B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

ava i l ab l e a t www.sc i enced i rec t . com

www.e l sev i e r. com/ l oca te /b ra in res

Research Report

Distribution of vesicular glutamate transporters in ratand human retina

Jie Gonga, Abdeljelil Jellalib, Jérome Muttererc, José A. Sahela,d,e,Alvaro Rendona, Serge Picauda,d,⁎aINSERM U-592, Université Pierre et Marie Curie-Paris6,Laboratoire de Physiopathologie Cellulaire et Moléculaire de la Rétine, F-75571 Paris, FrancebClinique de la souris, IGBMC, F-67404 Illkirch, FrancecInstitut de Biologie Moléculaire des Plantes, Université Louis Pasteur, F-67084 Strasbourg, FrancedFondation Ophtalmologique Adolphe de Rothschild, F-75012 Paris, FranceeCentre Hospitalier National d'Ophtalmologie des Quinze-Vingts, F-75012 Paris, France

A R T I C L E I N F O

⁎ Corresponding author. INSERM U-592, Bâtim149284605.

E-mail address: picaud@st-antoine.inserm

0006-8993/$ – see front matter © 2006 Elsevidoi:10.1016/j.brainres.2006.01.111

A B S T R A C T

Article history:Accepted 25 January 2006Available online 6 March 2006

Central nervous system neurons have traditionally been thought to express exclusivelymembrane transporters and/or vesicular transporters for their transmitter. Three vesicularglutamate transporters have recently been cloned: BNPI/VGLUT1 (a brain-specific sodium-dependent inorganic phosphate (Pi) transporter), and its homologs DNPI/VGLUT2(differentiation-associated sodium-dependent Pi transporter) and VGLUT3. We investigatedthe subcellular distributions of these three vesicular transporters in rat and human retina.VGLUT1 was present in the outer and inner plexiform layers (OPL and IPL), as shown bypunctate staining in both human and rat retina. In the OPL, it was colocalized withsynaptophysin, consistent with its expression in glutamatergic photoreceptor terminals, andit was present in PKC-α-labeled glutamatergic bipolar cell terminals in the IPL. By contrast,VGLUT2 was present in horizontal cells and ganglion cells in rat and human retina. In humanretina, VGLUT2 was also found in some amacrine cells, including GAD-immunopositiveamacrine cells. VGLUT3 was present in glycine-releasing amacrine cells in rat retina but wasrestricted to a few ganglion cells in human retina. The distribution of VGLUT1 in excitatorysynaptic terminal was consistent with its involvement in glutamate release at excitatorysynapses, whereas the cellular distributions of VGLUT2 and VGLUT3 suggested that thesemolecules may be involved in functions other than glutamate release, such as glutamatestorage for GABA synthesis in non-glutamatergic neurons.

© 2006 Elsevier B.V. All rights reserved.

Keywords:RetinaSynapseGlutamateGABAVesicular glutamate transporter

1. Introduction

Neurotransmitters were originally characterized in terms oftheir Ca2+-dependent vesicular release and their reuptake into

ent Kourilsky, 184 rue du

.fr (S. Picaud).

er B.V. All rights reserved

synaptic terminals via membrane transporters and intosynaptic vesicles via vesicular transporters. These classicsteps in synaptic transmission do not apply to unconventionalneurotransmitters such as nitric oxide (NO), which diffuses

Faubourg Saint-Antoine, F-75571 Paris Cedex 12, France. Fax: +33

.

Fig. 1 – Distribution of VGLUT1 and VGLUT2 in rat retina.(A, B) Rat retina sections immunolabeled for VGLUT1 (A)and visualized under Nomarski optics (B).VGLUT1-immunopositive structures were present in theouter plexiform layer (OPL) and throughout the laminae ofthe inner plexiform layer (IPL) and were detected aspunctate staining. (C, D) Rat retinal sectionsimmunolabeled for VGLUT2 (C) and visualized underNomarski optics (D). Faint VGLUT2 immunolabeling wasobserved in the OPL and IPL and in cell bodies in theinner nuclear layer (INL) and ganglion cell layer (GCL).Scale bars in panels A, C represent 10 μm.

74 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

through membranes. The traditional model of neurotrans-mitters being taken up into glial cells and presynapticterminals via membrane transporters has also been stronglychallenged over the last decade. For example, glutamatetransporters are now known not to be restricted to glutama-tergic neurons; instead, they are also expressed by cerebellarGABAergic Purkinje cells at a postsynaptic location (Rothsteinet al., 1994), in which they may accelerate postsynapticresponse recovery (Takahashi et al., 1996). Retinal horizontalcells have also been reported to accumulate glutamate(Schutte and Schlemermeyer, 1993) and to express glutamatetransporters, even though these neurons are generallythought not to release glutamate, with GABA release consid-ered more likely. Glutamate transporters have also beenreported to act as presynaptic receptors in photoreceptors(Picaud et al., 1995) or as postsynaptic receptors mediating thelight response in fish bipolar retinal cells (Grant and Dowling,1995).

Vesicular transporters for several amino-acid neurotrans-mitters have recently been cloned or identified (Gasnier,2000). Brain-specific Na+-dependent inorganic phosphatetransporter I (BNPI) (Ni et al., 1994) was found to loadglutamate into synaptic vesicles and was therefore renamedVGLUT1 (Takamori et al., 2000). Mutation of the eat-4 gene,encoding a BNPI/VGLUT1 homolog in C. elegans, completelysuppressed glutamatergic transmission, depleting synapticvesicles of glutamate (Lee et al., 1999). Conversely, inducingVGLUT1 expression in cultured GABAergic neurons elicitedglutamate release and receptor activation in postsynapticcells (Takamori et al., 2000). “Differentiation-associatedBNPI” (DNPI) or VGLUT2 was identified on the basis of itssimilar sequence and was shown to transport glutamateinto synaptic vesicles (Bai et al., 2001; Hayashi et al., 2001;Herzog et al., 2001). In the mammalian brain, VGLUT1 andVGLUT2 are localized in different subsets of glutamatergicneurons (Bellocchio et al., 1998; Takamori et al., 2000;Aihara et al., 2000; Herzog et al., 2001). Surprisingly, athird vesicular transporter, VGLUT3, also identified bysequence analogy, was recently identified in cholinergic,serotoninergic and GABAergic neurons, suggesting a possi-ble corelease of glutamate with another neurotransmitter(Fremeau et al., 2002; Gras et al., 2002; Schafer et al., 2002;Haverkamp et al., 2004; Haverkamp and Wässle, 2004;Somogyi et al., 2004).

