aminoapche

MajorsourceofantigenicpeptidesfortheMHCclassIpathwayisproducedduringthepioneerroundofmRNAtranslationSebastienApchera,ChrysoulaDaskalogiannia,FabriceLejeuneb,BénédicteManouryc,GabrielaImhoosa,LeaHeslopc,1,andRobinFhraeusa,2aCiblesTherapeutiques,InstitutNationaldelaSantéetdelaRechercheMédicaleU940,InstitutdeGénétiqueMoléculaire,UniversitéParis7,HpitalSt.
Louis,75010Paris,France;bInstitutPasteurdeLille,59019Lille,France;andcInstituteCurie,75245ParisCedex05,FranceEdited*byPeterCresswell,YaleUniversitySchoolofMedicine,NewHaven,CT,andapprovedJune3,2011(receivedforreviewMarch15,2011)TheMHCclassIantigenpresentationpathwayallowstheimmunesystemtodistinguishbetweenselfandnonself.
Despiteextensiveresearchontheprocessingofantigenicpeptides,littleisknownabouttheirorigin.
Here,weshowthatmRNAscarryingprematurestopcodonsthatpreventtheproductionoffull-lengthproteinsviathenonsense-mediateddecaypathwaystillproduceamajorityofpeptidesubstratesfortheMHCclassIpathwaybyanoncanonicalmRNAtranslationprocess.
Blockingtheinteractionofthetrans-lationinitiationfactoreIF4Ewiththecapstructuresuppressesthesynthesisoffull-lengthproteinsbuthasonlyalimitedeffectontheproductionofantigenicpeptides.
TheseresultsrevealanessentialcellbiologicalfunctionforaclassoftranslationproductsderivedduringthepioneerroundofmRNAtranslationandwillhaveimportantimplicationsforunderstandinghowtheimmunesystemdetectscellsharboringpathogensandgeneratestolerance.
MHCclassI-restrictedantigenpresentation|pioneerroundoftranslationPresentationofendogenouspeptidesonMHCclassImole-culesservestoallowtheimmunesystemtodepleteTcellsthatreactagainstself-antigensandtodetectpathogen-infectedcells.
Muchisknownaboutthemechanismsthatcontroltheprocessingofantigenicpeptidesubstrates(1–6),butlittleisknownaboutthemechanismsthatgoverntheproductionofpeptidematerialfortheMHCclassIpathway.
Whentheas-sumptionthatdegradationof"old"functionalproteinswasthesolebasisforantigenicpeptidematerialfortheendogenousMHCclassIpathwaywaschallengedbynewmodelssuggestingthatpeptidematerialmight,infact,bederivedfromalternativesources,suchasdefectiveribosomalproducts(7),theinterestinndingtheoriginofthesepeptidesandwhatgovernstheirsyn-thesiswasinitiated.
Observations,suchasthatantigenicpeptidescanbederivedfromcryptictranslationproducts(8),thatthepresenceofmRNAandnotthefull-lengthproteindeterminesantigenpresentation,andthatvirusestargetmRNAtranslationtoevadetheimmunesystem(9–11),haveledtoashiftoffocusfromproteindegradationtomRNAtranslationasbeingthecriticalprocess.
However,effortstotaketheseearlierobserva-tionsfurtherandidentifywhatsuchpolypeptidesmightactuallybe,orwheretheycomefrom,havebeenhamperedbythefactthatnotranslationproductsderivedfromanORFofanmRNAthatcandirectlybelinkedtoantigenpresentationhavebeenidentiedsofar.
Cap-dependentmRNAtranslationisinitiatedbythecap-bindingeIF4E,whereascap-independenttranslationischarac-terizedbythedirectrecruitmentofthe40Ssubunittointernalribosomeentrysites(12,13).
Eventhoughmethionine(AUG)isbyfarthemostcommonlyusedinitiationcodontogeneratefull-lengthproteins,othercodonscanalsobeused,mostnotablyleucine(CUG)(14,15).
Inadditiontothesetwotypesoftranslationinitiationprocessesthatgiverisetofull-lengthpro-teins,therearetranslationeventsthatarelesswellunderstood.
NewlysynthesizedmRNAsaresubjecttoqualitycontrolscan-ning,anddetectedfaultstriggerthenonsense-mediateddecay(NMD)pathway(16).
Bestdescribedisthedetectionofapre-terminationcodon(PTC)atleast50ntupstreamofthe3′ofthesplicejunctionbythescanningribosome.
RibosomesthatarestalledataPTCrecruitup-frameshiftprotein1(Upf1)(17)toformalinkwithUpf2andUpf3oftheexonjunctioncomplex(EJC).
ThispreventsthemRNAfromtranslatingfull-lengthproteinsandsinglesoutthefaultymRNAfortransporttopro-cessingbodiesanddegradation(18).
ThepioneerroundoftranslationthatprecedesNMDrequirestheribosometobeinitiatedaroundthecap-bindingcomplex(CBC),includingCBP20andCBP80(19),butitisnotknowniftheribosomesproducepeptidesduringthisscanning.
Here,weprovideevi-dencethatpeptidesproducedduringthepioneerroundofmRNAtranslationprovidesamajorsourceofantigenicpeptidesubstratesfortheMHCclassIpathway.
ResultsandDiscussionAntigenicPeptidesfortheMHCClassI-RestrictedPathwayAreDerivedfromNoncanonicalmRNATranslation.
