Numbers, densities, and colocalization of AMPA- and NMDA-type glutamate receptors at individual synapses in the superficial spinal dorsal horn of rats

Cellula /Molecula Numbe s, Densi ies, and Colocaliza ion o AMPA- and NMDA-Type Glu ama e Recep o s a Indi idual Synapses in he Supe icial Spinal Do sal Ho n o Ra s Miklo´s An al, 1 * Yugo Fukazawa, 4 *Ma´ ia Eo¨ do¨gh, 1 Do´ a Muszil, 1 Elek Molna´ , 2 Mako o I aku a, 3 Masami Takahashi, 3 and Ryuichi Shigemo o 4 1 Depa men o Ana omy, His ology, and Emb yology, Facul y o Medicine, Medical and Heal h Science Cen e , Uni e si y o Deb ecen, H-4012 Deb ecen, Hunga y, 2 Medical Resea ch Council, Cen e o Synap ic Plas ici y, Depa men o Ana omy, School o Medical Sciences, Uni e si y o B is ol, B is ol BS8 1TD, Uni ed Kingdom, 3 Depa men o Biochemis y, Ki asa o Uni e si y School o Medicine, Sagamiha a, Kanagawa 228-8555, Japan, and 4 Di ision o Ce eb al S uc u es, Na ional Ins i u es o Physiological Sciences, Myodaiji, Okazaki 444-8787, Japan Iono opic glu ama e ecep o s play impo an oles in spinal p ocessing o nocicep i e senso y signals and induc ion o cen al sensi- iza ion in ch onic pain. He e we applied highly sensi i e eeze- ac u e eplica labeling o laminae I–II o he spinal do sal ho n o a s and in es iga ed he numbe s, densi ies, and colocaliza ion o AMPA- and NMDA- ype glu ama e ecep o s a indi idual pos synap ic memb anespecializa ionswi hahigh esolu ion.Allglu ama e gicpos synap icmemb anesinlaminaeI–IIexp essedAMPA ecep o s, and mos o hem (96%) we e also immuno eac i e o he NR1 subuni o NMDA ecep o s. The numbe s o gold pa icles o AMPA and NMDA ecep o s a indi idual pos synap ic memb anes showed a linea co ela ion wi h he size o pos synap ic memb ane specializa- ions and a ied in he ange o 8–214 and 5–232 wi h median alues o 37 and 28, whe eas hei densi ies a ied in he ange o 325–3365/ ␮ m 2 and 102–2263/ ␮ m 2 wi h median alues o 1115/ ␮ m 2 and 777/ ␮ m 2 , espec i ely. Vi ually all (99%) glu ama e gic pos synap ic memb anes exp essed GluR2, and mos o hem (87%) we e also immuno eac i e o GluR1. The numbe s o gold pa icles o pan-AMPA,NR1,andGluR2 subuni s showed a linea co ela ion wi h he size o pos synap ic su ace a eas. Conce ning GluR1, he e may be wo popula ions o synapses wi h high and low GluR1 densi ies. In synapses la ge han 0.1 ␮ m 2 , GluR1 subuni s we e eco e ed in e ylownumbe s.Di e en ial exp ession o GluR1 and GluR2 subuni s sugges s egula ion o AMPA ecep o subuni composi ion by p esynap ic mechanism. Key wo ds: iono opic glu ama e ecep o s; NR1; GluR1; GluR2; molecula ana omy; pos synap ic ac i e zones; SDS-FRL In oduc ion Fas exci a o y neu o ansmission in he CNS is mainly media ed by glu ama e. In he supe icial spinal do sal ho n, glu ama e is eleased by nocicep i e p ima y a e en s and local axon e mi- nals o spinal exci a o y neu ons and, by ac i a ing iono opic and me abo opic glu ama e ecep o s, plays majo oles in no- cicep i e in o ma ion p ocessing. In case o hei epe i i e s im- ula ion, cen al e minals o nocicep i e p ima y a e en s dis- cha ge glu ama e oge he wi h neu opep ides and neu o ophins. The in e play o glu ama e, neu opep ide, and neu o ophin ecep o mechanisms hen esul s in inc eased Ca 2⫹ in lux in o he pos synap ic spinal neu ons and he conse- quen de elopmen o cen al sensi iza ion, which p esen s com- mon ea u es wi h he ea ly phase o long- e m po en ia ion (LTP). The p opaga ion o enhanced ac i i ies o spinal nocicep- i e ci cui s o highe b ain cen e s hen leads o hype algesia and a ious pain synd omes (Wool and Sal e , 2000). A long line o e idence indica es ha AMPA and NMDA ecep o s a e in ol ed in he induc ion o spinal LTP and consequen pain ( o e iew, see Ji e al., 2003). I has been demons a ed ex ensi ely ha AMPA and NMDA ecep o s a e exp essed in high densi ies in he supe icial spinal do sal ho n (Fu uyama e al., 1993; Pelleg ini-Giampie o e al., 1994; Tachibana e al., 1994; Jakowec e al., 1995a,b; Ha is e al., 1996; A. Pop a ilo e al., 1996, 1998; S. A. Pop a ilo e al., 1998; Ke e al., 1998). The majo i y o AMPA ecep o s in he do sal ho n consis o GluR2 and GluR1 subuni s ha may o m unc- ional AMPA ecep o ion channels in bo h homome ic and he - e ome ic a angemen s. In con as , he exp ession o GluR3 and GluR4 subuni s and hei mRNAs was ound o be weak (Sa o e al., 1993; To¨lle e al., 1993, 1995). Al hough a long line o expe imen s ha e been pe o med o de ec AMPA and NMDA ecep o s in he spinal do sal ho n, Recei ed Dec. 19, 2007; e ised Aug. 15, 2008; accep ed Aug. 19, 2008. This wo k was suppo ed by he Hunga ian Na ional Resea ch Fund (OTKA 46121; M.A.), Hunga ian Scien i ic Council o Heal h (ETT 079/2006; M.A.), Hunga ian Academy o Sciences (MTA-TKI 242; M.A.), Solu ion O ien ed Resea ch o Science and Technology om he Japan Science and Technology Agency (R.S.), Minis y o Educa ion, Cul u e, Spo s, Science, and Technology (Y.F., R.S.), and he Uni ed Kingdom Medical Resea ch Council (E.M.). *M.A. and Y.F. con ibu ed equally o his wo k. Co espondence should be add essed o Miklo´s An al, Depa men o Ana omy, His ology, and Emb yology, Facul yo Medicine,Medical andHeal h ScienceCen e , Uni e si yo Deb ecen,Nagye dei k 98, H-4012Deb ecen, Hunga y. E-mail: [email p o ec ed]. DOI:10.1523/JNEUROSCI.1551-08.2008 Copy igh © 2008 Socie y o Neu oscience 0270-6474/08/289692-10$15.00/0 9692 •The Jou nal o Neu oscience, Sep embe 24, 2008 •28(39):9692–9701 hese s udies could e eal he ecep o s mos ly, i no exclusi ely, in he cy oplasm and no a synap ic apposi ions. Pos synap ic memb anes a e poo ly accessible o an ibodies in p eembedding immunos aining p o ocols because o he p esence o ex ensi e p o ein meshwo k in he synap ic cle and pos synap ic densi y (Baude e al., 1995; O e sen and Landsend, 1997; F i schy e al., 1998; Wa anabe e al., 1998). Al hough pos embedding immu- nogold labeling has also been used o in es iga e synap ic AMPA and NMDA ecep o s in he spinal co d (A. Pop a ilo e al., 1996, 1998; S. A. Pop a ilo e al., 1998; Ragna son e al., 2003), he pa e n o AMPA and NMDA ecep o exp ession a indi id- ual synapses emains la gely elusi e. Recen ly an SDS-diges ed eeze- ac u e eplica labeling (SDS-FRL) me hod was de eloped o in es iga e AMPA ecep- o s a indi idual synapses (Tanaka e al., 2005; Masugi-Toki