interaction of lecithin-cholesterol acyltransferase with lipid … · 2017. 12. 15. · interaction...

Interactionoflecithin-cholesterolacyltransferasewithlipidsurfacesandapolipoproteinA-Iderivedpeptides:implicationsforthecofactormechanismofapolipoproteinA-IMarcoG.Casteleijn§,PetteriParkkila§,TapaniViitala,andArtturiKoivuniemi*

§co-firstauthorsDivisionofPharmaceuticalBiosciences,FacultyofPharmacy,UniversityofHelsinki,Finland*CorrespondenceArtturiKoivuniemi,TheDivisionofPharmaceuticalBiosciences,FacultyofPharmacy,UniversityofHelsinki,Helsinki,FinlandTel:+358451217612E-mail:[email protected]:Lecithin-cholesterolacyltransferase(LCAT)isanenzymeresponsiblefortheformationofcholesterylestersfromcholesterol(CHOL)andphospholipid(PL)moleculesinhigh-densitylipoprotein(HDL)particlesthatplayacrucialroleinthereversecholesteroltransportandthedevelopmentofcoronaryheartdisease(CHD).However,itispoorlyunderstoodhowLCATinteractswithlipoproteinsurfacesandhowapolipoproteinA-I(apoA-I)activatesit.Thus,herewehavestudiedtheinteractionsbetweenLCATandlipidsthroughextensiveatomisticandcoarse-grainedmoleculardynamicssimulationstorevealmechanisticdetailsbehindthecholesterolesterificationprocesscatalyzedbyLCAT.Inaddition,westudiedthebindingofLCATtoapoA-Iderivedpeptides,andtheireffectonLCATlipidassociationutilizingexperimental surface sensitive biophysical methods. Our simulations show that LCAT anchors itself to lipoproteinsurfacesbyutilizingnon-polaraminoacidslocatedinthemembrane-bindingdomainandtheactivesitetunnelopening.Meanwhile,themembraneanchoringhydrophobicaminoacidsattractcholesterolmoleculesnexttothem.Theresultsalsohighlighttheroleofthelid-loopinthelipidbindingandconformationofLCATwithrespecttothelipidsurface.TheapoA-I derived peptides from the LCAT activating region bind to LCAT and promote its lipid surface interactions,althoughsomeofthesepeptidesdonotbindlipidsindividually.Bymeansoffree-energycalculationsweprovidedahypotheticalexplanation for thismechanism.Wealso foundthat thetransfer free-energyofPL to theactivesite isconsistentwiththeactivationenergyofLCAT.Furthermore,theentryofCHOLmoleculesintotheactivesitebecomeshighlyfavorablebytheacylationofSER181.TheresultsprovidesubstantialmechanisticinsightsconcerningtheactivityofLCATthatmayleadtothedevelopmentofnovelpharmacologicalagentspreventingCHDinthefuture.Keywords: Lecithin-cholesteryl acyltransferase, apolipoprotein A-I, high-density lipoprotein,lipoproteins,lipidmembrane,activationmechanismINTRODUCTIONCoronaryheartdisease(CHD)anditscomorbiditiesareglobalhealththreatsshowinganincreasedprevalenceinindustrialaswellasindevelopingcountries.AlthoughmanydrugsfortreatingCHDexist,e.g.thecholesterol-loweringstatins,asubstantialresidualvascularriskremains(Fruchartetal.,2014).AccordingtotheResidualRiskReductionInitiativethisrepresentsaparamountpublichealthchallengeinthe21stcentury(Fruchartetal.,2014).Thus,moreeffectivemedicalstrategiesare needed to slow down and reverse this trend. It has been attributed that low high-densitylipoproteincholesterol(HDL-C)andelevatedtriglycerideslevelsarethetwokeycontributorsforthehighresidualCHDrisk(Fruchartetal.,2008).Therefore,inthelastdecade,anincreasedfocushasbeen on elevating the plasma concentration of HDL-C as a generally accepted intervention topreventorreducethedevelopmentofCHD(Perketal.,2012;Reiner,2013).Thisopposedtothelowering of high low-density lipoprotein cholesterol (LDL-C), which in turn accelerates the fataccumulation into arterial walls. Traditionally, a high HDL-C concentration in blood has been

.CC-BY-NC-ND 4.0 International licenseavailable under awas not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

The copyright holder for this preprint (whichthis version posted December 15, 2017. ; https://doi.org/10.1101/225037doi: bioRxiv preprint

https://doi.org/10.1101/225037

http://creativecommons.org/licenses/by-nc-nd/4.0/

regardedasapreventivemeasure reflecting theabilityofHDLparticles to transportcholesterolfromtheperipheraltissuesbacktotheliver,includingthemainlyLDL-derivedcholesterolfromthearterialwalls.Thisprocessistermedasreversecholesteroltransport(RCT)(LewisandRader,2005;Tall,1998).However,thereisagrowingbodyofliteratureshowingthattheelevationoftheHDL-Cby niacins or cholesterol ester transfer protein inhibitors does not or onlymoderately improvecardiovascularevents(Group,2017;RaderandHovingh,2014;TallandRader,2017).Inaddition,recentgeneticassociationstudieshaveraiseddoubtstowardstheHDL-ChypothesisandcausalityofHDL-CinthedevelopmentofCHD(Haaseetal.,2012;Pelosoetal.,2014;RaderandHovingh,2014).ThefailureofalargenumberofHDL-CraisingtherapiesandthefindingsofgeneticstudieshaveledtoasignificantuncertaintyconcerningthebenefitofraisingHDL-ClevelsinthetreatmentofCHD.Still,severalepidemiologicalstudiessupportaninversecorrelationbetweenHDL-CandCHD(Després et al., 2000; Gordon et al., 1977; Rader and Hovingh, 2014). Hence, the in-depthunderstandingofthestructure-functionrelationshipofHDLparticlesandtheenzymesprocessingthematthemolecularlevelisallthemoreimportantfortreatingCHD.

The lecithin cholesterol acyl transferase (LCAT; E.C. 2.3.1.43) is one of the key enzymesdrivingthematurationofnascentHDLparticlesinserum(Jonas,2000).Namely,LCATisresponsibleforlinkingtheacylchaincleavedfromaphospholipidtoacholesterol(CHOL)moleculeresultinginacholesterolester(CE)moleculewhichinturntransformsdiscoidalHDLparticlestosphericalones,which isa crucial step inRCT (Asztalosetal.,2007;FieldingandFielding,1995).Essentially, theesterificationofCHOLmoleculesincreasesthecholesterolloadingcapacityofHDLparticlesenablingamoreefficientRCT. Inadditionto its importance inRCT,mutations intheLCATgeneresults inmetabolicdisorderssuchasfamilialLCATdeficiencyandfish-eyediseaseinwhichthebody’sabilityto metabolize cholesterol is severed, leading to corneal lipid deposition, haemolytic anaemia,accumulationoffattymaterialinarterialwalls,andfinallyrenalfailure(CarlsonandPhilipson,1979;Kuivenhovenetal.,1997,1995).

TheinitialstepinthereactioncycleofLCATisitsbindingtoalipoproteinsurface.Previousresearch studies have established that apolipoproteins, especially the principal LCAT activatorapolipoprotein A-I (apoA-I), are playing a negligible role in the attachment rate of LCAT tolipoproteinsurfaces,butdecreasesitsdetachmentratefromsurfaces(Adimoolametal.,1998;Jinetal.,1999).ThissuggeststhatLCATinitiallybindstothelipidmoietyoflipoproteinparticleswhichisthenfollowedbyinteractionswithapoA-Ipromotingstrongerbindingandactivation.However,thespecificaminoacidsresponsible foranchoringLCAT initially to lipoproteinsurfaceshavenotbeen revealed, although deletion mutant studies have shown that amino acids 53-71 and thedisulphide bond formed between cysteines 50 and 74 of LCAT are crucial for the interfacialrecognition of LCAT (Adimoolam and Jonas, 1997; Jin et al., 1999; Frank Peelman et al., 1999).Furthermore, itwas shownbyMurrayetal. thatwhenLCATwasbound toHDLorhydrophobicsurfacesduring sink immunoassays, theepitopesof the121-136 regionwerenot accessible forantibodies(Murrayetal.,2001).Moreover,itisnotknownhowdeeptheLCATisburiedintothelipidmatrix,ifamembranebindingregionhasspecificinteractionswithdifferentlipidsthatcouldfacilitateitsactivity,andhowLCATisorientatedwithrespecttolipoproteinsurfaceswhenboundandnot-boundtoapoA-I.ThenextstepsinthereactioncyclearetheapoA-IdrivenactivationofLCATbyanunknownmechanismandthediffusionofphospholipids (PLs)andcholesterol totheactive site (Jonas, 2000; Sorci-Thomas et al., 2009). Consequently, the sn-2 chain of PLs ispreferentially lipolyzed and the acyl-intermediate of LCAT is formed in which the acyl chain iscovalentlyboundtoSER181,whichisapartoftheASP-HIS-SERcatalytictriadfoundalsoinotherlipasesbelongingtothea/bhydrolasefamily(FranconeandFielding,1991;Jonas,2000;FPeelmanetal.,1999;Winkleretal.,199AD).Next,theacylchainboundtoSER181istransferredtoafree



https://doi.org/10.1101/225037


cholesterol. Finally, the newly synthesized CE molecule diffuses from the active site into thelipoprotein.WhiletheLCATactivatorregionofapoA-I(centralhelixes5,6,and7,orresidues121-142,143-164.And165-186, respectively) is roughlyknownbasedontheapoA-ILCATdeficiencymutationsandvastamountofexperimentaldata,themorespecificLCATinteractionsiteandroleofthisinthedifferentreactionstepsremainsunclear(Sorci-Thomasetal.,1993,2009;Sorci-ThomasandThomas,2002).ByrevealingthemechanismsbehindthedifferentreactionstepsofLCATthewaymaybepavedforinventingnovelpositiveallostericmodulatorsofLCATthataimtoraiseHDL-CinamannerthatwouldbeneficialforthetreatmentofCHD.

For years, investigations concerning the LCAT lipid bilayer interaction and activation byapoA-IhavebeenhinderedbyalackofdetailedatomisticstructureofLCAT.Recently,however,X-raystructuresofLCAThavebecomeavailableenablingforexamplethecomputationalstudiesofLCATinteractingwithlipidsurfacesandapoA-Itoelucidatethemechanisticdetailsconcerningthecholesterolesterificationprocess (Glukhovaetal., 2015;RuwanthiNGunawardaneetal., 2016;Mantheietal.,2017;DerekE.Piperetal.,2015).FromthestructuresitisevidentthatthetertiaryandsecondarystructureofLCATishomologicaltolysosomalphospholipaseA2aspointedoutbythreesimilarfolds:thecap-domain,themembrane-bindingdomain,andthea/bhydrolasedomain(seeFig1A)(Glukhovaetal.,2015).Mostimportantly,structuraldetailsrevealthatLCATpossessesalid-loopthatcanmoveasidefromthetunnelopeningenablingtheentryoflipidsintotheactivesitewherethecatalytictriad is located in,thusbeingconsistentwithother lipases(RuwanthiN.Gunawardaneetal.,2016;DerekEPiperetal.,2015).

In this study, we carried out several atomistic and coarse-grained molecular dynamicssimulationsto investigatetheeffectofdifferentconformationalstatesofLCATto its interactionwithalipidbilayercomprisedofdioleylphosphatidylcholines(DOPC)andCHOLmolecules.TogaininsightsintoapoA-ImediatedLCATactivation,weutilizedquartz-crystal-microbalance(QCM)andmulti-parametric surface plasmon resonance (MP-SPR) experiments to investigate the effect ofapoA-IderivedpeptidesonthebindingofLCATtolipidbilayers,andthebindingofLCATtothesepeptides, respectively. These experiments were complemented with extensive free-energysimulations to reveal the role of the secondary structure of apoA-I derived peptides in lipidinteractions.Finally,westudiedtheenergeticsoflipidligandentrytotheactivesiteofLCAT.

