Download Chapter 28 Discovery and Classification of Glycan

Chapter 28
Discovery and Classification of GlycanBinding Proteins
Essentials of Glycobiology 3rd edition
DiscoveryandClassificationofGlycan-BindingProteins
Glycansserveavarietyofbiologicalfunctionsbyvirtueoftheirmass,shape,charge,
andotherphysicalproperties.Manyoftheirmorespecificbiologicalrolesare
mediatedviarecognitionbycomplementaryglycan-bindingproteins(GBPs).Nature
hastakenadvantageofthediversityofglycansexpressedinorganismsbyevolving
proteinmodulestorecognizediscreteglycansthatmediatespecificphysiologicalor
pathologicalprocesses.Thischapterprovidesageneralclassificationandoverview
ofnaturallyoccurringGBPs,thehistoryoftheirdiscovery,someoftheirbiological
functionsandwaysinwhichnewGBPsareidentified.Chaptersthatfollowdescribe
theanalysisofglycan–proteininteractions(Chapter29),thephysicalprinciples
involved(Chapter30)andthestructuresandbiologicalfunctionsofseveralGBPs
subclasses(Chapters31-38).
TWODISTINCTCLASSESOFGBPs
GBPsarefoundinalllivingorganisms,andfallintotwooverarchinggroups–lectins
andsulfatedglycosaminoglycan(GAG)-bindingproteins(onlineAppendix28A).
Lectinsarefurtherclassifiedintoevolutionarily-relatedfamiliesidentifiedby
“carbohydrate-recognitiondomains”(CRDs)basedonprimaryaminoacidand/or
three-dimensionalstructuralsimilarities(Figure28.1).CRDscanexistasstandaloneproteinsorasdomainswithinlargermulti-domainproteins.Theytypically
recognizeterminalgroupsonglycans,whichfitintoshallowbutwell-defined
bindingpockets(Chapters29,30).Incontrast,proteinsthatbindtosulfatedGAGs
(heparan,chondroitin,dermatanandkeratansulfates,Chapter17)dosoviaclusters
ofpositivelychargedaminoacidsthatbindspecificarrangementsofcarboxylicacid
andsulfategroupsalongGAGchains(Chapter38).Mostoftheseproteinsare
evolutionarilyunrelated.GBPsthatbindtothenon-sulfatedGAGhyaluronicacid
(hyaladherins)shareanevolutionarilyconservedfoldthatbindstoshortsegments
oftheinvarianthyaluronanrepeatingdisaccharide(Chapter16),soarebest
classifiedaslectinsratherthangroupedwithsulfatedGAG-bindingproteins.The
restofthischapterconsiderslectins,differentfamiliesofwhicharedetailedin
Chapters31-37.
DISCOVERYANDHISTORYOFLECTINS
Lectinswerediscoveredinplantsin1888whenextractsofcastorbeanseedswere
foundtoagglutinateanimalredbloodcells.Subsequentlyseedsofmanyplantswere
foundtocontainsuch"agglutinins",laterrenamedlectins(Latinfor“select”)when
theywerefoundtodistinguishhumanABObloodgroups(Chapter14)importantfor
bloodtransfusions.Lectinsareparticularlycommonintheseedsofleguminous
plantsandthese"L-type"lectins,includingconcanavalinAandphytohemagglutinin,
havebeenextensivelystudied.Althoughtheirspecificglycan-bindingactivities
makeplantlectinsextremelyusefulscientifictools,theirbiologicalfunctionsin
plantsremainmostlyunknown.
Thefirstanimallectindiscoveredwastheasialoglycoproteinreceptor(ASGPR)
identifiedbyAnatolMorellandGilbertAshwellinthelate1960’sduringtheir
investigationsoftheturnoverofaserumglycoprotein,ceruloplasmin.Likemost
glycoproteinscirculatinginblood,ceruloplasminhascomplexN-glycanswithsialic
acidtermini.Toprepareradiolabeledceruloplasmin,theterminalsialicacidswere
removed,leavinganexposedgalactose.Surprisingly,asialoceruloplasminhada
circulationhalf-life(inrabbits)ofminuteswhereasintactceruloplasminremained
inthebloodforhours.GlycoproteinswithexposedGalresidueswererapidly
clearedintolivercellsviaanendocyticcellsurfacereceptorthatspecificallybound
toterminalβ-linkedGalorGalNAc.ASGPRwaspurifiedbyaffinitychromatography
usingacolumnofimmobilizedasialoglycoprotein.
Otherglycan-specificreceptorsinvolvedinglycoproteinclearanceandtargeting
weresubsequentlydiscovered,includingmannose6-phosphatereceptorsfor
targetinglysosomalenzymestothelysosomesandmannosereceptorsthatclear
glycoproteinswithterminalmannoseorGlcNAcresiduesfromtheblood.Small
solublelectinsspecificforβ-linkedgalactose(nowcalled“galectins”,Chapter36)
wereisolatedbyaffinitychromatographyinextractsfrommanybiologicalsources
rangingfromtheslimemoldDictyosteliumdiscoideumtomammaliantissues.Bythe
1980’s,theconceptofvertebratelectinsthatrecognizespecificglycanswaswell
established.Althoughthefirstanimallectinsidentifiedwerespecificforendogenous
glycans,manylectinsspecificforexogenousglycansofmicroorganismswerelater
found.Lectinsrecognizingexogenousglycansincludesolubleproteinsthatcirculate
inthebloodofmanyspeciesaswellasmembrane-boundreceptorsoncellsofthe
immunesystem.
Lectinsarealsowidespreadinmicroorganisms,althoughtheytendtobecalledby
othernamessuchashemagglutininsandadhesins.Theinfluenzavirus
hemagglutinin,whichbindstosialicacidonhostcells(Chapter15)wasthefirstGBP
isolatedfromamicroorganism.Theviralhemagglutinins,likemanyplantlectins,
canagglutinateredbloodcells.Manybacteriallectinshavebeendescribed.Theyfall
intotwogeneralclasses:adhesinsonbacterialsurfacesthatrecognizeglycanson
hostcellmembraneglycolipidsorglycoproteinstofacilitatebacterialadhesionand
colonization,andsecretedbacterialtoxins(Chapter37).
DISCOVERYOFSULFATEDGAG-BINDINGPROTEINS
AlargegroupofGBPsthatdefyclassificationbasedonsequenceorstructure
recognizesulfatedGAGs(Chapter38).Thebest-studiedexampleistheinteractionof
heparinwithantithrombin.Heparinwasdiscoveredin1916byJayMcLean,a
medicalstudent,butitwasnotuntil1939thatheparinwasshowntobean
anticoagulantinthepresenceof“heparincofactor”,whichwasthenidentifiedas
antithrombininthe1950s.ManyothersulfatedGAG-bindingproteinswerelater
discoveredbyaffinitychromatographyoncolumnsofimmobilizedheparin.Growth
factorsandcytokinesbearingclustersofpositivelychargedaminoacidsalongtheir
proteinsurfaceinteractwithsulfatedGAGsinalooserfashion—i.e.,theydonot
alwaysshowthehighspecificityseenwithantithrombin.However,insomecases,
specificGAGsequencesmediatetheformationofhigher-ordercomplexes,actingas
atemplateforoligomerizationorpositioningofproteinssuchasFGFanditscell
surfacereceptor.
MAJORBIOLOGICALFUNCTIONSOFGBPs
GBPsfunctionincommunicationbetweencellsinmulticellularorganismsandin
interactionsbetweenmicrobesandhostsandcanalsobeinvolvedinbinding
growthfactorsorcytokines.Theseinteractionscantakevariousforms,resultingin
movementofmolecules,cells,andinformation.
Trafficking,targetingandclearanceofproteins
Directingmovementofglycoproteinswithinandbetweencellsisacommon
functionforlectinsinmanyorganisms.Ineukaryoticcells,includingyeastaswellas
highereukaryotes,severalgroupsoflectinsareimportantinglycoprotein
biosynthesisandintracellularmovement(Chapter39).Intheendoplasmic
reticulum(ER),twolectins,calnexinandcalreticulinbindmonoglucosylatedhigh
mannoseglycanspresentonnewlysynthesizedglycoproteins,formingpartofa
qualitycontrolsystemforproteinfolding.Bindingtocalnexinorcalreticulinkeeps
proteinsintheERuntiltheyarecorrectlyfolded.OthergroupsoflectinsintheER,
includingM-typelectinsandproteinscontainingmannose6-phosphatereceptor
homologydomainstakepartintheprocessofER-associatedglycoprotein
degradation(ERAD),bindingpartiallyprocessedhighmannoseglycanson
terminallymisfoldedglycoproteins,causingthemtoberetrotranslocatedintothe
cytoplasmfordeglycosylation,followedbydegradationintheproteasome.Oneof
thebestcharacterizedfunctionsofGBPsisindeliveryofnewlysynthesized
lysosomalenzymesfromthetrans-GolgitolysosomesbyP-typelectins(Chapter33)
thatrecognizemannose6-phosphateresiduesthathavebeenaddedtoN-glycanson
lysosomalenzymesintheGolgiapparatus,targettingthemtoendosomesforfusion
withlysosomes.
