蛋白质糖基化

来源：暴趣科技网

PROTOCOL

Analysisofproteinglycosylationbymassspectrometry

WillyMorelle&Jean-ClaudeMichalski

´MixtedeRechercheCNRS/USTL8576,GlycobiologieStructuraleetFonctionnelle,IFR147,BaˆtimentC9,Universite´desSciencesetTechnologiesdeLille1,Unite

59655Villeneuved’AscqCedex,France.CorrespondenceshouldbeaddressedtoW.M.(willy.morelle@univ-lille1.fr).Publishedonline21June2007;doi:10.1038/nprot.2007.227

©2007 Nature Publishing Group http://www.nature.com/natureprotocolsWepresentadetailedprotocolforthestructuralanalysisofprotein-linkedglycans.Inthisapproach,appropriateforglycomicsstudies,N-linkedglycansarereleasedusingpeptideN-glycosidaseFandO-linkedglycansarereleasedbyreductivealkalineb-elimination.Usingstrategiesbasedonmassspectrometry(matrix-assistedlaserdesorption/ionization–timeofﬂightmass

spectrometryandnano-electrosprayionizationmassspectrometry/massspectrometry(nano-ESI-MS-MS)),chemicalderivatization,sequentialexoglycosidasedigestionsandlinkageanalysis,thestructuresoftheN-and/orO-glycansaredeﬁned.Thisapproachcanbeusedtostudytheglycosylationofisolatedcomplexglycoproteinsorofnumerousglycoproteinsencounteredinacomplexbiologicalmedium(cells,tissuesandphysiologicalﬂuids).

INTRODUCTION

Glycosylationisacommonpost-translationalmodiﬁcation.Therearetwomaintypesofproteinglycosylation:N-glycosylation,inwhichtheglycanisattachedtoanAsnresiduepresentinthetripeptideconsensussequonAsn-X-Ser/Thr(whereXcanbeanyaminoacidexceptPro),andO-glycosylation,inwhichtheglycanisattachedtoaSerorThrresidue.Glycansareimplicatedinawiderangeofintracellular,cell–cellandcell–matrixrecognitioneventsandarethereforeofgreatbiologicalinterest1,2.Tocorrelatefunc-tionalfeatureswithdeﬁnedstructuralparameters,detailedstruc-turalanalysesofglycanchainsarerequired.Themostcommonmannerofcharacterizingproteinglycosylationinvolvesthefollow-ingsteps:ﬁrst,anenzymaticorchemicalreleaseoftheattachedglycans;second,derivatizationofthereleasedglycansviareductiveaminationwitharomaticoraliphaticaminesorpermethylation;third,analysisoftheglycans.Thecompletestructuralelucidationofglycansrequiresthedeterminationofthesugarcomposition,sugarsequence,monosaccharidebranching,interglycosidiclinkagesandanomericconﬁguration.Inaddition,glycoproteinsmaycarryseveraldifferentglycans,oftenamixtureofN-andO-linkedglycans,andtheseglycansmayoccurinvariableamountsatglycosylationsites,withvaryingdegreesofsiteoccupancy.Asaconsequence,glycananalysisrequiresthesequentialemploymentofseveralanalyticaltechniques.

Here,wepresenttheapproachweuseinourlaboratorytostudytheN-andO-glycanchainsfromnative,recombinantglycopro-teinsormorecomplexbiologicalsamplessuchasunfractionatedcellularextractsorphysiologicalﬂuids.Thisapproachismostappropriateforglycomicsstudies.N-linkedglycansarereleasedusingpeptideN-glycosidaseF(PNGaseF)andO-linkedglycansarereleasedfromtheN-deglycosylatedproteins/peptidesbyalkalineb-elimination.AportionofeachpoolofisolatedN-and/orO-glycansisderivatizedbypermethylationandthemethylatedderivativesareanalyzedusingmatrix-assistedlaserdesorption/ionization–timeofﬂightmassspectrometry(MALDI-TOF-MS),bynano-electrosprayionizationmassspectrometry/massspectro-metry(nano-ESI-MS-MS)andbylinkageanalysisbeforeandaftersequentialexoglycosidasedigestions.Partiallymethylatedalditolacetates(PMAAs)areobtainedfromthepermethylatedsamplesforlinkageanalysisbygaschromatography(GC)-MS3.Structuralassignmentsarebasedonmolecularweight,fragmentionforma-

tion(thelatterderivedfromtandemMS(MS-MS)experiments)4,susceptibilitytoexoglycosidasedigestionsandlinkagedata3(Fig.1).Thisrobustandhighlysensitivemassspectrometricapproachforcharacterizingtheglycosylationpatternofproteins,cells,tissuesandphysiologicalﬂuidsallowsarapidcharacterizationoftheglycoformsaswellastheirrelativequantitation5–10.Releaseofglycans

Thedifferencesinprotein–sugarlinkagechemistryofN-andO-glycansrequiredifferentapproachesforthereleaseofintactoligosaccharides.SeveralenzymesareavailableforreleasingN-glycans.ThemostpopularisPNGaseF11.PNGaseFcleavesofftheintactglycanasglycosylamine,whichisreadilyconvertedtoregularglycan.Withfewexceptions,PNGaseFreleasespracticallyallprotein-boundN-linkedcarbohydratesexceptthosewithfucoseattachedtothethirdpositionoftheAsn-linkedGlcNAcresidue12.Theircorrespondingglycoproteinsarecommonlyfoundinplantsandinnematodes.SuchPNGaseF–resistantglycanshavebeenfoundtobesensitivetoPNGaseA,anenzymefoundinalmondemulsin.AfterreleasingN-glycansusingPNGaseF,weuseachemicalcleavagemethodforthereleaseofO-glycans.GlycanslinkedO-glycosidicallytoSerandThrresiduesarereleasedfromtheglycoproteins/glycopeptidesinthereducedformcontainingGalNAc-olbyalkalineb-eliminationinthepresenceofhighconcentrationsofsodiumborohydride(NaBH4),whichprevent‘peeling’ofthereleasedoligosaccharidesbyreducingeachterminalGalNAcresiduetoitsalditol13(Fig.2).

WeuseseveralmethodstoprepareN-glycansampleswithPNGaseF(Fig.1).Theselectionofthesemethodsdependsontheamountofglycoproteinavailable.Iflessthan50mgofproteinisavailable(Step1A),theproteinshouldﬁrstbedenaturedbyheatinginthepresenceofSDStoincreasetheaccessibilityofPNGaseFtothevariousglycosylationsites.TopreventthedenaturationofPNGaseF,anon-ionicdetergent(nonidetP40,NP40)isaddedbeforedigestionwiththedeglycosylationenzyme.AfterPNGaseFdigestion,theN-deglycosylatedproteinisprecipitatedandtheN-glycansarepuriﬁedusingaSep-PakC18cartridgetoremovethedetergents,whichmayconsiderablyaffectthequalityoftheMSspectra.Withlessthan50mgofprotein,sequentialMALDIproﬁling,MS-MSsequencing,linkageanalysisandexoglycosidase

NATUREPROTOCOLS|VOL.2NO.7|2007|1585

PROTOCOL

digestionsontheMALDItargetcannor-mallybeperformed.Ifmorethan50mgofproteinisavailable(Step1B),weuseanothermethod,whichdoesnotrequirethepresenceofSDSforthedenaturationofproteins.However,thethroughputofthisprotocolislower.Inthismethod,tofacilitatethereleaseofN-glycanswith-outresortingtodetergentdenaturation,glycoproteinsareﬁrstreducedandcarbox-amidomethylated.Reducedandcarbox-amidomethylatedglycoproteinsarethensubmittedtoaproteolyticdigestiontogeneratesmallpeptidesandglycopeptidesbeforePNGaseFdigestion.PNGaseF–releasedglycansarethenseparatedfrompeptidesandO-glycopeptidesusingaSep-PakC18cartridge.Step1Bisparticularlyappropriatefortheanalysisofglycansreleasedfromcellsandtissues.Inthiscase,approximately10millioncellsandseveralmilligramsoftissuesarenecessarytoper-formMALDIproﬁling,MS-MSsequencingandlinkageanalysis.Ifglycoproteinsaregelseparated,thenchooseStep1C14,15.Inthiscase,MALDIproﬁlingandMS-MSsequen-cingcanbeperformed.ExoglycosidasedigestionsneedtobecarriedoutontheMALDItarget(Step18).

O-glycansarereleasedfromtheN-degly-cosylatedprotein(Step1A)orO-glyco-peptides(Step1B)byreductiveeliminationanddesaltedthroughaDowex50Â8(H+form).Thismethodhasworkedwellinourhands.Othermethods,includinggelreduc-tiveb-elimination,havebeendescribed16–19.AfterPNGaseFdigestion,analiquot(typically50%)ofreleasedglycansissubjectedtosequentialexoglycosidasedigestions(Step18)monitoredbyMALDI-TOF-MSandlinkageanalysis.Aftereachexoglycosidasedigestion,analiquotisper-methylatedandanalyzedbyMALDI-TOF-MSandlinkageanalysis.Anotheraliquot(typically20%)ofreleasedglycansisper-methylateddirectly(Step19)andthemethylatedderivativesarecharacterizedbyMALDI-TOF-MS,nano-ESI-MS-MSandlinkageanalysis.IfPNGaseF–releasedgly-cansareisolatedfromgel-separatedpro-teins,exoglycosidasedigestionsarecarriedoutontheMALDItarget5,9,10,20.

Isolated glycoprotein/mixture of glycoproteins(1) SDS denaturation(2) PNGase F digestion(3) Precipitation of proteins(4) CentrifugationReleased N-glycansSep-Pak C18to remove SDSPurified N-glycansN-deglycosylated protein(s)(1) Reductive elimination(2) Dowex 50 × 8Released O-glycans(1) Permethylation(2) Sep Pak C18Permethylated O-glycansMALDI-TOF-MSnano-ESI-MS-MSGC-MS(1) Permethylation(2) Sep Pak C18(1) Exoglycosidase digestions(2) Permethylation(3) Sep Pak C18Permethylated N-glycansMALDI-TOF-MSnano-ESI-MS-MSGC-MS©2007 Nature Publishing Group http://www.nature.com/natureprotocolsb

Isolated glycoprotein/mixture of glycoproteins(1) Reduction/alkylation(2) Protease digestionPeptides/glycopeptides mixture(1) Permethylation(2) Sep Pak C18(1) PNGase F digest(2) Sep Pak C18Released N-glycans(1) Exoglycosidase digestions(2) Permethylation(3) Sep Pak C18Permethylated N-glycansMALDI-TOF-MSnano-ESI-MS-MSGC-MSPeptides/O-glycopeptides(1) Reductive elimination(2) Dowex 50 × 8Released O-glycans(1) Permethylation(2) Sep Pak C18Permethylated O-glycansMALDI-TOF-MSnano-ESI-MS-MSGC-MSc

Gel separated glycoprotein(1) Decoloration(2) Reduction/alkylation(3) PNGase F digestion(4) N-glycans extractionExtracted N-glycans(1) Chemical desialylation (2) Nonporous graphitized mini columnPurified N-glycansMALDI-TOF-MSFigure1|StrategyforthedeterminationofthestructuresoftheN-andO-linkedglycansof

glycoproteinsusingmassspectrometry.Thestrategiesinglycoproteinanalysisdependontheamountofavailablesample.SeveralmethodsforpreparingN-glycansaredescribed.Iflessthan50mgofproteinisavailable,choosestepsinpanela.AfterpeptideN-glycosidaseF(PNGaseF)digestion,ifyouarenotinterestedstudyingtheO-glycansoftheproteins,youcanproceeddirectlytothepuriﬁcationoftheN-glycansusinganon-porousgraphitizedcarboncolumn(Step2A).Ifmorethan50mgofproteinisavailable,weuseanothermethod(stepsinpanelb),whichdoesnotrequirethepresenceofSDSforthedenaturationofproteins.Forgel-separatedglycoproteins,choosestepsinpanelc.

Massspectrometryandpermethylation

Electrosprayionization(ESI)andMALDIMSplayacrucialroleinthecharacterizationofglycosylation21,22.MALDI-TOF-MSoffersasimplemeansforscreeningcomplexmixtures,andMS-MSmethodsusingelectrospraywithlow-energycollision–induceddissociation(CID)arevaluabletoolsfortheproduction

1586|VOL.2NO.7|2007|NATUREPROTOCOLS

offragmentionstoinvestigatethesequenceandevenlinkageofglycans.

