LO

AR

De9

AcHsnlttsrcg(noaaepnwislsordrwpadmP

Ks

Sc(es

*EAPw

Neuroscience 151 (2008) 995–1005

0d

ACK OF PHOTORECEPTOR SIGNALING ALTERS THE EXPRESSION

F SPECIFIC SYNAPTIC PROTEINS IN THE RETINA

Avaet1nbt2(cphta2

rheateh

sslpwtaiinPptm1civ

smaZ

. H. KIHARA,* T. O. SANTOS, V. PASCHON,

. J. B. MATOS AND L. R. G. BRITTO

epartment of Physiology and Biophysics, Institute of Biomedical Sci-nces, University of São Paulo, Av. Prof. Lineu Prestes 1524, 05508-00, São Paulo, SP, Brazil

bstract—Synaptic modulation by activity-dependent changesonstitutes a cellular mechanism for neuronal plasticity.owever, it is not clear how the complete lack of neuronalignaling specifically affects elements involved in the commu-ication between neurons. In the retina, it is now well estab-

ished that both chemical and electrical synapses are essentialo mediate the transmission of visual signaling triggered byhe photoreceptors. In this study, we compared the expres-ion of synaptic proteins in the retinas of wild-type (WT) vs.d/rd mice, an animal model that displays inherited and spe-ific ablation of photoreceptors caused by a mutation in theene encoding the �-subunit of rod cGMP-phosphodiesterasePde6brd1). We specifically examined the expression of con-exins (Cx), the proteins that form the gap junction channelsf electrical synapses, in addition to synaptophysin and syn-psin I, which are involved in the release of neurotransmitterst chemical synapses. Our results revealed that Cx36 genexpression levels are lower in the retinas of rd/rd when com-ared with WT. Confocal analysis indicated that Cx36 immu-olabeling almost disappeared in the outer plexiform layerithout significant changes in protein distribution within the

nner plexiform layer of rd/rd retinas. Likewise, synaptophy-in expression remarkably decreased in the outer plexiform

ayer of rd/rd retinas, and this down-regulation was also as-ociated with diminished transcript levels. Furthermore, webserved down-regulation of Cx57 gene expression in rd/rdetinas when compared with WT and also changes in proteinistribution. Interestingly, Cx45 and synapsin I expression ind/rd retinas showed no noticeable changes when comparedith WT. Taken together, our results revealed that the loss ofhotoreceptors leads to decreased expression of some syn-ptic proteins. More importantly, this study provides evi-ence that neuronal activity regulates, but is not essential toaintain, the expression of synaptic elements. © 2008 IBRO.ublished by Elsevier Ltd. All rights reserved.

ey words: plasticity, synapse, connexin, synaptophysin,ynapsin, retina.

ynaptic plasticity is usually defined as the ability of theonnection between neurons to modulate its strengthSheng and Kim, 2002; Blitz et al., 2004). There are sev-ral mechanisms participating in the modulation of a givenynapse, including its own activity (Faber et al., 1991;

Corresponding author. Tel: �55-11-3091-7242; fax: �55-11-3091-7426.-mail address: kihara@icb.usp.br (A. H. Kihara).bbreviations: CT, cycle threshold; Cx, connexin; GJ, gap junction;

tB, phosphate buffer; rd/rd, (retinal degeneration) mouse model; WT,ild-type.

306-4522/08$32.00�0.00 © 2008 IBRO. Published by Elsevier Ltd. All rights reseroi:10.1016/j.neuroscience.2007.09.088

995

barbanel et al., 2002). Classically, electrical coupling pro-ided by gap junctions (GJ) was considered “not plastic,”lthough being important for tuning network activity (Traubt al., 2001). GJ channels couple adjacent cells, allowingransfer of second messengers, ions and molecules up to kDa (for review, see Rozental et al., 2000). These chan-els are composed by a multigene family of integral mem-rane proteins called connexins (Cxs), with 20 members inhe murine genome (Willecke et al., 2002; Sohl et al.,004). Hexameric association of Cxs forms a hemichannelconnexon), and the docking of two connexons in apposedell membranes forms a functional GJ channel. It has beenroposed that in addition to form channels in the GJ,emichannels also function in nonjunctional membranes inhe absence of pairing with partners from adjacent cells,lthough this matter remains under debate (Jiang and Gu,005; Spray et al., 2006).

In the retina, cell coupling provided by GJ channelsesults in extensive networks (Vaney, 2002). Several Cxsave been identified in the adult rodent retina (Condorellit al., 1998; Guldenagel et al., 2000). For example, Cx36nd Cx45 were identified as neuronal Cxs participating inhe rod-mediated circuitry (Deans et al., 2002; Maxeinert al., 2005), whereas Cx57 is apparently restricted toorizontal cells (Hombach et al., 2004).

The retina has been considered as a “natural brainlice” and largely used in studies focusing on chemicalynapses. The mammalian retina contains two synaptic

ayers: the outer plexiform layer, which is primarily com-osed of ribbon synapses (Sterling and Matthews, 2005),hile the inner plexiform layer largely comprises conven-

ional synapses (Von Kriegstein et al., 1999). A consider-ble number of molecules have been described to partic-

pate in the chemical communication between neurons,ncluding those that are involved in the intricate process ofeurotransmitter release (Zimmermann, 1997; Parnas andarnas, 2007). Synaptophysin is an integral membranerotein of synaptic vesicles and has an important role inhe neurotransmitter release from the synaptic vesicles byaking an exocytotic fusion pore (Elferink and Scheller,995). Synapsin I is a phosphoprotein associated with theytoplasmic surface of the synaptic vesicle membrane and

s thought to function by increasing the number of synapticesicles in the releasable pool (Hilfiker et al., 1999).

Based on the premise that neuronal activity may shapeynaptic communication, we took advantage of the rd/rdouse model, which in the adulthood displays naturalblation of photoreceptors (Strettoi and Pignatelli, 2000;hu et al., 2007) caused by a mutation in the gene encoding

he �-subunit of rod cGMP-phosphodiesterase (Pde6brd1).ved.

