Features of wild-type human SOD1 limit interactions with misfolded aggregates of mouse G86R Sod1

MISSING IMAGE

Material Information

Title:
Features of wild-type human SOD1 limit interactions with misfolded aggregates of mouse G86R Sod1
Series Title:
Molecular Neurodegeneration
Physical Description:
Mixed Material
Creator:
David A Qualls
Mercedes Prudencio
Brittany LT Roberts
Keith Crosby
Hilda Brown
David R Borchelt
Publisher:
Molecular Neurodegeneration
Publication Date:

Notes

Abstract:
Mutations in the gene encoding superoxide dismutase 1 (SOD1) account for about 20% of the cases of familial amyotrophic lateral sclerosis (fALS). It is well established that mutations in SOD1, associated with fALS, heighten the propensity of the protein to misfold and aggregate. Although aggregation appears to be a factor in the toxicity of mutant SOD1s, the precise nature of this toxicity has not been elucidated. A number of other studies have now firmly established that raising the levels of wild-type (WT) human SOD1 (hSOD1) proteins can in some manner augment the toxicity of mutant hSOD1 proteins. However, a recent study demonstrated that raising the levels of WT-hSOD1 did not affect disease in mice that harbor a mouse Sod1 gene (mSod1) encoding a well characterized fALS mutation (G86R). In the present study, we sought a potential explanation for the differing effects with WT-hSOD1 on the toxicity of mutant hSOD1 versus mutant mSod1. In the cell culture models used here, we observe poor interactions between WT-hSOD1 and misfolded G86R-mSod1, possibly explaining why over-expression of WT-hSOD1 does not synergize with mutant mSod1 to accelerate the course of the disease in mice.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution.
System ID:
AA00020069:00001

Full Text

PAGE 1

Features of wild-type human SOD1 limit interacti ons with misfolded aggregates of mouse G86R SOD1 David Qualls*, Mercedes Prudencio*, Brittany L.T. R oberts, Keith Crosby, Hilda Brown, David R. Borchelt Supplementary figures Supplementary Fig. S1. WT-mSod1:RFP forms saponin-resistant inclusions. The indicated fusion construct was expressed in CHO cells as explained in t he legends of the primary figures. The images shown are representative of 3 independent experiments.

PAGE 2

Supplementary Fig. S2. Additional comparative data for WT-mSod1:RFP co-expressed with WTmSod1:YFP, WT-hSOD1:YFP, or G85R-hSOD1:YFP. The indicated fusion constructs were expressed in CHO cells as explained in the legends of the primary figures. The images shown are representative of 3 independent experiments. The diagrams illustrate the primar y observations. WT-mSod1:RFP inclusions do not interact with WT-mSod1:YFP or WT-hSOD1:YFP. G85R-hSOD1:YFP appears to deposit on the surface of the inclusions formed by WT-mSod1:RFP.

PAGE 3

Supplementary Fig. S3. Immunoblot to demonstrate similar levels of expression of SOD1:GLuc fusion proteins. Ten micrograms of protein from freeze-thaw lysate s of CHO cells transfected with vectors that express the fusion proteins were analyzed by imm unoblot with a rabbit polyclonal antibody raised against the whole human SOD1 protein. These lysates were used in one of the replicates of luciferase assays that generated the data in Figure 6. The identity of each protein in the doublet is provided above the figure; the orientation of the labels matches the orientation of the two bands. UT denotes untransfected cells. The products of the vectors expressing the GLuc fragments cannot be seen with the SOD1 antibody. Supplementary Fig. S4. WT-mSod1:YFP does not bind to G85R-hSOD1:RFP inclusions. The indicated fusion constructs were expressed in CHO cells as explained in the legends of the primary figures. The images shown are representative of 3 independent experiments.

PAGE 4

Supplementary Fig. S5. WT-hSOD1mon:RFP forms intermingled inclusions with G86R-mSod1:YFP. The indicated fusion constructs were expressed in CHO c ells as explained in the legends of the primary figures. The images shown are representat ive of 3 independent experiments. Supplementary Fig. S6. WT-mSod1mon:YFP forms intermingled inclusions with G85R-hSOD1:RFP. The indicated fusion constructs were expressed in CHO c ells as explained in the legends of the primary figures. The images shown are representat ive of 3 independent experiments.



PAGE 1

RESEARCHARTICLEOpenAccessFeaturesofwild-typehumanSOD1limit interactionswithmisfoldedaggregatesofmouse G86RSod1DavidAQualls1 †,MercedesPrudencio1,2 †,BrittanyLTRoberts1,KeithCrosby1,HildaBrown1andDavidRBorchelt1*AbstractMutationsinthegeneencodingsuperoxidedismutase1(SOD1)accountforabout20%ofthecasesoffamilial amyotrophiclateralsclerosis(fALS).ItiswellestablishedthatmutationsinSOD1,associatedwithfALS,heightenthe propensityoftheproteintomisfoldandaggregate.Althoughaggregationappearstobeafactorinthetoxicityof mutantSOD1s,theprecisenatureofthistoxicityhasnotbeenelucidated.Anumberofotherstudieshavenow firmlyestablishedthatraisingthelevelsofwild-type(WT)humanSOD1(hSOD1)proteinscaninsomemanner augmentthetoxicityofmutanthSOD1proteins.However,arecentstudydemonstratedthatraisingthelevelsof WT-hSOD1didnotaffectdiseaseinmicethatharboramouse Sod1 gene( mSod1 )encodingawellcharacterized fALSmutation(G86R).Inthepresentstudy,wesoughtapotentialexplanationforthedifferingeffectswith WT-hSOD1onthetoxicityofmutanthSOD1versusmutant mSod1 .Inthecellculturemodelsusedhere, weobservepoorinteractionsbetweenWT-hSOD1andmisfoldedG86R-mSod1,possiblyexplainingwhy over-expressionofWT-hSOD1doesnotsynergizewithmutantmSod1toacceleratethecourseofthediseaseinmice.IntroductionMutationsinthegeneencodingsuperoxidedismutase1 (SOD1)accountforabout20%ofthecasesoffamilial amyotrophiclateralsclerosis(fALS){http://alsod.iop.kcl. ac.uk/default.aspx}.TheSOD1proteinisarelativelysmall antioxidantenzymecomprisedof153aminoacids.Inits activestate,theproteindimerizestoformthematureenzymewitheachsubunitbinding1atomofZnand1atom ofCu[1-3].Todatemorethan165mutationsinmore thanhalfoftheaminoacidresiduesintheenzymehave beenidentifiedinpatientsdiagnosedwithfALS{http:// alsod.iop.kcl.ac.uk/default.aspx}.Theimpactofthesemutationsonitsenzymaticactivityvariesgreatly,andithas notyetbeenpossibletodefineasinglemechanismby whichthesemutationscausedisease[forreviewsee[4]. OnecommonfeatureofmutantSOD1proteinsisthat theyexhibitahightendencytoaggregateaberrantlyinto highmolecularweightstructures,whichcanbeisolated biochemicallyduetotheirinsolubilityinnon-ionicdetergents[5]. AlthoughaggregationofmutantSOD1isbysomemannerlinkedtothetoxicpropertyoftheproteinthatisresponsibleforinducingmotorneurondisease,thenatureof thetoxicpropertyremainsincompletelydefinedandmay bemultifactorial[4].Oneofthemoreinterestingavenues towardsimprovingourunderstandingofthebasisofmutantSOD1toxicityhasbeenrevealedinstudiesinwhich miceexpressingmutantSOD1havebeenmatedtomice expressingwild-typehSOD1.Anumberofstudieshave nowfirmlyestablishedthatraisingthelevelsofWThSOD1proteinsbycrossingastrainofmiceproducedby Gurneyetal.(GurWT[6])tomutanthSOD1micecan,in somemanner,acceleratetheonsetofparalysis[7-11].Importantly,theGurWTmiceexpressWT-hSOD1atvery highlevelsandlinesofWT-hSOD1micethatexpressat lowerlevelsdonotasuniformlyacceleratedisease[11,12]. Thebasisforthecurrentworkisarecentstudyby Audetandcolleaguesthataskedwhetheroverexpression ofWT-hSOD1,usingtheGurWTstrain,couldaugment thetoxicityofG86R-mSod1,findingnoeffect[13].The *Correspondence: drb1@ufl.edu†Equalcontributors1FromtheDepartmentofNeuroscience,CenterforTranslationalResearchin NeurodegenerativeDisease,SantaFeHealthCareAlzheimer ’ sDisease ResearchCenter,McKnightBrainInstitute,UniversityofFlorida,Box100159, 1275CenterDrive,RoomJ491,Gainesville,FL32610,USA Fulllistofauthorinformationisavailableattheendofthearticle 2013Quallsetal.;licenseeBioMedCentralLtd.ThisisanopenaccessarticledistributedunderthetermsoftheCreative CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginalworkisproperlycited.Qualls etal.MolecularNeurodegeneration 2013, 8 :46 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 2

