We examined the distribution of the group I metabotropic glutamate receptors, mGluR1α and mGluR5a, in the adult rat retina and during postnatal development using receptor-specific antisera. In contrast to the restricted localization of group II and group III mGluRs to either the outer plexiform layer (OPL) or the inner plexiform layer (IPL), group I mGluRs are present in both synaptic layers in the rat retina. Double-labeling experiments and electron microscopy showed that in the OPL the two receptors are localized on the dendritic tips of bipolar cells postsynaptic to photoreceptor terminals. In the IPL the two mGluRs are localized on amacrine cell processes postsynaptic to bipolar cell terminals. These results suggest that group I mGluRs are involved in synaptic processing in both plexiform layers and in both the scotopic and photopic pathways in the rat retina. We propose that mGluR1α and mGluR5a play an important modulatory role in the responses of retinal neurons to inhibitory and excitatory neurotransmitters.
- bipolar cells
- amacrine cells
- outer plexiform layer
- inner plexiform layer
- rat retina
- retinal development
In the mammalian retina glutamate is the major excitatory neurotransmitter used in visual information processing. The flow of visual information in the vertical direction from the photoreceptors to bipolar cells and horizontal cells in the OPL and from bipolar cells to amacrine cells and ganglion cells in the IPL is mediated via glutamate (for review, see Massey, 1990). This flow of information is modulated by the laterally connecting inhibitory interneurons, horizontal cells and amacrine cells, that preferentially use GABA or glycine as their neurotransmitters (for review, see Brecha, 1983; Yazulla, 1986; Massey and Redburn, 1987; Marc, 1989).
Glutamate acts on two different classes of receptors, ionotropic and metabotropic glutamate receptors. Ionotropic glutamate receptors mediate fast excitatory synaptic transmission (for review, see Monaghan et al., 1989; Seeburg, 1993; Hollmann and Heinemann, 1994), whereas metabotropic glutamate receptors (mGluRs) are G-protein-coupled proteins that bind the ligand, influence a variety of intracellular second messenger systems, and modulate neuronal activity (for review, see Pin and Duvoisin, 1995). Glutamate can act in an excitatory or inhibitory manner via the action of mGluRs.
So far eight different members of the mGluR family have been cloned. Based on sequence homologies, pharmacological properties, and second messenger systems, mGluRs can be subdivided into three groups (for review, see Hollmann and Heinemann, 1994; Nakanishi, 1994; Pin and Duvoisin, 1995). MGluR2, 3 (group II) and mGluR4, 6, 7, 8 (group III) are functionally coupled to adenylyl cyclase and inhibit the formation of cAMP in artificial expression systems (Tanabe et al., 1992, 1993;Nakajima et al., 1993; Nakanishi, 1994; Okamoto et al., 1994; Saugstad et al., 1994; Duvoisin et al., 1995). MGluR1 and mGluR5 (group I), are activated most strongly by quisqualate. They are functionally coupled to stimulation of phospholipase C, increased synthesis of inositol 1,4,5-trisphosphate (IP3), and Ca2+ release from internal stores (Masu et al., 1991; Abe et al., 1992; Aramori and Nakanishi, 1992).
The presence of mGluR1 and mGluR5 in the rat retina was shown within situ hybridization at the mRNA level (Hartveit et al., 1995) and for mGluR1α with immunocytochemistry at the protein level (Peng et al., 1995). Several components of a possible pathway involving glutamate and IP3 signaling, e.g., phosphatidylinositol diphosphate (PIP2), IP3, and IP3receptors, also were shown to be present in the rat retina (Anderson et al., 1985; Das et al., 1986; Milani et al., 1990; Peng et al., 1991). However, knowledge on the exact distribution of group I mGluRs in the retina with respect to cellular and subcellular localization and their possible function in retinal synaptic circuitry is still lacking.
We have examined light and electron microscopically the cellular and subcellular distribution of mGluR1α and mGluR5a in the rat retina using receptor-specific antisera. We show that group I mGluRs are used in synaptic processing in both synaptic layers in the rat retina and suggest that these receptors play an important role in the modulation of the responses of retinal neurons to inhibitory and excitatory neurotransmitters.