Itwcttdmvec

A

ECmlwaMUcia

R

Rgtrawewfspcb(Iic(CaAa9we

P

Qtwtscccv

fipetucfopes

I

E1cA(aaeb

aT21pd(rb3ottpbtftwaCnIcpmbrepU

at1tUroc(

C2

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005996

n adult rd/rd mice, nuclear staining of vertical retinal sec-ions indicated no major abnormalities in inner layers,here only the reduction or the absence of the photore-eptor layer is evident. However, second-order neurons ofhe rd/rd may show morphological abnormalities, such ashe atrophy of dendrites of cone bipolar cells, mostly evi-ent at P90 (Strettoi et al., 2003). By studying this animalodel, we were able to determine whether the lack of

isual signaling triggered by the photoreceptors alters thexpression of proteins directly involved in electrical andhemical synapses.

EXPERIMENTAL PROCEDURES

nimal procedures

xperiments were carried out with C57BL/6J (wild-type (WT)) and3H/HeJ (rd/rd, The Jackson Laboratory, Bar Harbor, ME, USA)ice kept on a 12-h light/dark cycle (light phase 80–100 lux) with

ights on at 6:00 a.m. Young adult animals (3–4 weeks old, n�16)ere killed for different methodologies with an overdose of ket-mine (30 mg/100 g of body weight, i.m., Parke-Davis, Ann Arbor,I, USA) and xylazine (2 mg/100 g, i.m., Bayer, West Haven, CT,SA) between 10:00 a.m. and 12:00 a.m. All experiments wereonducted in accordance to the NIH and the Institute of Biomed-cal Sciences/USP guidelines. Thus, we minimized the number ofnimals used and their suffering.

NA isolation, cDNA synthesis and real time PCR

etinas were directly homogenized in 1 ml TRIzol reagent (Invitro-en, Carlsbad, CA, USA) and total RNA was extracted followinghe manufacturer-suggested protocol. In brief, following two chlo-oform extraction steps, RNA was precipitated with isopropanolnd the pellet washed twice in 70% ethanol. After air-drying, RNAas resuspended in DEPC-treated water and the concentration ofach sample obtained from A260 measurements. Residual DNAas removed using DNase I (Amersham, Piscataway, NJ, USA)

ollowing the manufacturer protocol. For each 20 �l reverse tran-cription reaction, 4 �g total RNA was mixed with 1 �l oligodTrimer (0.5 �g; Invitrogen) and incubated for 10 min at 65 °C. Afterooling on ice the solution was mixed with 4 �l 5� first stranduffer, 2 �l of 0.1 M DTT, 1 �l of dATP, dTTP, dCTP and dGTPeach 10 mM), and 1 �l SuperScript III reverse transcriptase (200 U;nvitrogen) and incubated for 60 min at 50 °C. Reaction wasnactivated by heating at 70 °C for 15 min. Real-Time PCR wasarried out using a Rotor-Gene 6000 Real-Time Rotary AnalyzerCorbett Life Science, Sidney, Australia) with specific primers forx36, Cx45, Cx57, synaptophysin and synapsin I (Table 1). cDNAbundance for GAPDH was also determined as internal control.ll PCR assays were performed as follows: after initial activationt 50 °C for 2 min and 95 °C for 10 min, cycling conditions were5 °C, 10 s and 60 °C, 1 min. Dissociation curves of PCR productsere obtained by heating samples from 60 °C to 95 °C, in order tovaluate primer specificity.

CR data and statistical analysis

uantification of gene amplification was performed by determininghe cycle threshold (CT) based on the fluorescence detectedithin the geometric region of the semi-log view of the amplifica-

ion plot. An amplification plot for each sample was generatedhowing the increase in reporter dye fluorescence (�Rn) with eachycle of PCR. From each amplification plot a CT value was cal-ulated, representing the PCR cycle number at which the fluores-ence was detectable above an arbitrary threshold, based on the

ariability of baseline data in the first 15 cycles. Relative quanti- fi

cation of target gene expression was evaluated using the com-arative CT method as previously described in detail (Medhurstt al., 2000). The �CT value was determined by subtracting thearget CT of each sample from its respective GAPDH CT value,sed as internal control. Calculation of ��CT involves using theontrol group mean �CT value as an arbitrary constant to subtractrom all other �CT mean values. Fold-changes in gene expressionf the target gene are equivalent to (1�E)���CT, where the Earameter corresponds to efficiency of amplification. Values werentered into a Student’s t-test and P values of �0.05 were con-idered to be significant.

mmunohistochemistry

yes were dissected out and the retinas were fixed for 30 min in% PFA in phosphate buffer 0.1 M pH 7.3 (PB) for 30 min andryoprotected with a 30% sucrose solution for at least 24 h at 4 °C.fter embedding in O.C.T. compound, they were cut transversally

12 �m) on a cryostat. Retinal sections were blocked for 30 min insolution containing 10% normal goat serum, 1% bovine serum

lbumin (BSA) and 0.5% Triton X-100 in PB. For whole-mountxperiments, retinas were fixed in 4% PFA in PB for 24 h at 4 °Cefore incubation with primary antibodies.

Retinal sections or whole-mounts were incubated overnightnd for 7 days, respectively, with the primary antibodies listed inable 1. A rabbit polyclonal antibody raised against amino acids96–304 of human Cx36 was used (36–4600, Zymed/Invitrogen,:250–1:500), which identifies in the retina a single band at ap-roximately 36 kDa (Mills et al., 2001). To identify Cx45, twoifferent antibodies were used: a mouse monoclonal antibodyMAB3101, Chemicon, Temecula, CA, USA, 1:100–1:250) and aabbit polyclonal antibody (AB1745, Chemicon, 1:250–1:500),oth raised against a peptide corresponding to amino acids 354–67 of human Cx45 and that does not possess sequence homol-gy with Cx43 (Coppen et al., 1998). Cx57 expression was inves-igated using a rabbit polyclonal antibody raised against a syn-hetic peptide derived from an internal region of the mouse Cx57rotein (40–5000, Zymed/Invitrogen, 1:100–1:250). In Westernlots, this antibody identifies the target bands at 57 and 64 kDa;he 64 kDa band likely represents a post-transcriptionally modifiedorm of Cx57, consistent with the multiple phosphorylation sites ofhis protein (manufacturer information). Expression of synapsin Ias investigated using a rabbit polyclonal antibody raised againstmixture of Ia and Ib isoforms purified from bovine brain (AB1543,hemicon, 1:250). According to manufacturer information, immu-olabeling is blocked by preadsorption of antibody with synapsin. Synaptophysin expression was determined using a rabbit poly-lonal antibody raised against human synaptophysin peptide cou-led to ovalbumin (A0010, DakoCytomation, Copenhagen, Den-ark, 1:250). Calbindin-D is a small acidic protein (28 kDa) that

elongs to a family of calcium-binding proteins, and in the mouseetina is a reliable marker for horizontal cells. In some double-labelingxperiments, we used a mouse monoclonal antibody raised againsturified calbindin-D from chicken gut (C8666, Sigma, St. Louis, MO,SA, 1:250) in order to identify retinal horizontal cells.