G86Rmouse[14],wascreatedbyasimilarapproachto whathasbeenusedformicethatexpressmutanthuman proteins;agenomicDNAfragmentofthemousegenewas mutatedtoencodeanfALSmutation.TheG86Rmutation inmouseSod1isequivalenttothehumanG85RhSOD1 mutation.Thedifferenceinnumberingisaconsequenceof thehistoricalnumberingofhumanSOD1residuesbased onthematurehumanproteinsequence(initiatormethionineremovedpost-translationally)whereasthemousesequencewasnumberedbyactualcodonposition.Thelack ofaneffectofco-expressedWT-hSOD1ondiseaseinthe G86Rmiceisnotapeculiarityofthemutationbecause micegeneratedbycrossingGurWTanimalstomiceexpressingG85R-hSOD1developmotorneurondiseaseconsiderablyearlierthanmiceexpressingthemutantprotein alone[10].Thedifferentoutcomesinthesetwoexperimentalsettingsimpliesthatimprovingourunderstanding ofthewayWT-hSOD1interactswithG85RhSOD1and G86R-mSod1couldprovideinsightintohowthepresence ofWT-hSOD1acceleratestheonsetofparalysis. TostudyinteractionsbetweenhSOD1andmSod1that mayoccurwhenfALSmutationsinducemisfolding, wehaveemployedanestablishedcellmodelofmutant SOD1aggregationinconjunctionwithco-transfectionapproaches.Usingadetergentextractionandsedimentation assay,weexaminewhetherWT-hSOD1interactswith misfoldedG86R-mSod1.Additionally,tovisualizesuchinteractions,andfollowingastrategypreviouslydescribed [11,15],wehavefusedtheSOD1variantstofluorescent tags;G86R-mSod1fusedtoredfluorescentprotein(RFP) andWT-hSOD1orWT-mSod1fusedyellowfluorescent protein(YFP).Usingbothbiochemicalandvisualmethods wecomparethedegreetowhichWT-mSod1andWThSOD1interactwithmisfoldedaggregatesofG86R-mSod1.MethodsGenerationofhSOD1/mSod1DNAexpressionplasmids andsplitluciferaseconstructsExpressionplasmids(pEF.Bos[16])thatencodewildtype(WT)hSOD1,WT-mSod1,A4V-hSOD1,andG85RhSOD1havebeenpreviouslydescribed[17-19].The G86RmutationinmSod1wasintroducedintopEF. Bos-WT-mSod1usingoligonucleotidesthatencodethe desiredmutationandtheQuickChangemutagenesis kit(Stratagene/LifeTechnologies/Thermo,GrandIsland, NY).WT-hSOD1monandWT-mSod1monwerecreated byintroducingmutationsatcodons50and51tochange aminoacidsFandGatthesepositionstoE.Mutations wereintroducedinpEF.BosvectorsforWT-hSOD1and WT-mSod1usingoligonucleotidesencodingthedesired mutationsandtheQuickChangeKit.pEF.Bosvectors encodingWT-hSOD1andG85R-hSOD1fusedtoturboRFP(Evrogen,Moscow,Russia)andeYFP(Invitrogen/ LifeTechnologies/ThermoGrandIsland,NY,)havealso beenpreviouslydescribed[11,15].TocreatepEF.BosexpressionplasmidsthatencodefusionsofG85R-hSOD1 andG86R-mSod1toYFPandRFP,andWT-hSOD1mon andWT-mSod1monfusedtoYFPandRFP,westarted withpEF.BosvectorsencodingWT-hSOD1:YFPorWThSOD1:RFPfusionsandremovedthesegmentofthevectorencodingSOD1bydigestionwithunique(engineered) Nco1andSal1restrictionendonucleasesites.Wethen insertedcDNAencoding,WT-mSod1,G86R-mSod1,WThSOD1monandWT-mSod1monengineeredwithcompatibleendstoproducethenewvectorsencoding in-framefusionproteins.Allvectorswereextensivelysequencedtoverifythepresenceofthedesiredmutations andtheabsenceofanyunwantedmutations. Thesplit,humanized,Gaussialuciferase(hGluc)constructs(NewEnglandBiolabs,Ipswich,MA,USA)were generatedbycleavingpEF.BosvectorsencodingWThSOD1orWT-hSOD1monwithSal1andinserting segmentsofcDNAencodinghGlucin-frameattheCterminusoftheSOD1cDNAs.EachhGlucsegment wasamplifiedusingstandardPCRprocedureswiththe primershGluc-S(cgcagctgcagtcgaccgatggtggcggtggctct) andhGluc1-BOS-AS(tcccaggtgggtaccttagcctatgccgccctgt) fortheN-terminalsegment(L1)orhGluc2-BOS-AS (tcccaggtgggtaccttagtcaccac cggcccc)fortheC-terminal segment(L2)ofluciferase.PCRproductswereanalyzed ona1%agarosegeltoconfirmthetargetfragmentwas amplified.InmostcasesasinglePCRproductwas present.IfthereweremultiplePCRproducts,thedesiredfragmentwascutfromapreparativeagarosegel, extracted,andpurifiedusingtheQIAquickGelExtraction kit(Qiagen,Germantown,MD,USA;Cat.No28706). PCRproductswereinsertedintothecleavedpEF.BosSOD1vectorsusingtheIn-FusionCloningkit(Clontech, MountainView,CA;Cat.No639601).Allplasmidswere thentransformedintoStellarco mpetentcells(Clontech)followingstandardtransformation strategies.TheFastPlasmid MiniKit(5Prime,Gaithersburg,MD,USA;Ref#2300010) wasusedtoisolatenewrecombinants,whichwereverified byDNAsequenceanalysis.UponverificationofthecorrectplasmidDNAs,largescalepreparationsofthedifferentplasmidsfortransfectionwerepreparedbystandard cesiumchloridegradientpurification.TransienttransfectionsandmicroscopyanalysesExpressionofhSOD1/mSod1constructsforanalysesof detergent-insolubilitywereperformedinHEK293FTcells asexplainedinfigurelegendsandfollowingprocedures describedinpreviousstudies[5,20].Briefly,theexpression plasmidsweretransientlytransfectedintoHEK293FT cells.Followingthetimesindicated(24or48hrs),thecells wereharvestedandsubjectedtodetergentextractionand sedimentationtoseparatesolubleandinsolubleformsof SOD1.DetergentfractionswereanalyzedbySDS-PAGEQualls etal.MolecularNeurodegeneration 2013, 8 :46 Page2of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 3