In the retina, photoreceptors, bipolar cells and ganglioncells are generally considered to be glutamatergic neurons,with ganglion cells having their synaptic terminals in thebrain. The distributions of the various VGLUT proteins in ratand mouse retina have been described, with VGLUT1 detectedin bipolar cells and photoreceptor terminals, whereas VGLUT2was found only in Müller cell processes and, to a lesser extent,in ganglion cell bodies (Johnson et al., 2003). Finally, VGLUT3was localized in a few retinal amacrine cells (Fremeau et al.,2002; Haverkamp and Wässle, 2004; Johnson et al., 2004).

In this study, we compared the distributions of the variousVGLUTs in rat and human retina with a great emphasis ontheir subcellular localizations. The objective was to gain moreinformation on their functional role in synaptic transmission.In both rat and human retinas, we observed that VGLUT1 waspresent in glutamatergic photoreceptor and bipolar cell

terminals. By contrast, VGLUT2 was found in horizontal cellsand non-glutamatergic amacrine cells. VGLUT3 was detectedin one subset of rat amacrine cells but was localized inganglion cells in human retina. Thus, VGLUT2 and VGLUT3may not constitute specific markers of glutamatergictransmission.

2. Results

2.1. VGLUT1 localization in rat retina

In rat retina (Figs. 1A, B), VGLUT1 was detected in the outerplexiform layer (OPL) and throughout the laminae of the innerplexiform layer (IPL), a distribution consistent with theexpected synaptic localization of the protein. The subcellulardistribution of VGLUT1 in the OPL was investigated by labelingthe photoreceptor terminals with an antibody directed againstsynaptophysin, a synaptic vesicle protein found exclusively in

75B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

the photoreceptor terminals of the OPL (Brandstatter et al.,1996). All VGLUT1-immunopositive structures in the OPLcolocalized with synaptophysin (Figs. 2A–C), confirming thatVGLUT1 was located in photoreceptor terminals. We tried tolocate VGLUT1 with respect to the photoreceptor ribbonsynapse by labeling retinal sections with the cytomatrix

Fig. 2 – Subcellular localization of VGLUT1 in the rat outer pimmunolabeled or stained for VGLUT1 (red in A, C, D, F–H, J–L, Nlectin (PNA, green in E, F), Bassoon (Bass, green in G, I, J), dystropcalbindin (Calb, green in R, S). Structures indicated by arrows inrespectively). VGLUT1 (A) and the synaptic vesicles, identified byimage, C), suggesting that VGLUT1 is expressed in photoreceptorof the PNA-stained cone photoreceptor terminals (E, F). Bassoon-VGLUT1-immunopositive structures (G, H, J). The dystrophin labenclosed in VGLUT1-immunopositive structures (K, L, N). Finallyapposition to PKC-α-positive rod bipolar cell dendrites (P, Q) andrepresent 5 μm in panels A–G, K, P, S and 0.5μm in panels H–J, L

protein Bassoon (Brandstatter et al., 1999; Dick et al., 2001).Double immunolabeling showed that Bassoon-immunoposi-tive synaptic ribbonswere included in the immunostaining forVGLUT1 (Figs. 2G–J), consistent with the localization ofVGLUT1 in synaptic vesicles at the photoreceptor ribbonsynapse and, more generally, in the photoreceptor terminal.

lexiform layer. Confocal microscopy of rat retinal sections–S), synaptophysin (SVP38, green in B, C), peanut agglutininhin (Dys, green in K, M, N), PKC-α, (PKC, green in P, Q) and(G, K, P, S) are enlarged in small insets (H–J, L–N, Q, R,staining for synaptophysin (B), were colocalized (mergedterminals. VGLUT1-immunopositive structures (D) lay on topimmunopositive synaptic ribbons (G, I) were enclosed ineling of photoreceptor terminals (K, M, N) was only partly, VGLUT1-immunopositive structures were found in closecalbindin-positive horizontal cell tips (R, S). Scale bars–N, Q, R.

76 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

The yellow color of the Bassoon-immunopositive ribbon in thesuperimposed images indicates its inclusion in the VGLUT1staining (Figs. 2G, J). We estimated the extent of VGLUT1distribution by labeling retinal sections with an antibodydirected against dystrophin, a cytoskeletal protein located inrod spherules and cone pedicles in the areas adjacent tobipolar cell membranes at locations distant from the synapticribbon (Ueda et al., 1997a,b). Dystrophin immunolabeling waspunctate and not entirely colocalized with VGLUT1-immu-nopositive structures in the OPL; it was instead apposed tothese structures (Figs. 2K–N). This observation suggested thatVGLUT1-positive synaptic vesicles did not entirely fill thephotoreceptor terminals and were instead concentratedaround the ribbon synapse. We investigated whetherVGLUT1 was present in rod spherules and cone pedicles byincubating retinal sections with peanut agglutinin lectin(PNA), which labels cone pedicles. PNA-positive cone pedicleswere observed at the base of VGLUT1-immunolabeled struc-tures (Figs. 2D–F). These observations are consistent withVGLUT1 expression in both PNA-stained cone pedicles androd photoreceptor spherules. VGLUT1 expression in rodspherules was further supported by the immunolabeling ofrod bipolar cells with an antibody directed against the α-isoform of protein kinase C (PKC-α). Rod bipolar cell dendritictips were indeed often seen in close apposition to VGLUT1-immunolabeled structures (Figs. 2P, Q). Similarly, calbindin-immunopositive horizontal cell dendritic tips were found inclose apposition to VGLUT1-immunolabeled structures (Figs.2R, S). These observations are consistent with the localizationof VGLUT1 in rod and cone photoreceptor terminals in ratretina.

