasopressin(V2R),theangiotensin(AT1R),andtheprostaglandin(FP)receptors.
Todoso,wemonitoredGprotein-mediatedsignaling,GRK2andβarr2recruitment,andcell-surfacedensity.
AsshowninFig.
5AandB,wefoundthatexpressionofMEK-CA,butnotMEK-DN,ledtoareductioninV2RandAT1Rsurfacedensity.
Thislossofcell-surfacereceptors(60%fortheV2Rand80%fortheAT1R)resultedindiminishedagonist-promotedrecruitmentofGRK2andβarr2aswellasabluntedsignalingresponse(cAMPproductionforV2RandCa2+mobilizationforAT1R).
Theextentofsignalreductiondifferedbetweenthetwore-ceptors,probablyreflectingdistinctlevelsofcouplingefficiencyandsparereceptorsforthesignalingpathwaysconsidered.
NosuchnegativeregulationwasobservedfortheFPreceptor(Fig.
5C,LeftandRight),indicatingthatonlycertainGPCRsarenegativelyregulatedbyERK1/2.
Interestingly,unlikeCXCR4,V2R,andAT1R,theagoniststimulationoftheFPreceptordoesnotpromoteβarr2translocationtothereceptor(42)(Fig.
5C,Center),suggestingthattheERK1/2-dependentregulatorypro-cessmayinvolveβarr2.
β-ArrestinIsRequiredfortheERK1/2-DependentCellularRedistributionofGPCRs.
Toconfirmtheroleofβarr2inthesubcellularre-distributionofCXCR4inresponsetoERK1/2activation,weanalyzedCXCR4cell-surfaceexpressioninmouseembryonicfibroblasts(MEFs)derivedfromWTandβarr1/2-KOanimals(43).
AswasconsistentwithourobservationsinHEK293T,HeLa,andSupT1cells,wefoundthatactivationofERK1/2ledtoasignificantdecreaseintheproportionofCXCR4atthecellsurfaceinWTMEFs(Fig.
6A).
Notably,thismodulationwaslostcompletelyinβarr1/2-KOMEFs(Fig.
6A).
Reintroducingβarr2orβarr1(Fig.
6B)inβarr1/2-KOMEFsbytransfectionrestoredtheERK1/2-promotedlossofcell-surfaceCXCR4,resultinginthesequestrationof30±3%and38±4%,respectively,com-paredwith52±4%fortheWTMEFs.
Coexpressionofbothβarr1andβarr2didnotleadtoastatisticallysignificantgreaterlossofreceptors(42±5%),indicatingthatbothβarr1andβarr2canpromotereceptorredistributionfollowingERK1/2activation.
Similartoobservationsinothercelltypes,expressionofRas-DNdidnotaffectCXCR4cell-surfacedensityinβarr1/2-KOMEFswhetherβarr2wasreintroducedornot(Fig.
6C),confirmingtheselectivityofactionofactiveRas.
Fig.
3.
ActivationoftheERK/MAPKpathwayleadstotheredistributionofCXCR4intointracellularvesicles(sequestration).
CXCR4-YFPlocalizationwasassessedbyfluorescenceconfocalmicroscopyinHeLacellstransfectedwithCXCR4-YFPwithout(A)orwith(B–E)theindicatedRas(BandC)andMEK(DandE)mutants.
(Scalebars,10μm.
)Intensityasafunctionofthedistancealongaselectedtransversalsection(redbar)wasassessedusingImageJsoft-ware.
Plotsgeneratedfromphase-contrastimageswereusedtoidentifythecoordinatesoftheplasmamembrane(PM)andtheintracellularcompartmentofthecell.
Thesecoordinatesareschematicallyin-dicatedabovetheYFPintensityplots.
Datashownarerepresentativeofmorethan20cellsobtainedfromthreeindependentexperiments.
Quantifica-tionsarepresentedinFig.
S2.
A.
U.
,arbitraryunits.
Paradisetal.
PNAS|PublishedonlineAugust31,2015|E5163CELLBIOLOGYPNASPLUSDownloadedbyguestonJanuary22,2021β-ArrestinPhosphorylationIsRequiredforERK/MAPK-InducedCXCR4Redistribution.
TodetermineifCXCR4subcellularredistributionresultsfromthedirectphosphorylationofβarr2byERK1/2,wefirsttestedifactivatedERK1couldphosphorylatepurifiedβarr2andβarr1invitro.
AsshowninFig.
6D,ERK1inducedarobustincorporationof[32P]inbothβarr1andβarr2,withβarr2appearingtobeabettersubstrate.
Quantificationof32Pincorporationre-vealedsaturablephosphorylationwithstoichiometryapproaching2(mol/mol)forβarr1and2.
5forβarr2,consistentwiththepresenceofatleasttwophosphorylationsites.
Wethenanalyzedphospho-βarr2byMStoidentifytheERK1phosphorylationsites.
MS/MSsequencingrevealedthepresenceofβarr2-derivedphosphopep-tidescontainingThr276(Fig.
S3A).
ThispositioncorrespondstooneofthepotentialERK1/2consensussites(Ser/Thr-Pro)(44)inβarr2(Ser14,Ser264,andThr276)(Fig.
S3B).
Unphosphorylatedpeptidescontainingtheotherpotentialsitesweredetectedatmuchlowerabundance,butSer14hadbeenidentifiedpreviouslyasbeingphosphorylatedinJurkatcells(45).
Itshouldbenotedthatallputativesitesareevolutionarilyconservedinβarr2(Fig.
S3C),butonlySer14andThr276arefoundinbothβarr1andβarr2(Fig.
S3D),underscoringthelikelihoodoftheirimportanceforβarrfunction.
AsshowninFig.
S3B,insilicoanalysisofresidues'accessibilityusingmolecularmodelingofβarr2fromitscrystalstructure[ProteinDataBank(PDB)IDcode3P2D]withtheMolecularOperatingEnvironmentsoftware(MOE)(46)suggeststhatinthebasalinactivatedstateThr276isexposedtothesolventandconsequentlycouldbeeasilyaccessibletoproteinkinases.
Incontrast,Ser14isburiedinthecoreoftheproteinandprobablywouldrequirereorganizationofthemolecularenvironmenttobephosphorylated.
