ANALYTICALBIOCHEMISTRYARTICLENO.

247,185–192(1997)

AB972061

High-Ef ciencyBlottingofProteinsofDiverseSizesFollowingSodiumDodecylSulfate–PolyacrylamideGelElectrophoresis

MarkW.BoltandPaulA.Mahoney1

DivisionofBiologicalSciences,UniversityofMissouri,Columbia,Missouri65211

ReceivedSeptember11,1996

High-molecular-weightproteinsoftenblotontoni-trocellulosemembranespoorlyfollowingsodiumdo-decylsulfate–polyacrylamidegelelectrophoresis,re-sultinginlowlevelsofdetectiononimmunoblots,andhencedif cultyinanalyzingrareproteins.Moreover,optimizingconditionsforthetransferofhigh-molecu-lar-weightproteinstonitrocellulosefrequentlyresultsintheinef http://ingradiolabeledproteinstandardsandphos-phorimagingtechnology,wehavequantitatedtheef- cacyofmanydifferentproteingelelectrophoresisandblottingprotocols.Herewereportnovelgelandblottingconditionswhichsigni cantlyimprovethetransferandretentionofhigh-molecular-weightpro-teins,withoutsacri cingtheef http://ingthisnewlydescribedprocedure,wehavedetectedarare500-kDaproteininimmunoblotswhichwaspreviouslynotdetectablewithanyofthecommonlyusedblottingprocedures.Sincetheimprovedconditionsofferincreasedsensi-tivityacrossaspectrumofproteinsizes,theyshouldbewidelyapplicable. 1997AcademicPress

Theelectrophoretictransferofproteinsfrompoly-acrylamidegelstonitrocellulosesheets,initiallyde-scribedbyTowbin(1),hasenjoyedwidespreaduseasavaluabletoolinthe eldofproteinresearch.Oneapplicationofparticularimporthasbeenthesubse-quentemploymentofantibodyprobesdirectedtowardthenitrocellulose-boundproteins.Variouslytermedimmunoblotting,proteinblotting,or‘‘Western’’blotting

Towhomcorrespondenceshouldbeaddressed.Fax:573-882-0123.E-mail:mahoney@biosci.mbp.missouri.edu.

0003-2697/97$25.00

Copyright 1997byAcademicPress

Allrightsofreproductioninanyformreserved.

1

(2),themethodologyofthisapplicationhaschangedlittlesinceitwasoriginallyintroducedforusewithurea-orSDS2-containingpolyacrylamidegels(1).Nu-merousattemptstoimprovetheoriginalprotocolhavefocusedonincreasingtheamountofproteintrans-ferredtoandretainedonthenitrocellulosesheet,par-ticularlywhendealingwitheitherverysmallorverylargeproteins.Theobservationthatsmallproteinstendtomovethroughthenitrocellulose,withonlyafractionofthetotalamountactuallybinding,ledtoananalysisshowingthatnitrocellulosewithsmallerporesizeswasmoreeffectiveinretainingsmallproteins(3).Morerecently,Tooetal.describetheuseofgelatin-coatednitrocelluloseforthequantitativeretentionofsmallproteins(4).Conversely,investigatorsanalyzinglargeproteinshaveoftensoughttoincreasetheef -ciencyoftransferbyenhancingthedegreeofproteinmigrationoutofthegelduringthetransfer,usingtech-niquesrangingfromdisruptionofthegelmatrixtopartialproteolyticdigestionoftheproteinspriortotransfer(5–7).Theseexampleshighlighttwomajorfactorsin uencingtheef ciencyofproteinblotting:theelutionef ciencyofaproteinoutofagivengelmatrixandtheef ciencyofbindingbythemembrane.Im-provementsinoneareaoftenappeartobegainedatsomecostintheother;methanolhasbeenreportedtoincreasethebindingcapacityofnitrocellulose,perhapsthroughhydrophobiceffects(8),butitmaydecreasetheelutionef ciencyoflargeproteinsbypartially‘‘ xing’’theminthegel(forreview,see9).

Recently,oureffortstostudyalarge(ca.500kDa)membraneproteinwerestymiedbyourinabilitytode-tecttheproteinonimmunoblots.Suspectingthatthisre ectedapoorelutionef ciencyfromthegel,weem-barkedonasystematicstudyofsomeoftheparameters

Abbreviationsused:TEMED,N,N,N ,N -tetramethylenediamine;SDS,sodiumdodecylsulfate.

