ProcNatlAcadSciUSAVol81pp61296133October1984GeneticsOxygenregulationo - PDF document

1 Proc.Natl.Acad.Sci.USAVol.81,pp.6129-613
Proc.Natl.Acad.Sci.USAVol.81,pp.6129-6133,October1984GeneticsOxygenregulationofanaerobicandaerobicgenesmediatedbyacommonfactorinyeast(coordinateexpression/generegulation/genefusion/transcriptionalregulation)CHARLESV.LOWRYANDRICHARDS.ZITOMERDepartmentofBiologicalSciences,StateUniversityofNewYork,1400WashingtonAvenue,Albany,NY12222CommunicatedbyRollinD.Hotchkiss,June28,1984ABSTRACTTheexpressionofanumberofyeastgenesisregulatedbyoxygenlevels.Whilemanyoftheseareknowntobeinducedinthepresenceofoxygen,wehavedescribedagene,ANBI,thatrespondsintheoppositefashion,beingexpressedonlyunderanaerobicconditions.ToidentifygenesinvolvedinregulationofANBIandotheroxygen-regulatedgenes,weselectedmutationscausingconstitutiveexpressionofANBI,usingafusionoftheANBImodulatorsegmenttotheCYCIgene.Anumberoftrans-actingmutationsaffectingagenedesignatedROXIcausedconstitutiveexpressionofboththefusedandwild-typegenes,indicatingthattheROXIgeneproductoperatesthroughtheANBImodulatorsequenceattheleveloftranscription.ThemutantallelesofROXIfallintotwophenotypicclasses.Theroxl-aclassissemi-dominant,andtheroxl-bclassisrecessive.Onemutant,roxl-al,ispleiotropicandcausesconstitutiveexpressionofthreeoxygen-inducedgenes-CYCI,SOD(superoxidedismutase),andtr-I(anoxy-gen-inducedgenewithhomologytoANB1)-aswellascon-stitutiveexpressionoftheoxygen-repressedANBIgene.Allelesoftheroxi-bclasscauseconstitutiveexpressionofANBIbutdonotaffectexpressionoftheoxygen-inducedgenestested.Thepleiotropyoftheroxi-a)mutantindicatesthattheROXIgeneproductisinvolvedincoordinateexpressionofbothoxygen-inducedandoxygen-repressedgenes.Manyeukaryoticgenescanbegroupedintofamiliesdefinedbytheircoordinateexpressioninresponsetoextracellularconditionsorsignals(1).Thesegenefamiliesareusuallydispersedthroughoutthegenome.Fortheexpressionofadispersedgenefamilytoberegulatedbyacommonmecha-nism-afeatureseeminglydesirableforeconomyandeffi-cientcoordination-ithasbeensuggestedthatcommonfactors,specificforeachgenefamily,formregulatorycom-plexesatcommonrecognitionsequenceswithinthetran-scriptionalmodulatorslyingupstreamfromeachgene(1).Theroleofsuchconsensussequences(1-4)andofcommonregulatoryfactors(5-14)hasbeenindicatedinseveralsys-tems.WehavesoughttodeterminewhetheragroupofgenesinSaccharomycescerevisiaewhoseexpressionisregulatedbythepresenceofoxygenarecontrolledbycommonfactorsactingthroughcommonregulatorysites.Theexpressionofonesuchgene,CYCI,hasbeeninvestigatedindetail,asreportedfromthisandotherlaboratories.Regulationofexpressionbyoxygenlevelshasbeenshowntobemediatedatanupstreamtranscriptionalmodulatorsite(15,16),withtheinvolvementofhemeasaregulatoryeffector(16,17).Expressionofotheroxygen-inducedgenesisalsodependentonheme(17,18),regulatedbyitattheleveloftranscription(17).Thesefindingsareconsistentwiththehypothesisofcoordinateexpressionmediatedbyacommonmechanism.Inaddition,pleiotropicmutationsaffectingoxygenregula-tionofsomegenescodingforhemoproteinshavebeenre-ported(17).Wehavealsostudiedanothergenewhoseexpressionisregulatedbyoxygen,butintheoppositeway.TheanaerobicgeneANBJ(previouslydesignatedtr-2),whosefunctionisunknown,isexpressedonlyintheabsenceofoxygen(15).Wereasonedthat,sinceboththeaerobicallyandanaerobi-callycontrolledgenesareregulatedbyoxygen,commonfactorsmightbeinvolvedintheirexpression,sothattrans