Our simulations highlight the importance of specific non-polar amino acids in LCAT-lipidinteractions. In addition,we show that themembrane anchoring non-polar amino acids attractCHOLmoleculesadjacenttothem.Theresultsalsodemonstratethatthelid-loopplaysanimportantroleintheconformationofLCATwiththerespectoflipidsurface.Furthermore,theexperimentsindicatethatpeptidesderivedfromtheLCATactivatingregionofapoA-IbinddifferentlytoLCATand promote its lipid surface binding, although some of the peptides do not bind to lipidsindividually.We provide an explanation for thismechanism utilizing computational free-energycalculations.Itwasalsofoundthatthetransferfree-energyofPLfromthelipidbilayertotheactivesite is consistentwith the activation energy of LCAT. Finally, our results indicate that the acyl-intermediateofLCAThighlyfacilitatestheaccessibilityofCHOLmoleculesintotheactivesite.

RESULTSTheflexible lid-loopcoversnon-polaraminoacids locatedat thetunnelopeningofLCATfromwaterintheclosedstate

Toexplorethestructuralanddynamicalpropertiesofthelid-loopregionofLCAT,wecarriedoutmoleculardynamicssimulationsforLCATinthelidopenandclosedstatesinwater(Figure1A).



https://doi.org/10.1101/225037


Inthisfashion,westrivedtoestablishmechanisticinsightstotheconformationalswitchingofthelid-loop that, presumably, is aprerequisite for theesterification reactionof LCAT to takeplace.Moreover,theoverallstabilityoftheLCATstructurewasassessedagainsttheX-raystructurestovalidateourcomputationalmodels(RuwanthiN.Gunawardaneetal.,2016;Mantheietal.,2017;DerekEPiperetal.,2015).ToconstructapplicableLCATmodelsforthesepurposes,thelid loopregion (aa233-247)wascomputationallygraftedto thehighresolutionX-raystructuresofLCATrepresentingeithertheopenorclosedstate(RuwanthiN.Gunawardaneetal.,2016;DerekE.Piperetal.,2015).Duringthiswork,anotherhighresolutionX-raystructureofLCATwaspublished(PDBaccessioncode:5TXF)inthelid-closedconformationwithonlytwoaminoacidresiduesmissingfromthe lid-loop region (Manthei et al., 2017). This provided us an excellentmeans to validate ourgraftedLCATmodelintheclosedstate.AsshowninFigureS1ourmodelisingoodagreementwiththepublishedLCATstructureintheclosedstate.TheproducedLCATmodelsweresimulatedupto400nsinwatersurroundings.ThesesimulationsystemswerecoinedasLCAT-water-open-AAandLCAT-water-closed-AAsystems(SeemoredetailsinMaterialandMethodssection).

First, the simulation trajectorieswereutilized to investigate the conformational stability,dynamicsandsecondarystructurechangesofLCAT.AsitstandsoutfromFigure1A,thelid-loopexposestheactivesitefortheentryoflipidligandsinitsopenstateandshieldstheactivesitefromwaterintheclosedstate.Theroot-mean-squaredeviation(RMSD)profilesinFig1B(inset)indicatethatthebackboneatomsofLCATstabilizedafter50ns,andonlysmalldeviationsfromtheinitialstructuresoccurred.ThiswasreflectedbytheaverageRMSDsof0.14-0.17nmcomparedtotheinitialbackboneatomscaffolds.ByexaminingthemobilityofLCATasafunctionofresiduenumber(RMSFprofilesinFigure1B),itwasfoundthatthelid-loopregion(aminoacids233-247)showsthehighestconformationalfluctuationscomparedtotheotherstructuralpartsofLCATwhenneglectingthe first fewN-andC-terminal residues.Thiswasanexpectedresult.ConcerningthesecondarystructureofLCAT(Figure1C),theanalysisshowedastablestructurewithoutsignificantchangesduringthesimulations.Acloserinspectionrevealedthatthesecondarystructureofthelid-loopwasarandom-coil-richinbothconformationalstates,agreeingwiththeRMSFprofilesthatshowedahighflexibilityforthisregion.

Next, we analyzed the surface properties of LCAT and solvent accessible surface areas(SASAs)forthehydrophobicaminoacidslocatedattheactivesitetunnelopening(Figure2A)since

Figure1A)StructuresofLCATinthe lid-open(left)and–closed(right)statesrenderedascartoonmodels.Thecapdomainiscolouredwithpurple,thea/bhydrolasedomainwithorange,andthemembrane-bindingdomainwithred.The tunnelopeningof theactive site ismarkedwithadashedblack sphere.Thecatalytic triad residues: SER181,ASP345,andHIS377arerenderedwithsticksandcoloredaccordingtoelementtypes.Cyanatomsarecarbon,bluenitrogen,whitehydrogen,andredoxygenatoms.B)Rootmeansquarefluctuation(RSMF)profilesforthelid-openand–closedstates.Rootmeansquaredeviation(RMSD)resultsareshownintheinset.C)Secondarystructureprofilesforthelid-openand–closedstates.



https://doi.org/10.1101/225037


we hypothesized that the lid-loop shields these amino acids before LCAT detaches from thelipoproteinsurfacetothewaterphase.AsshowninFigure2B,thesurfacerepresentationsofLCATasafunctionofresiduetypes(non-polar,polar,andcharged)revealthathydrophobicaminoacidslocatedatthetunnelopeningoftheactivesiteareexposedtothewaterphaseintheopenstateandshieldedfromthewater in theclosedstate.This resultwassupportedbySASAcalculations(averagesfromthelast300ns)indicatingthatfouroffivehydrophobicresidueslocatedattheactivesite tunnel opening became shielded from the water in the closed state (Figure 2C). Morespecifically,theseaminoacidswereLEU62,PHE67,LEU117,andALA118.

Consideringthepreviousfindingthathydrogenbondsandsalt-bridgesplayaroleinthelid-loopconformationalswitchingofpancreaticlipase(vanTilbeurghetal.,1992),wealsoevaluatedtheaveragenumberofhydrogenbondsandsaltbridgesformedinthelid-closedand-openstates.Firstly,wecalculatedtheaveragenumberofhydrogenbondsformedbetweenthelid-loopandallLCATaminoacids.Secondly,weanalyzedtheaveragenumberofhydrogenbondsformedbetweenthelid-loopandthenon-lid-loopaminoacidsofLCAT.Theresultsrevealedthatthelid-loopformsfiveinternalhydrogenbondswhentheconformationalchangefromtheclosedtotheopenstatetakesplace(Figure2D).However,thesalt-bridgeanalysisindicatedthebreakageoftwosalt-bridgesformedbyARG244-ASP335andARG244-GLU241atthesametime(Figure2B).Ouranalysisdidnotregisterstablesalt-bridgesbetweenthelid-loopandthenon-lid-loopresiduesinthelid-openstate.Inaddition,nointernallyformedlid-loopsalt-bridgeswerefoundintheopenstate.

Figure2A)StructureofLCATshowingthehydrophobicaminoacidslocatedatthetunnelopeningoftheactivesite.LCATisrenderedwithcartoonrepresentation,andthedomainsarecoloredasinFig1.Additionally,thelidregionofLCATiscoloredwithblue(theopenstate)orred(theclosedstate).Non-polaraminoacidsinteractingwiththelidloopintheclosedstatearelabelledwiththecorrespondingresiduenamesandrenderedwithwhiteVanderWaalsspheres.B)ThesurfacepresentationsforLCATinthelid-openandlid-closedconformations.Thelocationofthehydrophobicregionismarkedwithblackanddashedsquares.Redindicatesnegativelycharged,bluepositivelycharged,greenpolar,andwhitehydrophobicaminoacidresidues.Thesalt-bridgesformedbythelid-looparemarkedwitharrows.C)Theaveragesolventaccessiblesurfaceareas(SASAs)fortheselectednon-polaraminoacidsinthelid-openand-closedstates.D)Theaveragenumberofhydrogenbondsbetweenthelidloopandallproteinresiduesornon-lidproteinresiduesinthelid-openand–closedstates.Errorsbarsarestandarddeviations.



https://doi.org/10.1101/225037


Thelid-openconformationenablesadeeperburialofthetunnelopeningnon-polaraminoacidsofLCATatlipidsurfaces.Intheprevioussection,weestablishedthatLCATexposesthenon-polaraminoacidslocatedatthetunnelopeningtowaterwhilethelidisopen.Next,weaskedifthesenon-polaraminoacidscouldalsointeractwithlipidbilayersresemblingthelipidmoietyofdiscoidalHDLparticles.Additionally,wehypothesizedthatthelid-closedconformationalsopreventstheburialofthetunnelopeningnon-polaraminoacidstoalipidmatrix,whichmightbeaprerequisitefortheentryoflipidligandsinadditiontothelid-openstate.Therefore,wecarriedoutatomisticandcoarse-grainedmoleculardynamicssimulationsofLCATtoinvestigatehowthesenon-polaraminoacidsinteractwithabilayercomposedofdioleoylphosphatidylcholines(DOPC)andcholesterol(CHOL).LCATwasinitiallyplacedon the surface of an equilibrated bilayer in awaywhich enabled the interaction of the tunnelopeningnon-polarresidueswiththelipidmatrix.Boththelid-closedand-openstatesofLCATweremodelled. The atomistic simulations (termed as LCAT-mem-open-AA and LCAT-mem-closed-AA)enabledustostudytheatom-scaleinteractionsbetweenLCATandindividuallipidsinamicrosecondtimewindow.On the other hand, the coarse-grained representations (LCAT-mem-open-CG andLCAT-mem-closed-CG)extendedthetimewindowuptoseveralmicrosecondsrenderingitpossibletoestimatee.g.thefree-energyofbindingofLCATtoalipidbilayerinlid-openand–closedstatesutilizingumbrellasamplingsimulations.

SimulationtrajectoriesrevealedthatLCATstaysatthesurfaceoflipidbilayerupto1µsor20µsinatomisticandcoarse-grainedsimulations,respectively(seeMoviesS1-S4).Whatstandsoutfromthetrajectoriesisthatthenon-polaraminoacidsofthetunnelopeningandthemembrane-binding region stay buried in the lipid matrix in LCAT-mem-open-AA and LCAT-mem-open-CGsimulations (Figure 3A and Movies S1-S2). However, the lid-closed conformation prevents thedeeperburialoftunnelopeningnon-polaraminoacidstothe lipidmatrix(MoviesS3-S4).Closerinspection by utilizing distance calculations with respect to the phosphorous atoms of DOPCs,revealedthatthehydrophobicaminoacids:L117,L64,F67,W48,L68,andL70wereclearlylocatedintheacylchainregionofphospholipidsinLCAT-mem-open-AAsimulation(seetheleftpanelinFig3B).Fromthedistanceanalysis,wecanseethattheseaminoacidsbecamelessburiedinthelipidsinthelid-closedconformation,especiallyaminoacidsA118,L117,L64,F67,W48,L68,andL70(seealsoMovieS3).AsshowninFigure3B,thistrendwasalsoseeninCG-simulationsfurtherverifyingourinitialhypothesisthatthelidconfigurationplaysanimportantroleinregulatingtheinteractionmode of LCAT with lipids. Further analysis revealed that the lid-closed state also changes theorientationofLCATwithrespecttothelipidbilayersurfacethatexplainsthedecreasedburialofA118,L117,andL64locatedatthehydrophobictunnelopening.ThisisshownbyahigheraveragetiltangleofLCATcalculatedbetweenthevector,formedbytheCa-atomsofMET49andASN131,andthelipidbilayernormal(Figure3BandMoviesS2-S4).

