- Email: [email protected]
Intracellular chloride concentration is higher in rod bipolar cells than in cone bipolar cells of the mouse retina
Neuroscience Letters 310 (2001) 161±164
www.elsevier.com/locate/neulet
Intracellular chloride concentration is higher in rod bipolar cells than in cone bipolar cells of the mouse retina Hiromasa Satoh*, Makoto Kaneda, Akimichi Kaneko Department of Physiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan Received 9 July 2001; accepted 17 July 2001
Abstract Bipolar cells (BCs) have antagonistic center-surround receptive ®eld. Surround illumination evokes depolarization in the OFF-type cone BC, and hyperpolarization in the rod BC and the ON-type cone BC. Surround illumination reduces gamma-aminobutyric acid (GABA) release from horizontal cells. If GABA hyperpolarize BCs, the polarity of the GABAinduced effect agrees with the light-evoked surround response in the OFF-type BC, but contradicts in the rod BC and the ON-type cone BC. Immunohistochemical study of the Cl 2 transporter of BCs has suggested that the intracellular Cl 2 concentration is different among BC subtypes. We examined the reversal potential of GABA-induced current of BCs using gramicidin-perforated patch clamp technique in the mouse retina, and found that GABA depolarizes rod BC and hyperpolarizes cone BCs. Our results are consistent with the GABAergic input to rod BC dendrite. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Bipolar cells; Perforated patch; Gamma-aminobutyric acid; Intracellular chloride; Retina; Mouse
Bipolar cells (BCs) in the vertebrate retina receive gamma-aminobutyric acid (GABA)ergic inputs from horizontal cells at the dendrite and from amacrine cells at the axon terminal. In the outer plexiform layer (OPL) of the retina, darkness depolarizes horizontal cells and thereby increases the GABA release [17]. In BCs, the direct glutamatergic input from photoreceptors generates the center response, while the GABAergic input from horizontal cells is considered to produce the surround response. Since increase in Cl 2 conductance by GABA is thought to hyperpolarize BCs, the reduction of GABAergic input by surround illumination is thought to produce depolarization. This idea explains the depolarization of OFF-type BCs by surround illumination. On the other hand, surround illumination induces hyperpolarization in ON-type BCs [7]. This is contradictory to what we expect in the horizontal to bipolar feed-forward connection. It has been a puzzle how the surround hyperpolarization in ON-type BCs is accounted for in relation to GABA released from horizontal cells. It is generally thought that the intracellular chloride concentration ([Cl 2]i) is regulated primarily by two chloride transporters, an Na-K-Cl cotransporter (NKCC) that * Corresponding author. Tel.: 181-3-5363-3749; fax: 181-33359-0437. E-mail address: [email protected] (H. Satoh).
normally accumulates Cl 2 [14], and a K-Cl cotransporter (KCC2) that normally extrudes Cl 2 [11,15]. Immunohistochemical localization of Cl 2 transporters in the mammalian retina revealed that dendrites of the OFF-type BC express KCC2 and the dendrites of rod and ON-type BC express NKCC [20]. They hypothesized that [Cl 2]i of the rod BC and the ON-type BC is higher than that of the OFF-type BC. The aim of the present study is to examine the above hypothesis by comparing ECl in three types of BCs using gramicidin-perforated patch clamp technique [9]. We found that ECl of rod BCs is statistically more positive than ECl of cone BCs of both ON-type and OFF-type. All experiments were performed on retinal slices prepared from mice (C57BL/6J, 6±8-weeks-old, male). The care of the animals was in accordance with the Guidelines for the Care and Use of Laboratory Animals, Keio University School of Medicine, and our experiments were approved by the University Animal Welfare Committee. For retinal slices preparation (thickness, 150 mm) we followed the method described in detail previously [16]. Slices were continuously superfused with oxygenated external solution of the following composition: (in mM) NaCl 135; KCl 5; CaCl2 2; MgCl2 1; HEPES 5 and glucose 10 (pH 7.4). The pipette solution contained (in mM): KCl 120; CaCl2 0.5; MgCl2 1; EGTA 5; HEPES 10; ATP±tris 5 (pH 7.2); Lucifer yellow (0.2 %) and gramicidin D (14.8 mg/ml). Lucifer
0304-3940/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 1) 02 12 0- 6
162
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