MALDI-TOF-MSisveryoftenusedasaﬁrststepbecauseitisuniqueinitscapacitytogeneraterapidlyinformationaboutthenatureanddiversityofglycansreleasedfromnative,recombinantglycoproteinsorevenmorecomplexbiologicalsamples.Neutralglycansyieldintensesignalsinthepositiveionmodecorrespondingtosodium-cationizedmolecularspecies[M+Na]+.Thisionis

oftenaccompaniedbyaweaker[M+K]+ion.SialylatedglycansareusuallymoredifﬁculttoanalyzeusingMALDI-TOF-MSandgiveamixtureofionssuchas[M+Na]+,[M+K]+,[M-nH+(n+1)Na]+and[M-nH+(n+1)K]+.Furthermore,sialylatedglycanscanveryeasilyloseasigniﬁcantamountofsialicacidintheionsourceoraftertheionextractionfromtheionsource,considerablydistortingtheglycanproﬁles.Toreducethisloss,sialylatedglycanscanbeanalyzedwithTOFinstrumentsinthelinearnegativeionmode.Neutralglycansarenot,how-ever,detectedinthenegativeionmode.Inaddition,thefragmentionsformedwithintheionsourcearestillobservedintheMALDI-TOF-MSspectrum.Severalmethodsofderivatizationtosloimprovethestabilityofsialicacidresiduesbyneutralizingtheircotnegativechargearecurrentlyinuse.Methylesteriﬁcationoforpthecarboxylgroupofsialicacidresiduesbyreactionoftheirerusodiumsaltswithmethyliodideallowssimultaneousanalysisoftaneutralandsialylatedoligosaccharidesinthepositiveionmode23.n/mPermethylationofglycansusingthemethoddescribedbyocCiucanuandKerekalsostabilizesthesialicacidresidues24.Several.erurecentmodiﬁcationstothismethodhavesubstantiallyimprovedtanitsefﬁciency25,26.

w.Inourlab,wehavechosentostabilizesialylatedoligosaccharideswwusingpermethylation,whichisbyfarthemostimportanttypeof//:pderivatizationusedinglycanMS.Despitethedrawbackofinvol-tthvinganadditionalwetchemistrystep,permethylationderivatiza- ptionoffersseveraladvantages:(i)sampleclean-upusinglipophilicuextractionmakesitpossibletoremovesaltsveryeasily;(ii)thisroG derivatizationsigniﬁcantlyimprovesthesensitivityofdetectionofgnmolecularions;(iii)itallowssimultaneousanalysisofneutralandihssialylatedoligosaccharidesinthepositiveionmode;(iv)thisilbapproachalsoleadstopredictablefragmentationsthatgiveuPcharacteristic‘maps’offragmentionsateachaminosugarresidue; er(v)furthermore,themethylationapproachcanbeusedfortheutaelucidationofallinterglycosidiclinkagessinceGC/MSanalysisofN chemicallymodiﬁedmonosaccharidesthatarederivedfrom70methylation,hydrolysis,reductionandacetylationofglycanscan02beperformedafterMALDI-TOF-MSanalysisofpermethylated©glycans.

Sinceglycanmoietiesarecomposedofarelativelysmallnumberofdifferentmonosaccharideconstituentswithuniqueincrementalmasses,proﬁlingofpermethylatedglycanstructuresusingMALDI-TOF-MSallowsusrapidlytoobtaininformationaboutthenatureanddiversityofstructuresofglycanspresent.Inaddition,asthesignalstrengthofpermethylatedglycansappearstoreﬂectaccu-ratelytheamountofmaterialonthetarget,weareabletodeterminetherelativequantitiesofeachneutralandsialylatedglycanpresentinamixturebyMALDI-TOF-MSpermethylatedglycanproﬁling.Themassofamolecularioniscalculatedbyaddingthemassofamethylgroupcorrespondingtothenon-reducingendofthemoleculeandthemassofthegroupcorres-pondingtothereducingendofthemoleculetothesumoftheincrementmassesofthesugarresiduesconstitutingthemolecularion.Fornon-reducedglycans(PNGaseF–releasedN-glycans),thegroupcorrespondingtothereducingendofthemoleculeisamethyl-glycoside(CH3O)withasodiumtogivethecharge.Forreducedglycans(O-glycansreleasedusingreductiveelimination),thegroupcorrespondingtothereducingendofthemoleculehastheformulaC2H7Owithasodiumtogivethecharge.TheseareillustratedinFigure3.Table1givestheaccurateandaveragevalues

PROTOCOL

R1NHCH2OHOOCOOH2CCHOH–

ONHNHCOCHR3CONaOH 50 mM NaBH4 1 M45 °C overnight

R1NHCH2OHOOHCO–OCCOHNHNHCOCHR3COR1CH2OHOOOHONHCOCHR3HOCCH2OH

HCNHCOCH3HCNHCOCH3ROCHNaBH4ROCHHOCHHOCHAlditol

HCOHHC

CH2OR1

Figure2|ReactionstepsforreductiveeliminationofO-linkedglycans.R,3-linkedsaccharide;R1,6-linkedsaccharide.

necessarytointerpretMALDI-TOF-MSspectraofpermethylatedN-andO-glycans.

ThepermethylatedcomponentsobservedintheMALDI-TOF-MSspectrumcanbefurtheranalyzedusinganano-electrosprayquadrupoleTOF(Q-TOF)-MStoassistsequenceassignment27,28.Thisinstrumentisusedtoobtainfragmentationspectrainashorttimewithhighmassaccuracythatrevealdetailedstructuralandinsomecaseslinkageinformationoncomplexoligosaccharides.Fragmentationofcarbohydratesisdependentonseveralfactors,suchasthetypeofionformation,itschargestateandtheenergydepositedintotheion29,30.Twomajortypesofionsareusuallyobserved:glycosidiccleavageswherebondruptureoccursbetween

NATUREPROTOCOLS|VOL.2NO.7|2007|1587

PROTOCOL

thesugarringsandinvolvesahydrogenmigration,andcross-ringcleavagesthatinvolvetheruptureoftwobondsonthesamesugarresidue.ThenomenclaturegenerallyusedfordescribingthesefragmentionsisthatproposedbyDomonandCostello31.Inthisnomenclature,ionsretainingthechargeonthenon-reducingterminusarenamedA,BandC,whereastheionsretainingchargeonthereducingterminusareX,YandZ(Fig.4).AandXcorrespondtocross-ringcleavages,whereasB,C,YandZcorres-pondtoglycosidiccleavages.Subscriptnumbersdenotethecleavageposition,startingatthereducingterminusfortheX,YandZions,andatthenon-reducingterminusfortheothers.Inthecaseofringcleavages,superscriptnumbersaregiventoshowthecleavedbonds.Permethylatedglycansusuallyyieldintensesodium-cationizedandprotonatedmolecularspecieswhenanalyzedintheESImode.Themainfragmentationprocessofsodium-cationizedspeciesinvolvesthecleavageofglycosidicbondsandgivesrisetopertinentdataonsequenceandbranch-ing4,27,28.Underlow-energyCIDconditions,protonatedspecies,signiﬁcantlymorelabilethanthecorrespondingsodiumadducts,yieldverysimplefragmentationpatterns,allowingimmediateandunambiguousassignmentofsequenceandbranchinginfor-mationandeveninformationonsomeinterglycosidiclinkages28,32.FragmentationofthepermethylatedglycansgavepredominantlyB-typeglycosidicfragments.Y-typeglycosidicfragmentsarealsoobserved(Fig.5).B-typeionsresultprimarilyfromcleavageoftheglycosidicbondattheN-acetylhexosaminemoietywithlocalizationofthechargeonthenon-reducingend.Thesefragmentionsareextremelyhelpfulforantennaesequenceassign-ments.Themassofthesefragmentsiscalculatedbyaddingthenon-reducingendincrement(Table1)withthesodiumminushydrogentothesumoftheresiduemasses.ThemostcommonlyencounteredB-typeionsfromsodiatedpermethylatedglycansarelistedinTable2.

Exoglycosidasedigestions

Todeﬁnethemonosaccharidesandtheiranomericconﬁgurations,andtoconﬁrmtentativesequences,additionalexperimentsarerequired,andglycanscanbetreatedwithexoglycosidases.Exogly-cosidasesremovemonosaccharidesfromthenon-reducingterminiofglycanchains.Theseenzymesarespeciﬁctothestereochemistry,

Nonreduced permethylated glycan

CH2OMeOOMeOMeOMeOCH2OMeOOMeOOMe204

CH2OMeOOMeOMe204

Reducing end

OMe15204

Nonreducing end

Reduced permethylated glycan CH2OMeOOMeOMe15Nonreducing endOMe204O31

CH2OMeOOMeOOMe204CH2OMeCHOMeCHOMeCHCHOMeCH2OMe204 + 1 + 15

theanomericconﬁgurationofthemonosaccharidebeingreleasedanditslinkagesitewithrespecttotheremainderoftheglycanchain33.Anon-exhaustivelistofcommonlyusedexoglycosidasesandtheconditionstoperformthedigestsisgiveninTable3.Itisrecommendedthathighlyspeciﬁcexoglycosidasesareused.PNGaseF–releasedN-glycansareusuallysubjectedtotreatmentwitha-sialidase,followedbyb-galactosidase,b-N-acetylhexosami-nidase,a-fucosidaseanda-mannosidasedigestion.Afterenzymatictreatment,themixtureisanalyzedbyMALDI-TOF-MStodeterminethenumberofmonosaccharidesreleasedbytheenzyme.Usingseveralwell-deﬁnedenzymesinsequence,themolecularmassinformationaftereachdigestionstepthusrevealsthesequenceofthemonosaccharideconstituentsoftheglycanchains20,34,35.Thesetreatmentscanbeapplicablebothtomixturesofglycansandtoaspeciﬁcglycan.Ifthetreatmentisperformedonamixtureofglycans,therelativeabundanceofeachionobtainedontheMALDI-TOF-MSspectrumshouldbetakenintoaccountininterpretingtheexoglycosidaseresults.

Exoglycosidasedigestionsareperformedinbufferedsolutions;consequently,asampleclean-upprocedureisrequiredbeforethe

TABLE1|Residuemassesofcommonmonosaccharidesandterminalgroups.

Native

Accuratemass

MonosaccharidesPentoseDeoxyhexoseHexose

N-acetylhexosamineN-acetyl-neuraminicacidN-glycolyl-neuraminicacid

TerminalgroupsNon-reducingendReducingend

Reducedreducingend

SumofmasseswiththesodiumforN-glycans

SumofmasseswiththesodiumforreducedO-glycans

1588|VOL.2NO.7|2007|NATUREPROTOCOLS

Permethylated

Averagemass132.1161146.1430162.1424203.1950291.2579307.25731.007917.007319.023241.014143.0300

Accuratemass160.0736174.02204.0998245.1263361.1737391.184215.023531.018447.049769.031685.0629

Averagemass160.1699174.1968204.2230245.2756361.3923391.418615.034831.034247.077069.067985.1107

132.0432146.0579162.0528203.0794291.0954307.09031.007817.002719.018441.000243.0159

PROTOCOL

1,51,5MALDI-TOF-MSanalysis.AliquotsareFigure4|Nomen-Y2Y0Z0X0X1Y1Z2takenaftereachdigestion,permethylatedclaturefordescribingZ1thefragmentationofCHOHCH2OHandexaminedbyMALDI-TOF-MSand2CH2OHcarbohydrates31.InthisOOlinkageanalysisafterreverse-phaseSep-Onomenclature,ionsOROHOHOOOHPakC18puriﬁcation.Sep-PakC18isaretainingthechargeon

HOsimple,rapidmethodforseparatingper-thenon-reducing

OHOHOHmethylatedglycansfromtheremainingterminusarenamedA,

0,2A1salts.ComparisonoflinkagedatabeforeBandC,andtheions

B1B32,5C3B2A3andafterexoglycosidasetreatmentmakesretainingchargeonC12,4C2A2

itpossibletoestablishwheretheremovedthereducingterminus

areX,YandZ.AandX

monosaccharideswereattached.Iflinkage

correspondtocross-ringcleavages,whereasB,C,YandZcorrespondtoglycosidiccleavages.Subscript

analysisisnotnecessary,analiquottakennumbersdenotethecleavageposition,startingatthereducingterminusfortheX,YandZions,andataftereachdigestioncanbepuriﬁedusingathenon-reducingterminusfortheothers.Inthecaseofringcleavages,superscriptnumbersaregivenminiaturizedporousgraphitizedcarbontoshowthecleavedbonds.solid-phaseextractionmethodandana-AfterMALDI-TOF-MSanalysis,permethylatedglycansarelyzedwithoutderivatizationbyMALDI-TOF-MSafterﬁrstremov-acid-hydrolyzed,deuteroreduced,andperacetylated.Theresultingingsialicacids.

