MOLECULARANDCELLULARBIOLOGY,JUlY1992,p.3006-30140270-7306/92/073006-09 - PDF document

MOLECULARANDCELLULARBIOLOGY,JUlY1992,p.3
MOLECULARANDCELLULARBIOLOGY,JUlY1992,p.3006-30140270-7306/92/073006-09$02.00/0CopyrightC1992,AmericanSocietyforMicrobiologyFusedProteinDomainsInhibitDNABindingbyLexAERICAA.GOLEMISANDROGERBRENT*DepartmentofMolecularBiology,MassachusettsGeneralHospital,50BlossomStreet,Boston,Massachusetts02114,*andDepartmentofGenetics,HarvardMedicalSchool,Boston,Massachusetts02115Received19December1991/Accepted8April1992ManystudiesoftranscriptionactivationemployfusionsofactivationdomainstoDNAbindingdomainsderivedfromthebacterialrepressorLexAandtheyeastactivatorGALA.SuchstudiesoftenimplicitlyassumethatDNAbindingbythechimericproteinsisequivalenttothatoftheproteindonatingtheDNAbindingmoiety.Todirectlyinvestigatethisissue,wecomparedoperatorbindingbyaseriesofLexA-derivativeproteinstooperatorbindingbynativeLexA,byusingbothinvivoandinvitroassays.WeshowthatoperatorbindingbymanyproteinssuchasLexA-Myc,LexA-Fos,andLexA-Bicoidisseverelyimpaired,whilebindingofotherLexA-derivativeproteins,suchasthosethatcarrybacteriallyencodedacidicsequences("acidblobs"),isnot.OurresultsalsoshowthatDNAbindingbyLexAderivativesthatcontaintheLexAcarboxy-terminaldimerizationdomain(aminoacids88to202)isconsiderablystrongerthanbindingbyfusionsthatlackitandthatheterologousdimerizationmotifscannotsubstitutefortheLexA88202function.TheseresultssuggesttheneedtoreevaluatesomepreviousstudiesofactivationthatemployedLexAderivativesandmodificationstorecentexperimentalapproachesthatuseLexAandGALAderivativestodetectandstudyprotein-proteininteractions.ChimericproteinsthatcontaintheDNAbindingdomainsofthebacterialLexArepressor(10)ortheyeastGAL4activator(36)havefacilitatedthestudyoftranscriptionregulation.WeandothershaveusedtheabilitytoseparateDNAbindingfromotherfunctionstoidentifyandmapactivationdomains(10,30,34,43,61,65),ligandbindingdomains(20),anddomainsthatinteractwithotherproteins(16,49,54).LexAandGAL4fusionproteinshavealsobeenusedinscreeningprocedurestoidentifyactivatingmotifsencodedbyrandombacterialopenreadingframes(50),tocharacterizethestrengthofactivationdomainsofproteinsinvariouscelltypes(34,58),andtoprovideaninternalstandardinstudiesofsiterecognitionbyproteinsthatcontainasecondDNAbindingdomain(24,33).Veryre-cently,weandothershaveemployedLexAandGALAderivativesas"baits"ininteractortrapassaystoidentifyproteinsthatcomplexwithknownproteins(12,23).LexAfusionproteinstypicallycontaineithertheLexAamino-terminalDNAbindingdomain(LexA1l87)(10,45,69)orthecompleteprotein(LexA1l202)whichalsoincludesadimerizationdomain(4,46,69,74).Likemanyprokaryoticrepressors(26,32,56,66,67,72),nativeLexAbindsasadimertoanoperatorthatconsistsoftwodyadsymmetrichalf-sites(consensussequenceCTGTNNNNNNNNACAG)(5,8,77).LexAderivativesareassayedfortranscriptionactivationbyusingreportergenesthatcarryoneormoreLexAoperatorsupstreamofthetranscriptionstartsiteofagenesuchaslacZ(foryeastassays)orCAT(formammaliancellassays)(10,20).FornonactivatingLexAderivatives,DNAbindingcanbeassayedbyusingarepressionorblockingassay,inwhichbindingoftheLexAderivativetooperatorsequenceslocatedbetweenanupstreamactivationsite(UAS)andthetranscriptionstartsiteofareportergenediminishesitstranscription(6,9,36).OnecommonuseofLexAorGALAfusionproteinshasbeentocomparetherelativestrengthofactivationdomains*Correspondingauthor.betweendifferentproteinsorbetweendifferentdeletionderivativesofthesameprotein.ImplicitinsuchexperimentsisanassumptionthatheterologousfusiondomainsdonotaffectDNAbindingbytheLexAorGAL4moiety;thatis,theDNAbindingmoietyfunctionsasanindependentdo-main.However,inthecourseofconductingadetailedanalysisofaseriesofLexA-Mycderivatives(22),weobtainedresultsthatsuggestedthatthisassumptionisnotcorrect.HerewepresentastudyofvariablesthataffectoperatorbindingbyseveraldifferentLexAderivativeproteins.IntheseexperimentswecomparedoperatorbindingbynativeLexAtobindingbyanumberofLexAderivatives,invivobyusingatranscriptionactivationassay,andinvitrobyusingagelmobilityshiftassay.WethenexaminedthebindingofanumberofLexAderivativestoasetofmutantoperators.Theresultsshowthatmanyfusedmoietiesdra-maticallyreducetheabilityoftheLexAmoietytobindtheLexAoperator.TheyalsoshowthattheLexA88-202dimer-izationdomainpromoteshigh-affinityoperatorbindingandthatdimerizationfunctionsprovidedbyheterologouspro-teinscannotsubstituteforthisLexA88-202-specificfunction.TheseresultssuggesttheorganizationofnativeLexAisnotstrictlymodular,inthattheidentityofonedomaincanaffectthefunctionoftheother;inasimplemodel,aspecificgeometryoftheLexA88202dimerizationdomainencouragesaspatiallyprecisealignmentoftheLexAl-87domainsonoperators.WediscusstherelevanceofthesefindingsfortheuseofLexAfusionproteinstostudytranscriptionregulationandfortheiruseinrecentlydevelopedmethodstodetectandstudyprotein-proteininteractions.MATERIALSANDMETHODSFusionproteins.Manyoftheconstructionsusedinthisstudyhavebeenpreviouslydescribed;allfusionsandreport-ersarerepresentedschematicallyinFig.1.Allconstructionsweremadebyusingstandardmethods(1,63).Fusionproteinswereexpressedfromthestrongconstitu

tiveADH3006Vol.12,No.7DNABINDINGBYLexAF