Todeterminetheroleofβarr2phosphorylationinERK1/2-dependentGPCRregulation,Ser14,Ser264,andThr276werereplacedindividuallybyunphosphorylatablealanineresidues(S14A,S264A,andT276A).
Thentheseβarr2mutantsweretestedfortheirabilitytorestoreCXCR4modulationuponERK1/2ac-tivationinβarr1/2-KOcells.
AsshowninFig.
6E,incontrasttoβarr2-WTorβarr2-S264A,theexpressionofβarr2-S14Aorβarr2-T276AfailedtorestoreanysignificantERK1/2-promotedCXCR4sequestration,suggestingthatphosphorylationofbothresidueshasaroleinthismechanism.
Theseresultsaresupportedfurtherbytheobservationthattheexpressionofamutantformofβarr2harboringphosphomimeticresiduesatpositionsSer14andThr276(S14D/T276D;S/T2D)wassufficienttodecreaseCXCR4cell-surfacelocalizationintheabsenceofactivation(Fig.
6F).
BothS14A/T267A(S/T2A)andS/T2Dβarr2mutantformsmaintainedtheirabilitytoberecruitedtoCXCR4uponactivationofthereceptorbyitsnaturalligandCXCL12(Fig.
S4),confirmingtheirfunctionality.
Together,theseresultssuggestthatβarr2phosphorylationatSer14andThr276viaanERK1/2-dependentmechanismpromotesCXCR4redistribution,thusdampeningGPCRsignaling.
DiscussionOurdatademonstratethatERK1/2activationleadstothese-questrationofGPCRsintointracellularvesicles,functionallyresultinginthelossofcellresponsivenesstoGPCRstimulation.
ThisphenomenonappearstobeconservedamongasubclassofGPCRsthatrecruitβarr2inresponsetoagonistactivation.
In-deed,ourresultsshowthatERK1/2activationalsoaffectsthedistributionofV2RandAT1R(43),butnotFP,whichun-dergoesendocytosisinaβarr-independentmanner(42).
Con-sistently,wefoundthatreceptorredistributionrequiresthephosphorylationofβarr2byERK1/2,unravelinganuncharac-terizedmolecularmechanismforcrosstalkbetweenERK1/2andGPCRsignaling.
TheERK1/2-dependentintracellularredistributionofGPCRswasobservedinresponsetoconstitutiveactivationoftheMAPKpathway(byactiveformsofMEK1andH-Ras),stimulationofERK1/2-activatingGPCRs(V2R)orRTKs(EGFR),orpharmaco-logicalactivationoftheERK/MAPKpathway(byPMA).
ReceptorFig.
4.
AcutepharmacologicalactivationoftheERK/MAPKpathwaybluntsCXCR4signalingbyreducingreceptorcell-surfacelocalization.
(A)CXCL12-promotedβarr2translocationismeasuredbyBRETinHEK293TcellstransfectedwithCXCR4-RLucIIandβarr2-GFP10.
CellswerepretreatedwithPD184352for30minbeforestimulationwithEGFfor15minorwithPMAfor30min.
BRET400-GFP10betweenCXCR4-RLucIIandβarr2-GFP10wasmeasuredaftertheadditionofcoel-400a,15minfollowingtheadditionofCXCL12.
(B)CXCR4cell-surfaceexpressionlevelswereassessedbydual-flowcytometryinHEK293TcellsstablyexpressingHA-CXCR4-vYFPthatwerepretreatedornotwithPD184352for30minbeforestimulationwithEGFfor15minorwithPMAfor30min.
(C)KineticsofCXCR4cell-surfacelevelreductionandERK1/2activationfollowingEGForPMAtreatments.
(D)KineticsofthereductionofCXCR4cell-surfacelevelsandERK1/2activationinHEK293TcellstransfectedwithHA-CXCR4-vYFPandmyc-V2RpretreatedornotwithPD184352for30minbeforestimulationwithAVPfortheindicatedtime.
(E)CXCR4cell-surfaceexpressionlevelinSupT1cellstreatedornotwithEGFfor15min.
Inallcases,datashownrepresentthemean±SEMofatleastthreeindependentexperi-mentsandarenormalizedto100%ofthecontrolconditions.
TotalERK1/2andERK1/2activationstatus[(p)ERK1/2]wereassessedbyimmunoblotting.
*P<0.
05;**P<0.
01;***P<0.
001;NS,notsignificant.
E5164|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1508836112Paradisetal.
DownloadedbyguestonJanuary22,2021sequestrationresultedinalossofcellresponsivenesstoGPCRag-onists,asassessedbyG-proteinactivation,second-messengergener-ation,andGRKandβarrrecruitment.
Thisphenomenonrepresentsaheterologousdesensitizationprocessthathadnotbeenappreciateduntilnow.
BecausemanyGPCRsactivateERK1/2,itcouldbehy-pothesizedthatthismechanismalsocontributestothehomologousdesensitizationofsuchreceptors.
PKCactivationwithdifferentphorbolesters,includingPMA,wasshownpreviouslytopromotetheendocytosisofseveralGPCRs,includingCXCR4(47–49).
TheseresultshavebeenattributedtoapotentialPKC-mediatedphosphorylationofthereceptoritself,butadirectlinkbetweenreceptorphosphoryla-tionanditsPMA-inducedendocytosishadnotbeenestablished.
OurdatademonstratethatPMA-stimulatedCXCR4endocytosisrequiresERK1/2activation,suggestingthatthemechanismelu-cidatedinthepresentstudymayexplaintheeffectofPMAontheendocytosisofmanyGPCRs.
TheERK1/2-inducedsequestrationofGPCRsintocytoplas-micvesiclescouldresultfromeitheracceleratedconstitutiveendocytosisorreducedrecyclingfromendosomestotheplasmamembrane.
Giventheclassicalroleofβarrintheendocytosisprocess,thefirstpossibilityappearsmorelikely.
ThisideaissupportedbytheobservationthatERK1/2activationledtothesequestrationofV2R,areceptorthatdoesnotundergosignifi-cantrecyclingfollowingendocytosis(50).
Whetherthemodula-tionofconstitutiveendocytosisresultsfromanincreasedaffinityofERK1/2-phosphorylatedβarrsforunligandedreceptorsorforotherproteinsinvolvedintheendocyticprocessremainstobedetermined.
Interestingly,oneoftheputativeERK1/2phos-phorylationsitesidentifiedinβarr2,Ser14,islocatedincloseproximitytotwolysineresidues,Lys11andLys12.