PMAAsareanalyzedbyelectronimpact–GC-MS(EI-GC-MS)36–38.TheretentiontimeonthecapillaryGCcolumnandthefragmentLinkageanalysis

Eveniftheoligosaccharidesareanalyzedasmixtures,datafromionsobservedintheEI-MSspectraareusedtoidentifythedifferentlinkageanalysisallowimportantstructuralconclusionstobedrawntypesofresidues.ThepositionoftheO-acetylandO-methylandcanbeusedtoconﬁrmMALDI-TOF-MSandnano-electro-groupsonthePMAAreﬂectstheringformandlinkagepositionssprayQ-TOF-MSdata.Forexample,linkagedatacanyieldinfor-oftheglycosylresidueintheoriginalglycan.Theretentiontimes

andthecharacteristicfragmentionsofseveralPMAAsaregiveninmationon(i)thedifferentclassesofN-glycans(highmannose

Table4.Althoughretentiontimesarespeciﬁctoauser’slab,thestructures,hybrid-typestructuresandcomplex-typestructures)

(Fig.6);(ii)thedifferenttypesofnon-reducingsugars;(iii)theorderofelutionofthecompoundswillbeidenticalinanotherabundanceofbiantennary,triantennaryand/ortetraantennarylaboratoryprovidedthatsimilarcolumnsareused.However,(Fig.6)complex-typeglycansintheN-glycanmixturesbycompar-usersneedtoprepareacollectionofPMAAstandardstoestablishinglevelsof2-linkedManwith2,4-linkedManand2,6-linkedMan;theirownlistsofretentiontimes.PMAAsarecharacterizedby(iv)theabundanceofbisectedstructureswiththe3,4,6-linkedMansimplefragmentationpatterns.Them/zoftheionsidentiﬁesthelevel;(v)thedifferenttypesofantennaeincomplex-typeglycans;numberofmethoxylatedandacetoxylatedcarbonsinthefragment

ions.Theprimaryfragmentsderivedfrompartiallymethylated(vi)thepresenceofcorefucosylatedstructures.

hexitolacetateareformedbyadditionofmethoxylatedcarbons,44U,and/oracet-CH2OMeCH2OMeCH2OMeoxylatedcarbons,72U.Consequently,the

OOOOMeOOfragmentionscontainingC-6givem/zCH3MeOOMeOMeOOMevaluesﬁttingtheequation(aÂ44+bÂMeOB2OOOMeOMe72+1)andthefragmentionscontainingC-OMeNMeCHOMeCHCHAc1givem/zvaluesﬁttingtheequation(aÂ222OOO44+bÂ72+2).Fora6-deoxysugar,theOMeOMeOMeOOMeOfragmentionscontainingC-6givem/zMeOCH2OMeCH2OMeCH2OMevaluesﬁttingtheformula(aÂ44+bÂNMeY1NMeOOOOMeOOMeOMeOMeOO72+15)(ref.38).AcAcInaddition,aminitolderivativesareOMeNMeOMereadilyidentiﬁedbythefragmentionthat

AccarriestheN-methylacetamidylatedcarbonandcontributes85Utothefragmention38.Figure7illustratestheprimaryfragmentsobtainedinEI-MSofseveralPMAAs,which

Oareveryusefulininterpretingdata.Itis

CH2OMeCH2OMeMeO+CH+ Na3importanttonotethatthesensitivityofthisO+ Na+MeOOMeOOOMeOMemethodismuchlowerthanthatofMALDI-OOMeTOF-MSornano-electrosprayQ-TOF-MS.CH2OMeNMeSeveralmicrogramsofpermethylatedgly-OAcOMeOMecansarerequiredforasatisfactoryanalysis.B2ionY1 ion

m/z 474

HONMeAc

Figure5|Fragmentationofsodiatedpermethylatedglycansbycleavageoftheglycosidicbonds.

Presentlimitations

Inthisglycomicsapproach,N-glycansarereleasedusingPNGaseF.Themechanismof

NATUREPROTOCOLS|VOL.2NO.7|2007|15

PROTOCOL

actionofthisenzymeinvolveshydrolysisoftheamidebondbetweentheﬁrstGlcNAcoftheN-glycanandthesidechainofAsx.Thislinkageisverywellconservedineukaryoticcells.How-ever,inbacterialproteins,carbohydrate–peptidelinkagesaredif-ferentandPNGaseFisnotcapableofhydrolyzingthislinkage.Forexample,thestructureoftheN-linkedglycanpresentonmultipleglycoproteinsintheGram-negativebacteriumCampylobacterjejuniisGalNAc-a1,4-GalNAc-a1,4-[Glc-b1,3-]GalNAc-a1,4-GalNAc-a1,4-GalNAc-a1,3-Bac-b1,N-Asn-Xaa,whereBacisbacillosamine,2,4-diacetamido-2,4,6-trideoxyglucopyranose39.Recently,anewglycomicsstrategy,basedonthecombinationofnon-speciﬁcproteolyticdigestionandpermethylation,wasdevisedthatcanbeusedforbacterialglycoproteins40.Theglycomicsapproachthatweproposeisnotapplicabletobacterialproteins.However,theothermethodsdescribedinthisprotocol,suchaspermethylation,linkageanalysis,exoglycosidasedigestionandMS,canobviouslybeusedtodeterminethecarbohydratestructureofbacterialglycoproteins.Inaddition,thesodiumhydroxidemethoddescribedinthisprotocolisthesimplestapproachtopermethylationofglycans.However,thismethodcannotbeusediftheglycansaresulfated,inwhichcasetheHakomoriprotocolmustbeused41.

Thestrategiesinglycoproteinanalysisdependontheamountofavailablesample.Iflessthan50mgofproteinisavailable,permethylationderivatizationonPNGaseF–releasedglycansfromglycoproteinscanbeperformedandpermethylatedPNGaseF–releasedglycansareanalyzedbyMALDI-TOF-MS,MS/MSsequencingandlinkageanalysis.ExoglycosidasedigestionscanbeperformedontheMALDItarget.Whenglycoproteinsareisolatingfromgelspotsinelectrophoresis,theyaremostoftenavailableatonlypicomolarquantities.Forthisreason,permethylationderivatizationandobviouslylinkageanalysiscan-notbeperformedonPNGaseF–releasedglycansfromglyco-proteinsisolatingfromgelspotsinelectrophoresisunlesstheseglycoproteinsareabundant.Therefore,PNGaseF–releasedglycansfromglycoproteinsisolatedfromgelspotsinelectrophoresisareanalyzedbyMALDI-TOF-MSwithoutderivatizationinthelinearnegativeionmodeorinthereﬂectronpositiveionmodeafterachemicaldesialylation.Exoglycosidasedigestionscanalsobeper-formedontheMALDItarget.AllstructuralassignmentsaremadeonthebasisoftheMALDImassandexoglycosidasediges-tions.Theresultsoftheseexperimentswouldnormallybesufﬁcientforidentiﬁcation.

TABLE2|MostcommonlyencounteredB-typeionscorrespondingtonon-reducingstructuresinsodiatedpermethylatedglycans.Mass2823984865276606907017318348479351,021

B-typeionHexNAc1+Na+NeuAc1+Na+

Hex1HexNAc1+Na+HexNAc2+Na+

Fuc1Hex1HexNAc1+Na+Hex2HexNAc1+Na+Fuc1HexNAc2+Na+

HexNAc1Hex1HexNAc1+Na+Fuc2Hex1HexNAc1+Na+NeuAc1Hex1HexNAc1+Na+Hex2HexNAc2+Na+

NeuAc1Fuc1Hex1HexNAc1+Na+

©2007 Nature Publishing Group http://www.nature.com/natureprotocolsDeterminationofsialicacidresiduesthatoftenoccuringlyco-proteinglycansisnotpossibleusingthelinkageanalysisdescribedinthisprotocol.Thepresenceandlinkagepositionsofthesesugarconstituentsmustbeindirectlyestablishedbymethylationanalysisbeforeandafterdesialylation.Toprovethepresenceofsialicacidresidues,methylatedoligosaccharidescanalsobedegradedbymethanolysis42.Itisalsoimportanttoknowthat,undertheexperimentalconditionsdescribed,alkali-labilesubstituantsarelostduringmethylation.

Applicationsofthisapproach

Recently,wehavesuccessfullyappliedthisapproachtotheanalysisofdifferentglycoproteins,includinggalactommano-proteinsofAspergillusfumigatus6,excreted/secretedglycoproteinsoftheparasiteGiardiaintestinalis7,bovinelysosomala-mannosi-dase5andrecombinanthumanthyrotropin8.Wehavealsoidenti-ﬁedacquiredmodiﬁcationsofglycosylationbyapplyingthisapproachtotheserumofpatientswithcirrhosis9.Inaddition,morerecently,thisapproachhasbeenusedtoproposearapidmassspectrometricstrategyforthecharacterizationofN-andO-glycanchainsinthediagnosisofcongenitaldisordersofglycosylation10.Thus,thisapproachcanbeusedtostudytheglycosylationofisolatedcomplexglycoproteinsorofnumerousglycoproteinsencounteredinacomplexbiologicalmedium(e.g.,ﬂuid,cells,organ,tissue).

TABLE3|Exoglycosidasesandconditionsusedtoconﬁrmstructureassignmentsmadefrommatrix-assistedlaserdesorption/ionization–timeofﬂightmassspectrometryexperiments.Enzymea-Sialidaseb-Galactosidase

b-N-Acetylglucosaminidasea-Fucosidasea-Mannosidase

SourceArthrobacterureafaciensBovinetestesJackbeanBovinekidneyJackbean

Speciﬁcity

a2,6-,a2,3-,a2,8-boundsialicacids

b-boundgalactose

b-boundN-acetylglucosaminea1,2-,a1,3-,a1,4-,a1,6-boundfucose

a1,2-,a1,3-,a1,6-boundfucose

Buffer

50mMsodiumacetate,pH5

50mMammoniumformate,pH4.6

50mMammoniumacetatebuffer,pH4.5

Amountofenzyme(mU)501031040525

Exoglycosidasedigestionsarecarriedoutonreleasedglycansat371Cfor48h.After24h,asecondaliquotofenzymeisadded.Theamountofenzymeindicatedinthetablecorrespondsonlytotheﬁrstaliquotofenzyme.Aftereachexoglycosidasedigestion,asmallaliquotispermethylatedandanalyzedbymatrix-assistedlaserdesorption/ionization–timeofﬂightmassspectrometryandlinkageanalysis.

1590|VOL.2NO.7|2007|NATUREPROTOCOLS

PROTOCOL

MATERIALS

REAGENTS

.2,4,6-trihydroxyacetophenone(THAP;Sigma).2.5-Dihydroxybenzoicacid(DHB;Sigma).Aceticacid(Fluka)

.Aceticanhydride(Sigma)

.Acetonitrile(MeCN;AcrosOrganics)!CAUTIONMeCNshouldbehandledinahoodusinggloves.

.Ammoniumhydroxide(NH4OH;AcrosOrganics).Ammoniumacetate(Sigma)

.Ammoniumhydrogencarbonate(Sigma).Ammoniumformate(Sigma)

.Arthrobacterureafacienssialidase(Roche).Bovinekidneya-fucosidase(Sigma).Bovinetestesb-galactosidase(Sigma).Coomassiebluedye(Biorad).Chlorhydrateguanidinium.Cyanogenbromide

.Carboxamidomethylatedglycoprotein.Chloroform

.Jackbeana-mannosidase(Sigma)

.Jackbeanb-N-acetylhexosaminidase(Sigma).b-mercaptoethanol(Biorad)

.Diammoniumcitrate(DAC;Sigma).DMSO(Fluka).DTT(Biorad)

.Dowexbeads(50Â8,50–100mesh;Biorad).Ethanol(CarloErbaReactifs-SDS)

.Methanol(CarloErbaReactifs-SDS)!CAUTIONMethanolshouldbehandledinahoodusinggloves..Hydrochloricacid(HCl;Fluka).Iodoacetamide(IAA;Biorad)

.Iodomethane(ICH3;Fluka)!CAUTIONICH3ishighlytoxicandshouldbehandledinahoodusinggloves.