a e al., 2007). By emo ing a ached molecules om he ac u ed plasma memb anes, hus making in amemb ane p o eins di- ec ly accessible o an ibodies, he SDS-FRL me hod seems o be ideal o immunocy ochemical in es iga ion o memb ane- bound molecules, including neu o ansmi e ecep o s (Masugi-Toki a and Shigemo o, 2007). In he p esen s udy, we applied he SDS-FRL me hod o AMPA and NMDA ecep o s in he supe icial spinal do sal ho n o a s, which allowed us o examine he numbe s, densi ies, and colocaliza ion o hese e- cep o s a indi idual pos synap ic memb anes wi h a high esolu ion. Ma e ials and Me hods Animals and p epa a ion o issue sec ions. Expe imen s we e pe o med on h ee adul male Wis a a s, wo GluR2 knock-ou (Ha mann e al., 2004), one GluR1 knock-ou (Ha mann e al., 2004), and wo wild- ype mice. Ca e and handling o he animals be o e and du ing he expe i- men s ollowed Eu opean Union and Japanese egula ions and was ap- p o ed by he animal ca e and use commi ees o he au ho s’ ins i u- ions. The animals we e deeply anes he ized by sodium pen oba bi al (3.5 mg/100 g o body weigh , i.p.), and ansca dially pe used wi h iso onic saline solu ion o 10–15 min ollowed by a ixa i e con aining 2% pa a o maldehyde and 15% sa u a ed pic ic acid dissol ed in 0.1 M phospha e bu e (PB; pH 7.4). The pe usion wi h he ixa i e las ed o 30–40 min, and hen he lumba segmen s (L2–L5) o he spinal co d we e immedia ely excised and hemisec ed along he midline. The he- misec ed co ds we e ans e ed in o 0.1 MPB, pH 7.4, o 1–2 h. Then 150- ␮ m- hick pa asagi al sec ions we e cu om he hemisec ed co ds wi h a ib a ome. A e ex ensi e washes in 0.1 MPB, he sec ions we e c yop o ec ed in 30% glyce ol dissol ed in 0.1 MPB o e nigh . The ixed issues we e kep a 4°C un il hey we e c yop ese ed by eezing. SDS-diges ed eeze- ac u e eplica labeling. The sec ions we e ozen quickly by a high-p essu e eezing machine (HPM 010; Bal-Tec). The ozen slices we e hen eeze ac u ed and coa ed wi h ca bon (5 nm), shadowed by pla inum (2 nm), and hen coa ed wi h ca bon (15 nm) again in BAF 060 (Bal-Tec). A e hawing, issue deb is a ached o he eplicas was diges ed wi h gen le s i ing a 80°C o e nigh in a solu ion con aining 2.5% SDS and 20% suc ose dissol ed in 15 mMT is bu e , pH 8.3. The eplicas we e washed in 25 mMT is-bu e ed iso onic saline (TBS; pH 7.4) washing bu e con aining 0.05% bo ine se um albumin (BSA) and incuba ed in a blocking solu ion con aining 5% BSA in 25 mM TBS o 1 h. Subsequen ly, he eplicas we e eac ed wi h he ollowing p ima y an ibodies: (1) pan-AMPA ecep o (GluR1–4) an ibody aised in abbi (1 ␮ g/ml) (Nusse e al., 1998), (2) mouse monoclonal an ibody agains he NR1 subuni o NMDA ecep o (1 ␮ g/ml; Millipo e Bio- science Resea ch Reagen s), (3) mix u e o he pan-AMPA and NR1 an ibodies, (4) an ibody agains he GluR1 subuni o AMPA ecep o aised in abbi (8 ␮ g/ml), o (5) mouse monoclonal an ibody agains he GluR2 subuni o AMPA ecep o s (15 ␮ g/ml; Millipo e Bioscience Re- sea ch Reagen s), dilu ed in 25 mMTBS con aining 1% BSA a ⫹15°C (in case o 1–3) o a oom empe a u e (in case o 4–5) o 2 d. The abbi an ibody agains GluR1 was aised agains an ex acellula epi ope (amino acid esidues 345–362) o he a GluR1 (C- RFEGLTGNVQFNEKGRRT), ollowed by a ini y pu i ica ion wi h he same pep ide. Subsequen ly, he an ibody was abso bed wice by GluR2- speci ic pep ide (C-QVEGLSGNIKFDQNGKRI) o a oid undesi ed con amina ion o GluR2- eac ing an ibodies. A e se e al washes, he eplicas we e eac ed wi h goa an i- abbi ( o pan-AMPA and GluR1) and goa an i-mouse ( o NR1 and GluR2) IgGs coupled o 5 nm gold pa icles (1:30; BioCell Resea ch Labo a o ies) dilu ed in 25 mMTBS con aining 5% BSA o e nigh a oom empe a u e. In he double- labeling p o ocols, one o he seconda y an ibodies was conjuga ed o 5 nm gold pa icle, whe eas he o he was coupled o 10 nm gold pa icle. The eplicas we e hen washed, picked up on 100-mesh g ids (Fig. 1), examined wi h an elec on mic oscope (JEOL1010), and pho og aphed a a magni ica ion o 40,000⫻. Con ols. To es he speci ici y o an ibodies aised agains GluR1 and GluR2 subuni s o AMPA ecep o s, eplicas o spinal co d sec ions ob- ained om GluR1 and GluR2 knock-ou and wild- ype mice we e im- munolabeled acco ding o he p ocedu e desc ibed abo e. Replicas o GluR2 knock-ou mice we e almos nega i e o GluR2 in he pos syn- ap ic memb ane specializa ion. On he a e age, he densi y alue o GluR2 in he knock-ou mice was 3.3 pa icles/ ␮ m 2 , whe eas ha in he wild- ype animals was 219 pa icles/ ␮ m 2 . On he o he hand, he an i- body aised agains GluR1 appea ed o be highly speci ic and eac ed selec i ely wi h GluR1 subuni s in Wes e n blo analysis (supplemen al Fig. 1, a ailable a www.jneu osci.o g as supplemen al ma e ial) How- e e , labeling densi y o GluR1 a indi idual pos synap ic memb ane specializa ions in wild ype (144 ⫾13.9; n⫽58) was only h ee imes highe han ha in knock-ou mice (54 ⫾7.1; n⫽50), al hough he di e ence was s a is ically signi ican ( p⬍0.001, Mann–Whi ney U es ) (supplemen al Fig. 2, a ailable a www.jneu osci.o g as supplemen al ma e ial). The e o e, we sub ac ed he densi y o non-GluR1 labeling om he labeling in wild ype, hus making he densi y 63% o he o ig- inal alue. Speci ici y o p ima y an ibodies agains NR1 and pan-AMPA was ex ensi ely cha ac e ized p e iously (Siegel e al., 1994; Nusse e al., 1998; Masugi-Toki a e al., 2007). To es he speci ici y o he immunolabeling p ocedu e, eplicas we e incuba ed acco ding o he p o ocol desc ibed abo e wi h p ima y an i- bodies omi ed o eplaced wi h 1% no mal goa se um. Labeling densi- ies on clus e s o in amemb ane pa icles we e ⬍2.8 pa icles/ ␮ m 2 in hese cases. Quan i a i e e alua ion o he immunolabeling. Immunogold pa icles we e coun ed in exci a o y pos synap ic a eas indica ed by clus e s o in amemb ane pa icles (IMP clus e s) (Ha is and Landis, 1986) on he exoplasmic ac u e ace (E- ace) o plasma memb anes. Agg ega ions o Figu e 1. Pho omic og aph o an SDS-FRL eplica on a 100-mesh coppe g id. The eplica was p epa ed om a pa amedian–sagi al sec ion o he do sal aspec o he a lumba spinal co d. Whi e lines indica e he p esumed ou e and inne bo de s o laminae I–II. Scale ba , 100 ␮ m. An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 • 9693 in amemb ane pa icles we e ega ded