Ithasbeendeter-minedthatrapidproteasome-dependentdegradationofpeptidesubstratesisrequiredfortheinitialstepsinantigenicpeptideprocessing,andithasalsobeenproposedthatthisinvolvestheubiquitin-dependentproteasomalpathway(1).
TheIκBαproteinisakeyregulatoroftheNF-κBpathwayandisrapidlydegradedviaubiquitin-dependent26Sproteasomeactivityinresponsetodifferentextracellularstimuli,includingTNF-α.
ThisismediatedbytheIκBkinase(IKK),whichphosphorylatesIκBαatSer32andSer36andleadstoitssubsequentubiquitinationandim-mediatedegradation(20)(Fig.
S1A).
Hence,inresponsetoex-tracellularstimuli,theIκBαproteinrequirestwoindependentmodicationsforbecomingtargetedfordegradation.
ThismakesIκBαasuitablereporterproteintotestifrapiddegradationoffull-lengthproperlyfoldedfunctionalproteinsviatheubiquitin-dependentproteasomalpathwayisasourceofMHCclassIantigenicpeptides.
WefusedtheSIINFEKL(SL8)immunodo-minantepitopefromchickenovalbumin(Ova)totheCterminusofWTIκBα(IκBα-wt-SL8)andtoanIκBαthatcarriesmutationsinboththeIKKphosphorylationsites(IκBα-mut-SL8)(Fig.
S1B).
TheseconstructswerecoexpressedwiththeMHCclassIKbmoleculeinhumanH1299cells,andthepresentationoftheAuthorcontributions:S.
A.
,B.
M.
,F.
L.
,andR.
F.
designedresearch;S.
A.
,C.
D.
,F.
L.
,G.
I.
,andL.
H.
performedresearch;S.
A.
,C.
D.
,F.
L.
,andR.
F.
analyzeddata;andS.
A.
,C.
D.
,andR.
F.
wrotethepaper.
Theauthorsdeclarenoconictofinterest.
*ThisDirectSubmissionarticlehadaprearrangededitor.
1Presentaddress:TumourImmunologyLaboratory,DivisionofImmunology,QueenslandInstituteofMedicalResearch,Brisbane,Queensland4006,Australia.
2Towhomcorrespondenceshouldbeaddressed.
E-mail:robin.
fahraeus@inserm.
fr.
Thisarticlecontainssupportinginformationonlineatwww.
pnas.
org/lookup/suppl/doi:10.
1073/pnas.
1104104108/-/DCSupplemental.
11572–11577|PNAS|July12,2011|vol.
108|no.
28www.
pnas.
org/cgi/doi/10.
1073/pnas.
1104104108DownloadedbyguestonJanuary10,2021SL8epitopewasdeterminedusingtheB3ZCD8+T-cellre-portersystem(21).
TreatmentwithTNF-αresultedinarapiddegradationoftheendogenousIκBαaswellastheIκBα-wt-SL8reporterprotein,whereastheIκBα-mut-SL8wasunaffected,asexpected(Fig.
1A).
Following30minoftreatmentwithTNF-α,thelevelsofIκBα-wt-SL8werereducedto50%,indicatingthathalfoftheproteinamount,includingnewlysynthesizedproteinsduringthistimeperiod,hadbeentargetedfordegradation.
However,upto240minaftertreatmentwithTNF-α,weob-servedonlyasmallnonspecicincreaseinSL8presentationfromtheIκBα-wt-SL8constructcomparedwiththeIκBα-mut-SL8orthecontrolproteininwhichtheSL8epitopehadbeenfusedtoGFP(GFP-SL8)(Fig.
1B).
Weconcludethattherapiddegradationofproperlyfoldedfull-lengthproteinsviathe26Sproteasomepathwaydoesnotnecessarilyconstituteasourceforantigenicpeptides,andwenextstartedtoinvestigatewhichtypeoftranslationproductsmightbeusedinstead.
Alternativecodons,particularlyCUG(leucine),canbeusedtoinitiatetranslation(14,22).
UpstreamoftheMHCclassIepitopeMBP(79–87)inthemyelinbasicprotein(MBP)(23)ORF,thereareleucinecodonsatpositions16(CUG),39(CUU),and73(CUG)andonecodondown-streamatposition110(CUG)(Fig.
1C,Left).
Metabolicpulselabelinginthepresenceofproteasomeinhibitors(MG132)showedthatsilentmutationsinanyofthesecodons(i.
e.
,dif-ferentcodonforthesameaminoacid)donotchangethesyn-thesisorstabilityofthefull-lengthMBP,orofasecondMBPproductthatisderivedfromanin-frameAUGatposition22(Fig.
1D,Upper).
Thecodon22productissynthesizedatasimi-larrateasthefull-lengthMBPbutishighlyunstableandisdifculttodetectunlesscellsaretreatedwithproteasomein-hibitor(Fig.
S2A).
Interestingly,silentchangesincodon39fromaCUUtoCUGresultedinthedetectionofahighlyunstabletranslationproduct(codon39product)thatcouldbeobservedinthepresenceofproteasomeinhibitor(Fig.
1D,UpperandFig.
S2B).
ByusingCD8+lacZTcellstowardtheMBP(79–87)epi-topepresentedonKkmolecules,wenoticedthatchangingcodon39fromCUUtoCUGresultedinanincreaseof35%inan-tigenpresentation(Fig.
1D,Lower).
IntroducingAUGorGUG(Val),whichcanalsobeusedfortranslationinitiation(14)atposition39,resultedinexpressionofthecodon39productatdifferentlevelsthatcorrelatedwiththepresentationoftheMBP(79–87)epitope(Fig.
