InstituteofBiologicalSciencesDepartmentofCellularBiology

TemporalProteomicAnalysisofRatHippocampusProvidesEvidenceforMemoryReconsolidationin

OperantConditioning

AnáliseProteômicaTemporaldeHipocampodeRatosForneceEvidênciasparaaReconsolidaçãodeMemórias

noCondicionamentoOperante

ArthurHenriquesPontes

Mentor:Ph.D.MarceloValledeSousaLaboratoryofProteinChemistryandBiochemistry(LBQP)

BrasiliaJune2018

2

3

TableofContents

Abbreviations.......................................................................................................................................................................................................................................4

Summary......................................................................................................................................................................................................................................................6

SummaryinPortuguese.......................................................................................................................................................................................................7

ExtendedsummaryoftheworkinPortuguese...............................................................................................................................8

Introduction.........................................................................................................................................................................................................................................11

a.Themolecularbasesofmemory....................................................................................................................................................11 b.Massspectrometry-basedproteomics…................................................................................................................................15

MotivationsforThisWork.............................................................................................................................................................................................20

Results..........................................................................................................................................................................................................................................................21

a. ParametersoftheOperantConditioningParadigm............................................................................................21b. SAXFractionationEnablesIn-depthProteomicsoftheHippocampus..........................................22c. QuantitativeProteomicsoftheHippocampusDuringMemoryFormation...........................26d. ProteinsDisplayingAbundanceChangesasEarlyas30minAfterOC...........................................27e. OC-BasedAssociativeMemoriesPossessParticularMolecularSignatures............................27f. IncreaseofDeNovoProteinSynthesisAfteraRecallSession………........................................................30g. QuantitativeHippocampalPhosphoproteome........................................................................................................33h. RT-qPCRMeasurementsofmRNAs………………………………………………………………........................................................................34

Discussion................................................................................................................................................................................................................................................36

MaterialsandMethods....................................................................................................................................................................................................44

a.EthicalCommittee............................................................................................................................................................................................44 b.AnimalsandBehavioralProtocol..................................................................................................................................................44 c.TissueDissection.................................................................................................................................................................................................44 d.Filter-aidedSampleDigestion..............................................................................................................................................................45 e.LabelingtheSampleswithMultiplexReagents..............................................................................................................46 f.OfflineAnionExchangeBasedFractionationwithSaltElution.....................................................................46 g.OfflineCationExchangeBasedFractionationwithSaltElution..................................................................47 h.EnrichmentofPhosphopeptidesUsingTiO2Beads...................................................................................................47 i.DesaltinginC18StageTips............................................................................................................................................................................48 j.LiquidChromatographicandMassSpectrometry.........................................................................................................48 k.ComputationalandStatisticalDataAnalysis......................................................................................................................49 l.RealTimePCRMeasurements..............................................................................................................................................................50

References…...........................................................................................................................................................................................................................................51

Acknowledgments…..................................................................................................................................................................................................................63

4

AbbreviationsACN–AcetonitrileAcOH–AceticacidAMPA–α-amino-3-hydroxy-5-methyl-4-isoxazolepropionicacidANI–AnisomycinCaMKII–Calcium/calmodulin-dependentproteinkinaseIIcAMP–CyclicadenosinemonophosphateCREB1–cAMPresponseelementbindingprotein1DG–DentategyrusDTT–DithiothreitolEC–EntorhinalcortexE-LTP–Early-phaseLTPFC–FearconditioningFDR–FalseDiscoveryRateIAA–IodoacetamideiCAT–Isotope-codedaffinitytagsICR–IoncyclotronresonanceIT–IontrapiTRAQ–IsobarictagsforrelativeandabsolutequantificationLC–LiquidchromatographyLC-MS/MS–Liquidchromatography-tandemmassspectrometryL-LTP–Late-phaseLTPLTP–Longtermpotentiation

MAPK–MitogenactivatedproteinkinaseMS–Massspectrometrym/z–MasstochargeratioNH4AcO–AmmoniumacetateNH4OH–AmmoniumhydroxideNMDA–N-methyl-D-aspartateOC–Operantconditioning

OT–Orbitrap PBS–PhosphatebufferedsalinePCR–PolymerasechainreactionPKA–ProteinkinaseAPSD–PostsynapticdensityPTM–PosttranslationalmodificationSAX–StronganionicexchangeSC–Subicularcomplex

5

SCX–StrongcationicexchangeSDS–SodiumdodecylsulfateSMT–SpatialmemorytasksTEAB–TriethylammoniumbicarbonateTFA–TrifluoroaceticacidTMT–TandemmasstagsTOF–Time-of-flightUPS–Ubiquitin-proteasomesystem

XIC–Extractedionchromatogram

6

Summary

Animals are able to store newly acquired information about the external world as

memories.Memoryformationoccursviarearrangementsofneuralcircuitries,whichareelicited

by changes in transcription, translation and post-translationmodifications (PTMs) in cells of

specificregionsofthecentralnervoussystemsuchasthehippocampus.Notably,themolecular

characterizationofmemoryformationhasbeencarriedoutprimarilyinthecontextofanimals

thathavebeensubjectedtofearorspatiallearningparadigms.Inthisstudy,weexaminedthe

molecular changes associated with information storage in rodents subjected to operant

conditioning(OC).Herein,weemployedstronganionicexchange(SAX)withsaltgradientelution

asafractionationstrategyfollowedbyliquidchromatography-tandemmassspectrometry(LC-

MS/MS)tomeasurechangesinhippocampalproteomeandphosphoproteomeatearlyandlate

stages ofmemory formation, aswell as after behavior recall.We identified a total of 8,951

proteinsand568phosphoproteins,makingthisstudythelargesthippocampalproteometodate.

Statisticallysignificantabundancechangeswereshownin465proteinsand64phosphoproteins

throughout the aforementioned time intervals. Furthermore, quantitative polymerase chain

reactionmeasurementsofmRNAabundancelevelsrevealedaweakinterdependencebetween

proteinandtranscriptlevels,givingcredencetothenotionofalowcorrelationbetweenproteins

andmRNAsindisturbedcellularstates.Also,theidentificationofdifferentiallyregulatedproteins

of the ubiquitin-proteasome system (UPS), as well as calcium/calmodulin-dependent protein

kinaseII(CaMKII),providesevidencefortheexistenceofatimewindowafterbehavioralrecall

in which stored information may become liable to further changes known as memory

reconsolidation.

7

SummaryinPortuguese

Osanimaissãocapazesdeguardarinformaçõesadquiridasdomundoexterioremforma

dememórias.Aformaçãodememóriasocorreatravésdeumrearranjodecircuitosneurais,que

éprovocadoporalteraçõesnatranscrição,traduçãoeadiçãodegruposquímicosemproteínas

emformademudançaspós-traducionais (PTMs)emcélulasderegiõesespecíficasdosistema

nervosocentralcomoohipocampo.Notavelmente,acaracterizaçãomoleculardaformaçãode

memórias tem sido realizada primariamente em animais submetidos a paradigmas

comportamentaisrelacionadosamemóriaespacialoudemedo.Nesteestudo,nósexaminamos

as mudanças moleculares associadas com o armazenamento de informações em animais

submetidos ao condicionamentooperante (OC). Aqui, empregamos a cromatografia de troca

aniônica (SAX)offline comeluiçãoatravésdeumgradiente crescentede sal seguidodeuma

cromatografia líquida acoplada a espectrometria de massas em tandem (LC-MS/MS) para

mensurarmudançasnoproteomaefosfoproteomahipocampaisemestágiosprecoceetardioda

formaçãodememórias,assimcomodepoisdaevocaçãodocomportamento.Identificamosum

totalde8.951proteínase568fosfoproteínas.Mudançasestatisticamentesignificativasforam

detectadasem456proteínase53fosfoproteínasaolongodosintervalosdetempomencionados

anteriormente. Ademais, mensurações de abundância de mRNA por reação em cadeia de

polimeraseemtemporealrevelouumafracainterdependênciaentreosníveisdetranscritose

proteínas,dandosuporteanoçãodeumabaixacorrelaçãoentreproteínasemRNAsemestados

celularesperturbados.Alémdisso,a identificaçãodeproteínasdiferencialmentereguladasdo

sistema ubiquitina-proteassoma (UPS), assim como calcium/calmodulin-dependent protein

kinase II (CaMKII), fornece evidência para a existência de uma janela de tempo depois da

evocação do comportamento na qual informações armazenadas se tornam sensíveis a

modificaçõesconhecidocomoreconsolidação.

8

ExtendedsummaryoftheworkinPortuguese

O aprendizado e a memória são processos fundamentais na vida de um animal. O

aprendizado refere-se à capacidade de adquirir novas informações sobre omundo exterior,

enquantoamemóriaéomecanismopeloqualessesdadossãocodificadosearmazenadospara

umaevocaçãoposterior.Aformaçãodenovasmemóriasdependedorearranjodeviascelulares

específicas em diferentes regiões do sistema nervoso central como o hipocampo. Esses

rearranjos incluem alterações morfológicas que são induzidas pela transcrição de genes

específicos, a tradução de distintas sequências de mRNA e a regulação de proteínas por

modificaçõespós-traducionais(PTMs).Asmoléculaseasregiõesdohipocampoquemedeiam

essas mudanças podem diferir dado o tipo de informação que está sendo codificada. Por

exemplo, animais submetidos a paradigmas comportamentais referentes amemória espacial

sofremmodificaçõesnogirodenteado(DG),enquantoorganismossujeitosaocondicionamento

contextual de medo (FC) possuem alterações nos neurônios da região CA1 do hipocampo.

Entretanto, pouco se sabe sobre asmudanças referentes a formaçãodenovasmemórias no

condicionamentooperante(OC)–umparadigmacomportamentalnoqualanimaisaprendema

associarumcomportamentocomasuaconsequência(e.g.roedorespressionandoumaalavanca

paraliberaráguaoucomidaemumacaixadeSkinner).

Esseprocessodeformaçãodememóriaspodelevardias.Duranteessetempo,memórias

instáveissetornamprogressivamenteestáveisaolongodotempoporumprocessoconhecido

comoconsolidação.Estudostemmostradoqueopotencialdelongaduração(LTP),quepodeser

definidocomoumfortalecimentodaconexãosinápticaentreneurônios,éoprincipalmecanismo

celulardoprocessodeconsolidaçãoemFCeparadigmascomportamentaisreferentesamemória

espacial.Opotencialdelongaduraçãotemsidodivididoemdoisestágios:precoceetardio.Em

ummodelo comumenteutilizado, a faseprecocedo LTP (E-LTP) é guiadaprimariamentepor

eventosdetransduçãodesinais.Duranteasprimeirashorasqueseguemaaquisiçãodeuma

memória, íons de Ca2+ entram no neurônio pós-sináptico e ativam diversas cinases como a

calcium/calmodulin-dependentproteinkinase II (CaMKII).CaMKIIativada fosforila resíduosde

receptores localizados na membrana pós-sináptica, aumentando a condutância deles. Essa

ativaçãodaCaMKIItambémlevaaotráficodereceptoresdotipoα-amino-3-hydroxy-5-methyl-

9

4-isoxazolepropionic acid (AMPA) para as densidades pós-sinápticas (PSDs), uma estrutura

especializada localizada na membrana dos neurônios pós-sinápticos. De 3 a 6 h depois da

aquisiçãodamemória,afasetardiadoLTP(L-LTP)inicia-se,desencadeandosínteseproteicae

grandes alterações do proteomado hipocampo. Por exemplo, aprotein kinaseA (PKA) ativa

fatores de transcrição incluindo cyclic AMP-responsive element-binding protein (CREB). CREB

ativadoaumentaatranscriçãodegenesalvoquemedeiammudançasestruturaisnosneurônios

pós-sinápticos.

Interessantemente, estudos iniciais mostraram que a administração de inibidores de

sínteseproteicacomoanisomycin(ANI)emcamundongosantesouimediatamenteapósserem

submetidosaparadigmascomportamentaisresultaramemumdéficitnaformaçãodememórias.

Entretanto,administraçãodeANI1hdepoisdoterminodasessãodetreinamentonãoafetoua

consolidaçãodememórias.Essesresultadosindicamquedeveexistirumajanelaprematurapara

asíntesedeproteínas–issoé,asíntesedenovodeproteínasduranteosestágiosiniciaisdoE-

LTP é necessário para a consolidação de memórias. Ademais, memórias consolidadas ainda

podem sofrer mudanças quando o comportamento é evocado novamente. Esse processo,

conhecidocomoreconsolidaçãodamemória,induzatraduçãodenovasproteínasquecausam

modificaçõesadicionaisnaarquiteturadosneurôniosdohipocampo.Acredita-sequeoprocesso

de reconsolidação da memória é uma adaptação dos organismos para fornecer novas

informaçõesaoconhecimentoquejáfoiadquiridoanteriormente,aumentandoaschancesdos

animaisdesobreviveremummundoemconstantemudança.Oprocessodereconsolidaçãoda

memória foi demonstrado em uma série de tarefas comportamentais,mas evidências desse

processoemratossubmetidosaoOCaindaestãoemdiscussão.

Comointuitodeentenderoprocessodeconsolidaçãodememóriasemratossubmetidos

aoOC,empregamos in-depth cromatografia líquidaacopladaaespectrometriademassasem

tandem(LC-MS/MS)paramedirmudançasemproteínasefosfoproteínashipocampaisa30min

e 12 h depois de submeter animais ao condicionamento operante. Ademais, sondamos o

proteomaefosfoproteomadessesanimaisdepoisdaevocaçãodessecomportamentoumdia

depois. Identificamos um total de 8.951 proteínas e 568 fosfoproteínas, das quais 465 e 64

mostraram mudanças estatisticamente significativas, respectivamente. Aqui, revelamos a

10

identidadede proteínas que exibiram variações significativas tão cedoquanto 1 h depois do

treinamento e demonstramos que as variações de abundância nas proteínas não foram

acompanhadas por uma mudança nos níveis de seus transcritos. Ademais, fornecemos

evidênciasmoleculares para a existência demecanismosde reconsolidação, indicandoqueo

armazenamentodeumcomportamentoOCprovavelmentetambémsofremodificaçõesafimde

atualizarmemóriasadquiridaspreviamente.

11

Introduction

a.TheMolecularBasesofMemory

Memory is one of the most important components of behavior and it refers to the

capacitytostoreandretrievenewinformationaboutthephysicalworld(Kandeletal.,2014).

Memoryhasemergedveryearly intheevolutionaryhistoryofthenervoussystem,andithas

playeda central role inhelpingorganisms to adapt and survive the challenges faced in their

environments(Emesetal.,2008;T.J.RyanandGrant,2009).Thiscognitiveprocesshasreached

oneofitsmostcomplexformsinhumans,wheredeficitscanhavedevastatingconsequencesfor

theindividual.InAlzheimerdisease,forinstance,thebuildupofmisfoldedproteinsinthebrain

disrupts theconnectivitybetweennervecells, leading tomemory impairmentsanddementia

(Musunuri et al., 2014; C. A. Ross and Poirier, 2004). In Huntington’s disease, likewise, the

expansionofaCAGtriplet in thehuntingtingene leads tomotorabnormalitiesandcognitive

declinesuchasmemoryloss(PontesanddeSousa,2016).Asaresultofthat,memoryhasbeen

thefocusofintensiveinvestigationindifferentresearchfieldssuchaspsychology,anthropology

andneuroscience (Rempel-Cloweretal.,1996;TronsonandTaylor,2007;Zola-Morganetal.,

1986).

Thefirstscientificstudiesonmemorywereperformedinpatientswithlesionsinspecific

brainregions.OneofthemostfamouscasesisofHenryMolaison,alsoknownaspatientH.M.

Sufferingfromconstantepilepticepisodes,Molaisonhadtwothirdsofhismedialtemporallobes

removed inanattempttocontrolhisseizures.Althoughthesurgerywasasuccess,H.M.was

incapableofformingnewmemories(Moseretal.,2015).Moreover,modelorganismswerealso

usedintheseinitialstudies.Typically,theseanimalsweresubjectedtobehavioralparadigmssuch

as fearconditioning (FC)–a formofassociative learning inwhichorganismspairanaversive

stimulus suchasa foot shockwithaneutral context suchasaboxora sound– followedby

administrationofdrugssuchasproteinsynthesisinhibitorsorlesionsinspecificregionsofthe

centralnervoussysteminanattempttoidentifybrainareasrelatedtomemory(Strekalovaetal.,

2003).Notably,theseassaysrevealedthatthehippocampuspossessedanimportantroleinthe

storageofnewlyacquiredinformation(PontesanddeSousa,2016).

12

Thehippocampus is a structure locatedbilaterally in themedial temporal lobeof the

vertebratebrain(BlissandCollingridge,1993).Inrodents,itcanbeanatomicallydividedintoa

fewdistinctregions:thedentategyrus(DG),theentorhinalcortex(EC),thesubicularcomplex

(SC)andthehippocampusproper,whichiscomposedoftheCA1,CA2,CA3andCA4areas(Figure

1; Strange et al., 2014). Each of these sections harbors different populations of cells that

communicate through two main pathways, namely the direct perforant pathway and the

trisynapticcircuit,alsoknownastheindirectperforantpathway(PontesanddeSousa,2016).

TheformertransmitsmultimodalsensoryandspatialinformationdirectlyfromtheECtotheCA1

area,amajoroutputofthehippocampus.Bycontrast,inthetrisynapticcircuit,informationis

sentfromtheECtotheCA1areafollowingthisroute:EC®DG®CA3®CA1(Figure1;Neves

etal.,2008;PontesanddeSousa,2016). Interestingly, thesubmissionofanimalstodifferent

types of behavioral paradigms elicits morphological changes in distinct regions of the

hippocampus.For instance, spatialmemory tasks (SMTs)generally inducechanges in theDG,

whereasFCelicitsalterationsintheCA1neurons(Monopolietal.,2011;Strekalovaetal.,2003).

FIGURE1IThetwomainpathwaystotheCA1areaofthehippocampusontheleft,andearly-phaseNMDAdependent-LTPontheright.Theredarrowsinthepictureontheleftshowthetrisynapticcircuitof the hippocampus, wheremultimodal sensory and spatial information coming from the entorhinalcortex(EC)isrelayedtotheCA1areafollowingthisroute:EC–DG–CA3–CA1.Inblue,weillustratedthe

Pontes and de Sousa Unveiling Memory through Mass Spectrometry

FIGURE 1 | The two main pathways to the CA1 area of the hippocampus on the left, and early-phase NMDA dependent-LTP on the right. The red arrows

in the picture on the left show the trisynaptic circuit of the hippocampus, where multimodal sensory and spatial information coming from the entorhinal cortex (EC) is

relayed to the CA1 area following this route: EC–DG–CA3–CA1. In blue, we illustrated the direct perforant pathway, which directly connects the EC to the CA1 region.

On the picture in right, we show an illustration of the early-phase LTP. Here, (1) glutamate from the presynaptic neuron is released into the synaptic cleft. (2) This

neurotransmitter reaches ionic channels of the postsynaptic cell causing depolarization of this neuron by the influx on sodium and calcium cations. (3) Calcium, in its

turn, activates CaMKII that (4) phosphorylates ionic channels in the PSDs and (5, 6) induces the addition of AMPA receptors to the postsynaptic membrane, increasing

synaptic efficiency.

the main excitatory neurotransmitter in the brain, diffusesthrough the cleft and reaches the post-synaptic neuron, whereit binds to ligand-gated ion channels. The interaction betweenthe neurotransmitter and the ionotropic NMDA receptor resultsin the influx of Ca2+ and Na+ into the cell (Schiller et al.,1998; Dingledine et al., 1999; Castillo, 2012). The Na+ helpsto bring about a depolarization of the postsynaptic neuronthat last a few milliseconds, while the Ca2+ promotes theactivation of protein kinases such as calcium/calmodulin-dependent protein kinase (CaMKII) (Lisman et al., 2012;Lüscher and Malenka, 2012). CaMKII and other kinases promptthe introduction of other ionotropic channels called α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)receptors into the postsynaptic density (PSD)—a structurelocated in the tip of the dendritic spine of postsynaptic cellscomposed of ionic receptors and a dense network of proteinsthat regulate synapse strength (Hayashi et al., 2000; Chen et al.,2008; Lu et al., 2010). This traffic of new AMPA receptors tothe membrane ultimately leads to an improvement in synapticcommunication (Figure 1).

If the activation of the pre and postsynaptic neuronspersists for longer periods of time—something that can alsobe accomplished in vitro by repeated stimulation of the cellsby high frequency tetanus pulses of 100 Hz, a number of

signaling cascades are activated, leading to protein synthesis andsynapse rearrangement (Nguyen et al., 1994; Hölscher et al.,1997; Ryan et al., 2015). This is known as the late phaseLTP, also called the expression phase. In this stage, the risein Ca2+ ions inside the cell, caused by the constant releaseof glutamate by the presynaptic cell, induces the increase inthe production of cyclic adenosine monophosphate (cAMP) byadenyl cyclase (Wong et al., 1999; Poser and Storm, 2001).cAMP, in turn, activates protein kinase A (PKA) that switcheson mitogen activated protein kinase (MAPK) (Abel et al., 1997;Roberson et al., 1999). This kinase is translocated to the nucleusand phosphorylates cAMP response element binding protein 1(CREB-1), an important transcription factor (Viola et al., 2000;Patterson et al., 2001). The phosphorylation activates CREB-1,resulting in increased transcription of a number of target genesand their subsequent translation into proteins responsible for theformation of new synaptic connections (Figure 2) (Deisserothet al., 1996; Ahmed and Frey, 2005; Benito and Barco, 2010).