as IMP clus e s i he pa icles showed a compac a - angemen in which he dis ances among adja- cen pa icles we e no la ge han 15 nm. We conside ed an IMP clus e as pos synap ic memb ane specializa ion i he clus e con- ained a leas 30 pa icles. Con i ming he ind- ings o Masugi-Toki a e al. (2007), all o he pos synap ic a eas on E- ace we e labeled wi h he an ibody aised agains pan-AMPA. Immu- nogold pa icles we e ega ded as associa ed wi h he pos synap ic memb ane specializa ion i hey we e abo e o in he immedia e icini y (no u he han 20 nm om he edge) o he IMP clus e (Ma suba a e al., 1996). Measu emen s we e pe o med in h ee ani- mals, and esul s we e pooled because he den- si y o immunogold pa icles on IMP clus e s was no signi ican ly di e en in he di e en animals. Immunogold pa icles on he p o o- plasmic ac u e ace (P- ace) o plasma mem- b anes should no be labeled wi h an ibodies di ec ed agains ex acellula epi opes, and hus we e de ined as backg ound labeling, which u ned ou o be 3.2 pa icles/ ␮ m 2 on he a e - age. The ex en o IMP clus e s on elec on mi- c og aphs was calib a ed by using a calib a ion g id (Ted Pella). The ou line o pos synap ic ac i e zones (IMP clus e s) was dema ca ed eehand, and he a ea o synap ic si es was measu ed by Scion Image. The numbe s o gold pa icles ma king ecep o molecules a synap- ic si es we e coun ed manually. S a is ical analysis o dis ibu ion pa e n o immunogold pa icles. To e alua e he dis ibu ion pa e n o gold pa i- cles labeling ecep o s wi hin a synapse, we used poin pa e n analysis using Ripley’s K- unc ion (Ripley, 1977, 1979; P io e al., 2003). Elec- on mic og aphs o six andomly selec ed IMP clus e s labeled o NR1, pan-AMPAR, GluR1, and GluR2 we e digi ized wi h an image scanne . A ec angula a ea ha co e ed a leas 50% o pa icles a indi idual IMP clus e s was selec ed, and he x–y coo dina es o indi idual pa icles in he a ea we e ob ained by iTEM image analysis so wa e (So Imaging Sys em). Dis ibu ion pa e ns o immunogold pa icles we e analyzed wi h he L( )⫺ alues, which exp ess he second-o de spa ial pa e n o pa icle dis ibu ion based on he in e pa icle dis ances wi hin he s udied a ea, by using a p og am kindly p o ided by John Hancock, Uni e si y o Queensland, B isbane, Aus alia (P io e al., 2003). To de ine whe he he dis ibu ion pa e ns o immunogold pa icles we e homogeneous o clus e ed, 99% con idence en elopes o comple e spa- ial andomness (CSR) we e gene a ed om 100 Mon e Ca lo simula- ions and plo ed wi h he L( )⫺ cu e (see supplemen al Fig. 3, a ail- able a www.jneu osci.o g as supplemen al ma e ial). When he L( )⫺ cu e was abo e he en elope a mo e han one poin , dis ibu ion was classi ied as clus e ed. When all poin s o he L( )⫺ cu e we e below he en elope, i s dis ibu ion was classi ied as CSR and as homogeneous. Fo mo e de ailed in o ma ion abou he applica ion o his s a is ical app oach o immunopa icle-labeled eplicas, please see Fuji a e al. (2007). Resul s Numbe s and densi ies o AMPA ecep o s a indi idual synap ic con ac a eas Using he SDS-FRL me hod, we analyzed he numbe s, dis ibu- ion, and densi ies o AMPA ecep o s a indi idual synap ic con- ac a eas in he supe icial spinal do sal ho n o a s wi h an an ibody agains highly conse ed ex acellula amino acid esi- dues o GluR1–4 (pan-AMPA). The pan-AMPA an ibody has been shown o eac selec i ely wi h all AMPA ecep o subuni s GluR1–4, bu no wi h kaina e ecep o subuni s (Nusse e al., 1998). The bo de be ween he do sal column and he do sal ho n was easily iden i ied on he eplicas wi h bo h ligh and elec on mic oscopy (Fig. 1). The bo de be ween laminae II and III was also clea because almos no myelina ed axons we e ound in lamina II, bu many we e in lamina III. Immuno eac i i y o IMP clus e s on E- ace p o iles o he eplicas was in es iga ed in a band be ween hese wo bounda ies (Fig. 1), in a zone ha had ea lie been iden i ied as a laye o he g ay ma e co esponding o laminae I and II o he spinal do sal ho n (McClung and Cas- o, 1978; Molande e al., 1984; McNeill e al., 1988). Pos synap ic memb ane specializa ions we e clea ly indica ed by agg ega ions o IMPs on E- aces (see Figs. 2, 5, 6), as desc ibed p e iously (Landis and Reese, 1974; Ha is and Landis, 1986). We de ined IMP clus e s as hose con aining a leas 30 in amem- b ane pa icles. Immunogold pa icles o AMPA ecep o s we e ound almos exclusi ely on IMP clus e s, and all IMP clus e s on E- ace p o iles we e immuno eac i e o he pan-AMPA an i- body (Fig. 2a,c). In he majo i y o pos synap ic memb ane spe- cializa ions, immunogold pa icles we e dis ibu ed o e he en- i e ield o indi idual IMP clus e s wi h no appa en clus e ing (Fig. 2a; supplemen al Fig. 2, a ailable a www.jneu osci.o g as supplemen al ma e ial), al hough he densi y o labeling a ied conside ably om clus e o clus e (Fig. 3a). The numbe o immunogold pa icles was coun ed in bo h comple e (n⫽97) and incomple e (pa ially ac u ed; n⫽65) pos synap ic mem- b ane specializa ions. In comple e pos synap ic memb ane spe- cializa ions, he numbe o gold pa icles a ied in he ange o 8–214 wi h a median alue o 37. The numbe o gold pa icles labeling AMPA ecep o showed a linea co ela ion ( ⫽0.86) wi h he size o he pos synap ic su ace a eas (Fig. 4a). Because Figu e 2. a– d, Elec on mic og aphs o SDS-FRL eplicas immunolabeled wi h an an ibody ha ecognizes all subuni s o AMPA- ype glu ama e ecep o s (pan-AMPA) (a,c) o wi h an an ibody aised agains he NR1 subuni o NMDA- ype glu ama e ecep o s (b,d). The mic og aphs illus a e pos synap ic memb ane specializa ions (IMP clus e s) ha show s ong immuno e- ac i i y o pan-AMPA (a,c) o NR1 (b,d) in he supe icial spinal do sal ho n o a s. All subuni s a e labeled wi h 5 nm gold pa icles. Scale ba s, 0.1 ␮ m. 9694 •J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s densi ies o immunogold labeling o he pan-AMPA an ibody ob ained om comple e and incomple e pos synap ic ac i e zones we e no signi ican ly di e en (supplemen al Fig. 4, a ail- able a www.jneu osci.o g as supplemen al ma e ial), hey we e pooled. The densi y alues a ied in he ange o 325–3365 pa - icles/ ␮ m 2 wi h a median o 1115 pa icles/ ␮ m 2 (Fig. 3a). The mean densi y alues o he ex asynap ic si es on E- ace and backg ound es ima ed on P- ace we e 2.7 and 3.2 pa icles/ ␮ m 2 , espec