After several washes, retinal sections were incubated for 2 hnd whole-mounts for 24 h with goat antiserum against rabbitagged to AlexaTM 488 (Molecular Probes, Eugene, Oregon,:500–1:1,000) and, in some cases, tetramethyl rhodamine iso-hiocyanate (TRITC, Jackson Laboratories, West Grove, PA,SA, 1:100–1:500) diluted in PB containing 0.5% Triton X-100 at

oom temperature. Controls for the experiments consisted of themission of primary antibodies; no staining was observed in theseases. After washing, the tissue was mounted using Vecta ShieldVector Laboratories, Burlingame, CA, USA).

PFA was reported to cause non-specific staining when somex antibodies were used (Feigenspan et al., 2001; Meier et al.,002). Thus, in some experiments cold ethanol was used as

xative instead of PFA. Immunoreactivity pattern was very similar

wiw(twtcb

I

DPrIUfi1atacdtipUdfipstic(

Gt

Lfvtwpawc

pnrPfyrII

Cv

gitW

T

G

C

C

C

S

S

G

C

sa

b

c

d

e

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005 997

ith both fixatives for Cx36, Cx45 and Cx57 antibodies, althoughn some cases, the concentration of the antibody was decreasedhen PFA was used as fixative to achieve same results as ethanol

data not shown). For this reason, and because immunostaining ofhe tissue for synaptophysin, synapsin I and, notably, calbindin-Dere much better with PFA, we decided to use this fixative with

hese antibodies. In order to better localize immunolabeling, weounterstained retinas using propidium iodide (1:1000), by incu-ating sections at room temperature for 15 min.

mage quantification

igital images (1024�1024 pixels) were acquired by using NikonCM-2000 microscope confocal scanning system (Nikon Corpo-

ation, Tokyo, Japan). Image analyses were performed withmage-Pro Plus software (Media Cybernetics, Bethesda, MD,SA). After channel separation (RGB) of color images, we per-

ormed three different types of computational analyses: 1) count-ng puncta, 2) bitmap analysis, and 3) co-localization index.) After proper setting of size and brightness, the software performsutomatic search of discrete elements, as punctate labeling. Au-omatically counted labeling was displayed in the image, thusrtifacts and background labeling could be identified and dis-arded. 2) x–y Axis bitmap analyses generated numerical appendedata file corresponding to pixel values. The bitmap analysis was usedo view the pixel values of the active window (or area of interest, AOI)n numeric format, where values correspond to the bright of theixels. This matrix was exported to Excel (Microsoft, Redmond, WA,SA) for the appropriate mathematical computation. The numericalata generated a histogram, essentially averaging the intensity pro-le in the different layers of the retina. 3) In some cases, we alsoerformed computational co-localization analysis. Since our confocalcanning system generated color images, the software determinedhe co-localization coefficient from different RGB channels, whichndicates the relative degree of overlap between signals. Photomi-rographs and charts were prepared using Adobe Photoshop CS2

able 1. Description of PCR primers and antibodies used in this stud

ene/protein (GenBank#) Primers

x36 (AF016190) Up: 5=-CCCCCAGTCTCTGTTTTATCDown: 5=-GACAGTAGAGTACCGGCG

x45 (AF283254) Up: 5=-TCACCCTGCCTTGTCACCTADown: 5=-TGTTGCTTTCCCAAATTCT

x57 (NM010289) Up: 5=-AGAGCCCAGATGGAGAATCC

Down: 5=-TCCTCGAGTCTCCTCAGC

ynaptophysin (BC014823) Up: 5=-AACAACAAAGGGCCAATGAT

Down: 5=-CCCAGGCGGATGAGCTAA

ynapsind (BC022954) Up: 5=-TGATCTCAATCTTGTGGCTC

Down: 5=-ACCTCACAACTTTGACTCC

APDH (M32599) Up: 5=-CTCAACTACATGGTCTACATDown: 5=-CCCATTCTCGGCCTTGAC

albindin-D (NM009788)

Sequences were designed using Primer Express software (version 3equences within mouse genome (http://www.ncbi.nlm.nih.gov/BLASTKihara et al., 2006b.Kihara et al., 2006a.Von Kriegstein et al., 1999; Mandell et al., 1990.Hombach et al., 2004.Haverkamp and Wassle, 2000.

Adobe Systems Inc., San Jose, CA, USA). a

RESULTS

ene expression levels of synaptic proteins inhe retinas of WT vs. rd/rd mice

inear regression analysis of amplification plots generatedrom cDNA serial dilutions ranging from 4 to 256 ng re-ealed a high correlation (R2�0.97), indicating amplifica-ion linearity for all sets of primers (Fig. 1A–E). In addition,e were able to calculate amplification efficiency (E) of therimers, which was used to determine and compare thebundance of gene transcripts. GAPDH gene expressionas used as an internal control, and displayed no signifi-ant differences between WT and rd/rd retinas (Fig. 1F).