.NP40(BDHChemicals,Poole,UK)(seeREAGENTSETUP).Potassiumferricyanide(Sigma)

.Pyridine(Fluka)!CAUTIONPyridineshouldbehandledinahoodusinggloves..NaBH4(Sigma)

.Sodiumborodeuteride(NaBD4;Sigma).SDS(Biorad)

.Sodiumhydroxide(NaOH;CarloErbaReactifs-SDS).Sodiumthiosulfate(Sigma)

.Triﬂuoroaceticacid(TFA;Fluka)!CAUTIONTFAshouldbehandledinahoodusingprotectiveeyewearandgloves..Trypsin(Sigma)

.DowexresinmCRITICALItisessentialtousethehighest-qualitysolventsforeverypurpose.EQUIPMENT

.Columnof150mgofnon-porousgraphitizedcarbon(Alltech)

20–4

63463±43High-mannoseN-glycans

2HybridN-glycans4± 6

3±3/2± 44± 6

±3/6±3/6±3/24± 6

4444624± 6

± 4432Tetraantennarycomplex N-glycansFigure6|ThedifferentclassesofN-linkedglycans.Symbols:opensquare,galactose;closedsquare,N-acetylglucosamine;opentriangle,fucose;closedtriangle,N-acetyl-neuraminicacid;closedcircle,mannose.

.Sep-PakC18cartridge(Waters)

.Spectra/Porregeneratedcellulosemembraneno.1(cutoff6–8kDa)(VWR).SpeedVacconcentrator(Eppendorf)

.MALDI-TOF-MS(VoyagerEliteDE-STRPro;PerSeptiveBiosystem,

AppliedBiosystems,Framingham,MA)equippedwithapulsednitrogenlaser(337nm)andagridlessdelayedextractionionsource

.Q-TOFMS(Q-STARPulsar;AppliedBiosystems/MDSSciex,Toronto,Canada)ﬁttedwithanano-electrosprayionsource(Protana,Odense,Denmark)

.Goldcoated‘medium-length’borosilicatecapillaries(Protana,Toronto,Ontario,Canada)

.GC-MSinstrument(CarboErbaGC8000gaschromatograph(CarboErba)coupledtoaFinniganAutomassIImassspectrometer,Finnigan)

.CP-Sil5CBlow-bleedcapillarycolumn,25mÂ0.32mm,0.25-mmﬁlmphase(Chrompack,France)REAGENTSETUP

GlycoproteinsampleThisprotocolenablestheanalysisof5–20mgof

glycoproteindirectlyfromSDS-PAGEor2D-gels(N-glycans)orfrom20mgto

TABLE4|GC-MSanalysisofpartiallymethylatedalditolacetatesobtainedfromPNGaseF–releasedN-glycansfromhumanserum.Retentiontime(min)15.2919.5520.4623.5524.4626.4228.0229.2929.5131.1032.1634.3335.2336.4637.58

Characteristicfragmentions(m/z)115,118,131,162,175

102,118,129,145,161,162,205102,118,129,145,161,162,205129,130,161,190118,129,161,234

99,102,118,129,162,1,233130,190,233

129,130,1,190118,129,1,234118,333

117,159,203,205117,159,233118,233118,129

117,159,261

AssignmentTerminalFucTerminalManTerminalGal2-linkedMan3-linkedGal6-linkedGal2,4-linkedMan2,6-linkedMan3,6-linkedMan3,4,6-linkedManTerminalGlcNAc4-linkedGlcNAc4-linkedHexNAc3,4-linkedGlcNAc4,6-linkedGlcNAc

Relativeabundance

0.060.020.060.540.040.420.030.030.210.010.051.000.020.010.08

NATUREPROTOCOLS|VOL.2NO.7|2007|1591

PROTOCOL

2mgofglycoproteinfollowingPNGaseFdigestionforN-glycansandreductiveeliminationforO-glycans.ThegelpiecescanbeCoomassieblue–stainedorsilver-stained.IfthesampleisnotanalyzeddirectlyfromSDS-PAGEor2D-gels,itcanbefreeze-driedorinsolution.Itisimportanttodesalttheglycoproteiniftheconcentrationofsaltsishigh.

In-solutiondenaturingbuffer50mMammoniumbicarbonate(pH8.3),1%SDSand1%b-mercaptoethanol.

NP4010%(vol/vol)NP40isdilutedin50mMammoniumbicarbonate.RecombinantPNGaseF(Roche)LyophilizedPNGaseFisdissolvedinpurewatertoobtainaﬁnalconcentrationof1UmlÀ1(stability:severalmonthsat41C).mCRITICALSomecommercialPNGaseFcanbestoredatÀ201Cinpresenceofglycerol.However,fortheMALDI-TOF-MSstudies,glycerolpreventstheco-crystallizationoftheDHBmatrixandthesample.Forthisreason,weonlyuseafreeze-driedPNGaseFwithoutglycerol.Inaddition,PNGaseFreleasespracticallyallN-glycansexceptthosewithfucoseattachedtothe3-positionoftheAsn-linkedGlcNAcresidue.Theircorrespondingglycoproteinsarecommonlyfoundinplantsandininsectcells.SuchPNGaseF–resistantglycanshavebeenfoundtobesensitivetoPNGaseA12.Reduction/alkylationbuffer0.6MTris–HClsolutioncontaining6M

chlorhydrateguanidinium,pH8.4.Bubblewithnitrogenfor30minbeforeuse.Reductionbuffer1500mMDDTsolutioninthereduction/alkylationbuffer.Alkylatingbuffer13MIAAsolutioninthereduction/alkylationbuffer.mCRITICALThissolutionmustbekeptinthedark.

Reductionbuffer210mMDTTsolutioncontaining100mMammoniumhydrogencarbonate.

Alkylatingbuffer255mMIAAsolutioncontaining100mMammoniumhydrogencarbonate.

Reductiveeliminationsolution50mMNaOH,1MNaBH4.

DHBmatrixsolutionMethanol/1mMsodiumacetate(1:1,vol/vol)containing10mgofDHBperml.

THAPmatrixsolutionMeCN/DAC20mM(1:1,vol/vol)containing20mgofTHAPmonohydrateperml.

TFAsolution0.1%TFAinwater(vol/vol).EthanolsolutionEthanol/water(1:1,vol/vol).

Reductionsolution2MNH4OHcontaining4mgofNaBD4perml.EQUIPMENTSETUP

DesaltingcolumnThedesaltingcolumnisaPasteurpipettepluggedatthetaperedendwithasmallamountofglasswool.Apieceofsiliconetubing(approximately2cm)isplacedatthetaperedend,andtheﬂowisblockedusinganadjustableclip.Fillthepipettewith5%(vol/vol)aceticacidandopentheclipslightlytolettheequilibrationbufferslowlyﬂowout.Whiletheaceticacidisrunningout,thecolumnisﬁlledwith2mloffreshlywashedDowexbeads(50Â8,H+form).TheuseofPasteurpipetteslimitsthevolumeofDowexresinto2ml.Intheseconditions,itispossibletoremovecationicmaterials(sodiumsalts,aminoacidsandpeptides)obtainedfromthereductivetreatmentofglycoproteinsfullyupto2mg.

LinkageanalysisofPMAAsbyGC-MSThetemperatureoftheRossinjectoris2601C.Samplesareanalyzedusingatemperatureprogramstartingwitha

t-FucCHDOAcMeOCH175131

118

CHOMe162CHOMeAcOCH

CH3

t-ManCHDOAcMeOCH205MeOCH161

CHOMeCHOAcCH2OMe

118162

277

3-Linked Gal

CHDOAcCHOMe118 AcOCH161MeOCH

CHOAcCH2OMe

234

6-Linked Gal

CHDOAc

CHOMe118

233MeOCH1MeOCH

CHOAcCH2OAc

162

3,6-Linked Man

CHDOAc

305MeOCH

AcOCH

CHOMe234CHOAcCH2OAc

118

3,4,6-Linked Man

CHDOAc

333MeOCH

AcOCH

CHOAcCHOAcCH2OAc

118

CHDOAcAcOCHMeOCH161

CHOMeCHOAcCH2OMe

190

t-GlcNAc

CHDOAcCHNMeAc159

205MeOCH161

CHOMeCHOAcCH2OMe

203

4-Linked GlcNAc

CHDOAcCHNMeAc159

233MeOCH

CHOAcCHOAcCH2OMe

3-Linked GlcNAc

CHDOAcCHNMeAc159AcOCH161

CHOMe275CHOAcCH2OMe

3,4-Linked GlcNAc

CHDOAcCHNMeAc159AcOCH

CHOAcCHOAcCH2OMe

4,6-Linked GlcNAc

CHDOAcCHNMeAc159

261MeOCH

CHOAcCHOAcCH2OAc

Figure7|Primaryfragmentsobtainedinelectronimpact(EI)massspectrometryofthemostcommonpartiallymethylatedalditolacetates(PMAAs)obtainedfromN-glycans.

gradientof21CminÀ1from130to1801C,after2minat1301C,followedbyagradientof41CminÀ1until2401C.ThecolumniscoupledtoaFinniganAutomassIImassspectrometer.AnalysesareperformedintheEImodeusinganionizationenergyof70eV.QuantiﬁcationofthevariousPMAAderivativesiscarriedoutusingtotalioncurrentoftheMSdetectorinpositiveionmode.

PROCEDURE

N-glycanrelease

1|PrepareN-glycansamples(Fig.1)inoneofthefollowingways:in-solutiondeglycosylationafterdenaturationof

glycoproteinswithSDS(optionA),in-solutiondeglycosylationafterreductionandalkylationofglycoproteins(optionB)orin-geldeglycosylation(optionC).ChooseoptionAiflessthan50mgofproteinisavailable.Ifmorethan50mgofproteinisavailable,weuseoptionB.Iflargeamountsofprotein(morethan1mg)areusedorifglycoproteinsofinterestcouldbecontaminatedbyoligosaccharidesoranythingelse(typicallyglycoproteinsextractedfromanyﬂuid,cells,organ,tissueetc.),weprefertouseoptionB.Afterreduction,alkylationanddigestionwithaprotease,theresultingpeptides/glycopeptidescanbepuriﬁedusingSep-PakC18beforePNGaseFdigestion.Thisoptionalstepremovespossiblecontaminants.TheprotocoltoconditiontheSep-PakC18columnandtopurifythesampleisdescribedinStep2B.Inthiscase,onlythefractionthatcorrespondstoamixtureofpeptidesandglycopeptidesiscollectedandfreeze-driedafterremovingMeCNunderastreamofnitrogeninahood.Beforein-geldeglycosylation(optionC),proteinbandsareexcisedfromthegelanddestained.GlycoproteinsneedtobereducedandcarbamidomethylatedinthegeltoensurecompletePNGaseFdeglycosylation.(A)In-solutiondeglycosylationafterdenaturationofglycoproteinswithSDS

(i)Dissolvetheglycoproteinsamplein200mlofin-solutiondenaturingbuffer(seeREAGENTSETUP).(ii)Incubateinaboilingwaterbathfor20min.

(iii)Leavethemixtureatroomtemperature(201C)for20min.

1592|VOL.2NO.7|2007|NATUREPROTOCOLS

PROTOCOL

(iv)Add200mlof50mMammoniumhydrogencarbonateand20mlof10%NP40(vol/vol)(seeREAGENTSETUP).(v)Leavethemixtureatroomtemperaturefor5min.

(vi)Add3mlofthePNGaseFenzymesolution(seeREAGENTSETUP).(vii)Incubateat371Covernight.

(viii)EvaporatethemixturetodrynessinaSpeedVacconcentratoratroomtemperatureforapproximately3h.Atthispoint,

ifyouarenotinterestedinstudyingtheO-glycansoftheproteins,youcanproceeddirectlytothepuriﬁcationoftheN-glycansusinganon-porousgraphitizedcarboncolumn(Step2A).Otherwise,itisnecessarytoseparatetheproteinsfromPNGaseF–releasedN-glycansbyprecipitatingtheN-deglycosylatedproteins.

’PAUSEPOINTThemixtureisstableforatleastseveralmonthswhenstoredatÀ201C.

(ix)Re-dissolvethesamplein200mlof50mMammoniumhydrogencarbonateandprecipitatetheN-deglycosylatedproteins

usingthreevolumesofethanolandincubatingfor2hatÀ201C.

