- Email: [email protected]
Developmental expression of excitatory amino acid transporter 5: a photoreceptor and bipolar cell glutamate transporter in rat retina
Neuroscience Letters 280 (2000) 21±24 www.elsevier.com/locate/neulet
Developmental expression of excitatory amino acid transporter 5: a photoreceptor and bipolar cell glutamate transporter in rat retina David V. Pow a,*, Nigel L. Barnett b a
Department of Anatomical Sciences, University of Queensland, Brisbane, Queensland 4306, Australia b VTHRC, University of Queensland, Brisbane, Queensland 4306, Australia Received 5 October 1999; received in revised form 3 November 1999; accepted 8 December 1999
Abstract Excitatory amino acid transporter 5 (EAAT5) is a retina-speci®c glutamate transporter which has an associated chloride conductance. Thus it is comparable in its functional properties to the glutamate transport systems previously described in photoreceptors and some bipolar cells. We have raised antibodies to the carboxyl- and amino-terminal regions of EAAT5. Labeling for both of these antisera was developmentally regulated: weak labeling appeared in photoreceptors around P7; by P10 strong labeling was present in photoreceptors and by P21 a population of bipolar elements were also weakly labeled. In adult retinae both antisera heavily immunolabeled all photoreceptors as well as a heterogeneous population of bipolar cell somata and their proximal axonal processes: synaptic terminals of these cells were also labeled after partial proteolytic digestion of the tissues. The positions and morphology of these terminals suggests that they are the terminals of both rod and cone rod bipolar cells. We conclude that in rat retina, EAAT5 is a photoreceptor and bipolar cell glutamate transporter. q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Bipolar cell; Glutamate; Excitatory amino acid transporter 5; Photoreceptor; Retina; Transporter
The retina expresses a variety of glutamate transporters including GLAST which is present in MuÈller cells [6,8,15,16], GLT-1 which is present in cone bipolar cells and cone photoreceptors [14], EAAC1 which is present in horizontal cells, some amacrine cells and ganglion cells [16] and excitatory amino acid transporter 5 (EAAT5). EAAT5 is a retina-speci®c glutamate transporter [1]. We have recently shown that in primate retina EAAT 5 is selectively associated with rod photoreceptors [9]. In this study we have investigated whether rat photoreceptors or other retinal neurons express EAAT5. Since there is always the possibility that an antibody to one end of a protein alone might recognize other related proteins this study used antisera to both the carboxyl and amino termini of EAAT5, thereby reducing the possibility of erroneously labeling a cell population which expresses a related molecular species other than EAAT5. Expression of EAAT5 was examined at a series of landmark developmental time points such as the day of birth, postnatal day 7 when photoreceptor synapses ®rst appear [19], and the day of eye opening (postnatal day 14). P21 and adult retinae were also examined. For each age, retinae * Corresponding author. Fax: 161-7-3365-1299. E-mail address: [email protected] (D.V. Pow)
from 5 rats were examined. The developmentally advanced central third of the retina was preferentially examined, and the descriptions that ensue refer to this region. However the peripheral areas of each retina were also examined. Our working hypothesis was that onset of EAAT5 expression might be associated with synaptogenesis and thus onset of glutamate release from photoreceptors. All experiments were carried out in accordance with the ethical guidelines of the NHMRC (Australia). Synthetic peptides were produced by Auspep (Melbourne, Australia). Animals were obtained from the Central Animal Breeding House, University of Queensland. Secondary antibodies were obtained from Amersham (Australia). Antisera were raised in rabbits using our standard protocols for production of antisera [11] and which have been used to generate antisera against other amino acid transporters including, glycine, glutamate, GABA and taurine transporters [2,8,12]. Peptides MVPHTILARGRDVCRRNGLLILSV and RDEELPAASLNHCTIQISELETNV corresponding to the amino and carboxyl terminal residues of the human sequence published by Arriza et al. [1] (GenBank accession number U76362) were conjugated to porcine thyroglobulin and rabbits immunized as previously described [11]. For immunocytochemistry, rats were anaesthetized with sodium pentobarbital (100 mg/kg, i.p.) and perfused transcardially
0304-3940/00/$ - see front matter q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 98 8- X
22
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) and then post®xed by immersion for 1 h. Vibratome sections (40 mm thick) of retinae and brains were immunolabeled using each antiserum at a dilution of 1:20 000. Labeling was detected using biotinylated secondary antibodies, streptavidin±horseradish peroxidase complex, and revealed using diaminobenzidine as a chromogen. Immunocytochemical controls consisted of: (i) use of pre-immune sera; (ii) pre-absorption of each antiserum (adding 20 mg of immunizing or irrelevant peptide to 1 ml of diluted antiserum); and (iii) omission of primary or secondary reagents. SDS-PAGE electrophoresis and Western blotting on nitrocellulose membranes was performed as previously described [12]. Each antiserum, when used at a dilution of 1:20 000, labeled a single band of approximately 63 kDa (Fig. 1), which