1EandFig.
S2B).
Othercodonchangesatposition39thathavebeenshowntoinitiatetranslationeitherFig.
1.
AntigenicpeptidesareproducedbynoncanonicalmRNAtranslationinitiation.
(A)Westernblotshowingtherapiddegradationofproperlyfoldedfull-lengthIκBα-wt-SL8andendogenousIκBαinH1299cellsfollowingtreatmentwithTNF-α(Fig.
S1).
TheexpressionoftheIκBα-mut-SL8,whichcarriesmutations(SertoAla)intheIKKphosphorylationsitesatSer32andSer36,isnotaffectedbyTNF-αtreatment.
Actinisshownasaloadingcontrol.
W.
B.
,Westernblot.
(B)EffectonSL8epitopepresentationafterIκBα-wt-SL8degradationinH1299cellsexpressingexogenousKbafterTNF-αtreatmentforindicatedtimeperiods,duringwhichcellswerewashedandincubatedwiththeB3ZreporterCD8+Tcells.
ControlcellswereexposedtoexogenousSL8peptidefor16h(peptidecontrol).
ThedifferenceinantigenpresentationbetweenIκBα-wt-SL8andIκBα-mut-SL8afterTNF-αtreatmentisnotsignicant.
N.
S.
,notsignicant.
(C)Cartoonillustratingin-dicatedcodonpositionsintheMBPorinachickenOvaconstruct.
WTcodons(Upper,black)andcodonchanges(Lower,red)areshown.
(D,Upper)Autoradiographof[35S]me-thioninemetabolicpulselabeling,followedbyMBPimmunoprecipitationinthepresenceof25μMproteasomeinhibitorMG132.
Thetwoup-perbandsarederivedfromAUGcodonsatpo-sition1(fulllength)orcodon22(MBP22AUG).
Changingcodon39fromCUUtoCUGresultsintheappearanceofatranslationproduct(Codon39)withoutaffectingthesynthesisofthetwoAUG-initiatedpolypeptides.
Thecorrespondingeffectofindicatedcodonchangesonthepre-sentationoftheMBP(79–87)epitopeinH1299cellscoexpressingtheKkmoleculeisshownbe-lowusingMBP(79–87)reporterCD8+Tcells.
F.
L.
,fulllength.
(E)EffectonMBP(79–87)antigenpresentationafterindicatedcodonchangesintheMBPORF.
(F,Upper)AutoradiographofmetabolicpulselabelingfollowedbyOvaIPinthepresenceofproteasomeinhibitorsofcellsexpressingOvaconstructswithindicatedsilentpointmutations.
(Lower)CorrespondingT-cellassayusingB3ZreporterTcells.
Antigenpre-sentationwascarriedoutinhumanH1299cellsexpressingthemouseKbMHCclassImolecule.
TheresultsarerepresentativeofatleastthreeindependentexperimentsplusSD.
**PApcheretal.
PNAS|July12,2011|vol.
108|no.
28|11573IMMUNOLOGYDownloadedbyguestonJanuary10,2021reducedorhadnoeffectonantigenpresentation.
Deletionofthesecondin-frameAUGpreventedtheproductionofthecodon22product,asexpected,withoutaffectingantigenpresentationfromtheWTorfromthecodon39CUGconstructs(MBPΔ22AUGandMBP39CUG-Δ22AUG),demonstratingthatthehalf-lifeofapolypeptide,perse,doesnotdetermineantigenpresentation(Fig.
1E).
Eventhoughitisnotpossibletoconcludethatthecodon39-dependenttranslationproductisindeedprocessedfortheMHCclassIpathway,theseresultsunderlinethefactthatthemRNAtranslationinitiationeventsthatgoverntheproductionoffull-lengthproductsandantigenicpeptidesaredistinguishable.
SimilartotheMBPORF,therearenumerousCUGcodonsupstreamanddownstreamoftheSL8epitopeintheOvamRNA(Fig.
1C,Right).
Whenweintroducedsilentmutations(CUGtoCUC)incodons114and124(relativeto+1),whicharelocatedupstreamoftheSL8epitope,weobservednochangesintheexpressionofthefull-lengthprotein(Fig.
1F,UpperandFig.
S2C)butasignicantdecreaseinSL8antigenpresentation(Fig.
1F,Lower).
However,whenwecarriedoutthesamemutationsatpositions173and182,whicharealsolocatedupstreamoftheepitope,weobservednosignicantchangesinSL8antigenpresentation,indicatingthatnotallCUGcodonswithintheOvaORFhavethesameimpactonantigenicpeptideproduction(Fig.
S2D).
Fromtheseresults,weconcludethattheeffectonantigenpresentationcausedbysilentchangeswithinanORFdoesnotdependontheepitopeortheMHCclassIhaplotype.
LittleispresentlyknownaboutthesynthesisofpeptidesduringtheearlyscanningeventsoftheimmaturemessengerRNAproteincomplex(mRNP),andwearecurrentlyinvestigatingthecontextofcodon39initiation;itisplausiblethatitdependsonRNAstructurestoalargerextentcomparedwiththesequencerequirementsthatarecharacteristicofthecanonicalAUGini-tiation.
ThiscouldofferanexplanationastowhynotallCUGcodonswithintheOvaORF,forexample,havethesameimpactintermsofantigenicpeptideproductionandwhysomealter-nativeinitiationcodons,suchasGUG,ACG,UUG,andAUU,havedifferenteffects.
