Activation of a synapse weakening pathway by human Val66 but not Met66 pro brain derived neurotrophic factor (proBDNF)

(1)

Contents lists available atScienceDirect

Pharmacological

Research

j o u r n a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / y p h r s

Activation

of

a

synapse

weakening

pathway

by

human

Val66

but

not

Met66

pro-brain-derived

neurotrophic

factor

(proBDNF)

Sumangali

Kailainathan

a,e,1

_,

_Thomas

_M.

_Piers

a,b,c,1

_,

_Jee

_Hyun

_Yi

a,b

_,

_Seongmin

_Choi

a,b,c

_,

Mark

S. Fahey

a

_,

_Eva

_Borger

d

_,

_Frank

_J.

_Gunn-Moore

d

_,

_Laurie

_O’Neill

a,b

_,

_Michael

_Lever

a,b

_,

Daniel

J. Whitcomb

a,b,c

_,

_Kwangwook

_Cho

a,b,c,∗,1

_,

_Shelley

_J.

_Allen

a,e,∗,1

a_Henry_Wellcome_Laboratories_for_Integrative_Neuroscience_and_{Endocrinology,}_School_of_Clinical_Sciences,_Faculty_of_Health_Sciences,_University_of_Bristol,

WhitsonStreet,BristolBS13NY,UK

b_Centre_for_Synaptic_Plasticity,_University_of_Bristol,_Whitson_Street,_Bristol_BS1_3NY,_UK

c_{Chonnam-Bristol}_Frontier_Laboratory,_Biomedical_Research_Institute,_Chonnam_National_University_Hospital,_Gwangju_501-757,_South_Korea d_Medical_and_Biological_Sciences_Building,_University_of_St._Andrews,_Fife_KY16_9TF,_UK

e_Learning_&_Research,_School_of_Clinical_Sciences,_Faculty_of_Health_Sciences,_Southmead_Hospital,_Bristol_BS10_5NB,_UK

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received23November2015

Receivedinrevisedform5December2015 Accepted6December2015

Availableonline11December2015

Chemicalcompoundsstudiedinthisarticle: GSK3␤inhibitorCT-99021

6-[2-[[4-(2,4-dichlorophenyl)-5-(5-methyl-1H-imidazol-2-yl)pyrimidin-2-yl]]amino] ethylamino]pyridine-3-carbonitrile.Cas no.252917-06-9

Keywords:

Pro-brain-derivedneurotrophicfactor (proBDNF)

Val66Metpolymorphism Long-termdepression(LTD) Bindingkinetics

Neurotrophinreceptors

Chemicalcompoundsstudiedinthisarticle

a

b

s

t

r

a

c

t

Thisstudydescribesafundamentalfunctionaldifferencebetweenthetwomainpolymorphismsofthe pro-formofbrain-derivedneurotrophicfactor(proBDNF),providinganexplanationastowhytheseforms havesuchdifferentage-relatedneurologicaloutcomes.HealthyyoungcarriersoftheMet66form(present in∼30%Caucasians)havereducedhippocampalvolumeandimpairedhippocampal-dependentmemory function,yetthesamepolymorphicpopulationshowsenhancedcognitiverecoveryaftertraumaticbrain injury,delayedcognitivedysfunctionduringaging,andlowerriskoflate-onsetAlzheimer’sdisease(AD) comparedtothosewiththemorecommonVal66polymorphism.Toexaminethedifferencesbetweenthe proteinpolymorphismsinstructure,kineticsofbindingtoproBDNFreceptorsandinvitrofunction,we generatedpuriﬁedcleavage-resistanthumanvariants.Intriguingly,wefoundnostatisticaldifferences inthosecharacteristics.Asanticipated,exogenousapplicationofproBDNFVal66torathippocampal slicesdysregulatedsynapticplasticity,inhibitinglong-termpotentiation(LTP)andfacilitatinglong-term depression(LTD).Wesubsequentlyobservedthatthisoccurredviatheglycogensynthasekinase3␤ (GSK3␤)activationpathway.However,surprisingly,wefoundthatMet66hadnosucheffectsoneither LTPorLTD.Thesenovelﬁndingssuggestthat,unlikeVal66,theMet66variantdoesnotfacilitatesynapse weakeningsignaling,perhapsaccountingforitsprotectiveeffectswithaging.

Abbreviations: JC-1 dye, 5_,6,6_{-tetrachloro-1,1}_,3,3 -tetraethylbenzimidazolylcarbocyanineiodide; AD,Alzheimer’sdisease;ANOVA, analysis of variance; aCSF, artificial cerebrospinal fluid; BDNF, brain-derived neurotrophicfactor;CD,circulardichroism;DIV,daysinvitro;DTT,dithiothreitol; ECD,extracellulardomain; fEPSP,fieldexcitatorypostsynapticpotentials; FCR, furincleavage-resistant;GSK3␤,glycogensynthase kinase3␤;IPTG,isopropyl ␤-D-1-thiogalactopyranoside;LDH,lactatedehydrogenase;LTD,long-term depres-sion;LTP,long-termpotentiation;LD,luminaldomain;NAD,nicotinamideadenine dinucleotide;p75NTR,pan-neurotrophinreceptor;PMS,phenazinemethosulfate; PONDR,predictorofnaturallydisorderedregions;RU,responseunits;SNP,single nucleotide polymorphism; SDS PAGE, sodium dodecyl sulfatepolyacrylamide electrophoresis;s.e.m.,standarderrorofmean;SPR,surfaceplasmonresonance; TrkBIg2, TrkB immunoglobulin-like domain 2; TrkB, tyrosine kinase B; WT,

wild-type.

∗_{Corresponding}_authors.

E-mailaddresses:[email protected](K.Cho),[email protected] (S.J.Allen).

1. Introduction

Theneurotrophinbrain-derivedneurotrophicfactor(BDNF)is ﬁrsttranslatedastheprecursorform,proBDNF.Acommonsingle nucleotidepolymorphism(SNP)rs6265 atcodon66 inthe pro-domainofhumanproBDNFchangestheaminoacidsequencefrom thecanonicalsequenceofvaline(Val66)tomethionine(Met66). Whilstgenerallylessprevalent,theMet66polymorphismcanbe foundin∼50%Asians,∼30%Caucasianand∼4%African–Americans

[1]. Interestingly,population studiesshow healthyyoung adult Met66carrierstohaveasigniﬁcantvolumetricreductionin spe-ciﬁc brain structures includingthe hippocampus[2–4]and are

1_These_authors_contributed_equally_to_this_work.

http://dx.doi.org/10.1016/j.phrs.2015.12.008

(2)

knowntosuffer fromhippocampal-dependent memorydeﬁcits

[2,5]. Intriguingly however, studies also showa slower overall cognitivedeclinewithincreasingageinhealthyproBDNFMet66 carriers[4–7]comparedwithVal66.Furthermore,Met66carriers haveincreasedrecoveryofhighorder executivefunctioning[8]

andgeneralcognitivefunctions[9] aftertraumaticbrain injury. Accordingly,evidencewouldsuggestthatMet66conferssome neu-roprotectiveeffectsthatarenotpresentwithVal66.Exactlywhy thisisthecaseandthemechanismsresponsible,however,remain tobefullyexplained.

BDNFmediatesneurotrophicfunctionsincludingneuronal sur-vivalanddifferentiation,mainlyviaitsspeciﬁcreceptor,tyrosine kinase B (TrkB) [10,11]. In contrast, studies have shown that thecommonproBDNF Val66predominantly facilitates unfavor-ablesignalingsuchasgrowthconecollapse[12,13],reductionin spinedensity[14,15]andapoptosis[14,16–19]throughits pan-neurotrophinreceptor(p75NTR)and itsco-receptorssortilin or thesortilin-relatedVPS10domainfamilymemberSorCS2.Synaptic plasticity,widelyconsideredtobeacellularcorrelateforlearning andmemoryinthebrain,isalsodifferentiallyregulatedbyBDNF andproBDNF Val66;whilstBDNF isimportantin theinduction oflong-termpotentiation(LTP) andsynapsestrengthening[20], bycontrast,proBDNFVal66facilitateslong-termdepression(LTD) and thedegradation ofsynapses [15,21]. Accordingly,proBDNF Val66appearstodirectlymodulatesynapsefunctionby promot-ingsynapseweakening.TherelationshipbetweenproBDNFMet66 andsynapsefunction,however,remainstobefullycharacterized.