MATERIALS AND METHODS
Antisera against mGluR1α and mGluR5a. The affinity-purified polyclonal antiserum against mGluR1α was obtained from Chemicon (AB1551, Chemicon, Temecula, CA) and was generated in rabbit against a C-terminal peptide of rat mGluR1α (PNVTYASVILRDYKQSSSTL) conjugated to keyhole limpet hemocyanin. The affinity-purified polyclonal antiserum against mGluR5a also was raised in rabbit against a peptide corresponding to the C-terminal amino acid sequence of rat mGluR5a (PSSPKYDTLIIRDYTQSSSSL) conjugated to ovalbumin (Vidnyánszky et al., 1994).
Animals and tissue preparation. Retinas of adult albino rats, 6–8 weeks of age, and of different postnatal stages, 1–30 d old, were investigated. For the developmental studies, only retinas from the same litter mates and only retinal pieces with the same eccentricity were compared. The rats were anesthetized deeply with halothane and decapitated. A detailed description of the preparation of the retinal tissue for light and electron microscopic immunocytochemistry is given in Brandstätter et al. (1996).
Light microscopic immunocytochemistry. In addition to the antisera against mGluR1α (0.2 μg/ml) and against mGluR5a (1 μg/ml), we used the monoclonal antibody MC-3A against PKCα (1:100; Seikagaku, Tokyo, Japan). Immunocytochemical labeling was performed by the indirect fluorescence method. To characterize the cellular distribution of mGluR5a in more detail, we combined the mAb MC-3A against PKCα, known to stain rod bipolar cells (Greferath et al., 1990), in double-labeling experiments with the specific anti-peptide antiserum against mGluR5a. The binding sites of the primary antibodies were revealed by secondary antibodies: goat anti-mouse or goat anti-rabbit IgG coupled either to carboxymethylindocyanine (Cy3, red fluorescence; Dianova, Hamburg, Germany) diluted 1:1000 or fluorescein isothiocyanate (FITC, green fluorescence; Sigma-Aldrich, Deisenhofen, Germany) diluted 1:100 or to Texas Red (red fluorescence; Amersham, Braunschweig, Germany) diluted 1:100. In the double-labeling experiments, sections were incubated in a mixture of the primary antibodies, followed by a mixture of the secondary antibodies.
Preembedding immunoelectron microscopy. After blocking, vibratome sections (50 μm thick) were incubated in primary antisera against mGluR5a and mGluR1α for 4 d at 4°C. The primary antisera were used at the same concentration and diluted in the same medium, but without Triton X-100, as used for light microscopy.
Detection of the immunostaining and microscopic analysis were performed as previously described (Brandstätter et al., 1996).
Characterization of the antisera against mGluR1α and mGluR5a.The antiserum against mGluR5a has been fully characterized and used previously for light and electron microscopic distribution studies (Vidnyánszky et al., 1994). The specificity of the antiserum against mGluR1α was assessed by immunoblotting of rat retina membrane proteins (Fig. 1 A,B). Albino rats were deeply anesthetized and decapitated. The retinas were dissected and homogenized in lysis buffer containing 4 mm HEPES, 220 mm d(+)-mannose, 70 mm sucrose, 1 mg/ml benzamidine hydrochloride, 0.5 mg/ml aprotinin (Merck, Darmstadt, Germany), and 0.25 mg/ml benzethonium chloride (Sigma-Aldrich) at pH 7.5, centrifuged for 3 min at 1000 × g and 4°C. Subsequently, the supernatant was centrifuged for 15 min at 15,000 × g. The pellet was resuspended in lysis buffer to obtain a crude retinal membrane protein fraction. After denaturation with SDS and 2-mercaptoethanol, crude retinal membrane proteins (80 μg/lane) and biotinylated SDS molecular weight markers were electrophoresed on 7.5% SDS-polyacrylamide gels. Proteins were transferred onto cationized nylon membranes by standard Western blotting technique. After incubation with blocking buffer [5% (v/v) normal goat serum (NGS) and 0.05% (w/v) Tween 20 in PBS (0.01 m), pH 7.4] for 1 hr at room temperature, blots were incubated with primary (0.1 μg/ml) and secondary antibodies for 1 hr each in blocking buffer. Binding of the antiserum against mGluR1α to polypeptides was detected by goat anti-rabbit IgG antibodies coupled to alkaline phosphatase (diluted 1:100; Dianova) with 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium (Boehringer Mannheim Biochemica, Mannheim, Germany) as substrates.