i ely. Numbe s and densi ies o NMDA ecep o s a indi idual synap ic con ac a eas To in es iga e he numbe s, dis ibu ion, and densi ies o NMDA ecep o s a indi idual synap ic con ac a eas in he supe icial spinal do sal ho n, eplicas we e eac ed wi h a monoclonal an- ibody agains he NR1 subuni , which is a cons i uen o all unc ional NMDA ecep o ion channels [ o e iew, see Cull- Candy e al. (2001) and Gibb (2004)]. The an ibody was di ec ed agains a usion p o ein co esponding o amino acid esidues 660–811, ep esen ing he ex acellula loop be ween ansmem- b ane egions III and IV o he NR1 subuni (Siegel e al., 1994). I has been ex ensi ely documen ed ha he an ibody is highly spe- ci ic and eac s selec i ely wi h NMDA ecep o s (Siegel e al., 1994). Immunogold pa icles o NR1 we e ound almos exclusi ely on IMP clus e s. The majo i y (96%) o IMP clus e s ha con- ained a leas 30 in amemb ane pa icles and hus we e e- ga ded as pos synap ic memb ane specializa ion on E- ace p o- iles we e immuno eac i e o he NR1 an ibody (Fig. 2b,d). Simila ly o he AMPA ecep o labeling, immunogold pa icles we e andomly dis ibu ed o e he en i e ield o IMP clus e s wi hou o ming clus e s, al hough he densi y o labeling a ied om clus e o clus e (Fig. 3b; supple- men al Fig. 3, a ailable a www.jneu osci. o g as supplemen al ma e ial). In com- ple e pos synap ic memb ane specializa- ions (n⫽99), he numbe o gold pa i- cles a ied in he ange o 5–232 wi h a median alue o 28. The numbe o gold pa icles o NR1 showed a linea co ela- ion ( ⫽0.87) wi h he size o he pos syn- ap ic su ace a eas (Fig. 4b). Because o he andom dis ibu ion o gold pa icles (supplemen al Fig. 3, a ailable a www. jneu osci.o g as supplemen al ma e ial), da a ob ained om comple e (n⫽99) and incomple e (n⫽48) pos synap ic mem- b ane specializa ion we e pooled o he calcula ion o densi ies o immunogold la- beling o he NR1 an ibody. The densi y alues a ied in he ange o 102–2263 pa - icles/ ␮ m 2 wi h a median o 777 pa icles/ ␮ m 2 (Fig. 3b). The shape o he densi y dis ibu ion his og am o NR1 labeling was e y simila o ha o pan-AMPA labeling. Colocaliza ion o AMPA- and NMDA- ype glu ama e ecep o s a indi idual synap ic con ac a eas A synch onized ain o epea ed nocicep- i e inpu e okes a pe iod o acili a ed ansmission in spinal do sal ho n neu- ons ha esembles LTP in o he pa s o he CNS. This ac i i y- dependen o m o cen al sensi iza ion (spinal co d equi alen o LTP) is he consequence o ac i a ion o mul iple in acellula signaling pa hways ha in ol es ac i a ion o bo h AMPA- and NMDA- ype glu ama e iono opic ecep o s [ o e iew, see Wol and Sal e (2000) and Ji e al. (2003)]. Because o i s unc- ional impo ance, we in es iga ed he colocaliza ion o AMPA and NMDA ecep o s a indi idual pos synap ic memb ane spe- cializa ions by simul aneous double labeling o eplicas wi h pan- AMPA and NR1 an ibodies. The colocaliza ion s udies appea ed o be e y consis en . Al- hough, because o spa ial compe i ion among he simul a- neously applied an ibodies, he signal in ensi y in double labeling o a gi en an ibody was app oxima ely one-hal o ha in single labeling, we ound e y simila dis ibu ion pa e ns o densi y in single and double labeling (da a no shown). Rega dless o whe he AMPA ecep o s we e labeled wi h 5 nm and NR1 sub- uni s wi h 10 nm gold pa icles o ice e sa, he majo i y o comple e pos synap ic memb ane specializa ions (179 o 186; 96.2%) displayed posi i e labeling o bo h ecep o s (Fig. 5a,b, Table 1). Se en o he 186 pos synap ic memb anes (3.8%) in es- iga ed in his s udy we e posi i e o pan-AMPA bu nega i e o he NR1 (Fig. 5c, Table 1). Howe e , pos synap ic memb anes posi i e o NR1 and nega i e o pan-AMPA we e no de ec ed (Table 1). Numbe s and densi ies o GluR2 subuni s a indi idual synap ic con ac a eas A long line o expe imen al e idence sugges s ha he mos p ominen ly exp essed AMPA ecep o subuni in he spinal do - sal ho n is GluR2 (Fu uyama e al., 1993; Henley e al., 1993; To¨lle e al., 1993, 1995). To in es iga e he dis ibu ion and densi ies o Figu e3. a– d,His og amsshowing hedis ibu iono densi ieso goldpa icles ha label a ioussubuni so AMPA-(a,c,d) and NMDA- (b) ype glu ama e ecep o s in SDS-FRL eplicas o indi idual pos synap ic memb ane specializa ions in he supe i- cial spinal do sal ho n o a s. Replicas om which da a we e ob ained we e immunolabeled wi h an ibodies ha ecognize all subuni s(pan-AMPA)(a) o AMPA ecep o s, NR1 subuni s o NMDA ecep o s(b),and GluR2 (c) and GluR1 (d) subuni s o AMPA ecep o s. An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 • 9695 GluR2 subuni s a indi idual synap ic con ac a eas in he supe icial spinal do - sal ho n o a s, eplicas we e eac ed wi h a monoclonal an ibody di ec ed agains a ecombinan usion p o ein co espond- ing o amino acid esidues 175–430, ep- esen ing he N- e minal po ion o he e- cep o molecule. The an ibody p o ed o be highly speci ic and appea ed o eac selec i ely wi h GluR2 subuni s, because he applica ion o he an ibody o eplicas o GluR2 knock-ou mice did no esul in any labeling (Fig. 6a), whe eas s ong im- munolabeling was obse ed a IMP clus- e s on eplicas o wild- ype animals (Fig. 6b). Immunogold pa icles o GluR2 sub- uni s we e ound almos exclusi ely on IMP clus e s, and i ually all IMP clus e s on E- ace p o iles we e immuno eac i e o he GluR2 an ibody (Fig. 6c). The en- i e ield o IMP clus e s was densely bu andomly labeled wi hou clus e ing (sup- plemen al Fig. 3, a ailable a www. jneu osci.o g as supplemen al ma e ial), al hough he densi y o immunogold pa - icles o e indi idual IMP clus e s was e y a iable (coe icien o a ia ion ⫽0.437) (Fig. 3c). In comple e pos synap ic ac i e zones (n⫽62), he numbe o gold pa i- cles a ied in he ange o 4–162 wi h a median alue o 21. The numbe o gold pa icles o GluR2 showed a linea co ela ion ( ⫽0.92) wi h he size o pos synap ic su ace a eas (Fig. 4c). Because o he andom dis ibu ion o gold pa icles, da a ob- ained om comple e (n⫽62) and incomple e (n⫽24) pos syn- ap ic memb ane specializa ions we e pooled o he calcula ion o densi ies o immunogold pa icles o he GluR2 an ibody. The densi y alues a ied in he ange o 144–1422 pa icles/ ␮ m 2 wi h a median o 733 pa icles/ ␮ m 2 . The shape o he densi y dis ibu ion his og am o GluR2 labeling showed a p ominen peak a he median