Antibacterial membrane attack by a pore-forming intestinal C

LETTER

doi:10.1038/nature12729

Antibacterialmembraneattackbyapore-formingintestinalC-typelectin

SohiniMukherjee1,HuiZheng2,MehabawG.Derebe1,KeithM.Callenberg3,CarrieL.Partch4,DarcyRollins1,DanielC.Propheter1,JosepRizo5,MichaelGrabe3{,Qiu-XingJiang2*&LoraV.Hooper1,6*

Humanbody-surfaceepitheliacoexistincloseassociationwithcomplexbacterialcommunitiesandareprotectedbyavarietyofantibacterialproteins.C-typelectinsoftheRegIIIfamilyarebac-tericidalproteinsthatlimitdirectcontactbetweenbacteriaandtheintestinalepitheliumandthuspromotetolerancetotheintestinalmicrobiota1,2.RegIIIlectinsrecognizetheirbacterialtargetsbybindingpeptidoglycancarbohydrate1,3,butthemechanismbywhichtheykillbacteriaisunknown.HereweelucidatethemechanisticbasisforRegIIIbactericidalactivity.WeshowthathumanRegIIIa(alsoknownasHIP/PAP)bindsmembranephospholipidsandkillsbacteriabyformingahexamericmembrane-permeabilizingoligo-mericpore.Wederiveathree-dimensionalmodeloftheRegIIIaporebydockingtheRegIIIacrystalstructureintoacryo-electronmicroscopicmapoftheporecomplex,andshowthatthemodelaccordswithexperimentallydeterminedpropertiesofthepore.Lipopoly-saccharideinhibitsRegIIIapore-formingactivity,explainingwhyRegIIIaisbactericidalforGram-positivebutnotGram-negativebacteria.OurfindingsidentifyC-typelectinsasmediatorsofmem-braneattackinthemucosalimmunesystem,andprovidedetailedinsightintoanantibacterialmechanismthatpromotesmutualismwiththeresidentmicrobiota.

RegIIIadamagesthesurfacesofGram-positivebacteria1,indicatingthatRegIIIamighttargetbacterialmembranes.Weassessedthecapa-cityofRegIIIatopermeabilizebacterialmembranesbyquantifyingbacterialuptakeofamembrane-impermeantfluorescentdye(SYTOXgreen).RegIIIaincreasedSYTOXgreenuptakewhenaddedtotheGram-positivespeciesListeriamonocytogenes,indicatingdamagedmembranes(Fig.1a,b).RegIIIahasananionicamino-terminalpro-segmentthatinhibitsbactericidalactivity(butnotpeptidoglycanbind-ing)bydockingtotheproteincorethroughcharge–chargeinteractions4.Thepro-segmentisremovedbytrypsinonsecretionintotheintestinallumen,yieldingbactericidallyactiveRegIIIa(ref.4).Bactericidallyinactivepro-RegIIIadidnotinduceSYTOXgreenuptake,indicatingminimalmembranepermeabilization(Fig.1a).Thus,RegIIIapermea-bilizesthebacterialmembrane,andthepro-segmentinhibitsthisactivity.TotestdirectlyformembranedisruptionbyRegIIIaweusedlipo-somescomposedof85%zwitterionicphospholipid(PC)and15%acidicphospholipid(PS).Theliposomesencapsulatedcarboxyfluorescein,afluorescentdye.RegIIIainducedrapiddyeeffluxfromPC/PSlipo-somes(Fig.1c),whichwasreducedwhenPC-onlyliposomeswereused(Fig.1d,e).Thisindicatesapreferenceforacidicphospholipidsthatisconsistentwiththeacidiclipidcontentofbacterialmembranes5andwiththesaltsensitivityofRegIIIamembranetoxicity(ExtendedDataFig.2a,b).ThesefindingsindicatethatRegIIIainteractionswithlipidbilayersaremediatedbyelectrostaticinteractions.pro-RegIIIayieldedadiminishedrateofdyerelease(Fig.1f),indicatingthatthepro-segmentinhibitsmembranepermeabilization.