tiveADH3006Vol.12,No.7DNABINDINGBYLexAFUSIONPROTEINS3007A87GAL487GCN487BIcoid87cFos87cMyc87vMyc202vMyc202vMycAC202BIcoid202-B6202-B7202-B42202-PRD202-PRD/HD202-PLLexABR-noopsR-lopR-2opsnoop1/2oplop79122812289I3944051425514255136222891422916224UASGALLaoUAS3ALUASGALopsLWZionsnna~~~~~Lc2ops---FIG.1.(A)LexAderivatives.BlackboxesinthefigureindicateLexAsequences.Openboxesindicatefusionmoieties.Numbersaboveopenboxesindicatetheresiduesfromthenativeproteinsincludedineachfusionprotein.Mostoftheexpressionplasmidsusedinthisstudyhavebeenpreviouslydescribed;seeMaterialsandMethodsfordetails.(B)Reporterplasmids.Thefirstthreeplasmidswereusedforrepressionassays;thenextfivewereusedforactivationassays.ClosedboxesindicatetheGAL4UAS.BoxesindicatetheLexAoperators,whicharedescribedinmoredetailinMaterialsandMethods.Openboxes,wild-typeLexAoperator;overlappingopenboxes,overlappingorColEloperator,boxwithrighthalfshaded,1/2op;boxwithasteriskinrighthalf,opm.promoter.The2,umplasmiddirectingsynthesisofLexA1l87-GAL4isRB1027(10);theCENARSformisKL1027(42).The2,umplasmidexpressingLexAl17-BicoidisSH32-3(24);theCENARSformandthe2gumplasmidexpressingLexA1l202-Bicoid(plasmid14)aregenerousgiftsofSteveHanes(25).The2gmplasmidexpressingLexA1_87-cFosisVR1001(43);theCENARSformisKL1001(42).The2gmplasmidexpressingLexAl-87-cMycisVR1004(43);theCEN-ARSformisKL1004(42).The2,umplasmidexpressingLexA1l7-GCN4containsresidues12to281ofGCN4(28).The2,umplasmidexpressingLexAl-87-vMycisKA409(43).The2,umplasmidexpressingLexAispRB500(8);cenLexAistheCENARSversion(64).The2,umplasmidsexpressingLexAl202-vMycandLexAl-202-vMycACcontainfull-lengthLexAfusedto1to415and1to362ofavianvMyc(57),respectively(22).The2,umplasmidsexpressingLexA1l202-B6,LexA1l202-B7,LexA1l202-B42,andtheirexpressionvectorLexA1_202-PLhavebeende-scribed(60).The2,umplasmidsexpressingLexAl202-PRDandLexAl-202-PRD/HDcontainresidues2to95encodingthePRDboxandresidues2to160encodingthePRDboxandhomeodomainofbicoid(3)fusedtoLexA1l202attheBamHIsiteofLexA1l202-PL(17).Reporterplasmids.Repressionreportersareasfollows:thereporterplasmidR-noopisA20B(79).R-1opisA20B+lop(9);itcontainsasingleLexAoperatorclonedintoanXhoIsitebetweentheGALlUASandTATAofA20B.R-2opsisJK1O1(33)andcontainstwooverlappingLexAoperators(asfoundupstreamofthecolElgene[15])intheA20BXhoIsite.Activationreportersareasfollows:theplasmidnoopisLR1A1(79).lopis1840(10,43);itcontainsasingleLexAoperatorinsertedintheXhoIcloningsiteofLR1A1.lopmisaweakOcmutation(75);itisidenticaltolop,exceptthatintheLexArecognitionsequenceCTG-TATGTACATACAGT,theGshowninboldfacehasbeenchangedtoaC(80).l/2opisalsoidenticalexceptthathalftheLexAoperatorhasbeendestroyedbymultiplebasesubstitutions(CTGTATCTCGATATCC)(25).2opscontainsthecolElLexAoperatorsinsertedintotheXhoIsiteofLR1A1.Expressionanddetectionoffusionproteins.TheSaccharo-mycescerevisiaestrainEGY40(ura3trplhis3leu2)wastransformedbystandardLiOAcmethods(31).ExpressionoffusionproteinswasmonitoredbyWesternimmunoblotting(65)of1mlofmid-log-phaseculture(opticaldensityat595nm[OD595],0.6)byusinganti-LexAantiserum(9)andalkalinephosphatedetection(65).AbsoluteconcentrationsofLexAandderivativeswereapproximatedfromWesternblotsbycomparisonwithserialdilutionsofpurified,quan-titatedLexA(8).Thisanti-LexAantiseruminteractsprepon-derantlywiththeLexA1_87moiety(21),sotheamountoffusionproteininextractsofLexAl-87derivativescouldbeaccuratelyestimatedbycomparisontotheamountofpuri-fiedLexA1l202usedasastandard.Intracellularandintranu-clearconcentrationsofLexAderivativeswereestimatedbyassumingacellvolumeof20,im3(78)andanestimatednuclearvolumeof2,im3(9).Activationandrepressionassays.Theassayforactivationwasperformedaspreviouslydescribed(10).P-Galactosidasevaluesshownaretheaverageofnineseparatetransfor-mants,derivedduringassaysperformedonthreeseparateoccasions.Cellstransformedwith2,umLexAl-7-GAL4andCENARSLexA1l7-GAL4wereassayedbothonglucoseandongalactose.Resultsonbothmediawerequalitativelyidentical,althoughactualp-galactosidasevalueswereslightlylowerongalactosemedia;valuesreportedarethosedeterminedonglucose.Aspreviouslyreported,�LexA17-GAL4expressedfromtheADH1promoteractivatesonglucosebecausetheC-terminaldomainofGAL4(41)isoverexpressedsufficientlytotitrateGAL80(10,49,62).Theassayforrepressionwasperformedessentiallyaspreviouslydescribed(9).Valuesweredeterminedforeightindependentcolonies,inassaysperformedontwoseparateoccasions.Cellsweregrownontheappropriateselectivemediacontainingglucosetosaturatedovernightcultures;thesecultureswerespundown,washedwithwater,andVOL.12,19923008GOLEMISANDBRENTusedtostartculturesatanOD595of0.15ingalactosemedium.CultureswereharvestedatanOD595of0.5to0.6,and0-galactosidasevaluesweredetermined.Inbothactiva-tionandrepressionassays,lessthan25%variabilitywasobtainedbetweenvaluesforindividualcoloniesexpressinggivenfusionproteins.Proteinextracts.Cultures(500ml)ofyeastexpressingappropriatefusionsweregrowntoanOD595of0.6.Yeastcellswerepelletedfor5minat3,000xg,washedonceindH2O,andresuspendedin5mloflysisbuffer(20mMHEPES[

N-2-hydroxyethylpiperazine-N'-2-ethanesu
N-2-hydroxyethylpiperazine-N'-2-ethanesulfonicacid;pH8.0],5mMEDTA[pH8.0],7mM3-mercaptoeth-anol,10%glycerol,100mMKCI,1mMphenylmethylsulfo-nylfluoride,0.6,uMleupeptin,1,ugofpepstatinAperml).Glassbeads(5g)wereaddedtoeachpreparation;yeastcellswerevortexedathighspeedat4°Cfor15minandthenspunfor10minat4°Cat10,000xgtoremovecrudedebris.Thesupernatantwasspunat250,000xgfor3hat0°C(50,000rpminanSW55);theresultingsupematantwasconcentrated-20-foldinaCentricon-10(Amicon,Inc.),aliquoted,andfrozenat-70°C.ConcentrationandintegrityofLexAanditsderivativesweredeterminedbyprobingWesternblotscon-taining2,ulofeachextractwithanti-LexAantiserumasdescribedabove.Gelmobilityshiftassay.Thegeneraloutlineoftheprotocolisthatdescribedpreviously(18,19).Briefly,thelopsite,adouble-stranded,blunt-endedoligonucleotidewhoseplus-strandsequenceis5'-AAAAGTACTA[CTGTATATACATACAG]TGATATCCCC-3',wasusedasawild-typeoperatorforbindingintheassay;inthissequence,theLexAbindingsiteisinbrackets,andGCbasepairswhoseG'sareprotectedbyboundLexAfrommethylationbydimethylsul-fateareindicatedinboldface(8).Thelopmsite,containingasinglebasechange(showninboldface)relativetothelopsite(CTGTATATACATACAC),wasusedinsomeexperi-ments(seeResults).Bindingreactionsweredonein20,ulfor30minat30°Cinabuffercomposedof4%glycerol,1mMEDTA(pH8.0),10mM,BME,10mMTris-HCl(pH7.6),5mMCaCl2,and100mMNaCltowhichextractwasaddedtogiveafinalfusionproteinconcentrationeitherof2to10nM(-0.1pmolperreaction)(Fig.2)orof0.2pMto2nM(Fig.3).Inbindingreactions,20,000to50,000cpm(-1fmol)of-y-32P-labelledoligonucleotidewasusedforbinding,and1,ug(-1,000-foldexcessoverlabelledprobebyweight)ofdouble-strandedpoly(dI-dC)waspresenttocompeteforbindingbynonspe-cificDNAbindingproteins.Inaddition,incompetitionassays,100ng(4pmol)ofthefollowingunlabelleddouble-strandedoligonucleotideswasalsoincludedinthebindingreactionslop;lopm;Xop,whichisidenticalto1/2exceptthatthethreeDMS-protectedG.CbasepairsshownaboveintheLexAoperatorbindingsitehavebeenalteredtothebasesshowninboldface(ATATATATACATACAT);orNS,ablunt-ended30-merthatcarriedaconsensuspalindromicthyroidresponseelement(GGGGATCAGGTCATGACCTGGATCCTCTAG)(52).Reactionswererunona6%poly-acrylamidegel(acryl-bis,29:1bufferedin0.25xTBE[1,63]).RESULTSManyLexAderivativesbindoperatorspoorlyinvivo.WefirstwishedtodeterminewhetherLexAderivativesboundtightlytotheiroperatorsinvivo.Invitro,nativeLexAhasahighaffinityfornaturallyoccurringoperators(Kdsforbind-ingtotheoperatorsusedinourreportersrangefromof2x10-10M[2ops][15]to2x10-9M[lop][8,9]).IfLexAderivativesboundwithsimilarefficiency,wewouldpredictthatevenrelativelylowintracellularconcentrations(-5x10'M)oftheseproteinsshouldcauseoperatorstobealmostcompletelyoccupied;thus,increasesabovesuchlevelsofLexAderivativeconcentrationwouldnotcauseincreasesintranscriptionalactivation.Tomeasurebindinginvivo,weexpressedfivedifferentLexAderivativesfromeitherlow-copy-number(CENARS)orhigh-copy-number(2,um)plasmidsinyeastcells(Fig.1;Table1)andmeasuredactivationofthelopand2opreporters.IntracellularconcentrationsoftheproteinswereestimatedfromWesternblotanalysisofyeastcellextractsbyusingpurifiedLexAasastandard(seeMaterialsandMethods);calculatedvaluesareshowninTable1.Anumberoftheproteins(LexA-cFos,-Bicoid,-vMyc,and-cMyc)containfunctionalnuclearlocalizationsequences(14,27,59,73),andatleasttheLexA-cFosandLexA-Mycfusionproteinsarenuclearlocalizedinmammaliancells(22).Sincethenuclearlocalizationsequencesofhighereukaryotesfunctioninyeasts(53,76),weexpectthattheintranuclearconcen-trationsoftheseproteinsshouldbeapproximately5-to10-foldhigherthaniftheywereuniformlydistributedthroughoutthecell.Wethusestimatedthatnuclearconcen-trationsofthefusionproteinsrangedfrom0.02x10-6to2.8x10-6MforproteinsexpressedfromCENARSplasmids(lowconcentrations)to0.7x10-6to9.0x10-6Mforproteinsexpressedfrom2pmplasmids(highconcentra-tions).Whenexpressedatlowconcentrations,allfiveLexAderivativesactivatedtranscriptionpoorlyornotatall(Table1).Incontrast,athigherconcentrations,allfiveproteinsactivatedtranscription,somemorestronglythanothers.Forexample,LexA1l87-cFosdoesnotactivatestronglyatthelowerconcentrationbutdoesactivatestronglyatthehigherconcentration(Table1).Wedonotknowwhethertheactivationobservedatthehigherconcentrationsreflectsoperatorsaturation,butweconcludefromthisresultthatatthelowerconcentrationsoperatorsarepredominantlyunoc-cupied.Onthebasisoftheinferrednuclearconcentrations,thissuggeststhatLexAderivativesdonotsaturateoperatorbindingatconcentrationsexceedingbyatleast2ordersofmagnitudethosepredictedtogivehalf-maximaloperatoroccupancyfornativeLexA.NativeLexAbindsoperatorsmoreefficientlythanLexAderivativesinvivo.TotestwhethernativeLexAbounditsoperatorswiththepredictedaffinityinvivo,weusedatranscriptionrepressionorblockingassay.ThisassayexploitsthefactthatLexAboundtoeitherasingleoperator(R-lop)ortwooperators(R-2op)placeddownstreamofUASGALblocksactivationbyendogenousGALAproteinboundtotheUAS(6,9).WeexpressedLexAfromCENARSor2,umplasmids;theserespectivelydirectedLexAexpressiontopresume

dintranu-clearconcentrationsof8x10-8Mand
dintranu-clearconcentrationsof8x10-8Mand2x10-6M,comparabletothoseoftheLexAderivativesdescribedabove.Inagreementwithpreviouswork,ona2operatortarget,nativeLexArepressedreportergenesbyafactorof20-fold,(Table2),suggestingthatevenatthelowconcen-trationofLexA,lessthan1in20oftheoperatorsareunoccupied.WealsodirectlycomparedoperatorbindingbyaLexAderivativeproteintothatofnativeLexAbyinvivocompe-tition(Table1).HighlevelsofLexA1l87-GAL4(9x10-7M)wereexpressedincellsthatalsoexpressedlowlevelsofnativeLexA(8x10-8M).Inthisexperiment,becausebothproteinsbindthesameoperators,thelevelsofactivationofMOL.CELL.BIOL.DNABINDINGBYLexAFUSIONPROTEINS3009AProteinCompetitorElopXopnsBProteinCompetitorlopxop_-nsLexA202PL202-B6202-B42I|IIIDABC++++F~++++++++12345678910o111213141516171819g20~~*,S1vIe!Lex20H202H/PR2OvMcAFTF1F1F1E2L34ex67A90H102D11D1234561112_Fi;_lpFIG.2.MobilityshiftassayofLexAandLexAderivatives.Thelopsitewasusedasaprobeintheseexperiments.Lane1ineachpanelcontainsunboundprobe.Allotherlanescontain2to10nMtheindicatedLexAderivative:LexA(panelsAandB,lanes2to5and18to20),LexA1l202-PL(panelA,lanes6to9and18),LexAl202-B6(panelA,lanes10to13and19),LexA1l202-B42(panelA,lanes14to17and20),LexA1202-HD(panelB,lanes6to9and18),LexA1,202-HD/PRD(panelB,lanes10to13and19),andLexA1l202-vMycAC(panelB,lanes14to17and20).ThelanesdesignatedAtoC(A)andDtoF(B)representmixturesoftheLexAextractwitheachoftheotherextractsused,todemonstratethatfailuretoobtainspecificbindingwithsomeextractswasnotduetothepresenceofinhibitorsintheextracts;thus,A,LexA+LexAl202-PL;B,LexA+LexAl202-B6;C,LexA+LexAl202-B42;D,LexA+LexAl202-HD;E,LexA+LexAl1202-HD/PRD;F,LexA+LexAl202-vMycAC.Laneseithercontainnocompetitor(lanes1,2,6,10,14,and17to20)or100ngofspecificlopcompetitor(lanes3,7,11,and15),nonspecificXopcompetitor(lanes4,8,12,and16),ornonspecificNScompetitor(lanes5,9,13,and17).Noneofthespecificbandsobservedcompetedwithsingle-strandedspecificlopcompetitor(notshown).thereportergeneshouldgivearelativemeasureofthesiteoccupancybyLexAandLexAl-7-GAL4.Wefoundthatactivationofthe2opsreporterbyLexAl-7-GAL4wascompletelyblocked(U)inthepresenceofnativeLexA,whereasintheabsenceofnativeLexA,thesamereportergenewasactivatedbyLexA1,7-GALAtohighlevels
