Fax16313443407Emailaddressstudierbnlgov AbbreviationsusedIPTGisopropylthiogalactosidePDBProteinDataBankSSAThumanspermidinespermineacetyltransferaseSeMetselenomethionineTRBterri ID: 182827
Download Pdf The PPT/PDF document "Proteinproductionbyauto-inductioninhigh-..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Proteinproductionbyauto-inductioninhigh-densityshakingculturesF.WilliamStudierBiologyDepartment,BrookhavenNationalLaboratory,Upton,NY11973,USAReceived7January2005Availableonline12March2005InducibleexpressionsystemsinwhichT7RNApolymerasetranscribescodingsequencesclonedundercontrolofaT7moterecientlyproduceawidevarietyofproteinsinEscherichiacoli.Investigationoffactorsthataectstability,growth,andinductionofT7expressionstrainsinshakingvesselsledtotherecognitionthatsporadic,unintendedinductionofexpressionincom-plexmedia,previouslyreportedbyothers,isalmostcertainlycausedbysmallamountsoflactose.Glucosepreventsinductionbylactosebywell-studiedmechanisms.Aminoacidsalsoinhibitinductionbylactoseduringlog-phasegrowth,andhighratesofaer-ationinhibitinductionatlowlactoseconcentrations.Theseobservations,andmetabolicbalancingofpH,alloweddevelopmentofreliablenon-inducingandauto-inducingmediainwhichbatchculturesgrowtohighdensities.Expressionstrainsgrowntosatura-tioninnon-inducingmediaretainplasmidandremainfullyviableforweeksintherefrigerator,makingiteasytopreparemanyfreezerstocksinparallelanduseworkingstocksforanextendedperiod.Auto-inductionallowsecientscreeningofmanyclonesinparallelforexpressionandsolubility,ascultureshaveonlytobeinoculatedandgrowntosaturation,andyieldsoftargetproteinaretypicallyseveral-foldhigherthanobtainedbyconventionalIPTGinduction.Auto-inducingmediahavebeendevelopedforlabelingproteinswithselenomethionine,NorC,andforproductionoftargetproteinsbyarabinoseinductionofT7RNApoly-merasefromthepBADpromoterinBL21-AI.SelenomethioninelabelingwasequallyecientinthecommonlyusedmethionineauxotrophB834(DE3)(foundtobe)ortheprototrophBL21(DE3).PublishedbyElsevierInc.Auto-induction;T7expressionsystem;Lactose;pBADpromoter;Arabinose;Proteinproduction;High-densitybatchcultures;MetaboliccontrolofpH;Selenomethioninelabeling;IsotopiclabelingBackgroundandintroductionDNAsequencingprojectshaveprovidedcodingse-quencesforhundredsofthousandsofproteinsfromorganismsacrosstheevolutionaryspectrum.Recombi-nantDNAtechnologymakesitpossibletoclonethesecodingsequencesintoexpressionvectorsthatcandirecttheproductionofthecorrespondingproteinsinsuitablehostcells.AninducibleT7expressionsystemishighlyeectiveandwidelyusedtoproduceRNAsandproteinsfromclonedcodingsequencesinthebacteriumEscherichiacolicoli.ThecodingsequenceforT7RNApolymeraseispresentinthechromosomeundercontroloftheinduciblepromoterinhostssuchasBL21(DE3).Thecodingsequenceforthedesiredpro- Fax:+16313443407.E-mailaddress:studier@bnl.gov Abbreviationsused:IPTG,isopropyl--thiogalactoside;PDB,ProteinDataBank;SSAT,humanspermidine/spermineacetyltrans-ferase;SeMet,selenomethionine;TRB,terricbroth;PTS,phospho-enolpyruvate:carbohydratephosphotransferasesystem.www.elsevier.com/locate/yprep ProteinExpressionandPurication41(2005)207 234 transcribealmostanyDNAcontrolledbytheT7pro-moter.T7RNApolymeraseissospecic,active,andprocessivethattheamountoftargetRNAproducedcanbecomparabletotheamountofribosomalRNAinacell.IfthetargetRNAcontainsacodingsequencewithappropriatetranslationinitiationsignals(suchasthesequenceupstreamofthestartcodonfortheT7ma-jorcapsidprotein),mostproteinsynthesiswillbedirec-tedtowardtargetprotein,whichusuallyaccumulatestobecomeasubstantialfractionoftotalcellprotein.AprobleminusinginducibleT7expressionsystemsisthatT7RNApolymeraseissoactivethatasmallbasallevelcanleadtosubstantialexpressionoftargetproteinevenintheabsenceofaddedinducer.Ifthetargetpro-teinissucientlytoxictothehostcell,establishmentofthetargetplasmidintheexpressionhostmaybedicultorimpossible,ortheexpressionstrainmaybeunstableoraccumulatemutationsions.Aneectivemeanstore-ducebasalexpressionistoplacetheoperatorse-quence(thebindingsiteforrepressor)justdownstreamofthestartsiteofaT7promoter,creatingaT7promoterer.Lacrepressorboundattheoper-atorsequenceinterfereswithestablishmentofanelonga-tioncomplexbyT7RNApolymeraseataT7promoterandsubstantiallyreducestheleveloftargetmRNAproduceduced.Ifsucientrepressorispresenttosaturateallofitsbindingsitesinthecell,thebasalleveloftargetproteininuninducedcellsissub-stantiallyreduced,butinductionunblocksboththelacUV5andT7promotersandleadstothetypicalhighlevelsofexpression.Thus,theT7promoterin-creasestheconvenienceandapplicabilityoftheT7sys-temforexpressingawiderangeofproteins.Structuralgenomicsisanareawheremulti-milligramamountsofmanywidelydierentproteinsaresoughtfordeterminationofproteinstructuresbyX-raycrystal-lographyornuclearmagneticresonance(NMR)(NMR).Notalltargetproteinswillbewellexpressedandsoluble,soitisdesirabletoscreeninparallelmanysmallculturesexpressingdierenttargetproteinstoidentifythoseuse-fulforscalingup.Asignicantdicultyinlarge-scalescreeningistoobtainalloftheculturesinacomparablestateofgrowth,sothattheycanbeinducedsimulta-neously.Dierencesinlagtimeorgrowthratetypicallygenerateasituationwheredierentcultureswillbereadyforinductionatdierenttimes.Evenifculturesweregrowninamulti-wellplateanddensitiescouldbereadsimultaneouslyinaplatereader,considerableef-fortwouldberequiredtofollowgrowthandaddinducertoeachcultureatthepropertime.Ifallofthecultureswerecollectedatonce,choosingacollectiontimewhenallhadbeeninducedtooptimallevelsandnonehadsuf-feredovergrowthbycellsincapableofexpressingtargetproteinmightbedicultorimpossible.Onestrategyforobtainingfairlyuniforminductionistoincubateaplateuntilallofthecultureshavegrowntosaturation,addfreshmedium,growforanappropriatetime,andaddinducertoallwellsatthesametime.Ifallculturesinaplatesaturateatcomparabledensityandgrowafterdilutionwithsimilarenoughkinetics,theculture-to-culturevariationindensityatthetimeofinductionmightbelowenoughthatmostcultureswillbeoptimallyinduced.However,inatestofthisstrategy,IencounteredtheunintendedinductiondescribedbyGrossmanetal.al.,whofoundthatculturesgrowingincertaincomplexmediainducesubstantialamountsoftargetproteinuponapproachtosaturation,intheab-senceofaddedinducer.Inductionatsaturationwouldstresscellstodierentextents,dependingonthelevelsofinductionandrelativetoxicityoftargetproteinstothehostcells,makingastrategyofsaturationfollowedbydilutionunworkableinmediathathavesuchinduc-ingactivity.Grossmanetal.al.concludedthattheknowninducerlactosewasnotresponsibleforunin-tendedinductionbutthatcyclicAMPisrequired,andtheyfoundthatusingahostmutantunabletomakecyc-licAMPimprovedplasmidstabilityandproteinproduc-tion.Consistentwitharoleforcataboliterepression,theyalsofoundthatadditionof1%glucosetothecom-plexmediumpreventedunintendedinduction.However,Iobservedthatadditionof1%glucosealsocausedsat-uratedculturestobecomeveryacidic,whichlimitssat-urationdensityandagainmakesitdiculttogetuniformgrowthupondilution.Uponfurtherinvestigation,IfoundthatmediamadewithN-Z-amineASfroma100-poundbarrelrecentlyacquiredforstructuralgenomicsworkshowedinductionatsaturationwhereasotherwiseidenticalmediamadefromtheprevious(almostexhausted)barrelfromthesamesupplierdidnot.ScreeningdierentlotsofN-Z-amineorotherenzymaticdigestsofcaseinforthosewithouttheinducingbehaviordidnotseemtobeanattractivesolution:besidestheobviousineciency,suchlotsmightnotalwaysbeavailable.Toaddresstheprob-lemofsporadic,unwantedinduction,Iundertookasys-tematicanalysisofthecomponentsofbothcomplexanddenedmediaandtheireectsongrowthandinduction.ThegoalwastodevelopformulationsforreliablegrowthofculturesofT7expressionstrainstosaturationwithlittleornoinductionandtodeneconditionssuit-ableforgrowthandinductionofmanyculturesinparallel.MaterialsandmethodsBacterialstrainsandplasmidsEscherichiacolistrainsusedfortestinggrowthandexpressionwereprimarilyBL21(DE3)andB834(DE3).B834isarestriction-modicationdefective,galactose-negative,methionineauxotrophofE.colicoli.BL21F.W.Studier/ProteinExpressionandPurication41(2005)207 234 isaMetderivativeofB834obtainedbyP1transduc--.DE3lysogenscontainaderivativeofphagelambdathatsuppliesT7RNApolymerasebytranscrip-tionfromthelacUV5promoterinthechromosomesome.BL21-AI(Invitrogen)isaderivativeofBL21thatsuppliesT7RNApolymerasebytranscriptionfromthearabinose-induciblepBADpromoterinthechromosome.Codingsequencesfortargetproteinswereclonedun-dercontroloftheT7promoterandtheupstreamtranslationinitiationsignalsoftheT7majorcapsidpro-pro-byplacingtheinitiationcodonattheposi-tionoftheIsiteofpET-13a13aorpET-24b(Novagen),ortheIsiteofpREXvectors(equivalenttotheIsiteofpET-11d11d;tobedescribedelse-where),allofwhichconferresistancetokanamycin.PlasmidscontainingtheT7promoteralsocontainacopyofthegenetoprovideenoughrepressortosaturateallofitsbindingsites.Avarietyofdierenttargetproteinswereusedindevelopingandtestingnon-inducingandauto-inducingmedia,includingasetofabout100yeastproteinsclonedforastructuralgenomicsproject(http://proteome.bnl.gov/targets.html).Forconvenience,specicyeastproteinsmentionedinthetextarereferredtobytheirtargetnumbers:P07referstoyeastproteinYBL036C,ProteinDataBank(PDB)1B54,structurallysimilartotheN-terminaldomainofanaminoacidracemasease;P19referstoyeastproteinYBR022W,ofunknownfunction;P21referstotheproteinspeciedbyyeast,atranslationreleasefactor;P35referstotheproteinspeciedbyyeastgenehem13,PDB1TXN,coproporphyrinogenIIIoxidase;andP89referstoyeastproteinYMR087W,PDB1NJR,proposedfromitsstructuretobeanADP-ribose-1-monophosphatasesphatase.Thecodingsequenceforhumanspermidine/sperm-ineacetyltransferase(SSAT)wasampliedbyreversetranscriptaseandPCRfromtotalRNAfromahumancellline(thekindgiftofPaulFreimuth)andclonedinpET-13a.BacteriophageT7proteinsspeciedbygenes(thewell-expressedmajorcapsidprotein),,(highlytoxicproteinsofunknownfunction)[3,4]wereexpressedfrompREXvectors.TheexpressionhostforclonedyeastproteinswasB834(DE3),inthemistakenbeliefthatamethionine-re-quiringhostwouldbebetterforlabelingproteinswithselenomethionine(SeMet)forcrystallography(seesec-tiononAuto-inductionforlabelingproteinswithSeMetforcrystallography).TheRILplasmidfromBL21-Gold(DE3)RIL(Stratagene)increasestheexpressionofsomeyeasttargetproteinsbysupplyingtRNAsforcodonsusedfrequentlyinyeastbutnotE.coli.T7proteinsandsomeotherproteinswereexpressedinBL21(DE3)orBL21-Gold(DE3)RIL(intowhichStratageneintroducedtheHtephenotypeforhightransformationeciencyandanmutationtoreduceendonucleaseactivity).TheRILplasmidisderivedfromapACYCplasmidandconfersresistancetochloramphenicol.Freezerstocksforlong-termstorageofexpressionstrainsaremadebyadding0.1mlof100%(w/v)glycerolto1mlofcultureinlogphaseorgrowntosaturationinnon-inducingmediasuchasPG,LSGorMDG(Table),mixingwell,andplacinginaCfreezer.Subcul-turesforuseasworkingstocksaremadebyscrapingupasmallamountoffrozenculturewithasterileplasticpipettortipwithoutmeltingtherestofthestockandinoculatingintonon-inducingmedia.Aftergrowthtosaturation,suchworkingstocksaretypicallystableforweeksintherefrigerator.GrowthmediaN-Z-amineAS,asolubleenzymaticdigestofcasein(in100-poundbarrels),andyeastextract(HY-YEST444ina55-poundbarrel)wereobtainedfromQuestInternational,5515SedgeBlvd.,HomanEstates,IL60192,telephone800-833-8308.Forconvenience,thedesignationN-Z-aminewillrefertoN-Z-amineAS,whichcouldbesubstitutedforbyotherenzymaticdi-gestsofcasein,suchastryptone,inthemediadescribedhere.Smallerquantitiesofenzymaticdigestsofcaseinoryeastextractaswellassugars,salts,aminoacids,vita-mins,andothercomponentsofgrowthmediawereob-tainedfromDifco,Sigma,Fisherorotherbiochemicalandchemicalsuppliers.MediapreviouslydescribedbedforgrowthofE.coliandproductionoftargetproteinswiththeT7expressionsystemincludeZB(10gN-Z-amineand5gNaCl/L),ZYB(previouslyZY)(10gN-Z-amine,5gyeastextract,and5gNaCl/L),M9(1gCl,3gKH,6gNa,4gglucose,and1mlof1MMgSO/L)andM9ZB,thecombinationofM9andZB.Forconvenience,concentrationsofcer-tainmediacomponentsaregiveninpercent(w/v).ThepreviouslynamedZYmediumwillherebecalledZYBmediumtoindicatethepresenceof0.5%NaCl,analo-goustoZBmedium.ThenameZYwillbereservedfor1%N-Z-amine,0.5%yeastextractwithnosaltadded.Thecompositionsofsomeofthenewlydevelopedmediaforgrowingculturestohighdensitywithoutinductionandforauto-inductionaregiveninTable1Mediaareconvenientlyassembledfromsterileconcen-tratedstocksolutionsaddedtosterilewaterorZYjustbeforeuse.StandardstocksolutionsofmixturesincludeP(1MNa,1MKH,and0.5M);50L(0.625MNa,0.625M,2.5MNHCl,and0.25MNa);50(1.25MNaHPO,1.25MKH,2.5MNHand0.25MNa);505052(25%glycerol,2.5%glu-cose,and10%-lactosemonohydrate);and100(50%glycerol,5%glucose).ThetermlactosewillreferF.W.Studier/ProteinExpressionandPurication41(2005)207 234 Table1Compositionsofnewlydevelopednon-inducingandauto-inducingmediaMediumPreviousCommentN-Z-amine18aminoPGP-0.5GNon-inducing50502520.2PAGNon-inducing50502520.20.50.36P-5052Auto-inducing50502520.20.50.050.2PA-5052Auto-inducing50502520.20.50.050.20.36PASM-5052Auto-inducingSeMet50502520.20.50.050.2SeMetZYP-5052Auto-inducing10.550502520.20.50.050.2LSGNIMSNon-inducing12.512.550520.20.520LS-5052Auto-inducing12.512.550520.20.50.050.220MDGNon-inducing252550520.20.50.25MDAGNon-inducing252550520.20.50.250.36MDA-5052Auto-inducing252550520.20.50.050.20.250.36ZYM-505Noinducer10.5252550520.20.50.05ZYM-5052Auto-inducing10.5252550520.20.50.050.2N-5052Auto-inducing505050520.20.50.050.2C-750501Auto-inducing505050520.20.750.050.01AnexplanationofthenamingconventionsisgivenunderGrowthmedia.WecurrentlyuseMDGforgrowingfreezerstocksandworkingcultures,ZYM-5052forroutineauto-induction,andZYM-505forplasmidpreps.1mMMgSO4isadequateandwastheconcentrationgiveninpreviouslydistributedrecipes.to5metalsmaybeused;additionofmetalsisoptionalbutrecommendedinmediacontainingZY.Twohundredmicrogramspermilliliterofeachof18aminoacids(nocysteineortyrosine).PASM-5052contains200g/mlofeachof17aminoacids(noC,Y,andM),10g/mlmethionine,125g/mlSeMet,and100nMvitaminBF.W.Studier/ProteinExpressionandPurication41(2005)207 234 -lactosethroughoutthepaper.Stocksolutionsofindividualcompoundsinclude40%(w/v)glucose;5%(w/v)asparticacidneutralizedwithNaOH;2.5%methi-onine;1Mdisodiumsuccinate;and1MMgSO.Heat-inginamicrowaveovenishelpfulfordissolvingconcentratedstocksolutionsthatareslowtodissolve.Thesestocksolutionsaresterilizedbyautoclaving15minandstoredatroomtemperature.The50Msolu-tionmaybeclosetosaturationorsupersaturated;althoughbottlesremainedclearforlongperiods,occa-sionallyasampleshoweredcrystals,whichredissolvedreadilyuponheatinginamicrowaveoven.Anaminoacidmixturecontaining1%ofeachof17ofthe20natural-aminoacids,lackingmethionine,tyrosine,andcysteine,wassterilizedbyltrationandstoredintherefrigerator.Methioninewasomittedforconvenienceinlabeling,tyrosinebecauseitisnotsolubleenoughtoincludeatthisconcentration,andcysteinebe-causeslowoxidationtothemuchlesssolublecystinecausesprecipitatetoform.Themixtureof18aminoacids(includingmethioninebutlackingtyrosineandcysteine)wasaseectiveinpromotinggrowthofBL21(DE3)asamixtureofall20aminoacids(anexam-pleisgiveninTable7).Freeaminoacidswereusedtomakethemixture,exceptformonosodiumglutamate,asparaginemonohydrate,argininemonohydrochloride,lysinemonohydrochloride,andhistidinemonohydro-chloridemonohydrate.Themolarityof0.5%ofeachaminoacidusedinthemixtureisgiveninTable4.Whenconcentrationsofaminoacidmixturegreaterthanabout200g/mlofeachareused,theaminoacidsmayhavetobeneutralizedwithNaOHtokeepthepHofthenalmediumnearneutral.Astocksolutionof0.1MFeClwasdissolvedina100-folddilutionofconcentratedHCl(nalconcentra-0.12MHCl).Thissolutionwascombinedwithautoclavedstocksolutionsofothermetalstomakeatracemetalmixturecontaining50mMFeCl20mMCaCl,10mMeachofMnClandZnSO,and2mMeachofCoCl,CuCl,NiCl,NaMoO,Na,andH60mMHCl.Thesesolutionswerestoredatroomtemperature.Uponprolongedstorage,smallamountsofprecipitateformedinthemixture.Antibioticstocksolutionswerekanamycin(25mg/ml),chloramphenicol(25mg/mlinethanol),andampi-cillin(50mg/ml).Kanamycinwasinitiallyusedatg/mlandsubsequentlyat100g/ml(seeHighphos-phatepromoteskanamycinresistance).Chlorampheni-colwasusedat25g/mlandampicillinat50ThenamingconventionformedialistedinTable1andrelatedmediaistogivealetterdesignationtoeachuniquelydierentcompositionofthesaltsthatsupplyphosphate,ammonium,andsulfateions(otherthan).P,M,andLidentifysetsofmediathatsupply100,50,and25mMphosphate,respectively;NandCidentifyvariantsusedforisotopiclabelingwithNorC.Allmediacontain2mMMgSOandtracemetalmix(althoughtracemetalmixcanbeomittedinmediacontainingN-Z-amineandyeastextract).Abbreviationsforcomplexcomponents,ifany,areplacedaheadoftheletterdesignation,andabbreviationsforaminoacids,glycerol,glucose,andlactoseareplacedafter.Thus,Zindicates1%N-Z-amine,Yindicates0.5%yeastextract,andPindicatesthesaltscompositioninZYPmedium.Thedesignation505refersto0.5%glycerol,0.05%glu-cose(asinZYM-505);5052refersto0.5%glycerol,0.05%glucose,and0.2%lactose(asinZYP-5052);and750501refersto0.75%glycerol,0.05%glucose,and0.01%lactose(inC-750501).Gindicates0.5%glucose,asinPG;Dindicates0.25%aspartate,asinMDG;andAindicates200g/mlofeachof18dierentaminoacids(0.36%totalaminoacids),asinPAG.TheSinLSGrepresents20mMsuccinateandtheSMinPASMisforselenomethionine(SeMet).Thenamesofsomemediahavebeenshortenedfromdesignationsinprevi-ouslydistributedrecipes,asindicatedinTable1CultureconditionsCulturesweregrowninsterileglassvesselsinanincubatorshaker(NewBrunswickG25series),usuallyat300 350rpm,asindicatedonthemeter.Theincuba-tiontemperaturewas37C,unlessstatedotherwise.Targetproteinswereexpressedattemperaturesaslowas18C.Thestandardcongurationforgrowingcul-turesinparallelwastoplace0.5mlofculturein100mmculturetubeswithplasticcaps.Whenmorethanabout0.2mlofculturewastoberemovedforfollowingthetimecourseofgrowth,pHorinduc-tion,1.5mlofculturewasgrownin18150mmcul-turetubesor5 10mlofculturein125-mlErlenmeyerasks.Thesecongurationsprovidedsucientaerationtosustainlogarithmicgrowthtoanapproaching10inappropriatemedia,andexpressionresultsseemedtotranslatewelltogrowthin400 500mlculturevol-umesin1.8-or2.8-LbaedFernbachasks(Bellco),convenientforproducingmulti-milligramamountsofproteinsinanincubatorshaker.Higherratesofaera-tioncouldbeobtainedwithsmallervolumesofculturepervessel.Thestandardmeasureofculturegrowthwasopticaldensityat600nm()afterdilutioninwatertocon-centrationsthatgavereadingsbelow0.25ina1-cmpath-lengthcuvetteinaBeckmanDU640spectropho-tometer.ThepHofcultureswasmeasuredafter10-folddilutioninwater.Viabilityandstabilityofculturesgrownunderdierentconditionsweretestedbyplatingon1%agarplatescontainingZB,exceptasnoted.Via-bleculturesofBL21(DE3)producedapproximatelycoloniespermilliliterperoveraratherwiderange,fromlogphasethroughdensesaturatedcultures.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 PlaqueassayforinductionofT7RNApolymeraseTotestinductionofT7RNApolymeraseinexpres-sionhostsintheabsenceofatargetplasmid,thebacte-riophageT7deletionmutant4107wasusedused.ThismutantlackstheentirecodingsequenceforT7RNApolymeraseandisunabletoformaplaqueonalawnofcellsunlessthehostsuppliesT7RNApolymerase.WhenBL21(DE3)isgrownandplatedonmediathathavenoinducingactivity,thebasallevelofT7RNApolymeraseislowenoughthatonlysmallplaquesdevel-opatloweciency,andtheytypicallytakemorethan3htobecomevisible.Incontrast,whenBL21(DE3)isinducedbyincluding0.4mMIPTGintheplate,4107ecientlyformsthelargeplaquestypicalofwild-typeT7,whichbecomeapparentinlessthan2h.This4107plaqueassaywasusedtotestwhetherT7polymerasewasinducedinculturesofBL21(DE3)grownindierentmedia.AnalysisofproteinsonslabgelsProductionoftargetproteinwasfollowedbygelelec-trophoresisoftotalcellproteinsinthepresenceofso-diumdodecylsulfateonprecast4 20%polyacrylamidegels(Cambrex).CellswerelysedinBugbusterplusBen-zonase(Novagen)in50mMTris Cl,pH8.0,andcon-tainingeggwhitelysozymeat20g/ml.Lysozymeimprovesthereleaseoflargeproteinsintothesolublefractionbutwasomittedwhenitmightinterferewithidenticationofproteinsofaboutthesamesizeinthegelelectrophoresispattern.BenzonaseisaDNasethatreducesviscositythatcouldotherwiseinterferewithloadingsamplesorcausebandstosmearonthegel.Eithera5lysismixturewasaddeddirectlytoanappro-priatedilutionofculture,orcellswerepelletedbycentri-fuging1mininamicrocentrifuge(1.5mltubes),thesupernatantaspirated,andthepelletsuspendedin1sismixture.Thenalvolumeofcellsuspensionwasl,usuallyataconcentrationcorrespondingtoacul-turedensityof5,butsometimeshalfortwicethisconcentration.Immediatelyaftermixing,20lofcellsuspensionwastransferredtoasecondtube,andbothtubeswereleftforapproximately30minatroomtem-peratureforlysis.Oneofthetubeswasusedasthesam-pleoftotalcells,towhichwasadded10lof3buer(containingsodiumdodecylsulfate).Theothertubewascentrifuged1minandthesupernatantre-movedwithapipetterandmixedwithanother10lofloadingbuertoconstitutethesolublefraction.Thepellet(insolublefraction)wassuspendedin30ofloadingbuer.Allthreetubeswereheatedfor1mininaboilingwaterbathand10lofeachloadedonthegelforelectrophoresis.Rapidstainingofthegelafterelectrophoresisusesthefollowingprotocol.Thegelissuspendedin50mlof50%ethanol,10%aceticacidinacoveredplasticbox,heatedalmosttoboilinginamicrowaveoven(withthelidajar),andthenplacedonarockerforatleast5minatroomtemperature,duringwhichthegelshrinks.Theliquidisdiscardedandthegelissuspended50mlof5%ethanol,7.5%aceticacid,and200lofa0.25%solutionofCoomassiebrilliantbluein95%eth-anol.Aftergentlerockingtodispersethestain,thegelisagainheatedalmosttoboilinginamicrowaveovenandplacedontherocker.Theproteinpatternusuallybe-comesvisiblewithinafewminutesandcontinuestointensifyoverafewhours.Theresultcanusuallybevisualizedinlessthan30minbutthegelisusuallyrockedovernightbeforescanninganimageintothecomputer.AKimwipeplacedinthesolutionandrockedforafewminutescanrapidlytakeuptheslightamountofexcessstaininthesolvent.ResultsGrowthofshakingculturestohighdensityShakingculturesareconvenientforgrowingmanyculturesinparallel,andrapidgrowthtohighdensitiesisdesirableformaximizingtheyieldandeciencyofproducingtargetproteins.Complexmediacontainingenzymaticdigestsofcaseinandyeastextractareexten-sivelyusedbecausetheysupportgrowthofawiderangeE.colistrainswithdierentnutritionalrequirements,andculturestypicallygrow2 3timesfasterthaninsim-plemineralsaltsmediawithglucoseasthesolecarbonsource.However,complexmediacanvaryfromlottolotinabilitytosupportgrowth,andsomecomplexmed-iahavebeenfoundtoinducehigh-levelproductionoftargetproteinintheT7expressionsystemuponap-proachtosaturationwithoutaddedinducerinducer.Todeterminewhatfactorsmightlimitgrowthtohighden-sity,andtotrytounderstandandmanageunintendedinduction,theeectsofdierentcomponentsofgrowthmediaonsaturationdensity,growthrate,andinductionwereanalyzed.ResultstypicalofexploratoryexperimentsareshownTable2.CulturesofBL21(DE3)grownovernightinZB,where1%N-Z-amineisthesolesourceofnutrition,saturatedat1.2andpH7.9to8.2.Additionof0.5%yeastextract(togiveZYB)morethandoubledthesaturationdensityto2.8.Saturationdensityin-creasedapproximatelyinproportiontoconcentrationofN-Z-amineuptoabout4%,reaching6.9at8%.TriplingtheconcentrationofZYBalmosttripledthesaturationdensityto7.6.Additionof1%glucosetoZB,ZYB,4ZBor8ZBhadlittleeectonsaturationdensity,apparentlybecausetheacidgener-atedbyglucosemetabolismoverwhelmedthelimitedbueringcapacityofthesemediaanddecreasedpHF.W.Studier/ProteinExpressionandPurication41(2005)207 234 sucientlytostopgrowth.AlthoughgrowthratewasslowerinM9(mineralsaltsand0.4%glucose),thesatu-rationdensityof2.5wascomparabletothatinZYB.AddingZBtoM9tripledthesaturationdensity7.5,butincreasingtheglucoseconcentrationofM9ZBto2%overwhelmedthebueringcapacityofthe66mMphosphatebuerinM9andstoppedgrowthatalowerdensity,5.8andpHInducingactivitywasalsoanalyzedbytheabilityofBL21(DE3)growntosaturationtosupportplaquefor-mationby4107,aT7deletionmutantcompletelyunabletoformplaquesintheabsenceofT7RNApolymerasesuppliedbythehost.MediamadewithN-Z-aminefromouroldbarrel(OldZBinTable2)hadlittleifanyinduc-ingactivity.MediamadewithN-Z-aminefromthenewbarrel(fromwhichallmediaweremadeunlessspeciedotherwise)hadappreciableinducingactivity,andhigherconcentrationsofN-Z-aminehadhigherinducingactiv-ity,asjudgedbyplaquesizeandtimeofappearance.Additionof1%glucosestronglysuppressedinducingactivity,asfoundpreviouslybyGrossmanetal.al.,but0.1%glucosehadlittleeect,presumablybecauseitwasdepletedwellbeforesaturation.ThisinducingactivityisdiscussedfurtherinNon-inducingmediaandAuto-induction.IncreasingtheconcentrationofN-Z-amineand/oryeastextractcanincreasesaturationdensitybutcanalsoincreaseinducingactivityandisexpensiverelativetodeterminingandsupplyingpreciselywhatisneededforgrowthtohighdensity.Simplyadding1mMMgSOtoeitherZBorZYapproximatelydoubledthesaturationdensity(Table2).Althoughexcessglu-cosepreventedinduction,culturescouldbecomeacidicenoughtostopgrowth.Determiningandsupplyingwhatisneededforgrowthtohighdensityinbatchculturesandunderstandingandmanagingunintendedinductionhasbeenaniterativeprocess.Thefollowingsectionssummarizerstthegrowthmediathatre-sultedandthentheexperimentsandrationalethatledtothem. Table2SaturationdensitiesandinductionofT7RNApolymeraseindierentgrowthmediaGrowthmediumAdditionpHT74107plaquesNumberSizeTimeofappearance(h)ZB1.28.25ZYB2.87.65ZYB7.67.35OldZB1.07.9039Small4.5ZBIPTGinplate188Large2.00.58.62144Variable3.0ZB1.18.22150Variable3.02.48.53189Variable2.55.48.32183Variable2.06.97.92222Large1.5ZB0.1%glucose1.37.58150Variable2.5ZB1%glucose1.25.1238Tiny4.56.08.05231Large1.51%glucose4.35.2765Tiny,turbid3.55.57.37220Large1.51%glucose5.85.2845Tiny,turbid3.5ZYB2.58.55175Small3ZYB1%glucose3.25.3022SmallOvernightM92.56.10M9ZB7.57.02M9ZB2%glucose(total)5.84.57ZBM9PO+NHCl1.27.36ZB1mMMgSO2.18.35ZBM9salts2.67.40ZB1%glucose1.25.121mMMgSOBL21(DE3)wasgrown17h,37Cfrom10dilution,10mlin125-mlErlenmeyerasks.N-Z-amineingrowthmediawasfromthenewbarrel,whichhadinducingactivity,exceptthattheOldZBcultureandtheplatesfortesting4107plaqueformationweremadefromtheoldbarrel,whichlackedinducingactivity.EquivalentnumbersofT74107deletionphageparticleswereplatedon0.25mlofeachculturein2.5mlZBtopagaronZBplates,bothmadewithN-Z-aminelackinginducingactivity(oldZB).Contained0.5%NaCl.Grown14h,37C,0.5mlin13100mmtubes.