The signaling pathways outlined above provide a broadunderstanding about the order and timing of molecular eventsgoverning the early and late phase of NMDA-dependent LTP.However, this model lacks information about regulatory changesthat might be occurring translationally and post-translationally.In addition, this picture only focuses on a limited number

Frontiers in Chemistry | www.frontiersin.org 3 October 2016 | Volume 4 | Article 40

13

directperforantpathway,whichdirectlyconnectstheECtotheCA1region.Onthepictureinright,weshowanillustrationoftheearly-phaseLTP.Here,(1)glutamatefromthepresynapticneuronisreleasedintothesynapticcleft.(2)Thisneurotransmitterreachesionicchannelsofthepostsynapticcellcausingdepolarizationofthisneuronbytheinfluxonsodiumandcalciumcations.(3)Calcium,initsturn,activatesCaMKII that (4) phosphorylates ionic channels in the PSDs and (5, 6) induces the addition of AMPAreceptorstothepostsynapticmembrane,increasingsynapticefficiency.

The morphological changes associated with the formation of new memories in the

hippocampusaretriggeredbythetranscriptionofspecificgenes,thetranslationofdistinctmRNA

sequences, and the regulation of proteins by post-translational modifications (PTMs)

(Kaltschmidtetal.,2006;Maetal.,2014;Parketal.,2006).Notably, the formationofanew

memorymaytakeuptoseveraldays.Duringthistime,liablememoriesbecomestableovertime

throughaprocessknownasmemoryconsolidation(Squireetal.,2015).Studieshaveshownthat

long-termpotentiation(LTP),whichcanbedefinedasanenhancementofsynapticstrength,is

themajorcellularmechanismunderlyingtheprocessofmemoryconsolidationinFCandSMTs

(Nabavietal.,2014;Uzakovetal.,2005;Whitlocketal.,2006).LTPhasbeengenerallydivided

intotwodistinctstages:anearlyphasethatisdrivenprimarilybysignaltransductioneventsand

alatephasewhichtriggersdenovoproteinsynthesis(Abeletal.,1997a;Grangeretal.,2013).

Early-phaseLTP (E-LTP),also referredas the inductionphase,beginsby the releaseof

glutamatefromthepresynapticterminalintothesynapticcleft.Glutamate,themainexcitatory

neurotransmitterinthebrain,diffusesthroughthecleftandreachesthepostsynapticneuron,

whereitbindstoligand-gatedionchannels.Theinteractionbetweentheneurotransmitterand

theionotropicN-methyl-D-aspartate(NMDA)receptorresultsintheinfluxofCa2+andNa+into

thecell(Castillo,2012;Dingledineetal.,1999;Schilleretal.,1998).TheNa+helpstobringabout

adepolarizationofthepostsynapticneuronthatlastafewmilliseconds,whiletheCa2+promotes

theactivationofproteinkinasessuchascalcium/calmodulin-dependentproteinkinase(CaMKII)

(Lisman et al., 2012; Lüscher and Malenka, 2012). CaMKII and other kinases prompt the

introduction of other ionotropic channels such as α-amino-3-hydroxy-5-methyl-4-

isoxazolepropionic acid (AMPA) receptors into the postsynaptic density (PSD) – a specialized

membraneboundstructurelocatedinpostsynapticneurons(X.Chenetal.,2008;Hayashietal.,

2000;W.Luetal.,2010).ThistrafficofnewAMPAreceptorstothemembraneultimatelyleads

toanimprovementinsynapticcommunication(Figure1).

14

Iftheactivationofthepreandpostsynapticneuronspersistsforlongerperiodsoftime–

somethingthatcanalsobeaccomplished invitrobyrepeatedstimulationofthecellsbyhigh

frequencytetanuspulsesof100Hz–,anumberofsignalingcascadesareactivated,leadingtode

novoproteinsynthesis(Hölscheretal.,1997;Nguyenetal.,1994;T.J.Ryanetal.,2015).Thisis

knownaslate-phaseLTP(L-LTP),alsocalledtheexpressionphase.Inthisstage,theriseinCa2+

ionsinsidethecell,causedbytheconstantreleaseofglutamatebythepresynapticcell,induces

theincreaseintheproductionofcyclicadenosinemonophosphate(cAMP)byadenylylcyclase

(PoserandStorm,2001;Wongetal.,1999).cAMP,inturn,activatesproteinkinaseA(PKA)that

switchesonmitogenactivatedproteinkinase(MAPK)(Abeletal.,1997a;Robersonetal.,1999).

ThiskinaseistranslocatedtothenucleusandphosphorylatescAMPresponseelementbinding

protein1(CREB-1),animportanttranscriptionfactor(Pattersonetal.,2001;Violaetal.,2000).

ThephosphorylationactivatesCREB-1,resultinginincreasedtranscriptionofanumberoftarget

genes and their subsequent translation into proteins responsible for the formation of new

synapticconnections(Figure2;AhmedandJ.U.Frey,2005a;Deisserothetal.,1996;Pontesand

deSousa,2016).

Pontes and de Sousa Unveiling Memory through Mass Spectrometry

FIGURE 2 | Late-phase NMDA dependent-LTP. In this stage, (1, 2) Ca2+ ions inside the cell recruit adenyl cyclase to produce cAMP. (3, 4) Cyclic adenosine

monophosphate, in turn, activates PKA that switches on MAPK. (5) This kinase is translocated to the nucleus and phosphorylates CREB-1, an important transcription

factor. (6, 7) The phosphorylation activates CREB-1, resulting in increased transcription of a number of target genes and their subsequent translation into proteins

responsible for the formation of new synaptic connections.

of molecular players and little or no information on theirstoichiometry is known. In the next sections, we describe theuse of mass spectrometry-based proteomics as a tool to help toelucidate those questions.

MASS SPECTROMETRY (MS)-BASEDPROTEOMICS

Proteomics is a system-wide analysis of the proteins expressedin a specific cell, tissue, or organism at a given time (Andersonand Anderson, 1996; Zhang et al., 2013). Although the termproteomics relates to the use of any technology that seeksto interrogate a large number of proteins, it is nowadaysused to refer to works where the central platform is massspectrometry. Currently, the gold standard strategy in MS-basedproteomics is shotgun proteomics (Mann and Kelleher, 2008;Domon and Aebersold, 2010). Here, a complex mixture ofproteins is digested into peptides with a protease of interest,usually trypsin, which cleaves on the C-terminal side of lysineand arginine. Subsequently, the peptides are separated online

by reverse-phase liquid chromatography (LC) and analyzed bymass spectrometers such as quadrupole/time-of-flight (QTOF),ion trap (IT), orbitrap (OT), or ion cyclotron resonance (ICR)(Marshall et al., 1998; Michalski et al., 2011; Thakur et al., 2011;Beck et al., 2015).

In a typical shotgun experiment, the LC-MS/MS run takesup to 120 min and is composed of thousands of cycles, eachone made of a MS1 scan—also known as a full scan—thatmeasures the peptides’ mass to charge ratio (m/z) and intensity,and a MS2 or MS/MS scan. During the MS2 scan, the 20 mostintense peptides in each cycle are fragmented in a collision cell,usually filled with an inert gas such as nitrogen or helium, andtheir spectra are again measured to obtain sequence information(Geiger et al., 2011; Thakur et al., 2011). Once acquired,the LC-MS/MS data are used in searches against databasescontaining peptides digested in silico to identify the proteinspresent in the sample(s) (Figure 3) (Sadygov et al., 2004).

In addition to protein identification, mass spectrometry can beused to extract quantitative information from samples. Proteinquantification can be absolute, if known amounts of a heavyanalog of the analyte of interest is added prior to the analysis in

Frontiers in Chemistry | www.frontiersin.org 4 October 2016 | Volume 4 | Article 40

15

FIGURE2ILate-phaseLTP.Inthisstage,(1,2)Ca2+ionsinsidethecellrecruitadenylcyclasetoproducecAMP. (3,4)Cyclicadenosinemonophosphate, in turn,activatesPKA that switchesonMAPK. (5)ThiskinaseistranslocatedtothenucleusandphosphorylatesCREB-1,animportanttranscriptionfactor.(6,7)ThephosphorylationactivatesCREB-1,resultinginincreasedtranscriptionofanumberoftargetgenesandtheirsubsequenttranslationintoproteinsresponsiblefortheformationofnewsynapticconnections.

Interestingly, however, consolidated memories can still undergo changes or even

disruptionuponrecallinaprocessknownasmemoryreconsolidation(J.L.C.Leeetal.,2017).

Duringreconsolidation,theretrievalofaspecificmemorycaninducethetranslationalofnew

proteinsand thedegradationofpre-existingones (S.-H.Leeetal.,2008;TronsonandTaylor,

2007).Thesemoleculareventsleadtoextramodificationsinthearchitectureofneuronsinthe

hippocampus.Memoryreconsolidationisbelievedtobeanadaptationoforganismstosupply

additionalinformationtoknowledgethathasbeenpreviouslyacquired,increasingthechances

ofanimalstosurviveaneverlastingchangingworld(J.L.C.Leeetal.,2017).Thisprocesshas

beendemonstratedinaplethoraofbehavioraltasks,butevidenceforitinanimalssubjectedto

operantconditioning(OC)–atypeofbehavioralparadigmwhereinanimalslearntoassociatea

behaviorwithitsconsequence(e.g.rodentspressingaleverforthereleaseofwaterorafood

pelletinaSkinnerbox)–isstillunderdebate(HernandezandKelley,2004;TronsonandTaylor,

2007).

b.MassSpectrometry(MS)-basedProteomics

Proteomicsisasystem-wideanalysisoftheproteinsexpressedinaspecificcell,tissueor

organismatagiventime(N.G.AndersonandN.L.Anderson,1996;Zhangetal.,2013).Although

thetermrelatestoanytechnologythatseekstointerrogatealargenumberofproteins,today

proteomics is used to name works where the central platform is mass spectrometry (MS).

Currently,thegoldstandardstrategyinMS-basedproteomicsisshotgunproteomics(Domonand

Aebersold,2010;MannandKelleher,2008).Here,acomplexmixtureofproteinsisdigestedinto

peptideswithaproteaseofinterest–usuallytrypsin,whichcleavesontheC-terminaloflysine

and arginine. Subsequently, the peptides are separated online by reverse-phase liquid

chromatography(LC)andanalyzedbyhighresolutionmassspectrometerssuchastime-of-flight

16

(TOF), orbitrap (OT) or ion cyclotron resonance (ICR)(Beck et al., 2015;Marshall et al., 1998;

Michalskietal.,2012;Thakuretal.,2011).

Inatypicalshotgunexperiment,theLC-MS/MSruntakesupto120minandiscomposed

ofthousandsofcycles,eachmadeofaMS1scan–alsoknownasafullscan–thatmeasuresthe

peptides’masstochargeratio(m/z)andintensity,andaMS2orMS/MSscan.DuringtheMS2

scan,the20mostintensepeptidesineachcyclearefragmentedinacollisioncell,usuallyfilled

withan inertgassuchasnitrogenorhelium,andtheir spectraareagainmeasuredtoobtain

sequenceinformation(Geigeretal.,2011;Thakuretal.,2011).Onceacquired,theMSdatais

usedinsearcheswithindatabasescontainingpeptidesdigestedinsilicotoidentifytheproteins

presentinthesample(s)(Figure3;Sadygovetal.,2004).

FIGURE3IWorkflowofthegoldstandardstrategyinshotgunproteomics.Here,peptidesareseparatedonline inareverse-phase liquidchromatographyandelectrosprayed intothemassanalyzer.Themassspectrometer measures the peptides’ m/z and intensity in the MS1 cycle. Upon fragmentation, theproductionsofeachpeptidearereanalyzedtoobtainsequenceinformationoftheanalyteintheMS2cycle.Oncethisexperimentaldataisacquired,informationissearchedagainstadatabaseoftheorganismofinteresttoidentifytheproteinsinthesample.

In addition to protein identification, mass spectrometry can be used to extract

quantitative information from samples. Protein quantification can be absolute – if known

amountsofaheavyanalogof theanalyteof interest isaddedprior to theanalysis inamass

spectrometer–orrelative, ifsamples indifferentstatesarecompared(e.g.braintissuefrom

rodentstrainedinabehavioralparadigmandcontrols)(Bantscheffetal.,2012;Kettenbachetal.,

2011).Themostpopularstrategiesforrelativequantificationarelabel-free,metaboliclabeling,

andchemicallabeling(Figure4).

Inlabel-freequantification,asthenamesuggests,nolabelisaddedtothesamples,which

aredigestedandrunseparatelyinthemassspectrometer–beingtheresultscombinedafterthe

acquisitionofthedata(Filiouetal.,2012).Inthisstrategy,quantificationtakesadvantageofthe

Pontes and de Sousa Unveiling Memory through Mass Spectrometry

FIGURE 3 | Workflow of the gold standard strategy in shotgun proteomics. Here, peptides are separated online in a reverse-phase liquid chromatography and

electrosprayed into the mass analyzer. The mass spectrometer measures the peptides’ m/z and intensity in the MS1 cycle. Upon fragmentation, the product ions of

each peptide are reanalyzed to obtain sequence information of the analyte in the MS2 cycle. Once this experimental data is acquired, information is searched against

a database of the organism of interest to identify the proteins in the sample.

a mass spectrometer, or relative, if samples in different states arecompared (e.g., brain tissue from rodents trained in a behavioralparadigm vs. controls) (Kettenbach et al., 2011; Bantscheff et al.,2012). The most popular strategies for relative quantification arelabel-free, metabolic labeling, and chemical labeling (Figure 4).

In label-free quantification, as the name suggests, no label isadded to the samples, which are digested and run individuallyin the mass spectrometer—being the results computationallycombined after the acquisition of the data (Filiou et al., 2012). Inthis strategy, quantification takes advantage of the area plottedover time for each ion as it elutes from the chromatographiccolumn. Later, this extracted ion chromatogram (XIC) is alignedacross different samples, and a ratio for each peptide is obtained.Another mode of quantification in label-free experiments isspectrum counting. Here, quantification is based on the numbera particular peptide is fragmented during a LC-MS/MS run,which serves as a proxy for abundance and can be comparedbetween conditions (Bantscheff et al., 2007; Hernández et al.,2012). Label-free is regarded as the least accurate strategy ofrelative quantification, but it has gained momentum due toits low cost, improvements in sample handling, refinement ofthe chromatographic setup, and development of software foraccurate data analysis (Ong and Mann, 2005; Altelaar and Heck,2012).

By contrast, in metabolic labeling, prior to protein extractionat least one of the conditions is labeled with a heavy stable isotopesuch as 15N or heavy amino acids such as lysine, arginine, orboth (Ong, 2002; Rauniyar et al., 2013). The use of a heavyanalog prevents the variation usually encountered in label-freeexperiments, since the samples are mixed, digested and analyzedsimultaneously in the LC-MS/MS run. This can be accomplishedbecause heavy (labeled) and light (non-labeled) peptides retainthe same physicochemical properties (e.g., retention time duringthe LC), but a mass shift between them enables their distinctionlatter in the data analysis. The only exception to this rule isdeuterium (2H), which is more hydrophilic than hydrogen; thiscreates a delay in the retention time between the labeled andnon-labeled conditions (Yi et al., 2005). Here, as it is also the casein label-free experiments, quantification is acquired by the peakarea ratios of the heavy and light peptides in the XIC (Ong, 2002;Bantscheff et al., 2007).

Even though metabolic labeling is the most accurate relativequantification strategy, it has a restricted capacity to multiplexing

due to the limitation on the isotopes that can be added toan amino acid and the increase in sample complexity in theMS1 (Hebert et al., 2013). Chemical labeling, in its turn, is ableto circumvent those limitations. To this day, many chemicallabeling reagents have been developed, but the most used areisotope-coded affinity tags (iCAT), tandemmass tags (TMT), andisobaric tags for relative and absolute quantification (iTRAQ)(Gygi et al., 1999; Thompson et al., 2003; Ross, 2004). iCAT,which labels samples at the protein level, uses the tags that arecomposed of a reactive group that binds to reduced cysteineresidues, a linker group that incorporates isotopes in the heavyreagent, and a biotin affinity group for the isolation of the iCAT-labeled peptides. In a typical experiment using iCAT, the reagentlabels protein samples at two different conditions (the lightand heavy version of the tags), which are then mixed togetherand enzymatically cleaved. Next, the peptides with the tags areenriched by avidin affinity chromatography and analyzed in aLC-MS/MS run. Here, quantification is obtained by the peak arearatios of the heavy and light peptides (Yi et al., 2005).

iCAT possesses the same limitation in multiplexing asmetabolic labeling, yet iTRAQ and TMT enables from 8 to 10samples, respectively, to be analyzed in a single experiment.iTRAQ and TMT are isobaric tags that label analytes at thepeptide level. Their tags are composed of a reactive group thatbinds to the N-terminal of peptides and lysine residues, a balancegroup—which ensures that the same peptides in the differentconditions elute together and are indistinguishable in the MS1scan—and a reporter group. Unlike other quantitative strategies,quantification on iTRAQ and TMT is based on the intensitysignal of the reporter group that is released from the analytesupon fragmentation in a collision cell. Nevertheless, chemicallabeling also has limitations and some considerations have to betaken to get around these drawbacks (For in depth informationon iTRAQ and TMT strategies see Bantscheff et al., 2008; Karpet al., 2010; Ting et al., 2011; Wenger et al., 2011).

PROTEOMIC STUDIES OF MEMORY

The shotgun quantitative strategies described above havefostered a revolution in many fields of biology such as cancer,immunology, and neuroscience by improving our understandingof the systemic cellular response of stimulated or disease states

Frontiers in Chemistry | www.frontiersin.org 5 October 2016 | Volume 4 | Article 40

17

areaplottedovertimeforeach ionas itelutes fromthechromatographiccolumn.Later, this

extractedionchromatogram(XIC)isalignedacrossdifferentsamplesandaratioforeachpeptide

isobtained.Anothermodeofquantificationinlabel-freeexperimentsisspectrumcounting.Here,

quantificationisbasedonthenumberaparticularpeptideisfragmentedduringaLC-MS/MSrun,

whichservesasaproxyforabundanceandcanbecomparedbetweenconditions(Bantscheffet

al.,2007).Label-free is the leastaccuratestrategyofrelativequantification,but ithasgained

momentum due to its low cost, improvements in sample handling, refinement of the

chromatographicsetup,anddevelopmentofsoftwarefordataanalysis(AltelaarandHeck,2012;

OngandMann,2005).

Pontes and de Sousa Unveiling Memory through Mass Spectrometry

FIGURE 4 | Different strategies to quantify peptides. In label free experiments, samples are digested and ran separately in a mass spectrometer; they are

combined only in the data analysis. In metabolic labeling, on its turn, one of the conditions is grown in medium containing amino acids labeled with heavy isotopes or

heavy nitrogen. Here, the samples are combined very early, and sample handling and analysis are done concomitantly. Lastly, in chemical labeling, tags are

incorporated at the protein level, as is the case of iCAT, or at the peptide level with iTRAQ and TMT. In these strategies, the samples are combined early in the

workflow of the experiment without increasing the complexity of the samples in the MS1, since the tags of iTRAQ and TMT are isobaric—being distinguished only

upon fragmentation in a collision cell.

vs. control (Krüger et al., 2008; Dahlhaus et al., 2011; Geigeret al., 2012; Boersema et al., 2013; Meissner and Mann, 2014;Nascimento and Martins-de-Souza, 2015). Yet, when it comes tothe proteomic study of memory, very few MS-based experimentshave been carried out so far. Some authors believe that suchdiscrepancy is due to the difficulty to characterize proteins thatare genuinely associated with this cognitive process due to highbiological variability among individuals within areas related tomemory such as the hippocampus (Dieterich and Kreutz, 2015).Nevertheless, this explanation falls short, since transcriptomics

assays to investigate memory have been conducted before withsuccess (Ponomarev et al., 2010; Bero et al., 2014).