VGLUT1-immunopositive puncta in the IPL were distribut-ed throughout the laminae of the IPL. Small puncta weredetected in the outer part of the IPL, with larger punctadetected close to the ganglion cell layer, which contains theaxon terminals of rod bipolar cells (Fig. 3A). All VGLUT1-immunopositive structures in the IPL were also synaptophy-sin-immunopositive. By contrast, not all synaptophysin-positive structures, which included bipolar cell and amacrinecell terminals (Brandstatter et al., 1996), were labeled with theVGLUT1 antibody (Figs. 3A–C). We assessed the cellulardistribution of VGLUT1 further by staining rod bipolar cellswith the PKC-α antibody. The VGLUT1-positive large puncta inthe inner part of the IPL were PKC-α-positive and thereforecorresponded to rod bipolar cell axon terminals (Figs. 3D–F).Inhibitory amacrine cell synapses were identified by Bassoonimmunostaining in the IPL (Dick et al., 2001). No colocalizationwith VGLUT1 staining was observed, but Bassoon-positivesynapses were often found in apposition to VGLUT1-positiveexcitatory terminals, in both the OFF and ON sublaminae ofthe IPL (Figs. 3G, I–N). This observation is consistent with theabsence of VGLUT1 in inhibitory amacrine cells. SmallVGLUT1-immunopositive puncta were also seen in closeapposition to neurofilament-immunostained ganglion celldendrites (Figs. 3H, O–Q), consistent with the absence ofVGLUT1 in ganglion cells. Since neither Bassoon staining forinhibitory amacrine cells nor neurofilament staining ofganglion cell dendrites was colocalized with VGLUT1 in theIPL, the results are consistent with the expression of VGLUT1in bipolar cell terminals.

2.2. VGLUT1 localization in human retina

As in rat retina, VGLUT1-immunopositive structures werepresent in both the OPL and IPL of human retina (Fig. 4A).Labeling was almost continuous in the OPL but restricted toisolated puncta in the IPL. When retinal sections wereimmunostained for the synaptic vesicle protein, synaptophy-sin, all VGLUT1-positive structures colocalized with synapto-physin but few synaptophysin-positive structures did notlabel for VGLUT1 (Figs. 4A–C). Labeling of the retina withcalbindin to identify cone photoreceptors in the OPL made itpossible to demonstrate that VGLUT1 immunolabeling wasobserved in cone photoreceptor terminals (Figs. 4D–F), extend-ing into the cell bodies and axons of these structures (arrowsin Figs. 4D–F). VGLUT1-immunopositive structures enclosedBassoon-immunopositive ribbon synapses (Fig. 4G) and werein close contact with rod bipolar cell dendritic tips, as shownby PKC-α immunolabeling (Fig. 4I). Some VGLUT1-positivestructures in the IPL were identified as rod bipolar cell axonterminals by PKC-α immunolabeling (Figs. 4J–L). MostVGLUT1-positive structures were found in close appositionto Bassoon-positive amacrine cell synapses (Fig. 4H). As in ratretina, VGLUT1 was thus located in the terminals of rods,cones and bipolar cells.

2.3. VGLUT2 localization in rat retina

The distribution of VGLUT2 in rat retina was very differentfrom that of VGLUT1. Weak VGLUT2 labeling was observedin the processes of the OPL and in the somata of the innernuclear layer (INL) and ganglion cell layer (GCL) (Figs. 1C, D).Double labeling with antibodies against VGLUT2 andcalbindin (to identify horizontal cells) showed that theVGLUT2-positive cell bodies in the INL were horizontalcells and that VGLUT2-immunopositive structures in theOPL corresponded to the processes of these cells (Figs. 5A–C). VGLUT2 has been detected in dopaminergic neurons(Trudeau, 2004). We therefore double-labeled the retinafor VGLUT2 and tyrosine hydroxylase (TH), an enzyme-regulating dopamine synthesis. In rat retina, VGLUT2 wasnot localized in TH-positive amacrine cells (Figs. 5D–F). Forthe identification of VGLUT2-positive cells in the GCL,retinal sections were immunolabeled with an antibodyagainst neurofilament 68 (NF68) that labels both horizontaland ganglion cells. VGLUT2 was localized in NF68-positivecells in both the INL and the GCL (Figs. 5G–L), consistentwith the presence of VGLUT2 in horizontal cells andganglion cells.

2.4. VGLUT2 localization in human retina

The distribution of VGLUT2 in human retina differed fromthat in rat retina. VGLUT2 was present in many cell bodies ofthe INL and GCL and throughout the IPL (Fig. 6A). As in ratretina, horizontal cells, identified with an antibody againstparvalbumin, were found to have VGLUT2-immunopositivecell bodies and processes (Figs. 6B–D). For the identification ofother VGLUT2-positive cells in the INL, amacrine cells wereidentified by incubation with an antibody against syntaxinand double labeling for VGLUT2 and syntaxin showed that

Fig. 3 – Subcellular localization of VGLUT1 in the rat inner plexiform layer. Confocal microscopy of rat retina sectionsimmunolabeled for VGLUT1 (red in A, C, D, F–I, K, L, N, O, Q), synaptophysin (SVP38, green in B,C), PKC-α (PKC, green in E, F),Bassoon (Bass, green in G, J, K, M, N) and neurofilament (NF68, green in H, P, Q). Frames shown in panels G, H are enlarged asinsets (I–Q). VGLUT1-positive structures were also synaptophysin-positive (C), but not all synaptophysin-positive structureswere VGLUT1-positive. VGLUT1 (D, F) and PKC-α (E, F) were colocalized in the inner part of the IPL, indicating VGLUT1expression in rod bipolar terminals. By contrast, VGLUT1 was found in apposition to Bassoon-positive inhibitory terminals inthe IPL. This appositionwas observed in the OFF (G, I–K) and ON sublaminae (G, L–N) of the IPL. VGLUT1-positive terminalswerealso found in close apposition to ganglion cell dendrites, labeled for neurofilament NF68 (H, O–Q). Scale bars represent 5 μm inpanels A–H and 0.5 μm in panels I–Q.

77B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

many amacrine cells contained VGLUT2 (Figs. 6E–G). VGLUT2-positive amacrine cells were further identified by labelingretinal sections with an antibody directed against the 65/67isoforms of the GABA-synthesizing enzyme glutamic aciddecarboxylase (GAD). The GAD antibody labeled a populationof amacrine cells in the inner part of the INL, some of whichwere VGLUT2-immunopositive (Figs. 6H–J). Conversely,

some amacrine cells were VGLUT2-immunopositive butGAD-immunonegative. Similarly, some VGLUT2-immunopo-sitive puncta colocalized with GAD-positive structures in theIPL, but not all VGLUT2-immunopositive processes were alsopositive for GAD (Fig. 6J). Thus, VGLUT2 is expressed in non-glutamatergic horizontal cells and in GABAergic amacrinecells.