Our results indicated that Cx36, Cx57, and synapto-hysin mRNA levels are statistically different in rd/rd reti-as when compared with WT (Fig. 1G). In rd/rd retinas,educed Cx36 transcript levels were found (�49%,�0.01) when compared with WT. Similar results were

ound for Cx57 mRNA expression (�51%, P�0.01). Anal-sis of synaptophysin gene expression also revealed aemarkable reduction (�69%, P�0.01) in rd/rd retinas.nterestingly, we found similar levels of Cx45 and synapsintranscripts among WT and rd/rd retinas.

omparative expression of Cxs in the WTs. rd/rd retinas

Cx36. In accordance to previous studies (Guldena-el et al., 2000; Kihara et al., 2006b), we observed Cx36

mmunoreactivity mainly as a bright punctate labeling inhe inner half of the inner plexiform layer (ON sublamina) in

T retinas. Although not so evident, immunolabeling was

Antibody Antibody specificity

-3= Zymed 36–4600 Staining eliminated in Cx36 KOa

3=b

Chemicon AB1745 Staining eliminated by peptidepreadsorption (5�)a

3=b Chemicon MAB3101Zymed 40–5000 Omission of primary antibody,

labeling in accordance togene reporter studiesd

b

DakoCytomation A0010 Omission of primary antibody,labeling in accordance toprevious studiesc

Chemicon AB1543 Omission of primary antibody,labeling in accordance toprevious studiesc-3=

-3=

Sigma C8866 Omission of primary antibody,labeling in accordance toprevious studiese

ed Biosystems). All primers were submitted to search for homologouso significant similarity was found.

y

ACTTATTTTCTC-

CT-3=CTATG-A-3=

TCTT-3=

G-3=

C-3=

ATG-3=

ATTCC

GTTCCAT-3=b

.0, Appli/), and n

lso observed in the outer half of the inner plexiform layer

((dDaaisaplc

gtoOlCdwcsb

rtz

tmaluiomrrserci

Ca

(oii(lcx

Flp ine andi xpression

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005998

OFF sublamina) and in the outer plexiform layer as wellFig. 2A–C). In the rd/rd retinas, we found a marked re-uction of the labeling in the outer plexiform layer (Fig. 2,–F). Accordingly, X–Y bitmap analysis revealed remark-ble differences in the Cx36 distribution pattern (Fig. 2G),s evidenced by the altered labeling proportion between

nner and outer plexiform layers (Fig. 2H). However, de-pite the overall alteration of Cx36 expression, we were notble to find changes in Cx36 distribution within the innerlexiform layer of rd/rd and WT retinas, given that the

abeling ratio of OFF and ON sublaminas did not signifi-antly differ (Fig. 2I).

Cx45. As we observed in previous studies (Guldena-el et al., 2000; Kihara et al., 2006b), Cx45 immunoreac-ivity was seen in both the outer and inner plexiform layersf the adult mouse retina as fine and bright puncta. BothN and OFF sublaminas of the inner plexiform layer were

abeled (Fig. 3A–C). Two different antibodies raised againstx45 provided similar results, endowing satisfactory evi-ence regarding Cx45 distribution. This labeling patternas also observed in rd/rd retinas, with no noticeablehanges (Fig. 3D–F). Accordingly, x–y axis bitmap analy-is revealed no significant differences in the Cx45 distri-ution pattern (data not shown).

Cx57. In the present study, using a purified antibodyaised against Cx57, we were able to provide immunohis-ochemical evidence that Cx57 is indeed present in hori-

ig. 1. Gene expression levels of synaptic proteins in the mouse reinearity for primers specifically designed for Cx36 (A), Cx45 (B), Cx5rimers was calculated based on linear regression and used to determ

nternal control (F). Means from rd/rd were normalized based on WT e

ontal cells. In addition, we observed spatial changes of l

his protein in rd/rd retinas. In vertical sections, Cx57 im-unolabeling was observed as a punctate pattern formingribbon, which typically filled the entire outer plexiform

ayer (Fig. 4A–C). In turn, in the rd/rd retinas, Cx57 labelingsually appeared as a disrupted, twisted streak, surround-

ng presumptive horizontal cell perikarya located in theuter border of the inner nuclear layer (Fig. 4D–F). Whole-ount experiments performed in WT (Fig. 5A) and rd/rd

etinas (Fig. 5B) showed that Cx57 immunoreactivity iseduced in the latter. Indeed, quantification of puncta den-ity in WT (red) and rd/rd (green) revealed that Cx57xpression is reduced in the outer plexiform layer of rd/rdetinas (Fig. 5C). Additionally, we observed that the Cx57/albindin-D co-localization ratio did not significantly differ

n WT and rd/rd retinas (Fig. 5D).

omparative expression of synaptophysinnd synapsin I in the WT vs. rd/rd retinas

Synaptophysin. In accordance to previous studiesMandell et al., 1990; Von Kriegstein et al., 1999), webserved synaptophysin immunolabeling as a bright streak

n the outer plexiform layer, and a moderate, diffuse label-ng filling up the entire inner plexiform layer of WT retinasFig. 6A–C). By contrast, staining in the outer plexiformayer was virtually absent in rd/rd retinas, without signifi-ant changes within the inner plexiform layer (Fig. 6D–F).–y Axis bitmap analysis confirmed that synaptophysin

l Time PCR plots from cDNA serial dilutions indicated amplificationaptophysin (D) and synapsin I (E). Amplification efficiency (E) of the

compare the abundance of gene transcripts. GAPDH was used as anlevels (G). Bars represent standard error of mean. * P�0.01 vs. WT.

tina. Rea7 (C), syn

abeling is completely redistributed in rd/rd retina (Fig. 6G).

Qili

ap

pepiAc

F(pigii(ocpr

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005 999

uantification data revealed substantial changes in thenner/outer plexiform layer labeling ratio (Fig. 6H). Nonethe-ess, labeling ratio between OFF and ON sublaminas locatedn the inner plexiform layer was not altered (Fig. 6I).

Synapsin I. Synapsin I immunoreactivity was seen assolid, intense staining virtually filling up the entire inner

ig. 2. Expression of Cx36 in transverse sections of WT and rd/rd rePI), showing Cx36 expression (B). Superposition of red (PI) and grelexiform layer (ON sublamina) in WT retinas. Immunolabeling was al

n the outer plexiform layer. Vertical section of rd/rd retina counterstareen (Cx36) channels (F). Notice that the outer nuclear layer is virtual

n the labeling from the outer plexiform layer. In order to evaluate imn numeric format, where values correspond to the brightness of the pixen�3) of Cx36 distribution in the outer and inner plexiform layers, in Wf this protein (H). Quantification of Cx36 distribution in the OFF andhanges in the distribution of this protein (I). * P�0.01 vs. WT. Outerlexiform layer (ipl), and ganglion cell layer (gcl). Scale bar�30 �m.eferred to the Web version of this article.