is similar to the predicted molecular weight of EAAT5. Labeling of this band was abolished by preabsorbtion of the diluted antisera with the appropriate immunizing peptide. Conversely, labeling was not in¯uenced by pre-absorption with thyroglobulin. Dot blots were carried out as previously described, using each of the peptides coupled to bovine serum albumin (rather than thyroglobulin). Each antiserum recognized the appropriate peptide but did not label the other peptide (not illustrated). Immunolabeling was performed on sections of adult brains of a variety of ages. Neither antisera showed any speci®c labeling of the brain sections (not illustrated). When used on sections of retina, consistent and identical immunocytochemical results were obtained with the carboxyl-and amino-terminal directed antisera, thus whilst space constraints allow only the depiction of results obtained with the C-terminal antiserum, the descriptions below apply to both antisera. No speci®c labeling was detected with either antiserum up to and including P5 (Fig. 2). At P7, weak labeling was detected with both antisera in the central retina examined (Fig. 3). Staining in peripheral regions of retina at this age was usually very weak or non-existent, suggesting that expression of EAAT5 in the developmentally less mature peripheral retina lagged behind that observed in central retina by about 1 day (data not shown). Labeling was associated with cells in the photoreceptor layer of the retina. By P10 the labeling in this
Fig. 1. Western blots showing labeling of a single band of approximately 63 kDa by the C-terminal directed antiserum (lane B) and the N-terminal directed antiserum (lane C). Relative positions of molecular weight markers are also indicated (lane A).
layer in both central retina (Fig. 4) and peripheral retina (not shown) was strong with no evidence for a center-periphery gradient. Whilst some immunonegative cells might have been present they would have been obscured by the strong labeling of the vast majority of the photoreceptors. The pattern of labeling at P14 was similar to that at P10. By P21 (Fig. 5), an additional change occurred, in that whilst photoreceptor labeling was still strong, weak labeling had also appeared in the ovoid somata and proximal processes of a population of cells in the outer third of the inner nuclear layer. On the basis of the soma shape and position within the inner nuclear layer these cells were interpreted as being bipolar cells [18,20]. By adulthood this bipolar labeling was prominent, but was still restricted to the somata and proximal processes of these cells (Figs. 6 and 7). Many protein antigens are not completely accessible to antibodies after ®xation; treatment with trypsin can sometimes reveal cryptic patterns of labeling (Pow et al. [13]). Trypsinization was performed on 15±20 Vibratome sections from three retinae of each age using standard protocols [10]. The duration of trypsin digestion was varied from 5±30 min to optimize exposure of epitopes whilst minimizing generalized degradation of the tissue. After trypsin treatment of adult retinae two distinct effects were observed. Depending on the extent of digestion by the trypsin, labeling was partially lost from the photoreceptors and some bipolar cell somata. Conversely however, new sites of labeling were revealed in populations of large rod bipolar cell terminals in adult retinae and in processes and terminals of some cone bipolar cells (Figs. 8 and 9). The positions of the terminals in the inner plexiform layer suggested that a subset of both ON and OFF cone bipolar cells were labeled. In some sections immunoreactive processes could be seen to extend between these labeled terminals and the labeled bipolar cell somata, con®rming that the labeled bipolar cell bodies give rise to the observed populations of terminals. Trypsinization did not reveal additional labeling in bipolar cells or other cellular compartments of P1±P21 retinae (not shown). Collectively our data show that, as in the primate retina, immunoreactivity for EAAT5 is associated with photoreceptors. However, in the adult rat, EAAT5 is also present in a population of bipolar cells. The rat retina is heavily rod dominated with few cone photoreceptors, thus we conclude that the photoreceptor labeling we observe is associated with rods. Given the scarcity of cones in the rat retina we were unable to determine if rat cone photoreceptors do or do not express EAAT 5. However in view of our primate data, (where the abundant and easily identi®able cones do not express EAAT5), we suggest that the rat cones probably lack EAAT5. This is reasonable since cones, but not rods, express the glutamate transporter GLT-1 [15], and suggests that these two glutamate transporters have distinct roles in these two different photoreceptors groups. Our conclusion that photoreceptors express EAAT5 concurs with previous assumptions that EAAT5 is the photoreceptor glutamate
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
23