185

2

186

BOLTANDMAHONEY

commonlyemployedduringdenaturingSDS–http://ingradiolabeledproteinsandphosphorimaginganalysis,wewereabletoquantitatetransferef cienc-iesundermanydifferentconditions,includingthosecommonlyusedforelectrophoresisonaLaemmli-typegel(10)followedbyproteinblotting(11,12).Ourre-sultsshowthatthemaindeterminantofproteinblot-tingef ciencyrestsnotintheabilityofproteinstoexitaLaemmli-typegelduringblotting,butratherintheef ciencyofproteinbindingtothenitrocelluloseblot-tingmembrane.Byalteringthecompositionofthebufferusedduringtheblottingstep,wewereabletoincreasetheretentiononnitrocelluloseofmostproteinstested.Mostimportantly,whenthismodi edblottingbufferwasusedincombinationwiththegelsystemdescribedbyFairbanksetal.(13),wefoundasigni cantincreaseintheoverallblottingef ciencyofproteins,overawiderangeofproteinsizes.Forthe200-kDamyosinheavychain,theamountofproteintransferredwasalmostfourfoldmorethanthatobtainedusingmoreconventionalconditions.Moreover,theuseofthisnewcombinationofgelsystemandmodi edbufferal-lowedustodetectahigh-molecular-weightproteininthe400-to500-kDarangethathadnotbeendetectedpreviouslyusingawidevarietyofexistinggelprepara-tionandblottingprocedures.Thus,webelievethattheconditionsdescribedhereforSDS–polyacrylamidegelelectrophoresis,followedbyblottingtoanitrocellulosemembraneresultinthehighlyef cienttransferofpro-teinfromthegeltothemembrane,arewidelyapplica-bletoproteinsofdiversesizesandareparticularlyusefulfordetectinglow-abundance,high-molecular-weightproteins.

MATERIALSANDMETHODS

Genie(IdeaScienti c,Minneapolis,MN)wettransferblottingapparatuswith10115-cmplateelectrodes.PolyacrylamideGelElectrophoresis

Gelpreparation.Analyticalgels(8018011mm)consistedofeithera5to20%(w/v)gradientresolvinggelcoupledtoa4%(w/v)stackinggel,preparedasde-scribedbyLaemmli(10),or3.3to15%(w/v)gradientgelsusingbuffersdescribedbyFairbanksetal.(13).Sincewewereprimarilyinterestedinhigh-molecular-weightproteins,wecastgelsusinga50:1ratioofacryl-amide:bisacrylamideinallexperiments.Allgelswerestoredatroomtemperatureovernighttoensurecom-pleteandconsistentpolymerization.

Laemmligels.Theresolvingportionofthesegradi-entgelsconsistedof5%(w/v)acrylamide,375mMTris–HCl(pH8.8),0.1%(w/v)SDS,5%(v/v)glycerol,0.15%(w/v)ammoniumpersulfate,0.2%(v/v)TEMEDastheuppergradientmixtureand20%(w/v)acrylamide,375mMTris–HCl(pH8.8),0.1%(w/v)SDS,15%(v/v)glyc-erol,0.06%(w/v)ammoniumpersulfate,0.18%(v/v)TEMEDasthelowergradientmixture.Thestackinggelwascomposedof4%(w/v)acrylamide,125mMTris–HCl(pH6.8),0.1%(w/v)SDS,0.1%(w/v)ammoniumpersulfate,0.12%(v/v)TEMED.Thegelrunningbuffercontained25mMTris,192mMglycine,0.1%(w/v)SDS.Gelswererunat100V,20mAforapproximately3h,untilthedyefrontreachedthebottomofthegel.

Fairbanksgels.TheFairbanksgelbufferswereusedtocastgradientgelsessentiallyasdescribedbyWangetal.(14).Theuppergradientmixtureconsistedof3.3%(w/v)acrylamide,11stockbuffer(below),5%(v/v)glycerol,1%(w/v)SDS,0.15%(w/v)ammoniumpersulfate,0.2%(v/v)TEMED;thelowergradientmix-tureconsistedof15%(w/v)acrylamide,11stockbuffer,15%(v/v)glycerol,1%(w/v)SDS,0.06%(w/v)ammo-niumpersulfate,0.18%(v/v)TEMED.A101stockbufferconsistingof400mMTrisbase,200mMsodiumacetate,and20mMEDTAwaspreparedandadjustedtopH7.4withglacialaceticacid.Gelswererunat55V,21mA,in11stockbufferwith1%(w/v)SDS,forapproximately3huntilthedyefrontreachedthebot-tomofthegel.