The antiserum against mGluR1α detected a single protein band with a molecular weight of ∼140 kDa in membrane preparations of rat retina (Fig. 1 A). This is in agreement with the published molecular weight of mGluR1α deduced from its cDNA sequence (Masu et al., 1991). Preincubation of the antiserum against mGluR1α with a 10-fold excess (w/w) of the antigenic peptide led to no labeling in the subsequent immunoblot analysis (Fig. 1 B).
For staining of retina sections, controls were made by preincubating the antisera against mGluR1α and mGluR5a with a 10-fold excess of their respective antigenic peptides for 1 hr at room temperature, resulting in a complete loss of specific staining (Figs. 1 C,3 B). In double-labeling experiments, controls were prepared by omitting one of the two primary antibodies during the incubation, and in this case, only the immunoreactivity for the remaining primary antibody was detected.
In addition, we tested the specificity of the antiserum against mGluR5a and mGluR1α by staining sections of adult rat brain. Unfixed rat brains were sectioned on a cryostat at 18 μm thickness, collected on gelatin-coated slides, postfixed for 5 min in 4% (w/v) paraformaldehyde in phosphate buffer, and immunostained as described for the retina sections. Immunolabeling was detected with a biotinylated goat anti-rabbit IgG (1:100; Sigma-Aldrich) and extravidin-coupled peroxidase (1:100; Sigma-Aldrich). After preincubation with 0.05% (w/v) 3, 3′-diaminobenzidine (DAB; Sigma-Aldrich), the sections were reacted by adding 0.01% (v/v) H2O2. The staining patterns obtained in rat brain (Fig. 2) were compared with published data of the distribution of mGluR1α and mGluR5a to ensure the specificity to the antibody reactivity. As has been shown for mGluR1α (Baude et al., 1993) and mGluR5 (Romano et al., 1995), mGluR1α immunoreactivity in the cerebellum was strong in the molecular layer and the Purkinje cell layer and weak in the granule cell layer (Fig. 2 A), whereas mGluR5a immunoreactivity was weak in the molecular layer and Purkinje cell layer and stronger in the granule cell layer (Fig.2 B).
Distribution of mGluR5a in the adult rat retina
In the adult retina, mGluR5a immunoreactivity showed a distinct labeling pattern in the IPL, the OPL, and in the inner nuclear layer (INL) (Fig. 3 A). Patches of punctate immunoreactivity were seen throughout the IPL with reduced staining in the outer half of the IPL (OFF sublamina) and intense staining in the inner half of the IPL (ON sublamina). A narrow immunoreactive band of processes in the outermost part of the IPL was labeled more strongly. Intense immunoreactivity also was detected in the OPL in putative dendrites of bipolar cells (Fig. 3 A). In the INL, somata and occasionally axons of putative bipolar cells were diffusely labeled (Fig. 3 A). Punctate distribution of the mGluR5a receptor staining at the light microscopic level indicated synaptic localization, as was shown for several other types of neurotransmitter receptors in the retina (Pourcho and Owczarzak, 1991; Yazulla and Studholme, 1991; Grünert and Wässle, 1993; Hartveit et al., 1994; Brandstätter et al., 1996). The diffuse labeling on putative bipolar cell somata in the INL, however, also showed extrasynaptic localization of mGluR5a immunolabel. In control experiments, preadsorption of the anti-mGluR5a antiserum with the respective antigenic peptide before applying to retina sections resulted in complete absence of specific staining (Fig.3 B).
To determine whether bipolar cell somata in the INL and their dendrites in the OPL are immunolabeled for mGluR5a, we performed double-labeling experiments with the antiserum against mGluR5a and an antibody against an isoform of protein kinase C (PKC), shown to stain rod bipolar cells in the rat retina (Greferath et al., 1990) (Fig. 4). These experiments clearly demonstrated that at least rod bipolar cell somata in the INL and their dendrites in the OPL were immunolabeled for mGluR5a (Fig. 4). In the IPL there was no detectable colocalization between the mGluR5a staining and the labeling of rod bipolar cell terminals by the anti-PKC antibody (Fig. 4). The micrographs in Figure4 were printed as mirror images, and the symmetry of staining that can be detected along the midline (arrows in Fig. 4) shows colocalization.
Subcellular distribution of mGluR5a
Immunoreactivity for mGluR5a was found intracellularly because of the epitope specificity of the anti-mGluR5a antiserum.