ProductionofAntigenicPeptidesandProductionofFull-LengthProteinsfromanmRNAAreTemporallyDifferent.
Undernormalconditions,apeptidepresentedtogetherwithanendogenousMHCclassIKbmoleculehasahalf-lifeofover6h(24),makingitdifculttostudythedynamicsofantigenicpeptideproductionvs.
antigenpresentation.
ThemouseKbmolecule,togetherwithasyntheticSL8peptide,wasfoundtohaveahalf-lifeofapproxi-mately10minonthesurfaceofhumancells(H1299),pre-sumablyattributabletospecies-relatedincompatibilitiesinthisMHC–peptideassemblycomplex(Fig.
2A).
Thiscircumstance,togetherwithtransfectionofinvitrotranscribedcappedOvamRNA,makesthisasuitablesystemtostudytheproductionofanantigenicpeptidesubstratevs.
thefull-lengthproteinfromagivenmRNAovertime.
ThelevelsofOvamRNA(2μg)fol-lowingtransfectionweredetectedusingquantitativeRT-PCR(qRT-PCR)andwerefoundtobeapproximatelyhalfafter8h(Fig.
2B).
ThepresentationoftheSL8epitopefromthetrans-fectedOvamRNAwasdeterminedbyxingthecellsafterin-dicatedtimepointsandrevealedapeakofantigenpresentationafterapproximately2.
5h,afterwhichitrapidlydeclined(Fig.
2C).
Ithasbeenestimatedthatthetimeittakesforapeptidetobeproduced,processed,loadedontotheMHCclassImolecules,andpresentedonthecellsurfaceisbetween30and60min(25).
Assumingthatthehalf-lifeofKb–SL8complexesloadedwithsyntheticvs.
endogenouspeptidesissimilar,ifonetakesintoaccounttheturnoverrateoftheKbmolecule,onecanestimatethattheproductionofSL8precursorpeptidesfromthetrans-fectedOvamRNAreachesitspeaklessthan1.
5hafterRNAtransfection.
Interestingly,atthesametime,wecoulddetectproductionoffull-lengthOvaproteinuntilatleast8haftertransfection(Fig.
2DandFig.
S3).
Hence,thesynthesisofan-tigenicpeptidesfromtheOvamRNAstopswellbeforethatofthesynthesisoffull-lengthproteins.
mRNAsTargetedfortheNMDPathwayProduceAntigenicPeptideSubstrates.
NewlysynthesizedandsplicedmRNAsundergoaqualitycontrolbeforetheycanproducefull-lengthproteins.
TotestifmRNAsthatundergothecontrolribosome-mediatedqualityscanningthatprecedestheNMDproduceantigenicpeptidesubstrates,wefusedtheSL8epitopeintherstexonofaβ-globinNMDreporterconstruct(26)thatcarriesaPTC53ntupstreamoftheEJC(Glob1-SL8-PTC)(Fig.
3A).
RNAquan-ticationrevealedthattheGlob1-SL8-PTCRNAisefcientlydegradedandthattheGlob1-SL8RNAisstable,asexpected(Fig.
3B).
WeobservedtheexpressionoftheYFP-globinfusionproteinfromthecorrespondingGlob1-SL8constructthatlacksthePTCbutnotfromtheGlob1-SL8-PTCconstruct(Fig.
3C).
WealsoinsertedtheSL8andMBP(79–87)codingsequencesinthesecondexonjustafterthePTC[Glob2-PTC-SL8/MBP(79–87)],whichleadstoalessefcientNMDprocess,allowingsomemRNAstoescapetheNMDandexpressatruncatedproteinthatisterminatedatthePTC(Fig.
3BandC).
ThemolecularmechanismbehindthispartialevasionofNMDisnotclearbutwasobservedusingbothepitopesequences.
ThepresentationoftheSL8epitopefromtheGlob1-PTC-SL8constructreaches85%ofthelevelsobservedfromthatoftheGlob1-SL8con-struct(Fig.
3D).
Similarresultswereobtainedusingdifferentcelllines(Fig.
S4).
TheSL8andtheMBP(79–87)epitopesarenotproducedwhenplaceddownstreamofthePTCcodonintheGlob2-PTC-SL8/MBPconstructsbecausethePTCpreventstheread-throughintotheSL8codingsequence.
ThepresentationoftheMBP(79–87)epitopeonKkmoleculesderivedfromtheGlob1-Fig.
2.
Synthesisofantigenicpeptidesandfull-lengthproteinsistemporallydistinct.
(A)TurnoverrateoftheendogenousKbMHCclassI-SL8complexinEL4mousecells,ormouseKb-SL8complexinhumanH1299cells,wasesti-matedbymeasuringSL8presentationonKbmolecules.
SyntheticSL8pep-tidewasaddedfor1h,followedbyextensivewashing(time0)andcellxationattheindicatedtimepointsbeforeincubationwithB3Zreportercells.
(B)LevelsofcappedOvamRNAinH1299cellsfollowingRNAtrans-fectionweredeterminedusingqRT-PCR.
Thearbitraryvalueof100%wasgivenat1h.
(C)AntigenpresentationaftertransfectionofOvamRNAinH1299cellsexpressingKbwasdeterminedonxedcellsattheindicatedtimepoints.
Maximumantigenpresentation(100%)wasobservedat2.
5h.
(D)WesternblotshowstheexpressionofOvaatdifferenttimepointsaftermRNAtransfectioninH1299cells.
Endogenousactinisusedasaloadingcontrol.
W.
B.
,Westernblot.