-actingmutationsselectedforaneffectononegenemightaffecttheothers.Totestthis,weselectedforconstitutivemutationscausingaccumulationofANBImRNAinthepresenceofoxygen.Themutantsobtaineddefinearegula-torygene,whichwedesignateROXJ.Inagreementwiththehypothesizedinvolvementofcommonregulatoryfactorsincoordinateexpression,oneofthemutantallelesofROXIprovedtobepleiotropic,causingoxygen-constitutiveexpres-sionofaerobicgenesaswellasofANBJ.Thisprovidesanexampleofcoordinatedregulationoftwosetsofgeneswhoseexpressionrespondsinanoppositewaytothesameenvi-ronmentalfactor.MATERIALSANDMETHODSStrains.TheparentalstrainwasGM-3C-2[MATa,cycl-1,cyp3-1(cyc7),trpl-J,his4-519,leu2-3,leu2-112,Gal-;seeref.19].Forgeneticanalysis,crossesweremadewithaRZ5(MATa,cycl-1,cyp3-1,ura3,his4-519).SomegenetictestswereperformedwithstrainsCL-2-15a(MATa,roxi-al,cycl-1,cyp3-1,trpl-1,ura3,his4-519)andCL1-1-13b(MATa,roxl-b3,cycl-1,cyp3-1,trpl-1,leu2-2,leu2-112,his4-519).PlasmidsandTransformation.ThefollowingplasmidswereusedfortransformationofGM-3C-2:YCpCYCI(2.4)(seeFig.1;ref.15);YCpCYCI(2.4)1/X(seeFig.1;ref.15);YEp13(20),whichcarriestheLEU2gene;andYEpHO[suppliedtousbyR.JensenandI.Herskowitz(21)],whichcarriestheHOmating-typeswitchinggeneclonedinYEp13.PlasmidswereintroducedintoyeastorEscherichiacolistrainHB101bytransformationasdescribed(15).Strainswerecuredofcentromericplasmidsbygrowingcellsfor20generationsinYPDmedium(2%peptone/1%yeastextract/3%glucose),spreadingthemonYPDplates,andreplica-platingonCM(-trp)platestoidentifycoloniesthathadlosttheplasmid.PlasmidpFL1-cIII-FeS-3(ref.22;suppliedtousbyL.Grivell)wasrecentlyshowntocarrythesuperoxidedismu-tasegene(SOD)(L.Grivell,personalcommunication)ona1.8-kilobase(kb)BamHIfragment,whichwasusedasahybridizationprobeforSODmRNA.OtherprobeswerederivedfromYCpCYCI(2.4)orpYeCYCI(0.6)(23)asindi-catedinthetext.MutagenesisandSelectionofConstitutiveMutations.GM-3C-2cellstransformedwithYCpCYCI(2.4)1/Xweremu-tagenizedwithethylmethanesulfonateorUVirradiation,asAbbreviation:kb,kilobase(s).6129Thepublicationcostsofthisarticleweredefrayedinpartbypagechargepayment.Thisarticlemustthereforebeherebymarke

2 d"advertisement"inaccordancewith18U.S.C.
d"advertisement"inaccordancewith18U.S.C.§1734solelytoindicatethisfact. 6130Genetics:LowryandZitomerdescribedbyShermanetal.(24).Aftermutagenesis,cellswereallowedtogrowfor-6generationsinYPDbeforeplating.MutantcolonieswereselectedforgrowthonCM(lactate)plates.MediaandGrowthConditions.Syntheticmedia[CM(glu-cose)andCM(lactate)]wereasdescribed(15).ForRNAanalysis,aerobiccultureswereshakenat30'Cat250rpminlargeflasks(10timestheculturevolume)inYPDorinYPRmedium(2%peptone/1%yeastextract/2%raffinose).Cellswereharvestedat107cellspermlafterswirlingfor5minonice.AnaerobicculturesinoculatedwithafreshstationaryphasecultureweregrowninYPDfromaninitialdensityof106cellspermltoafinaldensityof107cellsperml.After15minofvigorousbubblingwithultrapure(99.998%;Linde)nitrogen,thecultureswereshakenatmoderatespeedwithslowbubblingfor7-8hr.Beforeopening,theflaskswerechilledfor15minbyadditionofcrushedicetotheshakingwaterbath.Cellswereharvestedat0C.ExtractionandAnalysisofRNA.RNAwasextractedfromyeastcellsandanalyzedbygelblotasdescribed(15)ontoGene-Screen(NewEnglandNuclear)membranes.Whennecessary,theseblotswerestrippedofprobeasdescribedbyThomas(25)andrehybridizedaftertheusualperiodofpre-hybridization.Insomecases,thiswasrepeatedseveraltimeswithaseriesofdifferentprobes