EarlierstudiesshowedthatthepresenceofoneMet66allele issufﬁcient toimpairtrafﬁckingoftheproBDNF Met66variant tothedistal dendrites, leadingto a reduction in both activity-dependentsecretionandsynaptic targetingof BDNF-containing vesiclesinneurons[2,22,23].Indeed,invitroelectrophysiological studiesofMet66homozygousknock-inmiceshowedaselective impairmentinactivity-dependentsynapticplasticity[24]. How-ever,thesechangesseeninMet66micecouldbeattributedtothe reducedhippocampalBDNFlevelsassociatedwiththis polymor-phism[25],anditisyettobedeterminedwhethertheseobserved impairmentsareactuallydue,instead,todifferencesinbindingto thereceptorsp75NTR[16],sortilin[21]and SorCS2[12]and/or intracellularneuronalsignalinginducedbytheproBDNFVal66and Met66variants.

Thepresent studyaimedtoanalyzeand comparethe struc-tureof theproBDNF variants,their receptor bindingdynamics, neuronalintracellularsignaling,andcapacitytoregulateLTPand LTD.Althoughpreviousstudieshave examinedtheMet66 vari-antinknock-inmice[24],herewepurifiedtwocleavage-resistant human proBDNF variants and used the proteins in exogenous applicationassays.Doingsoallowedustoexcludeanypotential endogenousdifferencesinproteinsecretionandindirectBDNF sig-naling,aswellasavoidpossibleunknowngeneticfactorspresent inknock-inmodels.ConstructsofproBDNFwereproducedintheir non-glycosylatedformsinEscherichiacoli.Others[14]reportno functionaldifferencebetweenproBDNFobtainedfrombaculovirus andE.coli.Eventhoughwecannotcompletelyexcludethe influ-enceofpost-translationalmodificationsonproteinstructureand function,inthisstudytheVal66andMet66polymorphismswere expressed, purified and treated identically throughout and are comparedonthisbasis.

2. Methods

2.1. Sub-cloningofproBDNFexpressionconstructs

HumanproBDNFDNAconstructswereproducedas proteolyt-icallycleavablewild-type(WT)andfurincleavage-resistant(FCR) forms:(withmutationforFCR):127_RR128_to127_AA128, _numbering

asforhumanproBDNFcanonicalisoform1(ExPASy bioinformat-icsproteinidentiﬁer,P23560-1).ProBDNFMet66wasformedfrom Val66DNAconstructbysitedirectedmutagenesis(QuikChange®

Site-DirectedMutagenesisKit,Stratagene).GenomicDNAwasused asatemplatetogenerateproBDNFconstructswithoutthesignal sequenceanddevoidofextraneoustags.Sub-cloningwasintopET 24a(+)(Novagen)withtransformationintoXL1-BlueE.coli (Strata-gene).

2.2. ProBDNFproteinexpression,refoldingandpuriﬁcation

ProBDNFexpressionwasperformedwithmodificationstothe methodpreviously described forproNGF [26], pET 24a(+) plas-midsencodingforproBDNF WTandFCR weretransformedinto BL21 (DE3) RIPL E. coli (Stratagene). ProBDNF expression was inducedusing1mMIPTG(isopropyl_␤-D-1-thiogalactopyranoside) andcellsgrownfor3hat37◦Candthenharvestedby centrifu-gation.ProBDNF,expressedininclusion bodies,wasisolatedby celllysisusinganX-press(Nikehydraulics,Sweden).Theinclusion bodieswerewashedin100mMTrisbuffer(pH8.0)containingin succession,100mMNaCland2.5%(v/v)proteasecocktailinhibitor (Sigma#P2174),1MNaCl,1%TritonX-100andfinally100mM NaCl.Inclusion bodiesweresolubilized and purifiedwith mod-ifications to the previously described method for proNGF [27]. Solubilizationwascarriedoutin6Mguanidinehydrochloride con-taining50mMdithiothreitol(DTT),50mMTrispH8.0and1mM EDTA.Thesolubilizedproteinswererefoldedbyrapiddilutioninto 100mMTrispH8.5,750mMl_-arginine,₅_mM_EDTA,₅_mM_reduced glutathioneand 2mMoxidizedglutathione; sampleswerethen incubatedfor4hat10◦C.Thiswasdialyzedagainstbuffer contain-ing50mMsodiumphosphatepH7.2,1mMEDTAand0.05%(v/v) proteasecocktailinhibitorat4◦C.

2.3. TrkBimmunoglobulin-likedomain2(TrkBIg2)protein expressionandpuriﬁcation

TrkBIg2 was expressed, puriﬁed and refolded as previously described[27].

2.4. Preparationoflysates,sodiumdodecylsulfate

polyacrylamideelectrophoresis(SDSPAGE)analysisandwestern blotting

Hippocampipre-incubatedwithproBDNFweremicrodissected toisolatetheCA1dendritic-richregion,aspreviouslydescribed

[28], andresolvedonSDS PAGE.Primaryantibodieswere anti-BDNFN20(SantaCruz,#sc-546);anti-phosphoserine9GSK3␤(Cell Signaling,#9336);anti-GSK3␤(Santa Cruz,#sc-9166);anti-pan tau(Tau5,Invitrogen#AHB0042);anti-phosphorylatedtau(PHF-1, kindlyprovidedbyProfessorPeterDavies,FeinsteinInstitute,NY); anti-␤-actin(AbCam,#ab6276).

2.5. AcidicnativePAGEanalysis

BDNFand proBDNF are basicproteinswith highiso-electric points(pI)of8.5–9.6andwereanalyzedon15%acidicnativePAGE todeterminethepresenceofoligomersandsolubleaggregates.Gels werestainedwithCoomassieBluetovisualizethebands.

2.6. Furindigestion

(3)

#P8077)at30◦Cfor1h.SampleswereresolvedonSDSPAGEand analyzedbywesternblotting.

2.7. Circulardichroism(CD)

SecondarystructuresofproBDNFvariants wereevaluatedby Far-UV(195–260nm)CDanalysis.Dataacquisitionwasoutsourced toDepartmentofChemistry,UniversityofWarwick,Coventry,UK andwascarriedoutbyDr.VincentHallandProfessorAlisonRodger usingaJascoJ-815spectrometer.Spectrawererecordedin0.2nm incrementsat100nm/minscanningspeedwithresponsetimeof 1s.Datawereacquiredusinga1mmpathlengthcuvette, accumu-latingeightscansperprotein.MeanresiduemolarCD(␧;moles aminoacids−1_dm3_cm−1₎_were_plotted_as_a_function_of_wavelength (nm).

2.8. Surfaceplasmonresonance(SPR)

SPRwasperformedonaBiacoreT200(GEHealthcare),as previ-ouslydescribed[27]todeterminethekineticprofilesofproBDNF variantstotheirreceptorbindingdomainsTrkBIg2 [27],p75NTR extracellulardomain(p75NTRECD)(SinoBiologicalsInc., #3184-H08H),sortilinluminaldomain(LD)(R&Dsystems,#3154-ST-050) andSorCS2LD(R&Dsystems,#4238-SR-050).TheligandsTrkBIg2 (17.8±0.6fmol/mm2_), _p75NTR _ECD _(40.5_±_1.8_fmol/mm2_), sor-tilin LD(38.8±1.1fmol/mm2₎_and _SorCS2 _LD_(32.35_fmol/mm2₎ wereimmobilizedonCM5-SSeriessensorchips(GEHealthcare, #BR-1005-30)usingtheamine-couplingmethodaccordingtothe manufacturer’sinstructions.TheproBDNFvariants(analytes)ata rangeof0,0.625,1.25,2.5,5,10,20nMwereallowedtoflowover thechipsurfaceinrunningbuffer(10mMHEPES,150mMNaCl, 3mMEDTA,0.05%(v/v)surfactantP20atafinalpH7.4). Regener-ationofthechipsurfacewascarriedoutusing10mMGlycinepH 1.5.Kineticanalyzes(n≥3)wereperformedbyfittingthekinetic plotstotheLangmuir1:1mathematicalmodelusingBiacoreT200 evaluationsoftware.Alldataweredouble-referencedandplotted asafunctionoftime.Doublereferencingincludedsubtractionof theinactive referencesurfaceSPRresponseunits(RU)fromthe ligand-analyteRUandsubtractionofbufferassociationSPRRUat 0nManalyte ontheactivatedflowcell fromtheligand-analyte responses.