Fig. 2. P5 retina, immunolabelled with the C-terminal antiserum to EAAT5. No speci®c labelling is apparent. IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 3. P7 retina immunolabelled with the C-terminal antiserum to EAAT5, Uniformly weak labelling is associated with cells in the photoreceptor layer (P). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 4. P10 retina immunolabelled with the C-terminal antiserum to EAAT5, showing stronger labelling is associated with cells in the photoreceptor layer (P). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 5. P21 retina immunolabelled with the C-terminal antiserum to EAAT5, showing strong labelling of photoreceptors (P) and additional labelling in the somata of a population of bipolar cells (B, arrow). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 6. Adult retina immunolabelled with the C-terminal antiserum to EAAT5. Labelling is associated with photoreceptors (P) and somata and proximal processes of bipolar cells (B, arrow). The inner plexiform layer (IPL) is unlabelled. Scale bar, 50 mm. Fig. 7. High magni®cation views of adult rat retina, immunolabeled with antisera to the C-terminal region of EAAT5. Photoreceptors (P) are labeled as are the somata and proximal axonal segments of populations of bipolar cells (B, arrow). Scale bar, 10 mm. Fig. 8. Section of adult rat retinae which has been trypsinized and subsequently labeled for EAAT5 (C-terminal antiserum). Labeling is reduced in the photoreceptor layer but labeling is still evident in many bipolar cell somata (B, arrows). Additional labeling is also present in many large spherical or ovoid structures in the inner part of the inner plexiform layer; these large structures are interpreted as being rod bipolar terminals (RB). Scale bar, 25 mm. Fig. 9. Montaged section of an adult rat retina which has been trypsinized and subsequently labeled for EAAT5 (N-terminal antiserum). In this particular section, the somata, axonal segment and terminals of a cone bipolar cell (CB) are clearly labeled for EAAT5 (arrow). Rod bipolar terminals (RB) are also weakly labeled. Scale bar, 15 mm.
transporter [17]. However this view is in contrast to previous immunocytochemical studies of the salamander where immunolabeling for the two salamander homologues of EAAT5 was associated with MuÈller cells [4]. This difference is certainly plausible given the differences between mammalian and non-mammalian retinae in terms of amino acid transport systems. Thus as an example, mammalian MuÈller cells express a GABA transporter, whereas ®sh MuÈller cells do not [5,7]. The developmental pro®le of EAAT5 is entirely commensurate with the notion of it
being a photoreceptor glutamate transporter. Its appearance in central retina at P7 is concomitant with the appearance of anatomically demonstrable photoreceptor ribbon synapses and outer segments [3,19]. At this point we assume that the photoreceptor would have some competency to release glutamate in a physiologically evoked manner. Since glutamate release without subsequent recovery would lead to the rapid loss of cellular glutamate content, the concomitant expression of EAAT5 at this time point is signi®cant (though circumstantial) evidence for EAAT5 having an
24
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
active role in the homeostasis of glutamate in rod photoreceptors. Interestingly we have previously shown [8] that the MuÈller cell glutamate transporter GLAST is also switched on at P7. This requirement for homeostasis of glutamate is not particularly surprising since glutamate is a factor involved in shaping developmental events such as the patterns of retinal synaptogenesis [20]. The presence of immunoreactive EAAT5 in a population of bipolar cells was ®rst observed at P21. This rather late appearance is not without precedence since other markers such as protein kinase c show a delayed developmental expression, adult levels only being attained in the 4th postnatal week [20]. The results of our proteolytic digestion experiments raise a number of interesting issues. The differential labeling of somata and processes of bipolar cells strongly suggests that there is a conformational change in the epitope which masks the immunoreactive epitope in compartments distal to the cell body, unless revealed by partial proteolysis. We recognize the numerous pitfalls attendant to trypsin treatment and indeed parts of the peptide sequence used as our original immunogen have adjacent basic residues. If these residues were exposed to the trypsin by virtue of being at an exposed locus, they could be cleaved, a fact which probably explains the reduction in photoreceptor labeling after trypsinization. Conversely the retention of labeling in the bipolar cell terminals after trypsinization suggests that the epitope was not at the surface of the molecule, or the surface of the cell and thus not easily accessible to the trypsin. We conclude that EAAT5 is most probably a rod photoreceptor glutamate transporter in the rat retina. The time course of expression corresponds exactly with the time course of onset of activity in photoreceptors. The use of a different transporter (GLT-1) in cones poses the fundamental question as to why two distinct tranporter systems are needed for cells performing such similar roles. The answer presumably lies in the different temporal kinetics of these two systems but this awaits experimental investigation. Similarly EAAT5 probably also functions as a glutamate transporter in a population of cone and rod bipolar cells. D.V.P. is supported by an NHMRC Research Fellowship. Supported by NHMRC project grants to D.V.P. and to N.L.B. & D.V.P. [1] Arriza, J.L., Eliasof, S., Kavanaugh, M.P. and Amara, S.G., Excitatory amino acid transporter 5, a retinal glutamate transporter coupled to a chloride conductance. Proc. Natl. Acad. Sci. USA, 94 (1997) 4155±4160. [2] Barnett, N.L. and Pow, D.V., Immunocytochemical localisation of a taurine transporter in the mammalian retina. Invest. Ophthalmol. Visual Sci., 39 (1998) 4865.