alue (Fig. 3c). Numbe s and densi ies o GluR1 subuni s a indi idual synap ic con ac a eas Wi h a li le lag behind GluR2, he second mos abundan ly ex- p essed AMPA ecep o subuni in he supe icial spinal do sal ho n is GluR1 (Fu uyama e al., 1993; To¨lle e al., 1993; Pop a il- o e al., 1996; Engelman e al., 1999; B own e al., 2002; Ha - mann e al., 2004). Analogous o he p edominan ole o GluR1 in hippocampal LTP (Zamanillo e al., 1999) and spa ial memo y consolida ion (Reisel e al., 2002), he p esence o GluR1 subuni is equi ed also o a apid sensi iza ion o glu ama e gic synap ic e en s in he spinal do sal ho n (Ha mann e al., 2004). To in- es iga e he dis ibu ion o GluR1 subuni s, eplicas we e eac ed wi h an an ibody ha was di ec ed agains esidues 345–362, ep- esen ing he pu a i e N- e minal ex acellula po ion o GluR1. The an ibody appea ed o be highly speci ic and eac ed selec- i ely wi h GluR1 subuni s in Wes e n blo analysis (supplemen- al Fig. 1, a ailable a www.jneu osci.o g as supplemen al ma e- ial). Howe e , he applica ion o he an ibody o eplicas o GluR1 knock-ou mice showed subs an ial amoun o labeling a IMP clus e s (supplemen al Fig. 2, a ailable a www.jneu osci. o g as supplemen al ma e ial). The e o e we sub ac ed he e- maining densi y o labeling in GluR1 knock-ou animals om he labeling obse ed in a samples. Thus, we educed he numbe s o gold pa icles and densi y o labeling o 63% o he o iginal alues (see Ma e ials and Me hods). These educed alues can be ound in he nex pa ag aph. Immunogold pa icles o GluR1 subuni s we e ound almos exclusi ely on IMP clus e s, and 87% o IMP clus e s on E- ace p o iles we e conside ed immuno eac i e o GluR1 (Fig. 6 ). The en i e ield o IMP clus e s was andomly labeled, and no clus e ing o immunopa icles was de ec ed (supplemen al Fig. 3, a ailable a www.jneu osci.o g as supplemen al ma e ial), al- hough he densi y o immunogold pa icles o e indi idual IMP clus e s was e y a iable (coe icien o a ia ion ⫽0.67). In comple e pos synap ic ac i e zones (n⫽66), he numbe o gold pa icles a ied in he ange o 2–47 wi h a median alue o 9. The numbe o gold pa icles o GluR1 showed a less linea co ela- ion ( ⫽0.14) han ha o GluR2 wi h he size o he pos syn- ap ic su ace a eas. Sea ching o an explana ion o his poo linea co ela ion, we ound ha he e migh be wo popula ions o synapses, one wi h high and ano he wi h e y low GluR1 densi ies. When we selec ed synapses la ge han 0.06 ␮ m 2 , hey could be clea ly sepa a ed in o wo g oups, one wi h a high mean densi y o GluR1 (240–696 pa icles/ ␮ m 2 wi h a median o 349 pa icles/ ␮ m 2 ) and ano he wi h ⬍10 pa icles/synapse (12–96 pa icles/ ␮ m 2 wi h a median o 52 pa icles/ ␮ m 2 ) (labeled wi h open ci cles in Fig. 4d). Synapses la ge han 0.1 ␮ m 2 we e all wi hin he second g oup wi h e y low pa icle densi y. Sepa a - ing all synapses ha showed low pa icle densi y and we e la ge han 0.06 ␮ m 2 (labeled wi h open ci cles in Fig. 4d) om he o al popula ion, he numbe o gold pa icles o GluR1 in he es o he synapses showed a nice linea co ela ion ( ⫽0.74) wi h he size o pos synap ic su ace a eas (Fig. 4d). Figu e 4. a– d, Sca e plo s showing he co ela ion be ween su ace a eas o pos synap ic memb ane specializa ions and numbe so gold pa icleslabelingallsubuni s(pan-AMPA)(a)o AMPA ecep o s,NR1subuni o NMDA ecep o s(b),andGluR2 (c) and GluR1 (d) subuni s o AMPA ecep o s. Open ci cles in dlabel da a om pos synap ic memb ane specializa ions ha a e la ge han 0.06 ␮ m 2 andp esen lownumbe s o gold pa icles.These synapseswe eexcluded om hecalcula iono he alue o linea co ela ion. 9696 •J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s Sizes o pos synap ic memb ane specializa ions Finally, we pooled all comple e pos synap ic memb ane special- iza ions ha showed any kind o posi i e labeling o AMPA and/o NMDA ecep o subuni s. Thus we collec ed 740 pos syn- ap ic memb ane specializa ions o glu ama e gic synap ic con- ac s in laminae I–II o he spinal do sal ho n o a s, and mea- su ed hei su ace a eas by Scion Image. Al hough he ypes o neu ons ha p esen ed he pos synap ic memb ane specializa- ions as well as he o igin o p esynap ic axon e minals ha we e associa ed wi h he in es iga ed pos synap ic memb anes could be e y he e ogeneous in ou sample, he dis ibu ion his og am o su ace a eas o pos synap ic memb ane specializa ions showed only one peak (Fig. 7). The su ace a eas o indi idual pos synap ic memb ane specializa ions a ied in he ange o 0.0026–0.3609 ␮ m 2 (median, 0.0419 ␮ m 2 ); only 26 (3.5%) o he 740 pos synap ic memb ane specializa ions we e la ge han 0.15 ␮ m 2 . Discussion Applying he SDS-FRL me hod, which has almos one gold pa - icle–one unc ional channel sensi i i y (Tanaka e al., 2005) o laminae I–II o he spinal do sal ho n o a s, we showed ha i ually all glu ama e gic pos synap ic memb anes in laminae I–II exp essed AMPA and NMDA ecep o s. Fu he - mo e, mos o hem we e immuno eac i e o GluR2, whe eas GluR1 is exp essed in selec i e popula ions o he synapses. In synapses la ge han 0.1 ␮ m 2 , GluR1 sub- uni s we e eco e ed in e y low numbe s. This di e en ial exp ession o GluR1 and GluR2 subuni s was encoun e ed e en a synapses ha we e loca ed in a close icin- i y o each o he on he e y same neu on (supplemen al Fig. 5, a ailable a www. jneu osci.o g as supplemen al ma e ial). Ne e heless, he numbe s o gold pa i- cles o AMPA and NMDA ecep o s showed linea co ela ions wi h he size o pos synap ic su ace a eas. I is e y likely ha in ou p esen s udy we in es iga ed he pos synap ic mem- b ane specializa ions o a qui e he e oge- neous popula ion o synap ic con ac s. C- and A ␦ - ype p ima y a e en s as well as exci a o y local in e neu ons all o m glu- ama e gic synap ic con ac s wi h a ious ypes o neu ons in laminae I–II o he spi- nal do sal ho n. Despi e his he e ogene- i y, howe e , he densi ies o AMPA and NMDA ecep o subuni s as well as su ace a eas o he pos synap ic memb ane spe- cializa ions showed homogeneous dis i- bu ions. Synapses could be clea ly di ided in o wo g oups only on he basis o hei GluR1 subuni densi ies. Size o indi idual pos synap ic memb ane specializa ions Synap ic s uc u e plays a key ole in syn- ap ic ansmission (Walmsley e al., 1998). In addi ion o p esynap ic mechanisms, pos synap ic esponses may be in