yellow and gramicidin D were added immediately before use. The resistance of the ®lled pipette was 10±13 MV. Pipettes were connected to an EPC-9 patch-clamp ampli®er (HEKA Elektronik). The holding potential was corrected for the liquid junction potentials measured by the method previously reported [8]. Current and voltage signals were recorded on a thermal arraycorder (WR8500, Graphtec) and stored on a digital tape (PC204Ax, Sony). Sampled signals were analyzed of¯ine. GABA (Sigma) was applied by a puffer system (applied pressure, 60 kPa). Most experiments were carried out in room temperature (20± 258C, n 28). To examine whether the Cl 2 transporter was temperature sensitive, we carried out some recordings at 31±338C (n 6). Because no signi®cant difference in ECl was found, we also used the data obtained under the two temperature conditions. After recording, all recorded cells were ®lled with Lucifer yellow by rupturing the perforated patch membrane. BCs were morphologically classi®ed into rod BC (n 18), ONtype cone BC (n 13) or OFF-type cone BC (n 3). As demonstrated in the photomicrograph of a representative cell of each group (Fig. 1A), BC subtypes can be identi®ed by the location of the cell body, the position of the axon terminal in the inner plexiform layer (IPL), and the morphology of the axon terminals [3]. During recording, cells were visually monitored under the ¯uorescent microscope. If the cell was found stained during recording, we judged that the patch membrane was broken and the data from this cell were discarded. Cells which lacked dendrite and/or axon terminals were excluded from the analysis. GABA applied to OPL induced an outward current in a rod BC held at 255 mV under the gramicidin-perforated mode (Fig. 1Ba). Since the gramicidin-ionophore selectively permeates monovalent cations and is impermeable for anions [9], [Cl 2]i of the BC in situ was estimated by the reversal potential of GABA-induced current. Under the gramicidin-perforated voltage clamp condition, the reversal potential was 245 mV (®lled circle, Fig. 1C). The [Cl 2]i estimated by the Nernst equation was 19.1 mM. Immediately after this recording, the membrane patch at the pipette tip was ruptured and the recording con®guration was switched into the whole cell clamp mode. Under this new condition the polarity of GABA-induced current became inward within 10 s after membrane rupture (data not shown). The change is accounted for by the diffusion of Cl 2 from the recording pipette to the ruptured BC, resulting in an increase in [Cl 2]i. Under the gramicidin-perforated recording condition, application of GABA induced an inward current in both ON-type and OFF-type cone BCs held at 255 mV (Fig. 1Bb,Bc). The reversal potential was 267 mV for the ONtype cone BC (open circle, Fig. 1C), and 265 mV for the OFF-type cone BC (cross, Fig. 1C). Statistically the reversal potential of GABA-induced current differed between the rod BCs and the cone BCs (Student's t-test, P , 0:001) (Fig. 2A); 239 ^ 10 mV (mean ^ SD) in rod BCs,
255 ^ 10 mV in ON-type cone BCs and 267 ^ 7 mV in OFF-type cone BCs. The difference between two types of cone BCs was statistically insigni®cant. The estimated average [Cl 2]i was 23.4 mM in the rod BC, 12.6 mM in the ONtype cone BC and 7.9 mM in the OFF-type cone BC. Based on the polarized distribution of KCC2 and NKCC in rod BCs, Vardi et al. [20] hypothesized that there is a
Fig. 1. (A) Photomicrographs of a rod BC (a), an ON-type cone BC (b) and an OFF-type cone BC (c). The image was captured by a CCD camera (M-3204C, Olympus) and processed by `Photoshop' software (version 5, Adobe). We focused on axon terminal to show the characteristic shape, so the soma was out of focus in (b). The position of layers labeled in the left is only approximation because slices have different thickness. INL, inner nuclear layer; GCL, ganglion cell layer. (a) and (b) represent sublamina a and b of IPL. Other abbreviations are de®ned in the text. Scale bar: 10 mm. (B) GABA-induced responses recorded under the gramicidin-perforated patch clamp condition of (a) a rod BC, (b) an ON-type cone BC and (c) an OFF-type cone BC. Holding voltage was 255 mV in all cells. Horizontal bar above each trace indicates GABA application (10 mM) by puffer pipette to OPL. (C) I-V relationships of GABA-induced current recorded in the same cells under the perforated-patch clamp condition (®lled circle, rod BC; open circle, ON-type cone BC; cross, OFF-type cone BC).