Oncereleasedfromcells,glycoproteinscanalsobetakenupfordegradationin
lysosomes.Asnotedabove,theASGPRonmammalianhepatocytescontrols
turnoverofmanyserumglycoproteinsbyrecognitionofterminalGalorGalNAc
residues.Similarly,themannosereceptoronmacrophagesandsinusoidalcellsof
theliverbindsandclearsglycoproteinswitholigomannoseN-glycansthatare
releasedfromcellsduringinflammationandtissuedamage.
NotallGBP-mediatedtargetingleadstodegradation.Glycan-bindingsubunitsof
secretedbacterialandplanttoxinstargetthemtoglycolipidsoncellsurfacesand
facilitateentryofthetoxinsintocells(Chapter37).Manyenzymescontainglycanbindingdomainsthatbringanotherdomainwithenzymeactivityintoclose
proximitywithitssubstrates.Onenotablegroupincludesbacterialcellulasesin
whichcellulose-bindingmodulespositiontheenzymaticdomainforoptimal
degradationofcellulosefibers.Usingasimilarprinciple,GalNAc-bindingdomainsin
polypeptide-N-acetylgalactosaminyltransferasesthatinitiateO-linkedglycosylation
inanimalspositiontheseenzymestoaddfurtherGalNAcresiduestoregionsof
polypeptidesthatalreadybearO-glycans(Chapter10).
Celladhesion
Distinctiveglycansonthesurfacesofdifferentcells,botheukaryoticand
prokaryotic,makethemtargetsforGBPs.Bindingofglycansonthesurfaceofone
cellbyGBPsonanothercellcaninducerecognitionandadhesion,whereas
crosslinkingglycansondifferentcellsbymultivalentsolubleGBPsprovidesan
alternativemechanism.Suchinteractionsareexploitedinspecializedsituations
exemplifiedbytransientcontactsbetweenmovingcells.Theselectins,three
receptorsthatfunctionininteractionsbetweenwhitebloodcells,plateletsand
endothelia,providethebestcharacterizedexampleoflectin-glycaninteractionsin
cell-celladhesion(Chapter34).Forexample,L-selectinonlymphocytesbinds
glycansonthespecializedendothelialcellsoflymphnodestoinducelymphocytehoming,whereincirculatinglymphocytesleavethebloodstreamandenterthe
lymphnode.OthermammalianGBPsthatmediatebindingofcellstoeachotheror
thatcrosslinkligandsonthesamecellsurfaceincludeSiglecs(Chapter35)and
galectins(Chapter36).Lectinsinmulticellularorganismsalsoforminteractions
betweencellsandtheextracellularmatrixandsupporttheorganizationofmatrix
components.Forexample,proteinscontaining“linkmodules”thatbindspecifically
tohyaluronanincartilage(andothertissues)areessentialforstructuringthe
extracellularmatrix(Chapter38)andotherextracellularproteinsbindtosulfated
GAGstoorganizecell-cellandcell-matrixinteractions(Chapter38).
Manybacteriaalsouselectinstoadheretoglycansonhostcellsinsituationsin
whichtheywouldotherwisegetwashedaway.Thelectinsareusuallypresentatthe
endsoflongstructurescalledpiliorfimbriaethatprojectfromthesurfaceofthe
bacteria(Chapter37).Adhesioncanbepartoftheinfectionprocess.Forexample,a
mannose-specificadhesinonpathogenicstrainsofEscherichiacolithatcause
urinaryinfectionsbindstoepithelialcellsoftheurinarytract.Otherglycan-protein
interactionsbetweenhostcellsandbacteriaprovideanormalmechanismofcoexistence.Severalbacterialspeciesthatarepartofthenormalgutfloraincluding
non-pathogenE.coliuseadhesinstobindtoglycolipidspresentoncellsliningthe
largeintestine.
Immunityandinfection
Manylectinsareinvolvedinimmuneresponses,in“lower”vertebratesand
invertebratesaswellasinmammals.Differencesinglycansonhostandmicrobial
cellsurfacesarecommonlythebasisforinnateimmuneresponses.Phagocytosisisa
commonoutcomeofthebindingofglycan-specificreceptorsonmacrophagesto
glycanscommontobacteria,fungiandviruses.Otherlectinscirculatingintheblood,
suchasserummannose-bindingproteinandficolins,bindtopathogencellsurfaces
andactivatethecomplementcascade,leadingtocomplement-mediatedkilling.
Bindingofglycanstolectinsonimmunecellscanalsotriggerintracellularsignaling
thatactivatesorsuppressescellularresponses.Receptorsthatrecognizeselfglycanssuchassialicacid,aswellasseveralthatarespecificforglycans
characteristicofmicro-organismscaninitiatesuchsignaling.Forexample,binding
ofα2-6linkedsialicsacidtoCD22,amemberoftheSiglecfamilyofvertebrate
lectinsfoundonBlymphocytes,initiatessignalingthatinhibitsactivationtoprevent
self-reactivity(Chapter35).Incontrast,bindingoftrehalosedimycolate,aglycolipid
foundinthecellwallofMycobacteriumtuberculosistothemacrophageC-typelectin
mincle,inducesasignalingpathwaythatcausesthemacrophagetosecrete
proinflammatorycytokines.
Finally,virusesoftenuseGBPstoattachtohostcellsduringinfection(Chapter37).
Proteinsonvirussurfaces,includingthoseoninfluenzavirus,reovirus,Sendaivirus,
andpolyomavirus,bindtosialicacids.Inadditiontobringingthevirusintocontact
withtheircelltargets,thesehemagglutininstypicallyinducemembranefusion,
facilitatingvirusentryanddeliveryofnucleicacidsintothecytosol.Glycan-binding
receptorsonvirusesareoftenhighlyspecificforaparticularlinkage;human
influenzavirusbindstosialicacidslinkedα2-6toGal,whereasbirdinfluenzavirus
bindstoα2-3linkedsialicacid.Otherviruses,suchasherpessimplexvirus,have
GAG-bindingproteinsthatbindtoheparansulfateproteoglycansoncellsurfaces.
ORGANIZATIONOFLECTINS
Animportantconceptinidentifying,definingandclassifyinglectinsisthatsugarbindingactivityisembodiedindiscreteproteinmodulesordomains,referredtoas
carbohydrate-recognitiondomains(CRDs).CRDsaretypicallyindependentlyfolding
segmentsofproteins;oftenonecanseparatethesugar-bindingactivityfromother
activitiesoftheproteinbyexpressingitsCRDinisolation.Insomecases,theCRDs
constitutetheentireGBP(Figure28.2).
WhenalectiniscomprisedsimplyofitsCRD,itsfunctionsoftenaredependenton
multivalency,whichendowslectinswiththeabilitytocross-linksugar-containing
structures.Thisarrangementexplainstheabilityofmanyplantlectinstoagglutinate
cellsandtoclusterglycoproteinsoncellsurfaces,whichcaninducemitogenesis.
OtherGBPsthatfunctionthiswayincludethegalectins,whichcanbridgeglycanson
onecellsurfaceorbetweencells.Sometimesotheractivitiesareencodedwithinthe
structureofthesamedomainthatbindssugars;somecytokinescomprisedofa
singlefoldeddomainmayhavedistinctsitesforbindingglycansandothertarget
receptors.Morecommonly,otheractivitiesoflectinsresideinseparatemodulesin
multi-domainproteins(Figure28.2).Sucharrangementsarewidespreadandthe
domainsassociatedwithCRDsperformmanydifferentfunctions,includingbinding
othertypesofligands,performingenzymaticreactions,anchoringproteinsto
membranesanddirectingoligomerization.GBPsoftencontainmultiplemodules,
combiningseveralfunctionsinoneprotein.
Membraneanchorsinlectinscantakemultipleforms,buttheyoftenspanthe
membrane,linkingextracellularCRDswithcytoplasmicdomains.Thisarrangement
facilitatestheflowofinformationbetweenglycan-bindingsitesontheextracellular
surfaceandthecytoplasm.Simplesequencemotifsinthecytoplasmicdomainsof
transmembranelectinsoftencontroltraffickingofreceptorsandtheirboundglycan
ligands.Commonfunctionsofsuchintracellularmovementsareinternalizationof
cellsurfacereceptors,directingboundligandstoendosomesandlysosomes,and
movementthroughintracellularcompartmentssuchastheendoplasmicreticulum
andGolgiapparatustothecellsurface.Flowofinformationintheoppositedirection
canleadtostimulationofsignalingcomplexesonthecytoplasmicsideofthe
membraneinresponsetobindingofglycansatthecellsurface.
Clusteringofglycan-bindingsites(multivalency)isoftencriticaltobothrecognition
andbiologicalfunctionsoflectins.Clusteringofsitesisachievedindifferentways,
byformationofsimpleoligomersofCRDs,asaresultofthepresenceofmultiple
CRDsinasinglereceptorpolypeptideandthroughassociationofCRD-containing
polypeptidesthroughindependentoligomerizationdomains.Someoligomersare
stable,whileothers,suchasthoseformedbysomegalectins,areinequilibriumwith
monomers.Thesearrangementsfacilitatemultivalentbindingtoincreaseavidity
anddirectthegeometricalarrangementofbindingsites.MultipleCRDsmayfacein
thesamedirectionforsurfacerecognitionorinoppositedirectionstofacilitate
crosslinking.MultivalentCRDsmayhavefixedspacingorflexiblespacingto
accommodatedifferenttargetglycans.Insomecases,oligomerizationdomainsalso
formstructuralfeatures,servingsasstalksthatprojectCRDsfromthecellsurface.