In addition, two recent proteomic studies have demonstratedthat those kinds of experiments are feasible. In the first one,Borovok et al. used a radial arm maze (RAM) paradigmto understand the process of memory consolidation in thehippocampus of mice (Borovok et al., 2016). In summary, theRAM behavioral task works as follows: a central circular chamberis connected to eight long arms, which are open to the animalupon the removal of a guillotine door. In the end of four of the

Frontiers in Chemistry | www.frontiersin.org 6 October 2016 | Volume 4 | Article 40

18

FIGURE4IDifferentstrategiestoquantifypeptides.Inlabel-freeexperiments,samplesaredigestedandranseparatelyinamassspectrometer;theyarecombinedonlyinthedataanalysis.Inmetaboliclabeling,onitsturn,oneoftheconditionsisgrowninmediumcontainingaminoacidslabeledwithheavyisotopesorheavynitrogen.Here,thesamplesarecombinedveryearly,andsamplehandlingandanalysisaredoneconcomitantly.Lastly, inchemical labeling, tagsare incorporatedat theprotein level,as is thecaseofiCAT,oratthepeptidelevelwithiTRAQandTMT.Inthesestrategies,thesamplesarecombinedearlyintheworkflowoftheexperimentwithoutincreasingthecomplexityofthesamplesintheMS1,sincethetagsofiTRAQandTMTareisobaric—beingdistinguishedonlyuponfragmentationinacollisioncell.

Bycontrast,inmetaboliclabeling,priortoproteinextractionatleastoneoftheconditions

islabeledwithaheavystableisotopesuchas15Norheavyaminoacidssuchaslysine,arginineor

both(Ongetal.,2002;Rauniyaretal.,2013).Theuseofaheavyanalogpreventsthevariation

usually encountered in label-free experiments, since the samples are mixed, digested and

analyzedsimultaneouslyintheLC-MS/MSrun.Thiscanbeaccomplishedbecauseheavyandlight,

non-labeled,peptidesretainthesamephysicochemicalproperties(e.g.retentiontimeduringthe

LC),butamassshiftbetweenthemenablestheirdistinctionlatterinthedataanalysis.Theonly

exceptiontothisruleisdeuterium(2H),whichismorehydrophilicthanhydrogen;thiscreatesa

delay in the retention timebetween the labeledandnon-labeled conditions (Yi et al., 2005).

Here,asisalsothecaseinlabel-freeexperiments,quantificationisacquiredbythepeakarea

ratiosoftheheavyandlightpeptidesintheXIC(Bantscheffetal.,2007;Ongetal.,2002).

Eventhoughmetabolic labeling isthemostaccuraterelativequantificationstrategy, it

hasarestrictedcapacitytomultiplexingduetothelimitationontheisotopesthatcanbeadded

toanaminoacidandtheincreaseinsamplecomplexityintheMS1(Hebertetal.,2013).Chemical

labeling,initsturn,isabletocircumventthoselimitations.Tothisday,manychemicallabeling

reagentshavebeendeveloped,butthemostusedareisotope-codedaffinitytags(iCAT),tandem

masstags(TMT)andisotopetagsforrelativeandabsolutequantification(iTRAQ)(Gygietal.,

1999;P.L.Rossetal.,2004;Thompsonetal.,2003).iCATlabelssamplesattheproteinlevel,and

thetagsarecomposedofareactivegroupthatbindstoreducedcysteineresidues,alinkergroup

thatincorporatesisotopesintheheavyreagent,andabiotinaffinitygroupfortheisolationof

theiCAT-labeledpeptides.InatypicalexperimentusingiCAT,twoconditionsarelabeledwiththe

lightandheavyversionofthetags,mixed,andenzymaticallycleaved.Next,thepeptideswith

thetagsareenrichedbyavidinaffinitychromatographyandanalyzedinaLC-MS/MSrun.Here,

quantificationisobtainedbythepeakarearatiosoftheheavyandlightpeptides(Yietal.,2005).

19

iCATpossessesthesamelimitationinmultiplexingasmetaboliclabeling,yetiTRAQand

TMTenablesfrom8to10samples,respectively,tobeanalyzedinasingleexperiment.iTRAQ

andTMTareisobarictagsandlabelanalytesandthepeptidelevel.Theirtagsarecomposedofa

reactivegroupthatbindstotheN-terminalofpeptidesandlysineresidues,abalancegroup–

which ensures that the same peptides in the different conditions elute together and are

indistinguishableintheMS1scan–,andareportergroup.Unlikeotherquantitativestrategies,

quantificationoniTRAQandTMTisbasedontheintensitysignalofthereportergroupthatis

releasedfromtheanalytesuponfragmentationinacollisioncell.Nevertheless,chemicallabeling

alsohaslimitationsandsomeconsiderationshavetobetakentogetaroundthesedrawbacks

(ForindepthinformationoniTRAQandTMTstrategiesseeBantscheffetal.,2008;Karpetal.,

2010;Tingetal.,2011;Wengeretal.,2011).

Theshotgunquantitativestrategiesdescribedabovehavefosteredarevolutioninmany

fieldsofbiologysuchascancer,immunology,andneurosciencebyimprovingourunderstanding

of thesystemiccellular responseof stimulatedordiseasestatesvs. control (Boersemaetal.,

2013;Geigeretal.,2012;Krugeretal.,2008;MeissnerandMann,2014).Yet,whenitcomesto

theproteomicstudyofmemory,veryfewMS-basedexperimentshavebeencarriedoutsofar

(Borovoketal.,2016;Monopolietal.,2011).Someauthorsbelievethatsuchdiscrepancyisdue

tothedifficultytocharacterizeproteinsthataregenuinelyassociatedwiththiscognitiveprocess

duetohighbiologicalvariabilityamongindividualswithinareasrelatedtomemorysuchasthe

hippocampus. Nevertheless, this explanation falls short, since transcriptomics assays to

investigatememoryhavebeenconductedbeforewithsuccess(Cavallaroetal.,2002).Moreover,

afewproteomicexperimentshavebeencarriedoutinanimalssubjectedtoFCandSMTs,yet

noneinanimalssubjectedtooperantconditioning.

20

MotivationsforThisWork

Memoryisoneofthemostimportantcomponentsofcognition,yetIthasbeenlargely

studied in animals subjected to specific behavioral paradigms such fear conditioning (FC) or

spatialmemorytasks(SMTs).Moreover,mostoftheseassayshaveemployedtechniquesthat

arelimitedtoaspecificsequencesuchasgeneknockoutsoradministrationofproteininhibitors,

lackingunderstandingoftheprocessofmemoryconsolidationataglobal-scale.Transcriptomic

assays have been performed aswell, however transcripts are not the end products ofmost

biologicalfunctionsandcanonlybeusedasaproxyforproteindynamics.

Inordertoextendtheknowledgeofhowmemoriesbecomeconsolidatedovertimein

the hippocampus of animals subjected to an operant conditioning (OC) task, we measured

protein, phosphoproteins and mRNA levels in rats subjected to this behavioral paradigm at

differenttimepoints.Toaccomplishedthat,weperformedthefollowingstepsthroughoutthis

studied:

• Modify the standard behavioral protocol of OC to identify proteins,

phosphoproteinsandtranscriptswhicharegenuinelyassociatedwiththisformof

memory;

• Developandimplementapre-fractionationstrategyforin-depthidentificationof

theproteinsinthehippocampusviaLC-MS/MS;

• Identify statistically significant abundance changes at 30min, 12h and after a

recallsessionthetookplace24hfollowingbehavioralconditioning;

• SelectmRNAtargetsbasedontheseresultsandtheliteraturetounderstandthe

relationbetweenproteinandtranscriptlevels;

• Identifyandquantifythephosphoproteomeofratsattheaforementionedtime

pointsusingmassspectrometry.

21

Results

ParametersoftheOperantConditioningParadigm

Inordertomeasurequantitativechangesinprotein,phosphoproteinandmRNAduring

memoryconsolidationandafterrecall,werandomlydistributedrats(n=12)intofourdifferent

groups.Alloftheseanimals,exceptthecontrolindividuals,weresubjectedtoOCinaSkinner

box,withaminormodificationofthestandardprotocol(Figure5A).Inacommonlyusedtraining

schedule,rodentsareplacedintheboxfortwoorthree20minsessions(i.e.habituation)prior

toconditioning.Thisisdonetominimizestressresponsesthatcannegativelyaffectthetraining

procedure,sincesomeanimals,forinstance,mayfreezewhenplacedintoanewenvironment.

Duringhabituation,arewardsuchasafoodpelletisreleasedatrandombyafeeder.Afterthat,

habituatedanimals return to theboxandundergo training trials,whichconsistofexposinga

retractedleverevery20sfor60sandmeasuringthenumberofcorrectleverpressesofeach

animal(Exton-McGuinnessetal.,2014;Rapanellietal.,2011).

Inthisstandardprocedure,however,thecriterionforbehavioracquisitionisnotusually

reachedinasingletrainingday,suchthattheanimalshavetobesubjectedtoadditionaltrialsin

theSkinnerbox.Insomecases,conditioningsessionsmaytakeupto10days(Exton-McGuinness

etal.,2014;Jurado-Parrasetal.,2013).Weanticipatedthistobeproblematicbecausesubjecting

rodentstoconsecutivetrainingsessionscouldpotentiallydiluteproteomicandtranscriptomic

signalsassociatedwiththeprocessofmemoryconsolidationormolecularsignaturescorrelated

with memory reconsolidation mechanisms. We thus manually performed the training by

releasingwater (i.e. the reward) as animals got progressively closer to the lever. Eventually,

animalslearntthatpressingtheleverintheboxwouldresultinthereward(Figure5B).Here,we

set the criterion for behavioral acquisition to 15 lever presses in a row (Figure 5C). Animals

subjected to this modified training protocol were able to retain the information of a single

trainingdayforacoupleofdays,whileindividualsthatweresubjectedtoarecallsession(i.e.50

leverpressesinarow)rememberedtheeventforatleasttwoweeks(Figure5C;datanotshown).

TheseresultsshowthatratsareabletolearnOCtasksinonetrainingsession,establishingthe

useofthismodifiedprotocoltostudyassociativememoriesinthehippocampus.

22

FIGURE 5 I Parameters of the Operant Conditioning Paradigm. (A) Illustration of a Skinner box. (B)Behavioralprotocolusedinthedifferentgroupsoftheexperiment.(C)Numberofleverpressesofeachanimalinthedifferentgroups–30min,12handRecall–duringa20minsession.Animalspressed15timestheleverduringthetrainingsession,whilerodentssubjectedtoarecallsessionpressedthelever50times.Controlindividualswerenotsubjectedtotraining,butunderwenthabituationjustliketheotheranimalsoftheexperiment.

StrongAnionExchange(SAX)FractionationEnablesIn-depthProteomicsoftheHippocampus

Ratssubjectedtotheaforementionedtrainingprotocolwereeuthanizedatdistincttime

points–30min,12hand30minafterarecallsessionperformedonedayafterconditioning.

ControlanimalsdidnotundergoOC,butwereplacedintheSkinnerboxliketheotheranimals.

0

2

4

6

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Num

bero

fLeverPresses

Time(min)

GroupT30min (Training)

Individual1 Individual2 Individual3

02468

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Num

bero

fLeverPresses

Time(min)

GroupT12h (Training)

Individual1 Individual2 Individual3

02468

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time(min)

GroupTRecall (Training)

Individual1 Individual2 Individual3

0

5

10

15

20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Time(min)

GroupTRecall (Recall)

Individual1 Individual2 Individual3

B

LETTERdoi:10.1038/nature13294

EngineeringamemorywithLTDandLTPSadeghNabavi1*,RockyFox1*,ChristopheD.Proulx1,JohnY.Lin2,RogerY.Tsien2,3&RobertoMalinow1

Ithasbeenproposedthatmemoriesareencodedbymodificationofsynapticstrengthsthroughcellularmechanismssuchaslong-termpotentiation(LTP)andlong-termdepression(LTD)1.However,thecausallinkbetweenthesesynapticprocessesandmemoryhasbeendifficulttodemonstrate2.Hereweshowthatfearconditioning3–8,atypeofassociativememory,canbeinactivatedandreactivatedbyLTDandLTP,respectively.Webeganbyconditioningananimaltoassociateafootshockwithoptogeneticstimulationofauditoryinputstargetingtheamygdala,abrainregionknowntobeessentialforfearconditioning3–8.SubsequentoptogeneticdeliveryofLTDcondition-ingtotheauditoryinputinactivatesmemoryoftheshock.Thensub-sequentoptogeneticdeliveryofLTPconditioningtotheauditoryinputreactivatesmemoryoftheshock.Thus,wehaveengineeredinactiva-tionandreactivationofamemoryusingLTDandLTP,supportingacausallinkbetweenthesesynapticprocessesandmemory.

Toexaminetherelationbetweensynapticplasticityandmemory,weusedcued-fearconditioning3–8inrats,whereinaneutralconditionedstim-ulus(CS),suchasatone,whenpairedwithanaversiveunconditionedstimulus(US),resultsinatone-drivenconditionedresponse(CR)indi-catingmemoryoftheaversivestimulus.Temporally(butnotnon-temporally)pairingatonewithashockledtoarobustCR(reducedleverpressingtoapreviouslylearnedcuedlever-presstask9;ExtendedDataFig.1)duringsubsequenttestingwithatonealone3–8(Fig.1a).Toinves-tigatethesynapticbasisunderlyingthisassociativememory,wereplacedatonewithoptogeneticstimulationofneuralinputstothelateralamyg-dalaoriginatingfromauditorynuclei.Weinjectedanadeno-associatedvirus(AAV)expressingavariantofthelight-activatedchannelChR2,oChIEF,thatcanrespondfaithfullyto50–100Hzstimuli10,intothemedialgeniculatenucleusandauditorycortex(ExtendedDataFig.2).Afterthechannelreachedaxonalterminalsinthelateralamygdala(ExtendedDataFig.3),acannulapermittinglightdeliverywasplacedtargetingthedorsaltipofthelateralamygdala(ExtendedDataFig.4).AnopticalCSalone(a2min10Hztrainof2mspulses,seeMethods)hadnoeffectonleverpressing(ExtendedDataFig.5).However,tem-porally(butnotnon-termporally)pairingtheopticalCSwithafootshock(seeMethods)ledtoaCR(Fig.1b)thatwassensitivetoextinc-tion(seebelow)andblockedbyNMDAreceptorinhibitionduringconditioning(ExtendedDataFig.6),indicatingthegenerationofanassociativememory.

ToexamineifLTPoccurredafterpairingopticalCSwithfootshock3–8

(Fig.1d),wepreparedamygdalabrainslicesfromanimalsreceivingunpaired,pairedornoconditioning,andmeasuredtheAMPAreceptorcomponent(A)andNMDAreceptorcomponent(N)oftheopticallydrivensynapticresponse(Fig.1c).TheA/NratioincreasedinanimalsreceivingpairedconditioningindicatingthatLTPhadoccurred11,12atopticallydriveninputstolateralamygdalaneurons.

Canmemoriesbeinactivated?IfLTPoccurredattheopticallydrivensynapseontothelateralamygdala,andthisLTPcontributestothememory,reversingLTPwithLTD13,14shouldinactivatethememory.AnimalsthatdisplayedCRafterpairedopticalCS-shockconditioningwereexposedtoanopticalLTDprotocol(seeMethods).Onedaylater,animalsweretestedwithopticalCSanddisplayednoCR,indicatinginactivationof

thememoryoftheshockbyLTD(Fig.2a,b,f).Nextweexaminedifmemoriescanbereactivated.TotheseanimalswedeliveredanopticalLTPprotocol(seeMethods).Onedaylater,animalsdisplayedaCR

*Theseauthorscontributedequallytothiswork.

1CenterforNeuralCircuitsandBehavior,DepartmentofNeuroscienceandSectionofNeurobiology,UniversityofCaliforniaatSanDiego,California92093,USA.2DepartmentofPharmacology,UniversityofCaliforniaatSanDiego,California92093,USA.3HowardHughesMedicalInstitute,UniversityofCaliforniaatSanDiego,California92093,USA.

ToneShock

Paired conditioning

Tone

Shock + tone

No fearFear

ODIShock

Paired conditioning

ODI

Shock + ODI

No fearFear

Laser

abc

Unpaired conditioned

Paired conditioned

Lever presses (norm)

Time (min)

d

02460.00.51.01.5

02460.00.51.01.5

02460.00.51.01.5

02460.00.51.01.5

0.00.20.40.60.81.0**

0.00.20.40.60.81.01.2

123456

AMPA/NMDA

LA

d

m

Time (min)

UnpairedPaired

**

UnpairedPaired

**

Unpair

edPair

edNaiv

e

Figure1|Fearconditioningwithtoneoroptogenetics.a,Top,diagramofratreceivingtoneandshockduringconditioning.Ratsexposedtounpaired(n55,middle)ortemporallypaired(n55,bottom)toneandshockweretestedonedaylaterbyatone(green).Timeplotsshownormalizednumberofleverpresses(1minbins)toapreviouslylearnedcuedlever-presstask.Bargraphshowsnormalizednumberofleverpressesduringthefirstminuteoftone.b,Top,diagramofratreceivingoptogeneticallydriveninput(ODI)stimulationandshockduringconditioning.Rats(n58)receivedunpaired(middle)andonedaylatertemporallypaired(bottom)ODIandshock.Timegraphsasina,exceptanimalsweretestedby10HzODI(blue).Bargraphasinafor10HzODI.c,Top,experimentaldesign;averagedopticallydrivensynapticresponsesobtainedat260mV(blue),140mV(red)and0mVholdingpotentialforcellsfromanimalsthatreceivedunpaired(top)orpaired(bottom)conditioning.TraceswerescaledtomatchNMDA-mediatedcurrents.BargraphplotsaverageAMPA/NMDA(noconditioning,2.460.2,n511from6rats;unpairedconditioning2.160.2,n510from6rats;pairedconditioning4.460.6,n58from4rats).Scalebars,100pA,50ms,1mm.d,Synapticmodificationmodel.Temporallypairingoftone(left)orODI(right)andshockinputstolateralamygdalaneuronsleadstopotentiationoftone(left)orODI(right)input,whichcancontributeintriggeringCR.Hereandthroughout:NS,notsignificant;*P,0.05;**P,0.01;errorbars,s.e.m.SeeMethodsfordetails.

00MONTH2014|VOL000|NATURE|1

Macmillan Publishers Limited. All rights reserved ©2014

A

C

Control

30min

12h

Recall

Day1 Day2 Day3 Day4

HabituationwithH2Oatrandom

BehavioralTraining

H2Oatrandom

Recall Euthanized

30min

30min

30min

12h

23

Thesecontrolratsalsowentthroughthehabituationstepsdescribedearlier,beingeuthanized

30minafterleavingthebox(Figure5B).Thehippocampiofallanimalsweredissectedunder3

min and processed using the filter-aided sample preparation (FASP) method, with minor

modifications(Figure6A;seemethodsformoredetails).Aftertrypticdigestion,peptidesofeach

condition were labeled with isobaric tags for relative and absolute quantification (iTRAQ) in

biologicaltriplicates–control(tag114),30min(tag115),12h(tag116),recall(tag117).Next,

labeledpeptidesderivedfromeachconditionweremixedinequalamounts,andsmallaliquots

fromeachbatchwereanalyzedinshortLC-MS/MSruns.Thiswasperformedtoverifypotential

differencesintheefficiencyofthelabelingreactionsandtomeasuretheproportionamongthe

iTRAQtagsineachbatch.AsillustratedinFigure6B,thelabelingwascarriedouteffectivelyand

theproportionamongthetagshavebeenevenlydistributedafternormalization.

24

FIGURE6IStrongAnionExchange(SAX)EnablesIn-depthProteomicsoftheHippocampus.(A)Workflowof the experiment. Hippocampi samples from operant conditioned and control animals wereenzymaticallydigested,labelledwithdifferentiTRAQreagenttagsandcombinedinequalratiosbeforeSAXfractionation.Fractionatedsampleswererunseparatelyinthemassspectrometerandthedatawerecombinedcomputationally.(B)Priortofractionation,eachbatchofthebiologicalreplicatescontaininglabelledpeptideswereanalyzedinshortLC-MS/MStoverifylabelingefficiencyandproportionamongthetags.Afternormalization,theproportionamongthetagshavebeendistributedevenlyinallreplicates.

Sinceanimalproteomesexhibitgreatcomplexityandlargedynamicranges(Mannetal.,

2013;H.Wangetal.,2015),wedecidedtopre-fractionatesamplespriortomassspectrometry

analyses.BeforeperformingthisstepintheiTRAQlabelledsamples,webenchmarkedforthe

first timetwooffline fractionationstrategies: stronganionexchange (SAX)andstrongcation

exchange (SCX). Differently from previous assays that have used the isoelectric point (pI) to

separatepeptidesinSAX(Nagarajetal.,2011;Wisniewskietal.,2009),weelutedtheanalytes

fromthestationaryphasebyincreasingtheconcentrationofsaltintheelutionbuffers.Label-

freeanalysesofthefractionsshowedthatSAXfractionationwassuperiortoSCX,althoughthe

laterhasbeenthemethodofchoiceinmostproteomicexperimentstodate(Azimifaretal.,2014;

Villénetal.,2007).TheSAXapproachyieldedamuchlargernumberofidentifiableproteinsand

alsoresultedinhighlyreproducibletechnicalreplicateswhencomparedwiththeSCXstrategy

(averagePearsoncorrelationcoefficientof0.9;Figure7;supplementaryexcelfile1).Thus,we

usedSAXtoseparatetheiTRAQlabeledsamplesinto12fractionsperbiologicalreplicate.