[3] Braekevelt, C.R. and Hollenberg, M.J., The development of the retina of the albino rat. Am. J. Anat., 127 (1970) 281±302. [4] Eliasof, S., Arriza, J.L., Leighton, B.H., Kavanaugh, M.P. and Amara, S.G., Excitatory amino acid transporters of the salamander retina: identi®cation, localization, and function. J. Neurosci., 18 (1998) 698±712. [5] Lam, D.M.K. and Steinman, L., The uptake of (g-3H) aminobutyric acid in the gold®sh retina. Proc. Natl. Acad. Sci. USA, 68 (1971) 2777±2781. [6] Lehre, K.P., Davanger, S. and Danbolt, N.C., Localization of the glutamate transporter GLAST in rat retina. Brain Res., 744 (1997) 129±137. [7] Pow, D.V., Baldridge, W. and Crook, D.K., Activity-dependent transport of GABA analogues into speci®c cell types demonstrated at high resolution using a novel immunocytochemical strategy. Neuroscience, 73 (1996) 1129±1143. [8] Pow, D.V. and Barnett, N.L., Changing patterns of spatial buffering of glutamate in developing rat retinae are mediated by the MuÈller cell glutamate transporter GLAST. Cell Tissue Res., 297 (1999) 57±66. [9] Pow, D.V., Barnett, N.L. and Penfold, P., Roles of glial and neuronal transporters in retinal glutamate homeostasis. Neurochem. Int., (2000) in press. [10] Pow, D.V. and Clark, A., Localisation of peptide hormones by light and electron microscopy. In Hutton, J.C., Siddle, K. (Eds.), Peptide Hormone Secretion (Practical handbook series), IRL Press, Oxford, 1990, pp. 189±210. [11] Pow, D.V. and Crook, D., Extremely high titre polyclonal antisera against small neurotransmitter molecules: rapid production, characterisation and use in light- and electron-microscopic immunocytochemistry. J. Neurosci. Meth., 48 (1993) 51±63. [12] Pow, D.V. and Hendrickson, A., Distribution of the glycine transporter glyt-1 in mammalian and non-mammalian retinae. Vis. Neurosci., 16 (1999) 231±239. [13] Pow, D.V., Morris, J.F. and Rodgers, S., Tunicamycin, puromycin and brefeldin-A in¯uence the subcellular distribution of neuropeptides in hypothalamic magnocellular neurones. Cell Tissue Res., 269 (1992) 547±560. [14] Rauen, T. and Kanner, B.I., Localization of the glutamate transporter GLT-1 in rat and macaque monkey retinae. Neurosci. Lett., 169 (1994) 137±140. [15] Rauen, T., Rothstein, J.D. and Wassle, H., Differential expression of three glutamate transporter subtypes in the rat retina. Cell Tissue Res., 286 (1996) 325±336. [16] Rauen, T., Taylor, W.R., Kuhlbrodt, K. and Wiessner, M., High-af®nity glutamate transporters in the rat retina: a major role of the glial glutamate transporter GLAST-1 in transmitter clearance. Cell Tissue Res., 291 (1998) 19±31. [17] Spiridon, M., Kamm, D., Billups, B., Mobbs, P. and Attwell, D., Modulation by zinc of the glutamate transporters in glial cells and cones isolated from the tiger salamander retina. J. Physiol. (Lond.), 506 (1998) 363±376. [18] Weidmann, T.A. and Kuwabara, T., Development of the rat retina. Invest. Ophthalmol., 8 (1969) 60±69. [19] Wong, R.O.L., Effects of glutamate and its analogs on intracellular calcium levels in the developing retina. Vis. Neurosci., 12 (1995) 907±917. [20] Zhang, D.R. and Yeh, H.H., Protein kinase C-like immunoreactivity in rod bipolar cells of the rat retina: a developmental study. Vis. Neurosci., 6 (1991) 429±437.