Fig. 4 – VGLUT1 distribution in human retina. Confocal microscopy of human retinal sections immunolabeled for VGLUT1 (redin A, C, D, F–J, L), synaptophysin (SVP38, green in B, C), calbindin (calb, green in E, F), Bassoon (Bass, green in G, H) and PKC-α,(PKC, green in I, K, L). All VGLUT1-positive structures appeared synaptophysin-positive (C), in both the OPL and IPL. In the OPL(D), calbindin-positive cone photoreceptors contained VGLUT1 from the cell body (arrows in D–F) to the axon terminal(arrowheads in D–F). In the OPL, Bassoon-immunopositive synaptic ribbons were enclosed in VGLUT1-immunopositivestructures (G), whereas, in the IPL, Bassoon-immunopositive inhibitory synapses were apposed to VGLUT1-positive terminals(H). In the OPL (I), PKC-α-labeled rod bipolar cell dendrites were apposed to VGLUT1-positive photoreceptor terminals, whereas,in the IPL, VGLUT1was observed in rod bipolar cell terminals (arrows in J–L). Scale bars represent 5μm in panels G–L and 10μmin panels A–F.

78 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

2.5. VGLUT3 localization in rat retina

VGLUT3hasbeendetected in a subtypeof glycinergic amacrinecells in rat retina (Fremeau et al., 2002;Gras et al., 2002; Johnsonet al., 2004; Haverkamp and Wässle, 2004). We detectedVGLUT3 in cell bodies in the INL and their correspondingprocesses in the IPL (Fig. 7A). We immunolabeled retinalsections for glycine transporter 1 (GLYT1), a specific markerof glycinergic amacrine cells. The distribution of immunos-

taining for GLYT1 perfectlymatched that for VGLUT3 (Figs. 7A–C). This pattern of staining confirmed that VGLUT3 wasexpressed in a subtype of glycinergic amacrine cells (Haver-kamp and Wässle, 2004; Johnson et al., 2004).

2.6. VGLUT3 localization in human retina

The distribution of VGLUT3 in human retina was differentfrom that in rat retina. VGLUT3 was detected as puncta in

Fig. 5 – Cellular localization of VGLUT2 in rat retina. Confocalmicroscopy of rat retinal sections immunolabeled for VGLUT2 (redin A, C, D, F, G, I), calbindin (calb, green in B, C), tyrosine hydroxylase (TH, green in E, F) and neurofilament (NF68, green in H, I).(A–C) In the outer retina, VGLUT2-positive structures contained calbindin-immunopositive horizontal cells. (D–F)VGLUT2-positive structures did not colocalize with TH-immunopositive dopaminergic amacrine cells. (G–I) In the GCL,VGLUT2-positive structures were observed in neurofilament-positive ganglion cells. Scale bars represent 10 μm in panels A–I.

79B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

a few cell bodies in the GCL (Fig. 7D). We determinedwhether these VGLUT3-positive cells were ganglion cellsby double labeling retinal sections with an antibodydirected against neurofilament 200 that specifically identi-fies ganglion cells (Figs. 7D–I). VGLUT3-positive cells alsodisplayed NF200 staining, indicating that at least somehuman ganglion cells expressed VGLUT3. We checked thatthe observed labeling did not correspond to displacedamacrine cells by double labeling for VGLUT and GAD.GAD-positive cells in the GCL were VGLUT3-negative,whereas VGLUT3-positive cells were GAD-negative (Figs.

7J–L). Thus, VGLUT3 was expressed solely in ganglion cellsin the human GCL.

3. Discussion

The recently cloned vesicular glutamate transporter BNPI/VGLUT1 and its homologs DNPI/VGLUT2 and VGLUT3promote the uptake of glutamate into synaptic vesicles. Weused immunocytochemical methods to investigate thedistribution of these transporters in rat and human retina.

Fig. 6 – VGLUT2 distribution in human retina. (A) VGLUT2-immunopositive structures were present in cell bodies of INL andGCL and their processes in the IPL. Confocal microscopy of human retinal sections immunolabeled for VGLUT2 (red in B, D, E, G,H, J), parvalbumin (Parv, green in C, D), syntaxin (Synt, green in F, G) and glutamic acid decarboxylase (GAD, green in I, J). (B–D)Many parvalbumin-immunopositive horizontal cells contained VGLUT2 in their cell bodies and processes. (E–G) Manysyntaxin-labeled amacrine cells contained VGLUT2. (H–J) Some GAD-immunopositive amacrine cells (arrow in H–J), but not all(arrowhead in H–J) contained VGLUT2. Note thatmanyVGLUT2-positive structures in the IPLwere also GAD-positive. Scale barsrepresent 10μm in panels A–J.

80 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

VGLUT1 was localized to photoreceptor and bipolar cellterminals, where it probably transports glutamate intosynaptic vesicles to support synaptic transmission. Theseresults are therefore consistent with the function of photo-receptors and bipolar cells in vertical excitatory transmissionwith glutamate release in mammalian and human retina. Bycontrast, VGLUT2 immunoreactivity was restricted to hori-zontal and ganglion cells in rat retina. VGLUT was also found

in amacrine cells, including some GABAergic amacrine cells,in human retina. VGLUT3 immunoreactivity was detected insome ganglion cells in human retina and in glycinergicamacrine cells in rat retina. These locations suggest thatVGLUT2 and VGLUT3 may have functions other thansynaptic glutamate release, contributing to other functionsof glutamate and to GABA neurotransmission, such asstorage.

Fig. 7 – VGLUT3 distribution in rat (A) and human (D, G, J) retina. Rat retinal section immunolabeled for VGLUT3 (A) and glycinetransporter 1 (GLYT1) (B) showing colabeling of a cell body in the INL and of cell processes in the IPL. The merged image (C)demonstrated the complete colocalization of VGLUT3 and GLYT1 in rat retina. Human retinal section immunolabeled forVGLUT3 (D, G), NF200 (E, H) and GAD (K), with DAPI used to stain nuclei (blue in F, I, L). VGLUT3-immunopositive puncta (D, G)were present in NF200-positive ganglion cells in the human retina (E, F, H, I). A GAD-positive displaced amacrine cell (K) in theGCL was not positive for VGLUT3 (J, L). Scale bars represent 10 μm in panels A–L and 2 μm in panels G–I.