MGluR5a is localized to bipolar cell dendrites in the OPL
In the OPL, dendrites of bipolar cells postsynaptic to photoreceptor terminals were mGluR5a-immunoreactive (Fig.5). In the majority of cases the immunolabeled bipolar cell dendrites were detected postsynaptic to the terminals of rod photoreceptors (rod spherules), therefore belonging to rod bipolar cells (Dowling and Boycott, 1966) (Fig. 5 A,B). This corroborates the light microscopic finding of mGluR5a-immunoreactive rod bipolar cell somata and dendrites. Horizontal cells, the two lateral postsynaptic elements at the rod photoreceptor synapse in the OPL (Dowling and Boycott, 1966), were never found to be mGluR5a-immunoreactive (Fig. 5 A,B). Occasionally, we also found the dendrite of a cone bipolar cell stained for mGluR5a that made contact at the terminal of a cone photoreceptor (cone pedicle) but was not associated directly with the synaptic complex (Fig.5 C).
MGluR5a is localized to amacrine cell processes in the IPL
In the IPL, mGluR5a immunoreactivity was found on amacrine cell processes postsynaptic to bipolar cell ribbon synapses, consistent with the light microscopic finding of punctate receptor staining in the IPL (Fig. 3). We detected mGluR5a immunoreactivity on amacrine cell processes postsynaptic to OFF-cone bipolar cell terminals (Fig.5 D), ON-cone bipolar cell terminals (Fig. 5 E), and rod bipolar cell terminals (Fig. 5 F). Whereas at the cone bipolar cell synapses the two postsynaptic elements are preferentially a process of an amacrine cell and a dendrite of a ganglion cell (Dowling and Boycott, 1966), at the rod bipolar cell synapse both postsynaptic elements belong to amacrine cells (Famiglietti and Kolb, 1975; Chun et al., 1993). At the bipolar cell synapses only one of the postsynaptic processes was labeled.
Distribution of mGluR1α in the adult rat retina
MGluR1α, the second member of group I mGluRs, like mGluR5a was expressed in both synaptic layers, the OPL and the IPL, of the adult rat retina (Fig. 6 A). In the IPL, staining for mGluR1α revealed several distinct immunoreactive bands with a mixture of diffuse and punctate staining separated by bands of reduced or no immunoreactivity (Fig. 6 A). Whereas the outer half of the IPL (OFF sublamina) was subdivided into three smaller mGluR1α-immunoreactive bands, the complete inner half of the IPL (ON sublamina) showed a more or less homogenous distribution of mGluR1α immunoreactivity (Fig. 6 A). Clear labeling of a narrow band of mGluR1α-immunoreactive processes also was found in the OPL. In the INL, occasionally labeled somata of putative bipolar cells were detected (Fig. 6 A).
Subcellular distribution of mGluR1α
As with mGluR5a, the reaction product of the mGluR1α immunostaining was found intracellularly because of the epitope specificity of the antiserum.
MGluR1α is localized to processes postsynaptic to ribbon synapses in the OPL and IPL
Like the labeling for mGluR5a, mGluR1α immunoreactivity in the OPL was found in rod bipolar cell dendrites postsynaptic at ribbon synapses of rod photoreceptor cells (Fig.7 A). Only rarely did we find a cone bipolar cell dendrite, postsynaptic to the terminal of a cone photoreceptor cell, stained for mGluR1α (data not shown). Again, the two lateral horizontal cell processes at the photoreceptor synapse were never found to be mGluR1α-immunoreactive (Fig. 7 A).
In the IPL amacrine cell processes postsynaptic to OFF-cone bipolar cell terminals (Fig. 7 B), ON-cone bipolar cell terminals (Fig. 7 C) and rod bipolar cell terminals (Fig.7 D) were labeled for mGluR1α. Like mGluR5a immunoreactivity, mGluR1α staining was present in only one of the two postsynaptic amacrine cell processes at the rod bipolar cell ribbon synapse. From our material we are not able to say whether ganglion cell dendrites also were stained for mGluR1α.
Postnatal development of mGluR5a and mGluR1α expression
During postnatal development, mGluR5a-immunoreactive processes of neurons stratifying in the IPL were detected first at approximately postnatal day 3 (P3) (Fig. 8 B). At around P5 the homogenous mGluR5a labeling pattern in the IPL changed to a more stratified appearance, with stronger staining in the inner part and weaker staining in the outer part of the IPL (Fig. 8 C). In addition, the mGluR5a staining was no longer restricted to the IPL but appeared in the somata of putative bipolar cells in the INL and their dendrites in the OPL (Fig. 8 C). The adult labeling pattern showing a pronounced stratification of mGluR5a immunostaining in the IPL and clear labeling of bipolar cell somata in the INL and of their dendrites in the OPL was observed already at approximately P8 (Fig. 8 D).