Oligomerizationdomainscanalsoembodyotherfunctions,suchastheproteasebindingsitesinthecollagen-likedomainsofmannose-bindingprotein.
CLASSIFICATIONOFLECTINSBASEDONSTRUCTURAL
SIMILARITIES
ItisconvenienttoclassifylectinsbasedonthestructuresoftheCRDsthatthey
contain(Figure28.3).CRDsarefoundinalargenumberofdifferentstructural
categories,indicatingthatmanydifferentproteinfoldscanaccommodateglycan
binding(Chapter30).Basedonthisobservation,sugar-recognitionmusthave
evolvedindependentlymanytimesandthediversityofCRDstructuresmusthave
arisentoaddressadiversityoffunctions.
GBPsappearacrossallkingdomsoflife,butthetypesoflectinsineachkingdom
varyconsiderably.Severalfamiliesappearinbothprokaryotesandeukaryotes,but
theirdistributionssuggestdifferentevolutionaryhistories.Themalectindomain,
althoughconservedinstructureandwidelydistributedinprokaryotes,plantsand
animals,isfoundinproteinswithdistinctdomainorganizationanddifferent
functionsinthethreegroups.Animalmalectinisamembrane-anchoredCRDofthe
endoplasmicreticulumthatbindsN-linkedglycansduringglycoprotein
biosynthesis.Inplants,themalectinCRDisexpressedatthecellsurfaceandis
linkedtoacytoplasmickinasedomain.BacterialmalectinsconsistofCRDs
associatedwithglycohydrolasedomains.Similarly,R-typeCRDs(Chapter31)in
plantsformthecellsurface-bindingcomponentoftoxinssuchasricinandarelinked
toglycohydrolasegenesinbacteria,butinanimalstheyappearintwodistinct
contexts:inpolypeptide-N-acetylgalactosaminyltransferasesthatinitiateO-GalNAc
glycans(Chapter10)andinthemannosereceptorfamily.AlthoughtheseCRDshave
beenadaptedtoservedifferentfunctionsindifferentkingdoms,asugar-binding
functionappearstohaveevolvedearlyandbeenpreservedinsubsequentlineages.
IncontrasttoCRDswithbroadevolutionarydistribution,twoothergroupsof
lectinshavesporadicdistributions.B-lectindomainsarebroadlydistributedin
bacteriainassociationwithhydrolasedomains,arefoundasisolatedortandem
CRDsinmonocotplantsbutnotinotherplants,inbonyfishesbutnotinother
animals,andinandsomefungi.TheF-typelectinsappearinbacteriaandina
limitednumberofanimalspecies.Inthesecases,thepresenceofrelateddomainsin
evolutionarilydistantspeciesmayreflectlateralgenetransferratherthanthe
presenceofaprecursorlectininthedistantcommonancestorthattheyshare.A
differentpatternofevolutionisobservedforPA14domains,theonlyothertypeof
CRDfoundinbothbacteriaandeukaryotes.AlthoughthePA14foldisrelatively
widespread,suggestingthatitoriginatedearlyandwasretainedacrossspecies,only
asubsethavebeenshowntohavesugar-bindingactivity:CRDsassociatedwith
bacterialglycohydrolasesandinadhesinsandflocculationfactorsonthesurfaceof
yeast.
Theintracellularsortinglectinsmentionedearlier,suchascalnexin,calreticulin,and
M-typelectins,arethemostbroadlydistributedlectinsthatevolvedfromacommon
eukaryoticancestor.Theirdistributionandtheconservationoftheirfunctions
probablyreflectanancientandconservedroleinintracellulartraffickingof
glycoproteinsineukaryotes.TwoothergroupsofCRDsappeartobefoundin
metazoansbutnotsimplereukaryotes.TheL-typeCRDshavedivergedinfunction
betweenanimals,wheretheyfunctioninintracellularglycoproteinsortingand
trafficking,andplants,wheretheyserveaprotectivefunction(Chapter32).
Chitinase-likesugar-bindingdomainsacrossarangeofspeciesretaintheabilityto
bindpolymersofGlcNAc,buttheirbiologicalfunctionsarenotwellunderstood,soit
isuncleariftheyhavesharedrolesinplantsandanimals.
Inadditiontothewidelydistributedfamilies,certainCRDfamiliesareevolutionarily
restricted.Inadditiontoanimal-specificandvertebrate-specificlectingroups,there
arealsogroupssuchastheI-typelectinsfoundonlyinmammals(Chapter35).The
patternofevolutionofanimal-specificlectinsvaries.Galectinsseemtobesimilarin
organizationinvertebratesandinvertebratesanditmaybepossibletoidentify
orthologsinquitediversespecies(Chapter36).Incontrast,C-typeCRDshave
undergoneindependentradiationinvertebratesandinvertebrates,andidentifying
orthologsevenbetweenmouseandhumanproteinsinsomecasesisdifficult
(Chapter34).Ofthetwelvedifferentproteinfoldsfoundinplantlectins,nineappear
tobeuniquetoplants.Itisalsonoteworthythatvirusesseemtohavedeveloped
theirownapproachestobindingglycansratherthanborrowingfromhosts(Chapter
37).
InadditiontofamiliesofproteinsthatshareevolutionarilyrelatedCRDs,thereare
individualproteinsthatbindsugarsthroughdomainsthatarenotrelatedtoCRDsin
otherproteins.Examplesincludeproteinswithdedicatedsugarbindingdomains,
suchassomelamininGdomains,whichrecognizeglycansonα-dystroglycan
(Chapter45),pentraxins,whichbindmodifiedandphosphorylatedsugars,and
macrophageαMβ2integrin,whichbindsfungalglucansandexposedGlcNAcresidues
onglycoproteins.Otherproteinsbindtosugarsthroughdomainsthatalsohave
otherligands:annexinVbindsbisectingGlcNAcresiduesaswellasphospholipids
andseveralcytokineshavebeenreportedtobindsugarsaswellastheirtarget
receptors.Sulfated-GAGbindingproteinshavealsolargelyevolvedbyconvergent
evolution.
IDENTIFYINGGBPsBYBIOLOGICALANDBIOCHEMICALFUNCTION
ANDSTRUCTURALSIMILARITY
Therearemultiplewaysinwhichglycanrecognitioncanbeimplicatedinspecific
biologicalprocesses.Onecommonapproachistodemonstratetheabilityofsimple
monosaccharidesorsmallglycanstocompetewithaprocess.Informationcanoften
alsobegainedbymodifyingsugarsoncellsandglycoproteinswithenzymesthat
addorremovesugars,bygeneticmanipulation,andbychemicalinhibitorsofglycan
metabolism.Thesestrategieshaveprovidedinformationabouttheglycansinvolved,
forexamplethoseneededforvirusortoxinbindingorthoserequiredfor
endocytosisofglycoproteins.Basedonthisinformation,itisthenpossibletolook
forGBPsthattargettheseparticularsugarsandwhichcanthenbelinkedtothe
biologicalprocess.
Theabilitytobindspecificsugars,assessedinvariousbiochemicalassays,hasoften
beenthebasisfordirectidentificationofnovelGBPswithoutreferencetoa
particularbiologicalfunction.Inadditiontoformingabasisforbindingand
competitionassays,thebindingactivityiscommonlyusedasameansofisolating
theseproteinsbyemployingaffinitychromatographyonappropriateimmobilized
glycanligands.Awidevarietyofmethodsforcouplingmonosaccharidesand
complexglycanstocreateaffinityresinshavebeendeveloped.Asmentionedabove,
manysulfated-GAGbindingproteinshavebeendiscoveredbyaffinity
chromatographyonimmobilizedGAGchains.Alimitationoftheseapproachesis
thatbindingactivitydoesnotdirectlyindicateabiologicalfunctionandtherolesof
manywell-characterizedGBPshavenotbeenfullydetermined.
Theobservationthatmanylectinsfallintostructuralfamiliesprovidesan
alternativewaytoidentifynovelGBPsthroughanalysisofproteinsequences.
SequencemotifscharacteristicofCRDsareroutinelyusedtoscreensequencesfrom
wholegenomesequencing.ThesemotifscanalsobeusedtoscreenspecificcDNA
andgenesequencesofinterestbecauseoftheirassociationwithbiological
functions.Detectionofanappropriatemotifsuggeststhepresenceofafunctional
CRD,andstructuralknowledgeofknownsugar-bindingsitescansuggestwhethera
novelproteinislikelytoretainglycan-bindingactivity.Insomecases,itcaneven
suggestpotentialligands.Suchpredictionsoftenmotivatetestingforsugarbinding
activity,eitherbyspecificallyexaminingbindingtopredictedligandsorby
screeningmoregenerallyusingglycanarrays.
Althoughstructure-basedpredictionsdonotdirectlyyieldinformationabout
biologicalfunction,theorganizationofCRDsandtheirassociationwithother
domainsoftenprovideinformationaboutpotentialfunctions.Thistypeoftop-down
analysisislimitedtodiscoveryofGBPsthatcontaindomainsresemblingknown
CRDs.Asglycanarrayscreeningbecomesmorewidelyaccessible,morebroad-based
screeningcanbeenvisioned.