2.9. Primarycellcultureandlactatedehydrogenase(LDH)assay

DissociatedcorticalneuronsfromCD1miceembryonicday16 (E16)werepreparedaspreviouslydescribed[30].Cellswereplated inneurobasalmediawithB27supplementat1.25×105_cells/cm2 onpoly-d-lysinecoated96-wellplatesandmaintainedat37◦Cin ahumidiﬁed5%CO2incubatoruntilDIV11.Corticalneuronswere incubatedinserumfreemediafor2handsubsequentlytreated withproBDNFvariants(0,0.05–10nM)for18hat37◦C,andLDH releasemeasured.Cellculturesupernatantwastransferredinto 96-wellplatesandLDHassayreagent(0.2MTrispH8.2,5mg/mllactic acidand1.3mg/mlreactionmixcontaining0.35mgtetrazolium, 0.09mgphenazinemethosulfate(PMS)and0.9mgnicotinamide adeninedinucleotide (NAD))was added, incubatedat 37◦C for 10minandtheabsorbancemeasuredusinga FLUOstarOPTIMA platereaderat485nm.

2.10. PrimarycellcultureandJC-1assay

E16mousecorticalneurons(DIV11)wereusedtoassess mito-chondrialmembrane potentialinresponse toproBDNF.Cortical cellsculturedin96-wellplatesweretreatedwithpuriﬁedproBDNF WTVal66andMet66(0,0.05, 0.1,1,5,10nM).Eachtreatment wascarriedoutintriplicateandtheplateswereincubatedfor18h

at37◦C.Followingincubation,the96-wellplateswereprocessed bythe5,6,6-tetrachloro-1,1,3,3-tetraethyl benzimidazolylcarbo-cyanineiodide(JC-1dye)assaytomeasuremitochondrialpotential. TherelativeﬂuorescenceofJ-aggregateswasmeasuredusingthe FLUOstarOPTIMAplatereader,indicatingchangeinmitochondrial membranepotential.

2.11. Animals

Allproceduresinvolvinganimalswerecarriedoutinaccordance withtheUKAnimals(ScientiﬁcProcedures)Act,1986and associ-atedguidelines.MaleWistarrats(CharlesRiver,UK)wereusedto prepareacutehippocampalslices(P24–26).Ratswerehousedfour orﬁvepercageandallowedaccesstowaterandfoodadlibitum.The cagesweremaintainedataconstanttemperature(23±1◦C)and relativehumidity(60±10%)undera12hlight/darkcycle(lights onfrom07:30to19:30).

2.12. Slicepreparation

Animalswerekilledbycervicaldislocationanddecapitation. Followingthis, the brain wasrapidlyremoved and placed into ice-coldartiﬁcialcerebrospinalﬂuid(aCSF;continuouslybubbled with95%O2/5%CO2)containing124mMNaCl,3mMKCl,26mM NaHCO3,1.25mMNaH2PO4,2mMCaCl2,1mMMgSO4,and10mM d-glucose.Hippocampiwereextractedandtransverse hippocam-palslices (400␮mthickness)were cutusing a McIlwaintissue chopper(Mickle LaboratoryEngineeringCo.).Followingmanual separation, thesliceswerethen submergedinaCSF fora mini-mumof1hforrecoverybeforeexperimentscommenced.Slices werepre-incubatedwith1nMproBDNFVal66orMet66in5ml aCSFfor90min.

2.13. Electrophysiology

Followingpre-incubation, fieldrecordingsweremeasuredat 30◦Cwithcontinuousperfusionwith1nMproBDNF (3ml/min). LTPwasinducedbydeliveringtwotetanicstimulitotheSchaffer collateralinput(CA3-CA1pathway)ofacutehippocampusslices. NMDAR-dependentLTDwasinducedbydeliveringLFS(1Hz,900 pulses,15min)attheCA3-CA1pathway[28].Fieldexcitatory post-synaptic potentials (fEPSP) were measuredin the hippocampal CA1regionusingrecordingelectrodesfilledwith3MNaCl.The dataacquisitionandanalysiswasperformedusingWinLTP(www. winltp.com).Briefly,theslopeoftheevokedfEPSPswasmeasured, normalizedtothepre-conditioningbaselineandexpressedasa percentageofbaseline.Themeanvaluesoffiveconsecutive record-ings(EPSPs) were calculatedat approximately 60minafter the inductionofLFS.

2.14. Studyapproval

Allanimal experimentswerecarriedout inaccordance with theUKScientiﬁcProceduresAct,1986andassociatedguidelines; methodswerecarriedoutinaccordancewiththeapproved guide-lines.AllexperimentalprotocolswereapprovedbytheUniversity ofBristolAnimalWelfare&EthicalReviewBody.

2.15. Statistics

(4)

ortwo-wayANOVAfollowedbyBonferroni’spost-hoctest. Signiﬁ-cancewasdeﬁnedas*_P_<_0.05,**_P_<_0.01,***_P_<_0.001.

3. Results

3.1. NomajordifferencesinstructurebetweentheproBDNF Val66andMet66variants

BDNFisﬁrsttranslatedinaprecursorformasproBDNFandis thenN-terminallycleavedbyproteasessuchasfurin[29],plasmin

[20]andmatrixmetallo-proteinases[31,32]toformmatureBDNF (∼14kDa).TheproBDNFVal66andMet66variantsproducedinthis studyresolvedsimilarlyonSDSPAGE,showingcomparable molec-ularweightof∼26kDawith>95%purity(Fig.1a).Furthermore, acidicnativePAGEgelsshowedtherecombinantproteinstobe devoidofoligomerizationandaggregation(Fig.1b)(noaggregation wasseeninthestackinggel;supplementaryFig.S1showingfull uncutgelandblots)andproteolyticcleavageusingfurinshowed bothcleavage-resistantproteinvariantstobecompletelyresistant tofurincleavage(Fig.1c).

Supplementrymaterialrelatedtothisarticlefound,intheonline version,athttp://dx.doi.org/10.1016/j.phrs.2015.12.008.

To determine differences in variant structure, analysis was performedontwofronts:Firstly,insilicopredictorofnaturally dis-orderedregions(PONDR),aweb-basedneuralnetwork[33–35]and secondlyinvitrofar-UVcirculardichroism(CD).ThePONDR predic-tionusingtheaminoacidsequencescorrespondingtothecanonical humanproBDNFVal66isoform1,P23560-1(1–247aminoacids) anditsvariantMet66polymorphism(VAR004626)showedMet66 tobemoredisordered intheregionsurroundingtheVal66Met polymorphism(Fig.1d).Aminoacidresiduesarepredictedhere asdisorderedwhenthePONDRoutputvalue≥0.5.Far-UVCD anal-ysisshowedthatoverallbothvariants’spectrahadsimilarshapes inthefar-UVregionwithaminimumpeakat∼204nm,indicative ofapredominant␤-sheetstructure(Fig.1e).Spectraloverlay fur-thershowedproBDNFMet66tocontainasmallincreaseinnegative ellipticityat∼222nm,theregioncharacteristicforhelicalproteins. Thus,limiteddifferencesinstructurewereobservedbetweenthe twovariants.