luenced by he numbe and densi y o pos synap ic ecep o s, which in u n may be ela ed o he size o he pos syn- ap ic memb ane specializa ions (Nusse e al., 1997; Lim e al., 1999; Mackenzie e al., 1999; Nusse , 2000). Because o hei unc ional impo ance, mo phological pa ame e s o pos synap- ic ac i e zones ha e ex ensi ely been s udied in a ious b ain egions, and a conside able a iabili y has been obse ed (An al e al., 1992; Pie ce and Mendel, 1993; Ryugo e al., 1996; Taschen- be g e al., 2002). Fo ins ance, he sizes o pos synap ic mem- b ane specializa ions a ce ebella climbing ibe and pa allel i- be synapses a e among he la ges on a e age (0.14 ␮ m 2 ) (Xu- F iedman e al., 2001), whe eas his alue has been de ined a a ound only 0.04–0.07 ␮ m 2 on CA1 dend i ic spines (Schiko ski and S e ens, 1997; Shephe d and Ha is, 1998). The size dis i- bu ion o indi idual pos synap ic memb ane specializa ions ha we encoun e ed in laminae I–II o he spinal do sal ho n o a s appea s o be simila o he ones ha ha e been obse ed on CA1 dend i ic spines. Numbe s and densi ies o AMPA and NMDA ecep o s The numbe o iono opic ecep o s in synapses is an essen ial ac o o de e mining he e icacy o as neu al ansmission. As Figu e5. Elec on mic og aphs o SDS-FRL eplicassimul aneouslydouble immunolabeled wi h an ibodies aisedagains he NR1 subuni o NMDA- ype glu ama e ecep o s and an amino acid sequence ha is common in all subuni s o AMPA- ype glu ama e ecep o s(pan-AMPA).a– c,AMPA ecep o s a elabeledwi h10nmgoldpa icles,andNR1subuni sa e ma kedwi h 5 nm gold pa icles in a, whe eas 5 nm gold pa icles label AMPA ecep o s, and 10 nm gold pa icles ma k NR1 subuni s in band c. Pos synap ic memb ane specializa ions in he mic og aphs in aand bshow immuno eac i i y o bo h an ibodies, whe eas pos synap icmemb anespecializa ionsin hemic og aphinca eposi i e o pan-AMPAbu nega i e o NR1.Scaleba s,0.1 ␮ m. An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 • 9697 a gene al ule o bo h glu ama e gic and GABAe gic iono opic ecep o s, i appea s ha he numbe o unc ional ecep o s is p o- po ional o he synap ic a ea in a pa icula ype o synapse. The ecep o densi y, howe e , shows a g ea deal o a iabili y, which mayde i e om di e ence in ypes o synapsesin es iga ed o quan- i a i e me hods used in he di e en s udies (Nusse e al., 1998; Somogyi e al., 1998a; Nusse , 1999; Tanaka e al., 2005; Masugi- Toki a e al., 2007). The a e age immunogold pa icle densi y ha we obse ed in laminae I–II o he spinal do sal ho n o AMPA ecep o s was close o he igu e ha was ob ained a synap ic con- ac s be ween mossy ibe s and ce ebella g anule cells (⬃1000 e- cep o s/ ␮ m 2 ) (Sil e e al., 1996). Howe e , i is impo an o no e ha we in es iga ed a he e ogeneous popula ion o synap ic con ac s es ablished by nocicep i e p ima y a e en s, segmen al and in e - segmen al p op iospinal axons wi h a ious spinal neu ons ha may possess AMPA and NMDA ecep o s in di e en quan i ies. Ac u- ally, ou esul s sugges ha he densi ies o AMPA and NMDA ecep o s a indi idual pos synap ic memb ane specializa ions o a iousg oupso synap ic apposi ions in he supe icialspinal do sal ho n o a s can eally be he e ogeneous. Acco ding o he non- Gaussiandis ibu ion o immunogold pa icledensi ies (Fig. 4a), he densi y alues may a y om400–500 o 1800–2000 ecep o s/ ␮ m 2 o AMPA ecep o s. Simila a ia ion was also ound o NMDA ecep o densi y. Figu e 6. Elec on mic og aphs o SDS-FRL eplicas immunolabeled wi h an ibodies ha ecognize GluR2 o GluR1 subuni s. a– , The mic og aphs illus a e pos synap ic memb ane specializa- ions(IMP clus e s) in he supe icial spinaldo sal ho n o GluR2 (a)andGluR1 (d)knock-ou mice andwild- ype(w ) mice(b,e)and a s(c, ).SDS-FRL eplicas omGluR2 knock-ou mice(a) and wild- ypemice(b)and a s(c)we elabeled o GluR2.Thepos synap icmemb anespecializa ionsa e eeo labelingin heGluR2knock-ou animal(a),andshows ongimmuno eac i i y o GluR2 in he wild- ype mice (b) and a s (c). SDS-FRL eplicas om GluR1 knock-ou (d) and wild- ype (e) mice we e double labeled o GluR1 and GluR2, whe eas eplicas om a a we e labeled only o GluR1 ( ). The pos synap ic memb ane specializa ions om GluR1 knock-ou animals a e posi i ely labeled o GluR2 bu nega i e o GluR1 (d). The pos synap ic memb ane specializa ion om wild- ype mice a e posi i e o bo h GluR1 and GluR2 (e), whe eas pos synap ic memb ane specializa ion om a a show s ong immuno eac i i y o GluR1 ( ). GluR2 subuni s we e labeled wi h 10 nm (b,d,e)and5nm(c), whe eas GluR1 subuni s we e labeled wi h 5 nm (d,e) and 10 nm ( ) gold pa icles. Scale ba s, 0.1 ␮ m. Table 1. Numbe s and pe cen age o comple e pos synap ic memb ane specializa ions immuno eac i e o pan-AMPA and/o NR1 subuni o NMDA ecep o in he supe icial spinal do sal ho n pan-AMPA⫹, NR1⫹pan-AMPA⫹, NR1⫺pan-AMPA⫺, NR1⫹To al pan-AMPA–10 nm, NR1–5 nm 89 3 0 92 pan-AMPA–5 nm, NR1–10 nm 90 4 0 94 To al 179 7 0 186 (96.2%) (3.8%) (100%) 9698 •J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s I is also impo an o no e ha ex asynap ic memb ane com- pa men s showed ema kably weak, i any, immunolabeling o bo h AMPA and NMDA ecep o subuni s, indica ing ha ex a- synap ic iono opic ecep o s may play a negligible ole in glu a- ma e gic coupling among neu ons in he supe icial spinal do sal ho n. Colocaliza ion o AMPA and NMDA ecep o s We ound ha all synap ic con ac s ha exp essed NMDA ecep- o s also con ained AMPA ecep o s in laminae I–II o he a spinal g ay ma e . In addi ion, none o he pos synap ic mem- b anes eco e ed in his s udy ca ied less han eigh gold pa i- cles ma king AMPA ecep o s. These indings con i m ea lie physiological s udies, concluding ha he e is no di ec e idence o he p esence o “silen ” synapses (NMDA ecep o - con aining synapses wi hou unc ional AMPA ecep o s) in he spinal do sal ho n o adul animals, al hough hei p esence has been con i med in he neona al a (Ba doni e al., 1998; Li and Zhuo, 1998; Baba e al., 2000). Thus, i appea s ha pu e-NMDA ecep o -media ed EPSCs a e ansien , de elopmen ally egu- la ed phenomena, and al hough hey may ha e a ole in synap ic e inemen in he imma u e do sal ho n, hey a e unlikely o be in ol ed in nocicep i e in o ma ion p ocessing mechanisms in he adul . I has been sugges ed ha AMPA and NMDA ecep o s migh be a anged in unc ional mic odomains wi hin