Next,weasked if thetunnelopeningandmembraneanchoringregionscan interactwithDOPCorCHOLmoleculesinamorespecificway.Tocharacterizethis,weanalyzedthe2D-numberdensitiesby firstdividing theXY-plane into100x100or36x36squaresdependingon thesystemstudied, atomistic or coarse-grained representation, respectively. This was followed by the



https://doi.org/10.1101/225037


calculationofthenumberdensitiesofDOPCandCHOLatomsineachsquare(SeemoredetailsinMaterialsandMethodssection).WefoundnospecificbindinginthecaseofDOPCmolecules,butCHOLmoleculespreferredtoaccumulatenexttothemembranepenetratingregionofLCATinbothatomisticandcoarse-grainedsimulations(Figure3CandFigureS2).

LCAT-interactionregionderivedpeptidesofApoA-IfacilitatethebindingofLCATtoalipidsurface.

TogainadditionalmechanisticinsightsregardingtheinteractionsofLCATwithlipidsandthepossibleroleofapoA-Iinthis,westudiedthebindingofLCATtoalipidbilayerwithandwithoutapoA-Iderivedpeptidesbyemployingthequartzcrystalmicrobalance(QCM)technique.TheapoA-I derived peptides were selected from the previously proposed LCAT-activation region (Sorci-Thomasetal.,2009)andwerebasedontheapoA-Iaminoacidsof122-142,135-155,and150-170

Figure3A)SnapshotsfromtheendofLCAT-mem-open-AAsimulation(1µs).ThecoloringofLCATdomainsisfollowing:Orangeisthealpha/beta-hydrolase,purplethecap,andredthemembrane-bindingdomain.GreysticksrepresentDOPCmoleculesandgreenspheresphosphorousatomsofDOPC.Watermoleculeshavebeenremovedfromthesnapshotsforclarity. Themembranepenetratinghydrophobic residues of LCATaremarkedand labelled to the snapshots. B)TheaveragecenterofmassdistancesfromthephosphorousatomsofDOPCsforthelipid-buriednon-polaraminoacidsinthelid-openandlid-closedmembranesimulations(left).TheaveragetiltangleofLCATwithrespecttothenormaloflipidmembrane inLCAT-mem-open-CGandLCAT-mem-closed-CGsimulations(right). Inaddition,snapshotsapproximatingtheaveragetiltangleineachcaseareshown.ColoringisthesameasinpanelA,butthetiltvectorformingaminoacids(ASN131andMET49)havebeenmarkedwithgreenspheres.C)The2D-numberdensitymapsforLCAT-mem-open-AAandLCAT-mem-open-CGsimulations.The centerofmass ofPHE67 ismarkedbya star showing the locationof themembranepenetratingregionofLCAT.



https://doi.org/10.1101/225037


(Figure4A).Inaddition,acontrolpeptidewaschosenfromtheregionofapoA-Icomprisingaminoacids185-205whichisnotinvolvedintheactivationofLCAT.TheeffectofmutationY166Fonthepropertiesofpeptide150-170wasalsostudiedsinceithasbeendemonstratedthatthismutationhamperstheactivityofLCAT(Wuetal.,2007).

TheQCMresultsinFig4BindicatethatLCATbindstothePOPCmembranewithoutapoA-IagreeingwithpreviousexperimentalstudiesandoursimulationresultsshowingthatLCATstaysatthelipidmembraneutilizingthetunnelopeningnon-polaraminoacidsormembrane-bindingregionintheattachment.TheresultsalsopointedoutthatsomeoftheapoA-Iderivedpeptidesdidnotinteractwithlipids,namelyonlypeptides122-142,185-205,and150-170-Y166Fwerefoundtobindto the lipidbilayer (SeeFig4B, solidblackcurve).Thecombined interaction studiesof LCATandpeptidesrevealedthatpeptideregions122-142,135-155,and150-170clearlyincreasedtheaffinityofLCATtothelipidbilayer(Figure4B).Resultsalsoshowedthatpeptides185-205and150-170-Y166FdidnotincreasethelipidbilayeraffinityofLCAT.

Aspeptides150-170and135-155didnotinteractwiththelipidsurfaceindividually,butstillincreasedtheoverallbindingofLCAT,wehypothesizedthattheputativeapoA-IinteractionsiteofLCAT could induce amphipathic a-helical structures for the peptides driving stronger lipidinteraction due to increased lipophilicity of the LCAT-peptide complexes. Another hypothesisconsideredthepossibleroleofthepeptidesinchangingtheconformationofthelidfromtheclosedtotheopenstateenablingastrongerinteractionofLCATwithlipids.

Totestthefirsthypothesis,wecalculatedthepotentialmeanforceprofiles(PMFs)foreachpeptidewhentheyaretransferredfromthewaterphasetothelipid-waterinterfaceutilizingCG-simulations. We carried out calculations with both coil and a-helix secondary structures to

Figure 4A) Schematic illustration showing thepeptides studiedalong the sequenceofapolipoproteinA-IB)Quartzcrystalmicrobalance(QCM)responsesforthepureLCAT(bluecurve)andpeptides(blackcurves)onthesurfaceofthePOPCbilayer.SurfaceresponsesofLCATinconjunctionwithpeptidesareshowninred.ThesumcurveofpureLCATandpeptidemeasurementsismarkedasadashedblackcurve.Thehorizontaldashedlinesmarktheminimumresponsesfordifferentmeasurements,includingtheirvalue.Theresultforpeptide135-155isnotshown,buttheprofileissimilarasinthecaseofpeptide150-170(SeeSI).C)Potentialmeanforce(PMF)calculationsfordifferentpeptidesincoil(black)orhelical(red)secondarystructureasafunctionofdistancefromthecenterofalipidbilayer.Thebindingfree-energiesforthepeptides(kJ/mol)aremarkednexttothedashedblacklines.



https://doi.org/10.1101/225037


investigatetheroleofthesecondarystructureofapoA-Iderivedpeptidesregardingthemembraneinteractions.TheresultsdepictedinFigure4Cindicatethatallpeptidescanbindtothesurfaceofthelipidbilayeriftheyadopta-helixconformation.However,noneofthepeptidesshowedafree-energyminimumwithinthemembraneregioninthecoilform.Therefore,forapeptideregiontoinducestrongerbindingofLCATtothelipidmembranesurfacetheymustfullyorpartlyadoptanamphipathica-helicalsecondarystructuretointeractwiththeunknownapoA-IinterfaceofLCAT.

Toassessthepossibilityofthesecondhypothesis,wecalculatedthebindingfree-energiesforLCATinopenandclosedstates.AsshownbythePMFprofilesinFigure5A,thelid-openstate(DGwater->lipid=-28±1kJ/mol)isnotabletobindsostronglytolipidswhencomparedtothelid-closedstate (DGwater->lipid = -34±1 kJ/mol), although the previous findings indicated a deeper lipidinteractionfortheopen-state(Figure3B).Thevaluefortheclosedstateiswellinagreementwithexistingexperimentallydetermineddissociationconstantsgivingbindingafree-energyvalueof-36kJ/molforLCATwheninteractingwithsmallunilamellarvesiclesintheabsenceofapolipoproteins(Jinetal.,1999).

Toalsoelucidatethespecificityofdifferentpeptide-regionsagainstLCAT,wecarriedoutmulti-parametric surface plasmon resonance (MP-SPR) experiments to determine dissociationconstants(KD)fortheLCAT-peptidecomplexes.Thepeptideswereattachedtostreptavidin-coatedgold-sensorsviaabiotinanchorandLCATwasflowedovertheanchoredpeptidestomeasuretheKD:susingaLangmuirmodel(SeemoredetailsinMaterialandMethods).FromtheMP-SPRresults(Figure5BandTable1) it canbededuced that the three strongest affinitiesof LCATwerewithpeptide-regions150-170,122-142,and135-155,whilethelowestaffinitieswereagainstpeptides185-205(control)and150-170-Y166F.TheseresultssuggestthatLCATpossessesamorespecificbinding site for peptides 150-170, 122-142, and135-155. This is also corroboratedby theQCMresultswhichcorrelatewellwiththeaffinityvaluesobtainedfromSPRmeasurements(Table1).

Figure5A)PotentialmeanforceprofilesforLCATasafunctionofdistancefromthecenteroflipidbilayer.Thetransferfree-energiesofLCATfromwatertothelipidbilayersurfacearemarkednexttothedashedlines(kJ/mol).B)Surfaceplasmon resonance peak responses for different apolipoprotein A-I derived peptides as a function of LCATconcentration.



https://doi.org/10.1101/225037


Table1MP-SPRderiveddissociationconstants(Kd)andRmaxvaluesfordifferentLCAT-peptidecomplexes.Inaddition,theQCMresponsedifferencesbetweenthecombinedLCAT+peptideandthesumofseparate(LCATandpeptidealone)measurementsaregiven(QCM-diff.).Theacyl-intermediateofLCATlurescholesterolmoleculestotheactivesiteTheentryoflipidstotheactivesiteofLCATisstillmechanisticallyanunknownprocess(Jonas,2000;Sorci-Thomasetal.,2009).Especially,thepossibledirectroleofapolipoproteinsisunknown.Togainadditionalinsightintothismatter,wecalculatedpotentialmeanforce(PMF)profilesforDOPCandCHOLmolecules when they were pulled from the lipid matrix to the active site of LCAT. PMFcalculationswere carried out utilizing the equilibrated conformation of LCAT in the LCAT-mem-open-AAsimulation.Weassumedthatthelidshouldbeintheopenstatebeforeentryofligands.

TheresultsinFigure6indicatethatthefree-energycostsofpullingDOPCorCHOLmoleculefrom the lipid matrix to the active site are approximately 65±2kJ/mol and 35±4 kJ/mol,respectively. The transfer free-energy cost of DOPC matched well with the experimental LCATactivationenergiesthatrangefrom53to76kJ/molmeasuredforHDL-LCATcomplexescomprisedofPOPCorDOPC lipids (Jonas,1986;ParksandGebre,1997).Next,weasked if theacylationofSER181 decreases the free-energy cost needed for CHOL to enter the active site. Thus, weconstructedanatomisticsimulationsystemwhereSER181wasacylatedandthelidwasintheopenstate(LCAT-mem-acyl-AA).Surprisingly,wefoundthattheacyl-intermediateofLCATdidnotonlydecreasethefree-energycostbutrenderedtheentryofCHOLtotheactivesitealmostfavorablereflectedbythetransferfree-energyvalueof4-7kJ/mol(Figure6A).Toinvestigatethedynamicsofthisprocessfurther,wecarriedoutacoarse-grainedsimulationwithacylatedSER181(LCAT-mem-acyl-CG).Thesimulationtrajectoryrevealedthatapproximatelyafter1.6µs(scaledMartinitime,seeTableII)acholesterolmoleculespontaneouslydiffusedtotheactivesite(Figure6CandMovieS5)agreeingwithourfree-energycalculations.ThiswasalsoregisteredbycalculatingthecontactsbetweenthesidechainatomsofSER181andthehydroxylgroupofCHOLmolecules(SeeFig6B).Inaddition, a closer inspection unveiled that the exchange of cholesterolmolecules between theactivesiteofLCATandthelipidmembranecanoccur iftheesterificationreactiondoesnottakeplace(MovieS6).

DISCUSSION

Inthepresentstudy,wehavestudiedtheinteractionsofLCATwithlipidsurfacesandapoA-IderivedpeptidestogainnovelinformationregardingthemolecularmechanismbehindtheLCATcatalyzedcholesterolesterificationtakingplaceatthesurfaceoflipoproteinparticles.Toachievethis,weemployedextensiveatomisticandcoarse-grainedmoleculardynamicssimulationsaswellasexperimentalsurfacesensitivebiophysicalmethods.