Tj;&#x /F1;&#x 11.;€ T; 70;&#x.34 ;&#xTz -;�.42;&#x Ts ;.5; 0 ;&#xTd 0;(1,000U)(Table1).Thus,eventhoughLexAl87-GAIAwasexpressedinsubstantialexcessoverLexA,LexAcompletelyinhibiteditsabilitytoactivatetranscription.WeinterprettheseresultstomeanthatnativeLexAboundtheoperatorswithafargreateraffinitythandidLexA1l7-GAL4.ManyLexAderivativesbindoperatorswithlowaffinityinvitro.TheexperimentsdescribedabovestronglysuggestedthatLexAderivativesboundoperatorsitesinvivolesswellthandidnativeLexA.Thisdecreaseinbindingcouldindi-catethatLexAderivativesboundoperatorswithloweraffinityorthattheywerepreventedfrominteractingwithoperatorsbecausetheyweresequesteredbyaninteractionwithsomeothercellularcomponent(seeDiscussion).Todistinguishbetweentheseideas,weexaminedoperatorbindinginvitrowithagelmobilityshiftassay(18,19)byusingwholecellextractsofyeastexpressingdifferentLexAderivatives.FromWesternblots,weestimatedtheconcen-trationofLexAderivativesintheseextractstobeat2x10-9to10x10-9M(notshown).A36-bpdouble-strandedoligonucleotidethatcontainedthelopsitewasusedtoassaybinding.UndertheseconditionsnativeLexAbound50%ofthelopsiteatamonomerconcentrationof-2x10-9M(Fig.2)andbounddetectablequantitiesofoperatorataVOL.12,19923010GOLEMISANDBRENTProbeLexA[M]CompetitorlPloprnopoPmO)_(NmO)Nr(liOooOOOOaC)C)OcCCXXXX(XX(XX(XXXXXCX(N(N(N(N4N(N(N7N(NA(N('(NN++++i2345657BsI811112;714TABLE2.RepressionassaysP-Galactosidase(U)PlasmidR-lopR-2opVectoronly7183242,um23133CENARS18217FIG.3.ComparisonofLexAbindingtolopandlopm.Siteswerelop(lanes1to7)orlopm(lanes8to14);concentrationsofLexAusedineachbindingreaction(seeMaterialsandMethods)aregivenabovetherespectivelanes.Atotalof100ngofeitherlop(lanes6and13)orlopm(lanes7and14)wasusedinsomereactionsasacompetitor.concentrationof-2x10-11M(Fig.3).ThisaffinityisingoodagreementwithpreviouslyreportedvaluesdeterminedforpurifiedLexAproteininDNaseIprotectionexperimentsinvitro(4,15).Invitro,mostLexAderivativestested,includingLexAl-202-PRD,LexAl-202-PRD/HD,andLexAl202-vMycAC(Fig.2),andLexA187-GALAandLexAl202-vMyc(notshown),didnotgivespecificcomplexeswithoperator.TABLE1.Activationassays'LexAderivativeand3-Galactosidase(U)Protein/cellplasmidlop2p(nM)87GAL42,um1,1521,192900CENARS203220087Bicoid2,um122415rLCENARS20NL87cFos2,um7801,000900IvLCENARS1240NvL87cMyc2pLm81683rLCENARS35NL202vMyc2p.m64296700vLCENARS1648280NLLexA(native)2,um2,000CENARS8087GAL4(2,um)+LexAND(CENARS)aSeeMaterialsandMethodsandFig.1.NL,containsnuclearlocalizationsignals;ND,notdone.Westernblotsofthebindingreactionsshowedthesefusionproteinstobeintact(notshown).OnlythreeLexAderiva-tivesformedspecificcomplexeswithoperator:two"acidblob"proteins,LexAl202-B6andLexAl202-B42(60),andaLexAderivativethatcontainedashort24-amino-acidtailencodedbythepolylinkeroftheLexAfusionvector(pLexA

l202-PL).Bindingofthesethreederivativesw
l202-PL).BindingofthesethreederivativeswasonlyslightlyimpairedrelativetothatofnativeLexA,andweestimatetheirKdsforoperatorbindingtobe_10-8M.LexAl202-B6andLexAl202-B42reproduciblygavethreespecificoperator-containingbands.Westernblotanalysisonnativeanddenaturinggelsofmockbindingreactionsre-vealedthatbothoftheseLexAderivativesranassinglebandsintheabsenceofoperator(notshown);wedonotpresentlyunderstandtheoriginofthethreeoperator-depen-dentspecies.LexAl-87andLexA,-202derivativesinteractdifferentlywithoperators.WecomparedactivationbyanumberofLexAderivativesthatcontainedeitherLexA1l7orLexAl202(expressedfromhigh-copy-number2,umplasmids)ofasetofreporterconstructscontainingoperatorvariants.Thesere-porterswerelop;2ops(2overlappingLexAoperators[15]);lopm,anoperatorinwhichonehalf-siteismildlydisrupted(seebelow)byapointmutation(4,5);and1I2op,inwhicheveryconsensusnucleotideinonehalf-sitewaschanged.AsshowninTable3,LexAderivativesdisplayedseveraldifferentpatternsofactivationfromthesereporters.AllTABLE3.ActivationassaysoncomplextargetsaLexAP-Galactosidase(U)derivative'I/2oplopmlop2opClass187BicoidND293113587cMycND81687vMycND52Class287GAL43601,1521,19287GCN43032056087cFos807801,000Class3202vMycND7264296202BicoidND12193135202-B61681961,748202-B7ND148260860202vMycAC7369681,064202-B426489201,552202-PLNDLexA(native)NDaThefusionproteinsandreportersusedaredescribedinMaterialsandMethodsandareshowninFig.1.ND,notdone.Nofusionproteinactivatedthenoopreporter,whichlacksLexAoperators(notshown).bAllderivativeswereexpressedfromhigh-copy-number2p.mplasmids.MOL.CELL.BIOL.MO.DNABINDINGBYLexAFUSIONPROTEINS3011LexA1l87derivativesfellintotwoclasses.Class1proteins(LexA1_87-Bicoid,LexA1l87-vMyc,andLexA1l7-cMyc)ac-tivatedallthereporterspoorlyandarediscussedbelow.Class2proteins(LexAl187-GALA,LexAl-87-GCN4,andLexA1l87-cFos)didnotactivatethe1/2opreporter(withoneexception)andactivatedthelopmreporteronlyweakly.Theyactivatedthelopreporterstronglyandthe2opre-porteronlyslightlymorestrongly.Theexceptionalclass2protein,LexA1_87-GAL4,activatedthe1/2opreporterex-tremelyweaklybutreproducibly(Table3andDiscussion).AllLexA1l202derivativesfellintoclass3.NoneoftheseproteinsactivatedthelI2opreporter.Theseproteinsacti-vatedthelopmandlopreportersmoderatelystronglyandtocomparablelevels.Someclass3proteins(LexA1l202-B6,LexA1l202-Bicoid,LexA1_202-B7,andLexA1l202-vMyc)ac-tivatedthe2opreportersubstantiallymorestronglythanthelop,whileothers(LexAl1202-B42andLexAl-202vMAC)didnot;probablybecauseexpressionof,B-galactosidasedi-rectedbythelatterproteinsfromthelopreporterwasalreadysaturated(1,200to1,500U[21,25]).WethencomparedbindingbynativeLexAtotwoofthesesitesinvitro(Fig.3).Measuredby1/2maximaloperatorbinding,underourconditionstheaffinityofLexAforthelopmsitedifferedonlyslightlyfromitsaffinityforthelopsite.Incompetitionassays,thelopmsitewas5-to10-foldlesseffectiveasacompetitorthanthelopsite.However,byeitherassay,thismutationinthelopmsitedoesnotseverelyimpairoperatorbinding.Takenwithourinvivodata,theseresultsshowthatbindingbyLexAandLexAderivativescontainingtheLexA88-202domainisrelativelyinsensitivetoaweakoperatormutationwhilebindingbyderivativeslackingthisdomainissignificantlyaffected.DISCUSSIONOperatorbindingbymostLexAderivativesisimpaired.UsingtranscriptionactivationinyeasttomonitorDNAbindinginvivoandgelmobilityshiftassaystomeasureDNAbindinginvitro,wefoundthatoperatorbindingbymostLexA1-87andLexA1l202fusionproteinsweexaminedisimpairedrelativetothatofnativeLexA.Thisassertionisbasedonthefollowingobservations.First,fusionproteinsthatactivatedstronglywhenexpressedathighlevelsdidnotactivate(ordidsoverypoorly)whenexpressedatlowerconcentrations,despitethefactthattheestimatedlowernuclearconcentrationswereasmuchas2to3ordersofmagnitudehigherthantheoperatorbindingKdfornativeLexA.Thus,atthelowerconcentrations,eventhoughpresentinexcess,thefusionproteinsdidnotsaturatetheiroperatorsites.Second,aninvivocompetitionexperimentshowedthatLexA1_87-GAL4,evenwhenexpressedata10-foldexcessovernativeLexA,didnotcompeteefficientlywithnativeLexAforoperatorbinding.Third,invitrobindingassaysfailedtodetectbindingtooperator-containingoligonucleotidesformanyfusionproteins,includingfusionstobothLexA1l7andLexA1l202,underconditionsinwhichnativeLexAboundavidly.RelativetobindingofnativeLexA,evenbindingbymostLexA1l202derivativesexaminedwassubstantiallyimpaired(Table3andFig.2).Itisworthnoting,however,thattwoofthefusedmoietiesweexamined-vMycandBicoid-acti-vatedmorestronglywhentheywerefusedtoLexA1l202thantoLexA1l87(Table3);asweandothershaveargued,suchdifferencesinactivationmostlikelyresultfromimprovedoperatoroccupancyconferredbytheLexA88202domain(33,60)(seebelow).TheLexAcarboxy-terminaldimerizationdomainspecifi-callypromotesoperatorbinding.ManypreviousstudiesofLexAandLexAderivativeshavesuggestedthathigh-affinityoperatorbindingrequiresadimerizationfunction,whichinnativeLexAisprovidedbytheLexA88-202domain(8,45-47,60,68,69,74).Inonesimpleview,theenhancementofDNAbindingcontributedbytheLexA88-202domainrepr

esentsadimerization-mediatedincreaseinth
esentsadimerization-mediatedincreaseinthelocalconcentrationofLexAmoleculesnearanoperator,suchthatbindingofafirstLexAproteintoanoperatorhalf-sitewillautomaticallyprovidea"tethered"secondLexAproteintofillthesecondhalf-site.Inasecondview,theLexA88-202domaincontrib-utesmoreactivelytoDNAbindingbyhelpingtopositionLexA1_87domainsinaconformationrequiredforstablebindingonoperatorhalf-sites.Theresultsofthisstudyprovidesomesupportforthesecondview.Intherelevantexperiments(Table3),wecomparedoper-atorbindingbyLexAderivativesinwhichahomodimeri-zationmotifwasprovidedbynativeLexAsequences(LexAj_202fusions),wasprovidedonlybyaheterologousmoiety(forinstance,LexAl-87-GCN4[29]andLexAl-87-GAL4[10]),orinwhichnostrongdimerizationmotifwaspresent(e.g.,LexAl87-vMycandLexAl-87-cFos[55][seebelow]).Weassayedoperatorbindingbyexaminingactiva-tionbyLexAderivativesofreportersthatcontaineddiffer-entoperatorvariants.Wereasonedthatifthesecondviewofoperatorbindingdescribedabovewastrue,theninfusionsthatcontainLexA88202theenergygainedfromgeometri-callyfavorablemonomer-monomercontactsbetweenLexA88202moietiesmightcompensatefortheenergylostfromremovalofamonomer-DNAcontact;thismightberevealedexperimentallybyanimprovedabilitytointeractwithanoperatorwithaweakmutation(lopm).Ourresultssupportedthissecondinterpretation.AllfusionproteinsthatcontainedtheLexA88202domain(class3)activatedthelopmandlopreporterssimilarly.Incontrast,allfusionproteinsthatlackedtheLexA88202domain(class2)wereextremelysensitivetothelopmpointmutation(10-to20-folddifferencesinactivation),whethertheycontainedotherdimerizationmotifsornot.WhilethisresultdoesnotexcludethepossibilitythatheterologousdimerizationmotifsnotexaminedinthisstudymayinfactsubstitutefortheLexA88202domain,clearlynotalldoso.Theseexperimentsalsoraisearelatedpoint.Heretofore,strongactivationbyLexA_87fusionproteinshasbeentakentoimplythatthefusedmoietycontainsdimerizationse-quences.However,theaboveexperimentsstronglysuggestthatalldimerizationsequencesarenotequivalentinpromot-ingoperatorrecognition.Moreover,asurveyofthepub-lishedliteratureonLexA-activatorchimerasindicatesthatsuchproteinsmaynotneedtocontainanydimerizationsequences,LexAspecificornot,inordertoactivate.AlargenumberofLexAl87derivativesactivate;atleastsomeofthefusedmoietiespresentareeitherthoughttocontainweakdimerizationsequencesorarenotknowntocontainthematall(22,24,27,36,56).Sinceitisunlikelythatsuchproteinscanquantitativelyoccupyoperators,theirabilitiestoacti-vatemayreflecttheextremesensitivityofthetypicaltran-scriptionactivationassays,ratherthantheirpossessionofdimerizationsequences.WenotethatalthoughtheLexAaminoterminushasbeenshowntobindtoisolatedhalf-sitesinvitro(37),ourresultsdonotgenerallysupporttheideathatmonomersofLexAderivativesbindtoandactivatefromisolatedoperatorhalf-sitesinvivo.MostoftheLexA1-7derivativesfailedtoactivatethel/2opreporters,inwhichallconsensusbasesinVOL.12,19923012GOLEMISANDBRENTonehalfoftheLexAoperatorwerechanged.Oneprotein,however,LexAl187-GALA,reproduciblygaveveryweakactivationofthisreporter.WesuspectthatthisactivationiscausedbyLexA1l7-GAL4dimersbutthatthesedimersdonotoccupytheoperatorefficiently.Itisalsopossible,however,thattheactivationweseereflectsbindingbyanisolatedmonomerwhichactivatesinefficientlyeitherbe-causeitdoesnotbindahalf-sitewellorbecauseitdoesbindahalf-sitewell,butmultipleactivationdomainsareneces-saryforsubstantialactivation(i.e.,synergy[12,44]).Insum,ourresultsaremostsimplyinterpretedbypostu-latingthatthecarboxy-terminaldomainofamonomerofnativeLexA(andofLexA1l202derivatives)promotesaspatiallypreciseassociationwithanothermonomerontheDNA,inwhichthetwoamino-terminalDNAbindingdo-mainsareoptimallyalignedinordertointeractwiththeiroperatorwithhighaffinity.Lackingthispreciseprotein-proteininteraction,LexA1l7derivativesaremuchmoresensitivetotheremovalofasinglemonomer-DNAcontactinthelopmsite.SuchamodelforLexAbindingissimilartorecentproposalsforbindingbytheglucocorticoidreceptor(48).Inaddition,theideathatLexAmonomerscanassociateontheDNAhasrecentlyreceivedindependentbiochemicalsupport(37).Fusiondomain-dependentinterferencewithactivation.WecanimagineanumberofpossiblemechanismsbywhichfusiondomainsmightreduceactivationbyLexA1l7deriv-atives.Anyorallofthesemaybereflectedintheextremelylowactivationobservedwithclass1proteins.First,thefusedsequencesmightinterferewithproperfoldingoftheLexAmoiety.Weregardthispossibilityasunlikely,butintheabsenceofstructuraldatawecannotexcludeit.Second,iftheLexAderivativecontainsasecondDNA-bindingdomain,theproteinmightbesequesteredfromLexAoperatorsbybindingtononoperatorDNA;suchsequestrationwouldbemostsevereincellsthathadcom-plexgenomes.SequestrationontheDNAmaywellexplainwhytheDNAbindingregionofthev-relproductinhibitsactivationbyLexA-vRelderivativesinmammaliancellsbutnotinyeastcells(33,57),mayaccountforlackofactivationorinhibitionofactivationreportedfortheDNAbindingdomainsofthePit-1(30)andMyc(22,35)proteins,andmaycomplicatetheinterpretationofstudiesofLexA-HMGde-rivatives,whichalsob