andimmunoblotwithSOD1antibodies.Immunoreactive proteinswerevisualizedandquantifiedonaFujiImaging system.Eachexperimentwasrepeatedandquantifiedat least3times. ForstudiestovisualizeinclusionsformedbyYFPand RFPfusionconstructs,wetransfectedChineseHamster Ovary(CHO)cellsbecausethesecellsnormallyshowa veryflatmorphologywithadistinctnucleusandcytoplasm;allowingforagoodvisualizationofintracellular inclusions.Thesecellsalsoshowgoodadherencetocultureplatesandresistliftingaftersaponintreatment. Cellsweresplitinto12-wellplatescontainingPoly-LLysinecoatedcoverslips,andincubatedat37Cwith5% CO2for24hours.Cellsweretransientlytransfectedwith thevectorsofinterestusingLipofectamine-2000(single transfections:500ngtotalDNAused;Co-transfections: 500ngofeachconstructused).Twenty-fourhoursafter transfection,onesetofcellsweretreatedwith0.1%saponinin1xPBSfor30minutes.Thecellswerethen rinsedwith1xPBS,andfixedin4%paraformaldehyde (1xPBS).A1:2000solutionofDAPIin1xPBSwasused tostainnuclei.Coverslipswerethenmountedonslides foranalysisviafluorescencemicroscopy.Allexperiments wereperformedthreetimes,andeachsamplewasanalyzedforthepresenceandcompositionofinclusion-like structures.Representativeexamplesofcellsfromeach samplewerephotographed.Thecameraexposuresused tocaptureRFPandYFPimagesinco-transfectionswere recordedandcomparedtosingletransfectionstoensure thatthefluorescencewastheresultoftheYFP,andnot bleed-throughfromco-expressedRFP. Foreachconstructanalyzed,multipletransfections wereperformed(atleast3)andmultipleimagesforeach transfectionexperimentwerecaptured.Between250 and1,000cellsexpressingthefluorescentfusionprotein wereassessedforeachexperiment,dependinguponthe transfectionefficiencyforaparticularexperiment.The imagesshownintherelevantfiguresarerepresentative ofatleast3individualexperiments.LuciferaseassayCHOcellswereculturedinto6wellplatesandtransientlytransfected,at90-95%confluency,withequimolar amountsoftwosplitluciferaseconstructs(total2 g DNA,1 gperconstruct).Co-transfectionswereperformedusingLipofectamine2000,followingthemanufacturer ’ sprotocol(Invitrogen,Carlsbad,CA,USA).The cellswereincubatedat37CinaCO2incubatorfor 24hours,rinsedwith1xPBS,harvested,andthecells werepelletedbycentrifugingat5000xgfor2min.A coelenterazineassaywasusedinordertodemonstrate theluciferaseactivityofthesplitluciferaseconstructs. Thecellpelletswereresuspendedinapproximately5x estimatedpelletvolumein1xPBScontainingprotease inhibitorcocktail(Sigma,St.Louis,MO;cat#P8340) andlysedbythreefreeze-thawcyclesinadry-ice/100% alcoholsolutionanda42Cwaterbath.Thesamples werecentrifugedandthesupernatantwastransferredto newEppendorftubes.Thesupernatantwasdiluted1:20 and2 lofeachdilutedsamplewastransferredtoa 96-wellplate.Thecoelenterazineassaywasperformed followingthemanufacturer ’ sprotocol(NanolightTechnology,Pinetop,AZ,USA;Cat.No303 – 500)usingamicroplatereader(SynergyHT,Biotechinstruments).This assaywasrepeatedtwomoretimeswitheachrepeated transfection.StatisticalanalysesDataonrelativeaggregationpropensitywereanalyzed onGraphPadPRISM5.01Software(LaJolla,CA)todeterminestatisticaldifferences,usingunpairedstudent t tests,andgenerategraphicrepresentations.ResultsWT-hSOD1doesnotreadilyco-aggregatewithG86RmSod1InordertodeterminewhetherWT-hSOD1co-aggregates withorotherwisemodulatesaggregationofG86R-mSod1, weusedacellculturemodelofmutantSOD1aggregation thatwehaveusedextensivelyinthepast.Humanembryonickidneycells(HEK293FT)weretransientlytransfected withtheindicatedmouseandhumanSOD1constructs. Forty-eighthoursfollowingtransfection,cellswereharvestedanddetergent-soluble(S1)anddetergent-insoluble (P2)fractionswereobtainedaspreviouslydescribed[19]. NotethattheaggregationpropensityofSOD1proteinsis measuredbytheratioofdetergent-insolubletodetergentsolubleSOD1.AcommoncharacteristicforbothWThSOD1andWT-mSod1proteinsistheirlowinherentaggregationpropensitywhenexpressedinculturedcells [17,19].Asexpected,expressionofWT-hSOD1inourcell culturesysteminducedlittleaggregationofthisproteinas assayedbytheformationofdetergentinsolublecomplexes (hWT,Figure1AandB).Ineachexperiment,expression oftheA4VvariantofhSOD1providedapositivecontrol,whichrobustlyaggregated(A4V,Figure1AandB). ComparedtoWT-hSOD1,G86R-mSod1showedamuch highertendencytoaggregate(Figure1A,P2upperpanel; andB).At48hourspost-transfection,cellsco-expressing WT-hSOD1withG86R-mSod1showednostatisticallysignificantdecreaseinthelevelofdetergentinsolublemutant mSod1nordidweobserveastatisticallysignificantincreaseinthelevelofinsolubleWT-hSOD1incellsexpressingbothproteins(Figure1AandB).Thesedata suggestthatWT-hSOD1doesnotreadilyinteractwith misfoldedG86R-mSod1thatisorganizedintodetergentinsolublecomplexes.Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page3of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 4

VisualizationofhSOD1/mSod1inclusionformation PreviouscharacterizationofWT-hSOD1fusedtogreen fluorescentprotein(GFP)havedemonstratednormal dimerizationandfullactivityoftheSOD1protein,and thusfusionofSOD1withsuchfluorescentreportersdoes notnecessarilyalterthefoldingofSOD1[21].Wehave previouslyusedastrategyinwhichWTandmutantSOD1 wasfusedtoeitherredfluorescentprotein(RFP)oryellow fluorescentprotein(YFP)asameanstoobserveinteractionsbetweentheWTandmutantproteinswithinaggregates[11,15].Wehavealsodevelopedastrategyinwhich cellsaretreatedwithsaponinpriortofixationtodistinguishsoluble,ordiffusible,moleculesfrominsoluble,or immobile,molecules.Saponinisanamphipathicglycoside thatcreatesholesintheplasmamembranewithoutlysing thecell[forreviewsee[22].FusionsofWT-hSOD1to YFParediffuselydistributedthroughoutthecytosoland readilydiffuseoutofcellspermeabilizedwithsaponin; propertiesconsistentwithfullysolubleproteins[15,23]. UnliketheGFPandYFPtags,theRFPtaghasamajorimpactonthebehaviorofWT-hSOD1,causingtheprotein toformasinglelargeroundinclusion-likestructureinthe cytosolthatdoesnotdiffuseoutofpermeabilizedcells [15].Similarly,weobservedthatWT-mSod1fusedtoRFP alsoproducedinclusionsthatweresaponinresistant whetherexpressedaloneorincombinationwithother YFPtaggedproteins(Additionalfile1:FiguresS1andS2). WehavemorerecentlydeterminedthatWT-hSOD1:RFP formsinclusionsbecauseRFPhasahighpropensityto dimerizeandwhenpairedwithWT-hSOD1thefusion moleculebecomesbivalentwithanabilitytoformnetworksofinteractionsbetweenmolecules[23].However, whenRFPisfusedtomutanthSOD1,thentheinherent propensityofthemutanthSOD1toaggregatebecomes theprimaryforceinbehaviorandinclusionsformedby mutantSOD1fusedtoRFPproducemultiplesmallinclusionsthatringthenucleus;theseinclusionsareidentical inmorphologytoinclusionsformedbymutantSOD1 fusedtoYFP[23].Thus,inthepresentstudy,weuseda paradigminwhichG86R-mSod1wasfusedtoRFPand co-expressedwithWT-hSOD1orWT-mSod1fusedto YFP.SimilartoWT-hSOD1,WT-mSod1fusedtoYFP showedadiffusedistributionandallofthefluorescence wasreleasedbysaponintreatment(Figure2).Similarto G85R-hSOD1[23],G86R-mSod1fusedtoeitherRFPor YFPformedmultiplesmallinclusionsthatwerelocated nearthenucleusandresistanttoreleasebysaponin (Figure3).Animportantelementinthesestudiestonote isthattheRFPproteinismuchbrighterthantheYFPproteinandthustheexposuretimeswereadjustedtocapture theimagesatequivalentintensities.Typically,imagesof RFPfluorescencewerecapturedwithexposuresof1/200 to1/300secondswhereasexposuresofYFPfluorescence weretypically1/20to1/30seconds;extendedtoupto1/2 to1/3secondsinsomecases.Weobservedthatexposure timesofupto1/2to1/3secondsintheYFPchannelwere possible,withminimalbled-throughofRFPintotheYFP channel(seeFigure3Acolumn3,row2).Collectively, thesefindingsindicatethatWTandmutantmSod1behavesimilarlytoWTandmutanthSOD1intermsofpropensitytoaggregate. Inco-transfectionexperiments,weobservedthatinclusionsformedbyG86R-mSod1:RFPcontainedlittleifany WT-hSOD1:YFPwithintheaggregatesthatremainedafter Figure1 WT-hSOD1doesnotco-sedimentwithaggregatesformedbyG86R-mSOD1.A) Immunoblotofdetergent-insoluble(P2)and soluble(S1)fractionsofHEK293FTcellstransfectedfor48hourswithvectorsforWT-hSOD1(hWT),A4V-hSOD1(A4V),G86R-mSod1(G86R),or co-transfectedwith2vectors(hWT+G86RandGFP+G86R).TheGFPvectorservedasaco-transfectioncontrol.NotethatG86R-mSod1migrates fasterthanthehSOD1constructsasnotedonrightsideoftheFigure. B) Quantificationoftherelativeaggregationpropensityofeachtransfected protein.Pairedstudent t -testswereperformedtoestablishsignificantdifferenceswithhWT:* p 0.05, # p 0.005.BarsrepresentmeanSEM.Note thattheaggregationpropensityofWT-hSOD1incellsco-transfectedwithG86R-mSod1isnotsignificantlygreaterthanwhenexpressedalone. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page4of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 5