(x)Centrifugethemixtureat8,000gfor10minat41CandtransfertheN-glycan-containingsupernatantintoa4-mlvial.(xi)RepeatSteps1A(ix)and1A(x)twice.ThepooledN-glycan-containingsupernatantsareevaporatedunderastream

ofnitrogenandthenfreeze-dried.ThewashedpelletscontainingtheN-deglycosylatedproteinarealsofreeze-dried.

TheN-glycansarepuriﬁedusingaSep-PakC18toremoveSDS(seeStep2B).ThewashedpelletsaresubjectedtoreductiveeliminationtoreleaseO-glycans(seeStep3).

’PAUSEPOINTTheN-glycansandthewashedpelletscontainingtheN-deglycosylatedproteinsarestableforatleastseveralmonthswhenstoredatÀ201C.

(B)In-solutiondeglycosylationafterreductionandalkylationofglycoproteins

(i)Dissolvetheglycoproteinsamplein200mlofreduction/alkylationbuffer(seeREAGENTSETUP).Thevolumemaybeincreasedtoensuresolubility.(ii)Incubateat451Cfor90min.

(iii)Preparereductionbuffer1justbeforeuse.

(iv)Addtheappropriateamountoffreshlypreparedreductionbuffer1totheglycoproteinsolutiontoobtainaﬁnal

concentrationof20mMDTT.Flushthemixturewithargon(ornitrogen)andincubateat451Cfor4h.(v)Preparealkylatingbuffer1justbeforeuseandkeepitinthedark.

(vi)Addtheappropriateamountoffreshlypreparedalkylatingbuffer1totheglycoproteinsolutiontoobtainaﬁnalconcen-trationof110mMIAA.Flushthemixturewithargon(ornitrogen)andleaveinthedarkatroomtemperatureovernight.(vii)Desaltthecarboxamidomethylatedglycoproteinbydialyzingagainst4Â2lof50mMammoniumhydrogencarbonateat

41C,witheachchangeoccurringafterstirringfor12h.(viii)Freeze-drythecarboxamidomethylatedglycoprotein.

’PAUSEPOINTThesampleisstableforatleastseveralmonthswhenstoredatÀ201C.

(ix)Dissolvethecarboxamidomethylatedglycoproteinsamplein200mloffreshlyprepared50mMammoniumhydrogen

carbonatesolution.Thevolumemaybeincreasedtoensuresolubility.

(x)Weighouttheappropriateamountoftrypsinanddissolveinthe50mMammoniumhydrogencarbonatesolutiontoaconcentrationof5mgmlÀ1.

mCRITICALSTEPPreparethetrypsinsolutionjustbeforeuse.SincePNGaseFenzymedoesnotreleaseanyN-glycanswhenglycosylationsitesareontheﬁrstorlastaminoacidresidueofapeptide,theproteasemustbecarefullychosen8.Forthisreason,whenPNGaseFdigestionisperformedonaglycoprotein,theaminoacidsequenceofthisglycoproteinmustbeknown.Trypsincanbereplacedbyotherproteasesorbyachemicalreagentsuchascyanogenbromidetogenerateglycopeptidessuchthatglycosylationsitesarenotpresentontheﬁrstaminoacidresidue.

(xi)Addthetrypsinsolutiontothecarboxamidomethylatedglycoproteinsolutiontoobtainanenzyme-to-substrateratioof

1:20(wt/wt).

(xii)Incubateat371Cfor24hundergentleagitation.

(xiii)Afterheatingthesolutionat1001Cfor10min,freeze-drythesample.

mCRITICALSTEPItisimportanttodestroytrypsinbeforereleasingN-glycansusingPNGaseF.’PAUSEPOINTThesampleisstableforatleastseveralmonthswhenstoredatÀ201C.

(xiv)Dissolvetheresultingpeptide/glycopeptidemixturein200mloffreshlyprepared50mMammoniumhydrogencarbonate

solution.Thevolumemaybeincreasedtoensuresolubility.

(xv)Add3mlofthePNGaseFenzymesolution(seeREAGENTSETUP).(xvi)Incubateat371Covernight.(xvii)Freeze-drythesample.

’PAUSEPOINTThemixtureisstableforatleastseveralmonthswhenstoredatÀ201C.(C)In-geldeglycosylationandextractionofN-glycans

(i)Destainthesilver-stainedgelpieceswithasolutioncontaining30mMpotassiumferricyanidesolutionand100mMsodiumthiosulfatewithshaking.

NATUREPROTOCOLS|VOL.2NO.7|2007|1593

(ii)DestaintheCoomassieblue–stainedgelpieceswithasolutionof100mMammoniumhydrogencarbonatein50%MeCN(1:1,vol/vol)withshaking.mCRITICALSTEPSDShastobethoroughlyremovedfromthegeltopreventdenaturationofthePNGaseFenzyme14.’PAUSEPOINTGelpiecescanalsobeleftovernightat41Cinthissolutionwithoutshaking.ThishelpstoremovetheCoomassiestain.

(iii)Washthegelpiecesthreetimeswithpurewaterfor10minwithshaking.(iv)DehydratethegelpiecesthreetimeswithMeCNfor5minwithshaking.

(v)Drythegelpiecesatroomtemperatureforapproximately30mininaSpeedVacconcentrator.

’PAUSEPOINTThedriedgelpiecescanbeleftovernightat41Cifyoucannotproceedimmediately.(vi)Coverthedriedgelpieceswithreductionbuffer2andincubateat561Cfor30min.

mCRITICALSTEPPreparereductionbuffer2justbeforeuse.

(vii)Removereductionbuffer2,coverthegelpieceswithalkylatingbuffer2andincubateinthedarkatroomtemperaturefor

30min.

mCRITICALSTEPPreparealkylatingbuffer2justbeforeuseandkeepitinthedark.

(viii)Removethealkylatingbufferandwashthegelpiecesthreetimeswithpurewaterfor10minwithshaking.(ix)DehydratethegelpiecesthreetimeswithMeCNfor5minwithshaking.

(x)Drythegelpiecesatroomtemperatureforapproximately30mininaSpeedVacconcentrator.

’PAUSEPOINTThedriedgelpiecescanbeleftovernightat41Cifyoucannotproceedimmediately.

(xi)Add3mlofthePNGaseFenzymesolutiontothedriedgelpieces(seeREAGENTSETUP)andleavethegelpiecesatroom

temperaturefor5min.

(xii)Coverthegelpieceswith50mMammoniumhydrogencarbonatebufferandincubateat371Covernight.

(xiii)Extractglycansfromthegelpiecesbyremovingtheincubationbufferfollowedbythreeextractionswithapproximately

300mlofpurewaterandtwoextractionswith50%MeCN,withsonicationfor30mineach.AllextractsandtheremovedincubationbufferarecombinedanddriedinaSpeedVacconcentratoratroomtemperature.’PAUSEPOINTTheglycansarestableforatleastseveralmonthswhenstoredatÀ201C.N-glycanpuriﬁcation

2|PurifyPNGaseF–releasedN-glycansbyeitherN-glycanpuriﬁcationusinganon-porousgraphitizedcarboncolumn(optionA),N-glycanpuriﬁcationusingaSep-PakC18cartridge(optionB)orN-glycanpuriﬁcationusinganon-porousgraphitizedcarbonmini-column(optionC).ChooseoptionAifyoureleaseN-glycansafterdenaturationofglycoproteinswithSDSandifyouarenotinterestedinstudyingtheO-glycosylationofproteins.ChooseoptionBintwocases:(i)ifyoureleaseN-glycansafterdenaturationofglycoproteinswithSDSandifyouareinterestedinstudyingtheO-glycosylationofproteins(inthiscase,theSep-PakC18cartridgewillremoveSDSfromN-glycans);(ii)ifyoureleaseN-glycansafterreduction/alkylationandproteolyticdigestionofglycoproteins(inthiscase,theSep-PakC18cartridgewillseparatePNGaseF–releasedN-glycansfromthemixtureofpeptidesandO-glycopeptides).ChooseoptionCifyoureleaseN-glycansfromgel-separatingproteins.Whenglycoproteinsareisolatingfromgelspotsinelectrophoresis,optimizationofmethodsforpuriﬁcationandanalysisofglycansisrequired.Inthiscase,glycoproteinsaremostoftenavailableatonlypicomolarquantities.Forthisreason,thoughthenon-porousgraphitizedcarbonmini-columnmaybeusedtopurifysialylatedglycans,weﬁrstdesialylatePNGaseF–releasedN-glycans.ThechemicaldesialylationresultsinmaximumsensitivityfortheMALDI-TOF-MSanalysisofN-glycanswithoutusingpermethylationderivatizationinthepositivereﬂectronionmode.(A)N-glycanpuriﬁcationusinganon-porousgraphitizedcarboncolumn

(i)AfterPNGaseFdigestioninthepresenceofSDSandevaporatingthemixturetodryness,re-dissolvethesamplein500mlto1mlof0.1%TFAsolution(vol/vol)forloadingontoacolumnofnon-porousgraphitizedcarbon43.

(ii)Conditionthecolumnof150mgofnon-porousgraphitizedcarbonwith5mlofmethanolandwashitwith10mlof0.1%TFA.(iii)Loadthesampleontothecolumnandremovesaltswith15mlof0.1%TFA.(iv)RecoverN-glycansbypassingthroughthecolumn3mlof25%MeCNin0.1%TFAanddirectlyfreeze-dry.Thisprotocolefﬁcientlyremovessalts,SDSandN-deglycosylatedproteins.However,O-glycanscannotbereleasedfromN-deglycosy-latedproteinsusingthisprotocol.mCRITICALSTEPDonotuseamorehydrophobicsolutiontoeluteN-glycans;ifyoudo,SDScouldalsobeelutedwithN-glycans.’PAUSEPOINTTheglycansarestableforatleastseveralmonthswhenstoredatÀ201C.(B)N-glycanpuriﬁcationusingaSep-PakC18column(i)Dissolvethesamplein100–200mlof5%aceticacid.

(ii)ConditiontheSep-PakC18columnwith5mlofmethanolandthenwith10mlof5%aceticacid.

PROTOCOL

(iii)Loadthesampleontothecolumnandelutestepwisewith3mlof5%aceticacidand4mlof5%aceticacidin

80%MeCN.

(iv)Collecteachfraction.TheN-glycansareelutedwith5%aceticacid.Thefractionelutedwith5%aceticacidin80%MeCN

iscollectedonlywhenPNGaseFdigestionhasbeenperformedafterreduction/alkylationandproteolyticdigestionofglycoproteins.ThisfractionthatcorrespondstoamixtureofpeptidesandO-glycopeptideswillbesubjectedtoreductiveelimination(seeStep3).

(v)RemoveMeCNunderastreamofnitrogeninahood.(vi)Freeze-dryeachfraction.

’PAUSEPOINTSamplesarestableforatleastseveralmonthswhenstoredatÀ201C.(C)N-glycanpuriﬁcationusinganon-porousgraphitizedcarbonmini-column(i)Dissolvethesamplein200mlof2Maceticacid.(ii)Incubateat801Cfor2h.

(iii)Evaporatethesampletodrynessatroomtemperatureforapproximately1hinaSpeedVacconcentrator.

’PAUSEPOINTTheglycansarestableforatleastseveralmonthswhenstoredatÀ201C.

(iv)Preparemini-columnsbydismantlingacolumnof150mgofnon-porousgraphitizedcarbon(Alltech).(v)Pack200-mlﬁlterpipettetipswithapproximately5mgofnon-porousgraphitizedcarbon.

(vi)Conditionthemini-columnbysequentialwashingwith1mlofmethanoland2mlof0.1%TFA.(vii)Dissolvetheextractedglycansin0.1%TFAandloadthesampleontothecolumn.(viii)Purifythesamplewith2mlof0.1%TFA.

(ix)RecoverN-glycansbypassingthroughthemini-column500mlof25%MeCNin0.1%TFA.

(x)Evaporatethesampletodrynessatroomtemperatureforapproximately2hinaSpeedVacconcentrator.’PAUSEPOINTGlycansarestablewhenstoredatÀ201C.O-glycanrelease

3|DissolvetheN-deglycosylatedproteins(Step2A)ortheresultingpeptides/glycopeptides(Step2B)in200mlofreductiveeliminationsolution(seeREAGENTSETUP).

mCRITICALSTEPPreparethereductiveeliminationsolutionjustbeforeuse.4|Incubateat451Covernight.

5|Stopthereactionbyaddingaceticaciddropwiseuntilnoﬁzzingisobserved(approximatelythreedrops).