indnonspecificDNAandwhicharereportednott
indnonspecificDNAandwhicharereportednottoactivateatall(40).Third,anoligomerizationmotifinthefusionmoietymightcausetheLexAderivativetoformacomplexwithothercellularproteinsthateitherkeepsitfrombindingoperatororallowsittobindoperatorbutoccludestheactivationdo-main.Suchcomplexformationmightexplainwhy,forLexA-MycandLexA-Fosfusionproteins,removaloftheoligomerizationmotifs(thehelix-loop-helix/leucinezipperandleucinezipper,respectively)resultsina5-to10-foldincreaseintheirabilitiestoactivate(Table3)(22,35,42).Fourth,thefusionmoietymightstericallyinhibitoroth-erwiseincreasetheamountofenergyrequiredtopositiontheLexADNAbindingdomainsproperlyonadjacenthalf-sites.Ourinvitrodataindicatethatthosefusionproteinsleastimpairedforoperatorbindingwerethosethatcontainedrelativelysmallfusiondomains-theacidblobsB42,B6(Fig.3),andB7(notshown)andthesmallmoietyencodedbythepolylinkerofpLexA1_202-PL.Thislastresultraisesthepossibilitythattheirrelativelytighterbindingcontributestothepotencyofacidblobsintranscriptionactivationassays(50).Insupportofthis,wehavefoundthataLexA-VP16fusionproteinisasignificantlystrongeractiva-torthanaLexA-acidblobfusion(LexA-B112)whentheproteinsandaLexA-operator-CATreporterareexpressedathighlevelsinmammaliancells.However,whenthetwoLexAfusionproteinsarecomparablyexpressedatlowlevels,LexA-B112becomesastrongeractivatorthanLexA-VP16(21).Consequencesforfutureapplicationsoffusionproteins.OurresultssuggestseveralconsiderationsforuseofLexAfu-sions.First,ourdataclearlyshowthatoperatoroccupancyinvivodiffersgreatly,bothbetweendifferentchimericproteinsandbetweenidenticalchimericproteinssynthe-sizedfromdifferentexpressionplasmids.Inconsequence,ourresultssuggestthatthedegreeofoperatoroccupancyshouldbeexplicitlyconsideredwhendifferentLexAderiv-ativesarecompared.Inpractice,thebestwaytoensurecomparableoccupancywillusuallybetoensurethatthedifferentderivativescanfullyoccupyoperators.Fulloccu-pancyisclearlyfavoredbyuseofLexA1l202derivatives.Second,sincesaturationof,B-galactosidaseproductionoc-curswithverysmallnumbersofDNA-boundstrongactiva-tors(compareLexAl202-B42withLexAj_87-GAL4andLexA1-87-cFos[Table3]),activatorsshouldbecomparedonareporterthatcarriesasmallnumberofoperators,whichshouldideallybepositionedfarenough�(200bp)upstreamofthelacZgenetoensurethattranscriptioniswellbelowthemaximumlevel.Third,oneofthemoreimportantapplicationsofchimericproteinsinyeastsisastranscriptionallyinertbaitstodetectinteractingproteinsfromactivationdomain-taggedcDNAexpressionlibraries(13,17,23,70,71,80).Ourdatashowthat,atleastwhenLexAderivativesareusedforbaits,twoneedsmustbebalanced.Ononehand,manybaitsactivateweakly,anditisimportanttosaturateoperatorbindingsothatadventitiousincreasesinbaitexpressioninindividualcells(21)cannotcausespuriousactivationofthereportergeneduringtheexpressionlibraryscreen.Ontheotherhand,transcriptionofthereporterinaparticularcellisdependentonthetotalamountofoperator-boundbaitthatinteractswiththeactivationdomain-taggedproteinencodedbyamemberoftheexpressionlibrary;forthistobemaximized,thetotalconcentrationofbaitshouldnotexceedthetotalconcentrationoflibrary-encodedprotein.Ourre-sultssuggestthattheseneedsmaybebestmetbythechoiceofLexAl-202ratherthanLexAl-7asaDNAbindingdomainandtheuseofshort,nonactivatingheterologousmoietiesinthebait.Inaddition,sincenativeLexAlacksanuclearlocalizationsequence,gainsinthedegreeofoperatorbind-ingatlowbaitconcentrationsmightcomefromtheadditionofanuclearlocalizationmotiftobaits.Finally,althoughthisstudyhasconfineditselftoLexAderivatives,itisequallylikelythatsomeheterologousmoi-etiescanaffectDNArecognitionbyotherbindingdomainssuchasGAL4(11).GiventhatthestructuresofGAL4andLexAarequitedifferent(2,38,39,51),fuseddomainsthatareextremelydeleterioustoDNAbindinginthecontextofLexAmaybelesssointhecontextofGAL4,andviceversa.SuchDNAbindingdomain-specificeffectsmaycontributetoreporteddifferencesinactivationstrengthbetweenother-wisesimilarLexAandGAL4derivatives(60).ACKNOWLEDGMENTSWethankSteveHanesforthelI2opandBicoidplasmids,TonyZervosforthelopmplasmid,RussFinleyfortheLexAl-202-PRDandLexAl-202-PRD/HDplasmids,andTodGulickfortheNSoligonucleotides.WealsothankJohnLittleforcommunicatingresultspriortopublicationandforusefuldiscussions.WeareMOL.CELL.BIOL.DNABINDINGBYLexAFUSIONPROTEINS3013extremelygratefultoDavidGreenstein,TodGulick,SteveHanes,AnnHochschild,JohnLittle,andRobinWhartonforcommentsonthemanuscript.ThisworkwassupportedbyHoechstAG.R.B.wassupportedbythePEWscholarsprogram.E.A.G.wassupportedbypostdoctoralfellowshipCA08817fromtheNIH.REFERENCES1.Ausubel,F.M.,R.Brent,R.Kingston,D.Moore,J.Seidman,J.Smith,andK.Struhl.1987.Currentprotocolsinmolecularbiology.JohnWiley&Sons,Inc.,NewYork.2.Baleja,J.D.,R.Marmorstein,S.C.Harrison,andG.Wagner.1992.SolutionstructureoftheDNA-bindingdomainofCd2-GAL4fromS.cerevisiae.Nature(London)356:450-453.3.Berleth,T.,M.Burri,G.Thoma,D.Bopp,S.Richstein,G.Frigerio,M.Noll,andC.Nusslein-Volhard.1988.TheroleoflocalizationofbicoidRNAinorganizingtheanteriorpatternoftheDroso

philaembryo.EMBOJ.7:1749-1756.4.Brent,R.