Peptide Kd(µM) Rmax QCM-diff.122-142 1.0258 139.1 -0.5135-155 0.8193 301.0 -1.0150-170 0.2903 25.4 -0.9150-170-Y177F 1.8442 174.4 0185-205 4.6901 103.4 -0.3



https://doi.org/10.1101/225037


Thesimulationresultsshowedthatthelid-regionofLCATformsacoilsecondarystructureaccompaniedwithrelativelyhighresidualfluctuationswhencomparedtotheotherstructuralpartsofLCAT.ThefactthatthelidregionsarepartlymissingfromtheX-raystructuresofLCATisbackingour simulation results showing a highlymobile regionwith no definite structural conformation(RuwanthiN.Gunawardaneetal.,2016;Mantheietal.,2017;DerekE.Piperetal.,2015).Anotherimportantfindingwasthatthehighlydynamicallid-loopshieldsthenon-polaraminoacidslocatedatthetunnelopeningfromthewater.Moreover,wesuggestthatwhenthelid-loopundergoestheconformationalchangefromtheclosedtotheopenstatetwosalt-bridgesarebrokenandfourintra-lidhydrogenbondsareformed.Thesefindingssupporttheviewthatthelidfunctionsasadynamicalgateregulatingtheaccessofphospholipidsandcholesterolmoleculesintotheactivesite.Ingeneral,lipasesareknowntopossessopenandclosedstates(Glukhovaetal.,2015;Meadetal.,2002;YangandLowe,2000).Forexample,theX-raystructuresofpancreaticlipaseindicatethattheopeningoftheactivesiteisaccompaniedwithconsiderablestructuralchangesofthelidregiongovernedbynon-polarcontacts,hydrogenbondsandsaltbridges(VanTilbeurghetal.,1993).Althoughwehavestudiedastructurallydifferentenzyme,thesefindingsareconsistentwithourresults.BasedonthecurrentfindingsitistemptingtospeculatethattheclosedstateofLCATisstabilizedinwaterbythelid-mediated shielding of the non-polar amino acids at the active site tunnel opening and thepresence of two salt-bridges. It is possible, therefore, that the entropic cost associated withconformationalswitchingfromtheclosedtoopenstateinwater,mainlyarisesfromexposingthelid-covered non-polar amino acids to aqueous surroundings. After LCAT becomes bound tolipoproteinparticles,presumablythisentropiccostdecreasessincethenon-polarresiduesatthetunnelopeningmaybecomeburiedinahydrophobiclipidmatrixasshownbyouratomisticandCG-simulations.Thus,it ispossibletohypothesizethatthiscouldbeoneofthereasonswhythelid-open state could be favored when LCAT is bound to lipoprotein particles. Yet, CG-simulationsrevealed that LCAT can strongly interact with lipids even in the lid-closed state which wasaccompaniedbyagreaterdistancebetweenthea/bhydrolasedomainandthelipidbilayersurface.Thisconformationdoesnotallowtheburialofthetunnelopeningnon-polaraminoacidsinlipids,whichweproposetobeaprerequisitefortheentryoflipidstotheactivesite.

Figure 6 A) Potential mean force profiles (PMFs) for DOPC and CHOL calculated in LCAT-mem-open-AA system. In addition, the PMF profile for CHOL in LCAT-mem-acyl-AA systems is shown. B) The number of contacts between the oleate chain beads linked to SER181 and cholesterol beads as a function of time. C) Snapshot from LCAT-mem-acyl-CG simulation showing the location of CHOL molecule in the active site of LCAT. Cholesterol molecules are rendered as orange spheres, DOPC molecules as grey sticks, the phosphate beads of DOPC as green spheres, LCAT backbone atoms as red sticks, the oleate chain beads linked to SER181 as cyan spheres, and SER181 as violet spheres.



https://doi.org/10.1101/225037


Interestingly,accordingtooursimulations,CHOLmoleculesprefertoaccumulatenexttothemembranepenetratingregionofLCAT.ThisfindingmaybeduetopackingdefectsbetweenLCATandphospholipids.ThetendencyofCHOLmoleculestoconcentrateadjacenttothetunnelopeningnon-polaraminoacidsindicatesitsimportantroleinregulatingtheaccessibilityofCHOLmoleculesintotheactivesiteoverphospholipids.ThiscouldbeoneofthereasonswhyLCATisthesoleenzymeamonglipasescatalyzingtheesterificationofCHOLmolecules.Thisissuecanbefurtherclarifiedbyexaminingthelipidsurfaceinteractionofotherlipases.

Our QCM measurements revealed that the peptides derived from the LCAT-interactionregionofapoA-IincreasedthebindingofLCATtolipidsurfaces,whilethecontrolpeptide185-205and themutantpeptide150-170-Y166Fdidnot facilitate theLCAT interactionwith lipids.ThesefindingsagreewithpreviousexperimentalreportsshowingthatapoA-IincreasesthebindingofLCATtolipidscomparedtoapolipoprotein-freesmallunilamellarvesicles(Jinetal.,1999;Jonas,2000).EarlierithasbeenshownthattheapoA-ImutationY166FdecreasesthesurfacebindingandactivityofLCATwhichisinagreementwithourexperimentalresults(Guetal.,2016;Wuetal.,2007).Ingeneral,many studies have pointed out that the charged and polar amino acids located in thecentralpartsofapoA-IareimportantinLCATactivity(Sorci-Thomasetal.,2009;Sorci-ThomasandThomas,2002).Oneofourunanticipatedfindingwasthatalthoughsomeofthepeptidesdidnotinteractwithlipidsalone,theystillcontributedstronglytothelipidattachmentofLCAT.Namely,peptides135-155and150-170didnotinteractwithlipidsatall,andpeptide122-142showedonlyamodestinterfacialactivity.Nevertheless,thesepeptidesgreatlyincreasedthebindingofLCATtolipid surfaceswhencompared to themore surface-activepeptides150-170-Y166Fand185-205.Interestingly, the computational free-energy calculations revealed that all the peptides studiedmustadoptfullyorpartialamphipathicalpha-helicalconformationsbeforetheycaninteractwithlipidsurfaces.ItcanthusbesuggestedthattheLCAT-activatingcentralregionofapoA-ImayadoptwaterexposedloopconformationsbeforeinteractingwithLCAT.TheinteractionofloopstructureswithLCATcouldbefollowedbyaconformationalchangefromthecoiltoa-helixleadingtostrongerbindingofLCATtoHDLparticles.TheMP-SPRresultssupportthisviewsinceitwasshownthattheapoA-IderivedpeptidescanbindtoLCATwithoutthepresenceofalipidsurface.Further,basedonthefree-energycalculationsofLCAT,itisevidentthatapoA-IderivedpeptidesdonotincreasethebindingofLCAT-peptidecomplexestolipidsurfacesbymerelychangingthelidfromclosed-totheopen-statesincethelipidbindingfree-energyfortheopenstateofLCATwaslowercomparedtotheopenstate.Overall,thesefindingsstrengthentheresultsshowninthepreviousexperimentalstudiesdemonstratingthatthecentralregionofapoA-Ipossessesmuchlowera-helicalstabilitiesand,thus,increasedthetendencytoformloopedstructurescomparedtootherregions(Chettyetal.,2013;SevuganChettyetal.,2012;Wuetal.,2007).

Thefree-energycalculationsproducedinthepresentstudysupporttheviewthattheentryof lipids into theactivesiteofLCAT likelyoccurswithout thedirect involvementofapoA-I.Onemajorfindingsupportingthisviewwasthatthecalculatedtransferfree-energyofDOPCfromthelipid bilayer into the active site (65 kJ/mol) is congruent with the experimentally determinedactivationenergiesofLCAT inreconstitutedHDLparticleswithvaryingPLspecies(53-76kJ/mol)(Jonas,1986;ParksandGebre,1997).ThisresultisalsoconsistentwiththedatademonstratingthattheapolipoproteincontentofHDLparticlesdoesnotaffecttheactivationenergyofLCAT(Jonasetal.,1987). ThesecondmajorfindingwasthattheacylationofSER181rendersthetransferfree-energy of cholesterolmuch lowerwhen compared to the non-acylated case (4-7 kJ/mol vs. 35kJ/mol).Withrespecttothisfinding,weobservedthatCHOLmoleculesspontaneouslydiffusedintotheactivesiteinCG-simulationswhenSER181wasacylated.TheCG-simulationsfurtherrevealedthattheexchangeofCHOLmoleculesintheactivesitetookplacein20µs.Thus,theatomisticfree-



https://doi.org/10.1101/225037


energycalculationsandCG-simulationsshowedthatthediffusionoflipids,especiallyinthecaseofCHOLmolecules,totheactivesiteofLCATisnotdirectlyassistedbyapoA-I.Interestingly,arecentresearch study revealed that molecular agents targeted for the lipid binding site of LCAT canincreasetheVmaxofLCAT(Freemanetal.,2017).Surprisingly,thecompoundalsoactivatedLCATdeficiency causingmutantswhich raiseshopes for treating thesedisorders in the future. ItwasidentifiedwithmolecularmodellingthattheLCATactivatingcompoundAformedahydrophobicadductwithCYS31locatedintheactivesite.Thus,onereasonforthehigherLCATactivitycouldbethatthecompoundArenderstheactivesiteofLCATenergeticallymorefavorabletoPLstodiffusein.Inotherwords,thedruglowerstheactivationenergyofLCAT,whichisessentiallycoveredbythetransferofaPLfromthelipidmonolayertotheactivesite.

Toconclude,ourresultssuggestthattheinitialbindingofLCATtolipoproteinsurfaceoccursbythehelpofnon-polaraminoacidsinthemembrane-bindingdomain.TheinitiallipidinteractionofLCATislikelyfollowedbyayetunknowninteractionwithapoA-IwhichenablestheopeningofthelidandtheconformationalchangeofLCATtotheconfigurationwherethetunnelopeningnon-polar aminoacidsbecomeburied in lipids. In addition, free cholesterolmolecules are attractedadjacenttothetunnelopening.Meanwhile,theaminoacids135-170ofapoA-Iincreasesthebindingof LCAT to lipoproteins. Thus, apoA-I enables the diffusion of PLs to the active site viaconformationalandstructural changeswhich leads to the formationof theacyl-intermediateofLCAT.This,inturn,greatlyfacilitatesthediffusionofCHOLmoleculestotheactivesitewithoutthedirect involvementofapoA-I.Theevidence fromthis studysuggests that the roleofapoA-IandotherapolipoproteinsinactivatingLCATistoadjusttheconformationofLCATwithrespecttothelipidsurface,increasethebindingofLCATanddrivethelidtotheopenstate.ThereasonsforvaryingLCAT activation potencies of different apolipoproteins could be speculated to arise from theirdifferentcapacitytocontrolthesefeatures(Jonas,1986).WhatishighlysurprisingisthattheapoA-IderivedpeptidescanincreaseLCATbindingtolipids,evenifsomeofthemdonotinteractwithlipids individually. These findings encourage further investigations and modifications of thesepeptides aiming to increase the activity of LCAT. Eventually, this approachmay i.e. lead to thedevelopment of more advanced apolipoprotein mimetic peptides that could be used in thetreatmentofdifferentmetabolicdisorderssuchasdyslipidaemiaandLCATdeficiencies(Navabetal.,2015,2010).Overall,thecurrentresearchstudyprovidesanovelframeworkfortheexplorationofLCATactivationbyapolipoproteins,apolipoproteinmimeticpeptides,andotherpharmacologicalcompounds.MATERIALSANDMETHODSComputationalproceduresConstructionofsimulationsystemsThehigh-resolutionX-raystructuresofLCATwereacquiredfromtheBrookhavendatabankwiththeaccessioncodesof4XWGand5BV7(RuwanthiNGunawardaneetal.,2016;DerekE.Piperetal.,2015).Inaddition,allthemutatedresidues(C31Yin4XWG;L4FandN5Din5BV7)werechangedbacktoonesmatchingwiththenativeLCAT.The4XWGstructureisconsideredtobetheclosedand5BV7structuretheopenconformationofLCATbasedonthelid-looporientationwithrespecttotheactive site tunnel opening. TheModeller programwasutilized to construct themissing lid-loopregionsinbothstructures(WebbandSali,2014).ThedefaultmodellingparameterswereutilizedandthestructureswiththehighestDOPEscoreswerechoseninbothcasesforfurtherstudies.