O-glycanpuriﬁcation

6|Pre-treatDowexbeads(50Â8,H+form,50–100mesh).Washtwicewith3MNaOHfor30minatroomtemperature.7|WashextensivelywithwateruntilthepHofthewaterwashmeasuredbypaperpHisslightlybelow7.8|Washtwicewith3MHClfor30minatroomtemperature.

9|WashextensivelywithwateruntilthepHofthewaterwashisslightlybelow7.

10|Washthreetimeswith5%aceticacid.

’PAUSEPOINTBeadscanbestoredinthe5%aceticacidsolutionforseveralmonthsatroomtemperature.

11|PackaPasteurpipette,pluggedatthetaperedendwithasmallamountofglasswool,withtheDowexbeads(see

REAGENTS).TheuseofPasteurpipetteslimitsthevolumeofDowexresinto2ml.Intheseconditions,itispossibletoremovecationicmaterials(sodiumsalts,aminoacidsandpeptides)obtainedfromthereductivetreatmentofglycoproteinsfullyupto2mg.

12|Washthecolumnwith5volumesof5%aceticacid(approximately10ml).13|Loadthesampleontothecolumnandelutewith4mlof5%aceticacid.14|Collecttwofractionsof2ml.

15|Freeze-dryimmediately.

’PAUSEPOINTDriedglycansarestableforatleastseveralmonthswhenstoredatÀ201C.

16|Removeboratesaltsbyrepeatedevaporationwithmethanolcontaining5%aceticacid(5Â500ml)underastreamofnitrogeninahood.

PROTOCOL

17|Addthesecondfractiontotheﬁrstfractionandfreeze-dry.

’PAUSEPOINTDriedglycansarestableforatleastseveralmonthswhenstoredatÀ201C.

Exoglycosidasedigestions

18|IfPNGaseF–releasedglycansareisolatedfromgel-separatedproteins,carryoutexoglycosidasedigestionsontheMALDItarget(optionA);otherwiseexoglycosidasedigestionscanbecarriedoutinsolution(optionB).(A)ExoglycosidasedigestionsontheMALDItarget

(i)Mix1mlofglycans(20pmol)directlyon-platewith1mlof20mMammoniumformiate,pH4.6.

(ii)Add1mlofeachexoglycosidaseontoeachspottedsampleatthefollowingconcentration:a-sialidase,50mU;

a-galactosidase,1.25mU;a-N-acetylhexosaminidase,150mU;a-fucosidase,7.7mU.Foreachsample,thenumberofspotsdependsonthenumberofexoglycosidases,aloneorincombination,totest.

(iii)PlacetheMALDIplateat371Cfor6hinacrystallizationbeakercontainingwatertosaturateatmosphere.

mCRITICALSTEPHigh-puritywatermustberegularlyaddedtothespotstopreventair-drying.(iv)Stoptheenzymaticreactionsbyallowingthesamplestodryatroomtemperature.

(v)Add2mloftheDHBmatrixsolutionandallowtodryatroomtemperature.TheMALDI-TOF-MSanalysisisperformedasdescribedbelow(Step37).

(B)Exoglycosidasedigestionsinsolution

(i)Dissolveglycans(typically50%ofPNGaseF–releasedglycans)in200mlofappropriatebuffer(seeTable3).(ii)Addtheappropriateamountofexoglycosidase(seeTable3).

(iii)Incubateat371Cfor48h.After24h,asecondaliquotofenzymeisadded.(iv)Stopthereactionbyheatingthesolutionat1001Cfor10min.

(v)Takeasmallaliquotforpermethylation(Step19).Therestofsampleissubjectedtoanotherexoglycosidasedigestion.(vi)Freeze-drybothaliquots.

mCRITICALSTEPGlycansmustbefreeze-driedjustbeforepermethylationinordertoavoidunder-methylation.Permethylation

19|Placetheglasstubecontainingthefreeze-driedglycansinavacuumvesselsaturatedwithanargonatmosphere.20|Add500mlofDMSO.

21|PlaceﬁvetotenpelletsofNaOHinadrymortar.

Note:TransformthepelletsofNaOHintoaﬁnepowderusingthepestlerapidly.22|Addapproximately25mgofNaOHtothesample.

23|Add300mlofICH3andﬂushthetubewithastreamofargon.

mCRITICALSTEPDMSO,NaOHandICH3mustbekeptunderargonatmosphere.Aglasstubemustbeusedforthepermethylation.24|Mixvigorouslyandplacethereactionmixtureinanultrasonicbathfor90minatroomtemperature.

mCRITICALSTEPDonotleavethereactionmixtureintheultrasonicbathformorethan90mintoavoidartifacts.25|Quenchthereactionbyadditionofapproximately1mlof5%aceticacidat41C.Mixvigorously.26|Add600mlofchloroform.Mixvigorously,andallowthemixturetosettleintotwolayersat41C.27|Transferthelowerchloroformphaseintoanewglasstube.28|RepeatSteps26and27twice.

29|Washthechloroformphaseeighttimeswith1volumeofwaterat41Canddiscardaqueousphases.30|Drydownthechloroformphaseunderastreamofnitrogeninahood.

’PAUSEPOINTPermethylatedglycansarestableforatleastseveralmonthswhenstoredatÀ201C.Puriﬁcationofpermethylatedglycans

31|Dissolvethesamplein100–200mlofmethanol.

32|ConditiontheSep-PakC18columnwith5mlofmethanolandthenwith10mlofwater.

33|Loadthesampleontothecolumnandelutesequentiallywith15mlofwater,2mlof10%MeCNand3mlof80%MeCN.

PROTOCOL

34|Collectthe80%MeCNfractionandevaporateMeCNunderastreamofnitrogeninahood.35|Freeze-drythesample.

’PAUSEPOINTPermethylatedglycansarestableforatleastseveralmonthswhenstoredatÀ201C.

PreparationofglycansforMALDI-TOF-MSanalysis

36|AnalyzepermethylatedglycansusingtheDHBmatrix(optionA)anddesialylatednativeN-glycansisolatedfromgel-separatedglycoproteinsandpuriﬁedglycansusinganon-porousgraphitizedcarbonmini-columnusingtheTHAP

monohydratematrix(optionB)44.TheTHAPmonohydratematrixgivesbetterresultsthantheDHBmatrixforsialylatedandnativeglycans,especiallywhenyouworkonsmallamounts45.(A)PreparationofpermethylatedglycansforMALDI-TOF-MS

(i)Dissolvepermethylatedglycansinamethanol/water(1:1;vol/vol)solutiontoobtainaconcentrationof10pmolmlÀ1.(ii)Mix1mlofthesolutionwith1mloffreshlymadeDHBmatrixsolution(seeREAGENTSETUP)directlyontotheMALDItarget.

(iii)Allowthemixturetodryatroomtemperature.

(iv)Add0.5mlofethanolsolution(seeREAGENTSETUP).(v)Allowthemixturetodryatroomtemperature.

mCRITICALSTEPRecrystallizationofthesamplespotwithethanolpromotestheformationofsmallcrystalsandiscrucialforobtainingmaximumsensitivityandspot-to-spotreproducibility.Caremustbetakentopreventethanolfromspreadingthesampleoveralargerarea.

(B)PreparationofdesialylatedN-glycansreleasedfromgel-separatedproteinsforMALDI-TOF-MS(i)Dissolvedesialylatedglycansin10mlofwater.

(ii)Mix1mlwith1mloffreshlymadeTHAPmatrixsolution(seeREAGENTSETUP)directlyontotheMALDItarget.mCRITICALSTEPTheTHAPmatrixmustbepreparedjustbeforeuse.(iii)Allowthemixturetodryatroomtemperature.

(iv)Add1mloffreshlymadeethanolsolution(seeREAGENTSETUP)45.(v)Allowthemixturetodryatroomtemperature.AcquisitionofMALDI-TOF-MSspectra

37|Acquirethespectrabysubmittingeachspottomultiplelasershots(100–500)overarangeofm/z¼1,000–10,000forpermethylatedN-glycans,m/z¼300–10,000forpermethylatedO-glycansorm/z¼900–7,000fornativeN-glycans.Thelaserspotshouldbemovedtodifferentareasofthetargetduringacquisition,andspectrashouldbeaverageduntilasatisfactorysignal-to-noiseratioisobtained.

AssignmentsofmolecularionsobservedinMALDI-TOF-MSspectra

38|Calculate[M+Na]+valuesbyaddingthesumoftheresiduemassesandthesumofthesodiatedreducingandnon-reducingendincrements(Table1).TheobjectofthisexerciseistomatchcalculatedmassesfromTable1withthe

experimentalmassestoobtainthecompositionoftheglycan.Forexample,them/zofabiantennarydisialylatedN-glycanis2,792(2Â361.17+5Â204.09+4Â245.12+15.02+31.01+22.98).ForO-glycansithastobekeptinmindthattheyarereducedwithNaBH4duringthereleaseofO-glycansfromthepeptidicchain.Therefore,[M+Na]+valuesforO-glycansare

obtainedbyaddingthesumoftheresiduemassesandthesumofsodiatednon-reducingandreducedreducingendincrements.Forexample,them/zofaglycanlinkedO-glycosidicallytoSerandThrresiduesthathasbeenreleasedfromtheglycoproteins/glycopeptidesinthereducedformcontainingGalNAc-olandwiththefollowingcomposition(NeuAc1Hex1HexNAc-ol)is5(361.17+204.09+245.12+15.02+47.05+22.98).Thisprocedurecanalsobecarriedoutwithacomputerusingeitherin-housesoftwareortoolsthatareavailableontheInternet.Forexample,GlycoMod(http://www.expasy.ch/tools/glycomod/)isaweb-basedtoolthatcanﬁndallpossiblecompositionsofaglycanfromitsdeterminedmass46.Theprogramcanbeusedtopredictthecompositionofanyglycoprotein-derivedoligosaccharide.

Nano-ESI-MS-MSofpermethylatedglycans

39|Dissolvederivatizedglycansinasolutionof80%methanol;1%aceticacidcanbeaddedtopromoteformationofprotonatedadducts28.

40|Load4mlintoagold-coated‘medium-length’borosilicatecapillary.41|Optimizethesprayerpotentialtoachievemaximumsensitivity.

42|Recordtheconventionalmassspectra:thenumberofscansrequireddependsonthesampleconcentrationandtheresultingsignal-to-noiseratio.

NATUREPROTOCOLS|VOL.2NO.7|2007|1597

43|RecordMS/MSspectraovertherangem/z¼100–3,000.Collisionenergiesdependonthetypeofmolecularion

([M+2Na]2+,50–80V;[M+2H]2+,15–30V)andonthesizeofthecarbohydrate.Thenumberofscansrequireddependsonthesampleconcentrationandtheresultingsignal-to-noiseratio.Thepositiveionelectrospraymassspectrumofglycanscanbeveryvariable,dependingontheconcentrationofvariouscationsinthesolventandontheconevoltage29.PreparationofPMAAs

44|Dissolvepermethylatedglycansin300mlof4MTFAinaglasstube.45|Incubateat1001Cfor4h.

46|Drydownthesampleunderastreamofnitrogeninahood.

’PAUSEPOINTDriedsamplesarestableforatleastseveralmonthswhenstoredatÀ201C.

49|Stopthereactionbyaddingaceticaciddropwiseuntilnoﬁzzingisobserved(approximatelythreedrops).50|Drydownthesampleunderastreamofnitrogeninahood.

51|Removeboratesaltsbyrepeatedevaporationwithmethanolcontaining5%aceticacid(5Â500ml)underastreamofnitrogeninahood.

52|Evaporatethemixturetodryness.

mCRITICALSTEPItisimportantthatthesampleiscompletelydrybeforebeginningtheperacetylationreaction.53|Add20mlofpyridineand200mlofaceticanhydride.54|Incubateat1001Cfor2h.

55|Evaporatethemixturetodrynessunderastreamofnitrogeninahood.

’PAUSEPOINTDriedsamplesarestableforatleastseveralmonthswhenstoredatÀ201C.

56|Add400mlofchloroformand1mlofwater.Mixvigorously,andallowthemixturetosettleintotwolayers.57|Transferthelowerchloroformphaseintoanewglasstube.

58|Add400mlofchloroformintotheﬁrsttube.Mixvigorously,andallowthemixturetosettleintotwolayers.59|Transferthelowerchloroformphaseintothenewglasstube.60|RepeatSteps58and59once.