1

WenextassessedRegIIIalipid-bindingactivitybymeasuringchangesintheintrinsicfluorescenceoftryptophanresidues6.WeobservedincreasedtryptophanfluorescenceintensityonlywhenRegIIIawasaddedtoPS-containingliposomes(Fig.1g–i),indicatingthatRegIIIainteractswithacidicphospholipids.Furthermore,weobservedfluorescenceres-onanceenergytransfer(FRET)betweendonorRegIIIatryptophanresiduesanddansyl-labelledPC/PSliposomes7(Fig.1j,k).FRETwasinhibitedbythepro-RegIIIaN-terminalpro-segment(Fig.1j,k),indi-catingthatthepro-segmentinhibitsbactericidalactivitybyhinderinglipidbinding.Consistentwithitsinabilitytobindlipids,pro-RegIIIadidnotinhibitRegIIIabactericidalactivityinmixingexperiments(ExtendedDataFig.2c).

Severalmembrane-activetoxinsdestabilizemembranesbyformingmonomericormultimericpores8.TotestforRegIIIapores,weper-formedconductancestudiesinblacklipidmembranes,amodelsystemthatmimicsthepropertiesofacellmembrane9.RegIIIaproducedrapidsingle-channel-likecurrentsat280mVinthepresenceofMg21ions(Fig.2a),http://www.77cn.com.cningtheNernst–Planck

(Extendedequationweestimatedthediameteroftheporeat,12–14A

DataFig.3).Thecalculatedporesizeagreedwiththelackofeffluxoffluoresceinisothiocyanate-dextran-10(FD10)orFD4,whichhave

and,28A ,respectively(Fig.2b).Incon-Stokesdiametersof,44A

)passedreadilythroughtheporestrast,carboxyfluorescein(,10A

(Figs1cand2b).TheseresultsshowthatRegIIIaformsfunctionaltransmembraneporesandyieldanestimateoftheinnerporediameter.Whenvisualizedbynegative-stainelectronmicroscopy(EM),numer- diameterwereobservedonliposomesouscircularstructuresof,100A

incubatedwithRegIIIa(Fig.2candExtendedDataFig.4a).AlthoughRegIIIaisamonomerinsolution10,thesizeoftheporessuggestedthattheywereRegIIIamultimers.Wethereforetreatedliposome-associatedRegIIIawithacrosslinkingagent,solubilizedtheproductsindetergent,andseparatedthembysize-exclusionchromatography(Fig.2d).Inaddi-tiontoaprominentmonomerpeakwedetectedasecond,liposome-dependentpeakatalowerretentionvolume,indicatingtheformationofamultimericcomplex.WesternblottingshowedasingleRegIIIaspecieswithmobilitysimilartothatpredictedforahexamer(Fig.2d),suggestingthattheporewasaRegIIIahexamer.

Afterlongerincubationswithlipid,RegIIIaformedfilaments(ExtendedDataFig.4b)similartothoseinpancreaticsecretions11.Thefilaments

indiameter,correlatingwiththedimensionsofthewere,100A

RegIIIapore(Fig.2c).RegIIIafilamentationrequiredlipidandwasdependentonRegIIIaporeformation,aspro-RegIIIaformedneitherporesnorfilaments(ExtendedDataFig.4b,d).FilamentationpartiallyinhibitedtheabilityofRegIIIatopermeabilizemembranes(ExtendedDataFigs4cand5a–c),asobservedwithothermembranetoxichostdefenceproteinswherefilamentationtrapsporecomplexesandlimitsdamagetohostcells12.ThesefindingsindicatethattheRegIIIafilaments

DepartmentofImmunology,TheUniversityofTexasSouthwesternMedicalCenter,Dallas,Texas75390,USA.2DepartmentofCellBiology,TheUniversityofTexasSouthwesternMedicalCenter,Dallas,Texas75390 …… 此处隐藏：41177字，全部文档内容请下载后查看。喜欢就下载吧 ……