https://doi.org/10.1101/225037


TwoLCAT-in-watersystemswereconstructed:onefortheopenandonefortheclosedlid-loopconformation.Intherestofthearticle,werefertothesesystemsbyabbreviationsLCAT-water-open and LCAT-water-closed. The dimensions of LCAT-water-open and LCAT-water-closedsimulationboxeswere8.5x7x7nm.TheopenandclosedLCATenzymesweresolvatedwith12000watermolecules.Inaddition,ninesodiumionswereaddedtoneutralizethesimulationsystems.Inthe construction of LCAT-lipid bilayer systems, we used a pre-equilibrated DOPC/CHOL bilayerconsistingavailableinthewebsiteoftheSlipiddevelopers.Toconstructatomisticlipidmembranesystems,theLCATwasplacedonthesurfaceofDOPC/CHOLbilayersothatthenon-polarresiduesofthetunnel-openingandthemembranebindingdomainwereburiedinthelipidmatrixandtheactivesitetunnelopeningwaspointingtowardsthelipidbilayer.Thiswasdoneforboththelid-open and –closed LCAT structures. TheDOPC andCHOLmolecules overlappingwith LCATwereremovedfromthesystems.Followingthis,thesystemsweresolvatedapproximatelywith20000water molecules and nine sodium ions were added to neutralize the systems. One additionalatomistic systemwas constructed for the open LCAT lid-loop conformationwhere SER181wasacylatedwithanoleicacid.Werefertotheseall-atomlipidmembranemodelsbyLCAT-mem-open-AA, LCAT-mem-closed-AA, and LCAT-mem-acyl-AA abbreviations. The starting structures for thecorrespondingcoarse-grainedLCAT-lipidsystemsrepresentingthesethreeatom-scalemodelswereconstructed similarly (LCAT-mem-open-CG, LCAT-mem-closed-CG, and LCAT-mem-acyl-CG). InTable2,amoredetailedlistofmoleculesineachsimulationsystemispresented.Simulationforcefieldsandparameters.MoleculardynamicssimulationswerecarriedoutwiththeGROMACSsimulationspackage(version5.1.2)(Berendsenetal.,1995;Hessetal.,2008).TheSlipidsandAMBER99SB-ILDNforcefieldswereutilized for lipids and protein molecules in atomistic simulations, respectively (Jämbeck andLyubartsev,2012;Lindorff-Larsenetal.,2010).TheMARTINIforcefieldwasusedforthecoarse-grained(CG)representations(Marrinketal.,2007;Monticellietal.,2008).ThepolarizablewaterandproteinmodelsoftheMARTINIforcefieldwereutilized(DeJongetal.,2013;Yesylevskyyetal.,2010).

Concerning the atomistic simulations, all systems were first energy minimized with thesteepest descent algorithm. After this, short 10 ns equilibrium simulationswere carried out tostabilize pressure fluctuations that were followed by 400 ns (LCAT-water systems) or one-microsecondproductionsimulations(LCAT-memsystems).Thetimestepwassetto2fs,andthetemperature and pressure were maintained at 310 K and 1 bar, respectively. The Berendsencouplingschemeswereutilizedforachievingconstanttemperatureandpressureduringtheshortequilibriumperiodofatomisticsystems(Berendsenetal.,1984).AfterthistheNose-HooverandtheParrinello-Rahmancouplingalgorithmswereemployedfortreatingtemperatureandpressureintheatomisticsystems,respectively(EvansandHolian,1985;ParrinelloandRahman,1981).Theisotropic or semi-isotropic pressure coupling scheme was used in LCAT-water and LCAT-memsystems,respectively.Waterplusions,lipids,andproteinwereseparatelycoupledtoheatpaths.FortheLennard-Jonesinteractionsacut-offof1nmwasused,andtheelectrostatic interactionswerehandledwiththeparticle-meshEwaldmethodwitharealspacecut-offof1.0(Essmannetal.,1995).TheLINCSalgorithmwasappliedtoconstraincovalenthydrogenbondlengths(Hessetal.,1997).

Regardingthecoarse-grainedsimulations,thetemperatureandpressurewerehandledwiththev-rescaleandtheParrinello-Rahmanschemes,respectively (Bussietal.,2007;ParrinelloandRahman,1981).Couplingconstantsof1ps-1and12ps-1wereusedforthetemperatureandpressureschemes,respectively.Lipids,proteinandwatermoleculeswereseparatelycoupledtoaheatpath.



https://doi.org/10.1101/225037


Thereaction-fieldfieldelectrostaticsandLennard-Jonesinteractionswereemployedwithcut-offsof1.1nm.Therelativeelectrostaticscreeningwassetto2.5sincethepolarizablewaterandproteinmodels ofMARTINI force field were used. Time step of 20 fs was used and all coarse-grainedsimulationsweresimulatedupto20us(scaledMartinitime).TheElNeDynschemewasemployedforLCAT(Perioleetal.,2009).

Table2Simulatedsystemswiththeirmolecularcompositionsandsimulationtimes.

System Details DOPC/POPC CHOL Water(mol./beads) Totalsim.time(µs)

LCAT-water-open-AA1 Lidopen 0 0 12000 0.4

LCAT-water-closed-AA Lidclosed 0 0 12000 0.4

LCAT-mem-open-AA Lidopen 286 132 20000 1

LCAT-mem-closed-AA Lidclosed 286 132 20000 1

LCAT-mem-acyl-AA SER181acylated 286 132 20000 1

LCAT-mem-open-CG1 Lidopen 226 105 10000 202

LCAT-mem-closed-CG Lidclosed 226 105 10000 202

LCAT-mem-acyl-CG SER181acylated 226 105 10000 202

LCAT-mem-open-AA-DOPC-PMF1 Lidopen 286 132 20000 1.6LCAT-mem-open-AA-CHOL-PMF Lidopen 286 132 20000 2.0LCAT-mem-acyl-AA-CHOL-PMF SER181acylated 286 132 20000 1.7LCAT-mem-open-CG-LCAT-PMF Lidopen 226 105 10000 962

LCAT-mem-closed-CG-LCAT-PMF Lidclosed 226 105 10000 962

Pep_122-142-CG-PMF Helixorcoil 128 0 1800 82




Pep_150-170-CG-PMF-Y166F Helixorcoil 128 0 1800 82

1AAandCGstandforall-atomandcoarse-grainedrepresentations,respectively.PMFmeanspotentialmeanforcecalculations.2Thetotalsimulationtimeofthecoarse-grainedsystemswasmultipliedbyfoursincethediffusiondynamicsofMARTINIwaterbeadsisapproximatelyfourtimesfastercomparedtorealwater(Marrinketal.,2007;Yesylevskyyetal.,2010).

Free-energycalculationsForcalculatingthefree-energyprofilesforPCorCHOLmoleculesenteringtheactivesiteofLCAT,eitheraCHOLoraPCmoleculewaspulledfromthelipidbilayertotheactivesiteofLCATinthelid-open state togenerateumbrellawindows (these systemsare coinedas LCAT-mem-open-DOPC-PMF-AAandLCAT-mem-open-CHOL-PMF-AA).Thefree-energyprofilesforPCandCHOLmoleculeswere calculated utilizing atomistic simulations. In addition, the free-energy profile for a CHOLmoleculecalculatedintheLCAT-memsystemswhereSER181wasacylated(LCAT-mem-acyl-CHOL-PMF-AA).TheoxygenatomofCHOLorphosphorousatomofDOPCwaspulledtowardsthecenterofmassofSER181residuewithapullrateof0.001nm/psandaforceconstantof10000kJ/mol-nm2.Theresultingreactioncoordinatewasdividedinto39umbrellawindows.Theforce-constantofthespringwassetto2000kJ/mol×nm2.Eachumbrellawindowwassampledupto40-50ns,thusthetotalsimulationtimerangedfrom1.6to2.0µspersystem.Thelast34-28nsofthesimulationtrajectorieswereusedtoconstructthePMFprofiles.

Fordeterminingthelipid-bindingfree-energiesfordifferentapoA-Iderivedpeptidesatthecoarse-grainedlevel,eachpeptidewaspulledfromthewaterphasetothecenterofPOPCbilayerwithapullrateof0.001nm/psandaforce-constantof20000kJ/mol×nm2.Thiswasfollowedbyagenerationof20umbrellawindowsalongthebilayernormal.Eachwindowwassampledupto100ns(400nsinscaledMARTINItime)andthecentersofmassofpeptideswereconstrainedtothecenterofeachumbrellawindowbyusingaforceconstantof500kJ/mol×nm2.Thelast320nsofthe



https://doi.org/10.1101/225037


simulationtrajectorieswereusedtoconstructthePMFprofiles.Toestimatetheeffectofsecondarystructuretothebindingoffree-energies,potentialmeanforceprofileswerecalculatedforpeptidesfullyadoptingalpha-helicalorcoilsecondarystructures. In addition to the above systems, also the free-energy profiles for LCAT along the lipidmembrane normal was estimated. First, LCAT was pulled from the surface of DOPC-CHOLmembrane to thewater phase by using a constant pulling speed of 0.0002 nm/ps and a forceconstantof20000kJ/mol×nm2.ThepullingwasdoneforendstructuresofLCAT-mem-open-CGandLCAT-mem-closed-CG simulations. Next, the reaction coordinate was divided into 30 differentumbrellawindows.Eachwindowwassampledupto0.8µs(3.2µsinscaledMARTINItime)inwithaforceconstantof1000kJ/mol×nm2.Thus,thetotalsimulationtimeforeachLCATsystemwas96µs(scaledMartinitime).Thelast60µsofthesimulationtrajectorieswereusedintheconstructionofthePMFprofiles.Analysismethods