3.2. Nosigniﬁcantdifferenceinreceptorbindingafﬁnities betweentheVal66andMet66proBDNFvariants

Inthecontextofthepresentstudy,itwasimportantto char-acterizeandcomparethebindingproﬁlesandkineticconstants oftheproBDNFvariantstotheirreceptorsusingsurfaceplasmon resonance(SPR).Thiswouldgivedirectevidencefortheir com-parativeabilitytobindthetyrosinekinasereceptorTrkB,p75NTR, sortilinandSorCS2.RepresentativeSPR sensorgramsare shown in Fig. 2(a–h) and complete on–off rates and equilibrium dis-sociationconstants (KD) are provided in Table1.Both variants boundwithsimilarafﬁnity(KD)asdeterminedbytwo-tailedt -testtotheextracellularimmunoglobulin-likebindingdomainof TrkB(TrkBIg2)(Val66:2.32±0.09nM,n=4;Met662.03±0.14nM, n=3;P=0.780;Fig.2aandb)andp75NTRextracellulardomain (ECD)(Val66:2.42±0.18nM,n=5; Met662.31±0.22nM,n=5;

P=0.502; Fig. 2c and d). Binding to sortilin luminal domain (LD) (Val66: 2.90±0.37nM, n=5; Met66: 3.83±0.84nM, n=6;

P=0.312;Fig.2eandf)andSorCS2LD(KDVal662.23±0.21,n=3; Met662.23±0.30nM,n=3;P=0.614;Fig.2gandh)alsodidnot differbetweenthevariants.Overall,thisissuggestiveoflittleor nodifferenceinthereceptorbindingafﬁnities,indicatingthatany differenceincellulareffectsthesevariantscauseislikelytobedue tothedifferingactionsoftheproteinsthemselves.

BDNFwasusedasacontrolineachcaseandtheresultsindicate (seeTable1)thatBDNFboundwithsub-nanomolarafﬁnitytoTrkB andp75NTRwhileitboundonlywithlowafﬁnitytosortilinand SorCS2.

3.3. Signiﬁcantdoserelateddecreaseinmitochondrialmembrane potentialbybothvariantsbutnosigniﬁcantchangeinLDHrelease

Previously,cellbasedassays haveshownproBDNF Met66to bindless well tointracellularsortilin as determinedby immu-noprecipitation[36]. We were therefore particularly interested toknowwhether, incells,therewouldbeadifferencebetween Val66andMet66variantswithregardstopro-neurotrophin medi-ated apoptosis, which occurs via binding to the extracellular p75NTR–sortilinreceptorcomplex[37].Itseemedpossiblethatthe age-relatedresistancetocognitivedecline,whichhasbeen associ-atedwiththeMet66variant,maybeduetoaninabilityofMet66to bindsortilinattheplasmamembrane.Thus,anLDHreleaseassay wasinitiallyusedtodeterminetherelativeeffectsoftheproBDNF variantsonthecelldeathprocessinserumdeprivedmousecortical neurons(DIV11),followingtreatmentfor18h.However,compared totheuntreatedserum-deprivedcells,treatmentwitheither vari-ant(0.05–10nM)didnotshowstatisticallysigniﬁcantcelldeath (Fig.1f).One-wayANOVAfollowed byDunnett’smultiple com-parisonposthoc testfor comparison withingroupshowedthat proBDNFtreatmentdidnotaffectthereleaseofLDH fromcells (Val66:P>0.05;Met66:P>0.05;Fig.1f).Two-wayANOVAfollowed byBonferroni’smultiplecomparisonposthoctestforcomparison betweengroupsshowedthatthiswasindependentofVal66Met substitution(P=0.473).

The effects on change in mitochondrial membrane poten-tial(␺m) were then assessed in serum starved mouse cortical neurons(DIV11)using theJC-1assay.TreatmentwithproBDNF (0.05–10nM)for18hshowedasigniﬁcantdoserelateddecreasein mitochondrialmembranepotentialforbothvariantsasdetermined by one-way ANOVA followed by Dunnett’s multiple compari-sonpost hoctest (Val66:P<0.0001; Met66:P<0.0001; Fig.1g). However,statisticalanalysisusingtwo-wayANOVAfollowedby Bonferroni’spost hoc test betweenthetwo groups showedthe Val66Metpolymorphismwasnotassociated(P=0.473)withthe reductioninmitochondrialmembranepotential.Consistentwith theearlierﬁndings,theseresultssuggestnomajordifferencesin thefundamentalactionsoftheproteinvariants.

3.4. InhibitionofLTPandfacilitationofLTDbyVal66proBDNF variantbutnotMet66variant

The absence of any substantial difference between the two proBDNFvariants,eitherstructurallyorfunctionally,was surpris-ing,especiallygiventhewidelyreportedcontrastintheireffects oncognitivefunction.SinceproBDNF isknowntofacilitateLTD

viap75NTR[21],we thereforesought toexamineifthe exoge-nousapplicationoftheproteinshadanyconsequencesonsynaptic plasticity,a molecularcorrelate oflearning and memory,in rat hippocampalslices.Applicationof1nMproBDNFVal66causeda signiﬁcantinhibitionofLTP(control:164.9±0.66%,n=12;Val66: 119.1±0.79%,n=6;P<0.001),whereasunexpectedly,nosuch inhi-bitionwasobservedwhentissuewasexposedtoproBDNFMet66 (162.7±1.44%,n=6;P=0.290;Fig.3a.Wealsoexaminedthe induc-tionofLTD,aspreviouslyinvestigated[21].Wedelivereda1Hz low-frequency stimulation protocol, knownto induce NMDAR-dependentLTDinjuvenile,butnotadult,animals[38].Consistent withthis, LTDwasnotinducedinuntreated controlslicesfrom P24–26rats(control:83.80_±0.86%,n=6),butwasfacilitatedin slices treated with proBDNF Val66 (Val66:70.11±0.93%, n=7;

(5)

Fig.1. CharacterizationofproBDNFVal66MetFCRpolymorphicvariants.(Forinterpretationofthereferencestocolorinthisﬁgurelegend,thereaderisreferredtotheweb versionofthisarticle.)

(6)

Fig.2.ProBDNFvariantsaresimilarintheirbindingtoTrkB,p75NTR,sortilinandSorCS2.

Representativesensorgramsareshownfor0–20nMproBDNFFCRVal66(a,c,e,andg)andFCRMet66(b,d,f,andh)bindingtotheirisolatedreceptorbindingdomainsTrkBIg2,

p75NTRECD,sortilinLDandSorCS2LD.Datawasdoublereferenced(RU)andplottedasafunctionoftime.Allthekineticdatawereﬁttedto1:1Langmuirmathematical modelandkineticconstantswerecalculated.KDdenotesmeanequilibriumdissociationconstant.MeanconstantsforbothVal66andMet66variantswerenotstatistically different.FullkineticconstantsaregiveninTable1.

reﬂectedtheactivation ofLTDsignaling.However, surprisingly, exposuretoproBDNF Met66failed tofacilitate LTDwhen com-paredtotheproBDNF Val66treated slices(95.81_±0.58%,n=7;

Fig.3b).Thedifferencesseenbetweenthevariantsandthe con-trolswerestatisticallysigniﬁcant(one-wayANOVAfollowedby Tukey’smultiplecomparisonpost-hoctest,proBDNFVal66 com-paredtoproBDNFMet66(P<0.001),proBDNFVal66comparedto controls(P<0.001)).

3.5. Val66proBDNFvariantspeciﬁcallyactivatesGSK3ˇ

(7)

Table1

SummaryofproBDNFreceptorbindingkinetics.

KineticconstantsofproBDNFbindingtoitsreceptorsTrkBIg2,p75NTRECDandsortilinLDwerecalculatedusing1:1Langmuirmathematicalmodel.Allthekineticconstants

aregivenasmean±s.e.m.andkineticconstantswerecomparedusingunpairedt-test(2-tailed).n:numberofexperiments,ka:associationrateconstant,kd:dissociation

rateconstantandKD:equilibriumdissociationconstant.Pdenotesvaluesfromunpaired,twotailedt-testbetweenKDofVal66andMet66.Nonewassigniﬁcant.Valuesfor

individualkaandkdwerealsonon-signiﬁcant.BindingforBDNFwastoolowafﬁnitytogiveaccuratemeasurementforSorCS2.

n (ka±s.e.m.)×106(M−1s−1) (kd±s.e.m.)×10−3(s−1) (KD±s.e.m.)×10−9(M) PofKDVal66vsMet66