Theselysinesbindthephosphatemoietiesofthephosphorylatedreceptorsandcanfacilitatethetransitiontothehigh-affinityreceptor–βarrbindingstate(51).
ThereforeitcouldbehypothesizedthatthephosphorylationofSer14maymimicthephosphorylatedre-ceptor,thuspromotingahigh-affinityinteractionwithaninac-tive,nonphosphorylatedreceptor.
ThecausallinkbetweenERK1/2activationandsteady-statereceptorredistributionissupportedbythecompleteinhibitionofthePMA-andEGF-inducedCXCR4sequestrationbytheMEK1/2inhibitorPD184352.
Moreover,thekineticsofCXCR4redistributionparalleledthatofERK1/2activation,withbothERK1/2activationandreceptorsequestrationbeingsustainedforPMAbuttransientforEGFstimulation.
Therescueexperi-mentsusingnonphosphorylatableformsofhumanβarr2orβarr2harboringphosphomimeticresiduesfurtherestablishesthelinkbetweenβarr2phosphorylationandreceptorsequestration.
ThelackofredistributionofFP,areceptorthatcannotrecruitβarr2,alsoisconsistentwiththecentralroleofβarr2inERK1/2-pro-motedreceptorrelocalization.
OurresultssuggesttheinvolvementoftwoputativeERK1/2phosphorylationsitesinβarr2,Ser14andThr276,becausetheirmutationtoalanineresiduesblockedERK1/2-promotedCXCR4sequestration,whereastheirreplacementbyphosphomimeticaspartateresiduesmimickedtheeffectofERK1/2activation.
Oneofthesesites,Thr276,wasconfirmedtobeadirectERK1phosphorylationsitebyourMS/MSanalysis,whereasSer14previouslyhadbeenshowntobephosphorylatedinalarge-scalephosphoproteomicstudy(45),thussupportingthenotionthatERK1/2phosphorylatesβarr2onthesetworesidues.
Althoughthesephosphorylationsitesareconservedinβarr1,whetherβarr1phosphorylationonthesesitesalsocontributestoreceptorredistributionremainstobedemonstrateddirectly.
Using2Dtrypticphosphopeptidemapping,apreviousstudyidentifiedanotherserineresidue,Ser412,asadirectERK1/2phosphory-lationsiteinratβarr1(30).
However,becauseSer412isnotconservedineitherratorhumanβarr2,itisunlikelythatthissiteplaysanymajorroleinthemechanismsdescribedhere.
Morerecently,Ser178wasproposedasaphosphorylationsiteforERK1/2inratβarr2andwasfoundtomodulatethedynamicsofFig.
5.
ActivationoftheERK/MAPKpathwaybluntssignalingbyreducingthecell-surfaceexpressionofV2RandAT1R.
cAMPaccumulation(GFP10-EPAC-RLucII)(A,Left),calciummobilization(Obelin-Cherry)(BandC,Left),GRK2recruitment(GRK2-GFP10)(AandB,CenterLeft),andβarr2recruitment(βarr2-GFP10)(AandB,CenterRightandC,Center)weremea-suredinHEK293TcellstransfectedwithV2R(A),AT1R(B),orFP(C)alongwiththeappropriatebio-sensorsintheabsence(Control)orpresenceoftheindicatedMEKmutants.
CellularactivitiesweremeasuredbyBRET,asinFig.
1,aftertheadditionofAVP(A),AngII(B),orPGF-2(C).
Cell-surfaceex-pression(A–C,Right)wasmeasuredbydual-flowcytometryintheabsenceofligandstimulationasinFig.
2.
Datashownarethemean±SEMofatleastthreeindependentexperimentsandwerenor-malizedtothecontrolcondition(100%).
*P<0.
05;***P<0.
001;NS,notsignificant.
Paradisetal.
PNAS|PublishedonlineAugust31,2015|E5165CELLBIOLOGYPNASPLUSDownloadedbyguestonJanuary22,2021βarr2withintheendosomalcompartment(31).
However,thismechanismisdistinctfromtheonewedescribe,becauseratSer178isnotconservedinhumanβarr2.
Inthepresentstudy,weidentifiedtwophosphorylationsitesforERK1/2inβarr2thatplayakeyroleinregulatingsteady-statecell-surfaceexpressionofdifferentGPCRs.
GiventhelargeproportionofGPCRsthatinteractwithβarrs,theobservedre-ceptorsequestrationpromotedbytheERK1/2-mediatedphos-phorylationofβarrcouldhaveamajorimpactonthetonicregulationofcellresponsivenesstoavarietyofhormonalstimu-lations.
BecauseERK1/2activationisknowntobederegulatedinmanypathologicaldiseases,includingseveralcancers,thismecha-nismcouldhavefar-reachingconsequencesforourunderstandingofdifferentpathophysiologicalprocesses.
MaterialsandMethodsPlasmids.
Thefollowingplasmidsweredescribedpreviously:Flag-MEK1-S218/222D(52),Gβ1andGFP10-Gγ2(53),CXCR4-RLucII(54),CXCR4-YFP(55),Obelin-Cherry(56),Gαi1-91RLucII,HA-CXCR4,andHA-CXCR4-vYFP(36),GFP10-EPAC-RLucIIandHA-AT1R(57),HA-FP(58),andHA-V2R(59).
HA-FP-vYFPandFP-RLucIIwereprovidedbyTerrenceHébert,McGillUniversity,Montreal.
Theβarr2-GFP10constructwasbuiltbyreplacingtheRLucIIsequencefromthepreviouslypublishedβarr2-RLucII(36).
vYFPwasintroducedintopIRESP-HAtogeneratethepIRESP-HA-vYFPvector.
TheFP,V2R,orspAT1R(sp:signalpep-tideMKTIIALSYIFCLVFAattheNterminus)sequencewasintroducedintopIRESP-HA-vYFPtogeneratepIRESP-HA-FP-vYFP,pIRESP-HA-V2R-vYFP,orpIRESP-HA-AT1R-vYFP.
Humanβarr1andβarr2cDNAwereobtainedfromthecDNAResourceCenterandwereclonedinpCMV-Tag3B(Stratagene)toobtainmyc-taggedproteins,andH-RasandMEK1cDNAwereclonedinpCMV-Tag2A(Stratagene)toobtainFlag-taggedproteins.