saponintreatment(Figure4A).Bycontrast,WT-mSod1: YFPappearedtobetightlyboundtoG85R-mSod1:RFPinclusion(Figure4B).Toassesswhetherhumanandmouse SOD1werebysomemannerunabletoformintermingled inclusions,weco-expressedG86R-mSod1:RFPwithG85RhSOD1:YFP,findingthatthemutanthumanprotein formedco-mingledinclusionswiththemutantmouseprotein(Figure5A).Thedegreeofco-minglingwassimilar towhatwasobservedwhenG86R-mSod1:RFPwascoexpressedwithG86R-mSod1:YFP(Figure5B).Notably,in ourassessmentoftheimageswegenerallyobservedanall ornoneresponse;meaningthatincellsinwhichbothproteinswereover-expressedeitherallofproteinco-localized ornoneco-localized.Althoughsuchdataareeasilyquantified,theoutcomedoesnotproducearangeofvaluesand thusitwasunnecessarytotabulatequantifieddata(Table1 providesasummaryoftheobserveddata).Inthecontext oftheseobservations,theweakinteractionbetweenWThSOD1:YFPandG86R-mSod1:RFPisindicativeofsome degreeofincompatibilitybetweentheseproteinsingeneratingaggregates. RoleofnormaldimericinteractionsinmutantSOD1 aggregation Misfolded/aggregatedformsofmutantSOD1selectively reactwithantibodiesraisedagainstsequencesthatare normallyinaccessibleduetolocationwithinthedimer interface[15,24].Todeterminewhetherdisruptingthe dimerizationofWT-hSOD1orWT-mSod1mayaffect interactionwithmisfoldedmutantSOD1,weutilized experimentalmutationsknowntomonomerizeSOD1 (SOD1-F50E/G51E;[25,26]).ToconfirmthattheWT monomersofSOD1(termedSOD1mon)cannolonger formdimericinteractionswithnativeSOD1,wegeneratedaseriesofsplit-luciferasereporterconstructs[27] inwhichtheWTmonomerofhSOD1(WT-hSOD1mon)ornativeWT-SOD1wasfusedtotheN-terminal orC-terminalhalfofhumanizedGaussialuciferase (Figure6).Asexpected,co-expressionofWT-SOD1fused totheN-terminalhalfofluciferase(L1)withWT-SOD1 fusedtotheC-terminalhalfofluciferase(L2)produced highlevelsofluciferaseactivityasthehomodimerization ofWT-SOD1broughtthetwodomainsofluciferasein closeenoughproximitytoreconstituteenzymaticactivity (Figure6).Alsoasexpected,whenWT-hSOD1monwas fusedtoL1andL2domainsandco-transfected,then muchlessluciferaseactivitywasdetected(Figure6).For thislatterpairofconstructs,thelevelofactivitywassimilartowhatwasdetectedincellsexpressingthetwodomainsofluciferasealone(Figure6;L1/L2).Similarly,very lowlevelsofactivityweredetectedbyanycombinationof WT-hSOD1monandWT-hSOD1fusionstoL1orL2 (Figure6).Immunoblotsofcellstransfectedwiththesame constructsusedforluciferaseassaydemonstratedrelativelyequalexpressionoftheWT-hSOD1andWThSOD1monfusionconstructs(Additionalfile1:Figure S3).Thesefindingsindicatethattheintroductionofthe EEmutationsatresidues50/51ofSOD1producesavariantofSOD1thatisunabletohomodimerizewithitselfor heterodimerizewithnativeWT-SOD1. Figure2 FusionsofWT-mSod1withYFParefullysoluble. WT-mSod1:YFPproteinswereexpressedinCHOcellsfor24h.Cellswerefixed directlyortreatedwith0.1%Saponinpriorfixation.CellswereimmunostainedwithDAPItoallowvisualizationofcellsnuclei.Notethecamera exposuretimesusedtocaptureYFPimageswererecordedasindicatedinthefigure.Forthisfigureandtherelevantfiguresthatfollow,the imagesshownarerepresentativeof3independentexperiments. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page5of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 6

Weexaminedhowco-expressionofthemonomerversionsofWT-hSOD1andmSod1affectedtheaggregation ofG85R-hSOD1andG86R-mSod1inco-transfection studies.Interestingly,comparedtonativeWT,wefound thatmonomericWTmouseandhumanSOD1proteins werehighlypronetoformdetergentinsolublecomplexes (Figure7A,P2fraction;aggregationindexsimilartoA4VhSOD1at48hours).WhentheWTmonomerofhSOD1 (WT-hSOD1mon)wasco-expressedwithG85R-hSOD1, weobservedadecreaseinthelevelsofinsolublemutant SOD1thataccumulatedin24hours(Figure7A).ThisoutcomewassimilartowhatwepreviouslyobservedwithnativeWT-hSOD1whenco-e xpressedwithG85R-hSOD1 [19].Interestingly,intheseco-transfections,thelevelsof highlyinsolubleWT-hSOD1monwerealsosignificantlydecreasedwhenco-expressedwithG85R-hSOD1(Figure7A). Co-expressionofWT-hSOD1monwiththeG86R-mSod1 mutantfor24halsosignificantlyreducedtheaggregation Figure3 G86R-mSod1fusedwitheitherRFPorYFPformssaponin-resistantinclusions.(A) G86R-mSod1:RFPand (B) G86R-mSod1:YFP proteinswereexpressedinCHOcellsfor24h.Cellswerefixeddirectlyortreatedwith0.1%Saponinpriorfixation.Cellswereimmunostained withDAPItoallowvisualizationofcellsnuclei.NotethecameraexposuretimesusedtocaptureRFPandYFPimageswererecordedasindicated inthefigure.Digitallyenlargedinsetsareshowntotheright. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page6of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 7

oftheALSmutant(Figure7A);however,inthiscase, asignificantaccumulationofinsolubleWT-hSOD1mon remaineddetectable(Figure7A).Co-expressionofWTmSod1monwithG85R-hSOD1significantlyloweredthe accumulationofinsolublemutanthSOD1over24hourperiods(Figure7A;lanemarkedwithanasteriskwasprobed withanantibodyspecifictohSOD1).Co-expressionof WT-mSod1monwithG86R-mSod1alsoresultedindiminishedaccumulationofmutantmSod1withacoincidentreductionintheaccumulationofinsolubleWT-mSod1mon (Figure7A).Thus,intheseexperiments,whichassaythe formationofdetergent-insolublecomplexes,weobserve thatavariantofWT-SOD1thatlacksthenormaldimer interfaceretainstheabilitytomodulatetheaggregationof eithermouseorhumanmutantSOD1.Inthisregard,the monomerizedversionofWT-SOD1fromeitherspeciesis similartowhatwehavepreviouslyobservedfornativeversionsoftheseproteins[19]. ToassesswhethermonomerizedversionsofWTSOD1mayultimatelyco-aggregatewithmutantSOD1, weextendedtheintervalpost-transfectionto48hours beforeharvestingcellsfordetergentfractionation.By 48hourspost-transfection,co-sedimentationofdetergentinsolubleWT-hSOD1monwithG85R-hSOD1wasobserved Figure4 Co-transfectionofmutantmSod1fusedtoRFPwithYFPfusionsofWT-mSod1andWT-hSOD1.(A and B) Theindicatedfused constructswereexpressedinCHOcellsasexplainedinpreviousfigurelegendsandMethods.Representativeimagesof3independent experimentsareshownalongwithdigitallyenlargedinsetstotheright. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page7of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 8