TheresultsarerepresentativeofatleastthreeindependentexperimentsplusSD(Fig.
S3).
11574|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1104104108Apcheretal.
DownloadedbyguestonJanuary10,2021MBP-PTCconstructwasfoundtobe90%comparedwiththecorrespondingcontrolGlob1mRNAthatlacksthePTCcodon(Glob1-MBP)(Fig.
3E).
Theseresultsdemonstratethatthepro-ductionofdifferentantigenicpeptidesubstratesforvariousMHCclassImoleculescoincideswiththeearlyroundoftranslationthatprecedestheNMD.
eIF4EBindingtotheCapStructureIsNotRequiredfortheProductionofAntigenicPeptides.
ToinvestigatetheroleofearlymRNAtranslationeventsintheproductionofantigenicpeptidesfur-ther,weexpressedanRNAaptamerthatspecicallytargetseIF4E(α-eIF4Eaptamer)andpreventsitsinteractionwiththecapstructure(27,28).
MetabolicpulselabelingassaysandWesternblotsshowedthatincreasingamountsoftheα-eIF4EaptamerresultedinadecreaseintherateofsynthesisofYFP-globinfromtheGlob1-SL8andGlob2-SL8constructs(80%)andfromtheOvamRNA(Fig.
4AandB).
Theα-eIF4EaptamerhadnoeffectontheNMDprocess,demonstratingthatitdoesnotaffecttheinteractionbetweentheCBCandthecapstructure(Fig.
4C).
Theexpressionoftheα-eIF4Eaptamerhadlittleef-fectonthepresentationofantigenicpeptidesrelativetoMHCclassImoleculeexpression(Fig.
4D).
The4E-BP1proteinpre-ventstheinteractionbetweeneIF4EandeIF4Gandblockscap-dependenttranslationinitiation(29).
Accordingly,increasingamountsof4E-BP1resultedinadecreaseintheexpressionofYFP-globin(Fig.
4E).
However,aswiththeα-eIF4Eaptamer,thishadlittleimpactontheproductionofantigenicpeptidesubstratesfromanyofthetestedreporterconstructs(Fig.
4F).
TheeIF4GinitiationfactorisrequiredfortheassemblyofboththeeIF4E-andCBC-dependentribosomalcomplexesthatinitiateribosomalqualitycontrolscanning(19,30).
Thepoliovirus2AproteasespecicallycleavesandinactivateseIF4G(Fig.
S5);inlinewiththis,wefoundthatincreasingamountsofprotease2AresultedinapartiallossofYFP-globinexpression(Fig.
4G)andareductionof60%inantigenicpeptidepre-sentation(Fig.
4H).
Hence,synthesisofantigenicpeptidesrequireseIF4GbutnoteIF4E,indicatingthattheproductionofantigenicpeptidesisspatiotemporallydifferentfromthatoffull-lengthproteinsandisgovernedbyadifferentmechanismoftranslationinitiation(Fig.
5).
Takentogether,thepresentedresultsareinagreementwiththenotionthatproductionofanti-genicpeptidestakesplaceduringtheearlytranslationeventsthatprecedetheNMD.
ConcludingRemarksItislikelythatPTPsrepresentfragmentsofthecorrespondingORFs,andwouldthereforebetargetedforrapiddegradationbythecellstolimitanypotentialharmtheymightcausebyun-controlledinterferencewithcellularprocesses.
Hence,inthisrespect,ourndingsonPTPswouldtwellwithcriteriaforwhathasbeenpostulatedtobehallmarksofantigenicpeptidesub-strates,includingrapiddegradationandcryptictranslation.
TherewouldbeobviousbenetsfortheimmunesurveillanceprocessifithasbeenadaptedtorecognizePTP-derivedproductsonMHCclassImolecules.
BecausethesepeptidesarethersttobeproducedfrommRNAs,itwouldmaketransactingmech-anismsusedby,forexample,virusestomanipulatetheMHCFig.
3.
mRNAstargetedforNMDproduceantigenicpep-tidesbutnotfull-lengthproteins.
(A)CartoonillustratingdifferentNMDreporterconstructsandthepositionofthePTCandtheSL8orMBP(79–87)epitope.
(B)RatiosofnormalandPTC-carryingmRNAsfromconstructsinwhichtheSL8isinsertedinexon1or2tothemouseurinaryprotein(MUP)controlmRNAareshown.
NumbersshowtherelativeamountsofindicatedmRNAs48haftertransfectioninHeLacells.
MUP.
(C)WesternblotshowstheexpressionofYFP-globinfusionproteinsinHeLacellsfromindicatedconstructsusingα-YFPmAb.
InsertionofSL8inexon2justdownstreamofthePTC(Glob2-PTC-SL8)par-tiallydisruptstheNMDprocess,allowingtheexpressionofatruncatedproductthatterminatesatthePTC.
TP,trun-catedproduct;W.
B.
,Westernblot.
(D)Antigenpre-sentationoftheSL8epitopefromanincreasingamountofindicatedconstructsinHeLacellsexpressingmouseKb.
SimilarresultswereobtainedusingH1299,aHEK293-TKbstablecellline,andB6mousebroblast(Fig.
S4).
Approxi-mately85%antigenpresentationisretainedfromGlob1-SL8-PTCmRNAsthataretargetedforNMDcomparedwiththenon-NMDGlob1-SL8construct.
TheGlob2-PTC-SL8con-structdoesnotproduceantigenicpeptidesubstratesbe-causethePTCstopsribosomalread-throughoftheSL8epitope.