NATURALLIGANDSFORGBPs
Monosaccharidesorsmalloligosaccharidesinisolationtendtobelow-affinity
ligandsforGBPs,oftenwithdissociationconstantsinthemillimolarrange.These
intrinsicaffinitiesareenhancedinseveralways(Figure28.4).Atthelevelof
individualglycans,affinitycanbeenhancedbylinkingthesugartoothertypesof
structures.Typicalconjugationofglycanstoproteinsandlipidscanleadto
enhancedCRDbinding.Forexample,someGBPssuchasthemacrophagereceptor
minclebindtoglycolipidswithmuchhigheraffinitythantheybindtofree
oligosaccharides.Inthiscase,enhancedaffinitycanresultfromthepresenceofan
extendedoraccessorybindingsiteinaCRDadjacenttothesugar-bindingsite,
whichisabletoaccommodatethehydrophobictailofthelipid.OtherGBPsbind
selectivelytoaparticularglycanconjugatedtoaspecificpolypeptidemotif.Optimal
bindingofP-selectintotheligandPSGL-1requiresanO-linkedglycanbearinga
sialylLewisxstructureonapeptidewithadjacentacidicresiduesandsulfated
tyrosines(Chapter34).Inyetothercases,glycanrecognitioniscombinedwithother
bindingdomainsonaprotein.ThemannosereceptorcontainsC-typeCRDsthat
bindhighmannoseoligosaccharidesandafibronectintypeIIrepeatthatbindsto
triplehelicalpolypeptides.Together,thesetwomodalitiesfacilitatebindingto
fragmentsofcollagenreleasedatsitesofinflammation.
Amajordeterminantofbindingtonaturalligandsistheinteractionofmultivalent
glycanswithclusteredCRDs,resultinginhighaviditybinding.Clusteringofligands
canresultfromthepresenceofmultiplebindingepitopesinasingleoligosaccharide
orpolysaccharide,thepresenceofmultipleglycansattachedtoasingleprotein
scaffoldorthepresenceofadjacentglycoproteinsorglycolipidsinacellmembrane.
Similarly,clusteringofCRDscanreflectthepresenceofmultipleCRDsinasingle
polypeptide,formationofoligomersofpolypeptidethateachcontainsasingleCRD
andfromclusteringofCRD-containingproteinsinthecellmembrane.Eachofthese
levelsoforganizationofCRDshasthepotentialtoplacegeometricalconstraintson
theoptimalarrangementofligands,dependingonthedegreetowhichCRDsare
heldinafixedarrangementorareflexiblylinked.Clusteringofglycansattachedtoa
singlepolypeptide,particularlyinheavilyO-glycosylatedproteinssuchasmucins,
canalsoaffecttheirabilitytotakeondifferentconformations.SinceGBPstypically
interactwithasingleconformation,selectingoneofmultipleaccessible
conformations,thereisanentropicpenaltyassociatedwithbindingwhichmaybe
reducedwhentheglycanhasfewerpotentialconformations.Invitrobiochemical
assays,includingglycanarrays,reflectonlysomeofthesetypesofclusteringof
CRDsandligands,sotheymustbeinterpretedwithsomecaution.Insomecases,
bindingofaCRDtoisolatedglycansmaybeessentiallyundetectableeventhough
bindingoftheintactCRD-containingproteintoitsendogenousglycoconjugatemay
behighlyselectiveandquitestrong.Caremustalsobeexercisedinuseoftheterm
ligand,todistinguishtheglycanpartofaligandfromtheentirenatural
glycoconjugateorevencellsurface.
TERMINOLOGYFORSPECIFICGBPLIGANDS
Basedontheaboveconsiderations,GBPsmaybindselectivelytoaparticularglycan
onlywhenitisconjugatedtoaparticularglycoprotein.TheGBPligandisneitherthe
glycanitselfnortheproteinitself.ExamplesincludeP-selectinbindingtosialyl
LewisxonPSGL-1(seeabove)andE-selectinbindingtothesameglycan(sialyl
Lewisx)carriedonavariantformoftheproteinCD44.Thereisatpresentno
consistentwaytodesignateaglycanonparticularproteinasaligandforaspecific
lectin.SayingthatsialylLewisxorthatPSGL-1(protein)isthe“ligand”forP-Selectin
isnotaccurate.TheE-selectin-bindingformofCD44wasgivenadifferentname
(HCELL,hematopoieticstemcellligandforE-selectin)thatfailstoidentifythe
polypeptidecarrier.Thismatterhasyettoberesolved,especiallywhenasingle
proteinmightbetherequiredpolypeptidescaffoldthatcarriesglycansfordifferent
GBPs.Atthispoint,theconceptthatglycansareoftenligandsforGBPsonlyinthe
contextoftheirproteinorlipidcarriershasbeenwellestablished.
FIGURELEGENDS
FIGURE28.1.Representativestructuresfromfourcommonanimallectinfamilies.
Theemphasisisontheextracellulardomainstructureandtopology.Thefollowing
arethedefinedcarbohydrate-bindingdomains(CRDs)shown:(CL)C-typelectin;
(GL)Galectin;(MP)P-typelectin;(IL)I-typelectin.Otherdomainsare(EG)EGF-like
domain;(IG2)immunoglobulinC2-setdomain;(TM)transmembranedomain;and
(C3)complementregulatoryrepeat.ThenumberofdomainsaccompanyingtheCRD
variesamongfamilymembers.
FIGURE28.2.Arrangementsofcarbohydrate-recognitiondomains(CRDs)inGBPs.
ProteinscontainingjustCRDsorCRDsassociatedwithothertypesoffunctional
domains,withmembraneanchorsorwitholigomerizationdomainsaredepicted
schematically.AsingleGBPcancontainalloftheseadditionaldomains.
agglutinin;EDEM,Endoplasmicreticulum-associateddegradation-enhancingαmannosidase-likeproteins;GH,glycohydrolase;MRH,mannosereceptorhomology.
FIGURE28.4.SourcesofenhancedbindingofnaturalligandstoGBPs.Within
individualCRDs,secondaryinteractionsbeyondtheprimarybindingsitecanbe
withsugar,proteinorlipidportionsofglycoconjugateligands.Multivalent
interactionscanreflectinteractionofsinglebranchedoligosaccharidesormultiple
oligosaccharidesattachedtoaglycoproteinwithmultipleCRDsbroughttogether
withinreceptoroligomersorinGBPclustersonthecellsurface.
FURTHERREADING
StillmarkH.Inauguraldissertation.UniversityofDorpat,Dorpat(nowTartu);
Estonia:1888.UberRicin,EingiftigesFermentausdenSamenvonRicinuscommunis
L.undeinigenanderenEuphoribiaceen.
GoldsteinIJ,HughesRC,MonsignyM,OsawaT,SharonN.Whatshouldbecalleda
lectin?Nature.1980;285:66.
AshwellG,HarfordJ.Carbohydrate-specificreceptorsoftheliver.AnnuRevBiochem.
1982;51:531–554.PubMedPMID:6287920.
DrickamerK.Twodistinctclassesofcarbohydrate-recognitiondomainsinanimal
lectins.JBiolChem.1988;263:9557–9560.PubMedPMID:3290208
Powell,LD,Varki,A.I-typelectins.JBiolChem.1995;270:14243–14246.
LeeR.T.andLeeY.C.2000.Affinityenhancementbymultivalentlectin-carbohydrate
interaction.Glycoconj.J.17:543-551.
Casu,B,Lindahl,U.Structureandbiologicalinteractionsofheparinandheparan
sulfate.AdvCarbohydrChemBiochem.2001;57:159–206.
Esko,JD,Selleck,SB.Orderoutofchaos:assemblyofligandbindingsitesinheparan
sulfate.AnnuRevBiochem.2002;71:435–471.
DrickamerK.andTaylorM.E.2003.Identificationoflectinsfromgenomicsequence
data.MethodsEnzymol.2003,362:592-599.
RigdenD.J.,MelloL.V.,and,GalperinM.Y.2004.ThePA14domain,aconservedallbetadomaininbacterialtoxins,enzymes,adhesinsandsignalingmolecules.
TrendsBiochem.Sci.29:335–339.
SharonN.andLisH.2004.Historyoflectins:fromhemagglutininstobiological
recognitionmolecules.Glycobiology14:53R-62R.
Lee,J.K.,Baum,L.G.,Moremen,K.,andPierce,M.2004.TheX-lectins:anewfamily
withhomologytotheXenopuslaevisoocytelectinXL-35.Glycoconj.J.21:443450
BlundellC.D.,AlmondA.,MahoneyD.J.,DeAngelisP.L.,CampbellI.D.,andDayAJ.
2005.TowardsastructureforaTSG-6.hyaluronancomplexbymodelingand
NMRspectroscopy:insightsintoothermembersofthelinkmodulesuperfamily.
J.Biol.Chem.280:18189-201
Varki,A,Angata,T.Siglecs--themajorsubfamilyofI-typelectins.Glycobiology.
2006;16(1):1R–27R.
SchallusT.,JaeckhC.,FehérK.,PalmaA.S.,LiuY.,SimpsonJ.C.,MackeenM.,StierG.,
GibsonT.J.,FeiziT.,PielerT.,andMuhle-GollC.2008.Malectin:anovel
carbohydrate-bindingproteinoftheendoplasmicreticulumandacandidate
playintheearlystepsofproteinN-glycosylation.Mol.Biol.Cell.19:3404-3414.
VanDamme,E.J.M.,Lannoo,N.,andPeumans,W.J.2008.PlantLectins.Adv.Bot.Res.