Developmental expression of excitatory amino acid transporter 5: a photoreceptor and bipolar cell glutamate transporter in rat retina David V. Pow a,*, Nigel L. Barnett b a
Department of Anatomical Sciences, University of Queensland, Brisbane, Queensland 4306, Australia b VTHRC, University of Queensland, Brisbane, Queensland 4306, Australia Received 5 October 1999; received in revised form 3 November 1999; accepted 8 December 1999
Abstract Excitatory amino acid transporter 5 (EAAT5) is a retina-speci®c glutamate transporter which has an associated chloride conductance. Thus it is comparable in its functional properties to the glutamate transport systems previously described in photoreceptors and some bipolar cells. We have raised antibodies to the carboxyl- and amino-terminal regions of EAAT5. Labeling for both of these antisera was developmentally regulated: weak labeling appeared in photoreceptors around P7; by P10 strong labeling was present in photoreceptors and by P21 a population of bipolar elements were also weakly labeled. In adult retinae both antisera heavily immunolabeled all photoreceptors as well as a heterogeneous population of bipolar cell somata and their proximal axonal processes: synaptic terminals of these cells were also labeled after partial proteolytic digestion of the tissues. The positions and morphology of these terminals suggests that they are the terminals of both rod and cone rod bipolar cells. We conclude that in rat retina, EAAT5 is a photoreceptor and bipolar cell glutamate transporter. q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Bipolar cell; Glutamate; Excitatory amino acid transporter 5; Photoreceptor; Retina; Transporter
The retina expresses a variety of glutamate transporters including GLAST which is present in MuÈller cells [6,8,15,16], GLT-1 which is present in cone bipolar cells and cone photoreceptors [14], EAAC1 which is present in horizontal cells, some amacrine cells and ganglion cells [16] and excitatory amino acid transporter 5 (EAAT5). EAAT5 is a retina-speci®c glutamate transporter [1]. We have recently shown that in primate retina EAAT 5 is selectively associated with rod photoreceptors [9]. In this study we have investigated whether rat photoreceptors or other retinal neurons express EAAT5. Since there is always the possibility that an antibody to one end of a protein alone might recognize other related proteins this study used antisera to both the carboxyl and amino termini of EAAT5, thereby reducing the possibility of erroneously labeling a cell population which expresses a related molecular species other than EAAT5. Expression of EAAT5 was examined at a series of landmark developmental time points such as the day of birth, postnatal day 7 when photoreceptor synapses ®rst appear [19], and the day of eye opening (postnatal day 14). P21 and adult retinae were also examined. For each age, retinae * Corresponding author. Fax: 161-7-3365-1299. E-mail address: [email protected] (D.V. Pow)
from 5 rats were examined. The developmentally advanced central third of the retina was preferentially examined, and the descriptions that ensue refer to this region. However the peripheral areas of each retina were also examined. Our working hypothesis was that onset of EAAT5 expression might be associated with synaptogenesis and thus onset of glutamate release from photoreceptors. All experiments were carried out in accordance with the ethical guidelines of the NHMRC (Australia). Synthetic peptides were produced by Auspep (Melbourne, Australia). Animals were obtained from the Central Animal Breeding House, University of Queensland. Secondary antibodies were obtained from Amersham (Australia). Antisera were raised in rabbits using our standard protocols for production of antisera [11] and which have been used to generate antisera against other amino acid transporters including, glycine, glutamate, GABA and taurine transporters [2,8,12]. Peptides MVPHTILARGRDVCRRNGLLILSV and RDEELPAASLNHCTIQISELETNV corresponding to the amino and carboxyl terminal residues of the human sequence published by Arriza et al. [1] (GenBank accession number U76362) were conjugated to porcine thyroglobulin and rabbits immunized as previously described [11]. For immunocytochemistry, rats were anaesthetized with sodium pentobarbital (100 mg/kg, i.p.) and perfused transcardially
0304-3940/00/$ - see front matter q 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 9 9) 00 98 8- X
22
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) and then post®xed by immersion for 1 h. Vibratome sections (40 mm thick) of retinae and brains were immunolabeled using each antiserum at a dilution of 1:20 000. Labeling was detected using biotinylated secondary antibodies, streptavidin±horseradish peroxidase complex, and revealed using diaminobenzidine as a chromogen. Immunocytochemical controls consisted of: (i) use of pre-immune sera; (ii) pre-absorption of each antiserum (adding 20 mg of immunizing or irrelevant peptide to 1 ml of diluted antiserum); and (iii) omission of primary or secondary reagents. SDS-PAGE electrophoresis and Western blotting on nitrocellulose membranes was performed as previously described [12]. Each antiserum, when used at a dilution of 1:20 000, labeled a single band of approximately 63 kDa (Fig. 1), which is similar to the predicted molecular weight of EAAT5. Labeling of this band was abolished by preabsorbtion of the diluted antisera with the appropriate