Subsequently, each SAX fraction was analyzed using a 165-min organic gradient in a

hybrid-ion trap orbitrap mass analyzer (Orbitrap Elite). All the raw files were processed in

ProteomeDiscover(PD)version2.1(ThermoFisherScientific)usingSequestHTandAmandaMS

assearchingengines,settingthefalsediscoveryrate(FDR)attheproteinandpeptidelevelsat

1%.Using these searchparameters,we identified a total of 7,744proteinswith at least one

uniquepeptideandanaveragesequencecoverageof17.5%(supplementaryexcelfile2).The

datageneratedthroughthebenchmarkingofSAXandSCXincombinationwiththeresultsofthe

iTRAQlabeledsamplesproducedatotalof8,951proteins.Thisdatasetshowsa13%increaseof

coverageincomparisontoapreviouslymousehippocampalproteomestudy,makingthepresent

workthelargesthippocampalproteomedatasettodate(Sharmaetal.,2015).

25

FIGURE7IBenchmarkofSAXtoSCX.Hippocampalpeptidescomingfromthesamedigestionbatchweredividedequallyintosixmicrotubes–SCX(n=3)andSAX(n=3).TheyweredriedinaSpeedvacandre-suspendedwiththeappropriatebuffer(seemethodsformoredetails).Thesepeptideswereloadedintothecolumnsandelutedbyincreasingtheconcentrationofsaltinthemobilephase.Next,eachfractionwasdesaltedutilizingStageTipscontainingC18disksandanalyzedseparatelyinthemassspectrometer.(A) Venn diagrams of the technical replicates of SAX and SCX strategies. In total,we identified 7,000proteinsusingSAXand5,620proteinsusingSCX.(B)Densityplotsofthetechnicalreplicatesofthetwofractionation strategies; Pearson correlation is shown in blue in each quadrant. Average PearsoncorrelationforSAXwas0.9andforSCXwas0.91.Werepeatedtheseexperimentswithdifferentelutionbufferconcentrationanddistinctorganisms(i.e.TrypanosomacruziandHeLacells)(datanotshown).Weobtainedsimilarresultsinallexperiments.

26

QuantitativeProteomicsoftheHippocampusDuringMemoryFormation

Quantitative information about proteins identified at the three-time intervals was

obtainedbyextractingiTRAQintensitiesderivedfromhighconfidencehippocampuspeptides.

Evaluationofthese intensities intheRsoftwareenabledthequantificationof4,554proteins,

which spanned the fold-change range of -2 ⋜ X ⋝ 2 (supplementary excel file 3). Proteins

displaying significant abundance changes were identified combining limma test and rank

products. These statistical tools work well in experiments with a small number of biological

replicates (n = 3) and data sets with additional missing values such as transcriptomic and

proteomic assays (Schwämmle et al., 2013). These analyses yielded a total of 465 regulated

proteinsthroughoutthethree-timepointsunderstudyhere(p<0.05;supplementaryexcelfile

3).

Tofindsimilaritiesacrosstheexperimentalconditionsandtouncoverputativefunctional

relationshipsamongallthequantifiableproteins,wesubjectedthemtofuzzyc-meansclustering.

Bydeterminingthefuzzifieraccordingto(SchwämmleandJensen,2010),wecompiledthedata

setintothreeclusters(Figure8;supplementaryexcelfile4).Theproteinsinclusternumber1

exhibitedanincreaseinabundancethroughouttheprocessofmemoryconsolidationandafter

therecallsession,whileproteinsinclusternumber2showedanoppositetendency.Conversely,

proteinsinclusternumber3decreased30minafterconditioning,butincreasedonceagainat12

handafterbehaviorretrieval.

Cluster 1

Time

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Cluster 2

Time

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Cluster 3

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Cluster 1

Time

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Cluster 2

Time

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Cluster 3

Expr

essi

on c

hang

es

1 2 3 4

−1.5

−1.0

−0.5

0.0

0.5

1.0

1.5

Control30min12hRecall Control30min12hRecall Control30min12hRecall

Cluster 1 Cluster 2 Cluster 3

-1.5-1

.00.500.51.01.5

-1.5-1

.00.5

00.5

1.01.5

-1.5-1

.00.500.51.01.5

27

FIGURE8IFuzzyc-meansClustering.Determinationofthefuzzierenabledgroupingofproteinsintothreedistinct clusters, which show patterns of expression across time. The proteins in cluster 1 display anincrease in abundance throughout the process ofmemory consolidation and after the recall session.Proteins of cluster 2 exhibited an opposite trend to cluster 1. Proteins of cluster 3 decreased theirabundancelevels30minafterconditioning,butincreasedonceagainat12handafterbehaviorretrieval.

ProteinsDisplayingAbundanceChangesasEarlyas30minAfterOperantConditioning

Administrationofproteinsynthesisinhibitorspriorto,orjustaftersubjectinganimalsto

behavioral paradigms, leads to deficits in memory consolidation (Abel et al., 1997b;

Bourtchouladzeetal.,1998).Thisdemonstratesthatdenovoproteinsynthesisisrequiredduring

suchatimeintervalfortheestablishmentoflong-termmemory.However,becausetranslation

inhibitorspossessabroad-spectrum,theidentityofmanyprotein(s)necessaryatearlystagesof

memoryconsolidationremainsunknown.Toprobetheidentityofthesemolecularplayers,we

analyzed the proteomic data set at 30 min after conditioning and showed that 20 proteins

exhibitedsignificantabundancechangeswhencomparedwiththecontrolgroup.Fromthese,3

down-regulatedproteinsweregrouped in clusters2and3,while10up-regulatedoneswere

groupedincluster1(Figure8;supplementaryexcelfile4).Theidentificationoftheseproteins

reveals thatOC-basedassociativememorieshave anearly time sensitivewindow forprotein

synthesispriortothefirsthouraftertraining.

Operant Conditioning (OC)-Based Associative Memories Possess Particular Molecular

Signatures

Memory consolidation may take several days (Squire et al., 2015). During this time,

activatedcellsofneuralnetworksundergolargewavesoftranscriptionandtranslationchanges

(Borovoketal.,2016;Cavallaroetal.,2002;Monopolietal.,2011).Twoofthesemajorexpression

waveshavebeenidentifiedinhippocampalcellsofratssubjectedtoaspatialparadigmat3and

12hafterconditioning.Toexaminetheproteinchangestakingplaceatlatertimeintervals,we

analyzedthehippocampalproteomeofrats12hafterconditioning.Weidentifiedatotalof170

proteinswithstatisticallysignificantabundancechanges(supplementaryexcelfile3).Ofthese,

63weredown-regulatedand107wereup-regulated.Fuzzyc-meansclusteringanalysesgrouped

92oftheseproteinsincluster1and54incluster2(Figure8).

28

Toextractbiologicalmeaningfromtheproteinsinthisdataset,weclassifiedthemusing

theDatabaseforAnnotation,VisualizationandIntegratedDiscovery(DAVID)(D.W.Huangetal.,

2007).Byassigninggeneontology(GO)termsto166proteins(p<0.05),weobservedthatour

data was enriched for cytoplasmic, extracellular exosome, nucleus and membrane proteins

(supplementaryexcelfile5).Othertermssuchasproteincomplex,post-synapticdensity,axon,

dendrite and neuron projection were also assigned to proteins in this data set. Further, by

analyzingthefunctionalcategoriesoftheseproteins,wefoundthatthetopenrichedtermswere

phosphoprotein,cytoplasmandacetylation(p<0.05;supplementaryexcelfile5).Theseresults

reinforce the notion that PTMs play a central role in the process of memory formation

(RouttenbergandRekart,2005).

We inferred physical and functional protein-protein associations among these 170

proteinsbyanalyzingthemusingtheSTRINGdatabase(Szklarczyketal.,2017);thisinteraction-

predictiontoolindicatedthatonlyasmallsetofproteinsidentifiedinthistimeintervaldidnot

displayassociations(Figure9A).WefurthersubjectedtheproteinnetworkgeneratedbySTRING

to clustering with overlapping neighborhood expansion (ClusterONE), an algorithm which

predicts protein complexes from protein-protein interaction data (Nepusz et al., 2012). This

analysisgeneratedatotalof6proteinclusters(p<0.05),whichmightrepresentgenuineprotein

complexes incells (Figure9B).ClusterVhad the largestnumberofproteinswitha totalof8

members.Proteinsofthe14-3-3proteinfamily,whichhavebeenassociatedwiththeregulation

ofthecytoskeletonandtheformationofnewassociativememoriesinthehippocampus(Angrand

etal.,2006;Qiaoetal.,2014),wereenrichedinclusterIIIandIV.InclusterIV,membersofthe

14-3-3proteinfamilywereshowntointeractwithmemberofthecytoskeletonsuchastubulin.

Theseresultsareconsistentwiththepivotalroleofneuralcircuitryrearrangementsunderlying

thestorageofnewinformationinthebrain(Dianaetal.,2007;Lamprecht,2011).

Wealsocomparedtheidentityofthese170differentiallyregulatedproteinsat12hafter

OCtoaproteindatasetobtainedfromthehippocampiofratsat12hafterSMT(Monopolietal.,

2011).Wehypothesizedthattherewouldbelittleoverlapbetweenthesesetsduetothedistinct

natureofthesememories.(Monopolietal.,2011)identifiedatotalof18differentiallyregulated

proteinsat12haftersubjectingratstoSMT.Albeitwehaveidentified17oftheseproteinsinthe

29

totalhippocampalproteome,weonlyobtainedquantitative information for11of themafter

statisticalfilters.Bycomparingthedatasets,weidentifyasingleregulatedproteinbeingshare

bythetwoassays,namelyprotein/nucleicaciddeglycaseDJ-1(Park7).Thisobservationsuggests

thatOC-basedassociativememoriespossessparticularmolecularsignatures,inspiteofthefact

that memory trace from different kinds of behavioral paradigms share common molecular

mechanisms.

B

A

RNMT

SRSF4

EIF4G3EIF4G1

HNRNPAB

HIST1H2BO

COL1A1MYSM1

PTMA

TPM4VPS26B

43345

HBA-A3

H2AFV

HIST1H1DEPN2

AP3S1

TGOLN2

NAP1L1

VPS29UBQLN2

ZC3H15

MAGED2CACYBP

RAD23BTPT1

MRPL1

YWHAHPRKCSH

YWHAZPTMS

ACTA2

DNAJA4 YWHAG

RCN1

HSP90AB1

GAP43

NRGN

EIF2S2

EIF3G

HSD17B12TECR

EIF4E

RPL32

GNB5

RALA

TUBB1

ELMOD1

DIRAS2

TXNRD2

SF3B4

SRSF1

SNW1

CRNKL1

PHF5A

CPT1A

SERPINA1

AGTRAP

SFXN3AGK

CHCHD3UQCC

ACP1PTPN5

DDT

PFKLACAA2

PKM

CALR

PDK3

TMED9

ARFGAP1

TMED7

FETUB

SCN1B

GCSH

CAMK2B

PARK7

USMG5ATP5J

MPC1

ATP5H

UQCRBNDUFB6

CHCHD3

ATP5H

PARK7

ATP5J

UQCRB

NDUFB6

USMG5

-0.4

-0.2

0.0

0.2

0.4

Mea

n Lo

g R

atio

(12

h/C

ontro

l) Cluster I

PHF5A

SRSF4

HNRNPAB

SRSF1

CRNKL1SNW1

SF3B4

-0.4

-0.2

0.0

0.2

0.4

0.6

Cluster II

DIRAS2

YWHAZ

YWHAG

YWHAHGNB5

RALA

-0.2

-0.1

0.0

0.1

0.2

0.3

Cluster III

DIRAS2

YWHAZ

YWHAG

YWHAH

TUBB1

HSP90AB1

TPM4

-0.4

-0.2

0.0

0.2

0.4

Mea

n Lo

g R

atio

(12

h/C

ontro

l) Cluster IV

EIF3GRPL32

MRPL1TPT1

EIF2S2

EIF4G3

EIF4G1EIF4E

-0.4-0.20.00.20.40.60.8

Cluster V

MYSM1

RAD23B

UBQLN2

AP3S1

TGOLN2EPN2

-0.5

0.0

0.5

1.0

Cluster VI

30

FIGURE9IPutativeProtein-ProteinAssociationsAmongSignificantlyRegulatedProteinsat12hAfterConditioning. (A) Protein network displaying possible protein-protein interactions among significantlyregulatedproteinsat12hafterOC.(B)Clustersdisplayingpossibleproteincomplexes(p<0.05)amongproteinsdifferentiallyexpressed12haftertraining.

IncreaseofDeNovoProteinSynthesisAfteraRecallSession

Comparedtothetwofirsttimepoints,theproteomeanalysisofoperantconditionedrats

subjected to a recall session displayed the largest number of proteins exhibiting significant

abundancechanges.We identifieda totalof275differentially regulatedproteinsat this time

interval,ofwhich60%wereup-regulated. These results strongly suggest that someof these

proteinsareassociatedwiththeprocessofretrievalitself,sinceearlierstudieshaveshownthat

themajorityofmoleculesaredown-regulatedat24hafterconditioning.(Monopolietal.,2011),

for instance, showed that thenumberofproteinsbeingup-regulatedat thispointdecreases

whencomparedtoearliertimeintervals.(M.M.Ryanetal.,2012), likewise,measuredmRNA

levels at 24 h after LTP induction in the hippocampi of rats and observed that 68% of the

transcriptsexhibitingsignificantabundancechangesweredown-regulated.

Togaininsightintotherolesoftheregulatedproteinsaftertherecallsession,weassigned

themGOtermsusingDAVID(D.W.Huangetal.,2007).Cellularcomponent(CC)analysisrevealed

enrichmentsforthetermscytoplasm,extracellularexosome,membraneandmitochondrion(p<

0.05; supplementary excel file 6). Proteins associatedwith axon, neuronprojection, PSD and

synaptic vesiclewerealsodetected in theCCanalysis,but theywere represented toa lower

degree. Notably, proteasome complex was also a term amongst the CC analysis. Proteins

classifiedunder this termaremembersof theubiquitinproteasome system,whichhasbeen

associatedwiththeprocessofmemoryreconsolidation(S.-H.Leeetal.,2008).UsingDAVID,we

also carried out enrichment analysis for functional categories (FC) and KEGG pathway.

Interestingly, the tophit termsofFCwerephosphorylation,acetylationandmethylation (p <

0.05;supplementaryexcelfile6).TheseresultsshowthatPTMsmaybeplayingimportantroles

in shaping neural networks after recall of an operant conditioning behavior. KEGG pathway

analysisgroupedsynaptotagmin1(Styt1),complexin-2(Cplx2)andras-relatedproteinRab-3A

(Rab3A)undersynapticvesiclecycle.Notably,alloftheseproteinsarelinkedwithexocytosisand

31

were shown tobeup-regulated (C.-C.Huanget al., 2011;McMahonet al., 1995; Tuckerand

Chapman,2002).

Next,weusedtheSTRINGdatabasetoinferpossibleprotein-proteininteractionsamong

the differentially regulated proteins after the recall session (Figure 10A). Subsequently, we

subjectedthisnetworktotheclusteringalgorithmofClusterONEtopredictproteincomplexes

(Figure 10B). ClusterONE analysis generated a total of 11 clusters (p < 0.05). Among these

putative protein complexes,we detectedmembers of the eukaryotic initiation factors (eIFs).

Notably, it has been known that eIFs participate in the initiation phase of de novo protein

synthesis by recruiting mRNAs or stabilizing the ribosomal pre-initiation complex, which is

comprisedofthe40Sribosomalsubunitandthemethionyl-transferRNA(Jacksonetal.,2010).

TranslationinitiationprocessisorchestratedbyalargerangeofeIFssuchaseIF2,eIF3

andeIF4.Interestingly,amongtheproteinswithsignificantabundancechangesaftertherecall

session,weidentifiedalargenumberofeIFsisoforms.Whilesomeofthesewereup-regulated,

othersweredown-regulated.Thisapparentdiscrepancyinabundancelevelscanbeaccounted

for by site specificity. For instance, some regions of the hippocampus may require protein

synthesisandinducesthesynthesisofspecificeIFs;inothersections,theoppositemayhappen.

Notably, among theup-regulatedproteins in this dataset,we identified the eIF2b. This is an

isoformofeIF2,whichhelpstoassemblyofthe43Spre-initiationcomplex.Theseobservations

providefurthersupporttothehypothesisthatdenovoproteinsynthesisistakingplaceafterthe

recallsessionandisrelatedtomemoryreconsolidation.

32

FIGURE 10 I Putative Protein-ProteinAssociationsAmong Significantly Regulated ProteinsAfter theRecallSession.(A)Proteinnetworkexhibitingputativeprotein-proteininteractionsamongsignificantly

B

ATP6V1E1 ATP6V0A1

RGD1566320 RCG_38845

TECR

ATP6V1G1

RPL32

NAP1L1

HSD17B12

SCN1BABCE1

EIF3H

HIST1H2BO

SYT1

RAB3A

EIF2S2

MAP1AEIF4E

MAP2 FBXL20

PSMC4GAP43

MYSM1

YWHAB

PSMD6

TRA2B

SMIM26

MLST8

CIRBPCACYBP

SKP1

CDK18 HMGB1 CAMK2BYWHAZ

MRPL1MBP

YWHAESYT2

EIF4G3PSMA6EIF3E UBQLN2G3BP1 SH3GL2

EPN2CLTB

TIMM10CPLX2TPT1ATP5J

CALM1 USMG5

BIN1

PLAAATXN2ADARRANBP1

ND2

NDUFB5

NDUFB6

DDT

CHCHD3

NDUFA13 MPC1

TMCC1

MCCC1 GCSH

ACSS2

ACAA2

ACYP1

ACADVL

ACSS1

PC

PFKM

PFKP

PFKL

CLK2

DNAJA4SGTA

PKM

YARS

ITPA

NME1

GRM1

STX8

CALR

MPST

ABCF2HTRA2

TIMM9

TXNRD2

LDHB

ALDH18A1

PARK7

VDAC2

MPP5

NDUFS4

NDUFS8

ATP5H

UQCRBCOX5B

NDUFA10

ECHS1

HBA-A3

HBA2

ACSBG1

HBB

CSNK1E

RAP2B

EXOC5 YWHAG

RALAGNB5

YWHAH

HIST1H1B

HIST1H1D

STRN4

PFDN6PHF5A

SRSF3

RNMT

HNRNPF

SF3B4

IDE

HNRNPAB

SF3A3

PFDN4NUP205

CANXLAMP1

SNX6VPS26B

RAB7A

CDC42EP4VPS29

LMAN2

PTMS

PTMA

MAP1LC3ATPM4

RAB5CTUBB1

PPP1R1B

RAB2B

TUBB2B

43345SERPINA1

FETUB

STRN

TMED9

ARL1

TMED7

CD59

SNX12

GOLGB1

RAB18

RTN3

A

ATP5H

CHCHD3

NDUFB5

PARK7

ATP5J

COX5B

NDUFS8

NDUFA10

NDUFS4

USMG5

UQCRB

NDUFB6

-0.4

-0.2

0.0

0.2

0.4

Mea

n Lo

g R

atio

(Rec

all/C

ontro

l)

Cluster A

CIRBP

SRSF3

PHF5A

SF3A3

HNRNPAB

HNRNPF

TRA2B

SF3B4

-0.4

-0.2

0.0

0.2

0.4

Cluster B

YWHAE

CSNK1E

RAB2B

YWHAB

YWHAZ

RAP2B

YWHAGGNB5

SEPT2RALA

EXOC5SKP1

MLST8

-0.4

-0.2

0.0

0.2

0.4

Cluster C

ALDH18A1

ECHS1

ACSS1

MCCC1

ACSBG1

ACADVL

ACYP1

ACSS2PC

ACAA2

-0.4

-0.2

0.0

0.2

0.4

Cluster D

BIN1

SH3GL2STX8

CPLX2

UBQLN2CLT

BEPN2

RAB3ASYT1

-1.5

-1.0

-0.5

0.0

0.5

Mea

n Lo

g R

atio

(Rec

all/C

ontro

l)

Cluster E

BIN1

SH3GL2

UBQLN2CLT

BEPN2

SYT1

-1.5

-1.0

-0.5

0.0

0.5

Cluster F

G3BP1YA

RS

ABCF2

ABCE1EIF3H

EIF2S2

RPL32EIF3E

EIF4E

EIF4G3

-0.4

-0.2

0.0

0.2

0.4

Cluster G

LDHBNME1 PC

PFKPPFKM

PFKLPKM

ITPAMPST

-0.2

-0.1

0.0

0.1

0.2

Cluster H

LDHBGCSH

TXNRD2 PCPFKP

PFKMPFKL

PKMMPST

-0.20.00.20.40.60.81.0

Mea

n Lo

g R

atio

(Rec

all/C

ontro

l)

Cluster I

NME1PFKP

PFKMPFKL

PKMITPA

-0.2

-0.1

0.0

0.1

0.2

Cluster J

G3BP1

ABCE1EIF3H

EIF2S2

RPL32TPT1

EIF3EEIF4E

EIF4G3

RNMT

-0.4

-0.2

0.0

0.2

0.4

0.6

Cluster K

33

regulated proteins after the recall session that took place 24 h after training. (B) Clusters displayingpossibleproteincomplexes(p<0.05)amongproteinsdifferentiallyexpressedafterbehavioralretrieval.