Proteinsampleandpreparation.Themixtureofseven14C-labeledproteinsobtainedfromGibcoBRLcontainedlysozyme(14.3kDa),lactoglobulin(18.4kDa),carbonicanhydrase(29kDa),ovalbumin(43kDa),bovineserumalbumin(68kDa),phosphorylaseB(97.4kDa),andmyosinheavychain(200kDa).Analiquotofthe14C-labeledmolecularweightmarkerswasaddedtoanequalvolumeof21samplebuffer(fora nal11concentrationof75mMTris–HCl,pH6.8,3%(w/v)SDS,15%(v/v)glycerol,5%(v/v)b-mercapto-ethanol,0.0025%(w/v)bromphenolblue),incubatedina100 Cwaterbathfor3min,andcentrifugedat

SDSwaspurchasedfromBio-Rad(Hercules,CA).C-labeledmolecularweightmarkerswereobtainedfromGibcoBRL(Frederick,MD).Ammoniumpersul-fate,EDTA(ethylenediaminetetraaceticacid),glycerol,glycine,HCl,sodiumacetate,glacialaceticacid,metha-nol,andTEMED(N,N,N ,N -tetramethylenediamine)werepurchasedfromFisherScienti c(St.Louis,MO).Trizmabase(tris[hydroxymethyl]aminomethane),acryl-amide,andN,N -methylene-bis-acrylamidewerepur-chasedfromSigmaChemicalCo.(St.Louis,MO).The1.2-mmcelluloseacetatemembranewaspurchasedfromM.S.I.(Westborough,MA).ThenitrocelluloseusedwasBA-85fromSchleicherandSchuell(Keene,NH).GelswerepouredinaMightySmallSE245GelCaster(HoeferScienti cInstruments,SanFrancisco,CA)usingagradientformer.Electrophoreticsepara-tionwasachievedusingaMightySmallIISE-250(HoeferScienti cInstruments,SanFrancisco,CA).BlottingprocedureswerecarriedoutusingaMini-14

HIGH-EFFICIENCYBLOTTINGOFPROTEINS

187

12,000gjustpriortoloading.Theamountofproteinstandardandthesamplesizewereconstantforallgels.ProteinBlotting

Blottingbuffercomposition.Eachofthedescribedblottingproceduresutilizedoneofthefollowingbuff-ers:Buffer1(11),25mMTrisbase,192mMglycine,20%methanol,0.1%SDS;Buffer2(12),50mMTrisbase,380mMglycine,20%methanol,0.1%SDS;orBuffer3(BoltandMahoney,thispaper),40mMTris,20mMsodiumacetate,2mMEDTA,pH7.4,20%(v/v)methanol,0.05%(w/v)SDS.

Blottingprocedure.Sincewewereprimarilyinter-estedinhigh-molecular-weightproteins,andhenceusedarelativelylowpercentageofacrylamidefortheupperportionofthegradientgel,additionalstepsweretakentoprotectthefragilegelsfromstickingtothenitrocellulose.Onesheetofcelluloseacetatewasplacedbetweeneachgelandthenitrocellulosewhencon-structingthegel‘‘sandwich’’topreventthelow-per-centagegelintheupperportionsofthegradientfromadheringtothenitrocellulose(describedin14).Expo-sureofthecelluloseacetateonphosphorimagerplatesforextendedtimeperiodsfailedtodetectanylabelboundtothesesheets(datanotshown).Twosheetsofnitrocellulosewereusedforeachblot;thesheetnearestthegelisreferredtoasthe‘‘ rstsheet’’inthetextand gures.

Blottingconditions.Gelstransferredtonitrocellu-loseinthepresenceofBuffer1wereblottedateither10or20V.Gelswereblottedat20VusingBuffer2and10VusingBuffer3.Inallcases,gelswereblottedfor2h,withtheentireminiblottingapparatusim-mersedinicetocooltheunit.

Blotanalysis.Afterblotting,nitrocellulosesheetswererinsedbrie yindeionizedwaterandallowedtoairdryatroomtemperature.Thedriedsheetswereloadedina lmcassettewithatypeBAS-IIIsimagingplate(FujiFilmCo.Ltd.)for7h.TheimagingplateswereprocessedwithaFUJIXBAS-1000Bio-imaginganalyzer,andtheprocessedimageswereanalyzeduti-lizingtheMacBASv1.0softwareonaPowerMacintosh7100.ValuesreportedinTable1wereobtainedbydi-vidingtheradiolabeledcountsdetected,foreachsizestandard,onthe rstsheetofnitrocellulosebythetotalnumberofcountsdetectedonthe rstandsecondsheetsofnitrocellulose.ValuesreportedinTable2wereobtainedbydividingtheradiolabeledcountsde-tected,foreachsizestandard,bythecountsdetectedusingaLaemmligelandblottingBuffer1at20V.