TrkBIg2

Val66 4 1.61±0.31 3.67±0.52 2.32±0.09 0.780

Met66 3 1.99±0.47 3.92±0.67 2.03±0.14

BDNF 7 4.57±0.76 2.32±0.43 0.58±0.13

P75NTRECD

Val66 5 1.88±0.48 3.94±0.87 2.42±0.18 0.502

Met66 5 1.55±0.43 3.25±0.48 2.31±0.22

BDNF 4 4.44±0.68 2.51±0.12 0.62±0.12

SortilinLD

Val66 5 2.48±0.63 7.46±2.27 2.90±0.37 0.312

Met66 6 1.56±0.37 5.04±0.87 3.83±0.84

BDNF 7 3.80±1.37 21.7±9.01 6.44±1.61

SorCS2LD

Val66 3 1.79±0.23 4.00±0.84 2.23±0.21 0.614

Met66 3 2.135±0.45 4.53±0.36 2.23±0.30

BDNF 3 Lowaffinity Lowaffinity Lowaffinity

P<0.05; Fig.3c; SupplementaryFig.S2shows fulluncut blots). However,nosigniﬁcantincreaseinGSK3␤activationwasobserved whensliceswereexposedtoproBDNFMet66(74.49±19.36%,n=4;

P>0.05;Fig.3c).Theseeffectswereindependentofachangeintotal levelsofGSK3␤(control:100±13.13%;Met66:104.01±19.79%; Val66:100.65±24.41%;n=4;P>0.05;Fig.3c).Thissuggeststhat whilstproBDNFVal66canactivateGSK3_␤,theMet66variantdoes not,and maygosomewaytoexplain theirdifferingeffects on synapticplasticity.

Supplementrymaterialrelatedtothisarticlefound,intheonline version,athttp://dx.doi.org/10.1016/j.phrs.2015.12.008.

3.6. ProBDNFVal66variantinhibitsLTPviatheactivationof GSK3ˇ

Totest whetherproBDNF mightmediate itseffects on plas-ticity via GSK3␤, we exposed hippocampal slices to proBDNF after pre-incubation with the GSK3␤ inhibitor CT-99021, and testedwhetherLTPcouldbeinduced.Remarkably,CT-99021 pre-vented theproBDNF Val66-mediatedinhibition of LTP(control: 158.35±5.75%,n=6;proBDNF+CT:156.10±10.46%,n=6;Fig.3e). ThissuggeststhattheactionsofproBDNFVal66onsynaptic plas-ticityarelikelymediatedthroughtheactivationofGSK3␤.

3.7. TauphosphorylationinducedbyVal66proBDNFvariantnot apparentwithMet66variant

The presence of hyperphosphorylated tauprotein is a neu-ropathologicalhallmarkof ADpathology [40].Interestingly,the phosphorylationoftauhasrecentlybeenidentiﬁedasa critical stepinhippocampalLTDinduction[41].GiventhatGSK3␤isan importantregulatoroftauphosphorylation,wetherefore investi-gatedtheregulationofphosphorylationoftaubythetwoproBDNF variants.WefoundthattheproBDNFVal66variantsigniﬁcantly increased tau phosphorylation (control: 100±28.19%; Val66: 244.55±25.95%,P<0.01)whencomparedtocontrol(n=4;Fig.3d), whilsttheproBDNFMet66varianthadnoeffect(101.16±29.3%,

P>0.05whencomparedtocontrol,n=4).Together,thesedataare consistentwithaspeciﬁceffectoftheproBDNFVal66varianton synapticplasticity,whereGSK3␤isactivatedtocause dysregula-tionofsynapticplasticity.

4. Discussion

Previous studieshave identiﬁed divergent effects in healthy youngandolderproBDNFMet66carriersversustheVal66 homozy-gousindividuals.YoungadultMet66carriershadreducedcognitive functionandhippocampalvolumecomparedwiththosewithVal66

[2–5],whereastheMet66alleleconferredprotectionagainst cog-nitivedeclinewithagingandaftertraumaticbraininjury[6,7,9].

Inordertoestablishwhetherphysicaldifferencesbetweenthem couldaccountfortheircontrastingeffects incognitivefunction, thesepolymorphicproteinswereproducedascleavage-resistant formsandcompareddirectlyfortheﬁrsttimefordifferencesin structure, receptorbindingkineticsand function.By theoretical analysis,using PONDR,Met66 wasseento bemore disordered intheregionsurroundingthepolymorphism.Far-UVCDanalysis indicatedapredominanceof␤-sheetstructureforboth,although Met66showedaminordifferenceinaregioncharacteristicfor heli-calproteins.Neitherofthesefactorsaffectedtheresistancetofurin cleavage;bothformswerecleavage-resistant.

Wetheninvestigatedthepossibledifferencesinreceptor bind-ingusing SPR.Thesubtle changesseen inCD analysisbetween proBDNF Val66 and Met66 were not reﬂected in SPR binding assays,whichshowedsimilarbindingafﬁnitiesforboth polymor-phicvariantstotheirreceptorsTrkBandp75NTR,mostlikelyas aconsequenceofthosereceptorsbindingtothematurepeptide ofproBDNF.Thistoomayaccountforthelackofstatistical differ-enceinbindingbetweentheproBDNFpolymorphismsatTrkBand p75NTR.Pertinenttothis,apreviousmutagenesisstudyshowed theevolutionarilyconserved103-arginineresidueinthemature regionofBDNFtobeimportantforbothmatureandproBDNFto bindTrkB[42].Inaddition,recently,proBDNFbutnotitsisolated PROdomainwasshown toco-immunoprecipitatewithp75NTR

(8)

Fig.3.ProBDNFVal66activatesGSK3␤tomediateeffectsonsynapticplasticity.

(a)LTPwasinducedbydelivering2tetanicstimulitotheSchaffercollateralinputofacutehippocampusslices.Priortotetanus,sliceswerepre-incubated(90min)with1nM proBDNFVal66andMet66FCRvariantsandperfusedconstantlywithappropriate1nMvariants.StatisticalcomparisonofthemeanfEPSPsvalues,calculated∼45minafter applicationoftetanus,showedprominentdifferencesbetweenVal66andMet66ininhibitingLTP.(b)LTDwasinducedusingstandardLFSprotocolof1Hz,900pulsesfor 15mininrat(P24–26)hippocampalslices.PriortoLFS,sliceswerepre-incubated(90min)with1nMproBDNFVal66orMet66andperfusedconstantlywith1nMproBDNF variantsasappropriate.StatisticalcomparisonofmeanfEPSPsvaluescalculatedapproximately45minafterapplicationofLFSshowedprominentdifferencesbetweenVal66 andMet66infacilitatingLTD.(c)WesternblottinganalysisofphosphorylatedGSK3␤atserine9andtotallevelsofGSK3␤afterincubationwithproBDNFvariantsfor90min, n=4.(d)RepresentativewesternblotshowingPHF-1phosphorylationlevelsnormalizedtotau-5aftertreatmentwithproBDNFVal66orMet66(1nM;90min).Bar-chart showedquantificationofpooleddata(n=4).(e)QuantificationofLTPlevelsfollowing2tetanicstimuliwithproBDNFVal66inthepresenceandabsenceofCT99021(1␮M; n=6).Significance(*_P_<_0.05;**_P_<_0.01)_was_determined_using_one-way_ANOVA_with_Tukey’s_multiple_comparison_post-hoc_test,_comparing_{proBDNF-treated}_groups_to

non-treatedcontrolgroups.Throughout,controlexperimentswerewithoutexposuretoproBDNFanderrorbarsdenotes.e.m.

termsoffundamentalformandmodeofaction,andthereforeany observeddifferencesintermsofdownstreamsignalingeffectsof theproteinsislikelytoreﬂectspeciﬁcactionsofthevariantsatthe functionallevel.

Becausethefull-lengthsortilinreceptor,whichfacilitatesthe intracellulartrafﬁckingofproBDNF,isnotamenabletoSPR,itwas

(9)

intracellularsortilincomparedtoVal66duringbiosynthesisand secretion[23,36]althoughbindingafﬁnitieshavenotbeen deter-mined.Here,contrarytoourexpectations,wefoundnodifference inthebindingafﬁnitiesofVal66andMet66variantstosortilinusing SPRanalysis.Morerecently,SorCS2hasbeenshowntobea recep-torforproBDNF[12,44].SPRanalysisperformed,usingSorCS2in thisstudydidnotdetectanydifferenceinproBDNFVal66orMet66 binding.However,thismaybeduetotheuseofisolated recep-torbindingdomainsinSPR ratherthanthefull-lengthreceptor and/ortheroleofthepresenceofp75NTRco-receptorcomplexes inproBDNFbinding.