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
gradient of [Cl 2]i in the rod BC, high in the dendritic region and low in the axon terminal. To examine this hypothesis we assessed [Cl 2]i at the dendrite and the axon terminal in the same cell. Since high sensitivity regions to GABA at OPL and that at IPL were separated by a low sensitivity region as
Fig. 2. Reversal potential of GABA-induced current recorded under the gramicidin-perforated patch clamp con®guration. (A) Difference of reversal potential among subtypes. Each symbol represents reversal potential of individual BC. Reversal potential was obtained by GABA application (10 mM) to OPL. Circle represents rod BC, triangle ON-type cone BC and rhombus OFF-type BC. (B) Relationship between the reversal potential of the current responses to GABA given to OPL and that given to IPL. Circles represent rod BC (n 15) and triangles ON-type cone BC (n 9). The solid line representing rod BCs and the broken line representing ON-type cone BCs were drawn by the least square method.
163
has been reported in isolated BCs [18], we assumed that GABA applied to the scleral margin of OPL mainly activated dendritic GABA receptors, while GABA applied to IPL activated GABA receptors at the axon terminal. Data points of Fig. 2B represent the reversal potentials of a BC obtained by applying GABA to OPL (abscissa) and to IPL (ordinate). Although the data points representing rod BCs and those representing ON-type cone BCs make separate clusters because the reversal potentials are different, all data points fall close to the line with a unit slope (0.93 for rod BC and 0.91 for the ON-type cone BC). We were unable to ®nd evidence that [Cl 2]i is different between the dendrite and the axon terminal. It may re¯ect the rapid diffusion of Cl 2, since we observed ¯uorescent images of the axon terminal and the dendrite within 2±3 s after membrane rupture at the soma. We found that [Cl 2]i in rod BCs was higher than [Cl 2]i in cone BCs. In the hippocampal neuron [2] and retinal ganglion cells [5,6], it has been shown that the polarity of the GABA-induced response reverses from depolarization in neonate to hyperpolarization in adult, suggesting that [Cl 2]i decreases during development. Probably, switch of the Cl 2 transporter from NKCC to KCC2 is responsible for the developmental change of [Cl 2]i [11,14,15]. NKCC maintains [Cl 2]i high [14], while KCC2 functions to keep [Cl 2]i low [11,15]. Higher [Cl 2]i of rod BCs (23.4 mM) than that of OFF-type cone BCs (7.9 mM) is in parallel with the type of the Cl 2 transporters expressed in these BCs [20]. Although the immunohistochemical data suggest high [Cl 2]i of ON-type cone BCs, it was 12.6 mM. A possible interpretation is that expression of NKCC in rod BCs is stronger than that in ON-type cone BCs. Since the resting potential of BCs was approximately 245 mV, GABAergic input to BCs is depolarizing in rod BCs, while that is hyperpolarizing in cone BCs. The present observation is in favor of the idea that horizontal cells exert GABAergic feed-forward action to rod BCs. Formation of the surround response in ON-type cone BCs is still unanswered by the feed-forward input from horizontal cells. It may be accounted for rather by the feedback from cone horizontal cells to cones [1,10,13]. Cone horizontal cells are GABAergic [12], and release GABA upon depolarization [17]. Cone photoreceptors have high sensitivity to GABA at the axon terminal [19]. Surround illumination reduces the amount of GABA release from horizontal cells resulting in the reduction of GABA-induced feedback to cone photoreceptors. This, in turn, increases glutamate release from cones and glutamate hyperpolarizes ON-type BC. It is unlikely that such a feedback system exists between horizontal cells and rods [4]. Furthermore, rods have an extremely low sensitivity to GABA at their axon terminal [19]. All these observations are in favor of supporting the notion that receptive ®eld surround is formed by different mechanisms between ON-type cone BCs and rod BCs; feedback from horizontal cells to cones in ON-type cone BCs,