RESULTS

AdrawbackoftheoriginalblottingprotocolofTow-bin(1)isthathigh-molecular-weightproteinsare

transferredverypoorly,aconditionwhichisremediedconsiderablybytheadditionof0.1%SDStotheblottingbuffer(11).Nevertheless,itisstilloftendif culttoachievequantitativetransferofproteinsoveradiverserangeofsizes.Thisisparticularlyproblematicwhenanalyzinglargeproteinswhichmaybeprocessedtosmallerbiologicallyactiveforms.Inanefforttoim-provedetectionofhigh-molecular-weightproteinswithoutsacri cingthetransferandretentionoflow-molecular-weightproteins,wetestedmanydifferentproceduresfortheelectrophoreticseparationandsub-sequentblottingofproteinstomembrane lters.Theresultsdescribedheredemonstratetheef ciencyofproteintransferfromtwotypesofgradientpolyacryl-amidegels,asdescribedbyLaemmli(10)andFair-banksetal.(13,seeRef.14forreview),usingdifferentcombinationsofgelandblottingsystems.Followingtheelectrophoreticseparationofseven14C-labeledproteinstandards,ranginginsizefrom14.3to200kDa,theproteinbandsineachtypeofgelweretransferredtonitrocellulosemembranesusingtheblottingbuffersde-scribedintwopublishedandwidelyusedtransferpro-tocols(11,12)andthenewblottingbufferdescribedinthispaper(seeMaterialsandMethods).Afterblotting,theamountsofeachoftheradiolabeledproteinsboundtothenitrocelluloseandremaininginthegelwerequantitatedwithaphosphorimager.Inallcases,twosheetsofnitrocellulosewereemployedduringtheblot-tingproceduresothatwecoulddeterminehowmuchprotein,ifany,passedthroughthe rstsheetofnitro-celluloseduringblottingandboundtothesecondsheet.Althoughwewereparticularlyinterestedintheef -cienttransferoflarge(ca.200kDaandabove)proteins,ourgoalwastoachievethehighestleveloftransferofallproteinbandsfromthegelmatrixtothe rstnitrocellulosesheet.

TheresultsobtainedusingaLaemmli-typegradientgelandBuffer1(seeMaterialsandMethods),withatransfervoltageof20V,areshowninFig.1A.Anexaminationofthegelfollowingtheblottingprotocolshowedthatessentiallyalltheproteinshadexitedthegelunderthesetransferconditions(datanotshown).Fortheproteinmarkersbetween14.3and29kDa,mostoftheproteinboundtothe rstsheetofnitrocellu-lose.Fortheproteinsbetween43and97.4kDa,asig-ni cantamountofproteinpassedthroughthe rstsheetofnitrocellulose,withasmuchas38–43%ofthetransferredcountsbindingtothesecondsheet(Fig.1A,Table1).Thepassageofproteinsthroughthe rstsheetofnitrocellulosecouldbetakentoindicatethatthebindingcapacityofthatsheethadbeenreachedintheregionsoftheproteinbands.However,decreasingthetransfervoltageto10V(Fig.1C)demonstratedthatthiswasnotthecase.A10-VpotentialdifferenceacrosstheelectrodesusingblottingBuffer1wasnotsuf cientforalltheproteintoblotoutofthegelbytheendof

188

BOLTANDMAHONEY

TABLE1

FractionofRadiolabeledProteinDetectedontheFirstofTwoSheetsofNitrocellulose

Blottingbuffer

Proteinsize(kDa)

Laemmligel

14.318.429.043.068.097.4200.014.318.429.043.068.097.4200.0

Buffer1(20V)0.800.870.970.570.620.621.000.400.580.890.440.530.580.57

Buffer1(10V)0.640.710.930.530.620.690.890.270.430.850.330.460.480.53

Buffer20.850.931.000.520.650.681.000.560.680.980.440.540.601.00

Buffer31.001.001.000.940.980.971.001.001.001.001.001.001.001.00

Fairbanksgel

the2-hblottingperiod(datanotshown),andfewertotal14Ccountscorrespondingtoeachmarkerproteinweretransferredtothe rstnitrocellulosesheet.Nev-ertheless,signi cantamountsofproteinwerealreadypassingthroughthe rstsheetofnitrocelluloseandbindingtothesecondsheet(Fig.1C).Indeed,Table1showsthatforthefourmarkersbetween14.3and43kDa,theproportionofthetotaltransferredcountsthatboundtothe rstmembranesheetduringthe10-Vtransferactuallydecreasedincomparisontoresultsofthe20-Vtransfer.Interestingly,the29-andthe200-kDaproteinmarkersboundbesttothe rstnitrocellu-losesheetateithervoltage,suggestingthatproteinsizeisnotthemajorfactordeterminingwhichproteinspassthroughthe rstnitrocellulosemembraneduringtransfer.