InanefforttoexaminethedifferentialeffectoftheproBDNF variantsonapoptosisinaneuronalcontextweusedserumdeprived corticalneuronsfollowingproBDNFtreatmentfor18h.We per-formedanassaytomeasuretheamountofLDHreleasedintothe culturemediumbydyingordamagedcellswithcompromisedcell membraneintegrity.Treatmentwitheitherpolymorphismupto 10nM didnot increase cell death. However, although previous ﬁndingshaveshownthatproBDNF-inducedapoptosisoccursvia

bindingtosortilinandp75NTR[37],andproBDNFVal66mediate celldeathhasbeenshowninculturedcerebralganglionicneurons

[17]andsympatheticneurons[16],thetriggeringofapoptosisdoes notseemtobeinevitableinallcells;othersfoundthatproBDNFin hippocampalneuronsdidnotinduceapoptosis[14].ProBDNFhas alsobeenshowntohavesomeprotectiveeffectsonneurons[42]

makingitmoredifﬁculttointerpretitseffectsinvitro.Therefore corticalneuronsweretestedforchangesinmitochondrial mem-branepotential,a more subtle formof change,after18h. Both variantsofproBDNF mediateda reductioninmembrane poten-tial.ThusadditionofproBDNFshowsthatithasanegativeeffect onmitochondrialmembranepotential,whichmayactasa precur-sortoreleaseofcytochromeCandcaspase3cleavage,anevent upstreamofthatassessedbytheLDHassay.However,bothVal66 andMet66variantshadsimilareffects,suggestingthatthiswasnot variantspeciﬁc.

ThepossibilitythattheMet66variantmaydifferentiallyaffect neuronalplasticitywasthen tested.Asexpected,using electro-physiology,theproBDNFVal66variantwasshowntobeableto facilitateLTD.Comparedtotheuntreatedcontrolslicesthe facili-tationwassigniﬁcantbutmodestat∼14%.Thiswasinaccordwith previousstudiesperformedinmousehippocampalslices,which useda murinefurincleavage-resistantproBDNFVal66[21] and inmiceoverexpressingproBDNFVal66[15].However,incontrast withthestudybyWooetal.[21]thedatahereshowthatproBDNF Val66wasalsoabletoinhibit tetanicstimuli-induced LTP.This apparentcontradictionmayperhapsbeexplainedbydifferences betweenthetwo experimentalapproaches,thestudiesbyWoo etal.,wereperformedwithmouseproBDNFVal66cleavage resis-tantproteincoupledtoahis-tag,whilstourstudieswereperformed withhumanproBDNFcleavageresistantproteinwithouttags,but withdifferingincubationtime and concentration.However,the comparableLTDdatabetweenthetwostudieswouldsuggestthat anydifferencesinproteinutilizedarespeciﬁctoLTPnotLTD sig-nalling.Themostlikelyreasonisprobablythereforetheuseofrat hippocampusinthisstudywhereasmousetissuewasusedinthe studybyWooetal.[21].

Veryrecently,studieshaveshownthat GSK3␤mediated tau phosphorylation has an important role in LTD [41,45] and is reportedtobedysregulatedincognitivedisordersincludingAD

[46]andanon-pathologicalmodel[47].Furthermore,ithas previ-ouslybeenshownthattheimpairmentofsynapticplasticityinan ADandPDpathologymodelwasmediatedbyGSK3␤[48,49].Given theactivationofGSK3␤bytheproBDNFVal66variantspeciﬁcally, asshownherefortheﬁrsttime,itisproposedthatproBDNF medi-atesitseffectsonplasticityviaGSK3␤,thusprovidingamechanistic linkbetweenproBDNFVal66andLTPinhibition.Together,thedata

presentedheresuggestthatproBDNFVal66inhibitsLTPthrough anenhancementofLTDsignalingviatheGSK3␤-pTaupathway.

In contrast, we showthat proBDNF Met66has noeffect on theGSK3␤-pTausignalpathway.Takentogetherthesedata pro-vide aconvincingpictureofa proBDNF Val66speciﬁceffecton hippocampalsynapticplasticity.Here,proBDNFVal66activatesa GSK3␤-pTausignalingpathwayleadingtosynapseweakeningthat isentirelyabsentinthepresenceofthecommonMet66variant.

This studyfurther addsto theﬁndings of Ninan et al. [24], whichshowedMet66knock-inmicetoexhibitimpairedLTD,by circumventionofthereductioninacuteactivitydependentrelease of BDNFseenin thesemice.Thustheexogenous applicationof humanproBDNFusedinthepresentstudyprovidesnewevidence fortheVal66Metpolymorphismtohavedirectfunctionalrelevance inage-relatedcognitiveimpairment.Thisisinconjunctionwiththe knowndecreaseinthecatalyticformofTrkBwithage[50]andAD

[51]comparedtop75NTR[52]whichmakeslikelyanincreasedrole ofLTDinsynapticplasticitywithagingandcognitiveimpairment. Thismayexplainabiastowardsprotectionagainstcognitivedeﬁcit inlaterlife,ofproBDNFMet66,withitsreducedabilitytofacilitate synapticweakening.Thishighlightstheimportanceof understand-ingtheactivationofsynapseweakeningsignalinginbothnormal andpathologicalagingconditions,andthepotentialimpactthis mayhaveinthedevelopmentofnoveltherapeuticinterventions.

5. Conclusion

In order to try to explain the age-related functional differ-encesseen incarriersof thecommonVal66Metpolymorphism, theseparatehumanrecombinantproteinswereproducedandthe biophysical,kinetic,cellularandsynapticdifferencesbetweenthe twocommonproBDNFpolymorphismswereassessed.Theoretical analysisofstructure,usingPONDR,andanalysisoffar-UVCD mea-surementshowedonlyminorstructuraldifferencesbetweenthe polymorphisms.Neithercouldthepolymorphismsbe differenti-atedintermsoftheirbindingkineticstoreceptorbindingdomains. Furthermore,incellassays,bothvariantscausedasimilarly signif-icantdosedependentreductioninmembranepotential,although notinreleaseofLDHincorticalneurons.However,differenceswere noted byelectrophysiology,in rat hippocampalslices.ProBDNF Val66wasabletofacilitateLTDandinhibitLTP,whereasMet66 wasnot.FurtheranalysisrevealedthatVal66wasabletomediate itseffectsonsynapticplasticityviaGSK3␤,thisprovidesthe pro-cessofassociationbetweenVal66andLTPinhibition.ThusproBDNF Val66mayinhibitLTPbyenhancingLTDsignalingviatheGSK3␤ -pTaupathway.Thesedataalsosuggestapossibleexplanationasto whyMet66mayconferprotectionagainstcognitivedeﬁcitinlater life,duetoitsreducedabilitytofacilitatesynapticweakening.

Authorcontributions

S.K.,S.J.A.,D.W.,T.M.P.,F.G-M.andK.C.designedanddirected differentaspectsofthestudyandwrotethepaper.M.F.andS.K. carriedoutproteinexpressionandpuriﬁcation,S.K.,T.M.P.,J.H.Y., D.W.,E.B.,S.C.,M.L.andL.O’N.carriedoutexperiments.FinalFigs. wereproducedmainlybyS.K.,T.M.P.andJ.H.Y.Allauthorsreviewed themanuscript.

Acknowledgments

S.J.A.isaSigmundGestetnerSeniorResearchFellow.

S.K.wasaUniversityofBristolPostgraduateResearchOverseas Scholar.

(10)

Gestet-ner Trust Fund (University of Bristol), the Alzheimer’s Society andtheAlumnioftheUniversityofBristol.K.C.andD.J.W.were supportedby UK Wellcome Trust-MRCNeurodegenerative Dis-easeInitiativeProgramme.J.H.Y.wassupportedbytheKorea-UK Alzheimer’sResearch ConsortiumProgrammeundertheKorean Ministryof Healthand Welfare.T.M.P. andS.C.weresupported by Bristol-Chonnam Frontier Programme under the Chonnam NationalUniversityHospital.K.C.wassupportedbytheWolfson ResearchMeritAwardandRoyalSociety,London.