Allmutants(βarr2-S14A,-T126A,-S264A,-S/T2D;H-Ras-G12V,-S17N;andMEK1-K97A)weregeneratedusingtheQuikChange(Stratagene)methodologyandwereverifiedbysequencing.
ReagentsandAntibodies.
CXCL12waspurchasedfromCedarlane.
Argininevasopressin(AVP),PGF-2,andangiotensinII(AngII)werepurchasedfromSigma-Aldrich.
A23187waspurchasedfromTocris.
TheMEK1/2inhibitorPD184352waspurchasedfromSelleckChemicals.
TheERK1/2activatorsEGFandPMAwerepurchasedfromInvitrogenandFisherScientific,respectively.
Anti-ERK1/2,anti–phospho-ERK1/2(T202/Y204),andanti-βarr1/2antibodieswerepurchasedfromCellSignalingTechnologies.
Anti-mycandanti-Flag,antibodieswerepurchasedfromSigma-Aldrich.
Anti-humanCXCR4anti-bodywaspurchasedfromBioLegend.
Anti-HAantibody(HA-11)waspur-chasedfromCovance.
Anti-mouseAlexaFluor647secondaryantibodywaspurchasedfromInvitrogen.
AllsecondaryHRP-conjugatedantibodiesusedforimmunoblottingwerepurchasedfromChemicon.
AllcellculturereagentswerefromWisent.
CellCulture.
HEK293T,HeLa,βarr1/2-KOMEFs,andWTMEFswereculturedinDMEMsupplementedwith10%(vol/vol)FBS,100IU/mLpenicillin,and100μg/mLstreptomycin.
βarr1/2-KOmiceweregeneratedasdescribed(43),andestablishedMEFcultureswerepreparedaccordingtothe3T3protocolofTodaroandGreen(60).
ForHEK293THA-CXCR4-vYFPstablecelllines,aYFP+polyclonalcellpopulationwasobtainedbyflowcytometrycellsortingandwasmaintainedinDMEMsupplementedwith10%(vol/vol)FBS,100IU/mLpeni-cillin,100μg/mLstreptomycin,and3μg/mLpuromycin.
Sup-T1cellswerecul-turedinRPMI-1640mediumsupplementedwith10%(vol/vol)FBS,100IU/mLpenicillin,and100μg/mLstreptomycin.
Transfections.
Thedaybeforetransfections,HEK293TorHEK293THA-CXCR4-vYFPcellswereseededinsix-wellplatesatadensityof600,000cellsperwell.
Transienttransfectionswereperformedusinglinearpolyethylenimine25kDa(PEI)(Polysciences,Inc.
)astransfectionagent,ataPEI:DNAratioof3:1.
Cellsweremaintainedincultureforthenext48h,andBRETexperimentswerecarriedoutsubsequently.
ForthetransfectionofMEFs,cellswereseededinsix-wellplatesatadensityof600,000cellsperwellandonthenextdayweretransientlytransfectedusingFuGENEHD(Promega),ataFuGENEHD:DNAratioof3:1.
Cellsweremaintainedincultureforthenext24h,andFACSexperimentswerecarriedoutsubsequently.
BRETAssays.
BRETassayswereperformedasdescribedpreviously(36,61).
ForBRET480-YFP,proteinswerefusedtoRLucasenergydonorandYFPasac-ceptor.
Coelenterazine-h(coel-h)(NanoLightTechnology)wasusedastheluciferasesubstratetogeneratelightwithamaximalemissionpeakatFig.
6.
ERK1/2-dependentβarr2phosphorylationonS14andT276inducesCXCR4intracellularsequestration.
(A–C)CXCR4cell-surfaceexpressionwasassessedbydual-flowcytometryinWTMEFsorβarr1/2-KOMEFstransientlytransfectedwithHA-CXCR4-vYFPintheabsence(Control)orpresenceoftheindicatedRasmutants(A)andwithorwithoutβarr1and/orβarr2comple-mentation(BandC).
(D)Phosphorylationofβarrswasperformedwithrecombinantbovineβarr1orβarr2asasubstrateandrecombinantactiveERK1inthepresenceof[γ32P]-ATPfor15min.
ThegelwasstainedwithCoomassie,and32PincorporationwasquantifiedusingaPhosphorImager(Left)orduringkineticsusingscintillationcounting(Right).
Thedatashownarerepresentativeofthreeindependentexperiments.
(EandF)CXCR4cell-surfaceexpressionwasassessedbydual-flowcytometryinβarr1/2-KOMEFstransientlytransfectedwithHA-CXCR4-vYFP,βarr2WT,ortheindicatedβarr2mutants,withorwithoutRasCA.
Datashownrepresentthemean±SEMofatleastthreeindependentexperimentsandwerenormalizedtothecontrolcondition(100%).
Expressionofβarr2WTandmutantswereassayedbyimmunoblotting.
*P<0.
05;**P<0.
01;***P<0.
001;NS,notsignificant.
E5166|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1508836112Paradisetal.
DownloadedbyguestonJanuary22,2021480nm,allowingYFPexcitation(maximumat488nm).
ForBRET400-GFP10,proteinswerefusedtoRLucII(donor)andGFP10(acceptor).
Coelenterazine-400a(coel-400a)(Biotium)wasusedastheluciferasesubstratetogeneratelightwithamaximalemissionpeakat400nm,allowingGFP10excitation(maximumat400nm).
BRETwasmeasuredusingaMithrasLB940Multi-modeMicroplateReader(BertholdTechnologies)equippedwitheitheraBRET480-YFPfilterset(donor480±20nmandacceptor530±20nmfilters)oraBRET400-GFP10filterset(donor400±70nmandacceptor515±20nmfilters).
Briefly,cellswereseededthedayaftertransfectioninapoly-L-ornithine(Sigma-Aldrich)–pretreated96-wellmicroplate(CulturePlate;PerkinElmerInc.
)andwerestarvedovernightinDMEMforatleast8hafterseeding.
Thedayoftheexperiment,cellswereincubatedwithHBSS(Invitrogen)complementedwith0.
1%BSA(HBSS/BSA)andstimulatedwithligands(CXCL12:200nM;AVP:100nM;AngII:1μM;PGF-2:1μM)for15min.