All the analysis programs reported here are part of the GROMACS simulations packageunlessmentionedotherwise.Thegmxrmsfandgmxrmsdprogramswereusedtoproducerootmeansquarefluctuation(RMSF)anddeviation(RMSD)graphs.OnlythebackboneatomsofLCATwereincludedintheRMSDandRMSFanalysis.TheRMSFprofileswerecalculatedasafunctionofaminoacidresiduesofLCATaftertheRMSDprofileswerestabilized(after100ns).ThesecondarystructureofLCATwasmonitoredasafunctionoftimewiththesecondarystructurepluginoftheVMD(Humphreyetal.,1996).Theaveragenumberofhydrogenbondswascalculatedutilizingthegmx hbond analysis tool. Themaximumdistance between a donor and an acceptorwas set to0.35nm,and themaximumangle formedbyhydrogenbondingatomshydrogen-donor-acceptorwassetto30degrees.Thenumberofsalt-bridgesinthestructureofLCATconcerningthelid-regionwascalculatedbythegmxmindistprogramwithadistancecriterionof0.6nmbetweenthechargedsidechaincentralcarbonatomsofnegativelyandpositivelychargedaminoacids(GLU-ARG,GLU-LYS,ASP-ARG,andASP-LYS).Thesolventaccessiblesurfaceareas(SASA)fortheselectedaminoacidswerecalculatedusingthegmxsasaprogram.Thedistanceandtiltanalysiswerecarriedoutwiththegmxmindistandgmxgangleprograms.TheseanalyseswereconductedafterthenumberofcontactsbetweenproteinandheadgroupatomsofDOPCwereinequilibrium.Forproducing2D-spatialdensitymaps thegmxdensmapprogramwasused.TheXY-planesof simulation systemsweredividedinto100x100or36x36squarebins,inwhichthenumberofCHOLatomswascalculatedineachtimeframeandsummedoverthewholesimulationtimetoproducenumberdensityprofiles.Before thecalculation, themembranepuncturing regionofLCAT (thecenterofmassof residuePHE67)wasplacedtothecenterofthesimulationbox.Thegmx_whamprogramwasutilizedtoderivethepotentialmeanforceprofilesfromtheumbrellasamplingsimulations.Thebootstrappingmethodwith50binswasusedtoestimatetheerrorsforthePMFprofiles.AllvisualizationsinthearticleweremadeeitherwiththePythoncodesortheVMDvisualizationpackage(Humphreyetal.,1996).ExperimentalproceduresMaterialsPOPC(1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine),dissolvedinchloroform(25mg/mL)wasobtained from Avanti Polar Lipids (Alabaster, USA). Calcium chloride, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 3-[(3-)cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), sodium chloride, ethylene diaminetetraacetic acid (EDTA), sodium



https://doi.org/10.1101/225037


azide,andpotassiumdihydrogenphosphatewereobtainedfromMerckSigma-Aldrich(Darmstadt,Germany). Potassium chloride was obtained from Honeywell Riedel de Haën (Seelz, Germany),disodiumhydrogenphosphatewasobtainedfromFisherScientific(Hampton,USA).PeptidesandN-terminalbiotin-labeledpeptidesweresynthesizedbyPeptideProteinResearchLtd,Hampshire,U.K.andhadthefollowingaminoacidsequences:(i)122-142:LRAELQEGARQKLHELQEKLS(ii)135-155:HELQEKLSPLGEEMRDRARAH(iii)150-170:DRARAHVDALRTHLAPYSDEL(iv)150-170-Y166F:DRARAHVDALRTHLAPFSDEL (v) 185 - 205: GGARLAEYHAKATEHLSTLSE. Lyophilized, glycosylatedhumanLCATprotein(SinoBiologicalInc.,Beijing,China)wasreconstitutedwith400μlofultrapure,sterilewatertoyieldLCATat5μMinPBS(manufacturerspecifications).Ultra-purewaterusedforpreparationofthebufferandinallmeasurements,werepreparedwithaMilli-Qpurificationsystem,havingaresistivityof18MΩ·cmandTOClevelof<5ppm.Thefollowingbufferswereused:20mMphosphatebuffersaline(PBS),1mMEDTA,1mMNaAzide(pH7.4;bufferA);20mMHEPES150mMNaCl(pH7.4;bufferB).Themulti-parametricsurfaceplasmonresonancemeasurementsMeasurementswereperformedwithamulti-parameterSPRNavi™200(BioNavisLtd.,Tampere,Finland) instrument. The setupwasequippedwith two incident laserwavelengths, 670nmand785 nm, two independent flow channels, inlet tubing and outlet (waste) tubing, and an auto-sampler. Both of the flow channels were measured simultaneously with 670 nm and 785 nmincidentlight.Themeasurementtemperaturewaskeptconstantat20°C.BiotincoatedSPRsensors(Bionavis Ltd., Tampere, Finland) were placed in the flow-cell. The flow rates used for thestreptavidininteractionwithbiotin,biotin-peptideswithstreptavidin,andforLCATinteractionwithimmobilizedpeptideswere10μL/min.SPRspectrawererecordedafterintroductionofbufferAintotheflow-cellfor10–30minuntilastablebaselinewasachievedafterwhichthesurfacewasfullysaturatedwithstreptavidinduringa10minutesserialinjectionat1.9μMtosaturatetheSPRsensor.Itwasobservedthatafter5minutestheSPRsensorwasfullycoatedwithstreptavidin,thereforethe method was adapted to 5 minutes. In the second phase of the measurement, in DMSOsolubilizedbiotin-peptideswithafinalconcentrationof100µMwereinjectedinparallelintobothflowchannelsfor7minutesat10μMtocovertheSPRsensorfully,i.e.untilallstreptavidinbindingssidewerefullyoccupied.Parallelinjectionsensuredthat2peptideswereanalyzedsimultaneously.IncreasingconcentrationsofLCATweretheninjectedinbothchannelsviaserialinjection,toensurethelowestvariationofLCATinjectionfor15minutesatincreasingconcentrations(156,313,625,1250,2500,and5000nM)todeterminethebindingconstantoftheLCAT-peptidecomplexes.Thedata was fitted in OriginPro (v. 8.6, OriginLab Corp., Northampton,MA, USA) according to theLangmuirmodeltoobtainthedissociationconstants(KD)andtheresponsesofsaturatedbinding(Rmax):

𝑅 =𝑅$%&

1 + 𝐾* 𝑐,

whereRisthefinalresponseaftereachinjectionandcistheconcentrationofLCATineachinjection.ThequartzcrystalmicrobalancemeasurementsPriortotheQCMmeasurements,thePBSbufferformedafterreconstitutionLCATwasexchangedforbufferAduetotheextremehighsensitivityoftheQCMmethodtoslightdifferencesinionicstrength of the background buffer. The LCAT solution was placed in an Amicon Ultra-0.5 10KCentrifugalFilterDevice(MerckMilliporeLtd,TullagreenCarrigtohill,Ireland)andequalvolumesofbufferAwerefilteredthroughtentimesat14000xg,afterwhich400μlofbufferAwasaddedto



https://doi.org/10.1101/225037


fullyrecoverLCATat5μM.ThepeptidesweredissolvedintobufferAatafinalconcentrationof100µM.An impedance-based quartz crystal microbalance instrument (KSV Instruments Ltd, Helsinki,Finland)was used for themeasurements. Themeasurement temperaturewas kept constant at20 °C. Before measurements, silica-coated QCM sensors (Q-Sense Inc./BiolinScientific, VästraFrölunda,Sweden)werefirstflushedwith70%ethanolandultrapurewater,driedundernitrogenflow,andfinallyoxygenplasma-treated(PDC-002,HarrickPlasma,Ithaca,USA)for5minutesat29.6Wand133-173Pa.Sampleswereinjectedintothemeasurementchamberwithaperistalticpumpsystem (Ismatec/Cole-Parmer GmbH, Wertheim, Germany). Small unilamellar POPC liposomevesicles (SUVs)whichwereused to formsupported lipidbilayers inside theQCMmeasurementchamberweremadeusingthinfilmhydrationmethodfollowedbyextrusion(11times)througha50-nmpolycarbonatefiltermembraneat60°C.TheresultingSUVs(10mg/mLinbufferB)hadamean number average particle size of 65 ± 7 nm and polydispersity index of 0.167 ± 0.008(determined from three individual measurements by Zetasizer APS instrument, MalvernInstrumentsLtd,Worcestershire,UK).After thebaselineof the signalatdifferent frequencyovertones (3,5and7)was stabilizedwithbufferB,supportedlipidbilayers(SLBs)wereformedbyflowingliposomesolution(0.15mg/mLinbufferB+5mMCaCl2at250µL/min)overthecrystalsurfacefor5minutes.Afterwards,bufferAwas injected,and the signalwas left to stabilize.Anadditional injectionofultrapurewaterwasadded to ensure the complete formation of an SLB and absence of intact vesicles on themeasurementsurface.Theoverlapofthechangesinnormalizedovertonefrequencies(-26Hz)wasconsideredasaconfirmationofthepresenceofagood-qualitySLB.Theflowspeedwasreducedto50µL/minandthemeasurementswereperformedbyinjectingpeptidesat100µMinbufferA,LCATat37.5nMinbufferA,orcombinationofLCATandpeptidethroughthemeasurementchamberforadurationof10minutes.Betweeneachmeasurement,sensorswerecleanedinsitubysequential2minuteinjectionsof20mMCHAPS,2%Hellmanex,70%ethanol,andultrapureH2O.DataanalysiswasperformedusingOriginPro(v.8.6,OriginLabCorp.,Northampton,MA,USA).Foreachmeasurement,frequencyovertonesignals(3,5and7)werenormalized,averagedandbaselinecorrected. Neither LCAT nor peptides induced changes in viscoelastic properties of the bilayer,whichwasseenasnegligiblechangesintherecordedenergydissipation.ACKNOWLEDGEMENTSAcademyofFinlandisthankedforprovidingfinancialsupportforMC(keyprojectfunding,projectno.303884),AK(postdoctoralgrant,projectno.298863) andTV(projectno.137053and263861). PPacknowledgesagrantfromtheFinnishPharmaceuticalSociety.CSC–ITCentreforScienceLtd.(Finland)isacknowledgedforcomputationalresources.ThisarticleisdedicatedtothememoryofDr.MarjaT.Hyvönen.REFERENCESAdimoolam,S.,Jin,L.,Grabbe,E.,Shieh,J.J.,Jonas,A.,1998.Structuralandfunctionalproperties

oftwomutantsoflecithin-cholesterolacyltransferase(T123IandN228K).J.Biol.Chem.273,32561–32567.



https://doi.org/10.1101/225037


Adimoolam,S.,Jonas,A.,1997.Identificationofadomainoflecithin-cholesterolacyltransferasethatisinvolvedininterfacialrecognition.Biochem.Biophys.Res.Commun.232,783–787.https://doi.org/10.1006/bbrc.1997.6375

Asztalos,B.F.,Schaefer,E.J.,Horvath,K.V,Yamashita,S.,Miller,M.,Franceschini,G.,Calabresi,L.,2007.RoleofLCATinHDLremodeling:investigationofLCATdeficiencystates.J.LipidRes.48,592–9.https://doi.org/10.1194/jlr.M600403-JLR200

Berendsen,H.J.C.,Postma,J.P.M.,vanGunsteren,W.F.,DiNola,A.,Haak,J.R.,1984.Moleculardynamicswithcouplingtoanexternalbath.J.Chem.Phys.81,3684–3690.https://doi.org/10.1063/1.448118

Berendsen,H.J.C.,vanderSpoel,D.,vanDrunen,R.,1995.GROMACS:Amessage-passingparallelmoleculardynamicsimplementation.Comput.Phys.Commun.91,43–56.https://doi.org/10.1016/0010-4655(95)00042-E

Bussi,G.,Donadio,D.,Parrinello,M.,2007.Canonicalsamplingthroughvelocityrescaling.J.Chem.Phys.126,14101.https://doi.org/10.1063/1.2408420

Carlson,L.A.,Philipson,B.,1979.FISH-EYEDISEASEANEWFAMILIALCONDITIONWITHMASSIVECORNEALOPACITIESANDDYSLIPOPROTEINÆMIA.Lancet314,921–924.https://doi.org/10.1016/S0140-6736(79)92621-7

Chetty,P.S.,Nguyen,D.,Nickel,M.,Lund-Katz,S.,Mayne,L.,Englander,S.W.,Phillips,M.C.,2013.ComparisonofapoA-IhelicalstructureandstabilityindiscoidalandsphericalHDLparticlesbyHXandmassspectrometry.J.LipidRes.54,1589–97.https://doi.org/10.1194/jlr.M034785

DeJong,D.H.,Singh,G.,Bennett,W.F.D.,Arnarez,C.,Wassenaar,T.A.,Schäfer,L.V.,Periole,X.,Tieleman,D.P.,Marrink,S.J.,2013.Improvedparametersforthemartinicoarse-grainedproteinforcefield.J.Chem.TheoryComput.9,687–697.https://doi.org/10.1021/ct300646g

Després,J.P.,Lemieux,I.,Dagenais,G.R.,Cantin,B.,Lamarche,B.,2000.HDL-cholesterolasamarkerofcoronaryheartdiseaserisk:TheQuébeccardiovascularstudy.Atherosclerosis.https://doi.org/10.1016/S0021-9150(00)00603-1

Essmann,U.,Perera,L.,Berkowitz,M.L.,Darden,T.,Lee,H.,Pedersen,L.G.,1995.AsmoothparticlemeshEwaldmethod.J.Chem.Phys.103,8577–8593.https://doi.org/10.1063/1.470117

Evans,D.J.,Holian,B.L.,1985.TheNose–Hooverthermostat.J.Chem.Phys.83,4069–4074.https://doi.org/10.1063/1.449071

Fielding,C.J.,Fielding,P.E.,1995.Molecularphysiologyofreversecholesteroltransport.J.LipidRes.36,211–28.