*PApcheretal.
PNAS|July12,2011|vol.
108|no.
28|11575IMMUNOLOGYDownloadedbyguestonJanuary10,2021classIpathwaylesseffective.
Amechanisminwhichnewlysyn-thesizedviralmRNAsarescannedbyribosomesusinga"loose"initiationrequirementfortheproductionofantigenicpeptideswouldformamoredifcultchallengeforvirusestoovercome.
TransfectedmRNAs,despitenotbeingspliced,onlysupportthesynthesisofantigenpeptidesubstratesearlyaftertransfection,indicatingthatthematurationofmRNPcomplexestakesplaceirrespectiveofwhetherthemRNAsareexpressedendogenouslyorintroducedarticially.
Hence,itisconceivablethatnonsplicedviralmRNAswouldgothroughasimilarmaturationprocess.
Thereare,however,reportstosupportthatlatentviruseshavemechanismsinplacetoovercomethechallengeofPTPsbeingproducedduringthepioneerroundoftranslation.
EBNA1,encodedbytheEpstein–Barrvirus,inhibitsitsowntranslationincistosuppressantigenproduction,whereasthefull-lengthEBNA1hasalonghalf-lifetoensurethatitisexpressedatlevelsthataresufcienttosupportitsviralfunctions(9,31).
ThefactthatantigenicpeptidescanbesynthesizedfrommRNAs,althoughnotproducingfull-lengthproteins,hasotherinterestingimplications.
Oneofthecornerstonesinpreventingself-recognitionofCD8+Tcellsisthecomplexprocessofgen-eratingcentraltoleranceinthethymus,wherebymaturingTcellsareexposedtoself-peptidesand,ifreactive,becomeeliminated.
Thepossibilitythatthymiccellscouldproducetherequiredepitopesforthisselectionprocesswithouthavingtoproducethefull-lengthproteinsisattractive.
Theresultspresentedherebynomeansprovethishypothesis;however,importantly,theyshowthatitmightbepossible.
Finally,theseresultsimplythatsimilartodifferentRNApolymerasesproducingavarietyofRNAproductswithdenedcellularfunctions,differentribosomecomplexesproducepeptideproductswithdenedcellbiologicalfunctions.
Fig.
4.
eIF4GbutnoteIF4Eisrequiredforthesynthesisofantigenicpeptides.
(A)CellsweretransfectedwiththeindicatedYFP-globinreporterconstructfor24h,fol-lowedbyexpressionoftheα-eIF4Eaptamer(24h),whichpreventsinteractionbetweenthecapstructureandeIF4E.
Autoradiographofa[35S]methinoninemetabolicpulselabelingexperiment(Left)andcorrespondingWesternblot(Right)inHeLacellsexpressingindicatedamountsoftheα-eIF4ERNAaptamer.
W.
B.
,Westernblot.
(B)WesternblotofOvaexpressionundersimilarcon-ditions.
(C)RNAstabilityofindicatedglobinNMDre-porterconstructsinthepresenceorabsenceoftheα-eIF4Eaptamer.
ThenumbersindicatetherelativeamountsofmRNAswithSD.
(D)PresentationoftheSL8epitopeexpressedfromindicatedmRNAconstructsinthepresenceofα-eIF4Eaptamer.
Dataarepresentedasrel-ativetopresentationoftheexogenousSL8peptide.
(E)WesternblotofYFP-globinexpressionfromtheGlob1-SL8constructinthepresenceofincreasingamountsof4E-BP1.
(F)Correspondingantigenpresentationfromin-dicatedconstructsundersimilarconditionsasinE.
(G)Westernblotshowsthatincreasingamountsofpro-tease2A,whichcleavestheeIF4Gtranslationfactor(Fig.
S5),leadtoareductioninYFP-globinproteinexpressionlevels.
(H)RelativepresentationofSL8epitopederivedfromindicatedconstructscomparedwithcellsbeingex-posedtoexogenousSL8peptidefor16hinthepresenceofincreasingamountsofprotease2A.
AllresultsarerepresentativeofthreeindependentexperimentsandSD.
Fig.
5.
ModelforthesynthesisandselectionofantigenicpeptidesubstratesduringthepioneerroundoftranslationthatprecedesNMD.
mRNAsarecheckedforPTCsduringtheinitialqualitycontroltranslationthatfollowsnewlysynthesizedandsplicedmRNAs.
ThistranslationisinitiatedbytheCBC,includingCBP20/80,andproducesPTPsthatareinstantlydegradedandselectedforantigenicpeptideprocessingfortheMHCclassIpathway.
mRNAsfoundtocontainPTCsaretargetedfortheNMDpathway(1),andthosemRNAsthatpassthequalitycontrolmatureintomRNPcomplexesthatbindtheeIF4Ecap-bindingfactorandinitiatecanonicaltranslationcom-plexesthatgovernthesynthesisoffull-lengthproteins(2).
Degradationoffull-lengthproteinsplaysonlyaminorroleasasourceofantigenicpeptidesfortheMHCclassIpathway.
11576|www.
pnas.
org/cgi/doi/10.
1073/pnas.
1104104108Apcheretal.
DownloadedbyguestonJanuary10,2021MaterialsandMethodsT-CellHybridoma,CellCulture,andTransfection.
TheSIINFEKL/Kb-specic(B3Z)andtheMBP(79-87)/Kk-specicT-cellreporterhybridomashavebeende-scribedpreviously(32).
HumancelllineswerecultivatedunderstandardconditionsinRPMI1640.