synap ic mem- b anes. The no ion is based on esul s ob ained in independen labo a o ies, sugges ing ha he dis ibu ion o AMPA ecep o s may no be uni o m o e he pos synap ic densi y bu mo e con- cen a ed la e ally, whe eas NMDA ecep o s a e loca ed cen- ally (Ma suba a e al., 1996; Kha azia and Weinbe g, 1997; So- mogyi e al., 1998b; He e al., 2001). O he heo ies, howe e , claim jus he opposi e and a gue in a o o a uni o m dis ibu- ion o iono opic glu ama e ecep o s o e synap ic memb ane specializa ions (Nusse e al., 1994, 1998, Masugi-Toki a e al., 2007). Al hough we canno exclude he possibili y ha subsyn- ap ic o ganiza ion o ecep o s migh ha e changed du ing spec- imen p epa a ion, ou p esen esul s suppo he “homogeneous dis ibu ion” heo y o spinal co d synapses. The high sensi i i y and wo-dimensional esolu ion o SDS-FRL ha e e ealed a a he homogeneous dis ibu ion o bo h AMPA and NMDA ecep o s wi hin indi idual pos synap ic memb ane specializa- ions. This ype o ecep o dis ibu ion may ha e p o ound bio- logical signi icance; i may make he coope a ion be ween AMPA and NMDA ecep o mechanisms di ec and e y e ec i e. Subuni composi ion o AMPA ecep o s AMPA ecep o s wi h di e en subuni composi ion ha e cha - ac e is ic pha macological and physiological p ope ies ( o e- iew, see Ha is, 1995). Fo ins ance, he p esence o GluR2 sub- uni s ende s he e ome ic AMPA ecep o assemblies Ca 2⫹ impe meable, whe eas ecep o s lacking GluR2 subuni s a e Ca 2⫹ pe meable (Gasic and Heinemann, 1991; Hollmann e al., 1991; Ve doo n e al., 1991; Bu nashe e al., 1992; Somme and Seebu g, 1992; Pelleg ini-Giampie o e al., 1997; Swanson e al., 1997). A g ea deal o ea lie expe imen al e idence indica ed ha di e en popula ions o neu ons in he supe icial spinal do sal ho n show di e en ial AMPA ecep o exp ession. Kaina e- induced cobal -up ake and immunocy ochemical s udies s ongly indica ed ha a majo i y o neu ons immuno eac i e o GABA a e posi i ely s ained o GluR1 bu no GluR2 and do exp ess Ca 2⫹ -pe meable AMPA ecep o s, whe eas se e al sub- popula ions o pu a i e exci a o y in e neu ons do no exp ess Ca 2⫹ -pe meable AMPA ecep o s (Ke e al., 1998; Albuque - que e al., 1999; Engelman e al., 1999). The con as ing dis ibu- ion o GluR1 and GluR2/3 immuno eac i i y aised he possibil- i y ha some neu ons in he supe icial do sal ho n may exp ess only one o he wo ecep o subuni s. O he s udies, howe e , sugges ed ha indi idual do sal ho n neu ons can exp ess bo h Ca 2⫹ -pe meable and Ca 2⫹ -impe meable AMPA ecep o s (Golds ein e al., 1995; Gu e al., 1996; Albuque que e al., 1999). This no ion was s ongly ein o ced by a ecen immuno luo es- cence s udy showing ha i ually all GluR1-immuno eac i e punc a ha we e assumed o ep esen pos synap ic memb anes we e also GluR2 immuno eac i e h oughou he do sal ho n (Nagy e al., 2004). Al hough immunocy ochemical colocaliza ion s udies p o- ide only ci cums an ial e idence ega ding he subuni compo- si ion o indi idual AMPA ecep o complexes, ou p esen ind- ings a e in ag eemen wi h he epo o Nagy e al. (2004). We also ound ha he majo i y o comple e pos synap ic memb ane specializa ions (99%) displayed posi i e immunolabeling o GluR2, and mos o hem (87%) we e also immuno eac i e o GluR1. On he o he hand, howe e , al hough he o e lap be- ween GluR1 and GluR2 immuno eac i i y was high, he num- be s o gold pa icles labeling GluR1 and GluR2 molecules a ied a indi idual pos synap ic ac i e zones. I is he e o e likely ha he a ge ing o AMPA ecep o subuni s o indi idual pos syn- ap ic memb anes is ema kably complex. Di e en AMPA ecep- o ion channels wi hin he same pos synap ic memb ane may p esen di e en subuni composi ion. Ca 2⫹ -pe meable AMPA ecep o s can be in e mingled wi h GluR2-con aining ecep o s wi hin he same pos synap ic ac i e zones, and he a io be ween Ca 2⫹ -pe meable and Ca 2⫹ -impe meable AMPA ecep o s may a y in a wide ange. Mo eo e , we encoun e ed his di e en ial exp ession o GluR1 and GluR2 subuni s e en a pos synap ic memb ane specializa ions ha we e loca ed in a close icini y o each o he on he same neu on, indica ing ha he a ge ing o AMPA ecep o subuni s o pos synap ic memb anes migh be egula ed by p esynap ic mechanisms. Figu e 7. His og am showing he dis ibu ion o su ace a eas o comple e pos synap ic memb ane specializa ions exp essing AMPA- and/o NMDA- ype glu ama e ecep o s in he supe icial spinal do sal ho n o a s. An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 • 9699 Re e ences Albuque que C, Lee CJ, Jackson AC, MacDe mo AB (1999) Subpopula- ions o GABAe gic and non-GABAe gic a do sal ho n neu ons exp ess Ca 2⫹ -pe meable AMPA ecep o s. Eu J Neu osci 11:2758–2766. An al M, K a sik R, Sze´kely G, an de Loos H (1992) Synapses on mo- oneu on dend i es in he b achial sec ion o he og spinal co d: a compu e -aided elec on mic oscopic s udy o cobal - illed cells. J Neu- ocy ol 21:34–49. Baba H, Doubell TP, Moo e KA, Wool CJ (2000) Silen NMDA ecep o - media ed synapses a e de elopmen ally egula ed in he do sal ho n o he a spinal co d. J Neu ophysiol 83:955–962. Ba doni R, Maghe ini PC, MacDe mo AB (1998) NMDA EPSCs a glu a- ma e gic synapses in he spinal co d do sal ho n o he pos na al a . J Neu osci 18:6558–6567. Baude A, Nusse Z, Molna´ E, McIlhinney RA, Somogyi P (1995) High- esolu ion immunogold localiza ion o AMPA ype glu ama e ecep o subuni s a synap ic and non-synap ic si es in a hippocampus. Neu o- science 69:1031–1055. B own KM, W a hall JR, Yasuda RP, Wol e BB (2002) Quan i a i e mea- su emen o glu ama e ecep o subuni p o ein exp ession in he pos - na al a spinal co d. B ain Res De B ain Res 137:127–133. Bu nashe N, Monye H, Seebu g PH, Sakmann B (1992) Di alen ion pe - meabili y o AMPA ecep o channels is domina ed by he edi ed o m o a single subuni . Neu on 8:189–198. Cull-Candy S, B ickley S, Fa an M (2001) NMDA ecep o s subuni s: di- e si y, de elopmen and disease. Cu Opin Neu obiol 11:327–335. Engelman HS, Allen TB, MacDe mo AB (1999) The dis ibu ion o neu- ons exp essing calcium-pe meable AMPA ecep o s in he supe icial laminae o he spinal co d do sal ho n. J Neu osci 19:2081–2089. F i schy JM, Weinmann O, Wenzel A, Benke D (1998) Synapse-speci ic lo- caliza ion o NMDA and GABA(A) ecep o subuni s e ealed by an igen- e ie al immunohis ochemis y. J Comp Neu