Francone,O.L.,Fielding,C.J.,1991.Structure-FunctionRelationshipsinHumanLecithin:CholesterolAcyltransferase.Site-DirectedMutagenesisatSerineResidues181and216.Biochemistry30,10074–10077.https://doi.org/10.1021/bi00106a002

Freeman,L.A.,Demosky,S.J.,Konaklieva,M.,Kuskovsky,R.,Aponte,A.,Ossoli,A.F.,Gordon,S.M.,Koby,R.F.,Manthei,K.A.,Shen,M.,Vaisman,B.L.,Shamburek,R.D.,Jadhav,A.,Calabresi,L.,Gucek,M.,Tesmer,J.J.G.,Levine,R.L.,Remaley,A.T.,Remaley,A.T.,2017.Lecithin:CholesterolAcyltransferaseActivationbySulfhydryl-ReactiveSmallMolecules:RoleofCysteine-31.J.Pharmacol.Exp.Ther.362,306–318.https://doi.org/10.1124/jpet.117.240457

Fruchart,J.-C.,Davignon,J.,Hermans,M.P.,Al-Rubeaan,K.,Amarenco,P.,Assmann,G.,Barter,P.,Betteridge,J.,Bruckert,E.,Cuevas,A.,Farnier,M.,Ferrannini,E.,Fioretto,P.,Genest,J.,Ginsberg,H.N.,Gotto,A.M.,Hu,D.,Kadowaki,T.,Kodama,T.,Krempf,M.,Matsuzawa,Y.,Núñez-Cortés,J.,Monfil,C.,Ogawa,H.,Plutzky,J.,Rader,D.J.,Sadikot,S.,Santos,R.D.,Shlyakhto,E.,Sritara,P.,Sy,R.,Tall,A.,Tan,C.,Tokgözoğlu,L.,Toth,P.P.,Valensi,P.,Wanner,



https://doi.org/10.1101/225037


C.,Zambon,A.,Zhu,J.,Zimmet,P.,2014.Residualmacrovascularriskin2013:whathavewelearned?Cardiovasc.Diabetol.13,26.https://doi.org/10.1186/1475-2840-13-26

Fruchart,J.C.,Sacks,F.M.,Hermans,M.P.,Assmann,G.,Brown,W.V.,Ceska,R.,Chapman,M.J.,Dodson,P.M.,Fiorett,P.,Ginsberg,H.N.,Kadowaki,T.,Lablanche,J.M.,Marx,N.,Plutzky,J.,Reiner,Z.,Rosenson,R.S.,Staels,B.,Stock,J.K.,Sy,R.,Wanner,C.,Zambon,A.,Zimmet,P.,2008.TheResidualRiskReductionInitiative:acalltoactiontoresidualvascularriskindyslipidaemicpatients.DiabetesVasc.Dis.Res.5,319–335.https://doi.org/10.3132/dvdr.2008.046

Glukhova,A.,Hinkovska-Galcheva,V.,Kelly,R.,Abe,A.,Shayman,J.A.,Tesmer,J.J.G.,2015.StructureandfunctionoflysosomalphospholipaseA2andlecithin:cholesterolacyltransferase.Nat.Commun.6,6250.https://doi.org/10.1038/ncomms7250

Gordon,T.,Castelli,W.P.,Hjortland,M.C.,Kannel,W.B.,Dawber,T.R.,1977.Highdensitylipoproteinasaprotectivefactoragainstcoronaryheartdisease.TheFraminghamstudy.Am.J.Med.62,707–714.https://doi.org/10.1016/0002-9343(77)90874-9

Group,T.H.C.,2017.EffectsofAnacetrapibinPatientswithAtheroscleroticVascularDisease.N.Engl.J.Med.377,NEJMoa1706444.https://doi.org/10.1056/NEJMoa1706444

Gu,X.,Wu,Z.,Huang,Y.,Wagner,M.A.,Baleanu-Gogonea,C.,Mehl,R.A.,Buffa,J.A.,Didonato,A.J.,Hazen,L.B.,Fox,P.L.,Gogonea,V.,Parks,J.S.,Didonato,J.A.,Hazen,S.L.,2016.Asystematicinvestigationofstructure/functionrequirementsfortheapolipoproteinA-I/Lecithincholesterolacyltransferaseinteractionloopofhigh-densitylipoprotein.J.Biol.Chem.291,6386–6395.https://doi.org/10.1074/jbc.M115.696088

Gunawardane,R.N.,Fordstrom,P.,Piper,D.E.,Masterman,S.,Siu,S.,Liu,D.,Brown,M.,Lu,M.,Tang,J.,Zhang,R.,Cheng,J.,Gates,A.,Meininger,D.,Chan,J.,Carlson,T.,Walker,N.,Schwarz,M.,Delaney,J.,Zhou,M.,2016.Agonistichumanantibodiesbindingtolecithin-cholesterolacyltransferasemodulatehighdensitylipoproteinmetabolism.J.Biol.Chem.291,2799–2811.https://doi.org/10.1074/jbc.M115.672790

Gunawardane,R.N.,Fordstrom,P.,Piper,D.E.,Masterman,S.,Siu,S.,Liu,D.,Brown,M.,Lu,M.,Tang,J.,Zhang,R.,Cheng,J.,Gates,A.,Meininger,D.,Chan,J.,Carlson,T.,Walker,N.,Schwarz,M.,Delaney,J.,Zhou,M.,2016.Agonistichumanantibodiesbindingtolecithin-cholesterolacyltransferasemodulatehighdensitylipoproteinmetabolism.J.Biol.Chem.https://doi.org/10.1074/jbc.M115.672790

Haase,C.L.,Tybjærg-Hansen,A.,AliQayyum,A.,Schou,J.,Nordestgaard,B.G.,Frikke-Schmidt,R.,2012.LCAT,HDLCholesterolandIschemicCardiovascularDisease:AMendelianRandomizationStudyofHDLCholesterolin54,500Individuals.J.Clin.Endocrinol.Metab.97,E248–E256.https://doi.org/10.1210/jc.2011-1846

Hess,B.,Bekker,H.,Berendsen,H.J.C.,Fraaije,J.G.E.M.,1997.LINCS:Alinearconstraintsolverformolecularsimulations.J.Comput.Chem.18,1463–1472.https://doi.org/10.1002/(SICI)1096-987X(199709)18:12<1463::AID-JCC4>3.0.CO;2-H

Hess,B.,Kutzner,C.,VanDerSpoel,D.,Lindahl,E.,2008.GROMACS4:Algorithmsforhighlyefficient,load-balanced,andscalablemolecularsimulation.J.Chem.TheoryComput.4,435–447.https://doi.org/10.1021/ct700301q

Humphrey,W.,Dalke,A.,Schulten,K.,1996.VMD:Visualmoleculardynamics.J.Mol.Graph.14,33–38.https://doi.org/10.1016/0263-7855(96)00018-5

Jämbeck,J.P.M.,Lyubartsev,A.P.,2012.Derivationandsystematicvalidationofarefinedall-atomforcefieldforphosphatidylcholinelipids.J.Phys.Chem.B116,3164–3179.https://doi.org/10.1021/jp212503e

Jin,L.H.,Shieh,J.J.,Grabbe,E.,Adimoolam,S.,Durbin,D.,Jonas,A.,1999.Surfaceplasmon



https://doi.org/10.1101/225037


resonancebiosensorstudiesofhumanwild-typeandmutantlecithincholesterolacyltransferaseinteractionswithlipoproteins.Biochemistry38,15659–15665.https://doi.org/10.1021/bi9916729

Jonas,A.,2000.Lecithincholesterolacyltransferase.Biochim.Biophys.Acta-Mol.CellBiol.Lipids1529,245–256.https://doi.org/10.1016/S1388-1981(00)00153-0

Jonas,A.,1986.Syntheticsubstratesoflecithin:cholesterolacyltransferase.J.LipidRes.27,689–98.

Jonas,A.,Zorich,N.L.,Kézdy,K.E.,Trick,W.E.,1987.ReactionofdiscoidalcomplexesofapolipoproteinA-Iandvariousphosphatidylcholineswithlecithincholesterolacyltransferase.Interfacialeffects.J.Biol.Chem.262,3969–3974.

Kuivenhoven,J.A.,Pritchard,H.,Hill,J.,Frohlich,J.,Assmann,G.,Kastelein,J.,1997.Themolecularpathologyoflecithin:cholesterolacyltransferase(LCAT)deficiencysyndromes.J.LipidRes.38,191–205.

Kuivenhoven,J.A.,VanVoorsttotVoorst,E.J.G.M.,Wiebusch,H.,Marcovina,S.M.,Funke,H.,Assmann,G.,Pritchard,P.H.,Kastelein,J.J.P.,1995.Auniquegeneticandbiochemicalpresentationoffish-eyedisease.J.Clin.Invest.96,2783–2791.https://doi.org/10.1172/JCI118348

Lewis,G.F.,Rader,D.J.,2005.NewinsightsintotheregulationofHDLmetabolismandreversecholesteroltransport.Circ.Res.https://doi.org/10.1161/01.RES.0000170946.56981.5c

Lindorff-Larsen,K.,Piana,S.,Palmo,K.,Maragakis,P.,Klepeis,J.L.,Dror,R.O.,Shaw,D.E.,2010.Improvedside-chaintorsionpotentialsfortheAmberff99SBproteinforcefield.ProteinsStruct.Funct.Bioinforma.78,1950–1958.https://doi.org/10.1002/prot.22711

Manthei,K.A.,Ahn,J.,Glukhova,A.,Yuan,W.,Larkin,C.,Manett,T.D.,Chang,L.,Shayman,J.A.,Axley,M.J.,Schwendeman,A.,Tesmer,J.J.G.,2017.Aretractablelidinlecithin:cholesterolacyltransferaseprovidesastructuralmechanismforactivationbyapolipoproteinA-I.J.Biol.Chem.jbc.M117.802736.https://doi.org/10.1074/jbc.M117.802736

Marrink,S.J.,Risselada,H.J.,Yefimov,S.,Tieleman,D.P.,DeVries,A.H.,2007.TheMARTINIforcefield:Coarsegrainedmodelforbiomolecularsimulations.J.Phys.Chem.B111,7812–7824.https://doi.org/10.1021/jp071097f

Mead,J.R.,Irvine,S.A.,Ramji,D.P.,2002.Lipoproteinlipase:Structure,function,regulation,androleindisease.J.Mol.Med.https://doi.org/10.1007/s00109-002-0384-9

Monticelli,L.,Kandasamy,S.K.,Periole,X.,Larson,R.G.,Tieleman,D.P.,Marrink,S.J.,2008.TheMARTINIcoarse-grainedforcefield:Extensiontoproteins.J.Chem.TheoryComput.4,819–834.https://doi.org/10.1021/ct700324x

Murray,K.R.,Nair,M.P.,Ayyobi,A.F.,Hill,J.S.,Pritchard,P.H.,Lacko,A.G.,2001.Probingthe121-136domainoflecithin:cholesterolacyltransferaseusingantibodies.Arch.Biochem.Biophys.385,267–275.https://doi.org/10.1006/abbi.2000.2154