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 High-phosphatePmediaFullydenedandcomplexPmedia(Table1)cansup-portthegrowthofBL21(DE3)andotherE.colitosaturationdensitiesof10orgreaterinreason-ablywellaeratedcultures.InPmedia,anequi-molarmixtureofNaandKHprovidesbueringagainstmetabolicallygeneratedchangesinpHinbothdirectionsandisasourceofsodium,potassium,andphosphateions.Aphosphateconcentrationof100mMwaschosentoprovideasmuchbueringcapac-ityaspossiblewithoutstressingthecells.Higherphos-phatelevelscanbetoleratedbutgrowthbeginstoslow,presumablybecauseofthehighionicstrength.Anadequatesupplyofnitrogenandsulfurissuppliedby25mMNH.Therequirementformagnesiumionsissatisedby1mMMgSO,theconcentrationgi-veninrecipespreviouslydistributed,buttherecipesgi-veninTable1callfor2mMMgSO,toprovidealargercushionforgrowthtoveryhighdensities.Tracemetalsarerequiredformaximalgrowthinfullydenedmedia.Thecombinedconcentrationofglucose,glycerol,andothersugarsintherecipesgiveninTable1islowenoughthattheyshouldbedepletedbeforeculturesbe-comeirreversiblyacidic,andsaturatedculturesusuallyhaveapHgreaterthan6.0.InfullydenedmediasuchasPAGandPA-5052,amixtureof18puriedaminoacidsincreasesgrowthrateaswellashelpingtoattainapproximatelyneutralpHatsaturation.Thestandardg/mlofeachaminoacidsupportedasmoothgrowthcurvetosaturationatdensitiesofatleast10,whereasdiscontinuitieswereapparentatconcen-trationsof100g/mlorless,presumablybecausedeple-tionofoneormoreaminoacidsrequiredtheinductionofsynthesispathways.ThedoublingtimeofBL21(DE3)inlog-phasegrowthat37Crangedfromabout60to70mininminimalmediatoabout30to35mininmediacontainingZYorthemixtureof18puriedaminoacids.TherecipesforPmediahavebeenwidelydistributedandusedsuccessfullytogrowstablestockculturesofT7expressionstrainsandtoproducetargetproteinsbyauto-induction.HighphosphatepromoteskanamycinresistanceExpressionvectorsthatconferresistancetokanamy-cinwereselectedforourstructuralgenomicswork,toavoidpossibleovergrowthofinducedculturesbycellsthathavelostplasmid.Suchovergrowthcanoccurwhenexpressionvectorsconferresistancetoampicillin,be-causesecreted-lactamasecandegradealloftheampi-cillininthemedium[1,2].However,IwassurprisedtondthatBL21(DE3)withoutanyplasmidgrewtohighdensityovernightat37Cinauto-inducingZYP-5052mediumcontaining25g/mlkanamycin,aconcentra-tionthatecientlykillstheminZBorZYBculturesorplates.Theculturesthatgrewhadtypicalplatinge-cienciesandremainedsensitiveto25g/mlkanamycininZYBplates.Furthermore,BL21(DE3)plateddirectlyonZPorZPGplatescontaining25g/mlkanamycinformedsmallerbutuniformcoloniesatnormale-ciency,indicatingthatallcellssurvivedandgrewinthesemedia.Systematictestsrevealedthattheincreasedresistancetokanamycinisduetohighconcentrationsofphos-phatecombinedwithaminoacidsandperhapsothernutrientsinrichmedia.Atakanamycinconcentrationof25g/ml,BL21(DE3)didnotgrowinZYB,whichhasnoaddedphosphate,norintheminimalPG,whichcontains100mMphosphate,butitgrewquitewellinthefullydenedPAG,whichcontainsboth100mMphosphateandpuriedaminoacids.Growthat25mlwasalsoobservedinothermediathatcontainrela-tivelyhighconcentrationsofphosphateandaminoacids,suchasM9ZB(64mMphosphate)andterricbroth(89mMphosphate)phate)(hereabbreviatedTRBtoavoidconfusionwithtryptonebroth(TB)).Inrichmedia,thehighertheconcentrationofphosphate,thehighertheconcentrationofkanamycinneededtopre-ventgrowthandkillcells:BL21(DE3)failedtogrowinM9ZBandTRBat50g/mlandwaskilledeectivelyat100g/ml;PAGculturesstillbecameturbidatakanamycinconcentrationof50g/ml,killingwassome-whatfasterthangrowthat100g/mlandkillingwaseectiveat200g/ml;ZYP-5052culturesstillbecameturbidat100g/ml,killingwasslightlyfasterthangrowthat200g/ml,andkillingwaseectiveat400g/ml.Althoughmanyuninducedexpressionstrainsarerelativelystableevenintheabsenceofselectiveanti-biotic,havingrichmediainwhichBL21(DE3)ismoresensitivetokanamycinseemedpreferabletoresortingtoconcentrationsashighas400g/mlwhenselectionisneeded.Afewattemptstodevelopanaminoacidmix-turethatwouldpromoterapidgrowthwithoutsubstan-tiallyincreasingkanamycinresistancewerenotsuccessful.Reducingthephosphateconcentrationingrowthmediaseemedthemostattractivewayofincreas-ingthesensitivitytokanamycin.LowerphosphateMandLmediaAsdescribedinthenextsection,culturescanbegrowntohighdensitieswithonlyminimalbueringofpHbyphosphateorotherbuers.TheMandLsetsofmedia(Table1)havephosphateconcentrationsof50and25mM,respectively.TheirsaltcompositionwasmodiedfromthatusedinPmediatoallowinde-pendentvariationofphosphate,sulfate,andammoniumions,whichisusefulfortestingnutritionalrequirementsandforisotopiclabeling.Non-inducingandauto-induc-ingLmedia(25mMphosphate)havebeentestedexten-sivelyandaresatisfactoryformostpurposes,buttheMF.W.Studier/ProteinExpressionandPurication41(2005)207 234 media(50mMphosphate)havesmallervariationsinpHduringgrowthandarecurrentlyusedforroutinework.BL21(DE3)iskilledaboutasfastasitdividesinZYM-5052containingkanamycinat50g/mlandiskilledfairlyeectivelyat100g/ml.Akanamycinconcentra-tionof100g/mlwasadoptedforroutinework.MetaboliccontrolofpHCulturesgrowinginmediacontainingglucose(andinwhichnoothernutrientislimiting)willcontinuetogrowuntiltheglucosebecomesdepletedortheacidgeneratedbythemetabolismofglucoseexceedsthebueringcapacityofthemediumandcausesthepHtodroptoalevelthatstopsgrowth.Aslongassucientglucoseispresentinthegrowthmedium,catabolismofothercarbonandenergysourcesthatcouldbalancetheacidgeneratedbymetabolismofglucoseispreventedbythephosphoenolpyruvate/carbohydratephosphotrans-ferasesystem(PTS),actingthroughcataboliterepres-sionandinducerexclusion[16 20].Intheabsenceofglucose,glycerolcansupportgrowthaboutaseectivelybutsuppressestheuseofothercarbonsourceslessdra-maticallythanglucosebyamechanismaectingcyclicAMPproductionction.Excessglycerolcanalsogenerateenoughacidtostopgrowth,but,incontrasttoglucose,thepresenceofglyceroldoesnotsuppresstheinducingactivityfoundincomplexmedia.Anotherfactorwithaprofoundinuenceongrowthistheavailabilityofoxygen.Iftheculturebecomesdenseenoughthatoxygenconsumptionexceedstherateofaerationintheshakingvessel,oxygenlimitationtrig-gerscomplexregulatoryresponsesthatattempttoadjustthemetaboliccapacitiesofthecelltotheavailabilityofoxygenandthecarbonandenergysourcesinthemed-med-.Thehighertherateofaeration(oroxygena-tion)thehighertheculturedensityattainedbeforeoxygenlimitationtriggerstheseresponses.Acidproduc-tionfromglucoseorglycerolseemstoincreaseasare-sultofthemetabolicchangesasoxygenbecomeslimiting.Imbalancesinneedsforenergyandcarboningrowthwithglucoseascarbonandenergysourcearetypicallyrectiedbyexcretionofacetateandothercompoundsintothemedium[23 25].Ifglucoseisdepletedbeforethemediumgetstoofaroutofbalance,theexcretedace-tateandothercarbonandenergysourcesthatmaybepresentinthemediumcanthenbemetabolized,whichcanreturnthepHofthemediumtotheneutraloralkalinerange.ThedecreaseofpHuponmetabolismofglycerolcanalsobereversedbymetabolismofothercarbonandenergysourcesinthemedium.ExcursionsofexternalpHoutsidetheneutralrangeoneithertheacidoralkalinesidealsoinducecomplexregulatoryresponsesresponses.Thestringentcontroloftheorderofcatabolismofdierentcarbonandenergysourcesinthegrowthmed-ium,togetherwiththecomplexregulatoryresponsestootherenvironmentalconditions,makeitchallengingtodevelopmediainwhichthepHremainsinarangethatsupportsgrowthtohighcelldensitiesinshakingvessels.The100mMphosphateinPmediaprovidesenoughbueringcapacitytoallowgrowthtodepletionof0.5%glucosewithasaturationdensitygreaterthan5whilemaintainingpHabove6.0.However,sig-nicantincreasesinglucoseconcentrationordecreasesinphosphateconcentrationusuallyproducedculturesthatsaturatedatlowpHandlostviabilitywithinhoursordays.InanattempttoprovideastrongerbueragainstdecreasingpH,whichwouldallowtheuseofhigherglucoseandglycerolconcentrationsorlowerphosphateconcentrations,organicacidswithrelativelyhighpweretestedfortheirabilitytobuerthemed-iumandtherebyallowgrowthtohigherdensity.OrganicacidsSuccinatewasfoundtobeeectiveincounteringtheacidgeneratedbyglucoseinminimalLmedium(whichhasonly25mMphosphate).ItisapparentfromresultsshowninTable3that,ratherthanactingsimplyasabuf-fer,succinateismetabolizedasglucosenearsdepletionduringgrowth:culturesreachahighersaturationdensityandahigherpHthanintheabsenceofsuccinate.ThegrowthrateandthechangesinpHduringgrowth(notshown)areconsistentwithglucosebeingmetabolizedrstandthensuccinate,asglucoseisdepleted.Approxi-mately20mMsuccinateseemsoptimalforbalancing0.5%(28mM)glucose,usuallyproducingsaturatedcul-tureswithapHclosetoneutral.SubstantiallyhigherconcentrationsofsuccinatecancausethepHtorisewellbeyond8.0,whichcanstressthecellsandreduceviabil-ity.ThepresenceofsuccinatedoesnotcausedetectableinductionofT7RNApolymerase,asmeasuredbythe4107plaqueassayandasindicatedbytheviabilityandstabilityofsaturatedculturesofstrainsthatexpresshighlytoxictargetproteins.Culturesthatsaturatebe-tweenpH6and7.5arestableforweeksintherefrig-eratorwithlittlelossofviabilityorincreaseinlagtimewhengrowingsubcultures.Fumarate,-malate,andcitratewerealsoabletobalancetheacidproducedbyglucoseinmuchthesamewayassuccinate.Addedace-tatewaseectivetoalesserextent.MaleateprovidedsomebueringagainstthedropinpHbutwastoxictoBL21(DE3)atlowpH,atleastinsomemedia.AminoacidsN-Z-amine,yeastextractoramixtureof18pureami-noacids(noY,C)increasebothgrowthrateandsatura-tiondensityofglucose-orglycerol-containingmedia.Uptakeofaminoacidsfromthemediumandincorpora-tiondirectlyintoproteinssparesthecellsfromhavingtomakeenzymesforentiremetabolicpathwaysanddivertcarbonfromglucoseintosynthesisofproteinsratherF.W.Studier/ProteinExpressionandPurication41(2005)207 234 thanproductionofenergyorothermetabolites.Iftheconcentrationofaminoacidsishighenough,atleastsomeofthemwillremaintobecatabolizedforcarbonandenergyafterglucoseisdepleted,causingpHtoriseandpotentiallybalancingacidgeneratedfromglucose.IncontrasttoN-Z-amine,puriedaminoacidscontrib-utednoinducingactivitywhenaddedtodenedmedia.TodeterminewhichaminoacidsaremosteectiveinbalancingpH,eachofthe18pureaminoacidsusedinthemixturewastestedindividuallyataconcentrationof0.25%forabilitytobalancetheacidgeneratedby0.5%glucoseinLmedium(25mMphosphate)().Themosteectivesingleaminoacidwasaspartate,followedbyserine,asparagine,glycine,andglutamate,allofwhichincreasedthesaturationdensity60 115%andproducedapH�6.2atsaturation(comparedwith4.1inglucoseitself).Bycomparison,20mMsuc-cinate(0.24%)increasedsaturationdensityby90%andproducedapHof6.8atsaturation,andthemixtureof18aminoacids(0.27%)increasedsaturationdensity75% Table3GrowthrequirementsforBL21(DE3)inmodiedLGmediaLGmedium,nosuccinateLGmedium+25mMsuccinatepHTiter(pHTiter(0000.05%0.246.890.667.031.80.10%0.76.761.47.203.60.15%1.26.622.27.337.20.20%1.66.513.17.4711.80.25%1.86.425.03.27.4715.50.30%2.66.197.14.07.6414.50.35%3.05.918.14.27.7612.00.40%3.04.854.67.698.80.45%3.84.575.17.877.80.50%3.34.415.37.849.500.76.670.110.66.670.460.1mM2.06.060.141.86.892.10.2mM2.65.503.26.493.90.5mM3.84.916.16.7513.41mM3.84.865.96.7412.72mM3.64.845.86.7514.2006.9807.025mM1.56.511.56.7510mM3.05.502.86.504.715mM3.44.904.05.959.120mM3.44.874.96.5216.425mM3.44.875.36.9212.550mM3.54.905.57.1013.40.786.751mM3.86.392mM3.96.095mM5.17.7510mM5.77.9615mM5.88.1520mM5.98.1925mM5.88.1135mM5.57.8350mM5.67.15006.9006.790.1mM0.416.810.123.96.735.00.2mM1.16.610.425.76.285.10.5mM3.75.646.46.208.81mM3.75.086.16.5113.22mM3.74.815.86.8012.3Culturesweregrown14 15h,37Cfrom10dilution,0.5mlin13100mmtubes.Culturesweretiteredafter3 4weeksintherefrigerator.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 andproducedapH5.7.Oftheotheraminoacids,onlyglutamineandprolineproducedasmuchasa25%in-creaseinsaturationdensityandonlyalanineproducedapH�4.0atsaturation.OftheaminoacidsthatfailedtobalancepHataconcentrationof0.25%,onlyalaninewaseectiveinbalancingpHwhentestedat0.5%.Sev-eralaminoacidssubstantiallyincreasedthelagorde-creasedthegrowthrateinminimalLGmedium,mostnotablyserine,alanine,leucine,andvaline,presumablybyrepressingoverlappingmetabolicpathwaysthways.Additionof0.01%eachofleucine,isoleucine,andvalinerestorednormalgrowthinthepresenceof0.25%serine(TheslowgrowthinhistidinemayreectalowpHofthemedium.)TodeterminewhichaminoacidsaremosteectivelyutilizedasacarbonandenergysourceforBL21(DE3),culturesweregrowninLmediumwithaminoacidsassolecarbonsource(Table4).Amixtureofthe18aminoacids,eachat100g/ml(0.18%totalaminoacids)wasprovidedtopromotesomegrowthandtoalleviatepos-sibleinhibitoryeectsofindividualaminoacids,whichwereaddedataconcentrationof0.5%.Themixtureof18aminoacidsbyitselfsupportedgrowthto1.4withanalpH6.8.Oftheindividualaminoacids,prolinewasthemosteectivecarbonandenergysource,supportinggrowthto9.6andpHcomparableto9.0andpH5.7supportedby0.5%glycerol.Otheraminoacidsthatsubstantiallyin-creasedthesaturationdensitywereserine,glutamate,alanine,andaspartate,withsmallerincreasesbythreo-nineandasparagine.Eachoftheseaminoacidsalsoin-creasedthenalpHatleastsomewhat,indicatingthattheyweremetabolized.ThenalpH5.2ofthehisti-dine-containingculturerepresentedasubstantialde-creasefromaninitialpHestimatedtobe6.0byreconstitution(versus6.6fortheotheraminoacids),suggestingthatmetabolismofhistidinedecreasesthepHoftheculture.TheremainingindividualaminoacidsdidnotsignicantlyaecteitherorpH,suggestingthattheywerenotsignicantlycatabolized.Acredibletestoftryptophanwasnotdone.MinimumnutritionalrequirementsforgrowthtohighdensityMetabolicbalancingofpHmadeitpossibletotesttherequirementforanynutrientincludingphosphatetosupportthegrowthofBL21(DE3),withoutthecom- Table4EectivenessofindividualaminoacidsinbalancingpHfrom0.5%glucoseorinservingasacarbonandenergysourceAdditionGrowninLGmediumGrowninLmedium+0.18%aapHConcentration8h22h22h5.7h46h46hGlucose0.5%0.584.04.060.5%28mM4.68.24.280.5%54mM2.49.05.72Succinate20mM0.747.66.800.5%42mM0.804.58.8018aa0.27%6.26.95.660.18%24mM0.561.46.78D0.25%4.88.17.550.5%38mM0.843.98.28S0.25%0.126.96.950.5%48mM1.35.97.43S+100ILV0.25%0.568.76.93N0.25%2.67.06.860.5%33mM0.722.07.02G0.25%1.16.46.580.5%67mM0.321.07.43E0.25%3.47.56.280.5%30mM0.725.17.28A0.25%0.164.24.530.5%56mM0.605.17.21A0.50%6.27.32P0.25%2.05.33.630.5%43mM0.929.67.00T0.25%0.524.93.640.5%42mM0.763.07.09Q0.25%1.36.13.940.5%34mM0.781.96.71I0.25%1.24.33.860.5%38mM0.441.16.67L0.25%0.223.73.650.5%38mM0.741.56.70V0.25%0.263.53.700.5%43mM0.641.36.73M0.25%0.664.33.820.5%34mM0.721.46.74R0.25%0.884.93.630.5%24mM0.701.56.81K0.25%0.904.33.600.5%27mM0.701.36.75F0.25%0.642.33.890.5%30mM0.681.26.81W0.25%1.34.43.680.5%24mM0.25%0.164.53.690.5%24mM0.061.05.19BL21(DE3)wasgrownfrom10dilution,0.5mlin13100mmtubes.g/mlofeachof18aminoacids(noCorY).0.25%serineand100g/mleachofisoleucine,leucine,andvaline.ReconstitutionindicatedaninitialpH5.4for0.25%histidineinLGandaninitialpH6.0for0.5%histidineinL+0.18%aminoacids.