QuantitativeHippocampalPhosphoproteome

Phosphorylationplaysasignificantroleintheprocessofmemoryconsolidationandmay

alsoberelevanttoplasticitymechanismsunderlyingbehaviorretrieval,sincethisPTMwasthe

topenrichmenttermforcellularcomponentinboththe12handtherecallproteindatasets.

Nevertheless, no large-scale proteomic assay has been performed to understand the role of

phosphorylationduringmemory formationand retrieval. Toaddress thisquestion inoperant

conditionedrats,wecombinediTRAQlabeledpeptidesoncemoreandenrichedforphospho-

modified analytes using TiO2 beads, a step required prior to LC-MS/MS analysis due to the

substoichiometriclevelsofPTMsincells.

Using a single-shot strategy, we identified 568 phosphoproteins and extracted iTRAQ

intensitiesofthereportergroupstoobtainquantitativeinformationforthesephosphopeptides

(supplementary excel file 7). After correction for protein abundance, statistical analysis in R

showedthat53phosphopeptidesexhibitedsignificantabundancechangesthroughoutthethree

timepointsunderinvestigationhere(p<0.05;supplementaryexcelfile8).Interestingly,wedid

not detect any significantly regulated phosphoprotein at 30min after training, although we

showed that CaMKII was up-regulated at this time point. Conversely, 4 phosphopeptides

displayed different abundance changes at 12 hwhen comparedwith the control individuals.

Among these, we have detected the up-regulation of EIF3G phosphorylation. Notably,

(Martineauet al., 2014) demonstrated that EIF3Gphosphorylation is required for translation

initiation, which is consistent with the notion that hippocampi cells of conditioned rats are

synthesizingnewproteinsatlaterstagesofmemoryconsolidation.

Analysisofthephosphoproteomeofratssubjectedtoarecallsessiondemonstratedthe

differential regulation 49phosphopeptides. Surprisingly, 80%of thesephosphoproteinswere

down-regulated.To investigate the roleof theseproteins in rodentcells,weused theDAVID

database to assign terms for cellular component (CC) and functional category (FC) (p < 0.05;

supplementaryexcelfile9).AmongtheoverrepresentedtermsforCCwerepost-synapticdensity,

synapse,dendriteandneuralcellbody.Theseresultsre-inforcethatsuchlocationsareunder-

34

regulationaftermemoryretrieve.Conversely,thetopenrichmenttermsforfunctionalcategories

were phosphorylation, acetylation, alternative splicing, cytoskeleton, cell projection and

intermediate filament.ThisshowsthatnotonlyPTMsare important in this time interval,but

rearrangementofthecytoskeletonseemstoplayavitalroleuponrecall.

RT-qPCRMeasurementsofmRNAs

Recently,anumberofstudieshavesoughttoaddressthequestionofhowmRNAlevels

arerelatedtotheirproteincounterpartsindistinctbiologicalsystems.Thesestudiesfoundahigh

similaritybetween transcript andproteinabundances in steady-state conditions,whilehighly

dynamic cellular states exhibited large differences (Liu et al., 2016). To investigate if mRNA

abundancemirrorstheirproteincounterpartsduringmemoryconsolidationandafterrecall,we

selectedatotalof5targets(supplementaryexcelfileprimers)andmeasuredtheirconcentration

using real-timequantitativepolymerase chain reaction (RT-qPCR).Gria1 andCamklla protein

counterpartsdidnotexhibitsignificantabundancechangesatnoneofthetimeintervalunder

studyhere,buttheyhavebeenshowntoberegulatedinotherbehavioralparadigmssuchasfear

conditioning and spatialmemory tasks (M. C. Frey et al., 2009; Lisman et al., 2002). RAB3A,

YWHABand YWHAG, contrary, displayed significant abundance changes at 12 h and/or after

recall(supplementaryexcelfile3).

mRNAmeasurementsestablishedthatthelevelsofonlyonetargetedtranscript,namely

Gria1,wassignificantlyaltered(one-wayANOVA,p<0.05;Figure11).ThemRNAofthisanalyte

wasup-regulatedonlyat12h,althoughproteomicresultsshowedthattheproteincounterpart

ofGria1didnotexhibitedsignificantchangesinabundanceatanyofthemeasuredtimepoints.

Notably,theseRT-qPCRresultsareinagreementwithotherlarge-scalequantitativeproteomic

and transcriptomic studies (Cheng et al., 2016; R. Lu et al., 2009). In part, this low

interdependencecanbeaccounted forbydifferential translationratesamongmRNAspecies,

post-transcriptionalregulationofgeneexpressionandproteintransport(Liuetal.,2016).Taken

together,theseresultshighlightthemolecularcomplexityunderlyingbehavioralassaysandthe

needtodissectmoleculareventsunderlyingmemoryatmultiplelevels.

35

FIGURE11IRT-qPCRMeasurementsofmRNAs.Weselectedatotaloffivetranscriptsandmeasuredtheirconcentrationat30min,12handaftera recall session that tookplace24hafter training.Gria1 andCamkIIaproteincounterpartsdidnotexhibitsignificantabundancechangesthroughtheaforementionedtimeintervals,yetGria1wasup-regulatedat12hafterconditioning(one-wayANOVA,p<0.05).RAB3A,YWHABandYWHAGproteinsweredifferentiallyregulatedatleastinoneofthemeasuredtimepoints;however,theirtranscriptswerenot.Hence,RT-qPCRmeasurementsoftheselectedmRNAsshowalowcorrelationwiththeirproteincounterparts.

Control

30 m

in12

h

Recall

0.00

0.05

0.10

0.15

log

[ΔC

t]Gria1

p < 0.05

Control

30 m

in12

h

Recall

0.00

0.05

0.10

0.15

log

[ΔC

t]

Camkiiα

Control

30 m

in12

h

Recall

0.00

0.02

0.04

0.06

0.08

0.10

log

[ΔC

t]

Rab3a

Control

30 m

in12

h

Recall

0.00

0.02

0.04

0.06

0.08

0.10

log

[ΔC

t]

Ywhab

Control

30 m

in12

h

Recall

0.00

0.02

0.04

0.06

0.08

0.10

log

[ΔC

t]Ywhag

36

Discussion

Large-scale transcriptomic and proteomic studies have been carried out in animals

subjectedtobehavioralparadigmstounderstandtheprocessofmemoryconsolidation inthe

hippocampusandotherregionsofthebrain.Forinstance,(Cavallaroetal.,2002)exposedrats

toconsecutivetrainingsessionsinthewatermazeparadigm,atypeofspatialmemorytask(SMT)

andmeasuredmRNA levels at threedifferent timepoints after it. (Rao-Ruiz et al., 2015), by

contrast,quantifiedhippocampalmembraneproteinsoffearconditioned(FC)miceat1hand4

haftersubjectingtheseanimalstotwodistinctbehavioralprotocols:afootshockwasdelivered

immediatelyuponplacingtheanimalsintheboxorwasdelayedby3min.Theydidnotobserve

proteome changes at 1 h in neither of these experimental groups. At 4 h, however, they

identified 164 proteins exhibiting significant abundance changes in animals subjected to the

delayedshockprotocoland273proteinsintheimmediateshockgroup.

Interestingly, pharmacological, electrophysiological and genetic approaches have

revealedthatdifferentformsofbehavioralparadigmssharetheactivationofsimilarmolecular

andcellulareventsunderlyingtheretentionofnewlyacquiredinformation.Forexample,operant

andfearconditioninginducetheactivationofproteinkinaseA(PKA)and,initsturn,theincrease

ofCREBphosphorylation(Lorenzettietal.,2008;McGaugh,2000).Asitwasmentionedearlier,

CREB isatranscriptionfactorknowntoactivatea largenumberofgenes involved inmemory

consolidation.Nevertheless,globaltranscriptandproteinabundancechangesvarysignificantly

among thesedifferent typesofmemories. In their assay, for instance, (Rao-Ruizetal., 2015)

demonstratedthatlessthan20%ofalldifferentiallyregulatedmembranehippocampalproteins

identifiedat4hafterconditioningweresharebetweentheimmediateanddelayedfootshock

groups.

Toidentifyproteinsresponsiblefortheconsolidationofothertypesofmemoriesbesides

aversive and spatial memories, we performed the first temporal quantitative hippocampal

proteomeandphosphoproteomeanalysesofoperantconditioned rats.Afterestablishing the

superior in-depth coverage of the SAX fractionation strategy, we identified a total of 8,951

proteinswith at least one unique peptide.We showed that the levels of 465 proteinswere

significantlyalteredatthreedistincttimepoints–30min,12handafterarecallsessionthat

37

took place 24 h after training. Furthermore, enrichment of phosphopeptides using titanium

dioxide (TiO2) beads followedby single-shot analyses in a high-resolutionmass spectrometer

enabledtheidentificationof568phosphoproteins.Ofthese,53exhibitedsignificantabundance

changesduringtheaforementionedtimeintervals.

At30minafterconditioning,weidentifiedatotalof20regulatedproteins.Theseresults

supportthehypothesisofanearlytimesensitivewindowforproteinsynthesisduringmemory

consolidation (Bourtchouladze et al., 1998). In addition, they are consistent with previous

experiments which showed that pharmacological inhibition of protein synthesis prior to or

immediatelyaftertrainingcausesdeficitsinmemoryformationinrodents(Bourtchouladzeetal.,

1998).Furtherendorsementtothisnotioncomesfromtheidentificationoftheup-regulationof

asubunitoftheeukaryoticinitiationfactor(eIF),namelyeukaryotictranslationinitiationfactor

3subunitG(eIF3g)(Zhouetal.,2008),atthisearlytime-point.EIFsfacilitatethesynthesisof

newproteinsincellsandtheup-regulationofdifferentsubunitshavebeenlinkedwithsynaptic

plasticitymechanismunderlyingtheprocessofmemoryconsolidation(KlannandDever,2004).

Inparticular,ithasbeendemonstratedthatinactivationofEif3gleadstoproblemsintheprocess

ofscanningforthestartcodonofmRNAsequences(Aitkenetal.,2016).

Surprisingly,wedidnotidentifyanydifferentiallyregulatedphosphopeptidesduringthe

earliestmeasuredtime-point.Theabsenceofsignificantalterationsinphosphoproteinat30min

afterconditioningcouldhaveresultedfromacombinationoftwofactors.First,themethodused

toisolatephosphopeptidescouldhavedepletedspeciesthatunderwentabundancechanges.For

instance, (Bodenmiller et al., 2007) employed three different techniques to isolate

phosphopeptides – phosphoramidate chemistry, immobilized metal affinity chromatography

(IMAC)andTiO2beads–fromDrosophilamelanogasterKc167cellsandanalyzedthesamplesin

a mass spectrometer. They found that each isolation strategy enriches for a specific set of

phosphopeptides.Accordingly,datasetsgeneratedfromthesemethodsshowedaweakoverlap

(Bodenmilleretal.,2007).Second, it ispossible thatphosphoprotein levelsmaynotundergo

significantalterationsat30minafterconditioning,butatearlierorlatertimepoints.Thiswould

occurif,forinstance,theactivationofproteinkinaseshappenedatdifferenttimepointsduring

theconsolidationofnewlyacquiredmemories(Izquierdoetal.,2006).

38

Proteinkinaseshavebeenshowntoplayasingularroleinmemory.DuringLTPplasticity

mechanisms, for instance, depolarization of CA1 cells of the hippocampus through tetanic

stimulationleadstotheinfluxofcalciumandactivationofCaMKII,aproteinkinasethatplaysan

importantroleinLTP.Interestingly,repeatedstimulationofCA3cellsleadstotheactivationof

PKA in the CA1 post-synaptic neurons by the increase in cyclic adenosine monophosphate

(cAMP).ActivatedPKAturnsonmitogenactivatedproteinkinase(MAPK),whichtranslocatesto

the nucleus and phosphorylates CREB (Pontes and de Sousa, 2016). Furthermore, repeated

tetanicstimulationalso leadstotheactivationofproteinkinaseCzeta.Thisproteinkinase is

requiredforthepersistenceofmemory,butnotforitsinitialinduction(Tsokasetal.,2016).

Among the up-regulated proteins at 30 min after conditioning, we identified a fold-

changeincreaseof0.2intheb-subunitofCaMKII.Interestingly,previousFCorSMTsbehavioral

paradigm studieshavedetectedmajor changes in theexpressionof thea-subunit of CaMKII

(AhmedandJ.U.Frey,2005b). CaMKIIa-andb-subunitsarethemostabundant isoformsof

CaMKIIinthebrainofvertebrates(Lismanetal.,2002).Wequantifiedthea-subunitatalltime

points, but its abundance did not change significantly during the course of the experiment.

Consistent with these results, quantification of CaMKIIa transcript was not altered when

compared with the control group. By contrast, the protein levels of the CaMKIIb increased

significantlyasearlyas30minafterconditioning.Theseresultssuggestthat–inadditiontoglobal

scalevariationsintheproteome–differentmemorytypesalsoexhibitdifferencesinconserved

signalingmechanismsthataresharedamongthem.

Analysisofproteinchangesat12hafterexposingratstoOCrevealedthat170proteins

exhibited significant abundance changes between subject and control animals. We grouped

these proteins into three different clusters that display the temporal behavior of the

hippocampalproteomeduringmemoryformationandafterrecall(Figure8;supplementaryexcel

file 4). Among the 107 up-regulated proteins, we found the isoforms YWHAG, YWHAH and

YWHAZ, which are members of the 14-3-3 protein family. In mammals, this protein family

possessesatotalofsevenisoformsandaccountfor1%ofallproteinsexpressedinthebrain.

Notably,membersofthe14-3-3proteinfamilyhavebeenlinkedtoavarietyofcellularprocesses

suchascellcycle,intracellulartrafficking,transcriptionalcontrolandsignaltransduction(Berget

39

al.,2003).Inaddition,previousstudiesshowedthattheyinteractwithalargearrayofproteins,

rangingfromenzymesandproteinreceptorstocytoskeletalandstructuralproteins(Kentetal.,

2010).Moreover,ithasbeendemonstratedthattheseproteinsplayacrucialroleduringmemory

formation(Qiaoetal.,2014).

Initial evidence for the role of 14-3-3 protein family isoforms inmemory came from

studiesinDrosophilafliescontainingmutationsinthegeneleonardo(leo),whichencodedone

of the isoformsof thisprotein family. Interestingly, leo flies exhibitedmemorydeficits anda

reduction in basal synaptic transmission when they were subjected to an olfactory learning

paradigm(SkoulakisandDavis,1996).Inmammals,rodentshavebeenusedtostudyalterations

in memory formation associated with these proteins. (Qiao et al., 2014), for example, used

transgenicmiceexpressinganinhibitorofthe14-3-3proteinfamilyisoformsintheneuronsof

theCA3-CA1regionsofthehippocampustostudyalterationsinmemoryformationwithinaFC

context. They showed that expression of this protein inhibitor lead to deficits in memory

consolidationwhenanimalswereretested24haftertraining.Consistentwiththeinvolvement

ofthe14-3-3proteinfamilyinmediatingsynaptictransmissionwithincells,(Qiaoetal.,2014)

establishedthatLTPinductioninCA1cellsexpressingtheinhibitorwaseliminated1hafterhigh-

frequencystimulation.Thisstudy,however,wasnotabletopinpointtheexactcontributionof

eachisoformtotheprocessofmemoryformation,sincetheinhibitortargetedallisoformsofthis

proteinfamily.

Thequantitativetemporalanalysisofthehippocampalproteomeofoperantconditioned

rats carriedouthere showed that the isoformsof the14-3-3protein family are regulatedat

distincttimeintervals,whichsuggeststhattheseisoformicspeciesplaydifferentrolesduringthe

processofmemoryconsolidationandafterrecall.Weestablishedthatafold-changeincreaseof

0.23oftheisoformYWHAGwasnotaccompaniedbyanup-regulationofinitsmRNAcounterpart,

suggesting that the expression of this protein is regulated post-transcriptionally. Interaction

networkscontainingregulatedproteinsat12hafterconditioningindicatedthatYWHAG,YWHAH

andYWHAZwereassociatedwith structuralor cytoskeletalproteins suchasmembersof the

tubulinfamily(Figure9B).Theseresultsendorsepreviousstudiesandsuggestthatthisprotein

40

familyisinvolvedwiththemorphologicalchangeswithinhippocampalcellsthatarelinkedwith

theretentionofinformationatlatertimepoints(Bergetal.,2003;Qiaoetal.,2014).

Interestingly, recentobservationshave ledtotheconceptionthatproteindegradation

mightbeasimportantasproteinsynthesisfortheprocessofmemoryformation.Accordingwith

this emerging hypothesis, the breakdown of specific groups of proteins withdraws possible

inhibitoryconstraintsonmemoryconsolidation(S.-H.Leeetal.,2008).Thedegradationofthese

proteinscanoccurviathreemainproteolyticroutes–actionofproteasessuchasneurotrypsin,

lysosomeortheubiquitin-proteasomesystem(UPS),whichtagsproteinsfordegradationthrough

thetransferofubiquitinto lysineresidues(FioravanteandByrne,2011;Muelleretal.,2015).

Notably,ourdataanalysisrevealedthat37%oftheproteinswithsignificantabundancechanges

at12hafterconditioningweredown-regulated.Yet,wedidnotdetecttheup-regulationofany

subunitsoftheseproteindegradationcomplexes,suggestingthattheyarecontrolledatthelevels

ofactivityand/orthattheyaretemporallyexpressedatdifferenttimeintervals(X.Wangetal.,

2013).

Onthisnote,proteindegradationhasalsobeenimplicatedinmemoryreconsolidation–

aprocessbywhichinformationbecomeliabletochangesuponbehavioralretrieval.(S.-H.Leeet

al.,2008)investigatedtheroleoftheUPSinmemoryreconsolidation.Theyinfusedb-lactone(i.e

aproteasomeinhibitor),ANIorbothbilaterallyinthehippocampusoffearconditionedmiceafter

a recall session. Re-testing animals 24 h later revealed that while ANI infused mice had a

significantdecrease in fear levels,administrationofb-lactonealonedidnotaffect fear levels.

Surprisingly,co-administrationofb-lactonereversedtheamnesiceffectsofANI.(S.-H.Leeetal.,

2008) concluded that the proteasome inhibitor destabilized the reactivated memories, yet

protein synthesis is still required for the re-stabilization of consolidated memory. That is,

consolidatedmemories couldbe reconfiguredduring recallsbydegradationandnewprotein

synthesis.

Inorder toaccess the roleofproteindegradationduringmemory retrieval inoperant

conditionedrats,weperformedcellularcomponentanalysisoftheregulatedproteinsafterthe

recallsession.Resultsrevealedanenrichmentforproteinsassociatedwithubiquitinproteasome

system. Notably, some ubiquitin proteasome system components such as proteasome 26S

41

subunit, non-ATPase, 6 (Psmd6) and ubiquitin-like modifier-activating enzyme 2 (Uba2) and

proteasomesubunit alpha type-6 (Psma6)hada fold-changedecreaseof0.13,0.23and0.15

respectively, while other proteins like ubiquitin-conjugating enzyme E2 N (Ube2n) and 26S

proteasomeregulatorysubunit6B(Psmc4)wereup-regulated.Aplausibleexplanationforthese

observations is thatwithinneuralnetworks inthehippocampus,differentcomponentsofthe

proteasomehavedistinctspecificitiesandactivitiesforsubstrates(X.Wangetal.,2013).

Recently, (Jaromeet al., 2016) demonstrated that CaMKII is critical for an increase in

proteasomeactivity in theamygdala followingmemoryrecall.First, theyshowedthatCaMKII

inhibitioninvitroreversedtheincreaseofproteasomeactivityinanimalssubjectedtoarecall

sessionafterFCtraining.Second,usinginvivoblockageofthiskinase,theyabolishedproteolytic

activityfollowingrecall.TheseresultsuncoveredanotherfunctionofCaMKII inshapingnewly

acquiredexternalinformation–controlofproteasomeactivityduringmemoryreconsolidation.

Here,wehavealsodetectedafold-changeincreaseof0.17ofCAMKIIbafterbehaviorrecall,as

well as an increase in proteins that are a part of the proteasome system such as the ones

mentionedabove.Theseobservationsreinforceourclaimthatmemoryreconsolidationistaking

placeinthehippocampusofoperantconditionedrats.