lexiform layer of WT retinas (Fig. 7A–C). As described in n

revious studies (Mandell et al., 1990; Von Kriegsteint al., 1999), synapsin I labeling is absent from the outerlexiform layer. In rd/rd retinas we did not observe signif-

cant changes in the pattern described for WT (Fig. 7D–F).ccordingly, x–y axis bitmap analysis revealed no signifi-ant differences in the synapsin I distribution pattern (data

rtical section of WT retina (A) counter-stained using propidium iodide) channels (C). Cx36 puncta were seen in the inner half of the innered in the outer half of the inner plexiform layer (OFF sublamina) and

g PI (D) showing Cx36 distribution (E). Superposition of red (PI) andin the rd/rd retinas. Compared with WT, we found a marked reductioneling, x–y axis bitmap analysis was used to view the pixel valuesnalysis from images shown in B (red) and E (green) (G). Quantificationd rd/rd (green) retinas revealed significant changes in the expressionlamina, in WT (red) and rd/rd (green) retinas showed no significantlayer (onl), outer plexiform layer (opl), inner nuclear layer (inl), innerretation of the references to color in this figure legend, the reader is

tinas. Veen (Cx36so observined usinly absentmunolabls. Pixel aT (red) an

ON subnuclear

For interp

ot shown).

L

SprcagFoaTiaici(2ar

Ctg

tte2rlrtas

rso(cereetpet

FslPil glion celc f this arti

A. H. Kihara et al. / Neuroscience 151 (2008) 995–10051000

DISCUSSION

ack of photoreceptor signaling alters Cx expression

tudies using transgenic mice showed that Cx36 isresent in several retinal neuron types participating inod-mediated vision, including rod and/or cone photore-eptors, AII amacrine cells and cone bipolar cells, as wells in a small population of amacrine cells within the gan-lion cell layer (Deans et al., 2002; Degen et al., 2004;eigenspan et al., 2004). In the present study, we havebserved that ablation of photoreceptors caused a remark-ble reduction of Cx36 labeling in the outer plexiform layer.his result strongly suggested that Cx36 expression found

n this layer is mainly related to photoreceptor terminals,lthough the expression in this retinal layer could also

nvolve bipolar cell process (Guldenagel et al., 2000). Ac-ordingly, Cx36 was described in photoreceptors, althought is not consensus whether it occurs in rods and conesDeans et al., 2002; Dang et al., 2004; Feigenspan et al.,004). Nevertheless, GJs among cones and rods, as wells between rods and rods, have been identified in mouseetina by electron microscopy (Tsukamoto et al., 2001).

While ablation of photoreceptors might suggest thatx36 is expressed by multiple neuronal types, the lack of

he visual signaling triggered by the photoreceptors sug-

ig. 3. Expression of Cx45 in transverse sections of WT and rd/rd retinhowing Cx45 expression (B). Superposition of red (PI) and green (C

ayers as a fine punctate. Both ON and OFF sublaminas of the inner plI (D) showing Cx45 distribution (E). Superposition of red (PI) and gr

n the rd/rd retinas. We were not able to detect changes in Cx45 distribayer (opl), inner nuclear layer (inl), inner plexiform layer (ipl), and ganolor in this figure legend, the reader is referred to the Web version o

ests additional functional roles of Cx36 in the retina. In G

his regard, some studies determined that visual depriva-ion caused by prolonged dark adaptation changes Cxxpression in the retina (Yang and Wu, 1989; Kihara et al.,006a). Indeed, tracer coupling studies determined thategulation of GJ permeability in response to neuromodu-ators such as NO and dopamine, whose production andelease has been related to the ambient light levels, affectshe size of receptive fields (Becker et al., 1998). Functionalspects of the modulation of cell coupling have been con-idered in recent studies (Kihara et al., 2006c).

Particularly interesting is the expression of Cx45 in theodent retina. Whereas some authors claimed its expres-ion to be exclusively neuronal (Guldenagel et al., 2000),thers suggested that Cx45 is present in glial cells as wellZahs et al., 2003), a proposal supported by data fromentral (Kleopa et al., 2002) and in vitro evidence (Dermi-tzel et al., 2000). Despite these hypotheses, Cx45 in theetina was shown to be essential in the rod pathway (Max-iner et al., 2005) and therefore constitutes an obligatorylement in the visual circuitry. In this regard, the findinghat Cx45 expression was not regulated by the lack ofhotoreceptor signaling was quite surprising, but could bexplained as the resulting balance of distinct alterations inwo distinct cell populations, neurons and glial cell.

It is well established that horizontal cells exhibit strong

al section of WT retina (A) counterstained using propidium iodide (PI)nnels (C). Cx45 labeling was seen in both outer and inner plexiformyer were labeled. Vertical section of rd/rd retina counterstained using

5) channels (F). Notice that the outer nuclear layer is virtually absentparing WT vs. rd/rd retinas. Outer nuclear layer (onl), outer plexiform

l layer (gcl). Scale bar�30 �m. For interpretation of the references tocle.

as. Verticx45) chaexiform laeen (Cx4ution com

J coupling, forming extensive homologous network con-

ncc

bdp

Fslaal r plexifoi er is refe

Fchc(i

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005 1001

ection. Functional studies revealed that horizontal celloupling is virtually abolished in Cx57-deficient mice, indi-ating that Cx57 is the major GJ channel protein (Hom-

ig. 4. Expression of Cx57 in transverse sections of WT and rd/rd retinhowing Cx57 expression (B). Superposition of red (PI) and green (Cx

ocated in the outer plexiform layer. Vertical section of rd/rd retina counnd green (Cx57) channels (F). Notice that the outer nuclear layer is vs a disrupted streak, surrounding presumptive horizontal cell perikary

ayer (onl), outer plexiform layer (opl), inner nuclear layer (inl), innenterpretation of the references to color in this figure legend, the read

ig. 5. Expression of Cx57 in retinal whole-mounts of WT and rd/ralbindin-D (CB, red), a marker for horizontal cells in the mouse retina (orizontal cells. Confocal analysis of rd/rd retinal whole-mount showinounting and determined the co-localization coefficient between green

n�3, * P�0.01) when compared with WT (C), with no significant changes interpretation of the references to color in this figure legend, the reader is refe

ach et al., 2004), although one cannot rule out that thisisruption was caused by downstream alterations. In theresent study, by using purified antibodies specifically

al section of WT retina (A) counterstained using propidium iodide (PI)nels (C). Cx57 labeling was observed as a punctate forming a ribbon,d using PI (D) showing Cx57 distribution (E). Superposition of red (PI)bsent in the rd/rd retinas. In the rd/rd retinas, Cx57 labeling appeared

in the outer border of the inner nuclear layer (arrows). Outer nuclearrm layer (ipl), and ganglion cell layer (gcl). Scale bar�30 �m. Forrred to the Web version of this article.