References

[1]T.L.Petryshen,P.C.Sabeti,K.A.Aldinger,B.Fry,J.B.Fan,S.F.Schaffner,S.G. Waggoner,A.R.Tahl,P.Sklar,Populationgeneticstudyofthebrain-derived neurotrophicfactor(BDNF)gene,Mol.Psychiatry15(2010)810–815. [2]M.F.Egan,M.Kojima,J.H.Callicott,T.E.Goldberg,B.S.Kolachana,A.Bertolino,

E.Zaitsev,B.Gold,D.Goldman,M.Dean,B.Lu,D.R.Weinberger,TheBDNF val66metpolymorphismaffectsactivity-dependentsecretionofBDNFand humanmemoryandhippocampalfunction,Cell112(2003)257–269. [3]L.Pezawas,B.A.Verchinski,V.S.Mattay,J.H.Callicott,B.S.Kolachana,R.E.

Straub,M.F.Egan,A.Meyer-Lindenberg,D.R.Weinberger,Thebrain-derived neurotrophicfactorval66metpolymorphismandvariationinhumancortical morphology,J.Neurosci.24(2004)10099–10102.

[4]A.N.Voineskos,J.P.Lerch,D.Felsky,S.Shaikh,T.K.Rajji,D.Miranda,N.J. Lobaugh,B.H.Mulsant,B.G.Pollock,J.L.Kennedy,Thebrain-derived neurotrophicfactorVal66Metpolymorphismandpredictionofneuralriskfor Alzheimerdisease,Arch.Gen.Psychiatry68(2011)198–206.

[5]A.R.Hariri,T.E.Goldberg,V.S.Mattay,B.S.Kolachana,J.H.Callicott,M.F.Egan, D.R.Weinberger,Brain-derivedneurotrophicfactorval66metpolymorphism affectshumanmemory-relatedhippocampalactivityandpredictsmemory performance,J.Neurosci.23(2003)6690–6694.

[6]K.I.Erickson,J.S.Kim,B.L.Suever,M.W.Voss,B.M.Francis,A.F.Kramer, Geneticcontributionstoage-relateddeclineinexecutivefunction:a10-year longitudinalstudyofCOMTandBDNFpolymorphisms,Front.Hum.Neurosci. 2(11)(2008).

[7]P.D.Gajewski,J.G.Hengstler,K.Golka,M.Falkenstein,C.Beste,TheMet-allele oftheBDNFVal66Metpolymorphismenhancestaskswitchinginelderly, Neurobiol.Aging32(2327)(2011)2327–12319.

[8]F.Krueger,M.Pardini,E.D.Huey,V.Raymont,J.Solomon,R.H.Lipsky,C.A. Hodgkinson,D.Goldman,J.Grafman,TheroleoftheMet66brain-derived neurotrophicfactoralleleintherecoveryofexecutivefunctioningafter combat-relatedtraumaticbraininjury,J.Neurosci.31(2011)598–606. [9]A.K.Barbey,R.Colom,E.Paul,C.Forbes,F.Krueger,D.Goldman,J.Grafman,

Preservationofgeneralintelligencefollowingtraumaticbraininjury: contributionsoftheMet66brain-derivedneurotrophicfactor,PLoSOne9 (2014)e88733.

[10]D.J.Glass,S.H.Nye,P.Hantzopoulos,M.J.Macchi,S.P.Squinto,M.Goldfarb, G.D.Yancopoulos,TrkBmediatesBDNF/NT-3-dependentsurvivaland proliferationinﬁbroblastslackingthelowafﬁnityNGFreceptor,Cell66 (1991)405–413.

[11]S.P.Squinto,T.N.Stitt,T.H.Aldrich,S.Davis,S.M.Bianco,C.Radziejewski,D.J. Glass,P.Masiakowski,M.E.Furth,D.M.Valenzuela,etal.,TrkBencodesa functionalreceptorforbrain-derivedneurotrophicfactorandneurotrophin-3 butnotnervegrowthfactor,Cell65(1991)885–893.

[12]A.Anastasia,K.Deinhardt,M.V.Chao,N.E.Will,K.Irmady,F.S.Lee,B.L. Hempstead,C.Bracken,Val66MetpolymorphismofBDNFaltersprodomain structuretoinduceneuronalgrowthconeretraction,Nat.Commun.4(2490) (2013).

[13]Y.Sun,Y.Lim,F.Li,S.Liu,J.J.Lu,R.Haberberger,J.H.Zhong,X.F.Zhou, ProBDNFcollapsesneuriteoutgrowthofprimaryneuronsbyactivatingRhoA, PLoSOne7(2012)e35883.

[14]H.Koshimizu,K.Kiyosue,T.Hara,S.Hazama,S.Suzuki,K.Uegaki,G. Nagappan,E.Zaitsev,T.Hirokawa,Y.Tatsu,A.Ogura,B.Lu,M.Kojima, MultiplefunctionsofprecursorBDNFtoCNSneurons:negativeregulationof neuritegrowth,spineformationandcellsurvival,Mol.Brain2(2009)27. [15]J.Yang,L.C.Harte-Hargrove,C.J.Siao,T.Marinic,R.Clarke,Q.Ma,D.Jing,J.J.

Lafrancois,K.G.Bath,W.Mark,D.Ballon,F.S.Lee,H.E.Scharfman,B.L. Hempstead,proBDNFnegativelyregulatesneuronalremodeling,synaptic transmission,andsynapticpasticityinhippocampus,CellRep.(2014). [16]H.K.Teng,K.K.Teng,R.Lee,S.Wright,S.Tevar,R.D.Almeida,P.Kermani,R.

Torkin,Z.Y.Chen,F.S.Lee,R.T.Kraemer,A.Nykjaer,B.L.Hempstead,ProBDNF inducesneuronalapoptosisviaactivationofareceptorcomplexofp75NTR andsortilin,J.Neurosci.25(2005)5455–5463.

[17]H.Koshimizu,S.Hazama,T.Hara,A.Ogura,M.Kojima,Distinctsignaling pathwaysofprecursorBDNFandmatureBDNFinculturedcerebellargranule neurons,Neurosci.Lett.473(2010)229–232.

[18]Y.J.Fan,L.L.Wu,H.Y.Li,Y.J.Wang,X.F.Zhou,Differentialeffectsofpro-BDNF onsensoryneuronsaftersciaticnervetransectioninneonatalrats,Eur.J. Neurosci.27(2008)2380–2390.

[19]M.Volosin,W.Song,R.D.Almeida,D.R.Kaplan,B.L.Hempstead,W.J.Friedman, Interactionofsurvivalanddeathsignalinginbasalforebrainneurons:rolesof neurotrophinsandproneurotrophins,J.Neurosci.26(2006)7756–7766.

[20]P.T.Pang,B.Lu,Regulationoflate-phaseLTPandlong-termmemoryin normalandaginghippocampus:roleofsecretedproteinstPAandBDNF, AgeingRes.Rev.3(2004)407–430.

[21]N.H.Woo,H.K.Teng,C.J.Siao,C.Chiaruttini,P.T.Pang,T.A.Milner,B.L. Hempstead,B.Lu,Activationofp75NTRbyproBDNFfacilitateshippocampal long-termdepression,Nat.Neurosci.8(2005)1069–1077.

[22]Z.Y.Chen,D.Jing,K.G.Bath,A.Ieraci,T.Khan,C.J.Siao,D.G.Herrera,M.Toth,C. Yang,B.S.McEwen,B.L.Hempstead,F.S.Lee,GeneticvariantBDNF(Val66Met) polymorphismaltersanxiety-relatedbehavior,Science314(2006)140–143. [23]Z.Y.Chen,P.D.Patel,G.Sant,C.X.Meng,K.K.Teng,B.L.Hempstead,F.S.Lee,

Variantbrain-derivedneurotrophicfactor(BDNF)(Met66)altersthe intracellulartrafﬁckingandactivity-dependentsecretionofwild-typeBDNF inneurosecretorycellsandcorticalneurons,J.Neurosci.24(2004) 4401–4411.