Navab,M.,Reddy,S.T.,Meriwether,D.,Fogelman,S.I.,Fogelman,A.M.,2015.ApoA-IMimeticPeptides:AReviewofthePresentStatus,in:ApolipoproteinMimeticsintheManagementofHumanDisease.SpringerInternationalPublishing,Cham,pp.15–27.https://doi.org/10.1007/978-3-319-17350-4_2

Navab,M.,Shechter,I.,Anantharamaiah,G.M.,Reddy,S.T.,VanLenten,B.J.,Fogelman,A.M.,2010.StructureandfunctionofHDLmimetics.Arterioscler.Thromb.Vasc.Biol.https://doi.org/10.1161/ATVBAHA.109.187518

Parks,J.S.,Gebre,A.K.,1997.Long-chainpolyunsaturatedfattyacidsinthesn-2positionofphosphatidylcholinedecreasethestabilityofrecombinanthighdensitylipoproteinapolipoproteinA-Iandtheactivationenergyofthelecithin:cholesterolacyltransferase



https://doi.org/10.1101/225037


reaction.J.LipidRes.38,266–275.Parrinello,M.,Rahman,A.,1981.Polymorphictransitionsinsinglecrystals:Anewmolecular

dynamicsmethod.J.Appl.Phys.52,7182–7190.https://doi.org/10.1063/1.328693Peelman,F.,Vanloo,B.,Perez-Mendez,O.,Decout,A.,Verschelde,J.L.,Labeur,C.,Vinaimont,N.,

Verhee,A.,Duverger,N.,Brasseur,R.,Vandekerckhove,J.,Tavernier,J.,Rosseneu,M.,1999.Characterizationoffunctionalresiduesintheinterfacialrecognitiondomainoflecithincholesterolacyltransferase(LCAT).ProteinEng.12,71–78.https://doi.org/10.1093/protein/12.1.71

Peelman,F.,Verschelde,J.L.,Vanloo,B.,Ampe,C.,Labeur,C.,Tavernier,J.,Vandekerckhove,J.,Rosseneu,M.,1999.Effectsofnaturalmutationsinlecithin:cholesterolacyltransferaseontheenzymestructureandactivity.J.LipidRes.40,59–69.

Peloso,G.M.,Auer,P.L.,Bis,J.C.,Voorman,A.,Morrison,A.C.,Stitziel,N.O.,Brody,J.A.,Khetarpal,S.A.,Crosby,J.R.,Fornage,M.,Isaacs,A.,Jakobsdottir,J.,Feitosa,M.F.,Davies,G.,Huffman,J.E.,Manichaikul,A.,Davis,B.,Lohman,K.,Joon,A.Y.,Smith,A.V.,Grove,M.L.,Zanoni,P.,Redon,V.,Demissie,S.,Lawson,K.,Peters,U.,Carlson,C.,Jackson,R.D.,Ryckman,K.K.,MacKey,R.H.,Robinson,J.G.,Siscovick,D.S.,Schreiner,P.J.,Mychaleckyj,J.C.,Pankow,J.S.,Hofman,A.,Uitterlinden,A.G.,Harris,T.B.,Taylor,K.D.,Stafford,J.M.,Reynolds,L.M.,Marioni,R.E.,Dehghan,A.,Franco,O.H.,Patel,A.P.,Lu,Y.,Hindy,G.,Gottesman,O.,Bottinger,E.P.,Melander,O.,Orho-Melander,M.,Loos,R.J.F.,Duga,S.,Merlini,P.A.,Farrall,M.,Goel,A.,Asselta,R.,Girelli,D.,Martinelli,N.,Shah,S.H.,Kraus,W.E.,Li,M.,Rader,D.J.,Reilly,M.P.,McPherson,R.,Watkins,H.,Ardissino,D.,Zhang,Q.,Wang,J.,Tsai,M.Y.,Taylor,H.A.,Correa,A.,Griswold,M.E.,Lange,L.A.,Starr,J.M.,Rudan,I.,Eiriksdottir,G.,Launer,L.J.,Ordovas,J.M.,Levy,D.,Chen,Y.D.I.,Reiner,A.P.,Hayward,C.,Polasek,O.,Deary,I.J.,Borecki,I.B.,Liu,Y.,Gudnason,V.,Wilson,J.G.,VanDuijn,C.M.,Kooperberg,C.,Rich,S.S.,Psaty,B.M.,Rotter,J.I.,O’Donnell,C.J.,Rice,K.,Boerwinkle,E.,Kathiresan,S.,Cupples,L.A.,2014.Associationoflow-frequencyandrarecoding-sequencevariantswithbloodlipidsandcoronaryheartdiseasein56,000whitesandblacks.Am.J.Hum.Genet.94,223–232.https://doi.org/10.1016/j.ajhg.2014.01.009

Periole,X.,Cavalli,M.,Marrink,S.J.,Ceruso,M.A.,2009.Combininganelasticnetworkwithacoarse-grainedmolecularforcefield:Structure,dynamics,andintermolecularrecognition.J.Chem.TheoryComput.5,2531–2543.https://doi.org/10.1021/ct9002114

Perk,J.,DeBacker,G.,Gohlke,H.,Graham,I.,Reiner,Ž.,Verschuren,M.,Albus,C.,Benlian,P.,Boysen,G.,Cifkova,R.,Deaton,C.,Ebrahim,S.,Fisher,M.,Germano,G.,Hobbs,R.,Hoes,A.,Karadeniz,S.,Mezzani,A.,Prescott,E.,Ryden,L.,Scherer,M.,Syvänne,M.,ScholteOpReimer,W.J.M.,Vrints,C.,Wood,D.,Zamorano,J.L.,Zannad,F.,Cooney,M.T.,Bax,J.,Baumgartner,H.,Ceconi,C.,Dean,V.,Fagard,R.,Funck-Brentano,C.,Hasdai,D.,Kirchhof,P.,Knuuti,J.,Kolh,P.,McDonagh,T.,Moulin,C.,Popescu,B.A.,Sechtem,U.,Sirnes,P.A.,Tendera,M.,Torbicki,A.,Vahanian,A.,Windecker,S.,Aboyans,V.,Ezquerra,E.A.,Baigent,C.,Brotons,C.,Burell,G.,Ceriello,A.,DeSutter,J.,Deckers,J.,DelPrato,S.,Diener,H.C.,Fitzsimons,D.,Fras,Z.,Hambrecht,R.,Jankowski,P.,Keil,U.,Kirby,M.,Larsen,M.L.,Mancia,G.,Manolis,A.J.,McMurray,J.,Pajak,A.,Parkhomenko,A.,Rallidis,L.,Rigo,F.,Rocha,E.,Ruilope,L.M.,VanDerVelde,E.,Vanuzzo,D.,Viigimaa,M.,Volpe,M.,Wiklund,O.,Wolpert,C.,2012.EuropeanGuidelinesoncardiovasculardiseasepreventioninclinicalpractice(version2012).Eur.HeartJ.https://doi.org/10.1093/eurheartj/ehs092

Piper,D.E.,Romanow,W.G.,Gunawardane,R.N.,Fordstrom,P.,Masterman,S.,Pan,O.,Thibault,S.T.,Zhang,R.,Meininger,D.,Schwarz,M.,Wang,Z.,King,C.,Zhou,M.,Walker,N.P.C.,2015.Thehigh-resolutioncrystalstructureofhumanLCAT.J.LipidRes.56,1711–1719.



https://doi.org/10.1101/225037


https://doi.org/10.1194/jlr.M059873Piper,D.E.,Romanow,W.G.,Gunawardane,R.N.,Fordstrom,P.,Masterman,S.,Pan,O.,Thibault,

S.T.,Zhang,R.,Meininger,D.,Schwarz,M.,Wang,Z.,King,C.,Zhou,M.,Walker,N.P.C.,2015.Thehigh-resolutioncrystalstructureofhumanLCAT.J.LipidRes.56,1711–1719.https://doi.org/10.1194/jlr.M059873

Rader,D.J.,Hovingh,G.K.,2014.HDLandcardiovasculardisease.Lancet384,618–625.https://doi.org/10.1016/S0140-6736(14)61217-4

Reiner,Ž.,2013.ManagingtheresidualcardiovasculardiseaseriskassociatedwithHDL-cholesterolandtriglyceridesinstatin-treatedpatients:Aclinicalupdate.Nutr.Metab.Cardiovasc.Dis.https://doi.org/10.1016/j.numecd.2013.05.002

SevuganChetty,P.,Mayne,L.,Kan,Z.-Y.,Lund-Katz,S.,Englander,S.W.,Phillips,M.C.,2012.ApolipoproteinA-Ihelicalstructureandstabilityindiscoidalhigh-densitylipoprotein(HDL)particlesbyhydrogenexchangeandmassspectrometry.Proc.Natl.Acad.Sci.109,11687–11692.https://doi.org/10.1073/pnas.1209305109

Sorci-Thomas,M.,Kearns,M.W.,Lee,J.P.,1993.ApolipoproteinA-Idomainsinvolvedinlecithin-cholesterolacyltransferaseactivation.Structure:functionrelationships.J.Biol.Chem.268,21403–21409.

Sorci-Thomas,M.G.,Bhat,S.,Thomas,M.J.,2009.Activationoflecithin:cholesterolacyltransferasebyHDLApoA-Icentralhelices.Clin.Lipidol.4,113–124.https://doi.org/10.2217/17584299.4.1.113

Sorci-Thomas,M.G.,Thomas,M.J.,2002.TheeffectsofalteredapolipoproteinA-IstructureonplasmaHDLconcentration.TrendsCardiovasc.Med.https://doi.org/10.1016/S1050-1738(01)00163-3

Tall,A.R.,1998.Anoverviewofreversecholesteroltransport.Eur.HeartJ.19SupplA,A31–A35.Tall,A.R.,Rader,D.J.,2017.TheTrialsandTribulationsofCETPInhibitors.Circ.Res.

CIRCRESAHA.117.311978.https://doi.org/10.1161/CIRCRESAHA.117.311978VanTilbeurgh,H.,Egloff,M.P.,Martinez,C.,Rugani,N.,Verger,R.,Cambillau,C.,1993.Interfacial

activationofthelipase-procolipasecomplexbymixedmicellesrevealedbyX-raycrystallography.Nature362,814–820.https://doi.org/10.1038/362814a0

vanTilbeurgh,H.,Sarda,L.,Verger,R.,Cambillau,C.,1992.Structureofthepancreaticlipase-procolipasecomplex.Nature359,159–62.https://doi.org/10.1038/359159a0

Webb,B.,Sali,A.,2014.ProteinstructuremodelingwithMODELLER.MethodsMol.Biol.1137,1–15.https://doi.org/10.1007/978-1-4939-0366-5_1

Winkler,F.K.,D’Arcy,A.,Hunziker,W.,199AD.Structureofhumanpancreaticlipase.Nature343,771–774.https://doi.org/10.1038/343771a0

Wu,Z.,Wagner,M.A.,Zheng,L.,Parks,J.S.,Shy,J.M.,Smith,J.D.,Gogonea,V.,Hazen,S.L.,2007.TherefinedstructureofnascentHDLrevealsakeyfunctionaldomainforparticlematurationanddysfunction.Nat.Struct.Mol.Biol.14,861–868.https://doi.org/10.1038/nsmb1284

Yang,Y.,Lowe,M.E.,2000.Theopenlidmediatespancreaticlipasefunction.J.LipidRes.41,48–57.

Yesylevskyy,S.O.,Schäfer,L.V.,Sengupta,D.,Marrink,S.J.,2010.Polarizablewatermodelforthecoarse-grainedMARTINIforcefield.PLoSComput.Biol.6,1–17.https://doi.org/10.1371/journal.pcbi.1000810



https://doi.org/10.1101/225037


interaction of lecithin-cholesterol acyltransferase with lipid … · 2017. 12. 15. · interaction...

Documents