48:107–209
TaylorM.E.andDrickamerK.2009.Structuralinsightsintowhatglycanarraystell
usabouthowglycan-bindingproteinsinteractwiththeirligands.Glycobiology
19:1155-1162.
DamT.K.,GerkenT.A.,andBrewerC.F.2009.Thermodynamicsofmultivalent
carbohydrate-lectincrosslinkinginteractions:importanceofentropyinthebind
andjumpmechanism.Biochemistry48:3822-3827.
LindnerH.,MüllerL.M.,Boisson-DernierA.,andGrossniklausU.2012.CrRLK1L
receptor-likekinases:notjustanotherbrickinthewall.Curr.Opin.PlantBiol.
15:659-669.
GhequireM.G.K.,LorisR.,andDeMotR.2012.MMBLproteins:fromlectinto
bacteriocin.Biochem.Soc.Trans.40:1553-1559.
GilbertH.J.,KnoxJ.P.,andBorastonA.B.2013.Advancesinunderstandingthe
molecularbasisofplantcellwallpolysacchariderecognitionbycarbohydratebindingmodules.Curr.Opin.Struct.Biol.2013:669-677.
AdrangiS.andFaramarziM.A.2013.Frombacteriatohuman:ajourneyintothe
worldofchitinases.Biotechnol.Adv.31:1786–1795
TaylorM.E.andDrickamerK.2014.Convergentanddivergentmechanismsofsugar
recognitionacrosskingdoms.Curr.Opin.Struct.Biol.28:14–22.
NagaeM.andYamaguchiY.2014.Three-dimensionalstructuralaspectsofprotein–
polysaccharideinteractions.Int.J.Mol.Sci.15:3768-3783.
DrickamerK.andTaylorM.E.2015.Recentinsightsintostructuresandfunctionsof
C-typelectinsintheimmunesystem.Curr.Opin.Struct.Biol.34:26-34.
BishnoiR.,KhatriI.,SubramanianS.,andRamya,T.N.C.2015.PrevalenceoftheFtypelectindomain.Glycobiology25:888–901.
Table1.Appendix26A
Comparisonoftwomajorclassesofglycan-bindingproteins
Glycosaminoglycan-binding
Lectinsa
proteinsb
Sharedevolutionaryorigins
Sharedstructuralfeatures
DefiningAAresidues
involvedinbinding
Typeofglycansrecognized
yes(withineachgroup)
yes(withineachgroup)
oftentypicalforeachgroup
no
no
patchofbasicaminoacidresidues
N-glycans,O-glycans,
differenttypesofsulfatedglycosaminoglycans
glycosphingolipids(afew
alsorecognizesulfated
glycosaminoglycans)
Locationofcognateresidues typicallyinsequencesat
typicallyinsequencesinternaltoanextended
withinglycans
outerendsofglycanchains
sulfatedglycosaminoglycanchain
Specificityforglycans
stereospecificityhighfor
oftenrecognizearangeofrelatedsulfated
recognized
specificglycanstructures
glycosaminoglycanstructures
Single-sitebindingaffinity
oftenlow;highavidity
oftenmoderatetohigh
generatedbymultivalency
Valencyofbindingsites
multivalencycommon
oftenmonovalent
(eitherwithinnative
structureorbyclustering)
Subgroups
C-typelectins,galectins,Pheparansulfate–bindingproteins,chondroitin
typelectins,I-typelectins,L- sulfate–bindingproteins,dermatansulfate–
typelectins,R-typelectins
bindingproteins
etc.
Typesofglycansrecognized canbesimilar(e.g.,
classificationitselfisbasedontypeof
withineachgroup
galectins)orvariable(e.g.,
glycosaminoglycanchainrecognized
C-typelectins)
ModifiedfromVarkiA.andAngataT.2006.Glycobiology16:1R–27R.
aThereareotheranimalproteinsthatrecognizeglycansinalectin-likemanneranddonotappeartofallintoone
ofthewell-recognizedclasses(e.g.,variouscytokines).
bHyaluronan(HA)-bindingproteins(hyaloadherins)fallinbetweenthesetwoclasses.Ontheonehand,some
(butnotall)ofthehyaloadherinshavesharedevolutionaryorigins.Ontheotherhand,recognitioninvolves
internalregionsofHA,whichisanonsulfatedglycosaminoglycan.
Chapter 38
Proteins that Bind Sulfated Glycosaminoglycans
Essentials of Glycobiology, 3rd edition
ProteinsthatBindSulfatedGlycosaminoglycans
Authors:JeffreyD.Esko,JamesPrestegardandRobertJ.Linhardt
Glycosaminoglycansbindtomanydifferentclassesofproteinsmostlythrough
electrostaticinteractionsbetweennegativelychargedsulfategroupsanduronic
acidsandpositivelychargedaminoacidsintheprotein.Thischapterfocuseson
examplesofglycosaminoglycan-bindingproteins,methodsformeasuring
glycosaminoglycan-proteininteraction,andinformationaboutthree-dimensional
structuresofthecomplexes.
GLYCOSAMINOGLYCAN-BINDINGPROTEINSARECOMMON
Severalhundredglycosaminoglycan(GAG)-bindingproteinshavebeendiscovered,
whichmakeuptheGAG-interactomeandfallintothebroadclassespresentedin
Table38.1.Toalargeextent,studiesoftheGAG-interactomehavefocusedon
proteininteractionswithheparin,amorehighlysulfated,iduronicacid(IdoA)-rich
formofheparansulfate(HS;Chapter17).Thisbiasreflectsinpartthecommercial
availabilityofheparinandheparin-Sepharose,whicharefrequentlyusedfor
fractionationstudies,andtheassumptionthatbindingtoheparinmimicsbindingto
HSpresentoncellsurfacesandintheextracellularmatrix.Incomparison,relatively
fewproteinsareknowntointeractwithchondroitinsulfate(CS)orkeratansulfate
(KS)withcomparableavidityandaffinity.Insomecases,CSandtherelatedGAG,
dermatansulfate(DS),maybephysiologicallyrelevantbindingpartnersbecause
theseGAGspredominateinmanytissues.Determiningthephysiologicalrelevance
oftheseinteractionsisamajorareaofresearch.
Incontrasttolectins,whichtendtofallintoevolutionarilyconservedfamilies
(Chapters28-37),GAG-bindingproteinsdonothavecommonfoldsandinstead
appeartohaveevolvedbyconvergentevolution.AsshowninTable38.1,the
interactionbetweenGAGsandproteinscanhaveprofoundphysiologicaleffectson
processessuchashemostasis,lipidtransportandabsorption,cellgrowthand
migration,anddevelopment.BindingtoGAGscanresultinimmobilizationof
proteinsattheirsitesofproductionorintheextracellularmatrixforfuture
mobilization;regulationofenzymeactivity;bindingofligandstotheirreceptors;
proteinoligomerization;andprotectionofproteinsagainstdegradation.Insome
cases,theinteractionmayreflectcomplementarityofcharge(e.g.,histone-heparin
interactions)ratherthananyspecificbiologicallyrelevantinteraction.Inother
cases,theinteractionhasbeenshowntodependonrarebutveryspecificsequences
ofmodifiedsugarsintheGAGchain(e.g.,antithrombinbinding).
METHODSFORMEASURINGGLYCOSAMINOGLYCAN-PROTEIN
BINDING
NumerousmethodsareavailableforanalyzingGAG-proteininteractions,andsome
provideadirectmeasurementofKdvalues.Acommonmethodinvolvesaffinity
fractionationofproteinsonSepharosecolumnscontainingcovalentlylinkedGAG
chains,usuallyheparin.Theboundproteinsareelutedwithdifferentconcentrations
ofsodiumchloride,andtheconcentrationrequiredforelutionisgenerally
proportionaltotheKd.High-affinityinteractionsrequireatleast1MNaClto
displaceboundligand,whichtranslatesintoKdvaluesof10−7–10−9M(determined
underphysiologicalsaltconcentrationsbyequilibriumbinding).Proteinswithlow
affinity(10−4–10−6M)eitherdonotbindunder“normal”conditions(0.15MNaCl)or
requireonly0.3–0.5MNaCltoelute.Thismethodisbasedontheassumptionthat
GAG-proteininteractionisentirelyionic,whichisnotentirelycorrect.Nevertheless,
itcanprovideanassessmentofrelativeaffinity,whencomparingdifferentGAGbindingproteins.
Anumberofmoresophisticatedmethodsarenowinusethatprovidedetailed
thermodynamicdata(ΔH[changeinenthalpy],ΔS[changeinentropy],ΔCp[change
inmolarheatcapacity],etc.),kineticdata(associationanddissociationrates),and
high-resolutiondataonatomiccontactsinGAG-proteininteractions(Table38.2).
Regardlessofthetechniqueoneuses,itmustbekeptinmindthatinvitrobinding
measurementsarenotlikelytobethesameasthosewhentheproteinbindsto
proteoglycansonthecellsurfaceorintheextracellularmatrix,wherethedensity
andvarietyofGAG-bindingproteins,proteoglycansandotherinteractingfactors
variesgreatly.Todeterminethephysiologicalrelevanceoftheinteraction,one
shouldconsidermeasuringbindingunderconditionsthatcanleadtoabiological
response.Forexample,onecanmeasurebindingtocellswithalteredGAG
composition(Chapter49)oraftertreatmentwithspecificlyasestoremoveGAG
chainsfromthecellsurface(Chapter17)andthendeterminewhetherthesame
responseoccursasobservedinthepresenceofGAGchains.Theinteractioncanthen
bestudiedmoreintensivelyusingtheinvitroassaysdescribedabove.