(Figure7A);similartowhatwehadpreviouslyobserved whennativeWT-hSOD1isco-expressedwithG85RhSOD1[19].Interestingly,incellsco-expressingWThSOD1monwithG86R-mSod1weobservedapersistent suppressionofmutantmSod1aggregation(Figure7A). Thus,eventhoughWT-hSOD1monhasahigherinherent propensitytoformdetergent-insolublecomplexesthan nativeWT-hSOD1,anditisunabletodimerize,WThSOD1monshowedafargreaterabilitytoslowtheaggregationofG86R-mSod1thannativeWT-hSOD1(compare Figure1AwithFigure7A).Inthereciprocalexperiment, whenWT-mSod1monwasco-transfectedwithG85RhSOD1for48hourspost-transfection,weobservedthat co-expressionofWT-mSod1monnolongerinhibitedaggregationofG85R-hSOD1(Figure7A;lanemarkedwithan asteriskwasprobedwithanantibodyspecifictohSOD1). WhenWT-mSod1monwasco-expressedwithG86RmSod1,thenbothproteinscouldbedetectedindetergentinsolublefractions,similartowhenWT-hSOD1monand G85R-hSOD1areco-expressed(Figure7B).Theimages showninFigure7Aarerepresen tativeofmultipleexperiments,whichwerequantifiedtoconfirmtheeffectofWT monomerizedSOD1ofeitherspeciesontheaggregationof mutantSOD1ofeitherspecies(Figure7B).At24hours post-transfection,thecellsco-expressingmutantSOD1of eitherspecieswithWT-SOD1monofeitherspeciesaccumulatedlessinsolublemutantSOD1(Figure7B).At 48hourspost-transfection,thesuppressiveeffectofWTSOD1mononmutantSOD1aggregationdissipatedexcept forthecombinationofWT-hSOD1monwithG86R-mSod1 (Figure7B).Thesefindings indicatethatmonomerized WT-hSOD1ormSod1retainanabilitytointeractwith misfoldedformsofSOD1thatgeneratedetergentinsoluble aggregates. Inourlastsetofexperiments,weusedfusionproteins ofWT-mSod1monandWT-hSOD1monwithYFPto investigatefurtherwhethermonomericWTproteinofeitherspeciesinteractedwithmisfoldedG86R-mSod1.WThSOD1monfusionstoRFPorYFPremainsolubleand completelyreleasablebysaponin[23].Unexpectedly,fusionsofWT-mSod1montoeitherRFPorYFPproduced inclusionsthatweresaponinresistant(Figure8).Additionally,theseinclusionsweremorphologicallysimilarto thoseproducedbymutanthSOD1fusedtoeithertag. WhenG86R-mSod1:RFPwasco-expressedwithWThSOD1mon:YFPweobservedweakinteractionsthatwere onlypartiallyresistanttosaponin;whereastheinteractions betweenG86R-mSod1:RFPandWT-mSod1mon:YFPwere robustandsaponinresistant(Figure9).Thus,theinherentlyaggregationproneWT-mSod1monreadilycoassembleswithinclusionsformedbyG86R-mSod1:RFP butthemoresolubleWT-hSOD1mondoesnot.These findingsindicatethatmonomerizationofWT-hSOD1does notinduceittoassociatewithmisfoldedG86R-mSod1. Figure5 Co-transfectionofmutantmSod1fusedtoRFPwithG85R-hSOD1:YFP. TheindicatedfusedconstructswereexpressedinCHOcells asexplainedinpreviousfigurelegendsandMethods. A) Co-transfectionofvectorsforG86R-mSod1:RFPwithG85R-hSOD1:YFP. B) Co-transfection ofvectorsforG86R-mSod1:RFPwithG86R-mSod1:YFP. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page8of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 9

DiscussionInthepresentstudy,wehaveusedcellculturemodelsof mutantSOD1aggregationtoexa mineinteractionsbetween G86R-mSod1andhumanSOD1.Toconducttheseinvestigations,weusedtwoapproachesthatofferdifferentwaysof detectingSOD1misfolding.Instudiesthatassesstheformationofdetergent-insolublecomplexesbyuntaggedSOD1, weobservedthatnativehSOD1doesnotco-aggregatewith G86R-mSod1.InexperimentstovisualizeinteractionsbetweenmisfoldedG86R-mSod1andWT-hSOD1,weobservedthatascomparedtoWT-mSod1fusionproteins, WT-hSOD1fusionproteinsdidnotreadilyinteractwith G86R-mSod1inclusions.Despitesomenotedincongruitiesinthedataasdescribedbelow(summarizedin Table1),thesefindingsareconsistentwiththeideathat misfoldedmutantmSod1hasalowpotentialtotemplate Table1ComparisonofassaysforinteractionsbetweenWTandmisfoldedmutantSOD1ConstructDetergentinsolubleSaponin-resistantRFPorYFPtaggedinclusions WT-hSOD1NoWT-hSOD1:YFPNo1WT-hSOD1:RFPYes1WT-hSOD1monYesWT-hSOD1mon:YFPNo2WT-hSOD1mon:RFPNo2WT-mSod1No1WT-mSod1:YFPNo WT-mSod1:RFPYes3WT-mSod1monYesWT-mSod1mon:YFPYes WT-mSod1mon:RFPYes G86R-mSod1YesG86R-mSod1:YFPYes G86R-mSod1:RFPYes WT-mSod1?WT-mSod1:YFPYes + + G86R-mSod1Yes48hrG86R-mSod1:RFPYes WT-mSod1monYes48hrWT-mSod1mon:YFPYes + + G86R-mSod1Yes48hrG86R-mSod1:RFPYes WT-hSOD1No48hrWT-hSOD1:YFPNo + + G86R-mSod1Yes48hrG86R-mSod1:RFPYes WT-hSOD1monYes48hrWT-hSOD1mon:YFPNo + + G86R-mSod1Reduced48hrG86RmSod1:RFPYes G85R-hSOD1YesG85R-hSOD1:YFPYes2G85R-hSOD1:RFPYes2WT-hSOD1Yes48hr1WT-hSOD1:YFPNo2+ + G85R-hSOD1Yes48hr1G85R-hSOD1:RFPYes2WT-hSOD1monYes48hrWT-hSOD1mon:YFPNo2+ + G85R-hSOD1Yes48hrG85R-hSOD1:RFPYes2WT-mSod1?WT-mSod1:YFPNo4+ + G85R-hSOD1Yes48hrG85R-hSOD1:RFPYes4WT-mSod1mon?WT-mSod1mon:YFPYes5+ + G85R-hSOD1Yes48hrG85R-hSOD1:RFPYes51datafromPrudencioetal.[ 19 ];2datafromQuallsetal.[ 23 ];3Additionalfile 1 :FigureS1;4Additionalfile 1 :FigureS4;5Additionalfile 1 :FigureS6.Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page9of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 10

themisfoldingofWT-hSOD1andthushaveimplications ininterpretingpriorstudiesthatdemonstratedhighlevels ofWT-hSOD1co-expressedwithG86R-mSod1hadnoeffectondiseasecourseintransgenicmice[13]. WTandmutantSOD1interactionswithindetergentinsolubleaggregatesandcellularinclusions Previousstudiesoftransgenicmicehaveobservedthat whenWT-hSOD1proteinswereoverexpressedwithmutanthSOD1,thentheageatwhichmicedevelopeddisease phenotypeswasdramaticallyaccelerated[7-11].Importantly,thesestudieshaveclearlydemonstratedthatwhen WThumanSOD1isco-expressedwithfALSmutantsof SOD1thatareC-terminallytruncated,thenWTprotein canbedetectedindetergent-insolublecomplexesascosedimentingwithmisfoldedmutantSOD1[7,11].Overall, thesestudiessuggestthatmutantSOD1maybeableto templatemisfoldingofWT-hSOD1,inducingthelatter proteintoacquirethemisfoldedconformationofmutant SOD1. IncellculturemodelsofmutantSOD1aggregation, wehavepreviouslydemonstr atedthatco-expression ofWT-hSOD1transientlyinhibitstheaggregationof G85R-hSOD1(24hoursfollowingtransfection)butultimately,at48hoursaftertransfection,co-sedimentation ofWT-hSOD1withG85R-hSOD1indetergent-insoluble complexeswasobserved[19].Inco-transfectionsofG85RhSOD1withWT-mSod1wecouldnotdiscernwhether WT-mSod1ultimatelyco-sedimentedwithG85R-hSOD1 becausetheseproteinsessentiallyco-migrateinSDSPAGE[19].However,inco-transfectionsofWT-mSod1 withothermutants(A4VandG93AofhSOD1)itwasvery clearthatWT-mSod1didnotformco-sedimentingcomplexeswiththehumanmutantproteins[19].Inthecourse ofouranalysisofinteractionsbetweenhumanandmouse SOD1usingfluorescentfusionproteins,lookingatallpossiblecombinations,wenotedthatWT-mSod1:YFPdidnot co-aggregatewithG85R-hSOD1:RFP(Additionalfile1: FigureS4).Collectively,thesestudiesareconsistentwith theideathatnativeWT-mSod1doesnotreadilyinteract withmisfoldedformsofG85R-hSOD1;thesefindingsmay provideanexplanationforwhydeletingmSod1inmiceexpressingG85R-hSOD1hadnoobviouseffectondisease course[12]. Thepresentstudywaspromptedbytheobservation byAudetandcolleaguesinwhichco-expressionofWTFigure6 MutationsatresiduesF50E/G51EofSOD1producesavariantofSOD1thatisunabletodimerizewithitselforheterodimerize withnativeWT-hSOD1subunits. CHOcellsweretransientlytransfectedwithindicatedsplitluciferaseconstructs(illustration)andasexplained inMethods.Acoelenterazineassaywasusedinordertodemonstratetheluciferaseactivityofthesplitluciferaseconstructs.Eachassaywas repeatedatleast3timesandluminescencevalueswerequantified.Foreachindependentexperiment,thevalueswerenormalizedtothevalue obtainedwhenfusionsofeachsegmentofGlucfusedtoWT-hSOD1wereco-transfected. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page10of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 11