immunizing peptide. Conversely, labeling was not in¯uenced by pre-absorption with thyroglobulin. Dot blots were carried out as previously described, using each of the peptides coupled to bovine serum albumin (rather than thyroglobulin). Each antiserum recognized the appropriate peptide but did not label the other peptide (not illustrated). Immunolabeling was performed on sections of adult brains of a variety of ages. Neither antisera showed any speci®c labeling of the brain sections (not illustrated). When used on sections of retina, consistent and identical immunocytochemical results were obtained with the carboxyl-and amino-terminal directed antisera, thus whilst space constraints allow only the depiction of results obtained with the C-terminal antiserum, the descriptions below apply to both antisera. No speci®c labeling was detected with either antiserum up to and including P5 (Fig. 2). At P7, weak labeling was detected with both antisera in the central retina examined (Fig. 3). Staining in peripheral regions of retina at this age was usually very weak or non-existent, suggesting that expression of EAAT5 in the developmentally less mature peripheral retina lagged behind that observed in central retina by about 1 day (data not shown). Labeling was associated with cells in the photoreceptor layer of the retina. By P10 the labeling in this
Fig. 1. Western blots showing labeling of a single band of approximately 63 kDa by the C-terminal directed antiserum (lane B) and the N-terminal directed antiserum (lane C). Relative positions of molecular weight markers are also indicated (lane A).
layer in both central retina (Fig. 4) and peripheral retina (not shown) was strong with no evidence for a center-periphery gradient. Whilst some immunonegative cells might have been present they would have been obscured by the strong labeling of the vast majority of the photoreceptors. The pattern of labeling at P14 was similar to that at P10. By P21 (Fig. 5), an additional change occurred, in that whilst photoreceptor labeling was still strong, weak labeling had also appeared in the ovoid somata and proximal processes of a population of cells in the outer third of the inner nuclear layer. On the basis of the soma shape and position within the inner nuclear layer these cells were interpreted as being bipolar cells [18,20]. By adulthood this bipolar labeling was prominent, but was still restricted to the somata and proximal processes of these cells (Figs. 6 and 7). Many protein antigens are not completely accessible to antibodies after ®xation; treatment with trypsin can sometimes reveal cryptic patterns of labeling (Pow et al. [13]). Trypsinization was performed on 15±20 Vibratome sections from three retinae of each age using standard protocols [10]. The duration of trypsin digestion was varied from 5±30 min to optimize exposure of epitopes whilst minimizing generalized degradation of the tissue. After trypsin treatment of adult retinae two distinct effects were observed. Depending on the extent of digestion by the trypsin, labeling was partially lost from the photoreceptors and some bipolar cell somata. Conversely however, new sites of labeling were revealed in populations of large rod bipolar cell terminals in adult retinae and in processes and terminals of some cone bipolar cells (Figs. 8 and 9). The positions of the terminals in the inner plexiform layer suggested that a subset of both ON and OFF cone bipolar cells were labeled. In some sections immunoreactive processes could be seen to extend between these labeled terminals and the labeled bipolar cell somata, con®rming that the labeled bipolar cell bodies give rise to the observed populations of terminals. Trypsinization did not reveal additional labeling in bipolar cells or other cellular compartments of P1±P21 retinae (not shown). Collectively our data show that, as in the primate retina, immunoreactivity for EAAT5 is associated with photoreceptors. However, in the adult rat, EAAT5 is also present in a population of bipolar cells. The rat retina is heavily rod dominated with few cone photoreceptors, thus we conclude that the photoreceptor labeling we observe is associated with rods. Given the scarcity of cones in the rat retina we were unable to determine if rat cone photoreceptors do or do not express EAAT 5. However in view of our primate data, (where the abundant and easily identi®able cones do not express EAAT5), we suggest that the rat cones probably lack EAAT5. This is reasonable since cones, but not rods, express the glutamate transporter GLT-1 [15], and suggests that these two glutamate transporters have distinct roles in these two different photoreceptors groups. Our conclusion that photoreceptors express EAAT5 concurs with previous assumptions that EAAT5 is the photoreceptor glutamate
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
23