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 plicationoftheculturebecomingtooacidicorbasicforoptimalgrowth.Aseriesoftestsofmineralsaltsmediawithglucoseorglycerolasprimarycarbonsourceestab-lishednutrientconcentrationsthatlimitgrowthtolowdensities,whichcouldbeextrapolatedtodetermineapproximateminimumconcentrationsneededforgrowthtohighsaturationdensities.Table3showsre-sultsofoneseriesoftestsofminimalrequirementsforsulfur,nitrogen,phosphate,andmagnesiuminmodiedLGmedium,intheabsenceorpresenceof25mMsucci-nate.Thecultureswereinoculatedwitha1000-folddilu-tionofBL21(DE3)thathadbeengrowntosaturationinPG,and0.5mlculturesweregrownin13100mmtubesinashakingincubatorfor14 15hat37C.Con-clusionsfromtheseandsimilarexperimentsaresumma-rizedinthefollowingsections.Carryoverof0.026mMsulfateintheinoculumsup-portedgrowthto0.7withpH6.7.Theneedforsulfatesaturatedatapproximately0.5mM,inwhichBL21(DE3)grewto6.1atpH6.7.Asulfateconcentrationof0.5mMorgreateratnearneutralpHwasalsoenoughtoproduceverystablecultures,asmea-suredbyviabilityafterthreeweeksintherefrigerator.The5mMNainLandMmediaand25mMinPmediashouldsupplyenoughsulfurtosupportgrowthtoveryhighdensitiesinshakeasks.NitrogenSaturationdensitycontinuedtoincreasewithNHconcentrationuntilatleast50mM,whichsupportedgrowthto5.5atpH7.1.CulturesretainedhighviabilityforatleastthreeweeksintherefrigeratoratClconcentrationsof20mMorhigherandpHnearneutral.InminimalmediainwhichpHwasmaintainednearneutral,50mMNHreproduciblysupportedgrowthtoslightlyhigherdensitythan25mMandisthereforethestandardconcentrationusedinP,M,andLmedia.However,25mMNHClissucientformostpurposes,includinglabelingofproteinswithforNMRstudies.PhosphateCarryoverof0.1mMphosphateintheinoculumsup-portedgrowthto0.8atpH6.7.Thepresenceof1mMphosphateinthemediumsupportedgrowthto3.8atpH6.4buttheneedforphosphatedidnotappeartosaturateuntil10 15mMatandpHE.colicellshavecomplexregulatoryre-sponseswhenphosphatebecomeslimitinginthemed-med-,andalternativeusesofinternalphosphatemayaccountfortherelativelyslowincreaseinsatura-tiondensitybetween1and10mMphosphate.Thebuf-feringcapacityofphosphateinthemediumdidnotsignicantlyreducethepHincreaseduetosuccinatemetabolismuntil35 50mMphosphate.TheminimumphosphateconcentrationinthemediagiveninTable1is25mM,totrytoavoidaphosphatelimitationthatwouldinduceresponsemechanisms.Experimentsinwhichsaturationdensitieswerepushedwellabove10haveoccasionallysuggestedthateven25mMphos-phatemaybecomelimitingatdensitiesachievableinshakingvessels.MagnesiumNogrowthofBL21(DE3)wasapparentintheab-senceofmagnesium,but,interestingly,culturescontain-ingonlylimitingamountsofmagnesiumgrewtomuchhigherdensities(5-to10-fold)whenthegrowthmediumcontainedsuccinatethanwhenitdidnot.Theneedformagnesiumappearedtosaturateat0.5mM,whichgave6.4andpH6.2.However,viabilityafterthreeweeksintherefrigeratorseemedtoremainsomewhathigherinculturesgrownin1 2mMMgSOthaninthosegrownatlowerconcentrations.Magnesiumlevelsashighas10mM(thehighestconcentrationtested)showednoinhibitionofgrowth.PreviouslydistributedrecipesforPmediumcontain1mMmagnesium,but2mM(asgiveninTable1)mayprovideagreatermar-ginforgrowthtoveryhighdensities.TracemetalsFullydenedmediamadefrompuriedcomponentscontaincontaminatingtracemetalsinamountssucienttosupportgrowthtomoderatedensitybutnotsucientforgrowthtohighdensitywithgoodexpressionoftar-getproteinsbyauto-induction.Table5summarizesre-sultsfromanauto-inductionexperimenttotesttheeectsoftracemetals.Inthisexperiment,theexpressionstrainsaturatedinZYP-5052at18withthetargetproteinexpressedathighlevel.InslightlymodiedPA-5052withoutaddedtracemetals,saturationwasat4.4withlittleexpressionoftargetprotein.Additionoftracemetalsabouttripledthesaturationdensity,to13,andallowedhigh-levelexpressionoftargetprotein.Clearly,adeciencyoftracemetalslimitedculturegrowthandauto-inductionoftargetproteininthisfullydenedmedium.Individualmetalionsweretestedatconcentrationsof1,10,and100Mforabilitytoincreasesaturationden-sityandforpossibletoxicity(Table5).Thetracemetalswerechosenasbeinglikelytohaveafunctionalassoci-ationwithproteinsorparticipateinsomebiologicalpro-cess.Ironionsat10and100Mincreasedsaturationdensityto13but1Mincreasedthedensityonly7.8.Manganeseionsat1,10,and100Malsoincreasedsaturationdensityto13,asdidcobaltionsat1and10M.However,10Mcobaltionscausedalagofaboutanhourbeforeattainingnormalgrowthrate,and100Mcobaltpreventedgrowth.Zincionsappearedtohaveonlyaslightstimulatoryeect,F.W.Studier/ProteinExpressionandPurication41(2005)207 234 andnickel,molybdate,calcium,copper,selenateorbo-rateevenless.SelenatedidnotappeartobetoxicatMbutpreventedgrowthat100Manyproteinsofunknownfunctionarebeingpro-ducedinstructuralgenomicsprojects,anyoneofwhichmighthaveanunsuspectedmetalligand.Targetproteinsof50,000Daproducedat100mg/Lwouldhaveacon-centrationof2Mandproteinsof10,000Daaconcen-trationof10M.The1concentrationofmetalmixsupplies50Miron,20Mcalcium,10Mmanganeseandzinc,and2Mcobalt,copper,nickel,molybdate,selenate,andborate,amountsthatarenottoxictogrowthbutcouldsaturatepotentialbindingsitesinmanytargetproteins.Ofcourse,ifatargetproteinisknowntohaveametalligand,theappropriateconcen-trationofthatspecicmetalcanbeadded.Concentra-tionsbetweenabout0.1and2metalmixsupportedmaximumsaturationdensity,5wasslightlyinhibitoryand10markedlyslowedgrowthbuttheculturestillat-tainedhighdensityandahighlevelofauto-induction.Concentrationsof0.05metalmixorlowerdidnotsupportgrowthtohighdensityindenedmediaandproducedonlylowlevelsoftargetproteinbyauto-in-duction,primarilyduetoadeciencyiniron.Inthepresenceof0.02metalmix,anironconcentrationofMwassucientformaximumgrowthandauto-in-ductioninadenedmediumwithoutaminoacidsbutMwasneededinthepresenceofaminoacids.Thehighestironconcentrationtested,500M,showednoevidenceoftoxicity.InadenedmediumcontainingMFeCl,omissionofthemetalmixonlyslightlydiminishedthemaximumdensityandtheleveloftargetproteinproducedbyauto-induction,so100MFeClmaysuceformanypurposesifasuitablemetalmixisnotavailable.IncontrasttotheresultssummarizedinTable5manganeseorcobalt,aloneorincombination,didnotcompensateforadeciencyinironinsubsequentexper-iments.AdierencewasthatthemediausedinthetestsreportedinTable5containedsevendierentvitaminsbutsubsequentexperimentscontainednoaddedvita-mins.Whetherthepresenceofvitaminscouldaccountforthedierencehasnotbeentested.ComplexmediaTestsofnutritionalrequirementsforgrowthofBL21(DE3)tohighdensityincomplexmediaindicatethatmediacontainingonlyZYaredecientinmagne-sium,phosphate,carbon,andenergysources,aswellastheabilitytobuerpHchangesthatoccurduringgrowth.ThehighconcentrationsofaminoacidsinZYarealmostguaranteedtoprovidesucientnitrogenandsulfur,buttheknownvariabilityfromlottolotmakesitseemprudenttoadd0.2metalmix,oratleast Table5Eectsofmetalionsonsaturationdensityandauto-inductionatdierentmetalconcentrationsTargetproteinConcentrationoftracemetalmix00.1ZYP-505218.0PA-50524.4PA-5052Metalmix11.312.913.1+++MetalionconcentrationM10M100PA-5052FeCl7.812.713.7+++PA-5052MnCl11.813.212.9+++PA-5052CoCl11.113.6PA-5052ZnSO6.98.18.7++PA-5052NiCl4.87.75.2+PA-5052Na7.65.56.6+PA-5052CaCl6.04.75.3(+)PA-5052CuCl5.65.24.5(+)PA-5052Na5.86.3PA-5052H6.14.65.3(+)B832(DE3)RIL/P21wasgrown20h,37Cfrom10dilution,0.5mlin13100mmtubes.Themediacontained0.625%glycerolratherthantheusual0.5%in5052.ThePAmediumcontained200g/mlmethionineand100g/moftheother17aminoacids(noCorY).BothZYPandPAmediaalsocontained1Meachofnicotinicacid,pyridoxine,thiamine,vitaminB,biotin,riboavin,andfolicacid.Thetracemetalmixdieredfromthenalformulation.Thistracemetalmixcontained20MCaCl,10MeachofFeCl,MnCl,andZnSOMCoCl,and0.05MeachofCuClandNiClGrowthratewasnormalafteralagofaboutanhour.Thesecultureshadveryslowgrowth.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 Mofanironsalt,toensurethattracemetalrequire-mentsaremet.ThemineralsaltscomponentsofP,MorLmediaareincludedinallformulationsofcomplexmediainTable1toensurethatminimalrequirementsforgrowthtohighdensityandauto-inductionaremet.Fullydenedmediahavebeenformulatedwithwell-metabolizedaminoacidsatconcentrationshighenoughtoachievesaturationdensitiesequaltoorgreaterthanthoseobtainedincomplexmedia.However,yeastex-tractappearstosupplysomethingthatallowsslightlymorerapidinitialgrowththaninthosefullydenedmedia.Additionofvitamins,purines,andpyrimidinestothedenedmediahadlittleeectongrowthrateorsaturationdensity.Yeastextractsuppliesavarietyofmetabolites,includingfatsandcomplexcarbohydrates,anyofwhichmightberesponsibleforaslightlyfasterinitialgrowthrate.Non-inducingmediaBesidesournewbarrelofN-Z-amine,asampleofBactotryptone(Difco)alsohadinducingactivity,sug-gestingthatinducingactivitymaybefairlycommoninenzymaticdigestsofcasein.Additionofexcessglucosetocomplexmediathathaveinducingactivitypreventsinductionoftargetproteinn,butcultureseventuallybecomeacidenoughtostopgrowthandcanloseviabil-ity.Atintermediateglucoseconcentrations,culturesbe-cameinducedifthepHroseatsaturation,indicatingthatglucosewasdepleted,butnotiftheculturestayedacid,indicatingthatglucoseremainedintheculture.Therateofaerationalsohadasubstantialeectonsat-urationdensity,acidityandinduction.Itseemeddicultorimpossibletoformulatecomplexmediainwhichcul-turesreliablygrowtosaturationwithoutinductionanddonotbecomesoacidastoreduceviability.Therefore,thenon-inducingmediagiveninTable1arefullyde-ned,madewithpuriedcomponentsthathavenodetectableinducingactivity.WecurrentlyuseMDGmediumforroutinegrowthofnon-inducedculturesofBL21(DE3)expressionstrainsbuthavepreviouslyusedPGandLSGexten-sivelyforthispurpose.ThesemediasupportthegrowthofBL21(DE3)withadoublingtimeofapproximatelyanhour.Beingminimalmedia,theymustbeappropriatelysupplementedwhengrowingstrainswithnutritionalrequirements,suchasB834(DE3).Overnightculturestypicallysaturateat5to9andapHnearneutralwithoutdetectableinductionoftargetprotein.Whengrowntosaturationinthesemedia,evenstrainsthatex-presshighlytoxictargetproteinsremainstableandvia-bleforweeksintherefrigerator,andsubculturesgrowwithlittleornolag.Thismakesitconvenienttogrowbothfreezerstocksandworkingculturesovernighttosaturation,whereaspreviouslywetriedtocollectcul-turesinlogphasetominimizepotentialinstabilitiesifthetargetproteinistoxictothehost.Thecellsthatsettleoutofworkingculturesstoredintherefrigeratorusuallydispersereadily,butoccasionallytheyhavebeenstickyandmorediculttodisperse.Thereasonforthisocca-sionalstickinesshasnotbeendeterminedbutmaybeassociatedwithaslightlyalkalinepHinthesaturatedculture.Agarplatesmadewithfullydenednon-inducingmediasuchasMDAGorPAGenabledtheisolationofsomeBL21(DE3)transformantsthatwereunabletoformcoloniesontheZYBplatesweusuallyuseforselection.Apparently,theinducingactivityinZYBplatescausedenoughexpressionofhighlytoxictargetproteintopreventcolonyformation,butthelackofinducingactivityintheMDAGorPAGplatesallowedcoloniestoform.MDAGorPAGplatesarerichenoughthatinnocuousclonesformcoloniesonthemalmostasrapidlyasonZYBplates.Auto-inductionUnintendedinductionisalmostcertainlyduetolactoseinthemediumMediamadewithN-Z-aminefromtheoldbarreldidnothaveinducingactivity.Apparently,somethinginthenewN-Z-aminewascausinginduction(ratherthansomethingintheoldN-Z-aminepreventinginduction)becauseincreasingtheconcentrationofnewN-Z-amineinthemediumalsoincreasedtheinducingactivity,asjudgedby4107plaquesizeandtimeofappearance(ble2).Grossmanetal.al.hadconcludedthatunintendedinductionwasnotduetothepresenceoflactoseinthemedium.However,itseemedreasonabletotestwhetheradditionoflactosetomediamadewithN-Z-aminefromtheoldbarrelwouldproduceinducingbehaviorsimilartothatobservedinmediamadefromthenewbarrel.In-deed,theresultssummarizedinTable6showthatitdoes.Asexpected,noinductionofB834(DE3)P35wasapparentintheabsenceofaddedlactose.Thesmallestconcentrationoflactosetestedinthisset,0.005%M),gaveahighlevelofinductionofP35protein,buttheculturedensity,viability,andmaintenanceofplasmidwereallcomparabletowhatwasfoundintheabsenceofaddedlactose.Apparently,P35proteinisnotverytoxictothecell.Withincreasingamountsoflac-tose,productionofP35proteinremainedhighandthedensityofthesaturatedculturesdecreasedsomewhat,buttheviabilitydecreasedsubstantially,particularlyat0.05%lactoseandhigher.Atthesehigherlactoseconcen-trations,mostofthesurvivingcellshadlosttheexpres-sionplasmid.Highlevelsofinductionareknowntokillcellsthatcarryamulti-copyplasmidwithaT7pro-moter,evenifthetargetproteinisinnocuouscuous.220F.W.Studier/ProteinExpressionandPurication41(2005)207 234 Otherexperiments(notshown)foundthatproduc-tionofP35proteinwasstillappreciablewithaslittleas0.003%(83M)lactose,anddetectableonstainedgelsat0.001%(28M)butnotat0.0003%(8.3ThelimitofdetectionintheassayusedbyGrossmanetal.al.totestforpossiblelactoseintheirinducingmediumwasstatedbythemtobe0.002%,intherangewhereinductionofP35proteinwasobserved.IconcludethattheunintendedinductiondescribedbyGrossmanetal.andobservedbyusinmediamadewithN-Z-aminefromournewbarrelisduetosmallamountsoflactoseinthemedium.Thisseemsentirelyreasonable,asN-Z-amineisanenzymaticdigestofcasein,amilkprotein,andmilkcontainslactose.Thecaseinwouldhavebeenpuriedbeforedigestion,butdieringtraceamountsoflactoseremaininginthenalproductpresumablyac-countforthedierencesininducingactivityindierentlotsofN-Z-amineortryptone.Thendingthatglucosepreventsunintendedinductionisalsoconsistentwithalargebodyofworkshowingthatthepresenceofglucoseinthemediumpreventstheuptakeandutilizationoflac-[16 20].Inretrospect,wewereluckythatthebarrelofN-Z-amineusedformostofourpreviousworkindevelopingtheT7expressionsystemhadlowenoughlevelsoflactosetobefreeofunintendedinduction.Aminoacidssuppressinductionbylactoseinlog-phaseAlthough,thepresenceofasmallamountoflactoseinthemediumexplainsmostobservationsrelatedtounintendedinduction,itseemedcuriousthatB834(DE3)P35couldgrowtorelativelyhighdensityinZYPcontaining0.05 1%lactose,eventhoughhighlev-elsofinductionkillthecells(Table6).Indeed,thetiterindicatedthatmorethan90%ofthecellsinthesaturatedcultureswereincapableofformingacolony.SimilarresultswereobtainedwithB834(DE3)RILpro-ducingyeasttargetproteinP21,whichwasusedforanextensiveexplorationofinductionphenomena.TotalproteinsofcellsgrowinginZYPcontaining0.5%lactoseshowednodetectableP21proteininearlylogphasebutrapid,high-levelproductionasthegrowthrateslowedonapproachtosaturation(Fig.1A),similartothetim-ingobservedbyGrossmanetal.al..Thetimecourselookedsimilarwhetherthemediumcontained0.1,0.2,0.5,1or1.5%lactose,withinductionineachcasebegin-ningat1andreachingamaximumlevelofP21proteinper3,whichwasmaintainedto5to6.Whenincubationwascontinuedfor15hovernight,furtherincreasesinculturedensityweregreaterthehigherthelactoseconcentration,reachingashighas14.8in1.5%lactose.However,theamountoftargetproteinperwasmuchreducedFig.1A),andtitersshowedthatthedensityincreasesweredueprimarilytoovergrowthoftheculturebycellsthathadlostplasmid.SuchovergrowthcanoccurinZYPmediumevenatthekanamycinconcentrationofg/mlusedintheseexperiments(seesectionHighphosphatepromoteskanamycinresistance).SomethinginZYPmediumpreventsinductionbylactoseduringlog-phasegrowth.Conceivably,smallamountsofglucoseorotherPTSsugarscouldberesponsible,butN-Z-amineandyeastextractarebothrichinaminoacidsanditseemedpossiblethataminoacidssomehowpreventormodulatethelethallevelsofexpressionthatwouldotherwisebeinducedbylactose.Pmediumcontaining1.25%glycerolasacarbonanden-ergysourcewasusedtotesttheabilityofpuriedaminoacidstoallowgrowthinthepresenceof0.1%lactoseTable7).Nogrowthwasapparentintheabsenceofaminoacids,consistentwiththeinabilityofglyceroltopreventlactoseinductionthatisstrongenoughtopre-ventcellgrowth.However,additionof18aminoacids,eachataconcentrationof100g/ml,allowedgrowthtohighdensitywithfullinductionofP21protein.Ofthreesubgroupsofaminoacids,onlythegroupcontain-ingserinesupportedovernightgrowth,asdidserineit-selfbutnototheraminoacidsinthatsubgroup.Although,serineseemstobethemosteectiveaminoacidinsuppressinginductionandallowinggrowthinthepresenceoflactose,thecombinationof17amino Table6InductionasafunctionoflactoseconcentrationinZYPmadewithN-Z-aminethathasnoinducingactivityMediumLactoseconcentrationTargetproteinTiter(ZYP005.8012.614.0ZYP0.005%0.14mM5.8+++10.610.7ZYP0.01%0.28mM5.5+++6.96.7ZYP0.02%0.56mM4.9+++4.74.6ZYP0.05%1.4mM4.2+++0.120.48ZYP0.1%2.8mM4.4+++ZYP1%28mM4.1+++0.9B834(DE3)P35wasgrown13h,37Cfrom10dilution,0.5mlin13100mmtubes.indicatesthattiterwasbasedonfewerthan10colonies.Titerofcellsthatareresistanttokanamycinandthereforeretainplasmid.Titerintheabsenceofantibiotic,whichincludescellswithorwithoutplasmid.