.GeneticAnalysis.Forcomplementationanalysis,mutantswerecuredofYCpCYCI(2.4)1/XandconvertedfromMATatoMATabytransformationwithYEpHO(21).Theparentalstrain,GM-3C-2,isdefectiveinswitchingfromMATatoMATa,sothatYEpHOtransformantswereconvenientlyrecoveredinhomogeneousMATacolonies.AftercuringofYEpHOandretransformingwithYEp13(LEU2),theMATamutantswerematedwithappropriateMATamutantscar-ryingYCpCYCI(2.4)1/X(TRPI),andisogenicdiploids(LEU2,TRPI)wereselected.PairsofmutationswerejudgedtocomplementifdiploidsfailedtogrowonCM(lactate)plates.Becauseofnegligiblesporulationbythesediploids,tetradanalysiswasdoneondiploidsresultingfromcrossesbetweenaRZ5andmutantscarryingYCpCYCI(2.4).Thesecellssporulatedatafrequencyof20%-30%wheninducedtodosobystandardmethods(24).RESULTSCellsoftherespiratory-deficientstrainGM-3C-2(cycl,cyc7)areunabletogrowonnonfermentablemedia.ThisdeficiencycanbecomplementedbytransformationwithachimericplasmidYCpCYCI(2.4)(15),whichcarriestheCYCIgeneona2.4-kbfragment.Thisfragmentalsocontainsaneighboringgene,ANBI,whichiscompletelyrepressedduringaerobicgrowthandinducedduringanaerobicgrowth.Inearlierwork,amutationhadbeenconstructedintheYCpCYCJ(2.4)plasmidinwhichacontrolregionlyingbetweenthetwogeneswasinverted,placingCYCItranscriptionunderthecontrolofthetranscriptionalmodulatoroftheANBIgene,formingaregulatoryfusion(ANBI/CYCI)betweenthetwogenes(Fig.1;ref.15).Inthisconfiguration,CYCImRNAisproducedonlyintheabsenceofoxygen,sothatthephe-notypeofGM-3C-2cellscarryingtheANBI/CYCIfusiononplasmidYCpCYCI(2.4)1/Xisstilloneofcompleterespira-torydeficiency,givingrisetonegligiblegrowthonlactateplates.ThepropertiesofthisfusionprovidedanopportunitytoselectformutantsconstitutiveforexpressionoftheANBIgene.AmutationthatinterferedwiththeaerobicrepressionactingthroughtheANB1modulatorwouldpermitexpressionoftheANBI/CYCJfusioninthepresenceofoxygenand,consequently,allowgrowthonlactateplates.GM-3C-2cellscarryingYCpCYCI(2.4)1/Xweremutagen-izedwithUVradiationorethylmethanesulfonateandplatedonlactateplates.Lactate-positivemutantswereobtainedatSB////~/N~\MAMCSBANBIXSXCYCIH.~"..MCMASI8HABSXCYCIHMAI~MC-liMAMCYCpCYCI(2.4)YCpCYCI(2.4)I/x(ANBI/CYCIfusion)YCpCYCI(2.4)I/s(ANBIinversion)FIG.1.Plasmidscarryingthewild-typeANBIandCYCIgenes,theANBI/CYCIfusion,andtheANBIinversion.ConstructionofplasmidsYCpCYC1(2.4)andYCpCYCJ(2.4)1/Xhasbeende-scribed(15).PlasmidYCpCYCI(2.4)1/Swasconstructedbydigest-ingYCpCYCJ(2.4)withSmaI,religating,andscreeningE.colitrans-formantsforclonescontainingplasmidwiththe1.6-kbSmaIfrag-mentcarryingtheANBIgeneinsertedinthereverseorientation.Thetranscriptionalorientationofthetwogenesisindicatedbythedirectionalarrowsforthedifferentplasmidconstructions.Theel-ementsMAandMcrepresentthemodulatorregionsofthetwogenes,apairofregionsboundedbyXhoIandSmaIsites,whichcontainregulatorysequencesfortheANBIandCYCIgenes,re-spectively.Restrictionsitesareindicatedbythefollowingabbre-viations:S,SmaI;X,XhoI;B,BamHI.afrequencyof_-4.Thesedisplayedawiderangeofgrowthratesonlactateplates.Discriminationofcis-andtrans-ActingConstitutiveMuta-tions.Themutagenesiswascarriedoutoncellsthathadalreadybeentransformedwiththefusionplasmidtoobviatethenecessityoftransforminglargenumbersofcellsaftermutagenesis.Themainpurposeofthemutantselectionwastoobtaintrans-actingmutationsinchromosomalgenesin-volvedinregulationofANBIexpression,butitwasexpectedthatthelactate-positiveisolateswouldalsoincludesomecarryingcis-actingmutationswithintheYCpCYCJ(2.4)1/Xplasmid.Sinceallcis-actingmutantswereexpectedtobedominant,theyweretentativelyscreenedforbymatingwithaRZ5(cycl,cyc7)andbytestingthediploidsforlactate-positivephenotype.Bothdominantandrecessivemutationswerefound;theformerpredominatingafterUVmutagen-esis,thelatterafterethylmethanesulfonatemutagenesis.Inaddition,some"semi-dominant"mutantswerepresent,asrevealedbyslowbutdetectablegrowthofdiploidsonlactateplates.Amoredefinitivecis/transtestinvolvescuringthemutantoftheplasmiditcarries,retransformingwithunmu-tagenizedplasmid,andtestingformutantphenotype.Cellscarryingtrans-actingmutationswillstillconstitutivelyex-presstheANBI/CYCIfusion(lactate'),whilecellsthatoriginallycarriedcis-actingmutationsin

3 theplasmidreverttowild-typeexpression(la
theplasmidreverttowild-typeexpression(lactate-).Forthistest,agroupwasselected,includingdominant,semi-dominant,andrecessivemutants.Aftercuringandretransformation,thesewerenowfoundtofallintothreecategories:trans-recessive,trans-semi-dominant,andcis-dominant.Theselastplasmid-bornemutationswerefoundtobeduetolargedeletionsorinser-Proc.Natl.Acad.Sci.USA81(1984) Proc.Natl.Acad.Sci.USA81(1984)6131tionsandhavenotbeenstudiedindetail.Asubsetofthesemi-dominantandrecessivetrans-actingmutantswerese-lectedforfurthercharacterization.GeneticAnalysisofConstitutivetransMutations.Aftercon-versionofseveralmutantstrainsfromMATatoMATabytransformationwiththeYEpHOplasmid(21),theappropri-ateisogenicMATaandMATamutantsweremated,andtheresultingdiploidswereselectedandtestedforcomplemen-tation.Complementationbetweenmutantswasshownbytherestorationofwild-typeregulationoftheANBI/CYCIfu-sion,apparentasalackofgrowthonlactate.Twocomplementationgroupswerefoundandweredes-ignatedroxiandrox2(forregulationbyoxygen).Therox2mutants,althoughexpressingtheANBI/CYCIfusion,didnotprovetobeconstitutiveforANBIexpression(seebelow)andhavenotbeenstudiedindetail.Wenoted,however,thatallthesemutantsshowedflocculentgrowthmorphologyandthatconstitutiveexpressionofthefusionwasunstable,de-cliningduringmaintenanceonYPDagar.Theroximutants,whichwerestableandnonflocculent,fellintotwoclasses:roxi-a,whicharesemi-dominant,androxl-b,whicharerecessive.Diploidsconstructedbetweenroxi-alandvariousroxi-bmutantsweremuchmorestronglylactate-positivethandiploidsconstructedbetweenroxi-alandtheparentstrain,indicatingthatthetwogroupswereallelic.However,thislackofcomplementationwasambigu-ousbecauserox1-amutantsweresemi-dominant.Toconfirmlinkage,tetradanalysiswasundertakenusingdiploidscon-structedfromstrainsCL-2-15aandCL-1-13b,whichcarrytheroxi-alandroxl-b3mutations,respectively.ThesestrainswerederivedfromcrossesbetweenaRZ5andtheoriginalroxi-alandroxl-b3isolates.ThediploidscarriedYCpCYCI(2.4)1/X,sothathaploidsegregantscouldbescoredforconstitutiveexpressionoftheANBI/CYCIfu-sion.Allofthesegregants(totalof26)whichstillcarriedtheYCpCYCI(2.4)1/Xplasmidaftermeiosisshowedtheroxiphenotype,amonghaploidsfromsixfulltetradsandseveralpartialtetrads,demonstratinglinkagebetweentheroxl-aandroxl-bclassesofmutation.Takentogether,thefunc-tionalnoncomplementationandgeneticlinkageofthephe-notypicallydistinctroxl-aandroxl-bmutationsstronglyindicatethattheyareallelesofasinglegene.PhenotypicAnalysis:ConstitutiveExpressionofANBI.Ac-cordingtotheoriginaldesignoftheselection,mutantsex-pressingtheANBI/CYCIfusioninYCpCYCI(2.4)1/Xwereexpectedtoincludesomethatwouldbeconstitutiveforexpressionfromthewild-typeANBIgene.Totestforthisproperty,mutantsthathadbeencuredoftheYCp-CYCI(2.4)1/XplasmidwereretransformedwithYCp-CYCJ(2.4),whichcarriesthewild-typeANBIandCYCIgenes.RNAfromseveralaerobicallygrownisolateswassubjectedtoRNAblotanalysis.Fig.2showsablotofsomeofthesesamples,probedwithafragmentcarryingtheANBIgene,whichhybridizesbothtotheANBItranscriptandtoanotherhomologoustranscript,tr-J.Thetr-1gene(15)islocatedelsewhereinthegenomeandispositivelyregulatedbyoxygen,althoughitshareshomologywithANBI.Asshown,alltheroxlmutantswereconstitutiveforANBIexpression-i.e.,theyaccumulatedvariableamountsofANBItranscriptinthepresenceofoxygen,whereastherox2mutantswerenormallyregulated.PleiotropicEffects:ConstitutiveExpressionofAerobicGenesinroxl-abutNotinroxl-bCells.Inwild-typecells,expressionofanumberofgenesisinducedbyoxygenandisnegligibleunderanaerobicconditions.AmongthesearethreegenesforwhichcloneswereavailableasprobesforRNAblots:(i)CYCI,(ii)tr-1,and(iii)SOD,thesuperoxidedis-mutasegene.Toexaminetheregulationofexpressionofthesegenesinroxl-a,roxl-b,androx2mutants,repre-sentativestrainstransformedwiththewild-typeplasmidYCpCYCI(2.4)weregrownanaerobicallyandaerobicallyinglucose,andaerobicallyonthecatabolitederepressingsugarraffinose.RNAwasextractedandanalyzedbyRNAblot,usingamembranethatcouldbeefficientlystrippedandreprobed,permittingsemi-quantitativecomparisonofsev-eraldifferenttranscriptsinthesamegelsample.roxl-b,rox2,andGM-3C-2cellswereallsubjecttostrongoxygeninductionoftr-i(Fig.3a),SOD(Fig.3b),andCYCI(Fig.3c),evidentbytheincreasedlevelsinaerobicculturescom-paredtoloworundetectablelevelsinanaerobiccultures(lanes1and2,4and5,and10and11).Incontrast,theroxia-Imutantshowedequallevelsofthethreetranscriptsincellsfromglucoseculturesgrownunderanaerobicandaerobicconditions(lanes7and8).InthecaseofCYCImRNA,aerobiclevelswerelowerthanthosefoundintheparentstrainunderthesameconditions,buttheywereeasilyde-tectable.CataboliterepressionofCYCJmRNAandSODmRNAwasobservedinGM-3C-2andinallthemutantstested,asshownbycomparingRNAfromtheaerobicderepressed(raffinose)andrepressed(glucose)cultures.However,thederepressedlevelsofCYCImRNAinroxla-JwereagainsignificantlylowerthaninGM-3C-2.PleiotropicEffectsofroxi-alDuetoaSingleMutantAllele.Onetrivialexplanationfortheunusualpleiotropiceffectsoftheroxi-almutationisthattheisolatestudiedcarriestwomutationsaffectinganaerobicandaerobicgenesindepen-dently.Sincetightlinkageoftwoindependentnonallelicnmtationswouldnotbeexpected,arelativelyreliabletestoftruepleiotropywastofollowthesegregationofthetwodistinguishableregulatoryphenotypes-constitutiveexpres-sionofANBIandoftheaerobicgenes.Sporeanalysisofdiploidsresultingfromacrossoftheroxi-alandaRZ5wasdonetodeterminewhether.thetworegulatoryphenotypestr-IIANBI02---N202

4 2345678910111213wtFIG.2.ANBIandtr-itrans
2345678910111213wtFIG.2.ANBIandtr-itranscriptsinmutantscarryingYCpCYCI(2.4).roxlmutantsnumbered1,2,3,6,8,9,10,and11,androx2mu-tants4,5,7,12,and13weretransformedwithYCpCYCI(2.4)andgrownaerobicallyinYPD.RNAwasextractedandsubjectedtoRNAblotanalysisusinganick-translatedfiagment[the1.2-kbXhoI/BamHlfragmentfromYCpCYCI(2.4)]carryingtheANBIgeneasprobe.Theparentalcells(wt;GM-3C-2)carryingYCpCYCI(2.4)weregrownanaerobicallyandaerobicallyforcomparison.Genetics:LowryandZitomer 6132Genetics:LowryandZitomerrox-b3rox2rox-alWta.tr-Iibitsee__ANBb.SOD_!