In contrast to mGluR5a, low mGluR1α immunoreactivity was present already at birth (P1) (Fig. 9 A). The intensity of mGluR1α immunoreactivity and the level of stratification of the staining in the IPL increased during the first and second postnatal week, P7–P12 (Fig. 9 B,C), and reached the adult staining pattern at approximately P19 (Fig. 9 D). Like mGluR5a, the antiserum against mGluR1α also stained bipolar cell somata in the INL and their dendrites in the OPL, but to a lesser extent (Fig. 9 B,C). During later postnatal development, P12–P19, this staining increasingly became restricted to the dendrites of bipolar cells in the OPL (Fig. 9 D).
MGluR1α and mGluR5a are involved in synaptic processing in both synaptic layers in the rat retina
In contrast to the restricted localization of group II and group III mGluRs to either the IPL or the OPL in the rat retina (Nomura et al., 1994; Brandstätter et al., 1996; Koulen et al., 1996), group I mGluRs are present in both synaptic layers. The localization of mGluR1α and mGluR5a in the IPL and OPL of the rat retina coincides with the distribution of presumed components of second messenger systems involved in group I mGluR signaling. IP3 receptors were found on synaptic processes of amacrine cells as well as in the distal parts of bipolar cell somata (Peng et al., 1991). Other elements of group I mGluR second messenger systems such as IP3 and its precursors also were localized in the retina (Anderson et al., 1985; Das et al., 1986; Milani et al., 1990). Group I mGluRs are activated most strongly by l-quisqualate, and their activation leads to an increase in IP3 synthesis and mobilization of intracellular Ca2+ (for review, see Pin and Duvoisin, 1995). Osborne (1990) found a receptor in the rabbit retina that was activated by quisqualate and stimulated the accumulation of inositol phosphates. Other excitatory amino acid agonists that influence inositol phospholipid metabolism had no effect, and Osborne (1990) suggested a specific quisqualate receptor in the retina.
The postsynaptic localization of mGluR1α and mGluR5a on dendrites of bipolar cells in the OPL and on the postsynaptic partners of bipolar cells in the IPL suggests that group I mGluRs take part in synaptic processing in both synaptic layers and are involved in the modulation of synaptic signals in both the photopic and scotopic pathways in the rat retina.
Although activation of mGluR1α and mGluR5a leads to the same effects in neurons, both receptors were present in dendrites of bipolar cells in the OPL and processes of amacrine cells in the IPL. Examination of the different expression patterns of these two mGluRs in the IPL suggests that different types of amacrine cells express the two mGluRs. However, the bipolar cell dendrites in the OPL seem to label for both mGluRs. Thus, the question arises why a cell would express receptors with the same pharmacological profile at the same site. Only recently it has been shown by Kawabata and colleagues (1996) that, in cells transfected with mGluR1α, glutamate caused a single-peaked mobilization of intracellular Ca2+, whereas in mGluR5a-transfected cells, glutamate elicited Ca2+oscillations caused by differences in phosphorylation sites of the two receptors. These differences could have an impact on intracellular signaling mechanisms in glutamate transmission.
Involvement of group I mGluRs in synaptic processing in the OPL
MGluR1α and mGluR5a were found in the OPL of the rat retina preferentially localized to the dendritic tips of rod bipolar cells postsynaptic to the terminals of rod photoreceptor cells. Only on rare occasions staining was found in dendrites of cone bipolar cells postsynaptic to the terminals of cone photoreceptor cells. Rod bipolar cells are ON bipolar cells that are active under scotopic conditions. One of their functions is the transmission of the light ON signal from the rod photoreceptor cells to the ON ganglion cells via the AII amacrine cells.
To date, mGluR6 is the sole mGluR found on the dendrites of ON bipolar cells in the OPL (Nomura et al., 1994). The function of mGluR6, sensitive to l-2-amino-4-phosphonobutyrate (l-AP4), is the transmission of the light signal from the photoreceptor cells to the ON bipolar cells; the photoreceptors are hyperpolarized by light, and this hyperpolarization is converted via the action of mGluR6 to a depolarization of the ON bipolar cells that convey the light ON signal to the ON ganglion cells [Masu et al. (1995); for review, see Nakanishi (1995)].
