2017-Yanagitani-UBE2O is a quality control fac.pdf

REPORTPROTEOSTASISUBE2O is a quality control factor fororphans of multiprotein complexesKota Yanagitani,*Szymon Juszkiewicz,Ramanujan S.HegdeMany nascent proteins are assembled into multiprotein complexes of defined stoichiometry.Imbalancesinthesynthesisofindividualsubunitsresultinorphans.Howorphansareselectivelyeliminated to maintain protein homeostasis is poorly understood.Here,we found that theconserved ubiquitin-conjugating enzyme UBE2O directly recognized juxtaposed basic andhydrophobic patches on unassembled proteins to mediate ubiquitination without a separateubiquitin ligase.In reticulocytes,where UBE2O is highly up-regulated,unassembled a-globinmolecules that failed to assemble with b-globin were selectively ubiquitinated by UBE2O.Innonreticulocytes,ribosomalproteinsthatdidnotengagenuclearimportfactorsweretargetsforUBE2O.Thus,UBE2O is a self-contained quality control factor that comprises substraterecognition and ubiquitin transfer activities within a single protein to efficiently target orphansof multiprotein complexes for degradation.Cells have evolved a wide range of qualitycontrol pathways to monitor failures inprotein biogenesis and promptly targetdefective products for degradation(1,2).Ineffective quality control is implicated inaging and can cause neurodegeneration(3);con-versely,robust quality control may be especiallycrucial in cancer to support high rates of pro-tein synthesis in the face of an aberrant genomeandenvironmentalstresses(4).Thesuiteofproteinquality control pathways needed for cellular ho-meostasisinmetazoansisincompletelydefined(1).To identify additional cytosolic quality controlpathways,we sought an aberrant protein whoseubiquitination in a cell-free cytosolic extract oc-curs without engaging known quality controlfactors.A transmembrane(TM)domain inter-rupted by three basic residues(hereafter termed3R)(fig.S1A)is not recognized by quality controlfactors specialized for mislocalized membraneproteins,such as BAG6 or the ubiquilin familymembers(57);yet,3R was ubiquitinated sim-ilarly to a BAG6 substrate(fig.S1B).Although 3Ris an artificial mutant,we reasoned that mech-anistic dissection of its ubiquitination pathwaymight lead us to a quality control pathway andprovide tools that could be exploited to identifyphysiological substrates.In vitrotranslated35S-labeled 3R immunopu-rified under native conditions became modifiedwith His-ubiquitin in the presence of recombinantE1,E2(UBCH5),and adenosine 5-triphosphate(ATP)(fig.S1,C and D),indicating that the im-munopurified complexes contained a ubiquitinligase.The primary 3R-associated ubiquitinationactivity had a native molecular weight of 150 to300 kD(Fig.1,A and B).Cross-linking reactionsof the active fractions revealed interacting part-ners of 200,120,and 100 kD(Fig.1B).Massspectrometry identified UBE2O as the 200-kDproduct,Importin 5(IPO5)andImportin7(IPO7)as the 120-kD product,and Importin b as the100-kD product(fig.S2A).Although none oftheseareubiquitinligases,UBE2Oisaubiquitin-conjugating(E2)enzyme,withsuspectedubiquitintransfercapability(8).Furthermore,UBE2Oisup-regulated incells in response to induced misfold-ingofacytosolicprotein(9).Wethusinvestigatedits potential role in recognition and ubiquitina-tion of 3R and other aberrant proteins.RESEARCHYanagitani et al.,Science 357,472475(2017)4 August 20171 of 4Fig.1.UBE2O associates with an aberrantnascent protein in the cytosol.(A)Experi-mental strategy to identify quality controlfactors through cross-correlation of native size,physical interactors,and ubiquitination activity.