Whereindicated,cellswerepretreatedwithPD184352(10μM)for30minbeforestimulationwithEGF(25ng/mL)for15minorPMA(100ng/mL)for30min(maximalresponse)orfortheindicatedtimes(timecourse).
BRETvalueswerecollectedaftertheadditionoftheluciferasesubstrateatafinalconcentrationof2.
5μM.
BRETsignalsweredeterminedastheratioofthelightemittedbyacceptor(YFPorGFP10)overdonor(RLucorRLucII).
Theligand-promotedBRETsignal(ΔBRET)wascalculatedbysubtractingtheBRETvaluesobtainedinthevehiclecondi-tionfromthevaluesmeasuredwithligand.
ΔBRETwasexpressedasaper-centageoftheΔBRETobtainedinthecontrolconditioninwhichtheERK/MAPKpathwaywasnotmodulatedbyeitheramutantoratreatment.
CalciumMeasurements.
HEK293TcellsweretransientlytransfectedwithObelin-Cherryusedasacalciumreporter(34).
Thedayaftertransfection,cellswereseededinapoly-L-ornithine–pretreated96-wellmicroplatetoobtainduplicateplatesforligandorionophorestimulation.
Thedayoftheex-periment,cellswerewashedwithHBSSandincubatedfor2hinthedark,atroomtemperature,with1μMcoelenterazinecp(Biotium)dilutedinHBSS/BSA.
Luminescencemeasurementswererecordedfor60sonaFlexStationII(MolecularDevices)afterligandinjection(CXCL12:200nM;A23187:10μM;AngII:1μM;PGF-2:1μM).
Kineticswerenormalized,settingthemaximalresponseoftheligandat100%.
Bargraphsweregeneratedfromtheareaunderthecurveofeachkinetic.
FlowCytometry.
Briefly,48haftertransfection,HEK293THA-CXCR4-vYFPcellswerestarvedovernightinDMEM.
Whereindicated,cellswerepretreatedwithPD184352beforeEGForPMAstimulation.
Thenthecellswererinsedwithice-coldHBSSandresuspendedinHBSS/BSA.
Cell-surfacereceptorswerelabeledoniceusingamonoclonalanti-HAantibodyfollowedbyananti-mouseAlexaFluor647secondaryantibody.
SUP-T1cellswerestarvedovernightandthenwerestimulatedwithEGFfor15min,rinsedwithice-coldHBSS,resuspendedinHBSScontainingFcblocksolution24G2(anti-CD16/32-FcRγ),andstainedusingananti-humanCXCR4antibody(PerCP/Cy5.
5anti-humanCD184clone12G5;BioLegend)for45minonice.
Cellswerewashed,resuspended,andanalyzedthroughanLSRIIflowcytometer(BDBio-sciences)settodetectYFPandAlexaFluor647orPerCP/Cy5.
5.
DataanalysiswasperformedusingBDFACSDivasoftware.
ImageAcquisitionandAnalysis.
Briefly,HeLacellsweretransfectedandthenseededintoaμ-Dish(IBIDI;no.
81156).
Twodaysaftertransfection,epi-fluorescenceconfocalmicroscopyimagesfromsingleconfocalsections(0.
7μm)wereacquiredatroomtemperaturewithanLSM510Metalaser-scanningconfocalmicroscope(Zeiss)usingaPlanApochromat100*oil-immersionobjective.
Allimageswereacquiredusingidenticalparameters.
ThecolocalizationbetweenthemembraneandtheYFPsignalwasdeterminedusingImageJsoftware(NIH).
First,alinecrossingaYFP+cellwasdrawnonthephase-contrastimage,andthecoordinatesoftheplasmamembraneandtheintracellularcompartmentsweredeterminedfromthegrayintensityprofilealongthisline.
ThenthesamelinewastransposedtotheYFPimage,andmeanfluorescenceintensityprofilesalongthislinewereobtained.
Usingthecoordinatesoftheplasmamembraneandtheintracellularcompartment,totalYFPintensitieswerecalculated,andboththeplasmamembraneandinnercellintensitieswereexpressedasapercentageoftotalintensityalongthesectionineachcondition.
InVitroPhosphorylationAssays.
Phosphorylationassayswereperformedinvitrowithrecombinantbovineβarr1orβarr2(giftofStéphaneLaporte,McGillUniversity,Montreal)asasubstrate(3μgperassay)andrecombinantactiveERK1(100ngperassay;Millipore;no.
14-439).
ThereactionproductsweresubjectedtoSDS/PAGE,and32PincorporationwasquantifiedusingaBio-RadphosphorimagerandMultiGaugesoftware.
Tocalculatethephosphate:sub-strateratio,[32P]incorporationinβarr1orβarr2wasmeasuredusingascin-tillationcounterandwasnormalizedtothespecificactivityof[32P]ATPinthereaction.
Immunoblotting.
CelllysatesweresubjectedtoSDS/PAGEon10%acrylamidegelsandelectroblottedtoPVDFmembranes.
Blockingandprimaryandsec-ondaryantibodyincubationsofimmunoblotswereperformedinTris-bufferedsaline/Tween20[10mMTris(pH7.
4),150mMNaCl,and0.
1%Tween20]supplementedwith5%(wt/vol)dryskimmilkpowder.
Theprimaryantibodiesanti-HA,anti-myc,anti-Flag,anti–phospho-ERK1/2,andanti-βarr1/2wereusedaccordingtothemanufacturers'instructions.
HRP-conjugateddonkeyanti-rabbitandanti-mouseIgGswereusedatadilutionof1:5,000,andimmuno-reactivebandsweredetectedusingenhancedchemiluminescence.
SamplePreparationandTrypsinDigestionforLC-MS/MS.
Thegelsliceswerewashedtwotimeswith100%acetonitrilefor10minandwithH2Ofor20min.
Thenthesamplesweredriedusingaspeedvacuum.
Proteinswerereducedin50mMammoniumbicarbonate(pH7.
5)containing10mMDTTandwereincubatedfor1hat56°C,followedbyalkylationwith55mMiodoaceta-midefor1hat25°Cinthedark.
Thesampleswerewashedanddriedagainasdescribedabove.
Proteinsweredigestedovernightwithsequencing-grademodifiedtrypsin(enzyme:proteinratioof1:50)at37°C.
Peptidesfromthegelsliceswereextractedusing100%acetonitrile.