81B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

3.1. VGLUT1 localization in the terminals of glutamatergicneurons

VGLUT1 has been reported to be specifically located inglutamatergic neurons (Bellocchio et al., 1998; Takamori etal., 2000; Aihara et al., 2000; Herzog et al., 2001). Photorecep-tors and bipolar cells in the retina release the excitatoryneurotransmitter glutamate; ganglion cells are also glutama-tergic neurons, but their synaptic endings are in the brain(Thoreson and Witkovsky, 1999). Consistent with VGLUT1expression in glutamatergic neurons, VGLUT1 has beendetected in photoreceptors and bipolar cells in rat andmouse retina (Johnson et al., 2003). In the primate retina,

glutamate release by photoreceptors and bipolar cells wassuggested by the immunocytochemical labeling of theseneurons for glutamate (Kalloniatis et al., 1996) and theimmunolabeling of postsynaptic neurons for glutamatereceptor (Haverkamp et al., 2001). We confirm here thatVGLUT1 is localized in photoreceptor and bipolar cellterminals in rat retina and have extended this observationto human retina. All VGLUT1-positive structures in bothplexiform layers were synaptophysin-positive, whereasnot all synaptophysin-positive cells were stained forVGLUT1. These results are consistent with the expression ofVGLUT1 in synaptic vesicles of photoreceptors and bipolarcells, suggesting that glutamate acts as a transmitter in

82 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

photoreceptors and bipolar cells, probably being loaded intosynaptic vesicles by VGLUT1.

3.2. VGLUT2 and VGLUT3 localization in the cell bodies ofglutamatergic neurons

VGLUT2hasbeen reported to be expressed in ganglion cell bodiesin rat retina (Mimura et al., 2002). This finding is consistent withthe reported localization of VGLUT2 in glutamatergic neurons(Fremeau et al., 2001). Our study confirmed that VGLUT2 waslocalized in ganglion cell bodies. Similarly, VGLUT3was detectedin ganglion cell bodies in human retina, contrasting with itsreported presence in amacrine cells in rats (Haverkamp andWässle, 2004; Johnson et al., 2004). The distribution of VGLUT2and VGLUT3 in ganglion cell bodies contrasts strongly with thatof VGLUT1 in photoreceptors and bipolar cells because VGLUT1was absent from cell bodies and restricted to synaptic terminals.The non-synaptic distribution of VGLUT2 and VGLUT3 suggeststhat these transporters may be involved in functions other thanthe synaptic release of glutamate. Glutamate itself is evenlydistributed throughout the cell bodies of retinal glutamatergicneurons (Kalloniatis et al., 1996). Thus, VGLUT2andVGLUT3mayplay a role in glutamate storage in cell bodies in the retina,whereas VGLUT1 seems to play amore specific role in glutamaterelease.

3.3. Localization of VGLUT2 and VGLUT3 innon-glutamatergic neurons

VGLUT3 is specific for glutamate and does not transport otherneurotransmitters, such as GABA or glycine (Fremeau et al.,2002; Gras et al., 2002). However, this transporter has beendetected in serotoninergic and cholinergic neurons (Fremeauet al., 2002; Gras et al., 2002; Schafer et al., 2002; Haverkamp etal., 2004; Haverkamp andWässle, 2004; Somogyi et al., 2004). Inthe rodent retina, VGLUT3 is localized in glycinergic amacrinecells expressing glycine transporter I (GlyT1) and containingglycine (Haverkamp and Wässle, 2004; Johnson et al., 2004).Despite previous reports of VGLUT2 expression in cultured

Table 1 – List of markers, antibody (AB) or lectin used to stain sretina

Marker Rat

Parvalbumin AB Glycinergic AII amacrine cellsCalbindin AB Horizontal cellsDystrophin AB Photoreceptor terminalsSyntaxin AB Amacrine cellsProtein kinase C-α (PKC-α) AB Rod bipolar cellsSynaptophysin AB Synaptic vesicles in OPL and IPLBassoon AB Ribbon synapses in the OPL and

inhibitory amacrine cell synapsesin the IPL

Neurofilament 68 (NF68) AB Ganglion cells and horizontal cellsNeurofilament 200 (NF200) AB Ganglion cellsGlutamic acid decarboxylase

(GAD) ABGABAergic amacrine cells

Tyrosine hydroxylase (TH) AB Dopaminergic amacrine cellsPeanut agglutinin lectin (PNA) Cone pedicles and inner,

outer segment sheathGlycine transporter 1 (GLYT1) AB Glycinergic amacrine cells

dopaminergic neurons (Trudeau, 2004), we did not detect thistransporter in dopaminergic retinal neurons. Instead, wedetected VGLUT2 in retinal neurons like horizontal cells thatare presumably non-glutamatergic. VGLUT2 was detected notonly in horizontal cells in rat and human retina but also inamacrine cells in human retina. Horizontal cells are consid-ered putative GABAergic neurons, although GABA immunos-taining in various animal species, including primates, appearshighly volatile or restricted to specific areas (Nguyen-Legros etal., 1997; Johnson and Vardi, 1998). Consistent with the GABAhypothesis, horizontal cell tips were immunopositive for thevesicular inhibitory amino acid transporter (VIAAT), selectivefor GABA and glycine (Cueva et al., 2002; Jellali et al., 2002).Some amacrine cells immunolabeled for GAD, which synthe-sizes GABA, were also found to express VGLUT2. However, notall GAD-containing neurons contained VGLUT2, consistentwith the absence of VGLUT2 in GABAergic neurons from therat cerebral cortex (Fujiyama et al., 2001). These observationsindicate that retinal GABAergic neurons can contain VGLUT2,suggesting that a given cell may release both glutamate andGABA or that VGLUT2 may store glutamate for GABAsynthesis. GAD does indeed catalyze the synthesis of GABAfrom glutamate in GABAergic neurons (Petroff, 2002). Further-more, postsynaptic GABA neurons can also take up glutamatereleased from presynaptic neurons by glutamate transportersas EAAC-1 does in cerebellar GABAergic Purkinje cells(Rothstein et al., 1994). This glutamate uptake may not onlyaccelerate postsynaptic response recovery (Takahashi et al.,1996) but also be used for GABA release (Sepkuty et al., 2002;Wiessner et al., 2002). It was indeed reported to contribute tothe de novo synthesis of GABA in rat brain (Sepkuty et al., 2002;Wiessner et al., 2002). Horizontal cells in the retina have beenshown to take up glutamate (Schutte and Schlemermeyer,1993). Horizontal and amacrine cells both express the gluta-mate transporter EAAC-1 (Rauen et al., 1996). The presence ofVGLUT2 in these retinal GABAergic neurons suggests that thevesicular transporter loading glutamate into a subcellularcompartment may maintain a high concentration of thissubstrate for GABA synthesis and its subsequent release.