AnotherwidelyusedbufferissimilartoblottingBuffer1,butcontainsahigherconcentrationofelectro-lytes(12).Wetestedthisbuffer(blottingBuffer2),un-derthevoltageconditionswhereessentiallyalloftheproteinsweretransferredoutofaLaemmligel,tode-terminewhethertheincreasedconcentrationofelectro-

TABLE2

RelativeAmountofRadiolabeledProteinDetectedontheFirstofTwoSheetsofNitrocellulose

Blottingbuffer

Proteinsize(kDa)

Laemmligel

14.318.429.043.068.097.4200.014.318.429.043.068.097.4200.0

Buffer1(20V)1.001.001.001.001.001.001.000.260.420.720.400.590.640.67

Buffer1(10V)0.670.660.790.500.550.640.500.100.190.540.270.420.490.43

Buffer20.961.080.901.041.271.090.920.901.051.660.991.652.381.23

Buffer31.401.130.721.961.851.490.381.311.290.983.383.593.801.98

a

Fairbanksgel

a

NormalizedtoamountdetectedusingaLaemmligelandblottingBuffer1at20V.

HIGH-EFFICIENCYBLOTTINGOFPROTEINS

189

FIG.1.Quantitationof14C-labeledproteinsblottedontonitrocelluloseunderdifferentconditions.14C-labeledproteinswereelectrophoresedingradientgelspreparedaccordingtoeitherLaemmliorFairbanks(seeMaterialsandMethods).Afterblotting,theradioactivecountsboundtoasheetofnitrocellulosenearestthegel( rstsheet)orasecondsheetimmediatelybehindthe rstsheet(secondsheet)werequantitatedusingaphosphorimager.Theproteinsbandsquantitatedarelysosyme(14.3kDa),lactoglobulin(18.4kDa),carbonicanhydrase(29kDa),ovalbumin(43kDa),bovineserumalbumin(68kDa),phosphorylaseB(97.4kDa),andmyosinheavychain(200kDa).Errorbarsindicateonestandarddeviation.Thegelsystem,blottingbuffer,andblottingconditionsareasfollows(seealsoMaterialsandMethods):(A,C,E,G)Laemmligelsystem,(B,D,F,H)Fairbanksgelsystem.(A,B)BlottingBuffer1,20-Vtransfervoltage;(C,D)BlottingBuffer1,10-Vtransfervoltage;(E,F)BlottingBuffer2;(G,H)BlottingBuffer3.

190

BOLTANDMAHONEY

lytesaffectedtheef ciencyofproteinblottingtothenitrocellulose.Ourresultsshowthat,atatransfervolt-ageof20V,therewaslittleimprovementgainedinblottingef ciencybyusingBuffer2(Fig.1E)comparedtoBuffer1(Fig.1A).Moreover,bothBuffers1and2resultedinafairlylargeproportionoftheproteinsinthesizerangeof43–97kDapassingthroughthe rstsheetofnitrocelluloseandbindingtothesecondsheet(Fig.1,Table1).Thisresultisinkeepingwiththeresultsofotherswhohavereportedthatmodi cationoftheelectrolyteconcentrationintheblottingbufferhaslittleeffectontransferyields(15).

WhilediscontinuousgelspreparedaccordingtoLaemmli(10)arewidelyusedforseparatingproteins,othergelsystemshavealsobeenusedtoadvantage.Onewhichhasbeenreportedtoworkwellwithhigh-molecular-weightproteins(14)isthesystemdescribedbyFairbanksetal.(13).Sinceourinterestinblottingistheef cienttransferofahigh-molecular-weightpro-tein(ca.500kDa),werepeatedtheaboveexperimentsusingaFairbanksgradientgelandblottingBuffers1and2.BecauseFairbanks-typegelsdonotcontainastackinggel,thepercentagesofacrylamideusedtopourthegradientgelwerechosensothatthesepara-tionoftheproteinbandscloselymimickedthesepara-tionweachievedusingtheLaemmligelsystemofa4%stackinggelanda5to20%resolvinggel.Despitelack-ingastackinggeltofocustheproteins,wedidnotdetectanylossofresolutioninthesegels.Tothecon-trary,theresolutionofproteinstandardswhichcoverasizerangefrom14.3to200kDaisexcellentwhenusingtheFairbanks-typegradientgels.Theresultsob-tainedafterblottingFairbanksgelsusingtheconven-tionalblottingbuffersprovideamarkedcontrasttothosedescribedabovefortheLaemmli-typegels.