Notably,theaforementionedresultsstandincontrasttoearlierstudies,whichdidnot

detect clear evidence for memory reconsolidation in operant conditioned animals (Exton-

McGuinnessetal.,2014;HernandezandKelley,2004;TronsonandTaylor,2007).Theapparent

absenceofmemoryreconsolidationcouldhaveresultedfromtheadoptedtrainingprotocol–

subjecting rodents to 10 consecutive training sessions prior to retrieval might havemasked

processesassociatedwithmemory reconsolidation. Tobemoreprecise, several studieshave

shown that hippocampal dependent memories last for a couple of days. Subsequently, this

information is transferred to regionsof thecortex for long-termstorage (Lesburguèresetal.,

2011).Consistentwiththisview, (MilekicandAlberini,2002)trainedrats inFCandsubjected

theseanimalstorecallsessionsatdifferenttimepointsafterconditioning–2,7,14and28days

later. They found that animals became vulnerable to memory disruptions through the

administrations of protein inhibitors only at early time points, showing that the process of

reconsolidationcanonlyoccurinnewlyacquiredmemories.Theresultsof(MilekicandAlberini,

42

2002)supportbehavioralassaysinwhichmemoryreconsolidationwasdetected–inthemajority

ofthesestudies,animalsweresubjectedtoatmost3trainingsessions.Takingtheseobservations

into perspective, we reason that the inability to detect memory reconsolidation in operant

conditioned animals in early studies is a result of the training protocols performed in these

previousexperiments(HernandezandKelley,2004;TronsonandTaylor,2007).

Aftertherecallsession,wealsoidentifiedafold-changeincreaseof0.4inthelevelsof

syntaxin8(Stx8).ThisproteinisamemberoftheQ-solubleN-ethylmaleimide-sensitivefactor

attachment protein receptor (Q-SNARE) family that mediates vesicle trafficking inside cells

(Hong,2005).KnockdownandoverexpressionassaysshowedthatStx8facilitatestropomyosin

receptorkinaseA(TrKA)transportfromtheGolgitotheplasmamembrane(B.Chenetal.,2014).

ActivatedTrkAinducesdownstreamsignaltransductioncascadesincludingphosphoinositide-3–

protein kinase B/Akt (PI3K-PKB/Akt) and Ras/mitogen-activated protein kinase (MAPK)

pathways,whicharebothinvolvedinLTPmechanismsofsynapticplasticityinthehippocampus

(Giovanninietal.,2001;Suietal.,2008).Interestingly,(Cavallaroetal.,2002)showedthat, in

STMsubjectedrats,Stx8isup-regulatedat1hafterthelasttrainingsession,butnotat12hnor

24haftertraining.Thisindicatesthatthisproteinmightbeinvolvedinmemoryrecallprocesses

andnotintheformationofnewmemories.Giventhat(Cavallaroetal.,2002)subjectedanimals

to 3 training sessions, their measurements of Stx8 transcripts might be associated with the

functionofStx8inmemoryrecallevents.

Insummary,weprovidethefirsttemporallarge-scaleproteomeandphosphoproteome

datasetofratssubjectedtoanoperantconditioningtaskbyemployingSAXasapre-fractionation

strategypriortoLC-MS/MS.Werevealedtheidentifyofproteinsdifferentiallyregulatedasearly

as30minaftertraining, includingtheb-subunitCaMKII.At latertimepoints,weshowedthat

proteins associated with the regulation of the cytoskeleton such as members of the 14-3-3

protein family are important in the rearrangement of neural circuitries in the hippocampus.

Interestingly, regulation of these proteins was not accompanied by changes in their mRNA

counterparts,revelinga lowinterdependencebetweenthesemacromolecules.Moreover,the

identificationofdifferentiallyregulatedproteinsoftheubiquitin-proteasomesystem(UPS)after

a recall session provides strong evidence for memory reconsolidation in OC, wherein we

43

hypothesizedthatthisprocesshasnotbeenfullyobservedearlierduetothebehavioralprotocols

usuallyusedinconditioninganimals.Wehopethattheproteomedatasetofoperantconditioned

ratspresentedherewillhelptobridgethegap intheunderstandingofthedifferentmemory

types,aswellasdisease-associatedphenotypes.

44

MaterialsandMethods

EthicalCommittee

AllexperimentalproceduresperformedherewereapprovedbytheEthicsCommitteeof

theInstituteofBiologicalSciencesoftheUniversityofBrasilia(documentidentificationnumber:

23106.075295/2016-61).

AnimalsandBehavioralProtocol

Three-month-oldWistarfemalerats(n=12)weremaintainedina12/12hlight/darkcycle

withwaterandfoodadlibitum,beinghandledeverydayfortwoweekspriortoanyexperimental

procedure.Animalswerethenwaterrestrictedfor18hthroughouttheexperiment.Theirbody

weightwascheckeddaily toensure itdidnotdropbelow80%of its initialweight.Following

handling,waterrestrictedanimalswereplacedinastandardSkinnerbox(InsightLtda,SãoPaulo,

Brazil)fortwodaysofhabituation.Inthese20minhabituationsessions,animalswereputinthe

boxandwaterwasreleasedat randomfromtimetotime.Habituatedfemaleratswerethen

returnedtotheSkinnerboxandmanuallytrainedinasinglesession.Thetrainingconsistedon

releasingwaterasanimalsgotprogressivelyclosertothelever.Eventually,theserats learned

that pressing the lever in the box would result in water release. The criterion for behavior

acquisitionwassetat15leverpressesinarow.Moreover,animals(n=3)oftherecallgroup

wereplacedonceagainintheboxonedayaftertrainingtoretrievethebehavior.Thecriterion

intherecallsessionwassetat50leverpressesinarow.Femalerats(n=3)ofthecontrolgroup

didnotundergotraining,butwentthroughthehabituationstepsdescribedearlierandwerealso

placedintheboxfor20mininthedayofthetrainingsession.

TissueDissection

Operantconditionedfemaleratswereeuthanizedat30min,12hand30minafterarecall

sessionthattookplace24hafterconditioning.Controlanimalswereeuthanized30minafter

leavingtheSkinnerbox.Theirhippocampiwereanatomicallydissectedunder3min,washedin

iced-coldphosphate-bufferedsaline(PBS)andflash-frozeninliquidnitrogen.Next,thosefrozen

tissues were ground using a pestle and separated in two distinctmicrotubes. This step was

45

performedtoseparatesamplesfortheproteomicandRT-qPCRassays,whichrequiredifferent

reagentsfortheextractionofproteinsandmRNAs,respectively.

Filter-aidedSampleDigestion

The digestion of proteins into peptides was performed as previously described in

(Wisniewskietal.,2009),withsomemodifications.Briefly,groundfrozenhippocampalsamples

were homogenized in 0.3 mL of 4% (w/v) sodium dodecyl sulfate (SDS), 0.02 M

triethylammoniumbicarbonate(TEAB),0.1Mdithiothreitol(DTT),phosphataseinhibitorcocktail

(Roche)andpH7.9.Tubeswerethenincubatedfor10minat90˚CandsonicatedtoshearDNA

and RNA molecules which might affect digestion efficiency. Subsequently, samples were

centrifugedat16,000gfor15minatroomtemperature.Thesupernatantwassavedandprotein

concentrationwasmeasuredinallconditionsusingtheQubitProteinAssayKit(ThermoFisher

Scientific).

Aliquotscontaining100µgoflysedproteinsweremixedwith0.2mLofureabuffer(UB)

–8Murea,0.02MTEABandpH8.5–inVivacon500(Sartorius)withanominalcutoffof30KDa.

In order to perform the comparison between SCX and SAX pre-fractionation strategies, we

digested500µgofproteinlysatesandmadeadjustmentstothevolumeofthebuffersolutions

ofthestepsdescribedbelow.Vivacondeviceswerethencentrifugedat10,000gfor15minat20

˚C – all the following centrifugations were performed at this temperature, unless otherwise

stated.Sampleswerewashedwith0.2mLofUBandcentrifuged.Subsequently,0.1mLofUB

containingiodoacetamidetoafinalconcentrationof0.05Mwasaddedtothesamplesandthe

Vivacon deviceswere incubated for 20min in the dark. Samples in the filtration unitswere

centrifuged,washedwith0.2mLofUBandre-centrifuged.ConcentratedsamplesintheVivacon

deviceswerebufferexchangedbywashingtwicewith0.1mLofdigestionbuffer (DB;0.02M

TEABandpH7.9)asdescribedabove. These sampleswere then trypsin-digested inDBwith

trypsin for 12 h at 37 ˚C with a concentration ratio of 1:100 (protease:sample). Following

proteolyticdigestion,thesampleswerecollectedbycentrifugation.Thefilterunitswererinsed

withDBandcentrifugedonemoretime.Theproteolyticreactionwashaltedbytheadditionof

46

trifluoroaceticacid(TFA)tothesamplestoafinalconcentrationof0.5%.Finally,theyieldofeach

digestionwasmeasuredwiththeQubitProteinAssayKit.

LabelingtheSampleswithMultiplexReagents

Aliquotsof50µgofeachreplicatewasvacuumdriedinaSpeedvacandthelabelingof

theconditionswascarriedoutwiththeiTRAQreagentkit(ABSciex,Framingham,MA)according

withthemanufacturer’sinstructions.Thelabelingwasperformedassuch–control(tag114),30

min(tag115),12h(tag116),recall(tag117).Thefourisobarictagsofeachbiologicalreplicate

weremixedinaratioof1:1:1:1,desaltedandanalyzedina90-mingradientviaLC-MS/MSto

verifytheefficiencyofthelabelingreactionsandtheratiobetweenthedifferenttags.

OfflineAnionExchangeBasedFractionationwithSaltElution

Offline pre-fractionation of peptideswas performed accordingwith (Rappsilber et al.,

2007), with some minor modifications. Briefly, strong anion exchange (SAX) stop-and-go

extractiontips(StageTips)weremadein-housebycutting3MEmporeAnionExchangediskswith

acutterandstackingthemembranesintoa200µLpipettetip.Aplasticringwassetaroundthe

StageTipsandtheywerepositionedinto2mLmicrotubes.Themembraneswereactivatedby

loading the pipette tips with 0.1 mL of 100% methanol and centrifuging them – all the

centrifugationssetat1,000gandtheentireprocedurewasperformedatroomtemperature.

Activatedmembranesweresubjectedtoanumberofwashesby loadingandcentrifugingthe

tips. First, they were washed with 0.1 mL of 0.1% ammonium hydroxide (NH4OH) and 80%

acetonitrile(ACN);second,0.1mLof0.1%NH4OH;andthird,0.1mLof20%ACN,0.1%NH4OH

and 2 M ammonium acetate (NH4AcO). Finally, the membranes were washed twice with a

solutionof0.1%NH4OH.Oncethosestepswereperformed,thesampleswereloadedintothe

StageTips.

Thefourdifferentchemicallylabeledconditionsweremixedagaininaratioof1:1:1:1to

atotalof30µgofsampleperbiologicalreplicate–intheexperimentdonetobenchmarkthe

two pre-fractionation strategies, aliquots of 30 µg were also used, yet the samples were

unlabeledandcamefromthesampleproteolyticreaction.Thepeptideswerethenvacuumdried

47

inaSpeedvac,resuspendedin0.1mLof0.5%NH4OHandloadedintotheStageTips.Thepipette

tipswerecentrifugedandtheflowthroughwascollectedinacleanmicrotube.Thefractionswere

elutedwith 0.03mL solutions of 20% ACN, 0.1%NH4OH and an increasing concentration of

NH4AcO–0.015to3M.Finally,the12fractionsofeachreplicate,includingtheflowthrough,

weredesaltedandanalyzedbyLC-MS/MS.IntheexperimenttobenchmarkSAXwithSCX,only7

fractionswereanalyzedperreplicate.

OfflineCationExchangeBasedFractionationwithSaltElution

Offline strong cation exchange (SCX) based fractionation was done with 30 µg of

proteolyticdigestpertechnicalreplicate.Thesesamplescamefromthesameenzymaticreaction

usedintheSAXfractionationexperiments–thiswasdoneintentionallytodecreaseanyvariation

arisingfromdistinctproteolyticcleavagereactions.Thefractionationwasperformedaccording

to(Rappsilberetal.,2007),withminormodifications.StageTipsweremadein-houseasdescribed

above,with3MEmporeCationExchangedisks,andall thecentrifugationwerecarriedoutat

1,000gatroomtemperature.

Inordertoactivatethemembranes,weadded0.1mLof100%ethanoltothetipsand

centrifugedthem.Next,weplacedan0.1mLsolutionof0.5%aceticacid(AcOH)and80%ACN,

followedbyacentrifugation.Subsequently, themembraneswereequilibratedwith0.1mLof

0.5%aceticacidsolution.Weadded0.1mLof20%ACN,0.5%AcOHand2MNH4AcOtothetips

andcentrifugedthem.Beforeplacingthesamplesinthepipettetips,wewashedthemembranes

twicewitha0.1mLsolutionof0.5%AcOH.Vacumdriedpeptideswereresuspendedin0.1mLof

0.5%TFAandloadedintotheStageTips.Fractionswerethenelutedwith0.03mLsolutionsof

20%ACN,0.5%AcOHandanincreasingconcentrationofNH4AcO–0.15to3M.The7fractions

ofeachreplicate,includingtheflowthrough,weredesaltedusingC18StageTips(seebelow)and

analyzedbyLC-MS/MS.

EnrichmentofPhosphopeptidesUsingTiO2Beads

PhosphopeptideenrichmentwasperformedwithiTRAQlabeledpeptidesaccordingwith

(Thingholmetal.,2008).Briefly,thechemicallabeledexperimentalandcontrolconditionswere

48

combined in a ratio of 1:1:1:1. Startingwith a total concentration of 100µg, each biological

replicatewasdilutedin80%acetonitrile,5%TFAand1Mglycolicacid.Next,titaniumdioxide

(TiO2)beadsof5µmindiameter(GLScience,Japan)wereaddedtothesamplesinaratioof0.6

mg/100 µg (bead/peptide) and the biological replicates were incubated for 20min at room

temperature. After that, the sampleswere spun down and the supernatantwas stored. The

beadswerewashedtwice,firstwithasolutionof80%ACNand1%TFAandthenwith10%ACN

and0.1%TFA.Thephosphopeptideswereelutedfromthebeadswithasolutionof1.5%NH4OH

ofpH11.ThesampleswerethenvacuumdriedandcleanedupwithC18StageTips.

DesaltinginC18StageTips

PriortoLC-MS/MS,allthesamplesweredesaltedaccordingwith(Rappsilberetal.,2007),

withsomeminormodifications.First,StageTipsweremadein-houseaspreviouslydescribedhere

using3MEmporeExchangediskscontainingC18bondedsilica.Next,activationandequilibration

ofthemembraneswereperformedasfollows:weadded0.1mLof100%ethanoltoeachStageTip

andcentrifugedthem.Subsequently,weplaced0.1mLofasolutioncontaining0.5%AcOHand

80%ACNintothetipsandcentrifugedthem.Finally,0.1mLof0.5%AcOHwasaddedtotheC18

exchangedisksandtheywerecentrifuged;thisstepwasrepeatedonemoretime.

OncetheStageTipshadbeenactivatedandequilibrated,thevacuumdriedsampleswere

resuspended in a solution of 1% TFA and placed into the tips. Next, the membranes were

centrifuged.Beforeelutingthepeptides,themembraneswerewashedtwicewith0.1mLof0.5%

AcOH.Afterthisstep,thesampleswereelutedwithasolutioncontaining0.5%AcOHwithan

increasingconcentrationofacetonitrile–25,50and100%.ThesamplesweredriedinaSpeedvac

andanalyzedviaLC-MS/MS.

LiquidChromatographicandMassSpectrometry

AllthesampleswereanalyzedbyaDIONEX3000nanoUPLC(ThermoScientific)system

coupled to anOrbitrap Elitemass spectrometer (Thermo Scientific). Briefly,we resuspended

vacuum dried samples in a solution of 0.1% formic acid. Then, peptideswere loaded into a

Reprosil-Pur120C18-AQin-housepackedtrapcolumn(5µmparticlesize,5.0cmlength,100µm

49

innerdiameter,360µmouterdiameter).Thetrapcolumnwaswashedfor5minwithsolventA

(0.1% formic acid, 2% ACN).Washed peptides were eluted into a 35 cm long (75 µm inner

diameter) homemade silica column packed with 3 µm C18 material (Dr.Maisch GmbH). The

peptideswereelutedfromthecolumnusingagradientofbufferB(0,1%formicacidand100%

ACN)from10to35%for155min,from35to90%for15minand90%for5minataflowrateof

230nl/min.Aftereachrun,thecolumnwaswashedandre-equilibrated.Massspectradatawere

acquiredinadata-dependent(DDA)mode,whereineachMSscan(350–1650m/zataresolution

of120,000FWHM)wasfollowedbyMS/MSscan(ataresolution15,000FWHM)withadynamic

exclusionof30s.InthebenchmarkingassayofSAXandSCX,wefragmentedthetop15most

intense peptides by higher energy collisional dissociation (HCD) with a normalized collision

energy(NCE)of35%.Inthephosphoproteomeexperiment,onlythetop12mostintenselabeled

ionswerefragmentedbyHCDwithaNCEof35%.ForthelabelediTRAQstronganionfractions,

wefragmentedthetop20mostintensepeptidesbyHCDwithaNCEof37%.

ComputationalandStatisticalDataAnalysis

Raw files of the LC-MS/MS runs have been deposited in the ProteomeXchange

Consortium through the PRIDE (Vizcaíno et al., 2016) partner repository with the data set

identifier PXD009782. These raw files were analyzed in Proteome Discover v2.1. Data were

searched against the Rattus norvegicus database (Proteome ID: UP000002494) from

Uniprot/SWISS-PROTwithaprecursormasstoleranceof10ppmandafragmentmasstolerance

of0.05Da.SearcheswereperformedusingMSAmanda2.0,aswellasSequestHT,andconducted

with methionine oxidation, deamidation, acetylation and iTRAQ labeling as variable

modifications.Cysteinecarbamidomethylationwassetasafixedmodification.Intheassaythat

webenchmarkedSAX toSCX, iTRAQ labelingwasnot setasa variablemodificationandonly

Sequest HT was used as a searching engine. In the phophosproteome experiment,

phosphorylationwasalsosetasavariablemodification.Datawerefilteredtoa1%FDRusingthe

percolator function of Proteome Discover, unique quantification and only high confident

peptideswereused infurtherprocessing.Phosphorylatedpeptideswerefilteredaccordingto

TODOforthelocalizationofthemodifications.

50

Quantifiedpeptide-spectrummatches fromenrichedandnon-enriched fractionswere

eachlog2-transformedandnormalizedbythemedianbeforecombiningthefractions.Multiple

peptide-spectrummatchesweresummarizedbytakingtheirmeanandscaledto0 levelafter

removingoutliersbyGrubb’stest(min.5values,thresholdp=0.05).Proteinsquantificationswere

obtainedbysummarizationofthepeptidesfollowingthesamemethodasforpeptide-spectrum

matches. Phosphorylated peptideswere adjusted by subtraction of protein expressions from

identical samples. Statistical testswere performed for proteins and phosphorylated peptides

separately. Differentially regulated features were determined by applying LIMMA and rank

product test and correction for multiple testing according to (Schwämmle et al., 2013).

Parameters for fuzzy c-means clustering of proteins were obtained using the method in

(SchwämmleandJensen,2010),providingaclusternumberofthreewhenusingtheminimum

centroiddistance.Proteinswithamembershipvaluebelow0.5wereassumednottobelongto

anycluster.

RealTimePCRMeasurements

RNA from the different conditionswere extracted in parallel using the TRIzol reagent

(Invitrogen)accordingwiththemanufacturer’sinstructions.Inshort,1mLofTRIzolreagentwas

addedtothetissuesamples,whichweresubsequentlyhomogenizedinanin-housemotor-driven

homogenizerfor5min.Duetothehighfatcontent,thelysateswerecentrifugedat12,000gfor

5minat4˚C.Thesupernatantsweretransferredtoanewtubeandtheywereincubatedfor5

minatroomtemperature.Next,0.2mLofchloroformwasaddedtothemixture, tubeswere

vortexedandincubatedfor5minatroomtemperature.Sampleswerecentrifugedat12,000g

for15minat4̊ C,andthecolorlessaqueousphasecontainingtheRNAmoleculeswastransferred

to a new tube. RNA was precipitated from solution by adding 0.5 mL of isopropanol and

incubatingthesamplesfor10minatroomtemperature.Finally,sampleswerecentrifugedat

12,000gfor10minat4˚Candthesupernatantwasdiscarded.TheRNApelletwasallowedto

air-dryandstoredat-20˚C.

RNAsamplesweresuspendedinTris-EDTA(TE)buffer(10mMTris-HCl,1mMEDTA,and

pH8)andtheirconcentrationweremeasuredusingNanoDrop(ThermoScientific).Inorderto

51

removepossibleDNAcontaminantsfromthesamples,DNAwasdigestedwithRQ1RNase-Free

DNase(Promega),accordingtomanufacturer’sinstructions.DNasetreatedmRNAsampleswere

used to synthesized cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo

Scientific).2µgofsampleRNAwereusedinreversetranscriptionreactionsasdescribedinthe

manufacturer’sprotocols.