CONFORMATIONALANDSEQUENCECONSIDERATIONS
Asmentionedabove,mostGAG-bindingproteinsinteractwithHSorheparin.The
likelybasisforthispreferenceisgreatersequenceheterogeneityandmorevariable
sulfation,comparedtootherGAGs.Theunusualconformationalflexibilityof
iduronicacid,whichisfoundinheparin,HS,andDS,alsohasaroleintheirabilityto
bindproteins.GAGsarelinearhelicalstructures,consistingofalternatingresidues
ofN-acetylglucosamine(GlcNAc)orN-acetylgalactosamine(GalNAc)with
glucuronicacid(GlcA)oriduronicacid(IdoA)(withtheexceptionofkeratan
sulfates,whichconsistofalternatingGlcNAcandgalactoseresidues;Chapter17).
Inspectionofheparinoligosaccharidescontaininghighlymodifieddomains
([GlcNS6S-IdoA2S]n)showsthattheN-sulfoand6-O-sulfogroupsofeach
disacchariderepeatlieonoppositesidesofthehelixfromthe2-O-sulfoand
carboxylgroups(Figure38.1).Analysisoftheconformationofindividualsugars
showsthatGlcNAcandGlcAresiduesassumeapreferredconformationinsolution,
designated4C1(indicatingthatcarbon4isabovetheplanedefinedbycarbons2,3,
and5andtheringoxygen,andthatcarbon1isbelowtheplane;Chapter2).In
contrast,IdoA2Sassumesthe1C4orthe2SOconformation(Figure38.1),which
reorientsthepositionofthesulfosubstituents,therebycreatingadifferent
orientationofchargedgroups.InmanycaseswhenaproteinbindstoanHSchain,it
inducesachangeinconformationoftheIdoA2Sresidueresultinginabetterfitand
enhancedbinding.IdoA2SresidueshavealwaysbeenfoundindomainsrichinNsulfoandO-sulfogroups(forbiosyntheticreasons;Chapter17),whichisalsowhere
proteinsusuallybind.Thus,thegreaterdegreeofconformationalflexibilityinthese
modifiedregionsmayexplainwhysomanymoreproteinsbindwithhighaffinityto
heparin,HS,andDSthantootherGAGs.ThepresenceofanN-acetylgroupinan
GlcNAcresiduechangesthepreferredconformationoftheneighboringIdoA2S
residue,showingthatevenminormodificationscaninfluenceconformationand
chainflexibility.BindingtoGAGsthathavealowdegreeofsulfationmayrequire
largerdomainsintheproteintointeractwithlongerstretchesofanoligosaccharide.
Moleculardynamicsimulationsonlargeheparinoligosaccharidesarepossiblewith
theavailabilityofsupercomputers(seeSimulation35.1ontheaccompanying
website).Suchsimulationscanbeusedtopredicttheconformationalflexibilityof
differentdomainswithinthechainandwhencombinedwithrecentadvancesin
protein-GAGdocking,canprovideadditionalinsightsintoGAG-proteininteractions.
HOWSPECIFICAREGLYCOSAMINOGLYCAN-PROTEIN
INTERACTIONS?
ThediscoveryofmultipleGAG-bindingproteinsledanumberofinvestigatorsto
examinewhetherthereisaconsensusaminoacidsequenceforGAGbinding.In
retrospect,thisstrategywasoverlysimplisticbecauseitassumedthatallGAGbindingproteinswouldrecognizethesameoligosaccharidesequencewithin
heparin,oratleast,sequencesthatwouldsharemanycommonfeatures.Wenow
knowthatsomeGAG-bindingproteinsinteractwithdifferentoligosaccharide
sequences.Thebindingsitesintheproteinalwayscontainbasicaminoacids(lysine
andarginine)whosepositivechargespresumablyinteractwiththenegatively
chargedsulfatesandcarboxylatesoftheGAGchains.However,thearrangementof
thesebasicaminoacidscanbequitevariable,consistentwiththevariable
positioningofsulfogroupsintheGAGpartner.
Mostproteinsareformedfromα-helices,β-strands,andloops.Therefore,toengage
alinearGAGchain,thepositivelychargedaminoacidresiduesmustalignalongthe
samesideoftheproteinsegment.α-Heliceshaveperiodicitiesof3.4residuesper
turn,whichwouldrequirethebasicresiduestooccureverythirdorfourthposition
alongthehelixinordertoalignwithanoligosaccharide.Inβ-strands,theside
chainsalternatesideseveryotherresidue.Thus,tobindaGAGchain,thepositively
chargedresiduesinaβ-strandwouldbelocatedquitedifferentlythaninanα-helix.
Onthebasisofthestructureofseveralheparin-bindingproteinsthatwereavailable
in1991,AlanCardinandHerschelWeintraubproposedthattypicalheparin-binding
siteshadthesequenceXBBXBXorXBBBXXBX,whereBislysineorarginineandXis
anyotheraminoacid.Fromthestructuralargumentsprovidedabove,itshouldbe
obviousthatonlysomeofthebasicresiduesinthesesequencescouldparticipatein
GAGbinding,theactualnumberbeingdeterminedbywhetherthepeptidesequence
existsasanα-helixoraβ-sheet.Wenowknowthatthepresenceofthesesequences
inaproteinmerelysuggestsapossibleinteractionwithheparin(oranotherGAG
chain),butitdoesnotprovethattheinteractionoccursunderphysiological
conditions.Infact,thepredictedbindingsitesforheparininfibroblastgrowthfactor
2(FGF2)turnedouttobeincorrectoncethecrystalstructurewasdetermined.Itis
likelythatbindinginvolvesmultipleproteinsegmentsthatjuxtaposepositively
chargedresiduesintoathree-dimensionalturn-richrecognitionsite.Inmanycases
thebindinginvolvesloopswhichmakethepositioningmorevariable.Anexampleof
thisphenomenonisobservedinthechemokineCCL5,whichcontainsaBBXBmotif
inaloop.Thespecificarrangementofresiduesshouldvaryaccordingtothetype
andfinestructureofthoseoligosaccharidesinvolvedinbinding.
Inplantandanimallectins,andinantibodiesthatrecognizeglycans,theglycan
recognitiondomainsaretypicallyshallowpocketsthatengagetheterminalsugars
oftheoligosaccharidechain(Chapters29,30and37).InGAG-bindingproteins,the
proteinusuallybindstosugarresiduesthatliewithinthechainornearthe
terminus.Therefore,thebindingsitesinGAG-bindingproteinsconsistofcleftsor
setsofjuxtaposedsurfaceresiduesratherthanpockets.TheseGAG-bindingsiteson
theproteinsurfacegiverisetomorerapidGAG-proteinbindingkineticsthanare
typicallyobservedforprotein-proteininteractions.GiventhatGAGchainsgenerally
existinahelicalconformation,onlythoseresiduesonthefacetowardtheprotein
interactwithaminoacidresidues;theonesontheothersideofthehelixare
potentiallyfreetointeractwithasecondligand(e.g.,asobservedinFGFdimers).
Alternatively,residuesinabindingcleftcouldinteractwithbothsidesofthehelix
(e.g.indengueenvelopeprotein).Finally,oneshouldkeepinmindthatbinding
occurstoonlyasmallsegmentoftheGAGchain.Thus,asingleGAGchaincan
potentiallybindmultipleproteinligandsfacilitatingcooperativebindingthatcan
leadtoproteinoligomerization(e.g.somechemokines).
ANTITHROMBIN-HEPARIN:APARADIGMFORSTUDYING
GLYCOSAMINOGLYCAN-BINDINGPROTEINS
Perhapsthebest-studiedexampleofprotein-GAGinteractionisthebindingof
antithrombintoheparinandHS(seecoverimageandFigure38.2).Thisinteraction
isofgreatpharmacologicalimportancebecauseheparinisusedclinicallyasan
anticoagulant.Bindingofantithrombintoheparinhasadualeffect:First,itcausesa
conformationalchangeintheproteinandactivationoftheproteaseinhibiting
action,resultingina1000-foldenhancementintherateatwhichitinactivates
thrombinandFactorXa.Second,theheparinchainactsasatemplate,enhancingthe
physicalappositionofthrombinandantithrombin.Thus,boththeprotease
(thrombin)andtheinhibitorhaveGAG-bindingsites.Heparinactsasacatalystin
thesereactionsbyenhancingtherateofthereactionthroughappositionof
substratesandconformationalchange.Aftertheinactivationofthrombinby
antithrombinoccurs,thecomplexlosesaffinityforheparinanddissociates.The
heparinisthenavailabletoparticipateinanotheractivation/inactivationcycle.
Antithrombinisamemberoftheserpinfamilyofproteaseinhibitors,manyofwhich
bindtoheparin.
Earlystudiesusingaffinityfractionationschemesshowedthatonlyaboutone-third
ofthechainsinaheparinpreparationactuallybindwithhighaffinityto
antithrombin.Comparingthesequenceoftheboundchainswiththosethatdidnot
bindfailedtorevealanysubstantialdifferencesincomposition,consistentwiththe
laterdiscoverythatthebindingsiteconsistsofonlyfivesugarresidues(Figure
38.2)(theaverageheparinchainisabout50sugarresidues).Thisobservationcan
beextendedtovirtuallyallGAG-bindingproteins,inferringthatthebindingsites
representaverysmallsegmentofthechains.