Wedetectedadecrease,acrosstheentiresizespec-trumofproteinstested,inthetotalamountofradioac-tiveproteinblottedfromFairbanksgels(Figs.1Band1D)ascomparedtoLaemmli-typegels(Figs.1Aand1C)withBuffer1astheblottingbuffer.Thiswasthecaseatatransfervoltageof10or20V.Moreover,arelativelylargeamountofproteinpassedfromtheFair-banksgelthroughthe rstsheetofnitrocelluloseandboundtothesecondsheet.At10V,theamountofpro-teinboundtothesecondnitrocellulosesheetexceededthatboundtothe rstsheetforallproteinmarkersexceptthoseof29and200kDa(Fig.1B,Table1).Evenat20V,theretentionofproteinonthe rstsheetwasgenerallylessthan60%ofthetotalproteintransferredfromthegel,withtheexceptionofthe29-kDaprotein.Thus,thetransferyieldisworseforproteinsblottedoutofaFairbanksgel,comparedtoaLaemmligel,whenusingBlottingBuffer1.

ThesituationisslightlyimprovedforFairbanksgelswhenblottingBuffer2issubstitutedforBuffer1.Inthisinstance,mostproteinstransferredfromthegelontothe rstsheetofnitrocellulose(Fig.1F)aboutaswellastheydidundersimilarconditionsusingLaem-mligels(Fig.1E).Infact,usingBuffer2atatransfervoltageof20V,thetotalamountofproteinblottedfromFairbanksgelsslightlyexceededthetotalamountofproteinblottedfromLaemmligels.However,asigni -cantproportionofmostproteinmarkerspassedfromtheFairbanksgelthroughthe rstsheetofnitrocellu-loseandadheredtothesecondnitrocellulosesheet(Fig.1F,Table1).

Alloftheresultsaboveindicatedthatsigni cantim-provementsinblottingef ciencycouldbeobtainedifmoreproteinboundtothe rstsheetofnitrocellulose,insteadofpassingthroughit.Underthecommonlyusedconditionswetestedasdescribedabove,relativelylargeamountsofproteinpassthroughthe rstsheetofnitrocellulose,evenwhenthebindingcapacityofthatsheetisnotreached(cf.Figs.1Aand1C,and1Band1D).Ifbindingconditionscouldbealteredsuchthatmoreproteinboundtothe rstsheetofnitrocellulose,theresultshouldbeanincreaseinblottingef ciency.Wethusexperimentedwithdifferentblottingbuffersandfoundthatsigni cantimprovementsinblottingcouldberealizedusingabufferformulation(blottingBuffer3)basedontherunninggelbuffercompositionusedintheFairbanksgels.WhenLaemmligelswereblottedusingBuffer3,asurprisinglylowamountofproteinwasdetectedonthesecondsheetofnitrocellu-lose(Fig.1G,Table1).Withtheexceptionofthe29-and200-kDaproteins,moreproteinblottedfromaLaemmligelontothe rstsheetofnitrocelluloseusingBuffer3thanusingBuffers1or2(Table2).

Thebestoverallresults,however,wereobtainedwhenwetestedblottingBuffer3withtheFairbanksgels.AsshowninFig.1H,highamountsofalltheproteinswereblottedfromFairbanksgelsontothe rstsheetofnitrocellulose.Moreover,acrosstheentiresizerangeofproteinstested,noproteinwasdetectedonthesecond,backingsheetofnitrocellulose.Forthe18.4-kDaproteinandthosemarkersbetween43kDaand200kDa,thetotalamountofproteinonthe rstnitrocellulosesheetwasgreaterthanunderanyothergelsystemandblottingbuffercombinationtested(Fig.1,Table2).Forthe14.3-kDaprotein,Buffer3gavebetterresultsthanBuffers1or2,andtheamountsof14.3-kDaproteinblottedfromLaemmliorFairbanksgelsusingBuffer3wereroughlyequivalent(Table2).Thus,assummarizedinTable2,thecombinationoftheFairbanksgelsystemwithblottingBuffer3appearstoofferaroughlytwo-tofourfoldincreaseintheabilitytoblotmedium-sizedtolarge-molecular-weightproteins(43–200kDa),whileallowingproteinsassmallas14.3and18.4kDatoblotatleastaswellasisobtainablewiththepresentwidelyusedmethods.

UsingtheFairbanksgelandblottingBuffer3combi-nationfollowedbyimmunolabeling,wehavesuccess-

HIGH-EFFICIENCYBLOTTINGOFPROTEINS

191

fullydetectedaca.500-kDaproteinwhichwaspre-viouslyundetectedusingawidevarietyofgelandblottingcombinations(datanotshown).Sincevirtuallynoproteinappearsonthesecondmembranewhenus-ingtheFairbanksgelsincombinationwithblottingBuffer3(Fig.1H),itappearsthattheincreaseintheblottingef ciencyoftheproteinsresultsfromanin-creaseintheabilityofthenitrocellulosemembranenearestthegeltobindandretainproteinsinthepres-enceofBuffer3,comparedtoBuffers1or2.Thus,thisproceduremaybemostusefulforothersattemptingtodetectmiddle-tohigh-molecular-weightproteinsoflowabundance.