(B)35S-labeled Sec61b(3R)was translated inreticulocyte lysate(RRL)for 15 min,which is justlong enough to synthesize the proteins butbefore appreciable downstream ubiquitination.The reaction was separated according to nativesize on a sucrose gradient,and each fractionwas analyzed for ubiquitination competence(middle)or physical interactions(bottom).Redasterisks indicate fractions with peak ubiquiti-nation activity,and red arrowheads indicatecross-linked proteins with a peak in these samefractions.(C)Sec61b wild type(WT)or mutantslacking the transmembrane domain(DTM)orcontaining 3R were translated in RRL,immuno-precipitated under native conditions,andanalyzed by means of immunoblotting foreither BAG6 or UBE2O.BAG6 interacts withthe intact TM domain,whereas UBE2Opreferentially interacts with the 3R-disruptedTM domain.on December 27,2018 http:/science.sciencemag.org/Downloaded from Immunoblotting of natively immunopurified3R verified its interaction with UBE2O.By con-trast,BAG6 interacted efficiently with an intactTMdomainbutpoorlywith3R(Fig.1C),whichisconsistent with its preference for long uninter-rupted hydrophobic domains(5,10).Immuno-precipitation of cross-linking reactions verifiedthat the 200-kD cross-linking partner from Fig.1B was indeed UBE2O,suggesting a direct inter-action with 3R(fig.S2B).Furthermore,the smallglobularproteinKRASdidnotinteractwithUBE2O,whereas mutants designed todisrupt its folding(termed K1 and K2)showed increased ubiquiti-nation and UBE2O interaction(fig.S3,A to C).GiventhatUBE2Owasobservedprominentlyinstainedgelsofaffinity-purifiedsamples(figs.S2AandS3D),weconcludeitisamajorandrelativelystableinteractor,withselectivityforaberrantpro-teins such as 3R and misfolded KRAS.Experiments with purifiedrecombinant factorsdemonstrated that UBE2O is sufficient for inter-action with and ubiquitination of its clients.Wefirst synthesized 3R in a“PURE”(protein syn-thesisusingrecombinantelements)invitrotrans-lation system reconstituted from recombinantEscherichia coli translation factors(11).Nascent3R normally precipitates in this chaperone-freePURE system but was prevented from aggrega-tion by including Calmodulin(CaM)during thetranslation(fig.S4A).CaM acts as a chaperoneforhydrophobicdomains(12)andcanbeinducedto retain or release 3R by using Ca2+or EGTA,respectively.Cross-linking assays verified thatpurified UBE2O(fig.S4B)engaged 3R releasedfrom CaM with EGTA but not with the CaM-3Rcomplex(Fig.2,A and B).In this system,UBE2O(together with E1,ubiquitin,and ATP)was suf-ficient for 3R ubiquitination(Fig.2C),whereasthe promiscuous E2 enzyme UBCH5 was ineffec-tive toward 3R on its own(Fig.2C and fig.S5A).Ubiquitination required the TM-3R region andwas markedly impaired if release from CaM wasinhibited with excess Ca2+(Fig.2C).Other hydro-phobic regions could also be directly recognizedand ubiquitinated by UBE2O in this purified sys-tem(fig.S5B),even though they would be recog-nizedbyotherfactors(suchasBAG6)inacompletecytosol(5,7,12).Themaximalnumberofubiquitinsadded to the client corresponded to the numberofavailableprimaryaminesandwasnotaffectedappreciably by using methyl-ubiquitin incapableof forming polyubiquitin chains(fig.S5C).C1040in the E2 domain of UBE2O was essential forubiquitinationactivity(Fig.2D)butnotforinter-actionwithaclient.Oftheconservedregions(CR)of UBE2O(13),deletion of CR2,and to a lesserextent CR1,reduced client ubiquitination withoutaffecting UBE2O activity,as judged by auto-ubiquitination(Fig.2D).Correspondingly,re-combinantCR2,andtoa lesserextentCR1,couldYanagitani et al.,Science 357,472475(2017)4 August 20172 of 4Fig.2.Reconstitution ofUBE2O-mediated clientubiquitination with purifiedfactors.(A)Experimentalstrategy to test UBE2O inter-action and ubiquitination in adefined system.