cal analysis of WT retinal whole-mount showing Cx57 (green) andpuncta were observed in both perikarya and processes of presumptivereen) and calbindin-D (B). Computational analyses produced punctahannels. In rd/rd retinas, density of puncta were significantly reduced

as. Vertic57) chanterstaineirtually aa located

d. ConfoA). Cx57g Cx57 (gand red c

n Cx57/calbindin-D co-localization ratio (D). Scale bar�30 �m. Forrred to the Web version of this article.

rdz

nmcest

ccepstsar

Fibciicd(dav

A. H. Kihara et al. / Neuroscience 151 (2008) 995–10051002

aised against mouse Cx57, we provided additional evi-ence that Cx57 protein is expressed exclusively by hori-ontal cells in WT retina.

Previous reports determined that horizontal cells areot conspicuously affected in retinas of young rd/rd adultice (Strettoi and Pignatelli, 2000), although morphologi-

al changes could be detected in older animals (Strettoit al., 2003). Accordingly, we were not able to detectignificant changes in horizontal cell morphology, even

ig. 6. Expression of synaptophysin in transverse sections of WT andodide (PI) showing synaptophysin (SYP) expression (B). Superpositioright streak in the outer plexiform layer, and a moderate, diffuseounterstained using PI (D) showing SYP distribution (E). Superpositios virtually absent in the rd/rd retinas. Staining in the outer plexiform lanner plexiform layer. In order to evaluate immunolabeling, x–y axis bitmorrespond to the brightness of the pixels. Pixel analysis from imageistribution in the outer and inner plexiform layers, in WT (red) and rd/H). Quantification of Cx36 distribution in the OFF and ON sublaminaistribution of this protein (I). * P�0.01 vs. WT. Outer nuclear layer (onnd ganglion cell layer (gcl). Scale bar�30 �m. For interpretation ofersion of this article.

hough our immunohistochemistry data revealed spatial p

hanges in Cx57 expression, the (sole) protein forming GJhannels in these cells. These results, combined with genexpression analysis, suggest that horizontal cell couplingrovided by Cx57 is affected by the lack of photoreceptorignaling. Functionally, the size of horizontal cells’ recep-ive field largely depends on coupling extent, which washown to be controlled by specific neuromodulators suchs NO, retinoic acid, and dopamine, whose production/elease in turn depend on visual activity (Weiler and Ako-

inas. Vertical section of WT retina (A) counterstained using propidium(PI) and green (SYP) channels (C). SYP labeling was observed as afilling up the inner plexiform layer. Vertical section of rd/rd retina

(PI) and green (SYP) channels (F). Notice that the outer nuclear layervirtually absent in rd/rd retinas, without significant changes within thesis was used to view the pixel values in numeric format, where valuesin B (red) and E (green) (G). Quantification (n�3) of synaptophysin) retinas revealed significant changes in the expression of this protein(red) and rd/rd (green) retinas showed no significant changes in thelexiform layer (opl), inner nuclear layer (inl), inner plexiform layer (ipl),nces to color in this figure legend, the reader is referred to the Web

rd/rd retn of redlabelingn of redyer wasap analy

s shownrd (green, in WT

l), outer pthe refere

ian, 1992). It is feasible that visual signaling also controls

t(aSma

Ep

Tvtbcdarrbdfrbg

dt

AgpesboditMsmtfJrrosd

Ic

Fsigc yer (opl),( is figure

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005 1003

he coupling by transcriptional and/or translation processesBecker et al., 1998; Kihara et al., 2006a), taking intoccount the fast turnover of Cxs (Beardslee et al., 1998;affitz et al., 2000). In accordance, we previously deter-ined that Cx57 is promptly down-regulated during darkdaptation (Kihara et al., 2006a).

ffect of photoreceptor loss in the expression ofresynaptic proteins

he retina contains two ultrastructurally distinct types ofesicle-containing synapses: in the outer plexiform layerhere are ribbon synapses, formed by photoreceptor andipolar cells, whereas in the inner plexiform layer thehemical communication between neurons is mainly me-iated by conventional synapses, made predominantly bymacrine cells (Von Kriegstein et al., 1999). In previouseports, it was determined that synapsin I is absent fromibbon synapses, whereas synaptophysin is expressed byoth ribbon and conventional presynaptic elements (Man-ell et al., 1990). In accordance to those reports, we haveound in the present study a similar pattern in both WT andd/rd retina for synapsin I. However, synaptophysin distri-ution was clearly altered in rd/rd retinas, as well as itsene expression levels.

Modification in the strength of synapses by activity-ependent changes constitutes a cellular mechanism for

ig. 7. Expression of synapsin I in transverse sections of WT and rd/rd rehowing synapsin I (SYS) expression (B). Superposition of red (PI) and gnner plexiform layer of WT retinas. Vertical section of rd/rd retina countereen (SYS) channels (F). Notice that the outer nuclear layer is virtually aomparing WT vs. rd/rd retinas. Outer nuclear layer (onl), outer plexiform lagcl). Scale bar�30 �m. For interpretation of the references to color in th

he plasticity of neuronal networks (Faber et al., 1991; e

barbanel et al., 2002; Derkach et al., 2007). In this re-ard, neuronal plasticity is achieved, at least in part, byresynaptic modulation of neurotransmitter release (Turnert al., 1999; Rettig and Neher, 2002). Despite the fact thatynaptophysin and synapsin I were differently affected,oth proteins could be easily detected in the retina, despitef the complete loss of visual signaling. This finding en-orsed the hypothesis that these genes might also be

nvolved in other cell functions, not necessarily associatedo the release of neurotransmitters (Schwarz, 1994; Mc-ahon et al., 1996; Ferreira and Rapoport, 2002). Notably,

tudies using synaptophysin- and synapsin I- deficientice revealed that both proteins perform essential func-

ions in synaptic plasticity without necessarily being crucialor neurotransmitter release itself (Rosahl et al., 1993;anz et al., 1999). Conversely, it is plausible that theelease of neurotransmitters in physiological conditionsegulates, but is not essential to maintain, the expressionf these proteins. This particular supposition may be theubject of additional investigation and finds supportingata in this study.