ThesampleswereevaporatedtodrynessinaSpeedVac.
LiquidChromatographyandNanoflowLC-MS/MS.
PeptideswereinjectedintoananoflowHPLCsystem(Eksigent;ThermoFisherScientific)foronlinereversed-phasechromatographicseparation.
Peptideswereloadedina5-mm-longtrapcolumn(i.
d.
300μm)inbufferA(0.
2%formicacid)andwereseparatedinan18-cm-longfusedsilicacapillaryanalyticalcolumn(i.
d.
150μm),bothpackedwith3-μm200-MagicC18AQreverse-phasematerial(Michrom).
PeptideswereelutedbyanincreasingconcentrationofbufferB(0.
2%formicacidinacetonitrile)from5–40%over100min.
Afterthegradientelution,thecol-umnwaswashedwith80%bufferBandre-equilibratedwith5%bufferB.
Peptideswereelutedintothemassspectrometerataflowrateof600nL/min.
Thetotalruntimewas70min,includingsampleloadingandcolumnconditioning.
Peptideswereanalyzedusinganautomateddata-dependentacquisitiononanLTQ-OrbitrapEliteorQExactivemassspectrometer.
ForLTQ-OrbitrapEliteanalysis,eachMSscanwasacquiredataresolutionof240,000FWHM(at400m/z)formassrange300–2,000withthelockmassoptionenabled(m/z:445.
120025)andwasfollowedbyupto12MS/MSdata-dependentscansonthemostintenseionsusingcollision-inducedactivation(CID).
Automaticgaincontrol(AGC)targetvaluesforMSandMS/MSscansweresetto1e6(maximumfilltime,500ms)and1e5(maximumfilltime,50ms),respectively.
Theprecursorisolationwindowwassetto2withCIDnormalizedcollisionenergyof35;thedynamicexclusionwindowwassetto60s.
ForQExactiveanalysis,eachMSscanwasacquiredataresolutionof70,000FWHM(atm/z200)formassrange300–2,000,andthe12mostin-tensepeakswerefragmentedinthehigher-energycollisionaldissociation(HCD)collisioncellwithnormalizedcollisionenergyof25%.
AGCtargetvaluesforMSandMS/MSscansweresetto1e6,andthedynamicexclusionwindowwassetto15s.
MSDataAcquisition.
MSdatawereanalyzedusingMascotDistillersoftwareandweresearchedagainsttheUniProtKBproteinknowledgebaserelease2014_11database(www.
expasy.
org)containing547,085entries.
Fordatabasesearching,theenzymespecificitywassettotrypsinwiththemaximumnumberofmissedcleavagessetto2.
Theprecursormasstolerancewassetto10ppm.
Searchcriteriaincludedastaticmodificationofcysteineresiduesof+57.
0214Da,avariablemodificationof+15.
9949Datoincludepotentialoxidationofmethionines,andamodificationof+79.
966onserine,threonine,ortyrosinefortheidentificationofphosphorylation.
IonscorecutoffandsignificancethresholdPvaluesweresetto25and0.
05,respectively.
DataAnalysisandStatistics.
AllgraphsweregeneratedandanalyzedusingGraphPadPrism(GraphPadSoftwareInc.
).
Datapresentedarethemeans±SEMofaleastthreeindependentexperiments.
StatisticalanalysiswasperformedusingeitherapairedStudent'sttestwhentwovalueswerecomparedorone-wayANOVAwithDunnett'sposthoctestwhenmorethantwovalueswerecomparedwiththecontrol.
*P<0.
05,**P<0.
01,***P<0.
001.
ACKNOWLEDGMENTS.
WethankMoniqueLagacéforcriticalreadingofthemanuscript;StéphaneLaporteforprovidingrecombinantβarr1/2;MartinAudetandViktoriyaLukashovafromtheM.
B.
laboratoryandDarlènePétrinandEugénieGoupilfromTerrenceHébert'slaboratoryforprovidingpIRESP-HA-V2R-vYFPParadisetal.
PNAS|PublishedonlineAugust31,2015|E5167CELLBIOLOGYPNASPLUSDownloadedbyguestonJanuary22,2021andpIRESP-HA-AT1R-vYFPandHA-FP-vYFPandFP-RLucIIplasmids,respec-tively;andDr.
RobertLefkowitzforprovidingβarr1/2-KOMEFs.
ThisworkwassupportedbyCanadianInstitutesofHealthResearchGrantsMOP10501(toM.
B.
)andMOP123408(toP.
P.
R.
),agrantfromtheHumanFrontierSci-enceProgram(toP.
P.
R.
),agrantfromtheCancerResearchFoundationFrance(ARC)(toM.
G.
H.
S.
),grantsfromtheLigueContreleCancerComitédel'Oise(toH.
E.
andM.
G.
H.
S.
),andagrantfromtheFondationpourlaRechercheMédicale(toS.
M.
).
M.
B.
holdsaCanadaResearchChairinSignalTransductionandMolecularPharmacology.
P.
P.
R.
holdsCanadaResearchChairinCellSignalingandProteomics.
1.
SegerR,KrebsEG(1995)TheMAPKsignalingcascade.
FASEBJ9(9):726–735.
2.
RoskoskiR,Jr(2012)ERK1/2MAPkinases:Structure,function,andregulation.
PharmacolRes66(2):105–143.
3.
SchlessingerJ(2002)Ligand-induced,receptor-mediateddimerizationandactivationofEGFreceptor.
Cell110(6):669–672.
4.
McKayMM,MorrisonDK(2007)IntegratingsignalsfromRTKstoERK/MAPK.
Oncogene26(22):3113–3121.
5.
MarinissenMJ,GutkindJS(2001)G-protein-coupledreceptorsandsignalingnet-works:Emergingparadigms.
TrendsPharmacolSci22(7):368–376.
6.
DelcourtN,BockaertJ,MarinP(2007)GPCR-jacking:FromanewrouteinRTKsig-nallingtoanewconceptinGPCRactivation.
TrendsPharmacolSci28(12):602–607.
7.
KolchW,etal.
(1993)ProteinkinaseCalphaactivatesRAF-1bydirectphosphoryla-tion.
Nature364(6434):249–252.
8.
KochWJ,HawesBE,AllenLF,LefkowitzRJ(1994)DirectevidencethatGi-coupledreceptorstimulationofmitogen-activatedproteinkinaseismediatedbyGbetagammaactivationofp21ras.