ThehumanHEK-293TKbstablecelllineandmouseB6celllineweregiftsfromL.
Eisenlohr(ThomasJeffersonUniversity,Phila-delphia,PA).
FormRNAtransfection,cellsweretransfectedwith0.
5μgofKbexpressionplasmidalongwith3μLofGenejuiceaccordingtothemanu-facturer'sprotocol(MerckBiosciences).
Twenty-fourhourslater,thecellsweretransfectedwith2μg(unlessstatedotherwise)ofcappedmRNAusinglipo-fectamineaccordingtothemanufacturer'sprotocol(Invitrogen).
DetailsofplasmidconstructionandqRT-PCRareprovidedinSIMaterialsandMethods.
InVitroTranscription.
CappedmRNAsweresynthesizedwithT7RNApoly-merase(mMessageMachine;Ambion)accordingtothemanufacturer'sprotocol.
ThepCDNA3-Ova(describedinSIMaterialsandMethods)linear-izedwithXba1wasusedina50-μLnalvolumeforgeneratingmRNAsandwaspuriedusingtheRNeasykit(Qiagen).
MetabolicCellLabelingandImmunoprecipitation.
AllmRNAtranslationassayswerecarriedoutinH1299andHeLacellstransfectedwithindicatedcon-structsasdescribedpreviously(9).
Transfectedcellswereculturedfor36hbeforebeingtreatedwith25μMMG132for1hinmethionine-freemediumcontaining2%dialyzedFCSinDMEM.
[35S]Methionine(0.
15mCi/mL;Perkin–Elmer)wasaddedinthepresenceofproteasomeinhibitor,andthecellswereharvestedatindicatedtimepoints.
ThepeptideproductswereimmunoprecipitatedusingindicatedantibodiesandseparatedonprecastBis-Tris4–12%acrylamideSDS/PAGEgels(Invitrogen).
Therelativeamountofproteinsynthesiswasdeterminedusingaphosphoimager.
T-CellAssayandFixation.
T-cellassaysinhumancelllineswerecarriedoutasdescribedpreviously(9).
Briey,humanH1299andHeLacelllineswerecotransfectedwiththeKborKkexpressionvector,togetherwiththedif-ferentreporterconstructs.
HumanHEK-293TKbisastablecellline,andthemouseB6celllineexpressesendogenousKb.
TheB3ZorMBPCD8+T-cellhybridomaexpressesLacZinresponsetoactivationofT-cellreceptorsspe-cicfortheSIINFEKLpeptide(Ova-immunodominantpeptide)inthecontextofH-2KbMHCclassImoleculesandfortheMBP(79–87)peptideinthecontextofH-2KkMHCclassImolecules,respectively.
H1299cell-expressingexogenousKb(Fig.
2A)waspulse-chasedfor48hwithsyntheticSIINFEKLpeptideandextensivelywashedbeforebeingxedin0.
05%glutaraldehydeand0.
1Mglycine.
Thecellswerethenwashedtwotimesin1*PBSandculturedwithB3ZT-cellhybridomafor16–20h.
RNAPreparation,RT-PCR,andqRT-PCR.
TotalcellularRNAswereextractedandpuriedwiththeRNeasykitfollowingthemanufacturer'sprotocol.
RTwascarriedoutwith1μgofRNAusingtheM-MLVreversetranscriptaseandoligo-dT(Invitrogen).
TheStepOne(AppliedBioSystems)real-timePCRsys-temwasusedforqRT-PCR,andthereactionwasperformedwiththePer-fectaSYBRgreenFastmixROX(Quanta)andspecicprimerpairsforeachgeneofinterest.
ACKNOWLEDGMENTS.
WethankJoanGovermanforCD8+TcellsagainsttheMBPclassIepitopeandSaumitraDasforthepoliovirusprotease2Aexpressionconstruct.
ThisworkwassupportedbyLaLigueContreleCancer,theInstitutNationaldelaSantéetdelaRechercheMédicale,theANR,andRegionalCentreforAppliedMolecularOncology(RECAMO)GrantCZ.
1.
05/2.
1.
00/03.
0101(fromtheEuropeanRegionalDevelopmentFund).
S.
A.
isfundedbytheFondationpourlaRechercheMédical.
1.
RockKL,etal.
(1994)InhibitorsoftheproteasomeblockthedegradationofmostcellproteinsandthegenerationofpeptidespresentedonMHCclassImolecules.
Cell78:761–771.
2.
RockKL,YorkIA,GoldbergAL(2004)Post-proteasomalantigenprocessingformajorhistocompatibilitycomplexclassIpresentation.
NatImmunol5:670–677.
3.
KloetzelPM(2004)GenerationofmajorhistocompatibilitycomplexclassIantigens:FunctionalinterplaybetweenproteasomesandTPPII.
NatImmunol5:661–669.
4.
SerwoldT,GonzalezF,KimJ,JacobR,ShastriN(2002)ERAAPcustomizespeptidesforMHCclassImoleculesintheendoplasmicreticulum.
Nature419:480–483.
5.
VanKaerL,Ashton-RickardtPG,PloeghHL,TonegawaS(1992)TAP1mutantmicearedecientinantigenpresentation,surfaceclassImolecules,andCD4-8+Tcells.
Cell71:1205–1214.
6.
FalkK,RtzschkeO,StevanoviS,JungG,RammenseeHG(1991)Allele-specicmotifsrevealedbysequencingofself-peptideselutedfromMHCmolecules.
Nature351:290–296.
7.