hSOD1hadnoeffectontheageatwhichmiceexpressing G86R-mSod1developparalysis[13].Usingabiochemical assayforaggregationinanestablishedcellmodel,we foundnoevidencethatnativeWT-hSOD1co-sediments withG86R-mSod1indetergent-insolublecomplexes.Similarly,usingfluorescentlytaggedSOD1constructsto visualizeaggregation,weobservedthatinclusionsformed byG86R-mSod1:RFPcontainlittleifanyWT-hSOD1:YFP; whereasWT-mSod1:YFPwasreadilyincorporatedinto thesestructures.Together,thesefindingsareconsistent withtheideathatnativeWT-hSOD1hasalowpropensity tobeinducedtomisfoldbyG86R. However,thestrengthofthedatafromexaminationof RFPandYFPtaggedvariantsofSOD1isunderminedbya numberofincongruentobservations(Table1).TheprimaryunderminingobservationisthatWT-hSOD1fused toYFPdoesnotrobustlyco-depositwithinclusionsformed byG85R-hSOD1fusedtoRFP[23](Table1).BasedoncotransfectionstudiesofuntaggedWTandG85R-hSOD1 [19],wewouldhavepredictedWT-hSOD1:YFPwouldcoFigure7 TheintroductionofexperimentalmutationsthatmonomerizeWT-SOD1producevariedeffectsoninteractionswithmutant SOD1.A) Immunoblotsofdetergent-insoluble(P2)andsoluble(S1)fractionsofHEK293FTcellstransfectedwithmouseand/orhumanSOD1 constructsfor24(leftpanels)or48(rightpanels)hours.BlotswereprobedwithanantibodythatrecognizesbothmouseandhumanSOD1 proteins.AsteriskdenotesprobingwithanantibodythatisspecificagainsthumanSOD1protein. B) Quantificationoftherelativeaggregation propensity,asdescribedinFigure1,forcellstransfectedfor24(blackbars)or48(whitebars)hours.Statisticalanalyseswasperformedtoestabl ish whetherthemeasuredamountofinsolublemutantSOD1incellsexpressingonlyG85R-hSOD1,oronlyG86R-mSod1,differedfromtheamount thatbecameinsolublewheneitherwasco-expressedwithmonomerizedvariantsofmSod1andhSOD1.Explanationsfornotationsonthegraph areasfollows.ThelevelsofinsolubleG85R-hSOD1incellsexpressingonlythemutantSOD1weresignificantly( p 0.05)higherthanincells co-transfectedwithWT-hSOD1mon(hWTmon)at24hrs(a)and48hrs(b)orWT-mSod1mon(mWTmon)at48hrs(b).ThelevelofG86R-mSod1 incellsexpressingonlymutantSOD1weresignificantly( p 0.05)higherthanincellsco-transfectedwithWT-hSOD1monorWT-mSod1monat 24hrs(c)orWT-hSOD1at48hrs(d).BarsrepresentmeanSEM. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page11of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 12

aggregatewithG85R-hSOD1:RFP.Thisincongruitymay bediscountedbyourpriorworkthathasdemonstrated thatthegenerationofdetergentinsolublecomplexesdoes notnecessarilyequatetoinclusionformation[15].Anothertroublesomeobservationwasthatinclusionsformed byG86R-mSod1:YFPcouldcapturemonomericvariants ofWThSOD1fusedtoRFP(Table1;Additionalfile1: FigureS5).WT-hSOD1mon:RFPonitsownappearsasa diffuselydistributedproteinthatisreleasedbysaponin, butdoesassociatewithinclusionsformedbymutant hSOD1fusedtoYFPincludingtheG85Rmutantof hSOD1[23].Inmultiplecombinationsofco-expressionof fluorescentlytaggedproteins(Table1),theG85Rvariant ofhSOD1andtheG86RvariantofmSod1showsimilar abilitiestointeractwithWThumanandmouseSOD1 proteins.ThemaindistinguishingfeaturewasthatWTmSod1:YFPreadilyco-depositedininclusionsformed byG86R-mSod1:RFP,providingaclearexampleinwhich homo-specificinteractionswerefavored.Althoughthese incongruitiesinthedatadiminishtheconclusivenessof thisportionofthestudy,thedataoverallareconsistent withthehypothesisthatnativeWT-hSOD1doesnotreadilyinteractwithmisfoldedmutantmouseSod1. RoleofdimerformationinmouseandhumanSOD1 interactions ExperimentalconversionofSOD1fromadimerictoa monomericenzymebythemutationofresidues50 (Phe)and51(Gly)toGluwasfirstdescribedbyBertini etal.[25,26].Itisthoughtthattheintroductioncharged residuesatthesesitesproducesarepulsiveeffectasthe twomonomersofSOD1attempttoalignasahomodimericenzymes.Thesemonomericenzymesretainactivityandcrystalstructuresofthisexperimentalvariant havedemonstratedthattheproteinscanfoldintoanear normalconformation[26].Thus,theengineeredmonomerofSOD1isthoughttobeWT-likeinitsproperties. ToconfirmthatthemutationsthatmonomerizeSOD1 inhibitinteractionsbetweensubunits,wedevelopedasplitluciferaseassay.Aspredicted,themonomerizationmutationsinWT-hSOD1producedaproteinthatfailedto homodimerize.Animportantconsiderationinourstudies ofinteractionsbetweenmutantSOD1andmonomerized WT-SOD1wasthatthemutantvariantspossessedintact dimerinterfacesandthustheeffectoftheengineered monomermutationscouldhavebeenlessrobust.However, datafromthesplitluciferaseassaydemonstratedthat Figure8 WT-mSod1monfusedtoRFPorYFPformssaponin-resistantinclusions. TheindicatedfusedconstructswereexpressedinCHO cellsasexplainedinpreviousfigurelegendsandMethods. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page12of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 13

SOD1harboringtheF50E/G51EmutationsdidnotefficientlydimerizewithWT-SOD1.Thus,weareconfident thatmonomerizedWT-hSOD1variantswerebehavingas monomericproteins. Ourdatademonstratethat,whenover-expressed, untaggedversionsofbothWT-hSOD1monandWTmSod1monwerepronetoformdetergentinsolublecomplexes.Thus,inboththemouseandhumanproteins,the monomerizingmutationsimpactedthetendencyofthese proteinstoaggregate.Thisfindingfitswithdatafrom severalstudiesthathavesuggestedlossofdimerizationis akeystepintheaggregationofSOD1[28-32].However, thebehaviorofRFPandYFPtaggedmonomericvariants ofhumanandmouseSOD1producedamorecomplicated picture.WeshowherethatmonomericvariantsofWTmSod1producedsaponinresistantinclusions;whereasthe monomerizedWT-hSOD1:YFPproteinshavebeenshown toexhibitpropertiesofafullysolubleprotein[23].ItappearsthatthemutationsthatmonomerizeWT-mSod1 hadamoreprofoundimpactonstructurethanthesame Figure9 Co-transfectionofG86R-mSod1:RFPwithWT-hSOD1monandWT-mSod1monfusedtoYFP.(A and B) Theindicatedfused constructswereexpressedinCHOcellsasexplainedinpreviousfigurelegendsandMethods.WT-hSOD1monweaklyinteractswithG86R-mSod1: RFPinclusionswhereasWT-mSod1monshowsastrongsaponinresistantinteraction. Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page13of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 14