DISCUSSION

Wehavecomparedseveralblottingprotocolsthatareinwidespreadusefortheirabilitytoquantitativelytransferproteinsfromapolyacrylamidegradientgeltoanitrocellulosesheet.Ourtrialsweredrivenbytheneedtodetectahigh-molecular-weightproteinthatwewereunabletodetectusinganyoftheconventionalmethods.Inthecourseofourexperiments,weidenti- edacombinationofgelsystemandtransferbuffercompositionwhichallowedustodetect,withantibod-ies,rare,high-molecular-weightproteinswhichwereundetectablebyanyothermethodstested(datanotshown).Here,wedocumentthattheuseofaFairbanksgelsystem(13),andanovelblottingbufferallowsef http://ingtheseconditions,signi cantlygreateramountsofproteinboundtoonesheetofnitro-cellulose,andextremelylow(orno)amountsofproteinappeartopassthroughthenitrocellulosetoasecondbackingsheet.Theentireprocedureincludingblottingtakesnolongerthanthetimeittakestoprepare,run,andblotaLaemmlitypegel,withtheaddedadvantagethattheFairbankstypegelsoffersequalresolutionwithouttheneedtopouraseparatestackinggel.

Thebindingofproteinstonitrocelluloseisnotpartic-ularlywellunderstood.Whileithasgenerallybeenas-cribedtohydrophobiceffects(8),littleisknownabouthowtheactualbindingofproteinstonitrocelluloseisin uencedbysuchvariablesaspH,ionconcentration,detergentconcentration,etc.Recently,Bhavsaretal.(16)demonstratedthatbindingofaproteinbandtonitrocellulosecouldbeenhancedbythepresenceofad-ditionalproteinsinthesameregionofthegelasthebandofinterest.Whilethis‘‘carriereffect’’ispoorlyunderstood,itillustratesthedegreeofuncertaintysur-roundingtheactualmolecularinteractionswhichoccurbetweenproteinsandnitrocellulose.Thepassageofproteinsthroughthe rstsheetofnitrocelluloseandontoasecondbackingsheetofnitrocellulose,asisob-servedinthetypicalLaemmli/Buffer1blot(Fig.1A),suggeststhatthebindingcapacityofthe rstsheetof

nitrocelluloseisexceededintheregionsoftheproteinbands.However,ourresultsshowthatthe rstsheetofnitrocelluloseisquiteabletobindadditionalprotein,inthepresenceofadifferentbuffer(Figure1H).Thus,whilethepassageofproteinsthroughasheetofnitro-celluloseisoftentakentoindicatethatthenitrocellu-loseissaturatedintheregionoftheproteinband(s),theterm‘‘saturated’’isarelativeone,dependinginpartonthecompositionofthetransferbuffer.Wesus-pectthatLaemmligelsdonotblotquiteaswellinthepresenceofBuffer3(Fig.1G),comparedtoFairbanksgels(Fig.1H),duetotheeffectsofthebufferthatiscontainedwithintheLaemmligelitself.

Incomparingtheblottingef cienciesofvarioussys-tems,caremustbetakenincomparingtheamountofradioactivityremaininginthegel,asopposedtotheamountofradioactivityblottedontothenitrocellulose.Inourexperiments,weobservedthatthegelmatrixquenchedthe14Csignaltoasubstantialdegree(datanotshown).Thiscanresultinoverlyoptimisticestima-tionsofthepercentageoftotalproteinwhichissuccess-fullyblotted,particularlyifattemptsaremadetocom-paretheamountofproteinblottedontonitrocellulosetoa‘‘startingamount’’presentinacontrollanethatwassubjectedtoelectrophoresisbutthen xedanddried.Aratioofblottedproteintoproteinintheunblot-tedgellanewilloverestimatetheamountofproteinblotted,duetothequenchingoftheradioactivityinthedriedgel.Thiscanbeobservedifgelstripsareremovedatvariousperiodsduringtheelectroblottingprocedure,dried,andquantitated.Agelslicewhichispartiallyblottedandthendrieddownwilloftenappeartohavehighercountsthanonewhichisnotblottedatallbeforedrying(datanotshown).Weinterpretthistomeanthatintheinitialstagesofblotting,thebulkofapro-teinbandmovesfromwithinthegelmatrix,tonearerthesurfaceofthegel.Thisresultsintheapparentpara-doxofapartiallyblottedgeldisplayinghigherradioac-tivitycountsthanacontrolunblottedgel,sincetheradioactiveproteinsarelesssubjecttothequenchingactionofthematrix.Withadditionalblottingtime,theradioactivitywithinthegeldrops,astheproteinseven-tuallyexitthegel.ThisinterpretationisconsistentwiththeresultsofSymingtonetal.(17)andEricksonetal.(11),whofoundthattransferof35S-or14C-labeledproteinsfromageltoa lterincreasedtheamountofsignalthatcouldbedetectedbyautoradiography.Becauseofthis,wehavenotreportedourresultsintermsofpercentageoftotalprotein(ormoreaccurately,radioactivecounts)inthegelthatisblotted,butonlyintermsofamountofradioactivitypresentinthe rstorsecondsheetofnitrocellulose.