CrystalsofantithrombinwerepreparedandanalyzedbyX-raydiffractionto2.6-Å
resolution.Thedockingsitefortheheparinpentasaccharideisformedbythe
appositionofhelicesAandD,whichbothcontaincriticalarginineandlysine
residuesattheinterface.ThesequenceintheDhelix
(124AKLNCRLYRKANKSSKLVSANR145)placesmanyofthepositivelycharged
residuesononefaceofthehelix,inproximitytothearginineresiduesintheAhelix
(41PEATNRRVW49)(Figure38.2).Thepentasaccharideissufficienttoactivate
antithrombinbindingtowardFactorXa,butitwillnotfacilitatetheinactivationof
thrombin.Forthistooccur,alargeroligosaccharideofatleast18residuesis
needed.Asmentionedabove,thrombinalsocontainsaheparin-bindingsite,andthe
largerheparinoligosaccharideisthoughttoactasatemplatefortheformationofa
ternarycomplexwiththrombinandantithrombin.Incontrasttoantithrombin,
thrombinexhibitslittleoligosaccharidespecificity.Asmightbeexpected,adding
highconcentrationsofheparinactuallyinhibitsthereaction,becausetheformation
ofbinarycomplexesofheparinandthrombinorheparinandantithrombin
predominate.Thisimportantprincipleof“activationatlowconcentrationsand
inhibitionathighconcentrations”alsooccursinothersystemswhereternary
complexesform(Chapters29and30).
Heparinisapharmaceuticalformulationproducedbypartialfractionationof
naturalGAGsderivedprimarilyfromporcineintestines(Chapter17).Mastcellsare
knowntoproduceahighlysulfatedversionofHSthatresemblesheparin;highly
sulfated,iduronicacid–richheparinoligosaccharidesarealsopresentinHSisolated
fromothertissuesaswell,especiallytheskin.Althoughheparinhasproventobeof
greattherapeuticuse,itsroleinvivoremainsunclear.Heparinandchondroitin
sulfateareoftenfoundinstoragegranulesalongwithbiogenicamines,proteases,
andotherproteins,possiblyenablingefficientstorage.Mastcellsdegranulatein
responsetospecificantigenstimulation,resultinginreleaseofstoredheparin,
histamine,andproteases.Whenthisoccurs,localanticoagulationmightoccur,but
localizedcoagulationdefectshavenotbeendescribedinanimalsbearingmutations
thataltermastcellsorheparin.AsmallpercentageofendothelialcellHScontains
antithrombin-bindingsequencesaswell.However,thesebindingsitesappeartobe
locatedontheabluminalsideofbloodvessels,andmicelackingthecentral3-OsulfatedGlcNSunit,ahallmarkoftheantithrombin-bindingsequence(Figure38.2),
donotexhibitanysystemiccoagulopathyafterbirth.Nevertheless,antithrombin
deficiencycausesmassivedisseminatedcoagulopathy.Perhapsthesefindings
indicatethatlower-affinitybindingsequencesaresufficienttoactivate
antithrombin.Thissystemillustratesanimportantcaveat:onecannotnecessarily
ascribefunctionstoendogenousproteoglycansbasedontheeffectsofGAGsadded
invitrotoexperimentalsystems.
FGF-HEPARININTERACTIONSENHANCESTIMULATIONOFFGF
RECEPTORSIGNALTRANSDUCTION
Alargenumberofgrowthfactorscanbepurifiedbasedontheiraffinityforheparin.
Theheparin-bindingfamilyoffibroblastgrowthfactorshasgrowntomorethan22
membersandincludestheprototypeFGF2,otherwiseknownasbasicfibroblast
growthfactor.FGF2hasaveryhighaffinityforheparin(Kd~10−9M)andrequires
1.5–2MNaCltoelutefromheparin-Sepharose.FGF2haspotentmitogenicactivityin
cellsthatexpressoneoftheFGFsignalingreceptors(fourFGFRgenesareknown
andmultiplesplicevariantsexist).Cell-surfaceHSbindstobothFGF2andFGFR,
facilitatingtheformationofaternarycomplex.Bothbindingandthemitogenic
responsearegreatlystimulatedbyheparinorHS,whichspromotedimerizationof
theligand-receptorcomplex.
ThecostimulatoryroleofHS(andheparin)inthissystemisreminiscentofthe
heparin/antithrombin/thrombinstory.Indeed,theminimalbindingsequencefor
FGF2alsoconsistsofapentasaccharide.However,thispentasaccharideisnot
sufficienttotriggerabiologicalresponse(mitogenesis).Forthistooccur,alonger
oligosaccharide(10mer)containingtheminimalsequenceandadditional6-O-sulfo
groupsareneededtobindFGFR.ThesequencethatbindstobothFGF2andFGFRis
prevalentinheparinbutrareinHS.Therequirementforthisrarebindingsequence
reducestheprobabilityoffindingthisparticulararrangementinnaturallyoccurring
HSchains.Thus,somepreparationsofHSareinactiveinmitogenesis,andthose
containingonlyonehalfofthebipartitebindingsequenceareactuallyinhibitory.
ThestructureofFGF2cocrystallizedwithaheparinhexasaccharidehassincebeen
obtained(Figure38.3).Theheparinfragment([GlcNS6Sα1-4IdoA2Sα1-4]3)was
helicalandboundtoaturn-richheparin-bindingsiteonthesurfaceofFGF2.Only
oneN-sulfogroupandthe2-O-sulfogroupfromtheadjacentiduronicacidare
boundtothegrowthfactorintheturn-richbindingdomain,andthenextGlcNS
residueisboundtoasecondsite,consistentwiththeminimalbindingsequence
determinedwitholigosaccharidefragments.Nosignificantconformationalchangein
FGF2occursuponheparinbinding,consistentwiththeideathatheparinprimarily
servestodimerizeFGF2andjuxtaposecomponentsoftheFGFsignal-transduction
pathway.ThecrystalstructureofacidicFGF(FGF1)hasalsobeensolvedandshows
similarsequencesonitssurface.However,theoligosaccharidesequencethatbinds
withhighaffinitytoFGF1contains6-O-sulfogroups.
Thecocrystalstructureofthecomplexof(FGF2-FGFR)2,firstsolvedintheabsence
ofheparin/HSligand,showedacanyonofpositivelychargedaminoacidresidues,
suggestiveofanunoccupiedheparin-bindingsite.Subsequently,theheparinoligosaccharide-containingcomplexwassolvedafterintroductionofheparin
oligosaccharides,suggestinga2:2:2complexofFGF2:FGFR:HS(Figure38.3).
Anotherimportantfeatureofthiscomplexistheorientationofthenon-reducing
endsoftheHSchainsthatterminateinanN-sulfoglucosamineresidue,whicharises
byendolyticcleavageofchainsbytheenzymeheparanase(Chapter17).The
structureoftheFGF-FGFR-HScomplexisnotwithoutcontroversy;structural
analysisofcomplexesformedinsolutionandpurifiedbygelfiltrationhassuggested
averydifferentstructureconsistingofa2:2:1complex(Figure38.3).
OTHERATTRIBUTESOFGLYCOSAMINOGLYCAN-PROTEIN
INTERACTIONS
Insomecases,theinteractionofGAGchainswithproteinsmaydependonmetal
cofactors.Forexample,L-andP-selectinshavebeenshowntobindtoasubfraction
ofHSchainsandheparininadivalent-cation-dependentmanner.Thisobservation
raisesthepossibilitythatotherexamplesofcation-dependentinteractionswithGAG
chainsmayexist.GAGbindingtoL-selectinhelpsinleukocyterolling.Furthermore,
theinteractioncanbepharmacologicallymanipulatedbyexogenousheparin,
includingchemicallymodifiedderivativesthatlackanticoagulantactivity.
CSproteoglycansinthecentralnervoussystem(CNS)influencecellmigrationand
axonpathfindingandregulateneuriteoutgrowth.Theinteractionofrare,highly
sulfateddisaccharidesequencesinCSchainswithmorphogensandgrowthfactors
impactCNSdevelopmentandplayrolesinCNSpathology.
HSproteoglycansareoftenexpressedinaspatiallyandtemporallylimitedfashion.
ThetemporaryplacementofanHSproteoglycanataspecifictissuesitemightor
mightnotcoincidewiththepresenceofitsappropriateproteinligand.Furthermore,
ifthebindingpartnerhasnoaccesstotheHSproteoglycan,itcannotinteract—
addinganadditionallevelofspecificity.Recentstudiesdemonstratethatthefine
structureofHSchainsalsochangesduringdevelopment,thusenablingordisabling
specificassociationsbetweenligandsandreceptors.