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 acidslackingserinepromotedgrowthinthepresenceoflactosealmostaswellas18aminoacidsincludingserine.Apparently,somethingabouttheuptakeandmetabo-lismofaminoacidsduringlog-phasegrowthpreventsormodulateslactoseinductionoftargetproteinsu-cientlytoallowcellstogrow,butthisinhibitionisre-laxedandfull-blowninductionoccursuponapproachtosaturation.Metabolicregulationenablesauto-inductionTherecognitionthatlactosecaninduceproductionoftargetproteinbutispreventedfromdoingsobycom-poundsthatcanbedepletedduringgrowthopenedthepossibilityofdevelopingmediainwhichtargetproteinisproducedautomatically,withouttheneedtomonitorgrowthandaddinduceratthepropertime.Icallthisauto-induction.Ideally,theexpressionstrainwouldgrowintheauto-inducingmediumwithoutexpressingtargetproteinuntilratherhighdensity,whendepletion Fig.1.Electrophoreticpatternsoftotalcellproteinsduringgrowthofauto-inducingculturesat37C.Equalculturedensitieswereanalyzedineachlaneofaset,andtheofthecultureatthetimeofsamplingisgivenaboveeachlane.(A)BL21(DE3)RIL/P21wasgrownin6mlZYPand0.5%lactoseina125-mlErlenmeyerask.Theculturewassampledevery30min,exceptthattheintervalbeforethelastsamplewas15h.Thecellsuspensionsbeforeprocessingforelectrophoresiswere10.(B)BL21(DE3)RIL/P21wasgrownin5mlZYP-5052ina125-mlask(exceptthattheglycerolconcentrationwas0.625%insteadof0.5%).Theculturewassampledevery30min.Thecellsuspensionswere10.(C)BL21(DE3)T7-10Awasgrownin2.5mlZYP-20052and25mMsuccinateina125-mlask(theglycerolconcentrationwas2%).Theculturewassampledevery30 40minuntil22.6,andthenintervalsof70min,55min,and13.5h.Thecellsuspensionswere5.(D)BL21-AI/T7-10Awasgrownin2mlZYM-5052and0.05%-arabinoseinan18150mmculturetube.Theculturewassampledevery30minuntil10.9,thenthreeintervalsof60minandanalintervalof16h.Thecellsuspensionswere Table7AbilityofaminoacidstosuppresslactoseinductionandallowgrowthofB834(DE3)RIL/P21inP+1.25%glycerol+0.1%lactose+g/mlmethionineAddition(100g/mleach)pHTargetprotein19h,3720aa12.46.28+++19aa(noC)12.75.18+++18aa(noC,Y)13.25.19+++GACPTKR0ILVSHNQ11.36.35+++FYWDE0ILV0S4.76.69+++14.5h,3718aa(noC,Y)10.66.20+++17aa(noS,C,Y)8.66.33+++Culturesweregrownfrom10dilution,0.5mlin13100mmtubes.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 ofinhibitoryfactorswouldallowthelactosepresentinthemediumtoinduceexpression,producinghighcon-centrationsoftargetprotein.Factorsthataecttheeciencyandreliabilityofauto-inductioninhigh-densitycultureswereexaminedsystematicallyinB834(DE3)andBL21(DE3),initiallytestingexpressionoftheyeasttargetproteinP21overawiderangeofconditionsandthenexpandingtootherproteins,includingbacteriophageT7proteinsthatareknowntobehighlytoxictothehostbacterium.Theexperimentsandconclusionsaresummarizedinthefol-lowingsections.Carbonandenergysourcesforhigh-levelproductionoftargetproteinbyauto-inductionAsdescribedinComplexmedia,growthinZYPislimitedbylackofacarbonandenergysource.Glucosecansupportgrowthtohighdensity,buttoomuchglu-cosepreventsinductionbylactose.LactoseitselfcansupportthegrowthofBL21(DE3),buttheinitialprod-uctsoflactosecatabolismareglucoseandgalactose,and,sinceBL21andB834cannotusegalactose,halfofthecarbonandenergyoflactoseisnotavailable.Per-hapsmoreimportant,inducedT7RNApolymerasecanbesoactivethatmosttranscriptionandproteinsynthe-sisinthecellisdirectedtowardtargetproteinprotein.Thiscompetitionmaylimittheproductionof-galactosidaseandlactosepermease,therebylimitingtheabilityoflac-tosetoserveasacarbonandenergysourceforcontin-uedproductionoftargetprotein.Glycerolsupportsgrowthaboutaswellasglucoseanddoesnotpreventinductionbylactose.Culturessup-plementedwithglycerolgrowtomuchhigherdensitiesbeforeandafterinductionthanwithlactoseascarbonandenergysource(forexample,compareFigs1AandB).BL21(DE3)cangrowonothereconomicalcarbonandenergysources,includingfructose,maltose,andsor-bitol(butnotsucrose).Inlimitedtests,maltoseandsor-bitolgavesomewhatinconsistentgrowthandinduction,oeringnoapparentadvantagesoverglycerol.There-fore,glycerolwaschosenasacarbonandenergysourceforbothfullydenedandcomplexauto-inducingmedia.Manycombinationsofglycerol,glucose,lactose,andpuriedaminoacidsweretestedtooptimizeauto-inductionandreliabilityinproducinghighconcentra-tionsoftargetproteinpervolumeofculture.Thestandard5052mixtureof0.5%glycerol,0.05%glucose,and0.2%lactosehasproducedreliableauto-in-ductionofawidevarietyofproteinsinarangeofmediaandgrowthconditions(Table1).ZYM-5052orZYP-5052isagoodchoicefortherstattempttoexpressal-mostanynewtargetprotein.Auto-inducedcultureswithhighlyexpressedproteins,suchasT7capsidproteinandyeastP21protein,oftenattaindensitiesgreaterthan20,morethantwicethedensityofBL21(DE3)orB834(DE3)themselvesgrowninthesamemedium.Microscopeobservationsofcellsfromsuchhighlyexpressingculturessuggestedthattheinducedcellscon-tinuedtoelongatefairlyuniformly,presumablywithoutdividing.Forsometargetproteins,higherglyceroland/orami-noacidconcentrationscanproducehighercultureden-sitiesandtargetproteinconcentrations,ifaerationandothermediacomponentsareappropriateformaintain-ingpH.Auto-inducedculturesexpressingT7capsidpro-teinhavereachedculturedensitiesof�40inlessthan24hinZYP-5052supplementedtoatotalof2%glyceroland25mMsuccinateinwell-aeratedculturesFig.1C).Comparably,highdensitieshavealsobeenreachedinfullydenedmediasupplementedwithpuri-edaminoacidsthatsupplycarbonandenergy.Eectofaerationontimingandlevelofauto-inductionoftargetproteinIntestingtheeectofdierentconcentrationsoflac-toseandglyceroloninductionofP21proteininZYPmedium,asubstantialdierencewasobservedintheamountofproteinproducedinZYPcontaining1.875%glycerolbutnoaddedlactoseontwodierentdays.Theonlyobviousdierencebetweentheculturesappearedtobethelevelofaeration:astandard0.5mlofcultureina13100mmtubereachedsaturationat13.9andpH5.6withahighleveloftargetpro-tein,buta5-mlcultureina125-mlErlenmeyerask,re-ducedtoahighlyaerated1.5mlbysampling,reachedsaturationat20.0andpH6.7withbarelydetectabletargetprotein.Totestmoresystematicallyhowgrowthandproteinproductionareaectedbylevelofaeration,a1000-folddilutionofB834(DE3)RIL/P21in80mlZYPcontain-ing0.625%glycerolbutnoaddedlactosewasdistributedas0.25,0.5,1or2mlsamplesin13100mmtubesand2.5,5,10,20or39mlsamplesin125-mlErlenmeyerasks,whichwereallgrownat37Cintheincubatorshakerat325rpmtoprovideafairlywiderangeofratesofaeration.Thetimecourseofgrowthandproteinpro-ductionintheErlenmeyeraskscontaining5mlormoreofculturewasfollowedbywithdrawingapproxi-mately12samplestotalingabout4mlfromeach,whichproducedaveryhighaerationrateallthewaytosatura-tionforthe5-mlcultureinthe125mlask.Twotimepointsandatotalvolumeofapproximately75 215weresampledfromtheremainingculturesbeforesatura-tion.Thesaturatedcultureswerealsotiteredwithandwithoutkanamycintotestforviabilityandplasmidretention.SaturationdensitiesandpH,relativetargetproteinlevels,andtitersaregiveninTable8AsshowninTable8,theleveloftargetproteinandviabilityofsaturatedculturesvariedtremendouslywiththerateofaeration:thehighestratesofaerationgavenoF.W.Studier/ProteinExpressionandPurication41(2005)207 234 apparentproductionofP21proteinorkillingofcellsandthelowestratesofaerationproducedveryhighlev-elsofP21proteinandsubstantialkillingofcells.Thedierentcultureswhosedensitiesweremeasuredinthegrowthphase(notshown)hadaboutthesamegrowthrateto1.0,wherelackofoxygenstartedtolimitgrowthrateinthecultureswiththelowestratesofaer-ation.Themosthighlyaeratedculturewhosegrowthratewasfollowed(5mlreducedto1mlina125-mlask)maintainedagraduallyslowingbutsteadyin-creaseindensityallthewaytosaturationat14.3,withlittleinductionoftargetprotein.Theleastwell-aeratedculturewhosegrowthratewasfollowed39mlina125-mlask)begansignicantproductionoftargetproteinby1.5andhadaccumulatedhighlevelsby3.Thedoublingtimeoftheculture33minbetweenof0.1and1butslowedmarkedlyto150minbetweenof3and5.Inthenext13hafterreaching5.3,theculturedensityreached10.2,withnoapparentdecreaseintheamountoftargetproteinper.Atthispoint,essentiallynocellsthatcarriedplasmidwerecapableofformingacol-ony,andcellsthathadlostplasmidhadnotyetover-growntheculture.Intermediateratesofaerationgavegrowthandinductionbehaviorintermediatebetweenthesetwoextremes.Thestandard0.5mlculturesin100mmtubesappearedtoprovideaerationcompa-rabletoabout5 10mlculturesin125-mlErlenmeyerasks,consideringthat4mlculturewasremovedfromthe10mlculturetofollowgrowthrateinthisexperi-ment.Inthissetofcultures,glycerolprobablybecamedepletedatthehigherlevelsofaeration,and,exceptforthelowestlevelsofaeration,mostculturesultimatelyreachedaboutthesamesaturationdensityandpHeventhoughtheamountsoftargetproteindieredmarkedly.ThefailuretoproducetargetproteinatthehighestratesofaerationintheaboveexperimentwasduetothelowconcentrationoflactosecontributedbytheN-Z-amine.Table9showsthesaturationdensities,targetproteinlevelsandtitersattainedatsaturationfortwosetsofculturesgrowninZYPcontaining0.625%glyc-erolanddierentconcentrationsoflactose.Intherstset,0.5mlculturesweregrownin13100mmtubes,providingthestandard,reasonablygoodrateofaera-tion;inthesecondset,1.5-mlculturesweregrownin125-mlErlenmeyerasks,providinganevenhigherrateofaeration.Intherstset,targetproteinwashighlyin-ducedevenintheabsenceofaddedlactose.Inthemorehighlyaeratedsecondset,littleinductionoftargetpro-teinorcellkillingwasapparentat0.001%orlessaddedlactoseandonlyminimalamountsoftargetproteinorcellkillingwereapparentbetween0.002and0.01%lac-tose.Thetypicalhighlevelsoftargetproteinandsub-stantialcellkillingseenwith0.5mlina13100mmtubewereproducedonlyat0.05%lactoseorhigher.Clearly,thehighertherateofaerationthemorelactoseisneededtoinducehigh-levelproteinproductioninauto-inducingmedia.Theconcentrationof0.2%lactosechosenforauto-inducingmediaseemslikelytobehighenoughtoinducefullexpressionoftargetproteinatal-mostanyrateofaerationlikelytobeencounteredwithshakingvessels.Inclusionofglucoseinauto-inducingmediaandexpressionoftoxicproteinsPreviousworkersusedlactosetoinducetheexpres-sionoftargetproteinsinT7expressionstrainsinfer-menters,addinglactoseafterglucosewasdepletededorusingafed-batchfermentationwithmixturesoflac-toseandglucose,whichappearedtoprovidelowerratesofinductionandimprovedsolubilityoftargetproteinn.However,intestingwhethermixturesofglucoseandlactosecouldproduceintermediateratesofproduc-tioninauto-inducingmedia,itwasclearthatthepres-enceofglucosecompletelypreventedinductionbylactoseandthatproductionoftargetproteinoccurredonlyaftertheglucosewasdepleted.Theseobservations Table8EectofaerationonsaturationdensityandproteinproductionVesselCulturevolumeSaturationTargetproteinTiter(pHPlasmid100mmtube0.2515.0+2129100mmtube0.515.47.15++1112100mmtube115.07.19+++2.66.5100mmtube28.56.34+++2.22.5125mlask2.513.07.0402525125mlask514.86.9902022125mlask1014.07.08++1517125mlask2014.37.09+++2.66.4125mlask3910.26.60+++0.13.4B834(DE3)RIL/P21wasgrown23h,37Cfrom10dilutioninZYP+0.625%glycerolwithoutaddedlactose.indicatesthattiterwasbasedonfewerthan10colonies.Titerofcellsthatareresistanttokanamycinandthereforeretainplasmid.Titerintheabsenceofantibiotic,whichincludescellswithorwithoutplasmid.