ol 390:194–210. Fuji a A, Cheng J, Hi akawa M, Fu ukawa K, Kusunoki S, Fujimo o T (2007) Gangliosides GM1 and GM3 in he li ing cell memb ane o m clus e s suscep ible o choles e ol deple ion and chilling. Mol Biol Cell 18:2112–2122. Fu uyama T, Kiyama H, Sa o K, Pa k HT, Maeno H, Takagi H, Tohyama M (1993) Region-speci ic exp ession o subuni s o iono opic glu ama e ecep o s (AMPA- ype, KA- ype and NMDA ecep o s) in he a spinal co d wi h special e e ence o nocicep ion. Mol B ain Res 18:141–151. Gasic GP, Heinemann S (1991) Recep o s coupled o ionic channels: he glu ama e ecep o amily. Cu Opin Neu obiol 1:20–26. Gibb AJ (2004) NMDA ecep o subuni ga ing-unco e ed. T ends Neu o- sci 27:7–10. Golds ein PA, Lee CJ, MacDe mo AB (1995) Va iable dis ibu ions o Ca 2⫹ -pe meable and Ca 2⫹ -impe meable AMPA ecep o s on emb yonic a do sal ho n neu ons. J Neu ophysiol 73:2522–2534. Gu JG, Albuque que C, Lee CJ, MacDe mo AB (1996) Synap ic s eng h- ening h ough ac i a ion o Ca 2⫹ -pe meable AMPA ecep o s. Na u e 381:793–796. Ha is EW (1995) Sub ypes o glu ama e ecep o s: pha macological classi- ica ion. In: CNS neu o ansmi e s and neu omodula o s (S one TW, ed), pp 95–125. Boca Ra on, FL: CRC. Ha is JA, Co si M, Qua a oli M, A ban R, Ben i oglio M (1996) Up egu- la ion o spinal glu ama e ecep o s in ch onic pain. Neu oscience 74:7–12. Ha is KM, Landis DM (1986) Memb ane s uc u e a synap ic junc ions in a ea CA1 o he a hippocampus. Neu oscience 19:857–872. Ha mann B, Ahmadi S, Heppens all PA, Lewin GR, Scho C, Bo cha d T, Seebu g PH, Zeilho e HU, Sp engel R, Kune R (2004) The AMPA e- cep o subuni s GluR-A and GluR-B ecip ocally modula e spinal synap- ic plas ici y and in lamma o y pain. Neu on 44:637–650. He Y, Ho PR, Janssen WG, Ro hs ein JD, Mo ison JH (2001) Di e en ial synap ic localiza ion o GluR2 and EAAC1 in he macaque monkey en o- hinal co ex: a pos embedding immunogold s udy. Neu osci Le 311:161–164. Henley JM, Jenkins R, Hun SP (1993) Localiza ion o glu ama e ecep o binding si es and mRNAs o he do sal ho n o he a spinal co d. Neu- opha macology 32:37–41. Hollmann M, Ha ley M, Heinemann S (1991) Ca2⫹pe meabili y o KA- AMPA-ga ed glu ama e ecep o channels depends on subuni composi- ion. Science 252:851–853. Jakowec MW, Fox AJ, Ma in LJ, Kalb RG (1995a) Quan i a i e and quali- a i e changes in AMPA ecep o exp ession du ing spinal co d de elop- men . Neu oscience 67:893–907. Jakowec MW, Yen L, Kalb RG (1995b) In si u hyb idiza ion analysis o AMPA ecep o subuni gene exp ession in he de eloping a spinal co d. Neu oscience 67:909–920. Ji RR, Kohno T, Moo e KA, Wool CJ (2003) Cen al sensi iza ion and LTP: do pain and memo y sha e simila mechanisms? T ends Neu osci 26:696–705. Ke R, Maxwell DJ, Todd AJ (1998) GluR1 and GluR2/3 subuni s o he AMPA- ype glu ama e ecep o a e associa ed wi h pa icula ypes o neu one in laminae I-III o he spinal do sal ho n o he a . Eu J Neu- osci 10:324–333. Kha azia VN, Weinbe g RJ (1997) Tangen ial synap ic dis ibu ion o NMDA and AMPA ecep o s in a neoco ex. Neu osci Le 238:41–44. Landis DM, Reese TS (1974) Di e ences in memb ane s uc u e be ween exci a o y and inhibi o y synapses in he ce ebella co ex. J Comp Neu ol 155:93–125. Li P, Zhuo M (1998) Silen glu ama e gic synapses and nocicep ion in mammalian spinal co d. Na u e 393:695–698. Lim R, Al a ez FJ, Walmsley B (1999) Quan al size is co ela ed wi h ecep- o clus e a ea a glycine gic synapses in he a b ains em. J Physiol 516:505–512. Mackenzie PJ, Kenne GS, P ange O, Shayan H, Umemiya M, Mu phy TH (1999) Ul as uc u al co ela es o quan al synap ic unc ions a single CNS synapses. J Neu osci 19:RC13(1–7). Masugi-Toki a M, Shigemo o R (2007) High- esolu ion quan i a i e isu- aliza ion o glu ama e and GABA ecep o s a cen al synapses. Cu Opin Neu obiol 17:387–393. Masugi-Toki a M, Ta usawa E, Wa anabe M, Molna´ E, Fujimo o K, Shige- mo o R (2007) Numbe and densi y o AMPA ecep o s in indi idual synapses in he a ce ebellum as e ealed by SDS-diges ed eeze- ac u e eplica labeling. J Neu osci 27:2135–2144. Ma suba a A, Laake JH, Da ange S, Usami S, O e sen OP (1996) O gani- za ion o AMPA ecep o subuni s a a glu ama e synapse: a quan i a i e immunogold analysis o hai cell synapses in he a o gan o Co i. J Neu- osci 16:4457–4467. McClung JR, Cas o AJ (1978) Rexed’s lamina scheme as i applies o he a ce ical spinal co d. Exp Neu ol 58:145–148. McNeill DL, Chung K, Hulsebosch CE, Bolende RP, Coggeshall RE (1988) Numbe s o synapses in laminae I-IV o he a do sal ho n. J Comp Neu ol 278:453–460. Molande C, Xu Q, G an G (1984) The cy oa chi ec onic o ganiza ion o he spinal co d in he a . I. The lowe ho acic and lumbosac al co d. J Comp Neu ol 230:133–141. Nagy GG, Al-Ayyan M, And ew D, Fukaya M, Wa anabe M, Todd AJ (2004) Widesp ead exp ession o he AMPA ecep o GluR2 subuni a glu ama- e gic synapses in he a spinal co d and phospho yla ion o GluR1 in esponse o noxious s imula ion e ealed wi h an an igen-unmasking me hod. J Neu osci 24:5766–5777. Nakanishi S (1992) Molecula di e si y o glu ama e ecep o s and implica- ions o b ain unc ion. Science 258:597–603. Nusse Z (1999) A new app oach o es ima e he numbe , densi y and a i- abili y o ecep o s a cen al synapses. Eu J Neu osci 11:745–752. Nusse Z (2000) AMPA and NMDA ecep o s: simila i ies and di e ences in hei synap ic dis ibu ion. Cu Opin Neu obiol 10:337–341. Nusse Z, Mul ihill E, S ei P, Somogyi P (1994) Subsynap ic seg ega ion o me abo opic and iono opic glu ama e ecep o s as e ealed by immu- nogold localiza ion. Neu oscience 61:421–427. Nusse Z, Cull-Candy S, Fa an M (1997) Di e ences in synap ic GABA A ecep o numbe unde lie a ia ion in GABA mini ampli ude. Neu on 19:697–709. Nusse Z, Lujan R, Laube G, Robe s JDB, Molna E, Somogyi P (1998) Cell ype and pa hway dependence o synap ic AMPA ecep o numbe and a iabili y in he hippocampus. Neu on 21:545–559. O e sen OP, Landsend AS (1997) O ganiza ion o glu ama e ecep o s a he synapse. Eu J Neu osci 9:2219–2224. Pelleg ini-Giampie o DE, Fan S, Aul B, Mille BE, Zukin RS (1994) Glu- ama e ecep o gene exp ession in spinal co d o a h i ic a s. J Neu osci 14:1576–1583. Pelleg ini-Giampie o DE, Go e JA, Benne MV, Zukin RS (1997) The 9700 •J. Neu osci., Sep embe 24, 2008 •28(39):9692–9701 An al e al. •Synap ic Localiza ion o AMPA and NMDA Recep o s