pecific cells or subcellular compartments in rat and human

Human Reference

Horizontal cells and amacrine cells Hamano et al., 1990Cone photoreceptors Rohrenbeck et al., 1987Photoreceptor terminals Ueda et al., 1995Amacrine cells Inoue et al., 1992Rod bipolar cells Wassle et al., 1991Synaptic vesicles in OPL and IPL Brandstatter et al., 1996Ribbon synapses inthe OPL and inhibitoryamacrine cell synapses in the IPL

Dick et al., 2001

Ganglion cells Drager et al., 1984Ganglion cells Drager et al., 1984GABAergic amacrine cells Brandon, 1985

Dopaminergic amacrine cells Nguyen-Legros et al., 1997Cone pedicles and inner,outer segment sheath

Sameshima et al., 1987

Glycinergic amacrine cells Pow and Hendrickson, 1999

83B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

4. Conclusion

Like plasma membrane transporters, vesicular transporterswere thought to be highly specific for the neurotransmitterreleased by a neuron. However, whereas VGLUT1 appears tobe expressed only in glutamatergic neurons in the retina,VGLUT2 and VGLUT3 are also found in putative non-gluta-matergic neurons. This suggests that VGLUTs may not bespecific markers of glutamatergic transmission and that thesemolecules may instead contribute to the synaptic release ofother neurotransmitters, such as GABA.

5. Experimental procedures

5.1. Immunohistochemistry

Human post-mortem retinal tissues were obtained from thehuman tissue bank in Strasbourg in accordance with Frenchlegislation on the use of human tissues for medical andscientific research 24–48 h after the patient death. Thesehuman tissues were fixed by incubation for 15 min at 4 °C in4% paraformaldehyde in phosphate-buffered saline (0.1M PBS,pH 7.4). Rat retinal tissue was obtained from adult Long Evansrats killed by cervical dislocation following anesthesia on dryice. Eyes were enucleated, the anterior segments removed andthe posterior eyecups fixed by incubation in 4% paraformal-dehyde in phosphate-buffered saline for 5min. The retina wasthen dissected from the eyecup, cryoprotected in gradedsucrose solutions (10%, 20%, 30%) and 12 μm vertical sectionswere cut on a cryostat.

Retinal sections were rinsed in PBS, permeabilized byincubation for 5min in 0.1% Triton X-100 in PBS and incubatedwith 0.1% BSA, 1% normal goat serum (NGS) and 0.1% NaN3 inPBS for 15min tominimize nonspecific labeling. Sections wereincubatedwith theprimarymonoclonal or polyclonal antibodyfor 2 to 3 h at room temperature or overnight at 4 °C. We usedthe following antibodies: a polyclonal anti-VGLUT1 antibody(1:4000, Herzog et al., 2001), a polyclonal anti-VGLUT2 anti-body (1:2000, Herzog et al., 2001), a guinea pig anti-VGLUT3antibody (1:5000, Chemicon), a monoclonal anti-dystrophinantibody (1:5, H5A3, generously provided by Dr. Mornet), amonoclonal anti-parvalbumin antibody (1:2000, clone PARV-19, Sigma), amonoclonal anti-calbindin antibody (1:1000, cloneCL-300, Sigma), a monoclonal anti-syntaxin antibody (1:500,clone HPC-1, Sigma), amonoclonal anti-PKC-α antibody (1:200,clone MC5, Sigma), a polyclonal anti-PKC-α antibody (1:500,C20, Santa-Cruz), a monoclonal anti-synaptophysin antibody(1:200, clone svp38, Sigma), a monoclonal anti-Bassoon anti-body (1:100, clone SAP7F407, StressGen), a monoclonal anti-neurofilament 68 antibody (1:1000, clone-NR4, Sigma), apolyclonal anti-neurofilament 200 antibody (1:500, Sigma), amonoclonal anti-GAD antibody (1:100, anti-glutamic aciddecarboxylase 65/67, Sigma), a polyclonal anti-GLYT1 antibody(1:5000, anti-glycine transporter 1, Chemicon) and a monoclo-nal anti-TH antibody (1:100, anti-tyrosine hydroxylase, Che-micon). Sections were washed and incubated for 1 h at 37 °C inthe dark with rabbit anti-mouse or goat anti-rabbit IgGantibodies conjugated to either Alexa TM 594 (red fluores-

cence) or Alexa TM 488 (green fluorescence) (Molecular Probes,EugeneOregon,USA), at a dilution of 1:500. Nucleiwere stainedwith diamidinophenyl indole (DAPI) by incubating the sectionsin the dye solution for 2 min. The sections were then washedfour times in PBS and observed. Non-immune control sectionswere treated similarly, but with the omission of primaryantibodies. Fluorescent labeling was observed with a NikonOptiphot 2 microscope, under epifluorescence illumination(AlexaTM594: excitation filter 510–560, dichroicmirrorDM575,barrier filter BA590; Alexa TM 488: excitation 470–490 nm,dichroicmirror XF22, barrier filter 530DF30). Sectionswere alsoobserved under an inverted Zeiss Axiovert 100 M microscopeequipped with the LSM510 laser scanning confocal module.High-resolution scanning was performed with a 63×, 1.4 oilimmersion objective with 1024 × 1024 or 2048 × 2048 pixelimages (minimum pixel size 0.04 μm × 0.04 μm) in multitrackmode. Excitation/emission wavelengths were 488 nm/505–530 nm and 543 nm/LP585 nm for Alexa TM 488 and Alexa TM594, respectively. All images were obtained from single opticalsections, 0.7 μm sections under blue excitation (488 nm) and0.9 μm sections under green excitation (543 nm). Images wereprocessed with LSM510 (version 2.5) and PhotoShop 5.0 LE(Adobe Systems, San Jose, CA) softwares (Table 1).

Acknowledgments

This work was supported by INSERM, University Louis Pasteur(Strasbourg), University Pierre and Marie Curie (Paris VI),Fédération des Aveugles de France, RETINA-France and theEuropean Economic Community (PRO-AGE-RET: QLK6-2001-00385, PRO-RET: QLK6-2001-00569), France-Regard and Infor-mation Recherche Rétinite Pigmentaire.