(B)35S-labeled3R in complex with CaM wastreatedwitheitherCa2+orEGTAin the presence or absence ofUBE2O and subjected to theamine-reactive cross-linkerdisuccinimidyl suberate(DSS).Positions of the starting 3R sub-strate and its cross-links to CaM(xCaM)and UBE2O(xUBE2O)are indicated.(C)35S-labeled 3Rin complex with CaM was mixedwith E1,ATP,ubiquitin,and theindicatedE2(UBCH5orUBE2O),then treated with EGTA so as toinduce substrate release fromCaM.One sample was incubatedwith Ca2+instead of EGTA.(D)Ubiquitinationreactionsasin(C),with the indicated UBE2Ovariants.The purified UBE2O pro-teins before and after the ubiqui-tinationreactionarealsoshowntoindicate auto-ubiquitination ofUBE2Oinallreactionsexceptwiththe C1040S mutant.(Single-letter abbreviations for the amino acid residues are as follows:A,Ala;C,Cys;D,Asp;E,Glu;F,Phe;G,Gly;H,His;I,Ile;K,Lys;L,Leu;M,Met;N,Asn;P,Pro;Q,Gln;R,Arg;S,Ser;T,Thr;V,Val;W,Trp;andY,Tyr.Inthemutants,otheraminoacidsweresubstitutedatcertainlocations;forexample,C1040S indicates that cysteine at position 1040 was replaced by serine.)Fig.3.Unassembleda-globin is a target forUBE2O ubiquitination.(A)Wild-type a-globin(HBA1)or two assembly mutants(H104R and F118S)weretranslated in RRL supple-mented with His-tagged ubiq-uitin and either analyzeddirectly(bottom;total)or afterubiquitin-pulldown(Ub-PD)viathe His tag(top).(B)Hemag-glutinin(HA)tagged HBA1 orthe H104R mutant were trans-lated in RRL,subjected tosulfhydryl-reactive cross-linking with bismaleimidohex-ane,and affinity-purified viathe HA tag.The purifiedproducts were denatured andanalyzed directly(input)orfurther immunoprecipitatedwith antibodies to eitherUBE2O or GFP.The positionsof ubiquitinated H104R and itscross-link to UBE2O are indi-cated.(C)35S-labeled HBA1 synthesized in the PURE translation system was incubated with E1,ATP,ubiquitin,and either UBCH5 or UBE2O and analyzed for ubiquitination.(D)35S-labeled WTor H104Rmutant HBA1 synthesized in the PURE system was incubated with or without UBE2O and AHSP asindicated.The positions of unmodified and ubiquitinated HBA1 are indicated.Medical Research Council(MRC)Laboratory of MolecularBiology,Francis Crick Avenue,Cambridge BiomedicalCampus,Cambridge,CB2 0QH,UK.*Present address:The Thomas N.Sato BioMEC-X Laboratories,Advanced Telecommunications Research Institute International(ATR),Hikaridai 2-2-2,Kyoto 619-0288,Japan.Correspondingauthor.Email:rhegdemrc-lmb.cam.ac.ukRESEARCH|REPORTon December 27,2018 http:/science.sciencemag.org/Downloaded from directly interact with in vitrotranslated 3R(fig.S6).Thus,the CR regions of UBE2O directly rec-ognize exposed hydrophobic patches on 3R andother clients to mediate E2 domaindependentmultimonoubiquitination.In considering potential physiologic qualitycontrol clients of UBE2O,we were struck by itsmarked up-regulation during erythrocyte differ-entiation(14),hinting that the massive increaseinhemoglobinsynthesisinpre-erythrocytesmightnecessitate greater quality control.During adulthemoglobin assembly(fig.S7,A and B),a-globin(also called HBA1)uses a-hemoglobinstabilizingprotein(AHSP)to temporarily shield the interfacethatiseventuallyoccupiedbyb-globin(15).Severalhuman mutations in a-globin impair interactionwith AHSP and cause a variant thalassemia-likedisease(16).Two such mutants were ubiquiti-nated at elevatedlevels compared with wild-typea-globin when translated in a reticulocyte lysate(Fig.3A).These a-globin mutants interacted di-rectly and specifically with UBE2O,as judged bycofractionation(fig.S7C)and cross-linking(Fig.3B).AHSP is abundant in reticulocyte lysate(thenative context for hemoglobin assembly)(15),ex-plaining why nascent wild-type a-globin was notubiquitinated in this experiment.By contrast,UBE2O ubiquitinated wild-type a-globin in thechaperone-free PURE translation system(Fig.3C).RecombinantAHSPpreventedUBE2O-mediatedubiquitination of wild-type