temporal proteomic analysis of rat hippocampus provides...
TRANSCRIPT
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.