philaembryo.EMBOJ.7:1749-1756.4.Brent,R.1982.RegulationandautoregulationbylexAprotein.Biochimie64:565-569.5.Brent,R.1982.RegulationofthecellularresponsetoDNAdamage.Ph.D.thesis.HarvardUniversity,Cambridge,Mass.6.Brent,R.1985.Repressionoftranscriptioninyeast.Cell42:3-4.7.Brent,R.,andM.Ptashne.1980.TheLexAgenerepressesitsownpromoter.Proc.Natl.Acad.Sci.USA77:1932-1936.8.Brent,R.,andM.Ptashne.1981.MechanismofactionofthelexAgeneproduct.Proc.Natl.Acad.Sci.USA78:4204-4208.9.Brent,R.,andM.Ptashne.1984.Abacterialrepressorproteinorayeasttranscriptionalterminatorcanblockupstreamactiva-tionofayeastgene.Nature(London)312:612-615.10.Brent,R.,andM.Ptashne.1985.AeukaryotictranscriptionalactivatorbearingtheDNAspecificityofaprokaryoticrepres-sor.Cell43:729-736.11.Carey,M.,H.Kakidani,J.Leatherwood,F.Mostashari,andM.Ptashne.1989.Anamino-terminalfragmentofGAL4bindsDNAasadimer.J.Mol.Biol.209:423-432.12.Carey,M.,Y.-S.Lin,M.R.Green,andM.Ptashne.1990.AmechanismforsynergisticactivationofamammaliangenebyGALAderivatives.Nature(London)345:361-364.13.Chien,C.-T.,P.L.Bartel,R.Sternglanz,andS.Fields.1991.Thetwo-hybridsystem:amethodtoidentifyandclonegenesforproteinsthatinteractwithaproteinofinterest.Proc.Natl.Acad.Sci.USA88:9578-9582.14.Driever,W.,andC.Nusslein-Volhard.1988.AgradientofbicoidproteininDrosophilaembryos.Cell54:83-93.15.Ebina,Y.,Y.Takahara,F.Kishi,A.Nakazawa,andR.Brent.1983.LexAproteinisarepressorofthecolicinElgene.J.Biol.Chem.258:13258-13261.16.Fields,S.,and0.Song.1988.Anovelgeneticsystemtodetectprotein-proteininteraction.Nature(London)340:245-246.17.Finley,R.,andR.Brent.Unpublisheddata.18.Fried,M.,andD.M.Crothers.1981.Equilibriaandkineticsoflacrepressor-operatorinteractionsbypolyacrylamidegelelec-trophoresis.NucleicAcidsRes.9:6505-6525.19.Garner,M.M.,andA.Revzin.1981.AgelelectrophoreticmethodforquantifyingthebindingofproteinstospecificDNAregions:applicationtocomponentsoftheE.colilactoseoperonregulatorysystem.NucleicAcidsRes.9:3047-3060.20.Godowski,P.J.,D.Picard,andK.R.Yamamoto.1988.Signaltransductionandtranscriptionalactivationbyglucocorticoidreceptor-LexAfusionproteins.Science241:812-816.21.Golemis,E.A.,andR.Brent.Unpublisheddata.22.Golemis,E.A.,R.Lee,andR.Brent.Unpublisheddata.23.Gyuris,J.,E.A.Golemis,andR.Brent.Unpublisheddata.24.Hanes,S.,andR.Brent.1989.DNAspecificityofthebicoidactivatorproteinisdeterminedbyhomeodomainrecognitionhelixresidue9.Cell57:1275-1283.25.Hanes,S.,andR.Brent.Unpublisheddata.26.Harrison,S.C.,andA.K.Aggarwal.1990.DNArecognitionbyproteinswiththehelix-turn-helixmotif.Annu.Rev.Biochem.59:933-969.27.Heaney,M.L.,J.Pierce,andJ.T.Parsons.1986.Site-directedmutagenesisofthegag-mycgeneofavianmyelocytomatosisvirus29:biologicalactivityandintracellularlocalizationofstructurallyalteredproteins.J.Virol.60:167-176.28.Hope,I.A.,S.Mahadevan,andK.Struhl.1988.StructuralandfunctionalcharacterizationoftheshortacidictranscriptionalactivationregionofyeastGCN4protein.Nature(London)333:635-640.29.Hope,I.A.,andK.Struhl.1987.GCN4,aeukaryotictranscrip-tionalactivatorprotein,bindsasadimertotargetDNA.EMBOJ.6:2781-2784.30.Ingraham,H.A.,S.E.Flynn,J.W.Voss,V.R.Albert,M.S.Kapiloff,L.Wilson,andM.G.Rosenfeld.1990.ThePOU-specificdomainofPit-1isessentialforsequence-specific,highaffinityDNAbindingandDNA-dependentPit-i-Pit-1interac-tions.Cell61:1021-1033.31.Ito,H.,Y.Fukada,K.Murata,andA.Kimura.1983.Transfor-mationofintactyeastcellstreatedwithalkalications.J.Bacteriol.153:163-168.32.Johnson,A.D.,C.0.Pabo,andR.T.Sauer.1980.Bacterio-phagelambdarepressorandcroprotein:interactionswithoperatorDNA.MethodsEnzymol.65:839-856.33.Kamens,J.,andR.Brent.1991.AyeasttranscriptionassaydefinesdistinctrelanddorsalDNArecognitionsequences.NewBiol.3:1005-1013.34.Kamens,J.,P.Richardson,G.Mosialos,R.Brent,andT.Gilmore.1990.OncogenictransformationbyvRelrequiresanamino-terminalactivationdomain.Mol.Cell.Biol.10:2840-2847.35.Kato,G.J.,M.Barrett,M.Villa-Garcia,andC.V.Dang.1990.Anamino-terminalc-Mycdomainrequiredforneoplastictrans-formationactivatestranscription.Mol.Cell.Biol.10:5914-5920.36.Keegan,L.,G.Gill,andM.Ptashne.1986.SeparationofDNAbindingfromthetranscription-activatingfunctionofaeukary-oticregulatoryprotein.Science231:699-704.37.Kim,B.,andJ.W.Little.1992.DimerizationofaspecificDNAbindingproteinontheDNA.Science255:203-206.38.Kraulis,P.J.,A.R.C.Raine,P.L.Gadhavi,andE.D.Laue.1992.StructureoftheDNA-bindingdomainofzincGAL4.Nature(London)356:448-450.39.Lamerichs,R.M.,A.Padilla,R.Boelens,R.Kaptein,G.Ottleben,H.Ruterhans,M.Granger-Schnarr,P.Oertel,andM.Schnarr.1989.Theamino-terminaldomainofLexArepressorisalpha-helicalbutdiffersfromcanonicalhelix-turn-helixpro-teins:atwodimensional'HNMRstudy.Proc.Natl.Acad.Sci.USA86:6863-6867.40.Landsman,D.,andM.Bustin.1991.Assessmentofthetran-scriptionalactivationpotentialoftheHMGchromosomalpro-teins.Mol.Cell.Biol.11:4483-4489.41.Laughon,A.,andR.F.Gesteland.1984.PrimarystructureoftheSaccharomycescerevisiaeGALAgene.Proc.Natl.Acad.Sci.USA79:6827-6831.42.Lech,K.1990.GeneactivationbyDNA-boundFosandMycproteins.Ph.