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 areinaccordwithawealthofpreviousliteratureshow-ingthatglucoseinthemediumpreventslactosefrominducingthe[16 20]Becauseglucosebothpreventslactosefrominducingexpressionoftargetproteinandismetabolizedpreferen-tiallyduringgrowth,Iexpectedthatsimplyadjustingtheconcentrationofglucoseinmediacontaininganinducingconcentrationoflactosecouldallowauto-in-ductionatanydesireddensityofanactivelygrowingculture.However,thendingthataminoacidsandoxy-genlevelmodulatethelactoseinductionoftargetpro-teinsmeantthatnetuningtheculturedensityatwhichauto-inductionoccurredwasnotstraightforward,particularlyinrichmedia.Mediacontainingaminoacidsandlactosebutnoglucoseoftenexhibitedarela-tivelyslowproductionoftargetproteinwellbeforeara-pid,high-levelinductionthatcoincidedwithslowinggrowthduetooxygenlimitation.However,thepresenceofglucosealwaysstronglypreventedproductionoftar-getprotein.Therefore,agoodapproachappearedtobetoincludeglucoseinauto-inducingmediaataconcen-trationthatwouldnotbedepleteduntiltheculturehadgrowntomoderatedensity,preferablyjustbeforetheoxygendepletionthatappearstotriggerhigh-levelproductionoftargetprotein.Theeectsofdierentcon-centrationsofglucoseontheleveloftargetproteinaccu-mulatedweretestedindierentfullydenedandcomplexmedia,instandard0.5-mlculturesin13tubes,intimecourseswithlargervolumesofculture,andatdierentlevelsofaeration.Aglucoseconcentra-tionof0.05%seemedtobeeectiveoverarangeofcon-ditionsandwasselectedforinclusionintheauto-inducingmediagiveninTable1Animportantquestioniswhetherthepresenceof0.05%glucosecompletelypreventslactosefromincreas-ingthebasalleveloftargetproteinintheearlystagesofgrowthinauto-inducingmedia.Whentargetproteinsarehighlytoxictothecell,evenasmallincreaseinbasalexpressionoverthatmaintainedinnon-inducingmediamighthaveasignicanteectontheabilityofanexpres-sionstraintogrowandmaintaininducibleplasmidsun-tilauto-inductiontakesplace.ThiswastestedwithclonescapableofexpressingT7geneteins,whosefunctionsareunknownbutwhicharehighlytoxictoBL21(DE3)anddiculttomaintainandexpressexpress.CertainplasmidclonescapableofexpressingproteinweretoxicenoughthatBL21(DE3)transformantswerenotobtainedonZYBplates,whichhadinducingactivity,buttheywereob-tainedonfullydenedPAGplates,whichlackinducingactivity.TheseexpressionstrainswerestablymaintainedinPGandMDGnon-inducingmedia,andcouldbegrownandauto-inducedinPA-5052,ZYP-5052,andZYM-5052mediatoproduceastrongdoublebandattheapproximatepositionexpectedforproteininelectrophoreticpatternsoftotalcellproteins.proteinisevenmoretoxictoBL21(DE3),andclonescapableofexpressingitcouldbeobtainedonlyinvectorsspecicallymodiedtoacceptandexpresshighlytoxicproteins(tobedescribedelsewhere).Again,theseexpressionstrainswerestableinnon-inducingmediaandcouldbegrownandinducedinauto-inducingmed-ia.Auto-inductionofactiveproteincausedthecul-turetostopincreasingindensitybeyond0.5to1.5,presumablybecauseofthetoxiceectofthetargetproteinonthehost.Amutantproteinhavingasingle Table9LactoseconcentrationneededforinductionasafunctionofrateofaerationLactoseconcentrationModerateaeration0.5mlin13100mmtubeHighaeration1.5mlin125-mlErlenmeyeraskpHTargetproteinpHTargetproteinTiter(0015.37.19+++15.36.94?12140.0001%2.8M13.86.96?0.0002%5.6M13.96.97?0.0005%14M14.06.98?0.001%28M14.26.98?20210.002%56M13.57.00(+)21180.005%0.14mM13.97.00+210.01%0.28mM16.27.19+++15.16.98+17160.02%0.56mM16.47.20+++16.96.99++10100.05%1.4mM16.47.20+++18.97.01+++3.36.20.1%2.8mM16.67.19+++17.97.02+++2.25.00.2%5.6mM16.66.93+++19.27.00+++1.54.10.5%14mM17.16.93+++20.46.94+++0.55.41%28mM18.06.33+++27.46.82+++7.7B834(DE3)RIL/P21wasgrown17h,37Cfrom10dilutioninZYP+0.625%glycerol.Titerofcellsthatareresistanttokanamycinandthereforeretainplasmid.Titerintheabsenceofantibiotic,whichincludescellswithorwithoutplasmid.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 aminoacidsubstitutionproducedarelativelystrongbandattheapproximatepositionexpectedforteininelectrophoreticpatternsoftotalcellproteins.However,wild-typeproteinwasnotdetected,pre-sumablybecauseproteinsynthesisstoppedbeforeen-oughproteinaccumulatedtobecomevisibleoverthebackground.Clearly,basalexpressionoftargetpro-teininauto-inducingmediacontaining0.05%glucoseislowenoughintheinitialstagesofgrowththatstrainscapableofexpressingtargetproteinsthatarehighlytoxictoBL21(DE3)canbegrownandtargetproteinex-pressedinauto-inducingmedia.Auto-inductioniswidelyapplicableandgenerallysuperiortoIPTGinductionforproteinproductionAuto-inductionismoreconvenientthanIPTGinduc-tionbecausetheexpressionstrainissimplyinoculatedintoauto-inducingmediumandgrowntosaturationwithouttheneedtofollowculturegrowthandaddindu-ceratthepropertime.Furthermore,theculturedensityandconcentrationoftargetproteinpervolumeofculturearetypicallyconsiderablyhigherthanwhatwehadbeenobtainingbyIPTGinduction.Therefore,evenastheauto-inductionphenomenonwasbeingexploredandmediawerebeingoptimized,auto-inductionwasbeingappliedwithgreatsuccesstotheproductionofproteinsofinteresttousandtocolleaguesinourdepartment.Nevertheless,amoresystematicandwidercomparisonofauto-inductionandIPTGinductionwasundertaken.Athandwereexpressionstrainsforthersthundredorsoyeastproteinsselectedforourstructuralgenomicspilotproject.ThecodingsequenceshadbeenclonedinpET-13aorpET-28b,bothofwhichtranscribethetar-getfromaT7promoter.TheexpressionhostwasB834(DE3),withorwithouttheRILplasmidthatsup-pliestRNAsforcodonsrarelyusedbyE.coli.Thepres-enceofRILsubstantiallyincreasedproductionofseveraloftheyeasttargetproteinsanddidnotdecreasetheproductionofany.Allofthesecloneshadalreadybeentestedforexpressionandsolubilityoftargetpro-teinsbyconventionalIPTGinductioninM9ZYBatboth37and20Inall,72oftheyeastclonesweretestedforexpres-sionandsolubilityofthetargetproteinbyauto-induc-tion,mostofthematboth37and20C,andtheresultswerecomparedwiththepreviousIPTGinduc-tions.For14clones,IPTGandauto-inductionwerecompareddirectlyinthesameexperiment.Ingeneral,thelevelofexpressionperofculturedensityandthesolubilityoftargetproteinsappearedtobecompara-blewhetherexpressionwasinducedbyaddingIPTGorbyauto-induction.Theauto-inducedculturestypicallyhadconsiderablyhigherdensitiesandthereforealsohadconsiderablyhigherconcentrationsoftargetproteinpermilliliterofculture.Continuedincubationofauto-inducedculturesforseveralhoursafterfullinductionusuallyseemedtohavelittleeectonthesolubilityorleveloftargetproteinperofculturedensity,whethertheculturewasinamediumwherethedensityremainedconstantorcontin-uedtoincreaseslowlyafterfullinduction.Thisstabilityofauto-inducedculturesatornearsaturation,togetherwiththerelativeuniformityoftheinoculatingculturesgrowntosaturationinnon-inducingmedia,makesitconvenienttoscreenmanystrainsinparallelforexpres-sionandsolubility(orlargerculturesforpurication)simplybyincubating1000-folddilutionsovernightatC,orsomewhatlongerat20C.IPTG-inducedcul-tures,ontheotherhand,wereusuallycollected3hafterinductionat37Ctoavoidovergrowthbyunproductivecells.Occasionally,continuedincubationformanyhoursat37Creducedtheapparentsolubilityofthetar-getproteininanIPTG-inducedorauto-inducedcultureorboth.Intherarecaseswheresuchbehaviorwasob-served,thetargetproteinappearedsolubleinaparallelCinduction.Itisimportanttonotethatauto-inductionandsatu-rationoftenoccuratconsiderablyhigherdensityatCthanat37C(perhapsduetothehighersolubilityofoxygenatthelowertemperature).Highersaturationdensitiescombinedwithslowergrowthat20Cmeansthatculturesmaybequitedenseafterovernightincuba-tionbutnotyetbeinduced,socaremustbetakennottocollectlow-temperatureculturesbeforetheyhavesatu-rated.Theincubationtimecanbeshortenedbyincubat-ingat37Cforafewhours,untilculturesbecomelightlyturbid,andthentransferringto20Cforauto-induction.Auto-inductionhasbecomethestandardprocedureinourlaboratoryfortestingexpressionandsolubilityofproteinsproducedbyT7expressionstrainsandforproducingtargetproteinsinlargeamountsforpuri-cation.IPTGinductionisrarelyifeverusedanymore.Severalyeastproteinswereproducedbyauto-induc-tionandpuriedforpossiblestructuredetermination.Threethatgavecrystalssuitableforstructuredetermi-nationwerelabeledwithSeMetbyauto-induction(asdescribedinthenextsection)andyieldedstructures(P35,PDB1TXN;P89,PDB1NJR;andP96,PDB1NKQ),ashasthehumanSSATprotein(throughcollaborationwithJ.FlanaganandM.Bewley).DaxFuofthisdepartment(personalcommunicationandd)hasfoundthatauto-inductionincreasedtheyieldsofvedierentbacterialintegral-membraneproteinsabout10-foldoverthepreviousIPTGinduc-tion,toapproximately30 50mg/L.Sofar,hehasdeterminedthestructureofoneofthem.Recipesandprotocolsforauto-inductionhavebeendistrib-utedtomanyotherlaboratories,includingstructuralgenomicscenters,andareprovingtobehighlysuccessful.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 Auto-inductionforlabelingproteinswithSeMetforcrystallographyLabelingproteinswithSeMetisastandardanduse-fulwaytoobtainphasesforstructuredeterminationbyX-raycrystallographygraphy.Auto-inductionseemedpromisingasawaytoproduceSeMet-labeledtargetproteinssimplyandeciently.Expectingthatamethio-nine-requiringhostwouldbeneededforecientincor-porationofSeMet,IstartedwithB834(DE3),whichhadamethioninerequirementofunknowngenotypegenotype.However,itturnedoutthatSeMetlabelingisequallyecientinBL21(DE3).Theconcentrationofmethioninerequiredforgrowthandauto-inductionofB834(DE3)RIL/P21wastestedinafullydenedauto-inducingmediumcomparabletoPA-5052.Culturesgrewatthenormalrateuntilthemethioninewasdepleted,whentheculturedensityabruptlystoppedincreasing.Concentrationsequaltoorgreaterthang/mlofmethionineweresaturat-ingforgrowthto10.6.Productionoftargetpro-teinwasnotapparentatmethionineconcentrationslessg/ml(wheredensitystoppedincreasingat3.8)butincreasedwithmethionineconcentrationuntilthemaximumamountoftargetproteinperwasreachedatapproximately90g/ml.Asforcomplexcomponents,N-Z-aminesuppliedsaturatingamountsofmethioninebutyeastextractdidnot.GrowthinYP-5052stoppedat3.9withoutinductionoftargetprotein,equivalenttog/mlofmethionine;addi-tionof100g/mlofmethioninetothemediumallowedsaturationat9.5withfullinduction.Theconcen-trationofmethionineinthislotofyeastextract,althoughnotenoughtosupportgoodauto-induction,isprobablytoohightomakeitausefulsupplementinauto-inducingmediaforlabelingwithSeMet.Growthandauto-inductioninSeMetisstimulatedbymethionineB834(DE3)expressingyeasttargetproteinP07,whichhadpreviouslybeenlabeledwithSeMetintheprocessofstructuredeterminationion,wasusedtotestthepotentialforlabelingwithSeMetinauto-inducingmediacomparabletoPA-5052.TotalreplacementofmethioninebySeMetwasnoteective:SeMetconcen-trationsof50or100g/mlsupportedgrowthrelativelypoorly,tooflessthan2,withnoinductionoftargetprotein,and150and200g/mlpreventedgrowthen-tirely.TotestwhethersmallamountsofmethioninemightalleviatethetoxiceectsofSeMet,growthandauto-inductionweretestedinPA-5052containing5 g/mlmethionineplus50 200g/mlSeMet.Indeed,culturescontainingasmuchas150g/mlSeMetat-tained5to8(considerablyhigherthanwitheithermethionineorSeMetalone)andinducedlargeamountsoftargetprotein.However,thetoxiceectsof200g/mlSeMetovercameeven30g/mlmethio-nine,reducingsaturationdensityto1.9andpre-ventingauto-induction.BothSeMetandmethionineseemtobeusedatallstagesofgrowth,becausegrowthratein30g/mlmethioninewasreducedbythepresenceof100g/mlSeMet,andthegrowthcurveseemednottohaveadiscontinuitythatmightindicateastrongprefer-entialuseofmethionineuntildepletion.Thestimulatoryeectofmethionineongrowthandauto-inductioning/mlSeMetwascomparablebetween10andg/mlmethioninebutwassignicantlydiminishedin5Theseresultssuggestedthatauto-inductionmightproducetargetproteinwithmorethan90%replacementofmethioninebySeMet,ifthemethionineneededtostimulategrowthandauto-inductioncouldbelessthan10%oftheamountofSeMetinthemedium.TotestthelevelofincorporationofSeMetintotargetprotein,100-mlculturesofB834(DE3)expressingHis-taggedP07weregrownindenedmediacomparabletoPA-5052,containingeither200g/mlmethionineor10methionineand100g/mlSeMet.Theculturessatu-ratedat8.8and6.7andyielded2.8and1.9mgofpuriedP07protein,respectively.MassspectroscopydeterminedthattheP07proteinfromtheSeMet-con-tainingculturewasmorethan90%labeledwithSeMet.YeasttargetproteinP89wasonlypartiallysolublebuthadbeenpuriedfromauto-inducedculturesandcrystallized,soitwasagoodcandidateforSeMetlabel-ingbyauto-induction.However,atestof0.5mlcultureinthemediumusedforSeMetlabelingofP07producedrathersmallamountsofsolubleP89.Inanattempttoimprovetheyield,dierentconcentrationsofmethio-nineandSeMetweretriedalongwithdierentconcen-trationsoftheother17aminoacidsplusamixtureof9vitamins.Interestingly,thevitaminshadnoeectinPA-5052itselfbutsignicantlyincreasedboththesatu-rationdensityandlevelofP89producedbyauto-induc-tioninthepresenceofSeMet.TestsoftheindividualvitaminsshowedthatvitaminBwastheonlyoneneededforthestimulation:amixtureoftheother8vita-minsprovidednostimulation.Aslittleas3nMvitaminwassucienttoprovidemaximumstimulation.Growthandauto-inductionofB834(DE3)RIL/P89in400mlPASM-5052(whichcontains10g/mlmethio-nine,125g/mlSeMet,and100nMvitaminB)pro-duced4mgofpuriedSeMetprotein,sucientforphasingandstructuredeterminationination.B834isametEmutantIncontrolexperiments,itwasdiscoveredthatthepresenceofvitaminBinthemediumallowsnormalgrowthofB834(DE3)intheabsenceofmethionine.ThisunexpectedresultshowsthatthemethioninedeciencyofB834isduetoamutationin,whichspeciesthevitaminB-independenthomocysteinemethylaseF.W.Studier/ProteinExpressionandPurication41(2005)207 234 E.coli,whichcatalyzesthelaststepofmethioninesynthesisis.(PreviousreportsrtsthatB834isametBmutantprovidednosupportingdataandmustbeincorrect.)E.colialsocontainsavitaminBdenthomocysteinemethylase,speciedby.How-ever,sinceE.coliisincapableofsynthesizingvitamin,thisenzymeisactiveonlywhenvitaminBispres-entinthegrowthmedium.Concentrationsofvitamingreaterthan0.75nMallowedmaximalgrowthandauto-inductionofB834(DE3)RIL/P21inPA-5052lackingmethionine.ThediscoverythatB834isamutantsuggestsapossibleexplanationforthestimulationofgrowthandauto-inductionbyvitaminBinthepresencebutnottheabsenceofSeMet.Thepresenceofmethionineinthegrowthmediumrepressesthesynthesisofalloftheenzymesspecicformethioninesynthesisexceptforenzymeyme.SeMetseemslikelytohavethesameeect,sinceitspresenceinthegrowthmediuminhibitsthegrowthofBL21(DE3)andB834(DE3)toaboutthesameextenteventhoughBL21(DE3)wouldbefullycompetenttosynthesizemethionine.Animportantroleformethionineisincorporationinto-adenosylmethionine,amethyldonorinreactionsthatgenerate-adenosylhomocysteineasaproduct,whichisultimatelymetabolizedtohomocysteineeine.Sincemethionineisnottoxic,concentrationsinthegrowthmediumcanalwaysbemadehighenoughtosupplyalloftheneedsformethioninewithoutrecyclinghomocys-teine.However,attheconcentrationsofSeMetthatcanbetoleratedinthegrowthmedium,asubstantialfrac-tionmayultimatelyendupinSe-homocysteine,andtheremainingSeMetmaybeinsucientforcontinuedgrowthandsynthesisoftargetproteins.ThestimulatoryeectofvitaminBmightbeduetoitsactivationofthemetHhomocysteinemethyltransferase,whichregener-atesSeMetfromSe-homocysteine.Ifthisinterpretationiscorrect,vitaminBmightalsobeexpectedtostimu-lategrowthandauto-inductionoftargetproteinsinBL21(DE3)growinginthepresenceofSeMet.SomestimulationoftargetproteinbythepresenceofvitamininPASM-5052wasapparentinonetestwithBL21(DE3)P19butnotinasecondtest.VitaminBisincludedinPASM-5052ataconcentrationof100nM.SeMetlabelinginBL21(DE3)TheecientsubstitutionofSeMetformethionineinB834(DE3)inthepresenceofvitaminBreinforcedtheconclusionthatthecombinationof10g/mlmethionineand125g/mlSeMetinPASM-5052mustrepresstheendogenoussynthesisofmethionine.Therefore,SeMetlabelingbyauto-inductioninBL21(DE3),whichdoesnotrequiremethionineforgrowth,shouldbejustase-cientasinB834(DE3).Indeed,auto-inductionofhumanspermidine/spermineacetyltransferase(SSAT)inPASM-5052producedgreaterthan90%substitutionofSeMetformethioninewhetherproducedfromB834(DE3)RILorBL21-Gold(DE3)RIL.Thus,targetproteinscanbeecientlylabeledwithSeMetbyauto-inductioninBL21(DE3),andtheuseofB834(DE3)isnotnecessary.GeneralityofSeMetlabelingToexplorethegeneralutilityofauto-inductionforSeMetlabeling,productionandsolubilityoftargetpro-teinsinPASM-5052relativetoZYP-5052andPA-5052weretestedfor10dierentyeastproteinsexpressedinB834(DE3)RILbyauto-inductionofculturesgrownfrom1000-folddilutionsatboth37and20C.