mutationsinWT-hSOD1,inducingthemouseproteinto forminclusions.Additionally,itisinterestingthatthe untaggedversionofWT-hSOD1monismoreproneto adoptdetergentinsolublecomplexes,butafusionofWThSOD1montoYFPexhibitspropertiesofasolubleprotein [23](Table1).Notably,wehavepreviouslyobserved thatformationofdetergent -insolublecomplexesdoes notnecessarilyequatetoformationofinclusions;WThSOD1:YFPisgenerallymorepronetolosesolubility innon-ionicdetergentwithoutproducinginclusions [15].ForWT-hSOD1weconcludethatthemutations thatmonomerizetheproteinmayheightenpropensity toaggregateinthepresenceofnon-ionicdetergents, butitdoesseemthattheproteinissobadlyfoldedthat iteasilyformsthelargeaggregatesthatproducevisible inclusions. FromtransfectionofuntaggedSOD1variants,weobservedthatmonomerizationofWT-hSOD1hadnoobviouseffectonitsinteractionsorinfluenceoverthe aggregationofG85R-hSOD1intodetergent-insolublecomplexes.Co-expressionofWT-hSOD1monwithG85RhSOD1producedeffectssimilartowhatwehavepreviously reportedinco-transfectionofnativeWT-hSOD1withmutanthSOD1[19]:initiallyin24hoursweobservedaninhibitionofmutantSOD1aggregation,butat48hourswe observebothproteinsco-sedimentinginthedetergentinsolublefractions(seeFigure7).Similarly,weobserved thatWT-mSod1montransientlyinhibitedtheformationof detergent-insolublecomplexesbyeitherG85R-hSOD1or G86R-mSod1.At48hoursfollowingtransfection,wecould clearlydiscernthatWT-hSOD1monwasinducedtoform co-sedimentingcomplexeswithG85R-hSOD1(Table1), butincellsco-expressingWT-hSOD1monandG86RmSod1therewassignificantreductioninlevelofaggregatedG86R-mSod1(seeFigure7).Forsomecombinations ofco-transfections,itwasmoredifficulttotellwhether bothproteinsco-sedimentedbecauseofco-migrationin SDS-PAGE.Forexample,inco-expressionofWT-mSod1 withG85R-hSOD1,wewereabletoconfirmthepresence ofG85R-hSOD1inthedetergent-insolublefractionusinga humanspecificSOD1antibodybutwewereunableto distinguishwhetherWT-mSod1wasalsopresent(see Figure7).Overall,inthosecasesinwhichwecouldmake definitiveobservations,ourstudyoftheformationofdetergentinsolublecomplexesindicatesthatmonomerization ofWT-SOD1ofeitherspeciesdidnotabrogateitsability tomakehomo-specificinteractionswithmutantSOD1 (mousetomouseorhumantohuman).Thisfindingindicatesnormaldimericinteractionsarenotrequiredtoinducetheco-aggregationofmutantSOD1withWT-like protein.Moreover,forboththenativefusionproteinsof WT-hSOD1ormSod1andthemonomericvariantsweobserveapreferenceforhomo-specificinteractionsover hetero-specificinteractions. Inourstudiesoffluorescently-taggedproteins,weobservedinstancesofagreementwithourbiochemicalassays aswellasinstancesofincongru ity.Inco-expressionstudies ofG86R-mSod1:RFPwithWT-hSOD1mon:YFPandWTmSod1mon:YFP,weobservedapreferentialrecruitmentof theWTmousefusionproteinint otheinclusions;afinding thatcorroboratedthebiochemicalassays.However,theapparentselectivityofmouseformousewasoffsetbyourpreviousobservationthatWT-hSOD1mon:YFPdidnotreadily incorporateintoinclusionsformedbyG85R-hSOD1:RFP [23](Table1).Moreover,wehaveobservedthatWThSOD1mon:RFP,whichbehavesasasolubleprotein,readilyinteractswithinclusionsformedbymutantSOD1fused toYFP[23](Table1).WT-hSOD1mon:RFPalsointeracts withinclusionsformedbyG86R-mSod1:YFP(Additional file1:FigureS5);afindingthatissomewhatincongruent withourbiochemicaldata.Thus,overall,thedatageneratedbystudieswithfluorescentlytaggedvariantsofhuman andmouseSOD1partiallycorroboratethebiochemical.ConclusionsInthepresentstudy,wehaveusedbothbiochemicaland visualassaystodeterminethedegreetowhichavariantof mouseSOD1carryingthefALSmutation(G85R/G86R)interactswithWT-hSOD1.Wehavealsoexaminedtherole ofnormaldimericinteractionsbetweenSOD1subunitsin thegenerationofaggregatesthatcontainbothproteins. Ourbiochemicalassaysforacquisitionofdetergentinsolubility,inwhichuntaggedSOD1wasover-expressed, indicatethatmutantmouseSOD 1preferentiallyinteracts withWTmouseSOD1andconverselymutanthuman SOD1preferentiallyinteractswithWThumanSOD1. MutationsthatmonomerizetheWTproteindonotsignificantlydiminishtheseinteractions.Thus,forhomo-specific interactionsinthegeneratio nofdetergent-insolublecomplexes,thenormaldimericinteractionisnotrequired.Our visualassaysofaggregationandco-aggregation,usingYFP taggedproteins,paintamorec omplexpicturepossiblybecausethefluorescenttagthatwasusedtotracktheprotein wasnotcompletelybenign[23](Table1).ThemostclearcutobservationthatcouldbemadewasthatG86R-mSod1 readilyinteractswithWT-mS od1whethernativeormonomerized.Overall,thedataareconsistentwiththehypothesis thatincreasingthelevelofWT-hSOD1inmiceexpressing G86R-mSod1doesnotacceleratetheonsetofdisease,as observedbyAudetetal.[13],becausethehumanprotein doesnotreadilyinteractwithmisfoldedmouseprotein.AdditionalfileAdditionalfile1:FigureS1. WT-mSod1:RFPformssaponin-resistant inclusions. FigureS2. AdditionalcomparativedataforWT-mSod1:RFP co-expressedwithWTmSod1:YFP,WT-hSOD1:YFP,orG85R-hSOD1:YFP. FigureS3. ImmunoblottodemonstratesimilarlevelsofexpressionofQualls etal.MolecularNeurodegeneration 2013, 8 :46 Page14of15 http://www.molecularneurodegeneration.com/content/8/1/46