~~~~~~~~~~~~~~CYCN202RN202RN202RN202R2345678910II12FIG.3.AnalysisofRNAinselectedmutantsgrownunderan-aerobic,aerobic,andcatabolitederepressedconditions.(a)RNAwasextractedfrommutantandwild-type(GM-3C-2)cellscarryingYCpCYCI(2.4),grownasindicated:N2,anaerobicallyinYPD;02,aerobicallyinYPD;R,aerobicallyinYPR(2%raffinosesubsti-tutedforglucose).RNAwasanalyzedbyRNAblot,usingthesameANBIgeneprobeasinFig.2.(b)Thesameblot,strippedofprobeandrehybridizedwithanick-translated1.2-kbBamHIfragmentcar-ryingtheSODgene,derivedfromplasmidpFL1-cIII-FeS-3(21).(c)Thesameblot,strippedofprobeandrehybridizedwiththenick-translatedpYeCYCI(O.6)plasmid(22),whichcarriesmostofthecodingregionoftheCYCIgene.cosegregated.Haploidsegregantsweredissectedfromsixtetradsandgrownunderanaerobicandaerobicconditions.RNAextractedfromthesecellswasanalyzedonRNAblotsbyhybridizingsuccessivelywiththeCYCJgeneandthenwiththeANBJgene.Althoughtherewasvariationamongthestrainstested,inallcasessegregantsthatwereconsti-tutiveforANBJexpressionwerealsoconstitutiveforexpres-sionoftheaerobicgenes.Conversely,segregantsthatshowedwild-typeregulationforANBJalsoexpressedtheaerobicgenesnormally.Thelinkageofthetworoxi-alphenotypesduringmeiosisindicatesthatthediverseregu-latoryeffectsareduetothepresenceofasinglemutantallele.Fig.4showsanRNAblotofsomeofthesamplesthatwereanalyzed.TetradsobtainedfromcrossesbetweenaRZ5andeithertheroxi-alorroxl-b3mutantsshowednormal2:2segregationofROX1alleles.Theroxi-alalleleremainedrox1-axROXpleiotropic(asshowninFig.4byconstitutiveexpressionoftr-JandANB1,withtheirtranscriptsevidentinanaerobicandaerobiccells,respectively).Theroxl-b3allelecontinuedtorenderonlyANB1expressionconstitutive.JointConstitutivityofANBIandCYCIExpressionDoesNotRequireProximity.Itwaspreviouslyfoundthatwild-typeregulationoftheANB1andCYC1genesisphysicallyinde-pendent,inthesensethattheirrespectiveregulatoryregionscanbeseparatedwithoutaffectingexpression(unpublisheddata).ThiswasshownbyanalysisoftheexpressionofANB1andCYC)fromtheplasmidYCpCYCl(2.4)1/S(Fig.1).Inthisconstruction,aSmaIfragmentcontainingtheANB1geneanditsmodulatorregionwasinsertedinthereverseorientation,producingaseparationof-1.6kbbetweentheANB1andCYCJregulatorysites.RegulationinGM-3C-2cellstransformedwiththisplasmidwasidenticaltothatincellstransformedwiththewild-typegenesonYCpCYCJ(2.4).Eventhoughregulationofthetwogenesthusappearedtobephysicallyindependentintheparentstrain,itstillremainedpossiblethatthenovelcoconstitutiveeffectonANB1andCYC1expressioncausedbytheroxi-almutationwastheresultofinteractionbetweenadisturbedregulatorycomplexattheCYCJmodulatorandtheproximalANB1modulator.ThiswastestedbyanexperimentinwhichRNAextractedfromanaerobicandaerobicroxl-alcellscarryingthe1/Splasmidwascomparedtothatobtained(Fig.3aandc,lanes7-9)fromroxi-alcellscarryingthewild-typeplasmidYCpCYCl(2.4).Anidenticalpattern(notshown)ofconsti-tutiveexpressionwasobserved,indicatingthatthepleio-tropiceffectdoesnotrequirephysicalproximity,butratherinvolvesacommontrans-actingelement.DISCUSSIONTheisolationofmutantsconstitutiveforexpressionoftheanaerobicANBIgenehasrevealedtheexistenceofatrans-actingfactor,theROXJgeneproduct,whichisinvolvedinregulationofANBIandanumberofoxygen-inducedgenes.Thisindicatesthatthereisageneralizedcoordinateregu-lationofoxygen-sensitivegeneseffected,atleastinpart,byacommonmechanism.TheROXIgeneproductmightactatanyofseveraldif-ferentlevels:(i)controloftranscription,(ii)post-transcrip-tionalmodification,(iii)specificbreakdownofmRNA,(iv)translationalcontrol,and(v)productionofaregulatoryef-fector.However,byprocessofelimination,adirectroleinregulationoftranscriptionappearsthemostlikely.Thethirdandfourthpossibilitiesarearguedagainstbyacomparisonbetweentheexpressionofwild-typeandfusedgenes.Thatis,thefactthatroximutationscauseconstitutiveexpressionofboththeANBIgeneandtheANBI/CYCIfusion,butnotofthenormalCYCIgene,indicatesthatROXI-mediatedregulationofexpressionisindependentofthenatureofthetranscript,beingeffectedthroughupstreamroxl-bxROXabcdabcdIIEl--lII01~~~~~~~~6S~~~~~~~~~~aSO~~~~~~~~~~@tr~~~~~~r-iW||*W~94IWO*W*ANBroxI-aroxI-aROXN202N202N202ROXROXN202N202roxl-broxl-bROXN?02N202N?02FIG.4.TetradsanalyzedforexpressionofANBIandtr-1.Mutantsrox1-alandroxl-b3carryingYCpCYCI(2.4)werematedwithaRZ5,diploidsweresporulated,andasciweredissected.SegregantscarryingYCpCYCI(2.4)weregrownanaerobicallyoraerobicallyinYPD.RNAwasextractedandanalyzedbyRNAblot,usingthesameANBIgeneprobeasinFig.2.Thegenotypesinferredfromtheregulationofexpres-sionareshownasroxlorROXI.Proc.Natl.Acad.Sci.USA81(1984) Proc.Natl.Acad.Sci.USA81(1984)6133controlsequencesinstead.Therefore,sinceregulationatthelevelofmRNAtranslationordegradationwouldbepre-sumedtobemRNAsequence-specific,participationbytheROXJgeneproductineitherprocessisunlikely.Controlatthelevelofpost-transcriptionalprocessingisapossibility;however,thereisnoevidenceof

5 ahighermolecularweightprecursor.Perhapst
ahighermolecularweightprecursor.Perhapsthemostlikelyalternativetoaroleintranscrip-tionisthefifthpossibility,thattheROXIgeneproductisinvolvedinthemetabolismofaregulatoryeffector.How-ever,itisdifficulttoexplainthepleiotropiceffectsoftheroxi-almutationintermsofchangesineffectorconcentra-tions,becausethemutationcausesconstitutiveexpressionofbothanaerobicandaerobicgenes.Inwild-typecells,thesegenesarenormallyregulatedinoppositewaysbyoxygenlevels,sothattheirexpressionwouldeitherhavetorespondinoppositewaystoacommoneffectororelsebecontrolledbydifferenteffectors.Ineithercase,amutationcausingachangeinthelevelofasingleeffectorwouldnotbeexpectedtocausetheobservedincreasedexpressionofbothtypesofgenes.Hence,mutationalchangesineffectorlevelscannotaccountinasimplewayforthepleiotropiceffectontheexpressionoftheseoppositelyregulatedgenes,althoughmorecomplexmetabolicexplanationscanstillbeconsid-ered.TheprecedingargumentfavorsarolefortheROXIgeneproductinregulationoftranscriptionofoxygen-inducedandoxygen-inhibitedgenes.AssumingthattheROXIgeneprod-uctinteractswith,orformsapartof,thetranscriptionalapparatus,anumberofplausibleexplanationsfortheob-servedpleiotropiceffectsofroximutationscanbeproposed.ThetwosimplestonesinvokeadualrolefortheROXIfactorinregulationofthetwoclassesofgenes.Bothmodelsareconsistentwiththeoccurrenceofpleiotropicandnonpleio-tropicmutantallelesofROXJ.Inonemodel,thefactorwouldactivateoneclassofgenesortheotherinresponsetodifferenteffectorspresentinaerobicoranaerobiccells(orinresponsetothepresenceorabsenceofasingleeffector).Inanalternativebimodelmechanism,theROXIfactorwouldsimultaneouslyparticipateinrepressionofANBIandac-tivationoftheoxygen-inducedgeneswhencomplexedwithoxygen-inducedeffector.Thepossibilitythatasinglefactorcanfunctionasapartofeitheranactivatingorarepressingcomplexwithdifferentgeneshasaprecedentintheversatilerolesuggestedforthealcomponentofthemating-typecontrolsystem(21).ItisnotpossibletoconcludeatpresentwhetherthecontrolofexpressionmediatedbytheROXIgeneproductispositiveornegative.Aninferencebasedonthedominanceorrecessivenessofregulatorymutationsisclearlynotpossible,becausetheroximutationsincludebothrecessiveandsemi-dominantalleles.Thepropertiesoftruenullmutationswillbeinformativeinthisregard.PossibleConsensusSequencesforOxygen-RegulatedGenes.Thepleiotropicregulatoryphenotypeoftheroxi-almuta-tionprovidesevidenceofacommonmechanismcontrollingexpressionofagroupofoxygen-regulatedgenes.Itmightbeexpectedthattherewouldbesequencehomologybetweentheregulatorysitesofthecoordinatelycontrolledgenes.Infact,astrikinghomologyhasbeenfoundbetweenDNAsegmentsintheupstreamregionsofthreeoxygen-regulatedgenes,CYCI(4),SOD(L.Grivell,personalcommunication),andCYC7(26),afindingthatsupportsthehypothesisofacommonregulatorycomplexactingthroughthesesites.TheupstreamregionoftheANBIgenecontainsnodis-cerniblehomologytothesegmentsalludedtoabove.Thissuggeststhatadifferentrecognitionsequencemaybein-volvedinROXI-mediatedregulationofANBIexpression,whichcouldhelpaccountfortheoppositeresponsetooxy-genintheexpressionoftheanaerobicandaerobicgenes.WearegratefultoRobertaLieberforassistanceingeneticanaly-sis.ThisworkwassupportedbyaNationalInstitutesofHealthandaNationalScienceFoundationresearchgrantandaResearchCa-reerDevelopmentAwardawardedtoR.S.Z.1.Davidson,E.H.,Jacobs,H.T.&Britten,R.J.(1983)Nature(London)301,468-470.2.Hinnebusch,A.G.&Fink,G.R.(1983)J.Biol.Chem.258,5238-5247.3.Donahue,T.,Daves,R.S.,Lucchini,G.&Fink,G.R.(1983)Cell32,89-98.4.Guarente,L.,Lalonde,B.,Gifford,P.&Alani,E.(1984)Cell36,503-511.5.Schurch,A.,Miozzari,J.&Hutter,R.(1974)J.Bacteriol.117,1131-1140.6.Wolfner,M.,Yep,D.,Messenguy,F.&Fink,G.R.(197^)J.Mol.Biol.96,273-290.7.Penn,M.D.,Galgoci,B.&Greer,H.(1983)Proc.Natl.Acad.Sci.USA80;2704-2708.8.Loison,G.,Losson,R.&Lacroute,F.(1980)Curr.Genet.2,39-44.9.Herskowitz,I.&Oshima,Y.(1981)inTheMolecularBiologyoftheYeastSaccharomyces:LifeCycleandInheritance,eds.Strathern,J.N.,Jones,E.W.&Broach,J.R.(ColdSpringHarborLaboratory,ColdSpringHarbor,NY),pp.181-209.10.Douglas,H.C.&Hawthorne,D.C.(1964)Genetics49,837-844.11.Oshima,Y.(1982)inTheMolecularBiologyoftheYeastSaccharomyces:MetabolismandGeneExpression,eds.Stra-them,J.N.,Jones,E.W.&Broach,J.R.(ColdSpringHar-borLaboratory,ColdSpringHarbor,NY),pp.159-180.12.Perlman,D.&Hopper,J.(1979)Cell16,89-95.13.Chisholm,G.&Cooper,T.G.(1982)Mol.Cell.Biol.2,1088-1095.14.Sledziewski,A.,Rytka,J.,Bilinkski,T.,Hortner,H.&Ruis,H.(1981)Curr.Genet.4,19-23.15.Lowry,C.V.,Weiss,J.L.,Walthall,D.A.&Zitomer,R.S.(1983)Proc.Natl.Acad.Sci.USA80,151-155.16.Guarente,L.G.&Mason,T.(1983)Cell32,1279-1286.17.Hortner,H.,Ammerer,G.,Hartter,E.,Hamilton,B.,Rytka,J.,Bilinski,T.&Ruis,H.(1982)Eur.J.Biochem.128,179-184.18.Saltzgaber-Muller,J.&Schatz,G.(1978)J.Biol.Chem.253,305-310.19.Faye,G.,Leung,D.W.,Tatchell,K.,Hall,B.D.&Smith,M.(1981)Proc.Natl.Acad.Sci.USA78,2258-2262.20.Broach,J.R.,Strathern,J.N.&Hicks,J.B.(1979)Gene8,121-133.21.Jensen,R.J.,Sprague,G.F.&Herskowitz,I.(1983)Proc.Natl.Acad.Sci.USA80,3035-3039.22.vanLoon,A.P.G.M.,Maarse,A.C.,Riezeman,H.&Gri-vell,L.A.(1983)Gene26,261-272.23.Smith,M.,Leung,D.,Gilham,S.,Astell,C.,Montgomery,D.&Hall,B.(1979)Cell16,753-761.24.Sherman,F.,Fink,G.&Hicks,J.(1982)MethodsinYeastGenetics(ColdSpringHarborLaboratory,ColdSpringHar-bor,NY).25.Thomas,P.S.(1980)Proc.Natl.Acad.Sci.USA77,5201-5205.26.Wright,C.F.&Zitomer,R.S.(1984)Mol.Cell.Biol.,inpress.Genetics:LowryandZitome