CONCLUSION

n the present study, we reported that the loss of photore-eptor signaling elicits specific, differential alterations of

rtical section of WT retina (A) counterstained using propidium iodide (PI)S) channels (C). SYS labeling was seen as a solid staining filling up theusing PI (D) showing SYS distribution (E). Superposition of red (PI) andhe rd/rd retinas. We were not able to detect changes in SYS distributioninner nuclear layer (inl), inner plexiform layer (ipl), and ganglion cell layer

legend, the reader is referred to the Web version of this article.

tinas. Vereen (SYrstainedbsent in t

lectrical and chemical synaptic proteins in the retina,

eube

AAf0r

A

B

B

B

C

C

D

D

D

D

D

E

F

F

F

F

G

H

H

H

J

J

K

K

K

K

M

M

M

M

M

M

P

R

R

R

S

S

S

S

A. H. Kihara et al. / Neuroscience 151 (2008) 995–10051004

specially in the outer plexiform layer. Our current resultsnequivocally revealed that neuronal signaling regulates,ut is not essential to maintain, the expression of synapticlements.

cknowledgments—The authors would like to thank Adilsonlves, Leandro Castro, Elisangela Vicentine and Sayami Akamine

or technical assistance. Grant sponsor: FAPESP (grant number:6/51094-1) and CNPq. A.H.K., T.O.S., R.J.B.M., and V.P. areecipients of fellowships from FAPESP.

REFERENCES

barbanel HD, Huerta R, Rabinovich MI (2002) Dynamical model oflong-term synaptic plasticity. Proc Natl Acad Sci U S A 99:10132–10137.

eardslee MA, Laing JG, Beyer EC, Saffitz JE (1998) Rapid turnoverof connexin43 in the adult rat heart. Circ Res 83:629–635.

ecker D, Bonness V, Mobbs P (1998) Cell coupling in the retina:patterns and purpose. Cell Biol Int 22:781–792.

litz DM, Foster KA, Regehr WG (2004) Short-term synaptic plasticity:a comparison of two synapses. Nat Rev Neurosci 5:630–640.

ondorelli DF, Parenti R, Spinella F, Trovato Salinaro A, Belluardo N,Cardile V, Cicirata F (1998) Cloning of a new gap junction gene(Cx36) highly expressed in mammalian brain neurons. Eur J Neu-rosci 10:1202–1208.

oppen SR, Dupont E, Rothery S, Severs NJ (1998) Connexin45expression is preferentially associated with the ventricular conduc-tion system in mouse and rat heart. Circ Res 82:232–243.

ang L, Pulukuri S, Mears AJ, Swaroop A, Reese BE, Sitaramayya A(2004) Connexin 36 in photoreceptor cells: studies on transgenicrod-less and cone-less mouse retinas. Mol Vis 10:323–327.

eans MR, Volgyi B, Goodenough DA, Bloomfield SA, Paul DL (2002)Connexin36 is essential for transmission of rod-mediated visualsignals in the mammalian retina. Neuron 36:703–712.

egen J, Meier C, Van Der Giessen RS, Sohl G, Petrasch-Parwez E,Urschel S, Dermietzel R, Schilling K, De Zeeuw CI, Willecke K(2004) Expression pattern of lacZ reporter gene representing con-nexin36 in transgenic mice. J Comp Neurol 473:511–525.

erkach VA, Oh MC, Guire ES, Soderling TR (2007) Regulatorymechanisms of AMPA receptors in synaptic plasticity. Nat RevNeurosci 8:101–113.

ermietzel R, Gao Y, Scemes E, Vieira D, Urban M, Kremer M,Bennett MV, Spray DC (2000) Connexin43 null mice reveal thatastrocytes express multiple connexins. Brain Res Brain Res Rev32:45–56.

lferink LA, Scheller RH (1995) Synaptic vesicle proteins and regu-lated exocytosis. Prog Brain Res 105:79–85.

aber DS, Lin JW, Korn H (1991) Silent synaptic connections and theirmodifiability. Ann N Y Acad Sci 627:151–164.

eigenspan A, Janssen-Bienhold U, Hormuzdi S, Monyer H, Degen J,Sohl G, Willecke K, Ammermuller J, Weiler R (2004) Expression ofconnexin36 in cone pedicles and OFF-cone bipolar cells of themouse retina. J Neurosci 24:3325–3334.

eigenspan A, Teubner B, Willecke K, Weiler R (2001) Expression ofneuronal connexin36 in AII amacrine cells of the mammalian ret-ina. J Neurosci 21:230–239.

erreira A, Rapoport M (2002) The synapsins: beyond the regulationof neurotransmitter release. Cell Mol Life Sci 59:589–595.

uldenagel M, Sohl G, Plum A, Traub O, Teubner B, Weiler R,Willecke K (2000) Expression patterns of connexin genes in mouseretina. J Comp Neurol 425:193–201.

averkamp S, Wassle H (2000) Immunocytochemical analysis of themouse retina. J Comp Neurol 424:1–23.

ilfiker S, Pieribone VA, Czernik AJ, Kao HT, Augustine GJ, Green-gard P (1999) Synapsins as regulators of neurotransmitter release.

Philos Trans R Soc Lond B Biol Sci 354:269–279.

ombach S, Janssen-Bienhold U, Sohl G, Schubert T, Bussow H, OttT, Weiler R, Willecke K (2004) Functional expression of con-nexin57 in horizontal cells of the mouse retina. Eur J Neurosci19:2633–2640.

anz R, Sudhof TC, Hammer RE, Unni V, Siegelbaum SA, BolshakovVY (1999) Essential roles in synaptic plasticity for synaptogyrin Iand synaptophysin I. Neuron 24:687–700.

iang JX, Gu S (2005) Gap junction- and hemichannel-independentactions of connexins. Biochim Biophys Acta 1711:208–214.

ihara AH, de Castro LM, Moriscot AS, Hamassaki DE (2006a) Pro-longed dark adaptation changes connexin expression in the mouseretina. J Neurosci Res 83:1331–1341.

ihara AH, Mantovani de Castro L, Belmonte MA, Yan CY, MoriscotAS, Hamassaki DE (2006b) Expression of connexins 36, 43, and45 during postnatal development of the mouse retina. J Neurobiol66:1397–1410.