164
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
and feedforward action from horizontal cell to BCs in rod BCs. This work was supported by grant in aid to A. Kaneko (Nos. 11480247, 12053260) from the Ministry of Education, Science and Culture and to H. Satoh from Keio University Grant-in-Aid for Encouragement of Young Medical Scientists, and from the Promotion and Mutual Aid Corporation for Private Schools of Japan. We thank W.R. Taylor for providing software, S. Miyoshi and A. Koizumi for their technical advice. [1] Baylor, D.A., Fuortes, M.G.F. and O'Bryan, P.M., Receptive ®elds of cones in the retina of the turtle, J. Physiol., 214 (1971) 265±294. [2] Ben-Ari, Y., Cherubini, E., Corradetti, R. and Gaiarsa, J.L., Giant synaptic potentials in immature rat CA3 hippocampal neurones, J. Physiol., 416 (1989) 303±325. [3] Euler, T. and WaÈssle, H., Immunocytochemical identi®cation of cone bipolar cells in the rat retina, J. Comp. Neurol., 361 (1995) 461±478. [4] Fain, G.L., Gerschenfeld, H.M. and Quandt, F.N., Calcium spikes in toad rods, J. Physiol., 303 (1980) 495±513. [5] Fischer, K.F., Lukasiewicz, P.D. and Wong, R.O.L., Agedependent and cell class-speci®c modulation of retinal ganglion cell bursting activity by GABA, J. Neurosci., 18 (1998) 3767±3778. [6] Huang, B. and Redburn, D.A., GABA-induced increases in [Ca 21]i in retinal neurons of postnatal rabbits, Vis. Neurosci., 13 (1996) 441±447. [7] Kaneko, A., Physiological and morphological identi®cation of horizontal, bipolar and amacrine cells in gold®sh retina, J. Physiol., 207 (1970) 623±633. [8] Kaneko, A. and Tachibana, M., Effects of g-aminobutyric acid on isolated cone photoreceptors of the turtle retina, J. Physiol., 373 (1986) 443±461. [9] Kyrozis, A. and Reichling, D.B., Perforated-patch recording with gramicidin avoids artifactual changes in intracellular
[10] [11]
[12] [13] [14]
[15]
[16] [17] [18] [19]
[20]
chloride concentration, J. Neurosci. Methods, 57 (1995) 27± 35. Lasansky, A., Synaptic action mediating cone responses to annular illumination in the retina of the larval tiger salamander, J. Physiol., 310 (1981) 205±214. Lu, J., Karadsheh, M. and Delpire, E., Developmental regulation of the neuronal-speci®c isoform of K-Cl cotransporter KCC2 in postnatal rat brains, J. Neurobiol., 39 (1999) 558± 568. Marc, R.E., The structure of vertebrate retinas, In Toyoda, J., Murakami, M., Kaneko, A., Saito, T. (Eds.), The Retinal Basis of Vision, Elsevier, Tokyo, 1999, pp. 3±19. Piccolino, M. and Gerschenfeld, H.M., Characteristics and ionic processes involved in feedback spikes of turtle cones, Proc. R. Soc. Lond. B, 206 (1980) 439±463. Plotkin, M.D., Snyder, E.Y., Hebert, S.C. and Delpire, E., Expression of the Na-K-2Cl cotransporter is developmentally regulated in postnatal rat brains: a possible mechanism underlying GABA's excitatory role in immature brain, J. Neurobiol., 33 (1997) 781±795. Rivera, C., Voipio, J., Payne, J.A., Ruusuvuori, E., Lahtinen, H., Lamsa, K., Pirvola, U., Saarma, M. and Kaila, K., The K 1 / Cl 2 co-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation, Nature, 397 (1999) 251±255. Satoh, H., Aoki, K., Watanabe, S.-I. and Kaneko, A., L-type calcium channels in the axon terminal of mouse bipolar cells, NeuroReport, 9 (1998) 2161±2165. Schwartz, E.A., Depolarization without calcium can release g-aminobutyric acid from a retinal neuron, Science, 238 (1987) 350±355. Suzuki, S., Tachibana, M. and Kaneko, A., Effects of glycine and GABA on isolated bipolar cells of the mouse retina, J. Physiol., 421 (1990) 645±662. Tachibana, M. and Kaneko, A., g-aminobutyric acid acts at axon terminals of turtle photoreceptors: difference in sensitivity among cell types, Proc. Natl. Acad. Sci., 81 (1984) 7961±7964. Vardi, N., Zhang, L.L., Payne, J.A. and Sterling, P., Evidence that different cation chloride cotransporters in retinal neurons allow opposite responses to GABA, J. Neurosci., 20 (2000) 7657±7663.