61|Washthechloroformphasetentimesintothenewglasstubewith1volumeofwateranddiscardaqueousphases.62|Drydownthechloroformphaseunderastreamofnitrogeninahood.

mCRITICALSTEPDonotleavethesampleunderastreamofnitrogenforalongtimewhenthesampleisdried.SomePMAAssuchaspermethylatedfucoseareveryvolatile.Thesampleshouldbeleftforapproximately30min.’PAUSEPOINTPMAAsarestableforatleastseveralmonthswhenstoredatÀ201C.LinkageanalysisofPMAAsbyGC-MS

63|InjectthePMAAsdissolvedinmethanol.|RuntheGCmethod(seeEQUIPMENTSETUP).

InterpretationofGC-MSdata

65|UsetheretentiontimeonthecapillaryGCcolumnandthefragmentionsobservedintheEI-MSspectratoidentifythedifferenttypeofresidues(Table4).

In-solutiondeglycosylationafterdenaturationofglycoproteinswithSDS:1dIn-solutiondeglycosylationafterreductionandalkylationofglycoproteins:1d

1598|VOL.2NO.7|2007|NATUREPROTOCOLS

󰀃TIMING

PROTOCOL

In-geldeglycosylationandextractionofN-glycans:2d

N-glycanpuriﬁcationusinganon-porousgraphitizedcarboncolumn:approximately3hN-glycanpuriﬁcationusingaSep-PakC18column:approximately20minReductiveeliminationandpuriﬁcationofO-glycans:2dExoglycosidasedigestionsontheMALDItarget:6h

Exoglycosidasedigestionsinsolution:2dforeachenzymePermethylation:2h

PuriﬁcationofpermethylatedglycansusingSep-PakC18:approximately20minPreparationofglycansforMALDI-TOF-MS:approximately10minAcquisitionofMALDI-TOF-MSspectra:approximately1min

Note:However,ifsamplesgiveverylowsignals,acquisitionofMALDI-TOF-MSspectracanrequireapproximately30min.Nano-ESI-MS-MSofpermethylatedglycans:approximately5minPreparationofPMAAs:5d

LinkageanalysisofPMAAsbyGC-MS:1h

ANTICIPATEDRESULTS

Figure1isasummaryoftheexperimentalstrategieswecurrentlyuseinourlaboratoryforthedeterminationofthestructuresofN-andO-linkedglycansofglycoproteins.ThreemethodsforpreparingN-glycansaredescribed.Theseexperimentalstrategies,whichallowimportantstructuralconclusionstobedrawnonthebasisofpicomolaramountsofcomponents,arebasedonderivatization,chemicalhydrolysis,exoglycosidasedigestions,MALDI-TOF-MS,nano-ESI-MS-MSandGC-MS.InFigure8,wepresentaMALDI-TOF-MSspectrumofpermethylatedN-glycansisolatedfromhumanserum.WehaveperformedtheprotocoldescribedinFigure1aon20mlofhumanserum.AfterdenaturationofproteinswithSDS,N-glycansarereleasedusingPNGaseFandpuriﬁedfromN-deglycosylatedproteinsusinganon-porousgraphitizedcarboncolumn(seeStep2A).AnaliquotofpuriﬁedN-glycans(25%)ispermethylated,puriﬁedonaSep-PakC18andanalyzedusingMALDI-TOF-MS(Fig.8).TheputativemajorstructureassignmentsarebasedonmolecularweightandknowledgeofN-glycanbiosynthesis.Fromtheirm/zratio,monosaccharidecompositionintermsofHex,dHex,NeuAcandHexNAcofeachPNGaseF–releasedoligosaccharidehasbeendeterminedandissummarizedinTable5.

Inmostcases,however,additionalexperimentsarenecessarytoassignthestructuresunambiguously.First,permethylatedglycansobservedintheMALDI-TOFspectrumaresubjectedtonano-ESI-MS-MSexperimentstoassistinsequenceassignment.Whenanalyzedusingthissystem,permethylatedglycansformpreferentiallysodiatedmolecularadductsinadditiontoprotonatedmolecularadducts.BothtypesofmolecularspeciesproducedthiswaycanbesubjectedtoCIDtostudytheirfragmentationsusingMS/MS.Evenunderlowcollisionenergy(typically20eV),[M+H]+ionsfrompermethylatedglycansfragmentveryeasilyandglycosidiccleavagesareprominent28.Thefragmentsoriginatingfromboththereducingandthe

non-reducingendsoftheglycanyieldinformationonsequenceandbranching.Here,weshowanexampleofanano-ESI-MS-MS

2,792% Intensity50

2,4312,0401,8361,5801,7840

1,2001,9602,7202,0702,2442,6052,9663,2123,2413,6033,4163,480m/z

3,7773,6913,8654,0524,22,4134,5874,2405,000Figure8|Matrix-assistedlaserdesorption/ionizationtimeofﬂight(MALDI-TOF)massspectrumofpermethylatedN-glycansderivedfrom20mlofhumanserum.One-quarterofthepeptideN-glycosidaseF(PNGaseF)–releasedN-glycanswerepermethylated,puriﬁedonSep-PackC18andanalyzedusingMALDI-TOF-MS.OnlythestructuresofthemajorN-glycansaregiven.Aminorportionofthemono-fucosylatedglycanscarriesfucoseonanantennaratherthanthecore.Symbols:opensquare,galactose;closedsquare,N-acetylglucosamine;opentriangle,fucose;closedtriangle,N-acetyl-neuraminicacid;closedcircle,mannose.

NATUREPROTOCOLS|VOL.2NO.7|2007|1599

PROTOCOL

spectrumoftheprecursorion[M+2H]2+ofthepermethylatedTABLE5|Assignmentsofmolecular[M+Na]+ionsobservedinthe

matrix-assistedlaserdesorption/ionizationspectrumofpermethylatedN-glycan(NeuAcFucHex5HexNAc4)observedintheMALDI-TOF

N-glycansofhumanserum.spectrum(Fig.9).Thispermethylatedglycanwasanalyzedto

Signal(m/z)Assignmentgainmoreinformationontheantennaeandtodetermine

whetherthecorewasfucosylated.Thediagnosticionsat1,580Hex5HexNAc2+Na+1,784Hex6HexNAc2+Na+m/z376(NeuAc)and825(NeuAc-Hex-HexNAc)areconsistent

1,836Hex3HexNAc4dHex1+Na+withtheantennaNeuAc-Hex-HexNAcascleavageatthe

2,040Hex4HexNAc4dHex1+Na+HexNAcresidueisafavoredevent.Theotherdiagnosticion

2,070Hex5HexNAc4+Na+

atm/z4isconsistentwiththeantennaHex-HexNAc.These

2,244Hex5HexNAc4dHex1+Na+

ionsatm/z376,4and825,togetherwiththeabsenceof

2,431NeuAc1Hex5HexNAc4+Na+

ionsatm/z638(Hex-[Fuc]-HexNAc)and999(NeuAc-Hex-2,605NeuAc1Hex5HexNAc4dHex1+Na+

[Fuc]-HexNAc),suggestthatthemajorityofthefucoseresidue2,792NeuAc2Hex5HexNAc4+Na+isonthecoreoftheglycan.2,966NeuAc2Hex5HexNAc4dHex1+Na+Second,permethylatedPNGaseF–releasedN-glycansfrom3,212NeuAc2Hex5HexNAc5dHex1+Na+humanserumarehydrolyzedandanalyzedastheirPMAA3,241NeuAc2Hex6HexNAc5+Na+derivativesusingGC-MS.ThedatapresentedinTable4conﬁrm3,416NeuAc2Hex6HexNAc5dHex1+Na+

3,603NeuAc3Hex6HexNAc5+Na+theMALDI-TOF-MSandnano-ESI-MS-MSdataandgivefurther

3,691NeuAc2Hex7HexNAc6+Na+informationaboutN-glycanstructures.Forexample,compari-3,777NeuAc3Hex6HexNAc5dHex1+Na+sonoftherelativeabundanceofthe2-linkedManwith

3,865NeuAc2Hex7HexNAc6dHex1+Na+

2,4-linkedManand2,6-linkedManindicatesthatbiantennary

4,052NeuAc3Hex7HexNAc6+Na+

structuresaremoreabundantthantri-and/ortetraantennary

4,226NeuAc3Hex7HexNAc6dHex1+Na+

structures.Therelativeabundanceof4,6-linkedGlcNAc

4,413NeuAc4Hex7HexNAc6+Na+

conﬁrmsthatthemajorityofN-glycansarecore-fucosylated.4,587NeuAc4Hex7HexNAc6dHex1+Na+Thepresenceof3,4-linkedGlcNAcsuggeststhatsome

antennaearea3-fucosylated.Thepresenceofbisectedcomplex

structuresisindicatedby3,4,6-linkedMan.Figure7illustratestheprimaryfragmentsobtainedinEI-MSofthemostcommonPMAAsobtainedfromN-glycans,whichareveryusefulininterpretingdata.TheprimaryfragmentsobtainedinEI-MSoftheotherPMAAscanbeobtainedonline(http://www.ccrc.uga.edu/specdb/ms/pmaa/).

Third,theseexperimentscanbecomplementedbyexoglycosidasedigestionsthatdeﬁnetheanomericconﬁgurationsandconﬁrmtentativesequences.N-glycansreleasedbyPNGaseFweresubjectedtodigestionwitha-sialidase,b-galactosidaseandb-N-acetyl-glucosaminidaseandaliquotsweretakenaftereachdigestion,permethylatedandexaminedbyMALDI-TOF-MSafterreverse-phaseSep-PakC18puriﬁcation(Fig.10).Afterneuraminidasetreatment,theMALDI-TOF-MSdataindicatedthat,asexpected,allsialylatedcomponentspreviouslydescribedwereconvertedtoothercompoundsofreducedmolecularweightconsistentwiththelossofone,two,threeorfoursialicacidresidues.The[M+Na]+ionsprovideevidenceforthischange(Fig.10a).Thus,theNeuAcresiduesareinnormala-linkages.Comparisonoflinkagedatabeforeandaftersialidasetreatmentindicatedthatremovalofterminala-NeuAcresiduesisaccompaniedbythelossofthe3-and6-linkedGal(Table4)andthatthereisaconcomitantincreaseofterminalGal,indicatingthatsialicacidresiduesweremainlyattachedtothe6-positionofGalresidues.Aftera-galactosidasetreatment,theMALDI-TOF-MSspectrumoftheN-glycans(Fig.10b)wascharacterizedbythepresenceofeightmolecularionsatm/z1,580(Hex5HexNAc2),1,662(Hex3HexNAc4),1,784(Hex6HexNAc2),1,836(Fuc1Hex3-HexNAc4),1,908(Hex3HexNAc5),2,082(Fuc1Hex3HexNAc5),2,154(Hex3HexNAc6)and2,328(Fuc1Hex3HexNAc6).Therefore,the

ˆ-linkages.Afterb-N-acetylglucosa-N-glycanswereefﬁcientlyde-galactosylated,indicatingthattheGalresidueswereinnormala

minidasetreatment(Fig.10c),the[M+Na]+ionsatm/z1,662(Hex3HexNAc4),1,908(Hex3HexNAc5),2,154(Hex3HexNAc6)and2,328(Fuc1Hex3HexNAc6)disappeared,concomitantwiththeappearanceoftwonewionsatm/z1,171and1,345,which

©2007 Nature Publishing Group http://www.nature.com/natureprotocols1,400344.24Figure9|TandemelectrospraymassspectrometryofpermethylatedN-glycanobservedatm/z2,605inthematrix-assistedlaserdesorption–ionizationtimeofﬂight(MALDI-TOF)massspectrum.Thedoublychargedionatm/z1,292.8,correspondingtom/z2,605(NeuAcFucHex5HexNAc4),wasselectedforcollisionalactivation.Assignmentsofthefragmentionsareshownontheschematics.Symbols:opensquare,galactose;closedsquare,N-acetylglucosamine;opentriangle,fucose;closedtriangle,N-acetyl-neuraminicacid;closedcircle,mannose.