R E F E R E N C E S

Aihara, Y., Mashima, H., Onda, H., Hisano, S., Kasuya, H., Hori, T.,Yamada, S., Tomura, H., Yamada, Y., Inoue, I., Kojima, I.,Takeda, J., 2000. Molecular cloning of a novel brain-type Na(+)-dependent inorganic phosphate cotransporter. J. Neurochem.74, 2622–2625.

Bai, L., Xu, H., Collins, J.F., Ghishan, F.K., 2001. Molecular andfunctional analysis of a novel neuronal vesicular glutamatetransporter. J. Biol. Chem. 276, 36764–36769.

Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M.,Edwards, R.H., 1998. The localization of the brain-specificinorganic phosphate transporter suggests a specificpresynaptic role in glutamatergic transmission. J. Neurosci.18, 8648–8659.

Brandon, C., 1985. Retinal GABAneurons: localization in vertebratespecies using an antiserum to rabbit brain glutamatedecarboxylase. Brain Res. 344, 286–295.

Brandstatter, J.H., Lohrke, S., Morgans, C.W., Wässle, H., 1996.Distributions of two homologous synaptic vesicle proteins,synaptoporin and synaptophysin, in the mammalian retina.J. Comp. Neurol. 370, 1–10.

Brandstatter, J.H., Fletcher, E.L., Garner, C.C., Gundelfinger, E.D.,Wässle, H., 1999. Differential expression of the presynapticcytomatrix protein bassoon among ribbon synapses in themammalian retina. Eur. J. Neurosci. 11, 3683–3693.

Cueva, J.G., Haverkamp, S., Reimer, R.J., Edwards, R., Wässle, H.,

84 B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

Brecha, N.C., 2002. Vesicular gamma-aminobutyric acidtransporter expression in amacrine and horizontal cells.J. Comp. Neurol. 445, 227–237.

Drager, U.C., Edwards, D.L., Barnstable, C.J., 1984. Antibodiesagainst filamentous components in discrete cell types of themouse retina. J. Neurosci. 4, 2025–2042.

Dick, O., Hack, I., Altrock, W.D., Garner, C.C., Gundelfinger, E.D.,Brandstatter, J.H., 2001. Localization of the presynapticcytomatrix protein Piccolo at ribbon and conventionalsynapses in the rat retina: comparison with Bassoon. J. Comp.Neurol. 439, 224–234.

Fremeau, R.T., Troyer, M.D., Pahner, I., Nygaard, G.O., Tran, C.H.,Reimer, R.J., Bellocchio, E.E., Fortin, D., Storm-Mathisen, J.,Edwards, R.H., 2001. The expression of vesicular glutamatetransporters defines two classes of excitatory synapse. Neuron31, 247–260.

Fremeau Jr., R.T., Burman, J., Qureshi, T., Tran, C.H., Proctor, J.,Johnson, J., Zhang, H., Sulzer, D., Copenhagen, D.R., Storm-Mathisen, J., Reimer, R.J., Chaudhry, F.A., Edwards, R.H., 2002.The identification of vesicular glutamate transporter 3suggests novel modes of signaling by glutamate. Proc. Natl.Acad. Sci. U. S. A. 99, 14488–14493.

Fujiyama, F., Furuta, T., Kaneko, T., 2001. Immunocytochemicallocalization of candidates for vesicular glutamate transportersin the rat cerebral cortex. J. Comp. Neurol. 435, 379–387.

Gasnier, B., 2000. The loading of neurotransmitters into synapticvesicles. Biochimie 82, 327–337.

Grant, G.B., Dowling, J.E., 1995. A glutamate-activated chloridecurrent in cone-driven ON bipolar cells of the white perchretina. J. Neurosci. 15, 3852–3862.

Gras, C., Herzog, E., Bellenchi, G.C., Bernard, V., Ravassard, P., Pohl,M., Gasnier, B., Giros, B., El Mestikawy, S., 2002. A thirdvesicular glutamate transporter expressed by cholinergic andserotoninergic neurons. J. Neurosci. 22, 5442–5451.

Hamano, K., Kiyama, H., Emson, P.C., Manabe, R., Nakauchi, M.,Tohyama, M., 1990. Localization of two calcium bindingproteins, calbindin (28 kD) and parvalbumin (12 kD), in thevertebrate retina. J. Comp. Neurol. 302, 417–424.

Haverkamp, S., Wässle, H., 2004. Characterization of an amacrinecell type of the mammalian retina immunoreactive forvesicular glutamate transporter 3. J. Comp. Neurol. 468,251–263.

Haverkamp, S., Grunert, U., Wässle, H., 2001. Localization ofkainate receptors at the cone pedicles of the primate retina.J. Comp. Neurol. 436, 471–486.

Haverkamp, S., Muller, U., Zeilhofer, H.U., Harvey, R.J., Wässle, H.,2004. Diversity of glycine receptors in the mouse retina:localization of the alpha2 subunit. J. Comp. Neurol. 477,399–411.

Hayashi, M., Otsuka, M., Morimoto, R., Hirota, S., Yatsushiro, S.,Takeda, J., Yamamoto, A., Moriyama, Y., 2001. Differentiation-associated Na+-dependent inorganic phosphate cotransporter(DNPI) is a vesicular glutamate transporter in endocrineglutamatergic systems. J. Biol. Chem. 276, 43400–43406.

Herzog, E., Bellenchi, G.C., Gras, C., Bernard, V., Ravassard, P., Bedet,C., Gasnier, B., Giros, B., El Mestikawy, S., 2001. The existence ofa second vesicular glutamate transporter specifiessubpopulations of glutamatergic neurons. J. Neurosci. 21, RC181.

Inoue, A., Obata, K., Akagawa, K., 1992. Cloning and sequenceanalysis of cDNA for a neuronal cell membrane antigen, HPC-1.J. Biol. Chem. 267, 10613–10619.

Jellali, A., Stussi-Garaud, C., Gasnier, B., Rendon, A., Sahel, J.A.,Dreyfus, H., Picaud, S., 2002. Cellular localization of thevesicular inhibitory amino acid transporter in the mouse andhuman retina. J. Comp. Neurol. 449, 76–87.

Johnson, M.A., Vardi, N., 1998. Regional differences in GABA andGAD immunoreactivity in rabbit horizontal cells. Vis. Neurosci.15, 743–753.

Johnson, J., Tian, N., Caywood, M.S., Reimer, R.J., Edwards, R.H.,

Copenhagen, D.R., 2003. Vesicular neurotransmittertransporter expression in developing postnatal rodent retina:GABA and glycine precede glutamate. J. Neurosci. 23, 518–529.