www.elsevier.com/locate/neulet
Intracellular chloride concentration is higher in rod bipolar cells than in cone bipolar cells of the mouse retina Hiromasa Satoh*, Makoto Kaneda, Akimichi Kaneko Department of Physiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan Received 9 July 2001; accepted 17 July 2001
Abstract Bipolar cells (BCs) have antagonistic center-surround receptive ®eld. Surround illumination evokes depolarization in the OFF-type cone BC, and hyperpolarization in the rod BC and the ON-type cone BC. Surround illumination reduces gamma-aminobutyric acid (GABA) release from horizontal cells. If GABA hyperpolarize BCs, the polarity of the GABAinduced effect agrees with the light-evoked surround response in the OFF-type BC, but contradicts in the rod BC and the ON-type cone BC. Immunohistochemical study of the Cl 2 transporter of BCs has suggested that the intracellular Cl 2 concentration is different among BC subtypes. We examined the reversal potential of GABA-induced current of BCs using gramicidin-perforated patch clamp technique in the mouse retina, and found that GABA depolarizes rod BC and hyperpolarizes cone BCs. Our results are consistent with the GABAergic input to rod BC dendrite. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Bipolar cells; Perforated patch; Gamma-aminobutyric acid; Intracellular chloride; Retina; Mouse
Bipolar cells (BCs) in the vertebrate retina receive gamma-aminobutyric acid (GABA)ergic inputs from horizontal cells at the dendrite and from amacrine cells at the axon terminal. In the outer plexiform layer (OPL) of the retina, darkness depolarizes horizontal cells and thereby increases the GABA release [17]. In BCs, the direct glutamatergic input from photoreceptors generates the center response, while the GABAergic input from horizontal cells is considered to produce the surround response. Since increase in Cl 2 conductance by GABA is thought to hyperpolarize BCs, the reduction of GABAergic input by surround illumination is thought to produce depolarization. This idea explains the depolarization of OFF-type BCs by surround illumination. On the other hand, surround illumination induces hyperpolarization in ON-type BCs [7]. This is contradictory to what we expect in the horizontal to bipolar feed-forward connection. It has been a puzzle how the surround hyperpolarization in ON-type BCs is accounted for in relation to GABA released from horizontal cells. It is generally thought that the intracellular chloride concentration ([Cl 2]i) is regulated primarily by two chloride transporters, an Na-K-Cl cotransporter (NKCC) that * Corresponding author. Tel.: 181-3-5363-3749; fax: 181-33359-0437. E-mail address: [email protected] (H. Satoh).
normally accumulates Cl 2 [14], and a K-Cl cotransporter (KCC2) that normally extrudes Cl 2 [11,15]. Immunohistochemical localization of Cl 2 transporters in the mammalian retina revealed that dendrites of the OFF-type BC express KCC2 and the dendrites of rod and ON-type BC express NKCC [20]. They hypothesized that [Cl 2]i of the rod BC and the ON-type BC is higher than that of the OFF-type BC. The aim of the present study is to examine the above hypothesis by comparing ECl in three types of BCs using gramicidin-perforated patch clamp technique [9]. We found that ECl of rod BCs is statistically more positive than ECl of cone BCs of both ON-type and OFF-type. All experiments were performed on retinal slices prepared from mice (C57BL/6J, 6±8-weeks-old, male). The care of the animals was in accordance with the Guidelines for the Care and Use of Laboratory Animals, Keio University School of Medicine, and our experiments were approved by the University Animal Welfare Committee. For retinal slices preparation (thickness, 150 mm) we followed the method described in detail previously [16]. Slices were continuously superfused with oxygenated external solution of the following composition: (in mM) NaCl 135; KCl 5; CaCl2 2; MgCl2 1; HEPES 5 and glucose 10 (pH 7.4). The pipette solution contained (in mM): KCl 120; CaCl2 0.5; MgCl2 1; EGTA 5; HEPES 10; ATP±tris 5 (pH 7.2); Lucifer yellow (0.2 %) and gramicidin D (14.8 mg/ml). Lucifer
0304-3940/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S03 04 - 394 0( 0 1) 02 12 0- 6
162
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