Because of the physiology of group I mGluRs, it is unlikely that they are involved in the direct transmission of synaptic signals in the OPL. Our immunocytochemical data showing group I mGluRs postsynaptic in the dendrites of bipolar cells in the OPL could explain, however, the modulatory action of mGluR agonists on the GABA responses of bipolar cells reported by Feigenspan and Bormann (1994). They showed that extracellular application of 1-amino-cyclopentane-1,3-dicarboxylate (ACPD), an agonist for group I mGluRs (for review, see Pin and Duvoisin, 1995), to retinal bipolar cells in culture stimulated the decline of GABAC responses. They suggested that glutamate released from photoreceptor cells in the OPL could act on mGluRs in bipolar cells, which in turn could decrease the response to inhibitory GABAergic inputs from horizontal or amacrine cells. Stimulation of phospholipase C, increased accumulation of IP3 and diacyl glycerol (DAG), and activation of PKC could lead to phosphorylation of the GABAC receptor channels. Group I mGluRs in dendrites of bipolar cells (present study), PKC in dendrites of bipolar cells (Greferath et al., 1990), and IP3 and GABACreceptors also found in the OPL in the rat retina (Peng et al., 1991;Enz et al., 1996) could be the anatomical substrates for such an action. For Purkinje cells it has been shown that the action of IP3 on the release of Ca2+ from intracellular stores remains restricted locally to a few micrometers within their dendrites and also is highly regulated temporally (Wang and Augustine, 1995).
Involvement of group I mGluRs in synaptic processing in the IPL
Both mGluR1α and mGluR5a were found in amacrine cell processes postsynaptic to bipolar cell ribbon synapses in the IPL. Amacrine cells do not express GABAC receptors in the rat retina (Enz et al., 1995), and therefore it is very unlikely that group I mGluRs in the IPL should have the same mode of action as that suggested for mGluR1α and mGluR5a in bipolar cell dendrites in the OPL. Most amacrine cells are inhibitory interneurons releasing either GABA or glycine (for review, see Wässle and Boycott, 1991). For group II and group III mGluRs present in amacrine cell processes in the IPL, it was suggested that they could influence the release of inhibitory neurotransmitter at reciprocal synapses back onto bipolar cells. Such a mechanism could modulate the release of glutamate from bipolar cells (Brandstätter et al., 1996; Koulen et al., 1996) indirectly. Because of the pharmacology of group I mGluRs, characterized by stimulation of PLC, increase of inositol phosphates, and release of Ca2+ from intracellular stores (for review, see Pin and Duvoisin, 1995), we think that a mechanism like that suggested for the action of group II and group III mGluRs in the IPL does not hold true for group I mGluRs. We suggest that group I mGluRs located in the processes of amacrine cells in the IPL could have a general effect on the activity of the cells, which in turn would influence their release probability for neurotransmitter.
Postnatal retinal development of group I mGluR expression
During postnatal development the expression patterns of mGluR1α and mGluR5a differed temporally and spatially. Although mGluR1α was expressed before mGluR5a, both receptors were first present in the IPL, where the earliest synapses form during retinal development (Horsburgh and Sefton, 1987). Only later in development both mGluRs were expressed in neuronal somata in the INL and their processes in the OPL, where synapse formation occurs later than in the IPL. In contrast to different brain regions, where a downregulation was observed for mGluR5 (Romano et al., 1996), mGluR1α and mGluR5a were upregulated during postnatal retinal development. It has been shown that intracellular levels of free Ca2+ influence neurite elongation and growth cone movement during development (Cohan et al., 1987) and that group I mGluRs are important for neuronal survival during development (Nicoletti et al., 1996; Pizzi et al., 1996). Because of the early appearance of group I mGluRs before synaptogenesis in the rat retina, they could influence synaptic differentiation during postnatal development and also might play a role in the final consolidation of synaptic connections in the retina (Redburn and Rowe-Rendleman, 1996).
This study was supported by Grant SFB 269/B4 from the Deutsche Forschungsgemeinschaft. We thank A. Leihkauf, G.-S. Nam, and W. Hofer for excellent technical assistance and Dr. E. Fletcher for critically reading and improving this manuscript.
Correspondence should be addressed to Dr. Johann H. Brandstätter, Max-Planck-Institut für Hirnforschung, Abteilung für Neuroanatomie, Deutschordenstrasse 46, D-60528 Frankfurt am Main, Germany.