TheconcentrationofSDSinBuffer3iscritical.IntheabsenceofSDS,largeproteinstypicallytransferpoorlyatbest,irrespectiveofthetransferbuffer.Hence,forthetransferoflargeproteins,0.1%SDS(w/

192

BOLTANDMAHONEY

v)isusuallyaddedtothetransferbuffer.WehavefoundthatBuffer3worksmosteffectivelywhenonly0.05%(w/v)SDSisused.Indeed,addingSDSto0.1%(w/v)markedlydecreasedtheamountofproteinsre-tainedonthenitrocellulose(datanotshown).

Theprocessingofalargeproteintosmallerfrag-mentsposesachallengeintermsofdetectingalloftherelevantfragmentsonthesameWesternblot.Ifconditionsareoptimizedtovisualizethehighestmolec-ularweightproteinband,thensmallerfragmentsarefrequentlylostviapassagethroughthenitrocellulosesheet.Wangetal.(14)observedthatifconditionswereoptimizedforblottingproteinsú200kDa,thenpro-teinssmallerthan40kDawererarelyretainedtoasigni cantextentbyevenseverallayersofnitrocellu-lose.Moreover,sizeisnotthesolefactorindeterminingthetransferef ciencyofaparticularprotein.Aswenotedabove,certainproteinsthatwetestedblottedwitharelativelyhighef ciencyundermanyofthecon-ditionstested.Clearly,individualproteincomposition,inadditiontosize,affectstransferand/orbindingef -ciency.However,wefoundournovelconditionseffec-tivefortransferofalltheproteinstested,ranginginsizefrom14to200kDa.Inthisrespect,theprocedurewedescribehereoffersanenhancedabilitytotransferhigh-molecular-weightproteins,whileretainingsig-ni canttransferofthelowermolecularweightpro-teins.Assuch,itshouldproveparticularlyusefulwhereitisdesirabletovisualizeproteinbandsofdi-versesizesonthesameblot.

ACKNOWLEDGMENT

ThisworkwassupportedbyUSPHSGrant5R29GM50568toP.A.M.

REFERENCES

1.Towbin,H.,Staehelin,T.,andGordon,J.(1979)http://A76,4350–4354.2.Burnette,W.N.(1981)Anal.Biochem.112,195–203.

3.Lin,W.,andKasamatsu,H.(1983)Anal.Biochem.128,302–311.4.Too,C.K.L.,Murphy,P.R.,andCroll,R.P.(1994)Anal.Bio-chem.219,341–348.5.Renart,J.,Reiser,J.,andStark,G.R.(1979)http://A76,3116–3120.6.Elkon,K.B.,Jankowski,P.W.,andChu,J-L.(1984)Anal.Bio-chem.140,208–213.7.Gibson,W.(1981)Anal.Biochem.118,1–3.

8.Gershoni,J.M.,andPalade,G.E.(1983)Anal.Biochem.131,1–15.9.Beisiegel,U.(1986)Electrophoresis7,1–http://emmli,U.K.(1970)Nature227,680–685.

11.Erickson,P.F.,Minier,L.N.,andLasher,R.S.(1982)J.Immu-nol.Methods51,241–249.12.Harlow,E.,andLane,D.(1988)Antibodies:ALabManual.Cold

SpringHarborLaboratoryPress,ColdSpringHarbor,NewYork.13.Fairbanks,G.,Steck,T.L.,andWallach,D.F.H.(1971)Bio-chemistry10,2606–2617.14.Wang,K.,Fanger,B.O.,Guyer,C.A.,andStaros,J.V.(1989)

MethodsEnzymol.172,687–696.15.Gershoni,J.M.,andPalade,G.E.(1982)Anal.Biochem.124,

396.16.Bhavsar,J.H.,Remmler,J.,andLobel,P.(1994)Anal.Biochem.

221,234–242.17.Symington,J.,Green,M.,andBrackmann,K.(1981)Proc.Natl.

http://A78,177–181.