1600|VOL.2NO.7|2007|NATUREPROTOCOLS

1,2001,000Intensity1,759800600376.24002000200400600432.24.3793.4825.41,067.11,000m/z

1,2001,400[M+2H]2+3768251,292.81,276.81,552.81,6001,759.91,8002,000800PROTOCOL

2,500b

100

1,662% Intensity1,83650

1,908and2,0822,0821,5801,78401,0002,1542,3281,5002,000m/z2,5003,0003,500c

100

1,171% Intensity50

1,3451,7841,5801,83601,000

2,0821,500

2,000

m/z

2,500

3,000

3,500

Figure10|Positivematrix-assistedlaserdesorption–ionizationtimeofﬂight(MALDI-TOF)massspectraofpermethylatedN-glycansfromhumanserumafter(a)a-sialidasedigestion,(b)b-galactosidasedigestionand(c)b-N-acetylhexosaminidasedigestion.Thereleasedglycans(50%)fromhumanserum(20ml)weresequentiallydigestedwitha-sialidase,b-galactosidase,andb-N-acetylhexosaminidase.Aliquotsweretakenaftereachdigestion,permethylatedandanalyzedbyMALDI-TOF-MSfollowingSep-PakC18puriﬁcation.Symbols:opensquare,galactose;closedsquare,N-acetylglucosamine;opentriangle,fucose;closedtriangle,N-acetyl-neuraminicacid;closedcircle,mannose.

correspondrespectivelytoHex3HexNAc2andFuc1Hex3HexNAc2.Thesignalsatm/z1,580(Hex5HexNAc2)and1,784(Hex6Hex-NAc2)wereunaffectedbytheaboveexoglycosidasedigestions,aresultthatisconsistentwiththeassignmentofhighmannosestructuretotheseions.

ACKNOWLEDGMENTSThisresearchwassupportedbytheCentreNationaldela

´MixtedeRechercheCNRS/USTL8576),theMiniste`reRechercheScientiﬁque(Unite

´rieur.ThemassspectrometryfacilitydelaRechercheetdel’EnseignementSupe

´gionusedinthisstudywasfundedbytheEuropeanCommunity(FEDER),theRe

´desSciencesetTechnologiesNord-PasdeCalais(France)andtheUniversite

deLille.

COMPETINGINTERESTSSTATEMENTTheauthorsdeclarenocompetingﬁnancialinterests.

Publishedonlineathttp://www.natureprotocols.com

Reprintsandpermissionsinformationisavailableonlineathttp://npg.nature.com/reprintsandpermissions

1.Ohtsubo,K.&Marth,J.D.Glycosylationincellularmechanismsofhealthand

disease.Cell126,855–867(2006).

2.Varki,A.Biologicalrolesofoligosaccharides:allofthetheoriesarecorrect.

Glycobiology3,97–130(1993).

3.Hellerqvist,C.G.LinkageanalysisusingLindbergmethod.MethodsEnzymol.193,

554–573(1990).

NATUREPROTOCOLS|VOL.2NO.7|2007|1601

PROTOCOL

4.Dell,A.FABmassspectrometryofcarbohydrates.Adv.Carbohydr.Chem.Biochem.

45,19–72(1987).

5.Faid,V.,Evjen,G.,Tollersrud,O.K.,Michaslki,J.C.&Morelle,W.Sitespeciﬁc

glycosylationanalysisofthebovinelysosomalalphamannosidase.Glycobiology16,440–461(2006).

6.Morelle,W.etal.GalactomannoproteinsofAspergillusfumigatus.Eukaryot.cell4,

1308–1316(2005).

7.Morelle,W.,Jimenez,J.C.,Cieniewski-Bernard,C.,Dei-Cas,E.&Michalski,J.C.

CharacterizationoftheN-linkedglycansofGiardiaintestinalis.Glycobiology15,549–559(2005).

8.Morelle,W.,Donadio,S.,Ronin,C.&Michalski,J.C.CharacterizationofN-glycans

ofrecombinanthumanthyrotropinusingmassspectrometry.RapidCommun.MassSpectrom.20,331–345(2006).

9.Morelle,W.etal.Massspectrometricapproachforscreeningmodiﬁcationsoftotal

serumN-glycomeinhumandiseases:applicationtocirrhosis.Glycobiology,16,sl281–293(2006).

oc10.Faid,V.,Chirat,F.,Seta,N.,Foulquier,F.&Morelle,W.Arapidmassspectrometric

otostrategyforthecharacterizationofN-andO-glycanchainsinthediagnosisofrpdefectsinglycanbiosynthesis.Proteomics7,1800–1813(2007).

er11.Tarentino,A.L.&Plummer,T.H.Enzymaticdeglycosylationofasparagine-linked

utglycans:puriﬁcation,properties,andspeciﬁcityofoligosaccharide-cleavinganenzymesfromFlavobacteriummeningosepticum.MethodsEnzymol.230,44–57/m(1994).

oc12.Tretter,V.,Altmann,F.&Marz,L.Peptide-N4-(N-acetyl-b-glucosaminyl)

.erasparagineamidaseFcannotreleaseglycanswithfucoseattacheda1–3totheutasparaginelinkedN-acetylglucosamineresidue.Eur.J.Biochem.199,7–652an(1991).

w.13.Carlson,D.M.Structuresandimmunochemicalpropertiesofoligosaccharides

wwisolatedfrompigsubmaxillarymucins.J.Biol.Chem.243,616–626(1968).//:14.Ku¨ster,B.,Wheeler,S.F.,Hunter,A.P.,Dwek,R.A.&Harvey,D.J.Sequencingof

ptN-linkedoligosaccharidesdirectlyfromproteingels:in-geldeglycosylationthfollowedbymatrix-assistedlaserdesorption/ionizationmassspectrometryand pnormal-phasehigh-performanceliquidchromatography.Anal.Biochem.250,u82–101(1997).

roG15.Wheeler,S.F.&Harvey,D.J.Extensionofthein-gelreleasemethodforstructural

ganalysisofneutralandsialylatedN-linkedglycanstotheanalysisofsulfatednglycans:applicationtotheglycansfrombovinethyroid-stimulatinghormone.ihsAnal.Biochem.296,92–100(2001).

il16.Patel,T.etal.Useofhydrazinetoreleaseinintactandunreducedformboth

buN-andO-linkedoligosaccharidesfromglycoproteins.Biochemistry32,679–693P(1993).

er17.Huang,Y.,Mechref,Y.&Novotny,M.V.Microscalenonreductivereleaseof

utO-linkedglycansforsubsequentanalysisthroughMALDImassspectrometryaNandcapillaryelectrophoresis.Anal.Chem.73,6063–6069(2001).

718.Huang,Y.,Konse,T.,Mechref,Y.&Novotny,M.V.Matrix-assistedlaserdesorption/

00ionizationmassspectrometrycompatiblebeta-eliminationofO-linked2oligosaccharides.RapidCommun.MassSpectrom.16,1199–1204(2002).

glycansreleaseddirectlyfromglycoproteinsingelsusingin-gelreductiveb-elimination.Proteomics5,2936–2946(2006).

20.Colangelo,J.&Orlando,R.On-targetexoglycosidasedigestions/MALDI-MSfor

determiningtheprimarystructuresofcarbohydratechains.Anal.Chem.71,1479–1482(1999).

21.Zaia,J.Massspectrometryofoligosaccharides.MassSpectrom.Rev.23,161–227

(2004).

22.Morelle,W.,Canis,K.,Chirat,F.,Faid,V.&Michalski,J.C.Theuseofmass

spectrometryfortheproteomicanalysisofglycosylation.Proteomics6,3993–4015(2006).

23.Powell,A.K.&Harvey,D.J.StabilizationofsialicacidsinN-linked

oligosaccharidesandgangliosidesforanalysisbypositiveionmatrix-assistedlaserdesorption/ionizationmassspectrometry.RapidCommun.MassSpectrom.10,1027–1032(1996).

24.Ciucanu,I.&Kerek,F.Asimpleandrapidmethodforthepermethylationof

carbohydrates.Carbohydr.Res.131,209–217(1984).

1602|VOL.2NO.7|2007|NATUREPROTOCOLS

25.Ciucanu,I.&Costello,C.E.Eliminationofoxidativedegradationduringtheper-O-methylationofcarbohydrates.J.Am.Chem.Soc.125,16213–16219(2003).26.Kang,P.,Mechref,Y.,Klouckova,I.&Novotny,M.V.Solid-phasepermethylationof

glycansformassspectrometricanalysis.RapidCommun.MassSpectrom.19,3421–3428(2005).

27.Wong,N.K.etal.Characterizationoftheoligosaccharidesassociatedwiththe

humanovariantumormarkerCA125.J.Biol.Chem.278,28619–28634(2003).28.Morelle,W.,Faid,V.&Michalski,J.C.Structuralanalysisofpermethylated

oligosaccharidesusingelectrosprayionizationquadrupoletime-of-ﬂighttandemmassspectrometryanddeutero-reduction.RapidCommun.MassSpectrom.18,2451–24(2004).

29.Harvey,D.J.Collision-inducedfragmentationofunderivatizedN-linked

carbohydratesionizedbyelectrospray.J.MassSpectrom.35,1178–1190(2000).

30.Morelle,W.&Michalski,J.C.Themassspectrometricanalysisofglycoproteinsand

theirglycanstructures.Curr.Anal.Chem.1,27–55(2005).

31.Domon,B.&Costello,C.E.Asystematicnomenclatureforcarbohydrate

fragmentationsinFAB-MS/MSspectraofglycoconjugates.Glycoconj.J.5,397–409(1988).

32.Viseux,N.,deHoffmann,E.&Domon,B.Structuralassignmentofpermethylated

oligosaccharidesubunitsusingsequentialtandemmassspectrometry.Anal.Chem.70,4951–4959(1998).

33.Jacob,G.S.&Scudder,P.Glycosidasesinstructuralanalysis.MethodsEnzymol.

230,280–299(1994).

34.Geyer,H.,Schmitt,S.,Wuhrer,M.&Geyer,R.Structuralanalysisof

glycoconjugatesbyon-targetenzymaticdigestionandMALDI-TOF-MS.Anal.Chem.71,476–482(1999).

35.Mechref,Y.&Novotny,M.V.MassspectrometricmappingandsequencingofN-linkedoligosaccharidesderivedfromsubmicrogramamountsofglycoproteins.Anal.Chem.70,455–463(1998).

36.Albersheim,P.,Nevins,D.J.,English,P.D.&Karr,A.Amethodfortheanalysisof

sugarsinplantcellwallpolysaccharidesbygas-liquidchromatography.Carbohydr.Res.5,340–345(1967).

37.Geyer,R.&Geyer,H.Saccharidelinkageanalysisusingmethylationandother

techniques.MethodsEnzymol.230,86–108(1994).

38.Hellerqvist,C.G.LinkageanalysisusingLindbergmethod.MethodsEnzymol.193,

554–573(1990).

39.Young,N.M.etal.StructureoftheN-linkedglycanpresentonmultiple

glycoproteinsintheGram-negativebacterium,Campylobacterjejuni.J.Biol.Chem.277,42530–42539(2002).

40.Liu,X.etal.Massspectrometry-basedglycomicsstrategyforexploring

N-linkedglycosylationineukaryotesandbacteria.Anal.Chem.78,6081–6087(2006).

41.Hakomori,S.Arapidpermethylationofglycolipid,andpolysaccharidecatalyzed

bymethylsulﬁnylcarbanionindimethylsulfoxide.J.Biochem.(Tokyo)55,205–208(19).

42.Fournet,B.,Strecker,G.,Leroy,Y.&Montreuil,J.Gas-liquidchromatographyand

massspectrometryofmethylatedandacetylatedmethylglycosides.Applicationtothestructuralanalysisofglycoproteinglycans.Anal.Biochem.116,4–502(1981).

43.Packer,N.H.,Lawson,M.A.,Jardine,D.R.&Redmond,J.W.Ageneralapproachto

desaltingoligosaccharidesreleasedfromglycoproteins.Glycoconj.J.15,737–747(1998).

44.Papac,D.I.,Wong,A.&Jones,A.J.S.Analysisofacidicoligosaccharidesand

glycopeptidesbymatrix-assistedlaserdesorption/ionizationtime-of-ﬂightmassspectrometry.Anal.Chem.68,3215–3223(1996).

45.Sagi,D.,Kienz,P.,Denecke,J.,Marquardt,T.&Peter-Katalinic,J.

GlycoproteomicsofN-glycosylationbyin-geldeglycosylationandmatrix-assistedlaserdesorption/ionisation-timeofﬂightmassspectrometrymapping:

applicationtocongenitaldisordersofglycosylation.Proteomics5,26–2701(2005).

46.Cooper,C.A.,Gasteiger,E.&Packer,N.H.GlycoMod—asoftwaretoolfor

determiningglycosylationcompositionsfrommassspectrometricdata.Proteomics1,340–349(2001).

因篇幅问题不能全部显示，请点此查看更多更全内容

查看全文