[24]I.Ninan,K.G.Bath,K.Dagar,R.Perez-Castro,M.R.Plummer,F.S.Lee,M.V. Chao,TheBDNFVal66MetpolymorphismimpairsNMDAreceptor-dependent synapticplasticityinthehippocampus,J.Neurosci.30(2010)8866–8870. [25]K.G.Bath,D.Q.Jing,I.Dincheva,C.C.Neeb,S.S.Pattwell,M.V.Chao,F.S.Lee,I.

Ninan,BDNFVal66Metimpairsﬂuoxetine-inducedenhancementofadult hippocampusplasticity,Neuropsychopharmacology37(2012)1297–1304. [26]O.Clewes,M.S.Fahey,S.J.Tyler,J.J.Watson,H.Seok,C.Catania,K.Cho,D.

Dawbarn,S.J.Allen,HumanProNGF:biologicaleffectsandbindingproﬁlesat TrkA,P75NTRandsortilin,J.Neurochem.107(2008)1124–1135.

[27]R.L.Naylor,A.G.Robertson,S.J.Allen,R.B.Sessions,A.R.Clarke,G.G.Mason,J.J. Burston,S.J.Tyler,G.K.Wilcock,D.Dawbarn,Adiscretedomainofthehuman TrkBreceptordeﬁnesthebindingsitesforBDNFandNT-4,Biochem.Biophys. Res.Commun.291(2002)501–507.

[28]R.C.Malenka,Synapticplasticityinthehippocampus:LTPandLTD,Cell78 (1994)535–538.

[29]N.G.Seidah,S.Benjannet,S.Pareek,M.Chretien,R.A.Murphy,Cellular processingoftheneurotrophinprecursorsofNT3andBDNFbythe mammalianproproteinconvertases,FEBSLett.379(1996)247–250. [30]G.H.Doherty,D.Beccano-Kelly,S.D.Yan,F.J.Gunn-Moore,J.Harvey,Leptin

preventshippocampalsynapticdisruptionandneuronalcelldeathinduced byamyloidbeta,Neurobiol.Aging34(2013)226–237.

[31]J.J.Hwang,M.H.Park,S.Y.Choi,J.Y.Koh,ActivationoftheTrksignaling pathwaybyextracellularzinc.Roleofmetalloproteinases,J.Biol.Chem.280 (2005)11995–12001.

[32]R.Lee,P.Kermani,K.K.Teng,B.L.Hempstead,Regulationofcellsurvivalby secretedproneurotrophins,Science294(2001)1945–1948.

[33]X.Li,P.Romero,M.Rani,A.K.Dunker,Z.Obradovic,Predictingprotein disorderforN-,C-,andinternalregions,GenomeInf.Ser.WorkshopGenome Inf.10(1999)30–40.

[34]P.Romero,Z.Obradovic,X.Li,E.C.Garner,C.J.Brown,A.K.Dunker,Sequence complexityofdisorderedprotein,Proteins42(2001)38–48.

[35]Romero,Obradovic,K.Dunker,Sequencedataanalysisforlongdisordered regionspredictioninthecalcineurinfamily,GenomeInf.Ser.Workshop GenomeInf.8(1997)110–124.

[36]Z.Y.Chen,A.Ieraci,H.Teng,H.Dall,C.X.Meng,D.G.Herrera,A.Nykjaer,B.L. Hempstead,F.S.Lee,Sortilincontrolsintracellularsortingofbrain-derived neurotrophicfactortotheregulatedsecretorypathway,J.Neurosci.25(2005) 6156–6166.

[37]A.Nykjaer,R.Lee,K.K.Teng,P.Jansen,P.Madsen,M.S.Nielsen,C.Jacobsen,M. Kliemannel,E.Schwarz,T.E.Willnow,B.L.Hempstead,C.M.Petersen,Sortilin isessentialforproNGF-inducedneuronalcelldeath,Nature427(2004) 843–848.

[38]P.V.Massey,Z.I.Bashir,Long-termdepression:multipleformsand implicationsforbrainfunction,TrendsNeurosci.30(2007)176–184. [39]S.Peineau,C.Taghibiglou,C.Bradley,T.P.Wong,L.Liu,J.Lu,E.Lo,D.Wu,E.

Saule,T.Bouschet,P.Matthews,J.T.Isaac,Z.A.Bortolotto,Y.T.Wang,G.L. Collingridge,LTPinhibitsLTDinthehippocampusviaregulationofGSK3beta, Neuron53(2007)703–717.

[40]K.Iqbal,F.Liu,C.X.Gong,I.Grundke-Iqbal,TauinAlzheimerdiseaseand relatedtauopathies,Curr.AlzheimerRes.7(2010)656–664.

[41]P.Regan,T.Piers,J.H.Yi,D.H.Kim,S.Huh,S.J.Park,J.H.Ryu,D.J.Whitcomb,K. Cho,Tauphosphorylationatserine396residueisrequiredforhippocampal LTD,J.Neurosci.35(2015)4804–4812.

[42]B.Fayard,S.Loefﬂer,J.Weis,E.Vogelin,A.Kruttgen,Thesecreted brain-derivedneurotrophicfactorprecursorpro-BDNFbindstoTrkBand p75NTRbutnottoTrkAorTrkC,J.Neurosci.Res.80(2005)18–28. [43]E.M.Quistgaard,P.Madsen,M.K.Groftehauge,P.Nissen,C.M.Petersen,S.S.

Thirup,Ligandsbindtosortilininthetunnelofaten-bladedbeta-propeller domain,Nat.Struct.Mol.Biol.16(2009)96–98.

[44]S.Glerup,D.Olsen,C.B.Vaegter,C.Gustafsen,S.S.Sjoegaard,G.Hermey,M. Kjolby,S.Molgaard,M.Ulrichsen,S.Boggild,S.Skeldal,A.N.Fjorback,J.R. Nyengaard,J.Jacobsen,D.Bender,C.R.Bjarkam,E.S.Sorensen,E.M. Fuchtbauer,G.Eichele,P.Madsen,T.E.Willnow,C.M.Petersen,A.Nykjaer, SorCS2regulatesdopaminergicwiringandisprocessedintoanapoptotic two-chainreceptorinperipheralglia,Neuron82(2014)1074–1087. [45]T.Kimura,D.J.Whitcomb,J.Jo,P.Regan,T.Piers,S.Heo,C.Brown,T.

Hashikawa,M.Murayama,H.Seok,I.Sotiropoulos,E.Kim,G.L.Collingridge,A. Takashima,K.Cho,Microtubule-associatedproteintauisessentialfor long-termdepressioninthehippocampus,Philos.Trans.R.Soc.Lond.BBiol. Sci.369(2014)20130144.

(11)

[47]C.L.Busceti,P.D.Pietro,B.Riozzi,A.Traﬁcante,F.Biagioni,R.Nistico,F.Fornai, G.Battaglia,F.Nicoletti,V.Bruno,5-HT2Cserotoninreceptorblockade preventstauproteinhyperphosphorylationandcorrectsthedefectin hippocampalsynapticplasticitycausedbyacombinationofenvironmental stressorsinmice,Pharmacol.Res.99(2015)258–268.

[48]J.Jo,D.J.Whitcomb,K.M.Olsen,T.L.Kerrigan,S.C.Lo,G.Bru-Mercier,B. Dickinson,S.Scullion,M.Sheng,G.Collingridge,K.Cho,A␤(1–42)inhibition ofLTPismediatedbyasignalingpathwayinvolvingcaspase-3,Akt1and GSK-3␤,Nat.Neurosci.14(2011)545–547.

[49]M.Golpich,E.Amini,F.Hemmati,N.M.Ibrahim,B.Rahmani,Z.Mohamed,A.A. Raymond,L.Dargahi,R.Ghasemi,A.Ahmadiani,Glycogensynthasekinase-3 beta(GSK-3␤)signaling:implicationsforParkinson’sdisease,Pharmacol.Res. 97(2015)16–26.

[50]M.J.Webster,M.M.Herman,J.E.Kleinman,C.ShannonWeickert,BDNFand trkBmRNAexpressioninthehippocampusandtemporalcortexduringthe humanlifespan,GeneExpr.Patterns6(2006)941–951.

[51]S.J.Allen,G.K.Wilcock,D.Dawbarn,Profoundandselectivelossofcatalytic TrkBimmunoreactivityinAlzheimer’sdisease,Biochem.Biophys.Res. Commun.264(1999)648–651.