YewdellJW,AntónLC,BenninkJR(1996)Defectiveribosomalproducts(DRiPs):AmajorsourceofantigenicpeptidesforMHCclassImoleculesJImmunol157:1823–1826.
8.
ShastriN,CardinaudS,SchwabSR,SerwoldT,KunisawaJ(2005)Allthepeptidesthatt:Thebeginning,themiddle,andtheendoftheMHCclassIantigen-processingpathway.
ImmunolRev207:31–41.
9.
ApcherS,DaskalogianniC,ManouryB,FhraeusR(2010)EpsteinBarrvirus-encodedEBNA1interferencewithMHCclassIantigenpresentationrevealsaclosecorrelationbetweenmRNAtranslationinitiationandantigenpresentation.
PLoSPathog6:e1001151.
10.
FhraeusR(2005)DopeptidescontroltheirownbirthanddeathNatRevMolCellBiol6:263–267.
11.
YinY,ManouryB,FhraeusR(2003)Self-inhibitionofsynthesisandantigenpresentationbyEpstein-Barrvirus-encodedEBNA1.
Science301:1371–1374.
12.
GilbertWV(2010)Alternativewaystothinkaboutcellularinternalribosomeentry.
JBiolChem285:29033–29038.
13.
GingrasAC,RaughtB,SonenbergN(1999)eIF4initiationfactors:EffectorsofmRNArecruitmenttoribosomesandregulatorsoftranslation.
AnnuRevBiochem68:913–963.
14.
PeabodyDS(1989)Translationinitiationatnon-AUGtripletsinmammaliancells.
JBiolChem264:5031–5035.
15.
SchwabSR,ShugartJA,HorngT,MalarkannanS,ShastriN(2004)Unanticipatedantigens:TranslationinitiationatCUGwithleucine.
PLoSBiol2:e366.
16.
ChangYF,ImamJS,WilkinsonMF(2007)Thenonsense-mediateddecayRNAsurveillancepathway.
AnnuRevBiochem76:51–74.
17.
IskenO,etal.
(2008)Upf1phosphorylationtriggerstranslationalrepressionduringnonsense-mediatedmRNAdecay.
Cell133:314–327.
18.
LeHirH,SéraphinB(2008)EJCsattheheartoftranslationalcontrol.
Cell133:213–216.
19.
LejeuneF,RanganathanAC,MaquatLE(2004)eIF4Gisrequiredforthepioneerroundoftranslationinmammaliancells.
NatStructMolBiol11:992–1000.
20.
MagnaniM,CrinelliR,BianchiM,AntonelliA(2000)Theubiquitin-dependentproteolyticsystemandotherpotentialtargetsforthemodulationofnuclearfactor-kB(NF-kB).
CurrDrugTargets1:387–399.
21.
KarttunenJ,SandersonS,ShastriN(1992)Detectionofrareantigen-presentingcellsbythelacZT-cellactivationassaysuggestsanexpressioncloningstrategyforT-cellantigens.
ProcNatlAcadSciUSA89:6020–6024.
22.
SchwabSR,LiKC,KangC,ShastriN(2003)ConstitutivedisplayofcryptictranslationproductsbyMHCclassImolecules.
Science301:1367–1371.
23.
HusebyES,OhlénC,GovermanJ(1999)Cuttingedge:Myelinbasicprotein-speciccytotoxicTcelltoleranceismaintainedinvivobyasingledominantepitopeinH-2kmice.
JImmunol163:1115–1118.
24.
LjunggrenHG,etal.
(1990)EmptyMHCclassImoleculescomeoutinthecold.
Nature346:476–480.
25.
EisenlohrLC,HuangL,GolovinaTN(2007)RethinkingpeptidesupplytoMHCclassImolecules.
NatRevImmunol7:403–410.
26.
ZhangJ,SunX,QianY,MaquatLE(1998)Intronfunctioninthenonsense-mediateddecayofbeta-globinmRNA:Indicationsthatpre-mRNAsplicinginthenucleuscaninuencemRNAtranslationinthecytoplasm.
RNA4:801–815.
27.
ElangoN,LiY,ShivshankarP,KatzMS(2006)ExpressionofRUNX2isoforms:Involvementofcap-dependentandcap-independentmechanismsoftranslation.
JCellBiochem99:1108–1121.
28.
MochizukiK,OguroA,OhtsuT,SonenbergN,NakamuraY(2005)HighafnityRNAformammalianinitiationfactor4EinterfereswithmRNA-capbindingandinhibitstranslation.
RNA11:77–89.
29.
HaghighatA,MaderS,PauseA,SonenbergN(1995)Repressionofcap-dependenttranslationby4E-bindingprotein1:Competitionwithp220forbindingtoeukaryoticinitiationfactor-4E.
EMBOJ14:5701–5709.
30.
AlvarezE,Menéndez-AriasL,CarrascoL(2003)Theeukaryotictranslationinitiationfactor4GIiscleavedbydifferentretroviralproteases.
JVirol77:12392–12400.
31.
ApcherS,etal.
(2009)mRNAtranslationregulationbytheGly-AlarepeatofEpstein-Barrvirusnuclearantigen1.
JVirol83:1289–1298.
32.
ShastriN,GonzalezF(1993)Endogenousgenerationandpresentationoftheovalbuminpeptide/KbcomplextoTcells.
JImmunol150:2724–2736.
Apcheretal.
PNAS|July12,2011|vol.
108|no.
28|11577IMMUNOLOGYDownloadedbyguestonJanuary10,2021