BeforeperformingtheRT-qPCRmeasurements,weconductedamplificationreactionsof

the targeted transcripts – Gria1, CamkIIa, Rab3a, Ywhab, Ywhag and Gapdh (endogenous

reference)–andvisualizedtheampliconsinan2%agarosegel.Eachreactiongeneratedaunique

band in thegelwhichmatchedwith itsexpectedmolecularweight.Wehavealsoperformed

dilutioncurvestomeasuretheefficiencyoftheprimers,showingthattheywereallcloseto100%

(Supplementary Table XXX). These dilution curve assays and the RT-qPCR experiments

themselves were performed with PowerUp SYBR GreenMasterMix (Thermo Scientific) in a

StepOnePlusReal-TimePCRSystem(AppliedBiosystems),whereinthecyclingconditionswere:

95˚Cfor20sfollowedby45cyclesof95˚Cfor3s,53˚Cfor10sand60˚Cfor30s.Gapdhand

CamkIIIaweretheonlytranscriptsthathadthesameannealingandextensiontemperatures:60

˚Cfor30s.Inaddition,meltingcurveswereperformedtoconfirmPCRproducts.Allthestatistical

analysesofthisdatasetwerecarriedoutinGraphPadPrismusingOne-WayANOVA.

References

Abel,T.,Nguyen,P.V.,Barad,M.,Deuel,T.A.,Kandel,E.R.,Bourtchouladze,R.,1997a.GeneticdemonstrationofaroleforPKAinthe latephaseofLTPandinhippocampus-basedlong-termmemory.Cell88,615–626.

Abel,T.,Nguyen,P.V.,Barad,M.,Deuel,T.A.,Kandel,E.R.,Bourtchouladze,R.,1997b.GeneticdemonstrationofaroleforPKAinthe latephaseofLTPandinhippocampus-basedlong-termmemory.Cell88,615–626.

Ahmed,T.,Frey,J.U.,2005a.Plasticity-specificphosphorylationofCaMKII,MAP-kinasesandCREBduring late-LTP in rat hippocampal slices in vitro. Neuropharmacology 49, 477–492.doi:10.1016/j.neuropharm.2005.04.018

Ahmed, T., Frey, J.U., 2005b. Plasticity-specific phosphorylation of CaMKII, MAP-kinases andCREBduring late-LTP in rathippocampal slices invitro.Neuropharmacology49,477–492.doi:10.1016/j.neuropharm.2005.04.018

Aitken,C.E.,Beznosková,P.,Vlčkova,V.,Chiu,W.-L.,Zhou,F.,Valášek,L.S.,Hinnebusch,A.G.,

52

Lorsch,J.R.,2016.Eukaryotictranslationinitiationfactor3playsdistinctrolesatthemRNAentry and exit channels of the ribosomal preinitiation complex. Elife 5.doi:10.7554/eLife.20934

Altelaar,A.F.M.,Heck,A.J.R.,2012.Trendsinultrasensitiveproteomics.CurrOpinChemBiol16,206–213.doi:10.1016/j.cbpa.2011.12.011

Anderson,N.G.,Anderson,N.L.,1996.Twentyyearsoftwo-dimensionalelectrophoresis:past,presentandfuture.Electrophoresis17,443–453.doi:10.1002/elps.1150170303

Angrand,P.-O.,Segura,I.,Völkel,P.,Ghidelli,S.,Terry,R.,Brajenovic,M.,Vintersten,K.,Klein,R.,Superti-Furga, G., Drewes, G., Kuster, B., Bouwmeester, T., Acker-Palmer, A., 2006.Transgenic Mouse Proteomics Identifies New 14-3-3-associated Proteins Involved inCytoskeletal Rearrangements and Cell Signaling. Mol. Cell Proteomics 5, 2211–2227.doi:10.1074/mcp.M600147-MCP200

Azimifar,S.B.,Nagaraj,N.,Cox,J.,Mann,M.,2014.Cell-type-resolvedquantitativeproteomicsofmurineliver.CellMetab.20,1076–1087.doi:10.1016/j.cmet.2014.11.002

Bantscheff,M.,Boesche,M.,Eberhard,D.,Matthieson,T.,Sweetman,G.,Kuster,B.,2008.Robustand sensitive iTRAQ quantification on an LTQ Orbitrap mass spectrometer. Mol. CellProteomics7,1702–1713.doi:10.1074/mcp.M800029-MCP200

Bantscheff,M.,Lemeer,S.,Savitski,M.M.,Kuster,B.,2012.Quantitativemassspectrometryinproteomics:criticalreviewupdatefrom2007tothepresent.AnalBioanalChem404,939–965.doi:10.1007/s00216-012-6203-4

Bantscheff, M., Schirle, M., Sweetman, G., Rick, J., Kuster, B., 2007. Quantitative massspectrometry in proteomics: a critical review. Anal Bioanal Chem 389, 1017–1031.doi:10.1007/s00216-007-1486-6

Beck, S., Michalski, A., Raether, O., Lubeck, M., Kaspar, S., Goedecke, N., Baessmann, C.,Hornburg,D.,Meier,F.,Paron,I.,Kulak,N.A.,Cox,J.,Mann,M.,2015.TheImpactII,aVeryHigh-Resolution Quadrupole Time-of-Flight Instrument (QTOF) for Deep ShotgunProteomics.Mol.CellProteomics14,2014–2029.doi:10.1074/mcp.M114.047407

Berg,D.,Holzmann,C.,Riess,O.,2003.14-3-3proteinsinthenervoussystem.Nat.Rev.Neurosci.4,752–762.doi:10.1038/nrn1197

Bliss,T.V.,Collingridge,G.L.,1993.Asynapticmodelofmemory:long-termpotentiationinthehippocampus.Nature361,31–39.doi:10.1038/361031a0

Bodenmiller, B., Mueller, L.N., Mueller, M., Domon, B., Aebersold, R., 2007. Reproducibleisolationofdistinct,overlappingsegmentsofthephosphoproteome.Nat.Methods4,231–237.doi:10.1038/nmeth1005

Boersema,P.J.,Geiger,T.,Wisniewski,J.R.,Mann,M.,2013.QuantificationoftheN-glycosylatedsecretomebysuper-SILACduringbreastcancerprogressionand inhumanbloodsamples.Mol.CellProteomics12,158–171.doi:10.1074/mcp.M112.023614

Borovok,N.,Nesher,E.,Levin,Y.,Reichenstein,M.,Pinhasov,A.,Michaelevski,I.,2016.Dynamics

53

ofHippocampalProteinExpressionDuringLong-termSpatialMemoryFormation.Mol.CellProteomics15,523–541.doi:10.1074/mcp.M115.051318

Bourtchouladze,R.,Abel,T.,Berman,N.,Gordon,R.,Lapidus,K.,Kandel,E.R.,1998.Differenttraining procedures recruit either one or two critical periods for contextual memoryconsolidation,eachofwhichrequiresproteinsynthesisandPKA.Learning&Memory5,365–374.

Castillo,P.E.,2012.PresynapticLTPandLTDofexcitatoryandinhibitorysynapses.ColdSpringHarbPerspectBiol4.doi:10.1101/cshperspect.a005728

Cavallaro,S.,D'Agata,V.,Manickam,P.,Dufour,F.,Alkon,D.L.,2002.Memory-specifictemporalprofiles of gene expression in the hippocampus. Proc. Natl. Acad. Sci. U.S.A. 99, 16279–16284.doi:10.1073/pnas.242597199

Chen,B.,Zhao,L.,Li,X.,Ji,Y.-S.,Li,N.,Xu,X.-F.,Chen,Z.-Y.,2014.Syntaxin8modulatesthepost-synthetictraffickingoftheTrkAreceptorandinflammatorypaintransmission.J.Biol.Chem.289,19556–19569.doi:10.1074/jbc.M114.567925

Chen, X., Winters, C., Azzam, R., Li, X., Galbraith, J.A., Leapman, R.D., Reese, T.S., 2008.Organizationofthecorestructureofthepostsynapticdensity.Proc.Natl.Acad.Sci.U.S.A.105,4453–4458.doi:10.1073/pnas.0800897105

Cheng, Z., Teo,G., Krueger, S., Rock, T.M., Koh,H.W.L., Choi,H.,Vogel, C., 2016.Differentialdynamicsof themammalianmRNAandproteinexpression response tomisfoldingstress.MolecularSystemsBiology12,855.

Deisseroth,K.,Bito,H.,Tsien,R.W.,1996.Signalingfromsynapsetonucleus:postsynapticCREBphosphorylationduringmultipleformsofhippocampalsynapticplasticity.Neuron16,89–101.

Diana,G.,Valentini,G.,Travaglione,S.,Falzano,L.,Pieri,M.,Zona,C.,Meschini,S.,Fabbri,A.,Fiorentini,C.,2007.EnhancementoflearningandmemoryafteractivationofcerebralRhoGTPases.Proc.Natl.Acad.Sci.U.S.A.104,636–641.doi:10.1073/pnas.0610059104

Dingledine,R.,Borges,K.,Bowie,D.,Traynelis,S.F.,1999.Theglutamatereceptorionchannels.Pharmacol.Rev.51,7–61.

Domon, B., Aebersold, R., 2010. Options and considerations when selecting a quantitativeproteomicsstrategy.Nat.Biotechnol.28,710–721.doi:10.1038/nbt.1661

Emes,R.D.,Pocklington,A.J.,Anderson,C.N.G.,Bayes,A.,Collins,M.O.,Vickers,C.A.,Croning,M.D.R., Malik, B.R., Choudhary, J.S., Armstrong, J.D., Grant, S.G.N., 2008. Evolutionaryexpansionandanatomicalspecializationofsynapseproteomecomplexity.Nat.Neurosci.11,799–806.doi:10.1038/nn.2135

Exton-McGuinness,M.T.J.,Patton,R.C.,Sacco,L.B.,Lee,J.L.C.,2014.Reconsolidationofawell-learnedinstrumentalmemory.Learn.Mem.21,468–477.doi:10.1101/lm.035543.114

Filiou,M.D.,Martins-de-Souza,D.,Guest,P.C.,Bahn,S.,Turck,C.W.,2012.Tolabelornottolabel:applicationsofquantitativeproteomicsinneuroscienceresearch.Proteomics12,736–747.

54

doi:10.1002/pmic.201100350Fioravante,D.,Byrne,J.H.,2011.Proteindegradationandmemoryformation.BrainRes.Bull.85,

14–20.doi:10.1016/j.brainresbull.2010.11.002Frey,M.C.,Sprengel,R.,Nevian,T.,2009.Activitypattern-dependentlong-termpotentiationin

neocortexandhippocampusofGluA1(GluR-A)subunit-deficientmice.J.Neurosci.29,5587–5596.doi:10.1523/JNEUROSCI.5314-08.2009

Geiger,T.,Wehner,A.,Schaab,C.,Cox,J.,Mann,M.,2012.Comparativeproteomicanalysisofelevencommoncelllinesrevealsubiquitousbutvaryingexpressionofmostproteins.Mol.CellProteomics11,M111.014050.doi:10.1074/mcp.M111.014050

Geiger,T.,Wisniewski,J.R.,Cox,J.,Zanivan,S.,Kruger,M.,Ishihama,Y.,Mann,M.,2011.Useofstableisotopelabelingbyaminoacidsincellcultureasaspike-instandardinquantitativeproteomics.NatProtoc6,147–157.doi:10.1038/nprot.2010.192

Giovannini,M.G.,Blitzer,R.D.,Wong,T.,Asoma,K.,Tsokas,P.,Morrison,J.H.,Iyengar,R.,Landau,E.M.,2001.Mitogen-activatedproteinkinaseregulatesearlyphosphorylationanddelayedexpression of Ca2+/calmodulin-dependent protein kinase II in long-term potentiation. J.Neurosci.21,7053–7062.

Granger, A.J., Shi, Y., Lu, W., Cerpas, M., Nicoll, R.A., 2013. LTP requires a reserve pool ofglutamate receptors independent of subunit type. Nature 493, 495–500.doi:10.1038/nature11775

Gygi,S.P.,Rist,B.,Gerber,S.A.,Turecek,F.,Gelb,M.H.,Aebersold,R.,1999.Quantitativeanalysisofcomplexproteinmixturesusingisotope-codedaffinitytags.Nat.Biotechnol.17,994–999.doi:10.1038/13690

Hayashi, Y., Shi, S.H., Esteban, J.A., Piccini, A., Poncer, J.C.,Malinow, R., 2000.DrivingAMPAreceptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domaininteraction.Science287,2262–2267.

Hebert,A.S.,Merrill,A.E.,Bailey,D.J., Still,A.J.,Westphall,M.S., Strieter,E.R.,Pagliarini,D.J.,Coon,J.J.,2013.Neutron-encodedmasssignaturesformultiplexedproteomequantification.Nat.Methods10,332–334.doi:10.1038/nmeth.2378

Hernandez, P.J., Kelley, A.E., 2004. Long-term memory for instrumental responses does notundergoprotein synthesis-dependent reconsolidationupon retrieval. Learning&Memory11,748–754.doi:10.1101/lm.84904

Hong, W., 2005. SNAREs and traffic. Biochim. Biophys. Acta 1744, 120–144.doi:10.1016/j.bbamcr.2005.03.014

Hölscher,C.,Anwyl,R.,Rowan,M.J.,1997.Stimulationon thepositivephaseofhippocampalthetarhythminduceslong-termpotentiationthatcanBedepotentiatedbystimulationonthenegativephaseinareaCA1invivo.JournalofNeuroscience17,6470–6477.

Huang,C.-C.,Yang,D.-M.,Lin,C.-C.,Kao,L.-S.,2011. InvolvementofRab3A invesicleprimingduring exocytosis: interaction with Munc13-1 and Munc18-1. Traffic 12, 1356–1370.

55

doi:10.1111/j.1600-0854.2011.01237.xHuang,D.W.,Sherman,B.T.,Tan,Q.,Kir, J., Liu,D.,Bryant,D.,Guo,Y.,Stephens,R.,Baseler,

M.W., Lane, H.C., Lempicki, R.A., 2007. DAVID Bioinformatics Resources: expandedannotationdatabaseandnovelalgorithms tobetterextractbiology from largegene lists.NucleicAcidsResearch35,W169–75.doi:10.1093/nar/gkm415

Izquierdo, I., Bevilaqua, L.R.M., Rossato, J.I., Bonini, J.S.,Medina, J.H., Cammarota,M., 2006.Differentmolecularcascades indifferentsitesofthebraincontrolmemoryconsolidation.TrendsNeurosci.29,496–505.doi:10.1016/j.tins.2006.07.005

Jackson, R.J., Hellen, C.U.T., Pestova, T.V., 2010. The mechanism of eukaryotic translationinitiation and principles of its regulation. Nat. Rev. Mol. Cell Biol. 11, 113–127.doi:10.1038/nrm2838

Jarome, T.J., Ferrara,N.C., Kwapis, J.L., Helmstetter, F.J., 2016. CaMKII regulates proteasomephosphorylation and activity and promotes memory destabilization following retrieval.NeurobiolLearnMem128,103–109.doi:10.1016/j.nlm.2016.01.001

Jurado-Parras,M.T.,Sánchez-Campusano,R.,Castellanos,N.P.,del-Pozo,F.,Gruart,A.,Delgado-García, J.M., 2013. Differential contribution of hippocampal circuits to appetitive andconsummatory behaviors during operant conditioning of behavingmice. J. Neurosci. 33,2293–2304.doi:10.1523/JNEUROSCI.1013-12.2013

Kaltschmidt,B.,Ndiaye,D.,Korte,M.,Pothion,S.,Arbibe,L.,Prüllage,M.,Pfeiffer,J.,Lindecke,A., Staiger, V., Israël, A., Kaltschmidt, C., Mémet, S., 2006. NF-kappaB regulates spatialmemoryformationandsynapticplasticitythroughproteinkinaseA/CREBsignaling.Mol.Cell.Biol.26,2936–2946.doi:10.1128/MCB.26.8.2936-2946.2006

Kandel,E.R.,Dudai,Y.,Mayford,M.R.,2014.Themolecularandsystemsbiologyofmemory.Cell157,163–186.doi:10.1016/j.cell.2014.03.001

Karp,N.A.,Huber,W.,Sadowski,P.G.,Charles,P.D.,Hester,S.V.,Lilley,K.S.,2010.Addressingaccuracy and precision issues in iTRAQ quantitation.Mol. Cell Proteomics 9, 1885–1897.doi:10.1074/mcp.M900628-MCP200

Kent,C.B.,Shimada,T.,Ferraro,G.B.,Ritter,B.,Yam,P.T.,McPherson,P.S.,Charron,F.,Kennedy,T.E., Fournier, A.E., 2010. 14-3-3 proteins regulate protein kinase a activity tomodulategrowth cone turning responses. J. Neurosci. 30, 14059–14067.doi:10.1523/JNEUROSCI.3883-10.2010

Kettenbach, A.N., Rush, J., Gerber, S.A., 2011. Absolute quantification of protein and post-translationalmodification abundancewith stable isotope-labeled synthetic peptides. NatProtoc6,175–186.doi:10.1038/nprot.2010.196

Klann, E., Dever, T.E., 2004. Biochemicalmechanisms for translational regulation in synapticplasticity.Nat.Rev.Neurosci.5,931–942.doi:10.1038/nrn1557

Kruger,M.,Moser,M.,Ussar,S.,Thievessen,I.,Luber,C.A.,Forner,F.,Schmidt,S.,Zanivan,S.,Fässler,R.,Mann,M.,2008.SILACmouseforquantitativeproteomicsuncoverskindlin-3as

56

an essential factor for red blood cell function. Cell 134, 353–364.doi:10.1016/j.cell.2008.05.033

Lamprecht,R.,2011.Therolesoftheactincytoskeletoninfearmemoryformation.FrontBehavNeurosci5,39.doi:10.3389/fnbeh.2011.00039

Lee,J.L.C.,Nader,K.,Schiller,D.,2017.AnUpdateonMemoryReconsolidationUpdating.TrendsCogn.Sci.(Regul.Ed.)21,531–545.doi:10.1016/j.tics.2017.04.006

Lee,S.-H.,Choi,J.-H.,Lee,N.,Lee,H.-R.,Kim,J.-I.,Yu,N.-K.,Choi,S.-L.,Lee,S.-H.,Kim,H.,Kaang,B.-K.,2008.Synapticproteindegradationunderliesdestabilizationofretrievedfearmemory.Science319,1253–1256.doi:10.1126/science.1150541

Lesburguères,E.,Gobbo,O.L.,Alaux-Cantin,S.,Hambucken,A.,Trifilieff,P.,Bontempi,B.,2011.Early tagging of cortical networks is required for the formation of enduring associativememory.Science331,924–928.doi:10.1126/science.1196164

Lisman,J.,Schulman,H.,Cline,H.,2002.ThemolecularbasisofCaMKIIfunctioninsynapticandbehaviouralmemory.Nat.Rev.Neurosci.3,175–190.doi:10.1038/nrn753

Lisman, J., Yasuda, R., Raghavachari, S., 2012. Mechanisms of CaMKII action in long-termpotentiation.Nat.Rev.Neurosci.13,169–182.doi:10.1038/nrn3192

Liu,Y.,Beyer,A.,Aebersold,R.,2016.OntheDependencyofCellularProteinLevelsonmRNAAbundance.Cell165,535–550.doi:10.1016/j.cell.2016.03.014

Lorenzetti, F.D., Baxter, D.A., Byrne, J.H., 2008. Molecular mechanisms underlying a cellularanalogofoperantrewardlearning.Neuron59,815–828.doi:10.1016/j.neuron.2008.07.019

Lu,R.,Markowetz,F.,Unwin,R.D.,Leek,J.T.,Airoldi,E.M.,MacArthur,B.D.,Lachmann,A.,Rozov,R., Ma’ayan, A., Boyer, L.A., Troyanskaya, O.G., Whetton, A.D., Lemischka, I.R., 2009.Systems-leveldynamicanalysesoffatechangeinmurineembryonicstemcells.Nature462,358–362.doi:10.1038/nature08575

Lu, W., Isozaki, K., Roche, K.W., Nicoll, R.A., 2010. Synaptic targeting of AMPA receptors isregulatedbyaCaMKIIsiteinthefirstintracellularloopofGluA1.Proc.Natl.Acad.Sci.U.S.A.107,22266–22271.doi:10.1073/pnas.1016289107

Lüscher,C.,Malenka,R.C.,2012.NMDAreceptor-dependentlong-termpotentiationandlong-term depression (LTP/LTD). Cold Spring Harb Perspect Biol 4.doi:10.1101/cshperspect.a005710

Ma,T.,Chen,Y.,Vingtdeux,V.,Zhao,H.,Viollet,B.,Marambaud,P.,Klann,E.,2014.InhibitionofAMP-activated protein kinase signaling alleviates impairments in hippocampal synapticplasticityinducedbyamyloidβ.J.Neurosci.34,12230–12238.doi:10.1523/JNEUROSCI.1694-14.2014

Mann,M.,Kelleher,N.L.,2008.Precisionproteomics:thecaseforhighresolutionandhighmassaccuracy.Proc.Natl.Acad.Sci.U.S.A.105,18132–18138.doi:10.1073/pnas.0800788105

Mann,M.,Kulak,N.A.,Nagaraj,N.,Cox, J.,2013.Thecomingageof complete,accurate,andubiquitousproteomes.Mol.Cell49,583–590.doi:10.1016/j.molcel.2013.01.029

57

Marshall,A.G.,Hendrickson,C.L.,Jackson,G.S.,1998.Fouriertransformioncyclotronresonancemass spectrometry: a primer. Mass Spectrom Rev 17, 1–35. doi:10.1002/(SICI)1098-2787(1998)17:1<1::AID-MAS1>3.0.CO;2-K

Martineau,Y.,Wang,X.,Alain,T.,Petroulakis,E.,Shahbazian,D.,Fabre,B.,Bousquet-Dubouch,M.-P., Monsarrat, B., Pyronnet, S., Sonenberg, N., 2014. Control of Paip1-eukayrotictranslationinitiationfactor3interactionbyaminoacidsthroughS6kinase.Mol.Cell.Biol.34,1046–1053.doi:10.1128/MCB.01079-13

McGaugh,J.L.,2000.Memory--acenturyofconsolidation.Science287,248–251.McMahon, H.T., Missler, M., Li, C., Südhof, T.C., 1995. Complexins: cytosolic proteins that

regulateSNAPreceptorfunction.Cell83,111–119.Meissner,F.,Mann,M.,2014.Quantitativeshotgunproteomics:considerationsforahigh-quality

workflowinimmunology.Nat.Immunol.15,112–117.doi:10.1038/ni.2781Michalski,A.,Damoc,E.,Lange,O.,Denisov,E.,Nolting,D.,Müller,M.,Viner,R.,Schwartz,J.,

Remes,P.,Belford,M.,Dunyach,J.-J.,Cox,J.,Horning,S.,Mann,M.,Makarov,A.,2012.Ultrahigh resolution linear ion trapOrbitrapmass spectrometer (Orbitrap Elite) facilitates topdown LC MS/MS and versatile peptide fragmentation modes. Mol. Cell Proteomics 11,O111.013698.doi:10.1074/mcp.O111.013698

Milekic, M.H., Alberini, C.M., 2002. Temporally graded requirement for protein synthesisfollowingmemoryreactivation.Neuron36,521–525.