Johnson, J., Sherry, D.M., Liu, X., Fremeau Jr., R.T., Seal, R.P.,Edwards, R.H., Copenhagen, D.R., 2004. Vesicular glutamatetransporter 3 expression identifies glutamatergic amacrinecells in the rodent retina. J. Comp. Neurol. 477, 386–398.

Kalloniatis, M., Marc, R.E., Murry, R.F., 1996. Amino acid signaturesin the primate retina. J. Neurosci. 16, 6807–6829.

Lee, R.Y., Sawin, E.R., Chalfie, M., Horvitz, H.R., Avery, L., 1999.EAT-4, a homolog of a mammalian sodium-dependentinorganic phosphate cotransporter, is necessary forglutamatergic neurotransmission in Caenorhabditis elegans.J. Neurosci. 19, 159–167.

Mimura, Y., Mogi, K., Kawano, M., Fukui, Y., Takeda, J., Nogami, H.,Hisano, S., 2002. Differential expression of two distinctvesicular glutamate transporters in the rat retina. NeuroReport13, 1925–1928.

Nguyen-Legros, J., Versaux-Botteri, C., Savy, C., 1997.Dopaminergic and GABAergic retinal cell populations inmammals. Microsc. Res. Tech. 36, 26–42.

Ni, B., Rosteck Jr., P.R., Nadi, N.S., Paul, S.M., 1994. Cloning andexpression of a cDNA encoding a brain-specific Na(+)-dependent inorganic phosphate cotransporter. Proc. Natl.Acad. Sci. U. S. A. 91, 5607–5611.

Petroff, O.A., 2002. GABA and glutamate in the human brain.Neuroscientist 8, 562–573.

Picaud, S., Larsson, H.P., Wellis, D.P., Lecar, H., Werblin, F., 1995.Cone photoreceptors respond to their own glutamate release inthe tiger salamander. Proc. Natl. Acad. Sci. U. S. A. 92,9417–9421.

Pow, D.V., Hendrickson, A.E., 1999. Distribution of the glycinetransporter glyt-1 in mammalian and nonmammalian retinae.Vis. Neurosci. 16, 231–239.

Rauen, T., Rothstein, J.D., Wässle, H., 1996. Differential expressionof three glutamate transporter subtypes in the rat retina. CellTissue Res. 286, 325–336.

Rohrenbeck, J., Wassle, H., Heizmann, C.W., 1987.Immunocytochemical labelling of horizontal cells inmammalian retina using antibodies against calcium-bindingproteins. Neurosci. Lett. 77, 255–260.

Rothstein, J.D., Martin, L., Levey, A.I., Dykes-Hoberg, M., Jin, L., Wu,D., Nash, N., Kuncl, R.W., 1994. Localization of neuronal andglial glutamate transporters. Neuron 13, 713–725.

Sameshima, M., Uehara, F., Ohba, N., 1987. Specialization of theinterphotoreceptor matrices around cone and rodphotoreceptor cells in the monkey retina, as revealed by lectincytochemistry. Exp. Eye Res. 45, 845–863.

Schafer, M.K., Varoqui, H., Defamie, N., Weihe, E., Erickson, J.D.,2002. Molecular cloning and functional identification of mousevesicular glutamate transporter 3 and its expression in subsetsof novel excitatory neurons. J. Biol. Chem. 277, 50734–50748.

Schutte, M., Schlemermeyer, E., 1993. Depolarization elicits, whilehyperpolarization blocks uptake of endogenous glutamate byretinal horizontal cells of the turtle. Cell Tissue Res. 274,553–558.

Sepkuty, J.P., Cohen, A.S., Eccles, C., Rafiq, A., Behar, K., Ganel, R.,Coulter, D.A., Rothstein, J.D., 2002. A neuronal glutamatetransporter contributes to neurotransmitter GABA synthesisand epilepsy. J. Neurosci. 22, 6372–6379.

Somogyi, J., Baude, A., Omori, Y., Shimizu, H., Mestikawy, S.E.,Fukaya, M., Shigemoto, R., Watanabe, M., Somogyi, P., 2004.GABAergic basket cells expressing cholecystokinin containvesicular glutamate transporter type 3 (VGLUT3) in theirsynaptic terminals in hippocampus and isocortex of the rat.Eur. J. Neurosci. 19, 552–569.

Takahashi, M., Sarantis, M., Attwell, D., 1996. Postsynapticglutamate uptake in rat cerebellar Purkinje cells. J. Physiol. 497(Pt. 2), 523–530.

85B R A I N R E S E A R C H 1 0 8 2 ( 2 0 0 6 ) 7 3 – 8 5

Takamori, S., Rhee, J.S., Rosenmund, C., Jahn, R., 2000.Identification of a vesicular glutamate transporter thatdefines a glutamatergic phenotype in neurons. Nature 407,189–194.

Thoreson, W.B., Witkovsky, P., 1999. Glutamate receptors andcircuits in the vertebrate retina. Prog. Retinal Eye Res. 18,765–810.

Trudeau, L.E., 2004. Glutamate co-transmission as an emergingconcept in monoamine neuron function. J. PsychiatryNeurosci. 29, 296–310.

Ueda, H., Tsukahara, S., Kobayashi, T., Ohno, S., 1995.Immunocytochemical study of dystrophin-related protein inthe rat retina. Ophthalmic Res. 27, 219–226.

Ueda, H., Baba, T., Terada, N., Kato, Y., Tsukahara, S., Ohno, S.,

1997a. Dystrophin in rod spherules; submembranous denseregions facing bipolar cell processes. Histochem. Cell Biol. 108,243–248.

Ueda, H., Kato, Y., Baba, T., Terada, N., Fujii, Y., Tsukahara, S.,Ohno, S., 1997b. Immunocytochemical study of dystrophinlocalization in cone cells of mouse retinas. Invest. Ophthalmol.Visual Sci. 38, 1627–1630.

Wassle, H., Yamashita, M., Greferath, U., Grunert, U., Muller, F.,1991. The rod bipolar cell of the mammalian retina. Vis.Neurosci. 7, 99–112.

Wiessner, M., Fletcher, E.L., Fischer, F., Rauen, T., 2002.Localization and possible function of the glutamatetransporter, EAAC1, in the rat retina. Cell Tissue Res. 310,31–40.