ihara AH, Tsurumaki AM, Ribeiro-do-Valle LE (2006c) Effects ofambient lighting on visual discrimination, forward masking andattentional facilitation. Neurosci Lett 393:36–39.

leopa KA, Yum SW, Scherer SS (2002) Cellular mechanisms ofconnexin32 mutations associated with CNS manifestations. J Neu-rosci Res 68:522–534.

andell JW, Townes-Anderson E, Czernik AJ, Cameron R, GreengardP, De Camilli P (1990) Synapsins in the vertebrate retina: absencefrom ribbon synapses and heterogeneous distribution among con-ventional synapses. Neuron 5:19–33.

axeiner S, Dedek K, Janssen-Bienhold U, Ammermuller J, Brune H,Kirsch T, Pieper M, Degen J, Kruger O, Willecke K, Weiler R (2005)Deletion of connexin45 in mouse retinal neurons disrupts the rod/cone signaling pathway between AII amacrine and ON cone bipo-lar cells and leads to impaired visual transmission. J Neurosci25:566–576.

cMahon HT, Bolshakov VY, Janz R, Hammer RE, Siegelbaum SA,Sudhof TC (1996) Synaptophysin, a major synaptic vesicle protein,is not essential for neurotransmitter release. Proc Natl Acad SciU S A 93:4760–4764.

edhurst AD, Harrison DC, Read SJ, Campbell CA, Robbins MJ,Pangalos MN (2000) The use of TaqMan RT-PCR assays forsemiquantitative analysis of gene expression in CNS tissues anddisease models. J Neurosci Methods 98:9–20.

eier C, Petrasch-Parwez E, Habbes HW, Teubner B, Guldenagel M,Degen J, Sohl G, Willecke K, Dermietzel R (2002) Immunohisto-chemical detection of the neuronal connexin36 in the mouse cen-tral nervous system in comparison to connexin36-deficient tissues.Histochem Cell Biol 117:461–471.

ills SL, O’Brien JJ, Li W, O’Brien J, Massey SC (2001) Rod pathwaysin the mammalian retina use connexin 36. J Comp Neurol 436:336–350.

arnas H, Parnas I (2007) The chemical synapse goes electric: Ca2�-and voltage-sensitive GPCRs control neurotransmitter release.Trends Neurosci 30:54–61.

ettig J, Neher E (2002) Emerging roles of presynaptic proteins inCa��-triggered exocytosis. Science 298:781–785.

osahl TW, Geppert M, Spillane D, Herz J, Hammer RE, Malenka RC,Sudhof TC (1993) Short-term synaptic plasticity is altered in micelacking synapsin I. Cell 75:661–670.

ozental R, Giaume C, Spray DC (2000) Gap junctions in the nervoussystem. Brain Res Brain Res Rev 32:11–15.

affitz JE, Laing JG, Yamada KA (2000) Connexin expression andturnover: implications for cardiac excitability. Circ Res 86:723–728.

chwarz TL (1994) Genetic analysis of neurotransmitter release at thesynapse. Curr Opin Neurobiol 4:633–639.

heng M, Kim MJ (2002) Postsynaptic signaling and plasticity mech-anisms. Science 298:776–780.

ohl G, Odermatt B, Maxeiner S, Degen J, Willecke K (2004) Newinsights into the expression and function of neural connexins with

transgenic mouse mutants. Brain Res Brain Res Rev 47:245–259.

S

S

S

S

T

T

T

V

V

W

W

Y

Z

Z

Z

A. H. Kihara et al. / Neuroscience 151 (2008) 995–1005 1005

pray DC, Ye ZC, Ransom BR (2006) Functional connexin“hemichannels”: a critical appraisal. Glia 54:758–773.

terling P, Matthews G (2005) Structure and function of ribbon syn-apses. Trends Neurosci 28:20–29.

trettoi E, Pignatelli V (2000) Modifications of retinal neurons in amouse model of retinitis pigmentosa. Proc Natl Acad Sci U S A97:11020–11025.

trettoi E, Pignatelli V, Rossi C, Porciatti V, Falsini B (2003) Remod-eling of second-order neurons in the retina of rd/rd mutant mice.Vision Res 43:867–877.

raub RD, Kopell N, Bibbig A, Buhl EH, LeBeau FE, Whittington MA(2001) Gap junctions between interneuron dendrites can enhancesynchrony of gamma oscillations in distributed networks. J Neuro-sci 21:9478–9486.

sukamoto Y, Morigiwa K, Ueda M, Sterling P (2001) Microcircuits fornight vision in mouse retina. J Neurosci 21:8616–8623.

urner KM, Burgoyne RD, Morgan A (1999) Protein phosphorylationand the regulation of synaptic membrane traffic. Trends Neurosci22:459–464.

aney DI (2002) Retinal neurons: cell types and coupled networks.

Prog Brain Res 136:239–254.

on Kriegstein K, Schmitz F, Link E, Sudhof TC (1999) Distribution ofsynaptic vesicle proteins in the mammalian retina identifies oblig-atory and facultative components of ribbon synapses. Eur J Neu-rosci 11:1335–1348.

eiler R, Akopian A (1992) Effects of background illuminations on thereceptive field size of horizontal cells in the turtle retina are medi-ated by dopamine. Neurosci Lett 140:121–124.

illecke K, Eiberger J, Degen J, Eckardt D, Romualdi A, GuldenagelM, Deutsch U, Sohl G (2002) Structural and functional diversity ofconnexin genes in the mouse and human genome. Biol Chem383:725–737.

ang XL, Wu SM (1989) Modulation of rod-cone coupling by light.Science 244:352–354.

ahs KR, Kofuji P, Meier C, Dermietzel R (2003) Connexin immuno-reactivity in glial cells of the rat retina. J Comp Neurol 455:531–546.

hu Y, Tu DC, Denner D, Shane T, Fitzgerald CM, Van Gelder RN(2007) Melanopsin-dependent persistence and photopotentiationof murine pupillary light responses. Invest Ophthalmol Vis Sci 48:1268–1275.

immermann H (1997) Are mechanisms of exocytosis neurotransmit-

ter specific? Neurochem Int 31:759–761.

(Accepted 18 December 2007)(Available online 8 December 2007)