Monopoli,M.P.,Raghnaill,M.N.,Loscher,J.S.,O'Sullivan,N.C.,Pangalos,M.N.,Ring,R.H.,Schack,von,D.,Dunn,M.J.,Regan,C.M.,Pennington,S.,Murphy,K.J.,2011.Temporalproteomicprofileofmemoryconsolidationintherathippocampaldentategyrus.Proteomics11,4189–4201.doi:10.1002/pmic.201100072

Moser,M.-B.,Rowland,D.C.,Moser,E.I.,2015.Placecells,gridcells,andmemory.ColdSpringHarbPerspectBiol7,a021808.doi:10.1101/cshperspect.a021808

Mueller,S.,Wahlander,A.,Selevsek,N.,Otto,C.,Ngwa,E.M.,Poljak,K.,Frey,A.D.,Aebi,M.,Gauss,R.,2015.ProteindegradationcorrectsforimbalancedsubunitstoichiometryinOSTcomplexassembly.Mol.Biol.Cell26,2596–2608.doi:10.1091/mbc.E15-03-0168

Musunuri,S.,Wetterhall,M.,Ingelsson,M.,Lannfelt,L.,Artemenko,K.,Bergquist,J.,Kultima,K.,Shevchenko,G., 2014.Quantification of the brain proteome in Alzheimer's disease usingmultiplexedmassspectrometry.J.ProteomeRes.13,2056–2068.doi:10.1021/pr401202d

Nabavi,S.,Fox,R.,Proulx,C.D.,Lin,J.Y.,Tsien,R.Y.,Malinow,R.,2014.EngineeringamemorywithLTDandLTP.Nature511,348–352.doi:10.1038/nature13294

Nagaraj,N.,Wisniewski,J.R.,Geiger,T.,Cox,J.,Kircher,M.,Kelso,J.,Pääbo,S.,Mann,M.,2011.Deepproteomeandtranscriptomemappingofahumancancercellline.MolecularSystemsBiology7,548–548.doi:10.1038/msb.2011.81

Nepusz, T., Yu, H., Paccanaro, A., 2012. Detecting overlapping protein complexes in protein-proteininteractionnetworks.Nat.Methods9,471–472.doi:10.1038/nmeth.1938

58

Neves,G.,Cooke,S.F.,Bliss,T.V.P.,2008.Synapticplasticity,memoryandthehippocampus:aneuralnetworkapproachtocausality.Nat.Rev.Neurosci.9,65–75.doi:10.1038/nrn2303

Nguyen,P.V.,Abel,T.,Kandel,E.R.,1994.RequirementofacriticalperiodoftranscriptionforinductionofalatephaseofLTP.Science265,1104–1107.

Ong,S.-E.,Blagoev,B.,Kratchmarova,I.,Kristensen,D.B.,Steen,H.,Pandey,A.,Mann,M.,2002.Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurateapproachtoexpressionproteomics.Mol.CellProteomics1,376–386.

Ong,S.-E.,Mann,M.,2005.Massspectrometry-basedproteomicsturnsquantitative.Nat.Chem.Biol.1,252–262.doi:10.1038/nchembio736

Park,M.,Salgado,J.M.,Ostroff,L.,Helton,T.D.,Robinson,C.G.,Harris,K.M.,Ehlers,M.D.,2006.Plasticity-induced growth of dendritic spines by exocytic trafficking from recyclingendosomes.Neuron52,817–830.doi:10.1016/j.neuron.2006.09.040

Patterson,S.L.,Pittenger,C.,Morozov,A.,Martin,K.C.,Scanlin,H.,Drake,C.,Kandel,E.R.,2001.SomeformsofcAMP-mediatedlong-lastingpotentiationareassociatedwithreleaseofBDNFandnucleartranslocationofphospho-MAPkinase.Neuron32,123–140.

Pontes,A.H.,deSousa,M.V.,2016.MassSpectrometry-BasedApproaches toUnderstand theMolecularBasisofMemory.FrontChem4,40.doi:10.3389/fchem.2016.00040

Poser, S., Storm, D.R., 2001. Role of Ca2+-stimulated adenylyl cyclases in LTP and memoryformation. International Journal of Developmental Neuroscience 19, 387–394.doi:10.1016/S0736-5748(00)00094-0

Qiao, H., Foote, M., Graham, K., Wu, Y., Zhou, Y., 2014. 14-3-3 proteins are required forhippocampallong-termpotentiationandassociativelearningandmemory.J.Neurosci.34,4801–4808.doi:10.1523/JNEUROSCI.4393-13.2014

Rao-Ruiz,P.,Carney,K.E.,Pandya,N.,vanderLoo,R.J.,Verheijen,M.H.G.,vanNierop,P.,Smit,A.B., Spijker, S., 2015. Time-dependent changes in the mouse hippocampal synapticmembrane proteome after contextual fear conditioning. Hippocampus 25, 1250–1261.doi:10.1002/hipo.22432

Rapanelli,M.,Frick,L.R.,Zanutto,B.S.,2011.Learninganoperantconditioningtaskdifferentiallyinducesgliogenesis inthemedialprefrontalcortexandneurogenesis inthehippocampus.PLoSONE6,e14713.doi:10.1371/journal.pone.0014713

Rappsilber, J.,Mann,M., Ishihama,Y.,2007.Protocol formicro-purification,enrichment,pre-fractionationandstorageofpeptidesforproteomicsusingStageTips.NatProtoc2,1896–1906.doi:10.1038/nprot.2007.261

Rauniyar,N.,McClatchy,D.B.,Yates,J.R.,2013.Stableisotopelabelingofmammals(SILAM)forin vivo quantitative proteomic analysis. Methods 61, 260–268.doi:10.1016/j.ymeth.2013.03.008

Rempel-Clower,N.L.,Zola,S.M.,Squire,L.R.,Amaral,D.G.,1996.Threecasesofenduringmemoryimpairment after bilateral damage limited to the hippocampal formation. Journal of

59

Neuroscience16,5233–5255.Roberson, E.D., English, J.D.,Adams, J.P., Selcher, J.C., Kondratick, C., Sweatt, J.D., 1999. The

Mitogen-ActivatedProteinKinaseCascadeCouplesPKAandPKCtocAMPResponseElementBindingProteinPhosphorylationinAreaCA1ofHippocampus.JournalofNeuroscience19,4337–4348.doi:10.1523/JNEUROSCI.19-11-04337.1999

Ross,C.A.,Poirier,M.A.,2004.Proteinaggregationandneurodegenerativedisease.Nat.Med.10Suppl,S10–7.doi:10.1038/nm1066

Ross,P.L.,Huang,Y.N.,Marchese,J.N.,Williamson,B.,Parker,K.,Hattan,S.,Khainovski,N.,Pillai,S., Dey, S., Daniels, S., Purkayastha, S., Juhasz, P., Martin, S., Bartlet-Jones, M., He, F.,Jacobson, A., Pappin, D.J., 2004. Multiplexed protein quantitation in Saccharomycescerevisiae using amine-reactive isobaric tagging reagents.Mol. Cell Proteomics 3, 1154–1169.doi:10.1074/mcp.M400129-MCP200

Routtenberg,A.,Rekart,J.L.,2005.Post-translationalproteinmodificationasthesubstrateforlong-lastingmemory.TrendsNeurosci.28,12–19.doi:10.1016/j.tins.2004.11.006

Ryan,M.M.,Ryan,B.,Kyrke-Smith,M.,Logan,B.,Tate,W.P.,Abraham,W.C.,Williams,J.M.,2012.Temporalprofilingofgenenetworksassociatedwiththelatephaseoflong-termpotentiationinvivo.PLoSONE7,e40538.doi:10.1371/journal.pone.0040538

Ryan,T.J.,Grant,S.G.N.,2009.Theoriginandevolutionofsynapses.Nat.Rev.Neurosci.10,701–712.doi:10.1038/nrn2717

Ryan,T.J.,Roy,D.S.,Pignatelli,M.,Arons,A.,Tonegawa,S.,2015.Memory.Engramcellsretainmemoryunderretrogradeamnesia.Science348,1007–1013.doi:10.1126/science.aaa5542

Sadygov,R.G.,Cociorva,D.,Yates,J.R.,2004.Large-scaledatabasesearchingusingtandemmassspectra: looking up the answer in the back of the book. Nat. Methods 1, 195–202.doi:10.1038/nmeth725

Schiller,J.,Schiller,Y.,Clapham,D.E.,1998.NMDAreceptorsamplifycalciuminfluxintodendriticspines during associative pre- and postsynaptic activation. Nat. Neurosci. 1, 114–118.doi:10.1038/363

Schwämmle,V.,Jensen,O.N.,2010.Asimpleandfastmethodtodeterminetheparametersforfuzzy c-means cluster analysis. Bioinformatics 26, 2841–2848.doi:10.1093/bioinformatics/btq534

Schwämmle,V.,León,I.R.,Jensen,O.N.,2013.Assessmentandimprovementofstatisticaltoolsforcomparativeproteomicsanalysisofsparsedatasetswithfewexperimentalreplicates.J.ProteomeRes.12,3874–3883.doi:10.1021/pr400045u

Sharma,K.,Schmitt,S.,Bergner,C.G.,Tyanova,S.,Kannaiyan,N.,Manrique-Hoyos,N.,Kongi,K.,Cantuti, L.,Hanisch,U.-K.,Philips,M.-A.,Rossner,M.J.,Mann,M., Simons,M.,2015.Celltype- and brain region-resolved mouse brain proteome. Nat. Neurosci. 18, 1819–1831.doi:10.1038/nn.4160

Skoulakis,E.M.,Davis,R.L.,1996.Olfactorylearningdeficitsinmutantsforleonardo,aDrosophila

60

geneencodinga14-3-3protein.Neuron17,931–944.Squire,L.R.,Genzel,L.,Wixted,J.T.,Morris,R.G.,2015.Memoryconsolidation.ColdSpringHarb

PerspectBiol7,a021766.doi:10.1101/cshperspect.a021766Strange, B.A., Witter, M.P., Lein, E.S., Moser, E.I., 2014. Functional organization of the

hippocampallongitudinalaxis.Nat.Rev.Neurosci.15,655–669.doi:10.1038/nrn3785Strekalova,T.,Zörner,B.,Zacher,C.,Sadovska,G.,Herdegen,T.,Gass,P.,2003.Memoryretrieval

aftercontextualfearconditioninginducesc-FosandJunBexpressioninCA1hippocampus.GenesBrainBehav.2,3–10.

Sui,L.,Wang,J.,Li,B.-M.,2008.Roleofthephosphoinositide3-kinase-Akt-mammaliantargetofthe rapamycin signaling pathway in long-term potentiation and trace fear conditioningmemoryinratmedialprefrontalcortex.Learn.Mem.15,762–776.doi:10.1101/lm.1067808

Szklarczyk,D.,Morris,J.H.,Cook,H.,Kuhn,M.,Wyder,S.,Simonovic,M.,Santos,A.,Doncheva,N.T.,Roth,A.,Bork,P., Jensen, L.J.,Mering,von,C.,2017.TheSTRINGdatabase in2017:quality-controlled protein-protein association networks,madebroadly accessible.NucleicAcidsResearch45,D362–D368.doi:10.1093/nar/gkw937

Thakur,S.S.,Geiger,T.,Chatterjee,B.,Bandilla,P.,Fröhlich,F.,Cox,J.,Mann,M.,2011.Deepandhighly sensitive proteome coverage by LC-MS/MS without prefractionation. Mol. CellProteomics10,M110.003699.doi:10.1074/mcp.M110.003699

Thingholm,T.E.,Jensen,O.N.,Robinson,P.J.,Larsen,M.R.,2008.SIMAC(sequentialelutionfromIMAC),aphosphoproteomicsstrategyfortherapidseparationofmonophosphorylatedfrommultiply phosphorylated peptides. Mol. Cell Proteomics 7, 661–671.doi:10.1074/mcp.M700362-MCP200

Thompson,A.,Schäfer,J.,Kuhn,K.,Kienle,S.,Schwarz,J.,Schmidt,G.,Neumann,T.,Johnstone,R.,Mohammed,A.K.A.,Hamon,C.,2003.Tandemmasstags:anovelquantificationstrategyforcomparativeanalysisofcomplexproteinmixturesbyMS/MS.Anal.Chem.75,1895–1904.

Ting,L.,Rad,R.,Gygi,S.P.,Haas,W.,2011.MS3eliminatesratiodistortioninisobaricmultiplexedquantitativeproteomics.Nat.Methods8,937–940.doi:10.1038/nmeth.1714

Tronson,N.C.,Taylor,J.R.,2007.Molecularmechanismsofmemoryreconsolidation.Nat.Rev.Neurosci.8,262–275.doi:10.1038/nrn2090

Tsokas,P.,Hsieh,C.,Yao,Y.,Lesburguères,E.,Wallace,E.J.C.,Tcherepanov,A.,Jothianandan,D.,Hartley, B.R., Pan, L., Rivard, B., Farese, R.V., Sajan,M.P., Bergold, P.J., Hernández, A.I.,Cottrell,J.E.,Shouval,H.Z.,Fenton,A.A.,Sacktor,T.C.,2016.CompensationforPKMζinlong-term potentiation and spatial long-term memory in mutant mice. Elife 5, 12677.doi:10.7554/eLife.14846

Tucker,W.C.,Chapman,E.R.,2002.RoleofsynaptotagmininCa2+-triggeredexocytosis.Biochem.J.366,1–13.doi:10.1042/BJ20020776

Uzakov,S.,Frey,J.U.,Korz,V.,2005.ReinforcementofrathippocampalLTPbyholeboardtraining.Learning&Memory12,165–171.doi:10.1101/lm.89305

61

Villén,J.,Beausoleil,S.A.,Gerber,S.A.,Gygi,S.P.,2007.Large-scalephosphorylationanalysisofmouseliver.Proc.Natl.Acad.Sci.U.S.A.104,1488–1493.doi:10.1073/pnas.0609836104

Viola,H.,Furman,M., Izquierdo,L.A.,Alonso,M.,Barros,D.M.,deSouza,M.M., Izquierdo, I.,Medina,J.H.,2000.PhosphorylatedcAMPresponseelement-bindingproteinasamolecularmarkerofmemoryprocessinginrathippocampus:effectofnovelty.J.Neurosci.20,RC112.

Vizcaíno,J.A.,Csordas,A.,del-Toro,N.,Dianes,J.A.,Griss,J.,Lavidas,I.,Mayer,G.,Perez-Riverol,Y.,Reisinger,F.,Ternent,T.,Xu,Q.-W.,Wang,R.,Hermjakob,H.,2016.2016updateofthePRIDE database and its related tools. Nucleic Acids Research 44, D447–56.doi:10.1093/nar/gkv1145

Wang,H.,Yang,Y.,Li,Y.,Bai,B.,Wang,X.,Tan,H.,Liu,T.,Beach,T.G.,Peng,J.,Wu,Z.,2015.Systematic optimization of long gradient chromatography mass spectrometry for deepanalysisofbrainproteome.J.ProteomeRes.14,829–838.doi:10.1021/pr500882h

Wang,X.,Pattison,J.S.,Su,H.,2013.Posttranslationalmodificationandqualitycontrol.Circ.Res.112,367–381.doi:10.1161/CIRCRESAHA.112.268706

Wenger,C.D.,Lee,M.V.,Hebert,A.S.,McAlister,G.C.,Phanstiel,D.H.,Westphall,M.S.,Coon,J.J.,2011.Gas-phase purification enables accurate,multiplexed proteomequantificationwithisobarictagging.Nat.Methods8,933–935.doi:10.1038/nmeth.1716

Whitlock, J.R., Heynen, A.J., Shuler, M.G., Bear, M.F., 2006. Learning induces long-termpotentiationinthehippocampus.Science313,1093–1097.doi:10.1126/science.1128134

Wisniewski, J.R., Zougman, A., Mann, M., 2009. Combination of FASP and StageTip-basedfractionationallowsin-depthanalysisofthehippocampalmembraneproteome.J.ProteomeRes.8,5674–5678.doi:10.1021/pr900748n

Wong, S.T., Athos, J., Figueroa, X.A., Pineda, V.V., Schaefer,M.L., Chavkin, C.C.,Muglia, L.J.,Storm,D.R.,1999.Calcium-stimulatedadenylylcyclaseactivityiscriticalforhippocampus-dependentlong-termmemoryandlatephaseLTP.Neuron23,787–798.

Yi,E.C.,Li,X.-J.,Cooke,K.,Lee,H.,Raught,B.,Page,A.,Aneliunas,V.,Hieter,P.,Goodlett,D.R.,Aebersold,R.,2005.Increasedquantitativeproteomecoveragewith(13)C/(12)C-based,acid-cleavable isotope-coded affinity tag reagent and modified data acquisition scheme.Proteomics5,380–387.doi:10.1002/pmic.200400970

Zhang, Y., Fonslow, B.R., Shan, B., Baek, M.-C., Yates, J.R., 2013. Protein analysis byshotgun/bottom-upproteomics.Chem.Rev.113,2343–2394.doi:10.1021/cr3003533

Zhou,M.,Sandercock,A.M.,Fraser,C.S.,Ridlova,G.,Stephens,E.,Schenauer,M.R.,Yokoi-Fong,T., Barsky, D., Leary, J.A., Hershey, J.W., Doudna, J.A., Robinson, C.V., 2008. Massspectrometryrevealsmodularityandacompletesubunitinteractionmapoftheeukaryotictranslation factor eIF3. Proc. Natl. Acad. Sci. U.S.A. 105, 18139–18144.doi:10.1073/pnas.0801313105

Zola-Morgan, S., Squire, L.R., Amaral, D.G., 1986. Human amnesia and the medial temporalregion:enduringmemoryimpairmentfollowingabilaterallesionlimitedtofieldCA1ofthe

62

hippocampus.JournalofNeuroscience6,2950–2967.NoteontheIntroductionofthisthesis:

TheintroductionwaspublishedasPontesAHanddeSousaMV(2016)MassSpectrometry-

Based Approaches to Understand theMolecular Basis ofMemory. Front. Chem. 4:40. doi:

10.3389/fchem.2016.00040;https://www.frontiersin.org/article/10.3389/fchem.2016.00040.

63

Acknowledgments

IwouldliketothankmymentorMarceloValledeSousaforsupportandguidance,aswellasforprovidingthefacilitiesandresourcesIneededthroughoutthisthesis.IwanttoexpressmygratitudetoVeitSchwämmle,WagnerFontes,BeatrizDolabela,EleniceHanna,FabianeHiratsukaandMaurícioPontesfor inspirationandfruitfulcollaborationsthroughwhichI learnedalot. IthankallmycolleaguesfromLBQP,speciallymyfriendsNuno,AntônioandEgmarwhohelpedmegetthroughfourimportantyearsofmylife.Specialthankstomyparents,mybrothersandmygirlfriendforsupportingmethroughoutmylifeandforprovidingmewiththefoundationforwhat I have accomplished today. Finally, I dedicate my thesis to my grandparents for theirunconditionalsupportandlove.