(Inthisset,PASM-5052contained100g/mlSeMet.)Allofthe37Cculturesappearedtobesaturatedandweresampledforgelelectrophoresisafteranovernightincubationof14h.The20CculturesinZYP-5052andPA-5052weresampledafter22h,butSeMetap-pearstoinhibitgrowthmuchmorestronglyat20thanat37C,andthe20CPASM-5052cultureswerenotsampleduntil65handagainat85h.Allbutoneofthese10targetproteinsappearedtobeproducedaboutaswellandtohavecomparablesolubilityinPASM-5052asintheothertwomediaatbothtempera-tures.Theseresultsindicatethatauto-inductioninPASM-5052shouldbegenerallyusefulforSeMetlabel-ingoftargetproteins.ASeMetconcentrationof125g/mlwaschosenforPASM-5052mediumbecauseitseemssucientbutnotmuchinexcessoftheamountneededtosupportgrowthandauto-inductioninthepresenceof10methionineand100nMvitaminB.CulturesgrowninsignicantlyhigherconcentrationsofSeMettendedtobecomeanorangebrowncoloruponprolongedincu-bationatsaturation.TheyieldofthefewSeMet-labeledproteinswehaveproducedbyauto-inductioninPASM-5052forstructuredeterminationhasbeencomparabletotheyieldoftheunlabeledproteins.Auto-inductionforlabelingproteinswithNandCforDeterminationofproteinstructuresbyNMRrequiressubstantialamountsofproteinlabeledwithNorAuto-inductioninfullydenedminimalmediaispoten-tiallyveryecientatincorporatingisotopiclabelsintotargetprotein.TargetproteinscanbeuniformlylabeledwithNsimplybyusingN-labeled(NHinP-5052orNHClinN-5052orLS-5052.Auto-inductionwithawellexpressedtargetproteinwillalmostdepletethe50mMammoniumioninthesemedia,souseoftheisotopeshouldbeveryecient.Reductionbelowabout25mMammoniumwillsignicantlyreducetheamountofwell-expressedtargetproteinobtained.GlycerolcanbeusedasasourceofCforlabelingtargetproteinsproducedbyauto-induction.Glucose,F.W.Studier/ProteinExpressionandPurication41(2005)207 234 themosteconomicalsource,cannotbeusedbecauseitpreventsauto-induction.Fortunately,Cglycerolisrel-ativelyeconomicalandundoubtedlywouldbecomecheaperifusageincreased.Theglucoseintheauto-in-ducingmediumwillhavebeendepletedbythetimetar-getproteinsynthesisbegins,butlactosemetabolismisnecessaryforauto-inductionandsomecarbonfromlac-toseislikelytobeavailableforincorporationintotargetprotein,atleastintheearlystagesofsynthesis.Theusualauto-inductionmediacontain0.2%lactoseand0.5%glycerol,concentrationschosentoensuremaximalproductionoftargetproteinevenatthehighestratesofaerationandtomakeitunlikelythattheculturewillgoirreversiblyacidevenatrelativelylowratesofaeration.ForecientClabelingoftargetprotein,theowofcarbonfromglycerolintotargetproteinshouldbemax-imizedandtheowfromlactoseminimized.AsdiscussedinthesectiononEectofaerationontimingandlevelofauto-inductionoftargetprotein,theconcentrationoflactoseneededformaximalinduc-tionoftargetproteindecreaseswithdecreasedrateofaeration.Usingaminimalmediumcontaining100mMphosphateforgoodbuering,auto-inductionofT7capsidproteinwasfollowedasafunctionofdecreasinglactoseconcentrationatthedierentratesofaerationprovidedbydierentvolumesofculturein13100mmtubes.Increasingconcentrationsofglycerolwerealsotested.Basedonthesetests,theauto-inducingmediumC-750501(Table1)shouldpro-videgoodClabelingoftargetproteinfromglycerol.Thismediumcontains0.75%glyceroland0.01%lac-tose,soalmostallofthecarbonenteringtargetproteinshouldbederivedfromglycerol.High-levelinductionwasobtainedattheaerationratedeliveredby0.75mlculturesin13100mltubes,asomewhatloweraera-tionratethanwiththestandard0.5mlpertube.In-ducedculturessaturatedat10inlessthan24hat37CwithapHusuallyabove6.0.Theseconditionsseemlikelytoscaletoapproximately200 400mlofcul-tureinanunbaed1-LErlenmyeraskorperhapsasmuchas1Lina1.8-LbaedFernbachask.SinceourstructuralgenomicsprojectdoesnotinvolveNMR,Ihavenottestedwhethertheeciencyoflabelinginthismediumisadequateforstructuredeter-mination.However,therecipehasbeendistributedtoseveralNMRgroupsinhopesthatitwillproveuseful.AtestofCincorporationasafunctionoflactosecon-centrationmightndthathigherlactoseconcentrationsandhigherratesofaerationalsoprovidesatisfactorylabeling.Auto-inductionwitharabinoseExpressionsystemsinwhichtranscriptioniscon-trolledbythepBADpromoterofthearabinoseoperonhaverelativelylowbasalexpression,whichcanmakethemusefulformaintainingandexpressingtoxicgenes[38,39].TheAraCproteinregulatesthepBADpro-moterbothpositivelyandnegatively,andbasalexpres-sionisfurtherreducedinthepresenceofglucose.ExpressionfromthepBADpromoterisinducedbyarabinoseandmodulatedbycataboliterepression.AtleasttwogroupshavereportedexpressionsystemsinwhichthesequenceforT7RNApolymerasehasbeenplacedundercontrolofthepBADpromotermoter,andInvitrogenmarketsBL21-AI,inwhichthepBADpromotercanexpressT7RNApolymerasefromthechromosomeofBL21.BasalexpressionofT7RNApolymerasefromthepBADpromoterinBL21-AIisex-pectedtobelowerthanthatfromthepromoterinBL21(DE3),whichisprobablyintrinsicallyleakybe-cause-galactosidaseactivityisneededtoconvertlac-tosetoallolactose,thenaturalinducerofthelactoseoperonron.Therefore,clonesexpressinghighlytoxictargetproteinsfromaT7promotermightbemain-tainedandexpressedmorereadilyinBL21-AIthaninBL21(DE3).Indeed,severalclonescapableofexpress-ingthehighlytoxicT7geneproteinthatcouldnotbeestablishedinBL21(DE3)werereadilyestablishedinBL21-AI.WhentranscriptionofthetargetgeneisfromaT7promoter,asintheexpressionclonesweareusing,fullexpressionrequiresbothinductionofT7RNApolymer-aseandreleaseofrepressorfromitsbindingsiteinthepromoter.InBL21(DE3),botheventsaretrig-geredbyreleaseoftherepressor,whichisconvention-allyinducedbyIPTGorauto-inducedbythepresenceoflactoseinthemedium.ThecombinationofexpressingT7RNApolymerasefromapBADpromoterinthechro-mosomeandthetargetgenefromtheT7promoterinamulti-copyplasmidprovidesenoughcontrolthatauto-inductionoftargetproteinproductionisfeasibleinBL21-AI.Auto-inductionofT7capsidproteininBL21-AIinZYM-5052and0.05%-arabinoseshowedbarelydetectablecapsidproteinat2.7,adistinctbandthatincreasedsteadilytohighlevelsbetween4and10,andanapproximatelyconstantamountperduringcontinuedincreaseinculturedensityto26(Fig.1D).Auto-inductionofthetoxicT7geneproteinstoppedgrowthat1.7withnoteinapparentingelpatterns,consistentwithresultsob-tainedinBL21(DE3).Thepresenceof0.05%glucoseintheauto-inducingmediumwasnecessarytoallowgrowthofthecloneinthepresenceof0.05%arabi-nose,andthegrowthwasaboutasrapidasintheabsenceofarabinose.Apparently,thepresenceofglucoseisnec-essaryandsucienttopreventsignicantinductionbyarabinoseinBL21-AIintheearlystagesofgrowthintheauto-inducingmedium.Incontrasttotheresultsob-tainedinBL21(DE3),theT7geneproteinwasnotapparentabovethebackgrounduponauto-inductioninBL21-AI.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 -Arabinoseconcentrationsbetween0.01and0.5%allinducedhigh-levelexpressionofT7capsidproteininZYM-5052(whichcontainslactosetoinduceunblockingoftheT7promoter).Capsidproteinwasdetectableinthepresenceofarabinoseandabsenceoflactose,butmuchlesswasproducedthaninthepres-enceofboth,providingameasureofhoweectivelyrepressorblockstranscriptionfromthepromoter(tobedescribedelsewhere).Togetabroadercomparisonofauto-inductioninBL21(DE3)andBL21-AI,42dierentclonesofyeastcodingsequencesundercontroloftheT7promoter,whichwereknowntobewellexpressedinBL21(DE3)RIL,werealsoplacedinBL21-AI/RIL.Inparalleltubes,theBL21(DE3)RILcloneswereauto-inducedinZYM-5052andtheBL21-AI/RILcloneswereauto-in-ducedinZYM-5052containing0.05%-arabinose.Lev-elsoftargetproteinweregenerallycomparableinthetwohosts,althoughafewclonesappearedtobeexpressedtoaslightlyhigherlevelortobeslightlymoresolubleinonehostortheother.Whetherthesedierencesrepresentexperimentalvariationoraremoresignicanthasnotbeenexplored.However,itisclearthatauto-inductionfromthepBADandT7promotersisgenerallyeec-tiveforproducingtargetproteinsinBL21-AI.OthergrowthmediaHigh-densityculturesforpreparationofplasmidsThehigh-densitycultureconditionsdevelopedforauto-inductionalsoareconvenientforpreparationofplasmidDNAs.RichmediasuchasZYM-505supportgrowthoftheplasmid-containingstrainsweworkwithtoculturedensitiesof10orhigherwhen1.5 2.5mlcultureisgrowninan18150mmtubeshakenat300 350rpm.Lactoseisomittedunlessauto-induc-tionisdesired.Thepresenceof0.05%glucoseensuresrapidinitialgrowthwithlittlelag.Typically,yieldsofplasmidDNAhavebeenseveralfoldgreaterthanob-tainedinmediapreviouslyusedforthispurpose,andasingle1.5-mlmicrofugetubeusuallyprovidesmoreplasmidDNAthanneededformostpurposes.Adequateaerationensuresgrowthtohighdensities,butevenmod-erateaerationgiveshighyields.JohnDunnofthisdepartment(personalcommunication)isusingauto-in-ducingmediatoobtainhighyieldsofasingle-copyplas-midthatcarriesaninduciblereplicationoriginundercontrolofapromoter.CommonlyusedcomplexmediacanbedecientinmagnesiumAsthispaperwasbeingwritten,Ithoughttocom-pareculturedensitiesattainedinZYM-505withterricbroth(TRB)and2YT,richmediacommonlyusedforhigh-densitygrowthsforpreparingplasmidDNAss.Allthreemediacontainanenzymaticdigestofcaseinplusyeastextract,butindierentconcentrations:YTcontains16gtryptone,10gyeastextract,and5gNaCl/LwhereasTRBcontains12gtryptoneand24gyeastextract,89mMphosphateand4mlglycerol(=0.5%w/v)perliter.TRBhasseveralofthesamecom-ponentsasZYM-505butaconsiderablyhigherconcen-trationofyeastextract.HavingpreviouslyfoundthatZYwasdecientinmagnesium,Itested2YTandTRBasdescribedandalsocontaining2mMMgSOTheresultsareshowninTable10ThemoststrikingresultwasthataddingmagnesiumtoTRBmadefromacommercialproduct(Gibco/BRL)increasedtheculturedensityfrom3.6toThestimulationfromaddingmagnesiumto2(madefromourownbarrelsofN-Z-amineandyeastex-tract)wasnotaslarge,increasingfrom5.7to8.3.ThedierenceindicatedthatourN-Z-amineandyeastextractprobablyhadhigherlevelsofmagnesiumthanthelotsusedtomaketheGibco/BRLproduct,andindeed,TRBmadefromourcomponentsproducedaconsiderablyhigher12.5,increasingtowithaddedmagnesium.TitrationoftheTRBfromGib-co/BRLindicatedthat0.5mMMgSOwassucienttosaturatethegrowthrequirement.Theseresultsconrmthatenzymaticdigestsofcaseinoryeastextractarelikelytobedecientinmagnesiumneededformaximumgrowth(todierentdegreesindierentlots),andthat1 2mMmagnesiumionshouldbeaddedtocomplexmed-iamadewiththesecomponentstoensuremaximumgrowth.The50%highersaturationdensityinTRB+magne-sium(18.5)relativetoZYM-505(12)isduethegreaterthan2-foldhigherconcentrationofcomplexcomponentsinTRB.Asimilarboostcouldbeachievedmoreeconomicallybyincreasingtheglycerolconcentra-tionofZYM-505,perhapsbalancingpHbyaddingsuc-cinateoraninexpensiveandwell-metabolizedaminoacidsuchasaspartateorglutamate. Table10Eectofmagnesiumonsaturationdensityin2YTandterricbrothGrowthmediumSourceZYM-505Local12.07.05YTLocal5.78.37YT+2mMMgSOLocal8.38.44TRBGibco/BRL3.67.73TRB+2mMMgSOGibco/BRL18.68.21TRBLocal12.58.06TRB+2mMMgSOLocal18.18.18BL21(DE3)wasgrown15h,37Cfrom10dilution,0.5mlin100mmtubes.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 DiscussionThephenomenonofunintendedinductionwasspo-radic,beingfoundinsomelotsofcomplexmediabutnotothersothers.Furthermore,dierentportionsofthesameculturemightproducewidelydierentlevelsoftargetprotein,dependingontherateofaeration().Therealizationthatlactoseisresponsibleforunin-tendedinductionmadeitpossibletodevelopnon-in-ducingmediainwhichT7expressionstrainsremainstableandviableallthewaytosaturation,andreliableauto-inducingmediathatproducehigh-density,fully-inducedculturescompletelyunattended.Thiswasaniterativeprocess,addressingfactorsthatlimitgrowthtohigh-densityinbatchmode,aecttheviabilityorstabilityofexpressionstrains,orinuencethelevelofproductionoftargetprotein.Interestingly,lackofmagnesiumlimitsgrowthintypicallotsoftraditionalrichmediasuchastryptonebrothorLB(andnewermediasuchasterricbroth).Withasuciencyofnutrients,themainlimitingfactorsbecomemainte-nanceofapHnearneutralandtheavailabilityofoxy-genastheculturebecomesdenseenoughthattherateofaerationbecomeslimiting.ThemediagiveninTablehavebeenformulatedtogivereliablenon-inducedculturesandgoodauto-inductionoverarangeofcon-ditions.WecurrentlyuseMDGasthenon-inducingmediumforgrowingculturesforfreezerstocksorworkingstocks,MDAGplatesforselectingstrainsthatexpresshighlytoxictargetproteins,ZYM-5052forauto-induction,andZYM-505forgrowinghigh-den-sityculturesforpreparingplasmids.Thesurprisingndingthathighphosphateconcen-trationinrichmediaprovidessubstantialresistancetokanamycinledtotheformulationoflowerphos-phatemediainwhichmetabolicbalancingmaintainsthepHofthemediumnearneutral.Growthinglucoseorglycerolproducesacidwhereasgrowthinaminoacidsortricarboxylicacidcycleintermediatessuchassuccinate,fumarate,malateorcitrateincreasesthepH.Although,thepresenceofglucoseorglycerollim-itsthecatabolismoftheseothercarbonandenergy[16 21],appropriatemixturescansupportgrowthtohighdensitieswithonlymoderateexcursionstowardacidpHfollowedbysaturationorextendedslowgrowthclosetoneutralpH.ByincreasingtheglycerolandaminoacidconcentrationswellabovethosegiveninTable1,auto-inductionofwell-ex-pressedtargetproteinshasproducedculturedensities�50inshakingbatchcultures,comparabletowhathasbeenreportedinafermenterter.Potentially,auto-inductioncouldproduceproteinseconomicallyonacommercialscale,ashigh-densityculturesfullyin-ducedfortargetproteincanbeobtainedwithoutcom-plexprocesscontrolsinmediamadeentirelyfrominexpensivecomponentssuchasmineralsaltsandmix-turesofglucose,glycerol,andlactose,supplementedwithfumarate,succinateorglutamate.Auto-inductiondependsonmechanismsbacteriausetoregulatetheuseofcarbonandenergysourcespresentinthegrowthmedium.Ifglucoseispresent,cataboliterepressionandinducerexclusionpreventtheuptakeoflactosebylactosepermease,theproductofthoughttobetheonlymeansoflactoseuptakeinwild-typecells[16 