D.thesis.HarvardUniversity,Cambridge,Mas
D.thesis.HarvardUniversity,Cambridge,Mass.43.Lech,K.,K.Anderson,andR.Brent.1988.DNA-boundFosproteinsactivatetranscriptioninyeast.Cell52:179-184.44.Lin,Y.-S.,M.Carey,M.Ptashne,andM.Green.1990.Howdifferenteukaryotictranscriptionalactivatorscancooperatepromiscuously.Nature(London)345:359-361.45.Little,J.W.,andS.A.Hill.1985.DeletionswithinahingeregionofaspecificDNA-bindingprotein.Proc.Natl.Acad.Sci.USA82:2301-2305.46.Little,J.W.,andD.W.Mount.1982.TheSOSregulatorysystemofEscherichiacoli.Cell29:11-22.47.Little,J.W.,D.W.Mount,andC.R.Yanisch-Perron.1981.PurifiedLexAproteinisarepressoroftherecAandlexAgenes.Proc.Natl.Acad.Sci.USA78:4199-4203.48.Luisi,B.F.,W.X.Xu,Z.Otwinowski,L.P.Freedman,K.R.Yamamoto,andP.B.Sigler.1991.CrystallographicanalysisoftheinteractionoftheglucocorticoidreceptorwithDNA.Nature(London)352:497-505.49.Ma,J.,andM.Ptashne.1987.Thecarboxy-terminal30aminoacidsofGALAarerecognizedbyGAL80.Cell50:137-142.50.Ma,J.,andM.Ptashne.1987.Anewclassofyeasttranscrip-tionalactivators.Cell51:113-119.VOL.12,19923014GOLEMISANDBRENT51.Marmorstein,R.,M.Carey,M.Ptashne,andS.C.Harrison.1992.DNArecognitionbyGAL4:structureofaprotein-DNAcomplex.Nature(London)356:408-414.52.Nelson,C.,E.B.CrenshawIII,R.Franco,S.A.Lira,V.R.Albert,R.M.Evans,andM.G.Rosenfeld.1986.Discretecis-activegenomicsequencesdictatethepituitarycelltype-specificexpressionofratprolactinandgrowthhormonegenes.Nature(London)322:557-562.53.Nelson,M.,andP.Silver.1989.ContextaffectsnuclearproteinlocalizationinSaccharomycescerevisiae.Mol.Cell.Biol.9:384-389.54.Oleson,J.T.,andL.Guarente.1990.TheHAP2subunitofyeastCCAATtranscriptionalactivatorcontainsadjacentdomainsforsubunitassociationandDNArecognition:modelfortheHAP2/3/4complex.GenesDev.4:1714-1729.55.O'Shea,E.K.,W.F.SaffordIII,andP.S.Kim.1989.Prefer-entialheterodimerformationbyisolatedleucinezippersfromfosandjun.Science245:646-648.56.Ptashne,M.1978.Theoperon,p.325-344.ColdSpringHarborLaboratory,ColdSpringHarbor,N.Y.57.Reddy,E.P.,R.K.Reynolds,D.K.Watson,R.A.Schultz,J.Lautenberger,andT.S.Papas.1983.Nucleotidesequenceanalysisoftheproviralgenomeofavianmyelocytomatosisvirus(MC29).Proc.Natl.Acad.Sci.USA80:2500-2504.58.Richardson,P.M.,andT.D.Gilmore.1991.v-RelisaninactivememberoftheRelfamilyoftranscriptionalactivatingproteins.J.Virol.65:3122-3130.59.Roux,P.,J.-M.Blanchard,A.Fernandez,N.Lamb,P.Jeanteur,andM.Piechaczyk.1990.Nuclearlocalizationofc-Fos,butnotv-Fosproteins,iscontrolledbyextracellularsignals.Cell63:341-351.60.Ruden,D.M.,J.Ma,Y.Li,K.Wood,andM.Ptashne.1991.Generatingyeasttranscriptionalactivatorscontainingnoyeastproteinsequences.Nature(London)350:250-252.61.Sadowski,I.,J.Ma,S.Triezenberg,andM.Ptashne.1988.GAL4-VP16isanunusuallypotenttranscriptionalactivator.Nature(London)335:683-689.62.Salmeron,J.M.,Jr.,K.K.Leuther,andS.A.Johnston.1990.GAL4mutationsthatseparatethetranscriptionalactivationandGAL80-interactivefunctionsoftheyeastGAL4protein.Genet-ics125:21-27.63.Sambrook,J.,E.F.Fritsch,andT.Maniatis.1989.Molecularcloning:alaboratorymanual,2nded.ColdSpringHarborLaboratory,ColdSpringHarbor,N.Y.64.Samson,M.-L.,andR.Brent.Unpublisheddata.65.Samson,M.-L.,L.Jackson-Grusby,andR.Brent.1989.GeneactivationandDNAbindingbyDrosophilaUbxandabd-Aproteins.Cell57:1045-1052.66.Sauer,R.T.,H.C.Nelson,K.Hehir,M.H.Hecht,F.S.Gimble,J.DeAnda,andA.R.Poteete.1983.ThelambdaandP22phagerepressors.J.Biomol.Struct.Dyn.1:1011-1022.67.Schleif,R.1988.DNAbindingbyproteins.Science241:1182-1187.68.Schnarr,M.,M.Granger-Schnarr,S.Hurstel,andJ.Pouyet.1988.Thecarboxy-terminaldomainoftheLexArepressoroligomerisesessentiallyastheentireprotein.FEBSLett.234:56-60.69.Schnarr,M.,J.Pouyet,M.Granger-Schnarr,andM.Daune.1985.Large-scalepurification,oligomerizationequilibria,andspecificinteractionoftheLexArepressorofEscherichiacoli.Biochemistry24:2812-2818.70.Shore,D.Personalcommunication.71.Silver,P.A.Personalcommunication.72.Steitz,T.A.1991.Structuralstudiesofprotein-nucleicacidinteractions:thesourcesofsequence-specificbinding.Q.Rev.Biophys.23:205-280.73.Stone,J.,T.deLange,G.Ramsay,E.Jakobovits,J.M.Bishop,H.Varmus,andW.Lee.1987.Definitionsofregionsinhumanc-mycthatareinvolvedintransformationandnuclearlocaliza-tion.Mol.Cell.Biol.7:1697-1709.74.Thliveris,A.T.,J.W.Little,andD.W.Mount.1991.Repres-sionoftheE.colirecAgenerequiresatleasttwoLexAproteinmonomers.Biochimie73:449-455.75.Thliveris,A.T.,andD.W.Mount.GeneticidentificationoftheDNAbindingdomainofE.coliLexAprotein.Proc.Natl.Acad.Sci.USA,inpress.76.Wagner,P.,J.Kunz,A.Koller,andM.N.Hall.1990.Activetransportofproteinsintothenucleus.FEBSLett.275:1-5.77.Walker,G.C.1985.InducibleDNArepairsystems.Annu.Rev.Biochem.54:425-457.78.Watson,J.D.,N.H.Hopkins,J.W.Roberts,J.A.Steitz,andA.W.Weiner.1987.Molecularbiologyofthegene,4thed.Benjamin/CumminsPublishingCompany,MenloPark,Calif.79.West,R.W.,Jr.,R.R.Yocum,andM.Ptashne.1984.Saccha-romycescerevisiaeGALl-GAL10divergentpromoterregion:locationandfunctionoftheupstreamactivatorsequenceUASG.Mol.Cell.Biol.4:2467-2478.80.Zervos,A.,andR.Brent.Unpublisheddata.MOL.CELL.B