PAGE 15

SOD1:GLucfusionproteins. FigureS4. WT-mSod1:YFPdoesnotbind toG85R-hSOD1:RFPinclusions. FigureS5. WT-hSOD1mon:RFPforms intermingledinclusionswithG86R-mSod1:YFP. FigureS6. WT-mSod1mon: YFPformsintermingledinclusionswithG85R-hSOD1:RFP. Competinginterests Theauthorsdeclarethattheyhavenocompetinginterests. Authors ’ contribution DQ,MP,andKCgeneratedthevariousYFPandRFPfusionconstructs described.DQcarriedoutthecellimagingexperiments.MPcarriedoutthe celltransfectionanddetergentextractionexperiments.BRandHBgenerated luciferasereporterconstructsandcarriedouttheexperiments.DBdesigned theexperimentsanddraftedthemanuscript.Allauthorsreadandapproved thefinalmanuscript. Authordetails1FromtheDepartmentofNeuroscience,CenterforTranslationalResearchin NeurodegenerativeDisease,SantaFeHealthCareAlzheimer ’ sDisease ResearchCenter,McKnightBrainInstitute,UniversityofFlorida,Box100159, 1275CenterDrive,RoomJ491,Gainesville,FL32610,USA.2Departmentof Neuroscience,MayoClinicJacksonville,4500SanPabloRdS,Jacksonville,FL 32224,USA. Received:5July2013Accepted:4December2013 Published:17December2013 References1.FridovichI: Superoxidedismutases. AdvEnzymolRelatAreasMolBiol 1974, 41: 35 – 97. 2.PargeHE,HallewellRA,TainerJA: Atomicstructuresofwild-typeand thermostablemutantrecombinanthumanCu,Znsuperoxidedismutase. ProcNatlAcadSciUSA 1992, 89: 6109 – 6113. 3.OgiharaNL,PargeHE,HartPJ,WeissMS,GotoJJ,CraneBR,TsangJ,SlaterK, RoeJA,ValentineJS, etal : Unusualtrigonal-planarcopperconfiguration revealedintheatomicstructureofyeastcopper-zincsuperoxide dismutase. Biochemistry 1996, 35: 2316 – 2321. 4.BruijnLI,MillerTM,ClevelandDW: Unravelingthemechanismsinvolvedin motorneurondegenerationinALS. AnnuRevNeurosci 2004, 27: 723 – 749. 5.PrudencioM,HartPJ,BorcheltDR,AndersenPM: Variationinaggregation propensitiesamongALS-associatedvariantsofSOD1:correlationto humandisease. HumMolGenet 2009, 18: 3217 – 3226. 6.GurneyME,PuH,ChiuAY,DalCantoMC,PolchowCY,AlexanderDD, CaliendoJ,HentatiA,KwonYW,DengH-X, etal : Motorneurondegeneration inmicethatexpressahumanCu,Znsuperoxidedismutasemutation. Science 1994, 264: 1772 – 1775. 7.DengHX,ShiY,FurukawaY,ZhaiH,FuR,LiuE,GorrieGH,KhanMS,Hung WY,BigioEH, etal : Conversiontotheamyotrophiclateralsclerosis phenotypeisassociatedwithintermolecularlinkedinsolubleaggregates ofSOD1inmitochondria. ProcNatlAcadSciUSA 2006, 103: 7142 – 7147. 8.DengHX,JiangH,FuR,ZhaiH,ShiY,LiuE,HiranoM,DalCantoMC, SiddiqueT: MoleculardissectionofALS-associatedtoxicityofSOD1in transgenicmiceusinganexon-fusionapproach. HumMolGenet 2008, 17: 2310 – 2319. 9.JaarsmaD,TeulingE,HaasdijkED,DeZeeuwCI,HoogenraadCC: Neuronspecificexpressionofmutantsuperoxidedismutaseissufficientto induceamyotrophiclateralsclerosisintransgenicmice. JNeurosci 2008, 28: 2075 – 2088. 10.WangL,DengHX,GrisottiG,ZhaiH,SiddiqueT,RoosRP: Wild-typeSOD1 overexpressionacceleratesdiseaseonsetofaG85RSOD1mouse. HumMolGenet 2009, 18: 1642 – 1651. 11.PrudencioM,DurazoA,WhiteleggeJP,BorcheltDR: Anexaminationofwild-typeSOD1inmodulatingthetoxicityandaggregationofALSassociatedmutantSOD1. HumMolGenet 2010, 19: 4774 – 4789. 12.BruijnLI,HouseweartMK,KatoS,AndersonKL,AndersonSD,OhamaE, ReaumeAG,ScottRW,ClevelandDW: Aggregationandmotorneuron toxicityofanALS-linkedSOD1mutantindependentfromwild-type SOD1. Science 1998, 281: 1851 – 1854. 13.AudetJN,GowingG,JulienJP: Wild-typehumanSOD1overexpression doesnotacceleratemotorneurondiseaseinmiceexpressingmurine Sod1G86R. NeurobiolDis 2010, 40: 245 – 250. 14.RippsME,HuntleyGW,HofPR,MorrisonJH,GordonJW: Transgenicmice expressinganalteredmurinesuperoxidedismutasegeneprovidean animalmodelofamyotrophiclateralsclerosis. ProcNatlAcadSciUSA 1995, 92: 689 – 693. 15.PrudencioM,BorcheltDR: Superoxidedismutase1encodingmutationslinked toALSadoptsaspectrumofmisfoldedstates. MolNeurodegener 2011, 6: 77. 16.MizushimaS,NagataS: pEF-BOS,apowerfulmammalianexpression vector. NucleicAcidsRes 1990, 18: 5322. 17.WangJ,SluntH,GonzalesV,FromholtD,CoonfieldM,CopelandNG, JenkinsNA,BorcheltDR: Copper-binding-site-nullSOD1causesALSin transgenicmice:aggregatesofnon-nativeSOD1delineateacommon feature. HumMolGenet 2003, 12: 2753 – 2764. 18.KarchCM,PrudencioM,WinklerDD,HartPJ,BorcheltDR: Roleofmutant SOD1disulfideoxidationandaggregationinthepathogenesisof familialALS. ProcNatlAcadSciUSA 2009, 106: 7774 – 7779. 19.PrudencioM,DurazoA,WhiteleggeJP,BorcheltDR: Modulationofmutant superoxidedismutase1aggregationbyco-expressionofwild-type enzyme. JNeurochem 2009, 108: 1009 – 1018. 20.KarchCM,BorcheltDR: Alimitedrolefordisulfidecross-linkinginthe aggregationofmutantSOD1linkedtofamilialamyotrophiclateral sclerosis. JBiolChem 2008, 283: 13528 – 13537. 21.StevensJC,ChiaR,HendriksWT,Bros-FacerV,VanMJ,MartinJE,JacksonGS, GreensmithL,SchiavoG,FisherEM: Modificationofsuperoxidedismutase1 (SOD1)propertiesbyaGFPtag – implicationsforresearchintoamyotrophic lateralsclerosis(ALS). PLoSONE 2010, 5:e9541. 22.FrancisG,KeremZ,MakkarHP,BeckerK: Thebiologicalactionofsaponins inanimalsystems:areview. BrJNutr 2002, 88: 587 – 605. 23.QuallsDA,CrosbyK,BrownH,BorcheltDR: Ananalysisofinteractions betweenfluorescently-taggedmutantandwild-typeSOD1inintracellular inclusions. PLoSOne .inpress. 24.RakhitR,RobertsonJ,VandeVC,HorneP,RuthDM,GriffinJ,ClevelandDW, CashmanNR,ChakrabarttyA: Animmunologicalepitopeselectivefor pathologicalmonomer-misfoldedSOD1inALS. NatMed 2007, 13: 754 – 759. 25.BertiniI,PiccioliM,ViezzoliMS,ChiuCY,MullenbachGT: Aspectroscopic characterizationofamonomericanalogofcopper,zincsuperoxide dismutase. EurBiophysJ 1994, 23: 167 – 176. 26.BanciL,BenedettoM,BertiniI,DelConteR,PiccioliM,ViezzoliMS: Solution structureofreducedmonomericQ133M 2copper,zincsuperoxidedismutase (SOD).WhyisSODadimericenzyme? Biochemistry 1998, 37: 11780 – 11791. 27.KimSB,SatoM,TaoH: SplitGaussialuciferase-basedbioluminescence templatefortracingproteindynamicsinlivingcells. AnalChem 2009, 81: 67 – 74. 28.RakhitR,CrowJP,LepockJR,KondejewskiLH,CashmanNR,ChakrabarttyA: MonomericCu,Zn-superoxidedismutaseisacommonmisfolding intermediateintheoxidationmodelsofsporadicandfamilial amyotrophiclateralsclerosis. JBiolChem 2004, 279: 15499 – 15504. 29.KhareSD,CaplowM,DokholyanNV: FALSmutationsinCu,Znsuperoxide dismutasedestabilizethedimerandincreasedimerdissociation propensity:alarge-scalethermodynamicanalysis. Amyloid 2006, 13: 226 – 235. 30.SvenssonAK,BilselO,KayatekinC,AdefusikaJA,ZitzewitzJA,MatthewsCR: Metal-freeALSvariantsofdimerichumanCu,Zn-superoxidedismutasehave enhancedpopulationsofmonomericspecies. PLoSONE 2010, 5: e10064. 31.AuclairJR,BoggioKJ,PetskoGA,RingeD,AgarJN: Strategiesforstabilizing superoxidedismutase(SOD1),theproteindestabilizedinthemost commonformoffamilialamyotrophiclateralsclerosis. ProcNatlAcadSci USA 2010, 107: 21394 – 21399. 32.RaySS,NowakRJ,StrokovichK,BrownRHJr,WalzT,LansburyPTJr: Anintersubunitdisulfidebondpreventsinvitroaggregationofa superoxidedismutase-1mutantlinkedtofamilialamytrophiclateral sclerosis. Biochemistry 2004, 43: 4899 – 4905.doi:10.1186/1750-1326-8-46 Citethisarticleas: Qualls etal. : Featuresofwild-typehumanSOD1limit interactionswithmisfoldedaggregatesofmouseG86RSod1. Molecular Neurodegeneration 2013 8 :46.Qualls etal.MolecularNeurodegeneration 2013, 8 :46 Page15of15 http://www.molecularneurodegeneration.com/content/8/1/46