20].Whenglucoseisdepleted,lactosecanbetakenupbyasmallamountofpresentinuninducedcellsandconvertedtoallolactose,thenaturalinducer,by-galactosidase,theproductofof.Thus,inductionoftheoperonbylactoseshouldre-quirethepresenceofatleastasmallamountoflactosepermeaseand-galactosidaseintheuninducedcell,andauto-inductionshouldnotbeeectivewithstrainsthatlackeitheroftheseactivities.Contrarytothisexpecta-tion,Grossmanetal.al.observedexpressionoffromtheT7promoterinamulti-copyplasmiduponapproachtosaturationinBL26(DE3),aderivativeofBL21fromwhichtheoperonhasbeendeleteddeleted.Perhapschangesthatoccuronapproachtosaturationmakecellspermeabletolactosebysomeothermecha-mecha-.Thepresenceof0.05%glucoseinauto-inducingmed-iablocksinductionbylactoseintheearlystageofgrowthsoeectivelythatevenstrainscapableofexpressingtargetproteinshighlytoxictothehostcellcangrowandmaintainfunctionalplasmiduntilinduc-tion.Infact,basalexpressionmaybelowenoughthatantibioticmightnotbeneededintheauto-inducingmediumtoobtainhigh-levelproductionofmanytargetproteins.Havingacarbonandenergysourceotherthanlac-tosetosupportcontinuedgrowthandproductionoftar-getproteinafterinductionenhanceshigh-levelproductionoftargetproteinsfromT7expressionstrains.T7RNApolymeraseissoactivethatinductioncandirectmosttranscriptionandtranslationtothetar-getproteintein,whichmightinterferewithfullinductionoftheabilitytometabolizelactoseforenergy.Glyceroldoesnotinterferewithinductionoftargetprotein,anditspresenceinauto-inductingmediamorethandoubledtheyieldoftargetproteinrelativetowhatwasobtainedwithequivalentamountsoflactoseastheprimaryen-ergysource.Intheabsenceofglucose,aminoacidsappeartomodulateorpreventinductionoftargetproteinsbylac-toseintheearlystagesofgrowth,untilgrowthslowsasoxygenbecomeslimitinguponapproachtosaturation.Complexmechanismschangemetabolisminresponsetoaminoacidavailabilityandoxygenlevelslevels,butIwasnotawarethattheyareknowntopreventlac-toseutilization.Serineappearstobeparticularlyeec-tiveinpreventinginductionoftargetproteinsbylactoseinlogphasebutevenhighconcentrationsofF.W.Studier/ProteinExpressionandPurication41(2005)207 234 serinedonotpreventinductionasculturesapproachsaturation.Interestingly,serineistherstaminoacidtobedepletedduringgrowthonthemixtureofaminoacidspresentinatrypticdigestofcasein[24,25].PerhapstheabilitytopreventinductionisrelatedtoaneedforhigherlevelsofallolactosetoinduceexpressionfromtheT7promoterinamulti-copyplasmid,becausehigherthannormallevelsofrepressorarepresenttoensuresaturationofalloftherepressorbindingsites[2,4].Slowingofgrowthuponoxygenlimitationmightallowhigherlevelsofallolactosetoaccumulate,becauseofincreaseduptakeoflactosefromthemedium,de-creasedcatabolismofallolactoseorsomeotherchangethatpromotesinduction.Although,developedforexpressingtargetproteinsintheIPTG-inducibleT7expressionsystem,auto-in-ductioncouldinprinciplebedevelopedforanyexpres-sionsysteminwhichtheelementsdrivingexpressionofthetargetproteinareinducedbyachangeinmetabolicstatethatisbroughtaboutbygrowthofaculture.Thiscouldincludenotonlypromoterswhoseinductionispreventedbycataboliterepressionorinducerexclusion,butalso,forexample,promotersactivatedbyapproachtosaturation,oxygenlimitation,ordepletionofacom-pound(suchasmethionine)whosesyntheticpathwayisblockedbyitspresenceinthemedium.Simplyadding-arabinosetotheauto-inductionmediaofble1allowsthemtobeusedforproducingtargetpro-teinsinBL21-AI,whereT7RNApolymeraseisexpressedfromthechromosomebythearabinose-in-duciblepBADpromoter.ConsistentwiththegeneralviewthatthepBADpromoterhaslowerbasalexpres-sionthanthelacUV5promoter,plasmidsexpressingbacteriophageT7proteinsthatarehighlytoxictothehostcellweremoreeasilytoleratedinBL21-AIthaninBL21(DE3).Levelofexpressionandsolubilityofmosttargetproteinstestedwerecomparableinthetwohosts.Auto-inductionhasprovedtobegenerallyusefulforproducingawiderangeofproteins,includingmembraneproteins.Thestabilityandviabilityofculturesgrowninnon-inducingmediamakesitpossibletoworkwithmanystrainsinparalleloveraperiodofweeks.Re-transformationorstreakingoutculturesforafreshsinglecolony,anunfortunateandtediouspracticeinmanylabs,isalmostnevernecessaryforreproduciblyexpressinghighlevelsoftargetprotein.Culturesforauto-inductionaresimplyinoculatedandgrowntosat-uration,whichismuchmoreconvenientthanIPTGinductionandespeciallyconvenientforhighthroughputtestingofmanydierenttargetproteinsforexpressionandsolubility.Thehighculturedensitiesattainedbyauto-inductionproducemoretargetproteinpervolumeofculturethanIPTGinductionandalsomakeecientuseofexpensivereagentswhenlabelingwithSeMetorisotopes.Auto-inductionisconvenient,ecientandeconomicalforproducingproteinsatalmostanyscale,fromanalysisofindividualproteinsinsmalllaboratoriestoproductionofmanydierentproteinsinlargepro-jects,andpossiblyevenforproductionofproteinsonacommercialscale.AcknowledgmentsIamgratefulfortheenthusiasmandexperttechnicalsupportofmyco-workersandcolleagues.ClonesforexpressingyeastproteinsforstructuralgenomicswereconstructedandtestedbyIPTGinductionforexpres-sionandsolubilitybySue-EllenGerchman,withhelpinthelaterstagesfromEileenMatz,whoconstructedtheclonesofT7andhumanproteins.Auto-inductionandpuricationofnormalandSeMet-labeledyeastpro-teinswasbyHelenKycia.NancyManningperformedinnumerablegelelectrophoreticanalysesofproteinexpressionandsolubilityinauto-inducedcultures.StructuresofyeastproteinsweredeterminedbyS.Swaminathan,S.Eswaramoorthy,andD.Kumaran.WorkonhumanSSATwasincollaborationwithJohnFlanagan,MariaBewley,andVitoGraziano,whoper-formedthemassspectroscopymeasurementstodeter-minelevelsofSeMetsubstitutions.TheworkwassupportedbytheOceofBiologicalandEnvironmen-talResearchoftheUSDepartmentofEnergyandtheProteinStructureInitiativeoftheNationalInstituteofGeneralMedicalSciencesoftheNationalInstitutesofHealth,aspartoftheNewYorkStructuralGenomicsResearchConsortium.References[1]F.W.Studier,B.A.Moatt,UseofbacteriophageT7RNApolymerasetodirectselectivehigh-levelexpressionofclonedgenes,J.Mol.Biol.189(1986)113 130.[2]F.W.Studier,A.H.Rosenberg,J.J.Dunn,J.W.Dubendor,UseofT7RNApolymerasetodirectexpressionofclonedgenes,MethodsEnzymol.185(1990)60 89.[3]F.W.Studier,UseofbacteriophageT7lysozymetoimproveaninducibleT7expressionsystem,J.Mol.Biol.219(1991)[4]J.W.Dubendor,F.W.Studier,ControllingbasalexpressioninaninducibleT7expressionsystembyblockingthetargetT7promoterwithlacrepressor,J.Mol.Biol.219(1991)45 59.[5]K.C.Kelley,K.J.Huestis,D.A.Austen,C.T.Sanderson,M.A.Donoghue,S.K.Stickel,E.S.Kawasaki,M.S.Osburne,Regula-tionofsCD4-183geneexpressionfromphage-T7-basedvectorsinEscherichiacoli,Gene156(1995)33 36.[6]T.H.Grossman,E.S.Kawasaki,S.R.Punreddy,M.S.Osburne,SpontaneouscAMP-dependentderepressionofgeneexpressioninstationaryphaseplaysaroleinrecombinantexpressioninstabil-ity,Gene209(1998)95 103.[7]T.J.Giordano,U.Deuschle,H.Bujard,W.T.McAllister,RegulationofcoliphageT3andT7RNApolymerasesbythelacrepressor-operatorsystem,Gene84(1989)209 219.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 [8]P.J.Lopez,J.Guillerez,R.Sousa,M.Dreyfus,OnthemechanismofinhibitionofphageT7RNApolymerasebylacrepressor,J.Mol.Biol.276(1998)861 875.[9]S.K.Burley,S.C.Almo,J.B.Bonanno,M.Capel,M.R.Chance,T.Gaasterland,D.Lin,A.Sali,F.W.Studier,S.Swaminathan,Structuralgenomics:Beyondthehumangenomeproject,Nat.Genet.23(1999)151 157.[10]W.B.Wood,HostspecicityofDNAproducedbyEscherichia:BacterialmutationsaectingtherestrictionandmodicationofDNA,J.Mol.Biol.16(1966)118 133.[11]A.H.Rosenberg,B.N.Lade,D.S.Chui,S.W.Lin,J.J.Dunn,F.W.Studier,VectorsforselectiveexpressionofclonedDNAsbyT7RNApolymerase,Gene56(1987)125 135.[12]S.E.Gerchman,V.Graziano,V.Ramakrishnan,ExpressionofchickenlinkerhistonesinE.coli:Sourcesofproblemsandmethodsforovercomingsomeofthediculties,ProteinExpr.Purif.5(1994)242 251.[13]S.Eswaramoorthy,S.Gerchman,V.Graziano,H.Kycia,F.W.Studier,S.Swaminathan,Structureofayeasthypotheticalproteinselectedbyastructuralgenomicsapproach,ActaCrystallogr.DBiol.Crystallogr.59(2003)127 135.[14]D.Kumaran,S.Eswaramoorthy,F.W.Studier,S.Swaminathan,StructureandmechanismofADP-ribose-1-monophosphatase-pase),aubiquitouscellularprocessingenzyme,ProteinSci.(2005),inpress.[15]J.Sambrook,E.F.Fritsch,T.Maniatis,MolecularCloning:ALaboratoryManual,2nded.,ColdSpringHarborLaboratoryPress,ColdSpringHarbor,NewYork,1989.[16]N.D.Meadow,D.K.Fox,S.Roseman,Thebacterialphospho-enolpyruvate:Glycosephosphotransferasesystem,Annu.Rev.Biochem.59(1990)497 542.[17]P.W.Postma,J.W.Lengeler,G.R.Jacobson,Phosphoenolpyr-uvate:Carbohydratephosphotransferasesystems,in:F.C.Nied-hardt,(Ed.),EscherichiacoliSalmonellatyphimurium:CellularandMolecularBiology,2nded.(Chapter75),AmericanSocietyforMicrobiology,WashingtonDC,1996.[18]M.H.SaierJr.,T.M.Ramseier,J.Reizer,Regulationofcarbonutilization,in:F.C.Neidhardt(Ed.),Escherichiacolinellatyphimurium:CellularandMolecularBiology,2nded.(Chapter85),AmericanSocietyforMicrobiology,WashingtonDC,1996.[19]T.Inada,K.Kimata,H.Aiba,Mechanismresponsibleforglucose-lactosediauxieinEscherichiacoli:ChallengetothecAMPmodel,GenesCells1(1996)293 301.[20]K.Kimata,H.Takahashi,T.Inada,P.Postma,H.Aiba,cAMPreceptorprotein-cAMPplaysacrucialroleinglucose-lactosediauxiebyactivatingthemajorglucosetransportergeneinEscherichiacoli,Proc.Natl.Acad.Sci.USA94(1997)12914 [21]T.Eppler,P.Postma,A.Schutz,U.Volker,W.Boos,Glycerol-3-phosphate-inducedcataboliterepressioninEscherichiacoli,J.Bacteriol.184(2002)3044 3052.[22]A.S.Lynch,E.C.C.Lin,Responsestomolecularoxygen,in:F.C.Neidhardt,(Ed.),EscherichiacoliSalmonellatyphimuriumCellularandMolecularBiology,2nded.(Chapter95),AmericanSocietyforMicrobiology,WashingtonDC,1996.[23]H.Holms,FluxanalysisandcontrolofthecentralmetabolicpathwaysinEscherichiacoli,FEMSMicrobiol.Rev.19(1996)85 [24]D.E.Chang,S.Shin,J.S.Rhee,J.G.Pan,AcetatemetabolisminaptamutantofEscherichiacoliW3110:Importanceofmaintain-ingacetylcoenzymeAuxforgrowthandsurvival,J.Bacteriol.181(1999)6656 6663.[25]S.Kumari,C.M.Beatty,D.F.Browning,S.J.Busby,E.J.Simel,G.Hovel-Miner,A.J.Wolfe,RegulationofacetylcoenzymeAsynthetaseinEscherichiacoli,J.Bacteriol.182(2000)4173 4179.[26]J.L.Slonczewski,J.W.Foster,pH-regulatedgenesandsurvivalatextremepH,in:F.C.Neidhardt(Ed.),EscherichiacoliSalmonellatyphimurium:CellularandMolecularBiology,2nded.(Chapter96),AmericanSocietyforMicrobiology,Washing-ton,DC,1996.[27]E.McFall,E.B.Newman,Aminoacidsascarbonsources,in:F.C.Neidhardt,(Ed.),EscherichiacoliSalmonellatyphimuriumCellularandMolecularBiology,2nded.(Chapter22),AmericanSocietyforMicrobiology,Washington,DC,1996.[28]F.C.Neidhardt,EscherichiacoliSalmonellatyphimuriumCellularandMolecularBiology,2nded.,SectionB1,BiosynthesisofAminoAcidsandNucleotides,AmericanSocietyforMicro-biology,Washington,DC,1996.[29]B.L.Wanner,Phosphorusassimilationandcontrolofthephos-phateregulon,in:F.C.Neidhardt,(Ed.),EscherichiacoliSalmonellatyphimurium:CellularandMolecularBiology,2nded.,(Chapter87),AmericanSocietyforMicrobiology,WashingtonDC,1996.[30]P.Neubauer,K.Hofmann,O.Holst,B.Mattiasson,P.Kruschke,Maximizingtheexpressionofarecombinantgeneinbymanipulationofinductiontimeusinglactoseasinducer,Appl.Microbiol.Biotechnol.36(1992)739 744.[31]B.J.Homan,J.A.Broadwater,P.Johnson,J.Harper,B.G.Fox,W.R.Kenealy,Lactosefed-batchoverexpressionofrecombinantmetalloproteinsinEscherichiacoliBL21(DE3):Processcontrolyieldinghighlevelsofmetal-incorporated,solubleprotein,ProteinExpr.Purif.6(1995)646 654.[32]Y.Chao,D.Fu,ThermodynamicstudiesofthemechanismofmetalbindingtotheEscherichiacolizinctransporterYiiP,J.Biol.Chem.279(2004)17173 17180.[33]W.A.Hendrickson,J.R.Horton,D.M.LeMaster,Selenomethi-onylproteinsproducedforanalysisbymultiwavelengthanoma-lousdiraction(MAD):Avehiclefordirectdeterminationofthree-dimensionalstructure,EMBOJ.9(1990)1665 1672.[34]G.N.Cohen,I.Saint-Girons,Biosynthesisofthreonine,lysine,andmethionine,in:F.C.Neidhardt(Ed.),EscherichiaColinellaTyphimurium:CellularandMolecularBiology,AmericanSocietyforMicrobiology,Washington,DC,1987,pp.429 444.[35]N.Budisa,B.Steipe,P.Demange,C.Eckerskorn,J.Kellermann,R.Huber,High-levelbiosyntheticsubstitutionofmethionineinproteinsbyitsanalogs2-aminohexanoicacid,selenomethionine,telluromethionineandethionineinEscherichiacoli,Eur.J.Biochem.230(1995)788 796.[36]M.Pieper,M.Betz,N.Budisa,F.X.Gomis-Ruth,W.Bode,H.Tschesche,Expression,purication,characterization,andX-rayanalysisofselenomethionine215variantofleukocytecollagenase,J.ProteinChem.16(1997)637 650.[37]C.H.Miller,J.A.Duerre,-ribosylhomocysteinecleavageenzymeEscherichiacoli,J.Biol.Chem.243(1968)92 97.[38]C.Cagnon,V.Valverde,J.M.Masson,Anewfamilyofsugar-inducibleexpressionvectorsforEscherichiacoli,ProteinEng.4(1991)843 847.[39]L.M.Guzman,D.Belin,M.J.Carson,J.Beckwith,Tightregulation,modulation,andhigh-levelexpressionbyvectorscontainingthearabinosePBADpromoter,J.Bacteriol.177(1995)4121 4130.[40]D.R.Wycu,K.S.Matthews,GenerationofanAraC-araBADpromoter-regulatedT7expressionsystem,Anal.Biochem.277(2000)67 73.[41]Y.P.Chao,C.J.Chiang,W.B.Hung,Stringentregulationandhigh-levelexpressionofheterologousgenesinEscherichiacoliusingT7systemcontrollablebythearaBADpromoter,Biotech-nol.Prog.18(2002)394 400.[42]J.Beckwith,Theoperon:Anhistoricalaccount,in:F.C.Neid-hardt(Ed.),EscherichiaColiSalmonellaTyphimuriumCellularandMolecularBiology,AmericanSocietyforMicrobi-ology,Washington,D.C.,1987,pp.1439 1452.F.W.Studier/ProteinExpressionandPurication41(2005)207 234 [43]R.E.Huber,G.Kurz,K.Wallenfels,Aquantitationofthefactorswhichaectthehydrolaseandtransgalactosylaseactivitiesofbeta-galactosidase(E.coli)onlactose,Biochemistry15(1976)[44]E.B.Newman,R.T.Lin,R.DAri,Theleucine/Lrpregulon,in:F.C.Neidhardt(Ed.),EscherichiacoliSalmonellatyphimu-:CellularandMolecularBiology,2nded.(Chapter94),AmericanSocietyforMicrobiology,Washington,DC,1996.F.W.Studier/ProteinExpressionandPurication41(2005)207 234