a-globin but was lesseffective for the mutant(Fig.3D).Thus,UBE2Oappears to recognize unassembled a-globin viathe region normally covered by AHSP.Consistentwith this,lower levels of ubiquitinated a-globinare observed in reticulocytes from UBE2O-nullmice(17).Although unassembled a-globin represents aphysiologically important UBE2O client,a moregeneralclientrangeissuggestedbyUBE2Osdeepconservation across eukaryotes and broad tissueexpressioninmammals.Wereasonedthataswiththe a-globinAHSP system,UBE2O may identifyother clients via elements that normally shouldhave engaged a protein biosynthesis factor.Majorfactors engaged by 3R included IPO5,IPO7,andImportinb(fig.S2A),allofwhich are involvedinnuclearimportofribosomalproteins(18).Chang-ing 3R to 3D strongly impaired interaction withboth UBE2O and IPO7(fig.S8A),supporting theidea that they recognize juxtaposed basic andhydrophobic residues,a feature also seen on thesurfaceofa-globinrecognizedbyUBE2O(fig.S8B).WethusinvestigatedwhethernascentribosomalproteinsmanyofwhichwouldcontainexposedbasicandhydrophobicsurfacesmightbeUBE2Oclients.We found that several(such as RPL3,RPL8,and RPL24)but not all(such as RPS10)riboso-mal proteins were highly ubiquitinated after syn-thesisinreticulocytelysate,andtheubiquitinatedproducts cofractionate with UBE2O(Fig.4A andfig.S9).Affinitypurificationofthesenascentribo-somal proteins showed that UBE2O was a majorinteractor of each protein that was ubiquitinatedbut not of RPS10(Fig.4B).Reconstitution exper-iments in the PURE system showed that RPL8was efficiently ubiquitinated by UBE2O,and thatits ubiquitination required the catalytic cysteineof the E2 domainand the CR2 region forefficientrecognition(Fig.4C).We used a green fluorescent protein(GFP)tagged RPL24(fig.S10A)that cannot assembleintoribosomesforstericreasonsandisdegradedinaproteasome-dependentmanner(fig.S10B)toverify its interaction with both UBE2O and IPO7in cultured cells(fig.S10C).Immunoprecipitationof UBE2O recovered unmodified and monoubi-quitinated GFP-RPL24,but not IPO7,suggestingthat they do not form a ternary complex(fig.S10D).CatalyticallyinactiveUBE2O(C1040S)alsorecovered endogenous and exogenous RPL24(but nodetectable ubiquitinated forms),whereasUBE2O(DCR2)interacted poorly with RPL24(fig.S10D).Thus,in vitro and in cells,nascent un-assembled ribosomal proteins engage either nu-clear import factors or UBE2O.The absence ofa ternary complex with IPO7 and UBE2Os ca-pacity to ubiquitinate bound clients suggest thatit recognizes the population of nascent riboso-malproteinsthatfailnuclearimportorassemblyinto the ribosome and targets it for proteasomaldegradation.Using a quantitative flow cytometry assay forGFP-RPL24 degradation(in which red fluores-cent protein serves as an expression control)(fig.S10A),we found that three independent smallinterfering RNAs(siRNAs)that reduce UBE2Oby 90%stabilize GFP-RPL24(Fig.4D and fig.S10E).By contrast,overexpression of UBE2Ostimulated GFP-RPL24 degradation,whereasoverexpression of a catalytically inactive mutant(C1040S)acted as a dominant-negative to in-crease GFP-RPL24(Fig.4E).Knockdown of IPO7to reduce GFP-RPL24 nuclear import resulted inits increased degradation,suggesting that cyto-solic degradation is more efficient than a recentlydescribed nuclear quality control pathway forribosomal proteins(19).Cytosolic degradation ofGFP-RPL24 was attributable to UBE2O becauseGFP-RPL24wasstabilizedbyadditionallyknock-ing down UBE2O(Fig.4F).Thus,UBE2O targetsfor degradation excess ribosomal proteins in theYanagitani et al.,Science 357,472475(2017)4 August 20173 of 4Fig.4.UBE2O facilitates degradation of cytosolic orphan ribosomal proteins.(A)Wild-typeKRAS or RPL8 was translated in RRL supplemented with His-tagged ubiquitin and eith