ProcNatlAcadSciUSA91(26):12706–12710.
9.
CrespoP,XuN,SimondsWF,GutkindJS(1994)Ras-dependentactivationofMAPkinasepathwaymediatedbyG-proteinbetagammasubunits.
Nature369(6479):418–420.
10.
LevS,etal.
(1995)ProteintyrosinekinasePYK2involvedinCa(2+)-inducedregulationofionchannelandMAPkinasefunctions.
Nature376(6543):737–745.
11.
Lopez-IlasacaM,CrespoP,PelliciPG,GutkindJS,WetzkerR(1997)LinkageofGprotein-coupledreceptorstotheMAPKsignalingpathwaythroughPI3-kinasegamma.
Science275(5298):394–397.
12.
LuttrellLM,etal.
(1999)Beta-arrestin-dependentformationofbeta2adrenergicreceptor-Srcproteinkinasecomplexes.
Science283(5402):655–661.
13.
DeFeaKA,etal.
(2000)beta-arrestin-dependentendocytosisofproteinase-activatedreceptor2isrequiredforintracellulartargetingofactivatedERK1/2.
JCellBiol148(6):1267–1281.
14.
LuttrellLM,Gesty-PalmerD(2010)Beyonddesensitization:Physiologicalrelevanceofarrestin-dependentsignaling.
PharmacolRev62(2):305–330.
15.
DaubH,WeissFU,WallaschC,UllrichA(1996)RoleoftransactivationoftheEGFreceptorinsignallingbyG-protein-coupledreceptors.
Nature379(6565):557–560.
16.
NomaT,etal.
(2007)Beta-arrestin-mediatedbeta1-adrenergicreceptortrans-activationoftheEGFRconferscardioprotection.
JClinInvest117(9):2445–2458.
17.
GaletC,AscoliM(2008)Arrestin-3isessentialfortheactivationofFynbythelutei-nizinghormonereceptor(LHR)inMA-10cells.
CellSignal20(10):1822–1829.
18.
HupfeldCJ,OlefskyJM(2007)RegulationofreceptortyrosinekinasesignalingbyGRKsandbeta-arrestins.
AnnuRevPhysiol69:561–577.
19.
LinFT,DaakaY,LefkowitzRJ(1998)beta-arrestinsregulatemitogenicsignalingandclathrin-mediatedendocytosisoftheinsulin-likegrowthfactorIreceptor.
JBiolChem273(48):31640–31643.
20.
Oligny-LongpréG,etal.
(2012)Engagementofβ-arrestinbytransactivatedinsulin-likegrowthfactorreceptorisneededforV2vasopressinreceptor-stimulatedERK1/2activation.
ProcNatlAcadSciUSA109(17):E1028–E1037.
21.
DalleS,RickettsW,ImamuraT,VollenweiderP,OlefskyJM(2001)Insulinandinsulin-likegrowthfactorIreceptorsutilizedifferentGproteinsignalingcomponents.
JBiolChem276(19):15688–15695.
22.
HinneyA,etal.
(2003)Melanocortin-4receptorgene:Case-controlstudyandtrans-missiondisequilibriumtestconfirmthatfunctionallyrelevantmutationsarecom-patiblewithamajorgeneeffectforextremeobesity.
JClinEndocrinolMetab88(9):4258–4267.
23.
KimJ,AhnS,GuoR,DaakaY(2003)RegulationofepidermalgrowthfactorreceptorinternalizationbyGprotein-coupledreceptors.
Biochemistry42(10):2887–2894.
24.
HollenbergMD(1995)Tyrosinekinase-mediatedsignaltransductionpathwaysandtheactionsofpolypeptidegrowthfactorsandG-protein-coupledagonistsinsmoothmuscle.
MolCellBiochem149-150:77–85.
25.
BenovicJL,StrasserRH,CaronMG,LefkowitzRJ(1986)Beta-adrenergicreceptorki-nase:Identificationofanovelproteinkinasethatphosphorylatestheagonist-occu-piedformofthereceptor.
ProcNatlAcadSciUSA83(9):2797–2801.
26.
PitcherJA,FreedmanNJ,LefkowitzRJ(1998)Gprotein-coupledreceptorkinases.
AnnuRevBiochem67:653–692.
27.
BouvierM,etal.
(1988)Removalofphosphorylationsitesfromthebeta2-adrenergicreceptordelaysonsetofagonist-promoteddesensitization.
Nature333(6171):370–373.
28.
BenovicJL,etal.
(1985)Phosphorylationofthemammalianbeta-adrenergicreceptorbycyclicAMP-dependentproteinkinase.
Regulationoftherateofreceptorphos-phorylationanddephosphorylationbyagonistoccupancyandeffectsoncouplingofthereceptortothestimulatoryguaninenucleotideregulatoryprotein.
JBiolChem260(11):7094–7101.
29.
HausdorffWP,etal.
(1989)Phosphorylationsitesontwodomainsofthebeta2-adrenergicreceptorareinvolvedindistinctpathwaysofreceptordesensitization.
JBiolChem264(21):12657–12665.
30.
LinFT,MillerWE,LuttrellLM,LefkowitzRJ(1999)Feedbackregulationofbeta-arrestin1functionbyextracellularsignal-regulatedkinases.
JBiolChem274(23):15971–15974.
31.
KhouryE,NikolajevL,SimaanM,NamkungY,LaporteSA(2014)Differentialregu-lationofendosomalGPCR/β-arrestincomplexesandtraffickingbyMAPK.
JBiolChem289(34):23302–23317.
32.
GalésC,etal.
(2006)Probingtheactivation-promotedstructuralrearrangementsinpreassembledreceptor-Gproteincomplexes.
NatStructMolBiol13(9):778–786.
33.
CarrièreA,etal.
(2008)OncogenicMAPKsignalingstimulatesmTORC1activitybypromotingRSK-mediatedraptorphosphorylation.
CurrBiol18(17):1269–1277.
34.
CampbellAK,DormerRL(1975)Studiesonfreecalciuminsidepigeonerythrocyte'ghosts'byusingthecalcium-activatedluminescentprotein,obelin.
BiochemSocTrans3(5):709–711.
35.
Vila-CoroAJ,etal.
(1999)ThechemokineSDF-1alphatriggersCXCR4receptordi-merizationandactivatestheJAK/STATpathway.