Fig. 2. P5 retina, immunolabelled with the C-terminal antiserum to EAAT5. No speci®c labelling is apparent. IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 3. P7 retina immunolabelled with the C-terminal antiserum to EAAT5, Uniformly weak labelling is associated with cells in the photoreceptor layer (P). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 4. P10 retina immunolabelled with the C-terminal antiserum to EAAT5, showing stronger labelling is associated with cells in the photoreceptor layer (P). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 5. P21 retina immunolabelled with the C-terminal antiserum to EAAT5, showing strong labelling of photoreceptors (P) and additional labelling in the somata of a population of bipolar cells (B, arrow). IPL indicates position of inner plexiform layer. Scale bar, 50 mm. Fig. 6. Adult retina immunolabelled with the C-terminal antiserum to EAAT5. Labelling is associated with photoreceptors (P) and somata and proximal processes of bipolar cells (B, arrow). The inner plexiform layer (IPL) is unlabelled. Scale bar, 50 mm. Fig. 7. High magni®cation views of adult rat retina, immunolabeled with antisera to the C-terminal region of EAAT5. Photoreceptors (P) are labeled as are the somata and proximal axonal segments of populations of bipolar cells (B, arrow). Scale bar, 10 mm. Fig. 8. Section of adult rat retinae which has been trypsinized and subsequently labeled for EAAT5 (C-terminal antiserum). Labeling is reduced in the photoreceptor layer but labeling is still evident in many bipolar cell somata (B, arrows). Additional labeling is also present in many large spherical or ovoid structures in the inner part of the inner plexiform layer; these large structures are interpreted as being rod bipolar terminals (RB). Scale bar, 25 mm. Fig. 9. Montaged section of an adult rat retina which has been trypsinized and subsequently labeled for EAAT5 (N-terminal antiserum). In this particular section, the somata, axonal segment and terminals of a cone bipolar cell (CB) are clearly labeled for EAAT5 (arrow). Rod bipolar terminals (RB) are also weakly labeled. Scale bar, 15 mm.
transporter [17]. However this view is in contrast to previous immunocytochemical studies of the salamander where immunolabeling for the two salamander homologues of EAAT5 was associated with MuÈller cells [4]. This difference is certainly plausible given the differences between mammalian and non-mammalian retinae in terms of amino acid transport systems. Thus as an example, mammalian MuÈller cells express a GABA transporter, whereas ®sh MuÈller cells do not [5,7]. The developmental pro®le of EAAT5 is entirely commensurate with the notion of it
being a photoreceptor glutamate transporter. Its appearance in central retina at P7 is concomitant with the appearance of anatomically demonstrable photoreceptor ribbon synapses and outer segments [3,19]. At this point we assume that the photoreceptor would have some competency to release glutamate in a physiologically evoked manner. Since glutamate release without subsequent recovery would lead to the rapid loss of cellular glutamate content, the concomitant expression of EAAT5 at this time point is signi®cant (though circumstantial) evidence for EAAT5 having an
24
D.V. Pow, N.L. Barnett / Neuroscience Letters 280 (2000) 21±24
active role in the homeostasis of glutamate in rod photoreceptors. Interestingly we have previously shown [8] that the MuÈller cell glutamate transporter GLAST is also switched on at P7. This requirement for homeostasis of glutamate is not particularly surprising since glutamate is a factor involved in shaping developmental events such as the patterns of retinal synaptogenesis [20]. The presence of immunoreactive EAAT5 in a population of bipolar cells was ®rst observed at P21. This rather late appearance is not without precedence since other markers such as protein kinase c show a delayed developmental expression, adult levels only being attained in the 4th postnatal week [20]. The results of our proteolytic digestion experiments raise a number of interesting issues. The differential labeling of somata and processes of bipolar cells strongly suggests that there is a conformational change in the epitope which masks the immunoreactive epitope in compartments distal to the cell body, unless revealed by partial proteolysis. We recognize the numerous pitfalls attendant to trypsin treatment and indeed parts of the peptide sequence used as our original immunogen have adjacent basic residues. If these residues were exposed to the trypsin by virtue of being at an exposed locus, they could be cleaved, a fact which probably explains the reduction in photoreceptor labeling after trypsinization. Conversely the retention of labeling in the bipolar cell terminals after trypsinization suggests that the epitope was not at the surface of the molecule, or the surface of the cell and thus not easily accessible to the trypsin. We conclude that EAAT5 is most probably a rod photoreceptor glutamate transporter in the rat retina. The time course of expression corresponds exactly with the time course of onset of activity in photoreceptors. The use of a different transporter (GLT-1) in cones poses the fundamental question as to why two distinct tranporter systems are needed for cells performing such similar roles. The answer presumably lies in the different temporal kinetics of these two systems but this awaits experimental investigation. Similarly EAAT5 probably also functions as a glutamate transporter in a population of cone and rod bipolar cells. D.V.P. is supported by an NHMRC Research Fellowship. Supported by NHMRC project grants to D.V.P. and to N.L.B. & D.V.P. [1] Arriza, J.L., Eliasof, S., Kavanaugh, M.P. and Amara, S.G., Excitatory amino acid transporter 5, a retinal glutamate transporter coupled to a chloride conductance. Proc. Natl. Acad. Sci. USA, 94 (1997) 4155±4160. [2] Barnett, N.L. and Pow, D.V., Immunocytochemical localisation of a taurine transporter in the mammalian retina. Invest. Ophthalmol. Visual Sci., 39 (1998) 4865.