Gradientsofmorphogens,factorsthatdeterminecellfatesbasedonconcentration,
alsodeterminethepatternsofcellandtissueorganizationduringdevelopment
(Chapter27).Themechanismofmorphogengradientformationiscontroversial,but
interestingly,virtuallyallmorphogenscaninteractwithheparinandHS.These
interactionscanaffecttransportofligands,receptorinteractions,endocytosis,and
degradation,whichtogethermayhavearoleindeterminingtherobustnessofthe
gradient.TheGAGchainsofproteoglycansalsoofferalineardomainoverwhich
proteinscandiffuse.Bylimitingthespaceavailabletotheseproteinsfromthethreedimensionalspaceofextracellularfluidsandtheextracellularmatrixtoonedimensionalspacealongthechains,thechanceofencountersamongheparinbindingproteins,suchasFGFanditsreceptor(FGFR),maybeenhanced.Thus,the
criticalroleofHSproteoglycansmaybeincontrollingthekineticsofprotein–
proteininteractionsratherthanthethermodynamicsofsuchencounters.
ACKNOWLEDGEMENTS
TheauthorsappreciatehelpfulcommentsandsuggestionsfromKristianSaied,
EathenRyan,PatienceWrightandKristinStanford.
FURTHERREADING
LiW,JohnsonDJ,EsmonCT,HuntingtonJA.2004.Structureoftheantithrombinthrombin-heparinternarycomplexrevealstheantithromboticmechanismof
heparin.NatStructMolBiol11:857-862.
MohammadiM,OlsenSK,GoetzR.2005.AproteincanyonintheFGF-FGFreceptor
dimerselectsfromanalacartemenuofheparansulfatemotifs.CurrOpin
StructBiol15:506-516.
DuchesneL,OcteauV,BearonRN,BeckettA,PriorIA,LounisB,FernigDG.2012.
Transportoffibroblastgrowthfactor2inthepericellularmatrixiscontrolled
bythespatialdistributionofitsbindingsitesinheparansulfate.PLoSBiol10:
e1001361.
KamhiE,JooEJ,DordickJS,LinhardtRJ.2013.Glycosaminoglycansininfectious
disease.BiolRevCambPhilosSoc88:928-943.
ThackerBE,XuD,LawrenceR,EskoJD.2013.Heparansulfate3-O-sulfation:Arare
modificationinsearchofafunction.MatrixBiol35:60-72.
XuD,EskoJD.2014.Demystifyingheparansulfate-proteininteractions.AnnuRev
Biochem83:129-157.
MizumotoS,YamadaS,SugaharaK.2015.Molecularinteractionsbetween
chondroitin-dermatansulfateandgrowthfactors/receptors/matrixproteins.
CurrOpinStructBiol34:35-42.
PominVH,MulloyB.2015.Currentstructuralbiologyoftheheparininteractome.
CurrOpinStructBiol34:17-25.
SmithPD,Coulson-ThomasVJ,FoscarinS,KwokJC,FawcettJW.2015."GAG-ingwith
theneuron":Theroleofglycosaminoglycanpatterninginthecentralnervous
system.ExpNeurol274:100-114.
DeshauerC,MorganAM,RyanEO,HandelTM,PrestegardJH,WangX.2015.
InteractionsofthechemokineCCL5/RANTESwithmedium-sizedchondroitin
sulfateligands.Structure(London,England:1993)23:1066-1077.
FigureLegends
FIGURE38.1.Conformationofheparinoligosaccharides.(A)Glucosamine(GlcN)and
glucuronicacid(GlcA)existinthe4C1conformation,whereasiduronicacid(IdoA)
existsinequallyenergeticconformationsdesignated1C4and2S0.(B)Space-filling
modelofaheparinoligosaccharide(14mer)deducedbynuclearmagnetic
resonance.(C)Thesamestructureinstickrepresentation.TherenderingsinBandC
weremadewithRASMOLusingdatafromtheMolecularModelingDatabase(MMDB
Id:3448)attheNationalCenterforBiotechnologyInformation(NCBI).
FIGURE38.2.Crystalstructureoftheantithrombin-pentasaccharidecomplex(from
ProteinDataBank).(A,D)α-Helicesthatmakecontactwithheparin;(RCL)the
reactivecenterloopthatinactivatesthrombinandFactorX;(F)anotherα-helixin
theprotein.(Lowerpanel)Interactionsbetweenkeyaminoacidresiduesand
individualelementsinthepentasaccharide.(Solidlines)Electrostaticinteractions
betweenpositivelychargedresiduesandsulfategroups;(brokenlines)hydrogen
bonds;(alternatelybrokenandsolidline)bridgingwatermolecule.
FIGURE38.3.CrystalandNMRsolutionstructuresofGAG-proteincomplexes.(A)
Crystalstructureofthe2:2:2FGF2:FGFR1:heparincomplex(sideview)anda2:2:1
complex;(B)StructureofthedimericV-C1domainsofRAGE(receptorforadvanced
glycationendproducts)(PDB4IM8).Thedodecasaccharideismanuallymodeled
intothestructureonthebasisoftheobservedpartialelectrondensity;(C)Structure
ofthedimericE2domainofamyloidprecursor–likeprotein1(APLP-1)andbound
oligosaccharide(PDB3QMK);(D)Structureofdimericinterleukin-8(PDB2IL8)and
amodeledoligosaccharide(degreeofpolymerization:20);(E)Arepresentative
framefromthemostenergeticallyfavoredmodelsoftheCCL5-chondroitinsulfate
complexdeducedbyNMR.Theribbonrepresentationofthecomplexwithside
chainsofselectiveaminoacidsisshowningray.Thechondroitin-4-sulfate(dp6)
ligandisshowninthestickrepresentationwiththenon-reducingendandthe
reducingendsugarlabeledGlcA1andGalNAc3,respectively(fromDeshaueretal.
(2015)Structure23,1066–1077)
TABLE38.1Examplesofglycosaminoglycan-bindingproteinsandtheir
biologicalactivity
Physiological/pathophysiological
Class
Examples
effectsofbinding
Enzymes
glycosaminoglycan
multiple
biosyntheticenzymes,
thrombinandcoagulation
factors(proteases),
complementproteins
(esterases),extracellular
superoxidedismutase,
lipases
Enzyme
antithrombinIII,heparin
coagulation,inflammation,
inhibitors
cofactorII,secretory
complementregulation
leukocyteproteinase
inhibitor,C1-esterase
inhibitor
Celladhesion
P-selectin,L-selectin,some celladhesion,inflammation,
proteins
integrins
metastasis
Extracellular
laminin,fibronectin,
celladhesion,matrixorganization
matrix
collagens,thrombospondin,
proteins
vitronectin,tenascin
Chemokines
plateletfactorIV,γ-andβ- chemotaxis,signaling,inflammation
interferons,interleukins
Growthfactors fibroblastgrowthfactors,
mitogenesis,cellmigration
hepatocytegrowthfactor,
vascularendothelial
growthfactor,insulin-like
growthfactor–binding
proteins,TGF-β-binding
proteins
Morphogens
hedgehogs,TGF-βfamily
cellspecification,tissue
members,wnts
differentiation,development
Guidance
Slits,ROBOreceptors,
axonguidance,endothelialtube
factors
neuropilins
formation
Tyrosinefibroblastgrowthfactor
Mitogenesis,axonguidance,
kinasegrowth receptors,vascular
inflammation
factor
endotheliumgrowthfactor
receptorsand receptor,receptorfor
coreceptors
advancedglycation
endproducts(RAGE),
receptorproteintyrosine
phosphatases(RPTPs)
Lipid-binding apolipoproteinsEandB,
lipidmetabolism,cellmembrane
proteins
lipoproteinlipase,hepatic functions
Plaque
proteins
Nuclear
proteins
Pathogen
surface
proteins
Viralenvelope
proteins
lipase,annexins
prionproteins,amyloid
proteins
histones,transcription
factors
malariacircumsporozoite
protein
herpessimplexvirus,
denguevirus,human
immunodeficiencyvirus,
hepatitisCvirus,vaccinia
viruscomplementcontrol
protein(VCP)
plaqueformation
unknown
pathogeninfections
viralinfections
TABLE38.2Methodstomeasureglycosaminoglycan-proteininteraction
Method
Type Throughput Principle
Affinity
M H/I
immobilizedligandor
chromatography
glycosaminoglycanchainsoncolumn
matrix
Affinity
M/S I
gelretardationthroughproteincoelectrophoresis
impregnatedgel
Analytical
S
L
equilibriumsedimentationat
ultracentrifugation
differentcarbohydrate:proteinratios
Circulardichroism
S
I/L
changeinrotationofplane-polarized
lightuponbinding
CompetitionELISA
M H
solution-andsolid-phaseligands
competeforbinding
Computational
S
L
calculatescomplexstructureand
bindingenergy
Fluorescence
S
H
conformationalchange,ligandbinding
spectroscopy
induceschangeinfluorescence
Ionmobilitymass
G
I/L
Measurescomplexshapeand
spectrometry
stoichiometry
Isothermaltitration
S
I
measuresenthalpyofbindingdirectly
calorimetry
andKdvalues
Laserlightscattering
S
I
intrinsicscatteringintensitiesof
carbohydrate-proteincomplexused
tocalculatestoichiometry
Nuclearmagnetic
S
L
chemicalshift,couplingconstant,and
resonance
nuclearOverhausereffectto
determinecontactpoints,distances,
andconformation
Surfaceplasmon
resonance
M
H/I
Xray
M
L
mass-inducedrefractiveindexchange
inrealtimefordirectmeasurementof
associationanddissociationrate
constants
solidstatecocrystalstructure
(M)Mixedphase;(S)solutionphase;(G)gasphase;(H)high;(I)intermediate;(L)low.
Lectin (handout)