FASEBJ13(13):1699–1710.
36.
QuoyerJ,etal.
(2013)PepducintargetingtheC-X-Cchemokinereceptortype4actsasabiasedagonistfavoringactivationoftheinhibitoryGprotein.
ProcNatlAcadSciUSA110(52):E5088–E5097.
37.
SchneiderCA,RasbandWS,EliceiriKW(2012)NIHImagetoImageJ:25yearsofimageanalysis.
NatMethods9(7):671–675.
38.
MindenA,etal.
(1994)DifferentialactivationofERKandJNKmitogen-activatedproteinkinasesbyRaf-1andMEKK.
Science266(5191):1719–1723.
39.
FrostJA,GeppertTD,CobbMH,FeramiscoJR(1994)Arequirementforextracellularsignal-regulatedkinase(ERK)functionintheactivationofAP-1byHa-Ras,phorbol12-myristate13-acetate,andserum.
ProcNatlAcadSciUSA91(9):3844–3848.
40.
CharestPG,Oligny-LongpréG,BoninH,AzziM,BouvierM(2007)TheV2vasopressinreceptorstimulatesERK1/2activityindependentlyofheterotrimericGproteinsig-nalling.
CellSignal19(1):32–41.
41.
EnglishJM,CobbMH(2002)PharmacologicalinhibitorsofMAPKpathways.
TrendsPharmacolSci23(1):40–45.
42.
GoupilE,etal.
(2012)BiasingtheprostaglandinF2αreceptorresponsestowardEGFR-dependenttransactivationofMAPK.
MolEndocrinol26(7):1189–1202.
43.
KohoutTA,LinFS,PerrySJ,ConnerDA,LefkowitzRJ(2001)beta-Arrestin1and2differentiallyregulateheptahelicalreceptorsignalingandtrafficking.
ProcNatlAcadSciUSA98(4):1601–1606.
44.
RouxPP,BallifBA,AnjumR,GygiSP,BlenisJ(2004)Tumor-promotingphorbolestersandactivatedRasinactivatethetuberoussclerosistumorsuppressorcomplexviap90ribosomalS6kinase.
ProcNatlAcadSciUSA101(37):13489–13494.
45.
HornbeckPV,etal.
(2012)PhosphoSitePlus:Acomprehensiveresourceforin-vestigatingthestructureandfunctionofexperimentallydeterminedpost-trans-lationalmodificationsinmanandmouse.
NucleicAcidsRes40(Databaseissue):D261–D270.
46.
2014,C.
C.
G.
I.
MolecularOperatingEnvironment(MOE)software.
Availableatwww.
chemcomp.
com.
AccessedJuly30,2015.
47.
SignoretN,etal.
(1997)PhorbolestersandSDF-1inducerapidendocytosisanddownmodulationofthechemokinereceptorCXCR4.
JCellBiol139(3):651–664.
48.
LilesWC,HunterDD,MeierKE,NathansonNM(1986)ActivationofproteinkinaseCinducesrapidinternalizationandsubsequentdegradationofmuscarinicacetylcholinereceptorsinneuroblastomacells.
JBiolChem261(12):5307–5313.
49.
KelleherDJ,PessinJE,RuohoAE,JohnsonGL(1984)Phorbolesterinducesde-sensitizationofadenylatecyclaseandphosphorylationofthebeta-adrenergicre-ceptorinturkeyerythrocytes.
ProcNatlAcadSciUSA81(14):4316–4320.
50.
InnamoratiG,SadeghiHM,TranNT,BirnbaumerM(1998)AserineclusterpreventsrecyclingoftheV2vasopressinreceptor.
ProcNatlAcadSciUSA95(5):2222–2226.
51.
GimenezLE,etal.
(2012)Roleofreceptor-attachedphosphatesinbindingofvisualandnon-visualarrestinstoGprotein-coupledreceptors.
JBiolChem287(12):9028–9040.
52.
CarriereA,etal.
(2011)ERK1/2phosphorylateRaptortopromoteRas-dependentactivationofmTORcomplex1(mTORC1).
JBiolChem286(1):567–577.
53.
GalésC,etal.
(2005)Real-timemonitoringofreceptorandG-proteininteractionsinlivingcells.
NatMethods2(3):177–184.
54.
BusilloJM,etal.
(2010)Site-specificphosphorylationofCXCR4isdynamicallyregu-latedbymultiplekinasesandresultsindifferentialmodulationofCXCR4signaling.
JBiolChem285(10):7805–7817.
55.
PercherancierY,etal.
(2005)Bioluminescenceresonanceenergytransferrevealsli-gand-inducedconformationalchangesinCXCR4homo-andheterodimers.
JBiolChem280(11):9895–9903.
56.
vanderWesthuizenET,BretonB,ChristopoulosA,BouvierM(2014)Quantificationofligandbiasforclinicallyrelevantβ2-adrenergicreceptorligands:Implicationsfordrugtaxonomy.
MolPharmacol85(3):492–509.
57.
ZimmermanB,etal.
(2012)Differentialβ-arrestin-dependentconformationalsignal-ingandcellularresponsesrevealedbyangiotensinanalogs.
SciSignal5(221):ra33.
58.
GoupilE,etal.
(2010)AnovelbiasedallostericcompoundinhibitorofparturitionselectivelyimpedestheprostaglandinF2alpha-mediatedRho/ROCKsignalingpath-way.
JBiolChem285(33):25624–25636.
59.
OakleyRH,LaporteSA,HoltJA,CaronMG,BarakLS(2000)Differentialaffinitiesofvisualarrestin,betaarrestin1,andbetaarrestin2forGprotein-coupledreceptorsdelineatetwomajorclassesofreceptors.
JBiolChem275(22):17201–17210.
60.
TodaroGJ,GreenH(1963)Quantitativestudiesofthegrowthofmouseembryocellsincultureandtheirdevelopmentintoestablishedlines.
JCellBiol17:299–313.
61.
MercierJF,SalahpourA,AngersS,BreitA,BouvierM(2002)Quantitativeassessmentofbeta1-andbeta2-adrenergicreceptorhomo-andheterodimerizationbybio-luminescenceresonanceenergytransfer.
JBiolChem277(47):44925–44931.
E5168|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1508836112Paradisetal.
DownloadedbyguestonJanuary22,2021