[3] Braekevelt, C.R. and Hollenberg, M.J., The development of the retina of the albino rat. Am. J. Anat., 127 (1970) 281±302. [4] Eliasof, S., Arriza, J.L., Leighton, B.H., Kavanaugh, M.P. and Amara, S.G., Excitatory amino acid transporters of the salamander retina: identi®cation, localization, and function. J. Neurosci., 18 (1998) 698±712. [5] Lam, D.M.K. and Steinman, L., The uptake of (g-3H) aminobutyric acid in the gold®sh retina. Proc. Natl. Acad. Sci. USA, 68 (1971) 2777±2781. [6] Lehre, K.P., Davanger, S. and Danbolt, N.C., Localization of the glutamate transporter GLAST in rat retina. Brain Res., 744 (1997) 129±137. [7] Pow, D.V., Baldridge, W. and Crook, D.K., Activity-dependent transport of GABA analogues into speci®c cell types demonstrated at high resolution using a novel immunocytochemical strategy. Neuroscience, 73 (1996) 1129±1143. [8] Pow, D.V. and Barnett, N.L., Changing patterns of spatial buffering of glutamate in developing rat retinae are mediated by the MuÈller cell glutamate transporter GLAST. Cell Tissue Res., 297 (1999) 57±66. [9] Pow, D.V., Barnett, N.L. and Penfold, P., Roles of glial and neuronal transporters in retinal glutamate homeostasis. Neurochem. Int., (2000) in press. [10] Pow, D.V. and Clark, A., Localisation of peptide hormones by light and electron microscopy. In Hutton, J.C., Siddle, K. (Eds.), Peptide Hormone Secretion (Practical handbook series), IRL Press, Oxford, 1990, pp. 189±210. [11] Pow, D.V. and Crook, D., Extremely high titre polyclonal antisera against small neurotransmitter molecules: rapid production, characterisation and use in light- and electron-microscopic immunocytochemistry. J. Neurosci. Meth., 48 (1993) 51±63. [12] Pow, D.V. and Hendrickson, A., Distribution of the glycine transporter glyt-1 in mammalian and non-mammalian retinae. Vis. Neurosci., 16 (1999) 231±239. [13] Pow, D.V., Morris, J.F. and Rodgers, S., Tunicamycin, puromycin and brefeldin-A in¯uence the subcellular distribution of neuropeptides in hypothalamic magnocellular neurones. Cell Tissue Res., 269 (1992) 547±560. [14] Rauen, T. and Kanner, B.I., Localization of the glutamate transporter GLT-1 in rat and macaque monkey retinae. Neurosci. Lett., 169 (1994) 137±140. [15] Rauen, T., Rothstein, J.D. and Wassle, H., Differential expression of three glutamate transporter subtypes in the rat retina. Cell Tissue Res., 286 (1996) 325±336. [16] Rauen, T., Taylor, W.R., Kuhlbrodt, K. and Wiessner, M., High-af®nity glutamate transporters in the rat retina: a major role of the glial glutamate transporter GLAST-1 in transmitter clearance. Cell Tissue Res., 291 (1998) 19±31. [17] Spiridon, M., Kamm, D., Billups, B., Mobbs, P. and Attwell, D., Modulation by zinc of the glutamate transporters in glial cells and cones isolated from the tiger salamander retina. J. Physiol. (Lond.), 506 (1998) 363±376. [18] Weidmann, T.A. and Kuwabara, T., Development of the rat retina. Invest. Ophthalmol., 8 (1969) 60±69. [19] Wong, R.O.L., Effects of glutamate and its analogs on intracellular calcium levels in the developing retina. Vis. Neurosci., 12 (1995) 907±917. [20] Zhang, D.R. and Yeh, H.H., Protein kinase C-like immunoreactivity in rod bipolar cells of the rat retina: a developmental study. Vis. Neurosci., 6 (1991) 429±437.