yellow and gramicidin D were added immediately before use. The resistance of the ®lled pipette was 10±13 MV. Pipettes were connected to an EPC-9 patch-clamp ampli®er (HEKA Elektronik). The holding potential was corrected for the liquid junction potentials measured by the method previously reported [8]. Current and voltage signals were recorded on a thermal arraycorder (WR8500, Graphtec) and stored on a digital tape (PC204Ax, Sony). Sampled signals were analyzed of¯ine. GABA (Sigma) was applied by a puffer system (applied pressure, 60 kPa). Most experiments were carried out in room temperature (20± 258C, n 28). To examine whether the Cl 2 transporter was temperature sensitive, we carried out some recordings at 31±338C (n 6). Because no signi®cant difference in ECl was found, we also used the data obtained under the two temperature conditions. After recording, all recorded cells were ®lled with Lucifer yellow by rupturing the perforated patch membrane. BCs were morphologically classi®ed into rod BC (n 18), ONtype cone BC (n 13) or OFF-type cone BC (n 3). As demonstrated in the photomicrograph of a representative cell of each group (Fig. 1A), BC subtypes can be identi®ed by the location of the cell body, the position of the axon terminal in the inner plexiform layer (IPL), and the morphology of the axon terminals [3]. During recording, cells were visually monitored under the ¯uorescent microscope. If the cell was found stained during recording, we judged that the patch membrane was broken and the data from this cell were discarded. Cells which lacked dendrite and/or axon terminals were excluded from the analysis. GABA applied to OPL induced an outward current in a rod BC held at 255 mV under the gramicidin-perforated mode (Fig. 1Ba). Since the gramicidin-ionophore selectively permeates monovalent cations and is impermeable for anions [9], [Cl 2]i of the BC in situ was estimated by the reversal potential of GABA-induced current. Under the gramicidin-perforated voltage clamp condition, the reversal potential was 245 mV (®lled circle, Fig. 1C). The [Cl 2]i estimated by the Nernst equation was 19.1 mM. Immediately after this recording, the membrane patch at the pipette tip was ruptured and the recording con®guration was switched into the whole cell clamp mode. Under this new condition the polarity of GABA-induced current became inward within 10 s after membrane rupture (data not shown). The change is accounted for by the diffusion of Cl 2 from the recording pipette to the ruptured BC, resulting in an increase in [Cl 2]i. Under the gramicidin-perforated recording condition, application of GABA induced an inward current in both ON-type and OFF-type cone BCs held at 255 mV (Fig. 1Bb,Bc). The reversal potential was 267 mV for the ONtype cone BC (open circle, Fig. 1C), and 265 mV for the OFF-type cone BC (cross, Fig. 1C). Statistically the reversal potential of GABA-induced current differed between the rod BCs and the cone BCs (Student's t-test, P , 0:001) (Fig. 2A); 239 ^ 10 mV (mean ^ SD) in rod BCs,
255 ^ 10 mV in ON-type cone BCs and 267 ^ 7 mV in OFF-type cone BCs. The difference between two types of cone BCs was statistically insigni®cant. The estimated average [Cl 2]i was 23.4 mM in the rod BC, 12.6 mM in the ONtype cone BC and 7.9 mM in the OFF-type cone BC. Based on the polarized distribution of KCC2 and NKCC in rod BCs, Vardi et al. [20] hypothesized that there is a
Fig. 1. (A) Photomicrographs of a rod BC (a), an ON-type cone BC (b) and an OFF-type cone BC (c). The image was captured by a CCD camera (M-3204C, Olympus) and processed by `Photoshop' software (version 5, Adobe). We focused on axon terminal to show the characteristic shape, so the soma was out of focus in (b). The position of layers labeled in the left is only approximation because slices have different thickness. INL, inner nuclear layer; GCL, ganglion cell layer. (a) and (b) represent sublamina a and b of IPL. Other abbreviations are de®ned in the text. Scale bar: 10 mm. (B) GABA-induced responses recorded under the gramicidin-perforated patch clamp condition of (a) a rod BC, (b) an ON-type cone BC and (c) an OFF-type cone BC. Holding voltage was 255 mV in all cells. Horizontal bar above each trace indicates GABA application (10 mM) by puffer pipette to OPL. (C) I-V relationships of GABA-induced current recorded in the same cells under the perforated-patch clamp condition (®lled circle, rod BC; open circle, ON-type cone BC; cross, OFF-type cone BC).
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
gradient of [Cl 2]i in the rod BC, high in the dendritic region and low in the axon terminal. To examine this hypothesis we assessed [Cl 2]i at the dendrite and the axon terminal in the same cell. Since high sensitivity regions to GABA at OPL and that at IPL were separated by a low sensitivity region as
Fig. 2. Reversal potential of GABA-induced current recorded under the gramicidin-perforated patch clamp con®guration. (A) Difference of reversal potential among subtypes. Each symbol represents reversal potential of individual BC. Reversal potential was obtained by GABA application (10 mM) to OPL. Circle represents rod BC, triangle ON-type cone BC and rhombus OFF-type BC. (B) Relationship between the reversal potential of the current responses to GABA given to OPL and that given to IPL. Circles represent rod BC (n 15) and triangles ON-type cone BC (n 9). The solid line representing rod BCs and the broken line representing ON-type cone BCs were drawn by the least square method.
163
has been reported in isolated BCs [18], we assumed that GABA applied to the scleral margin of OPL mainly activated dendritic GABA receptors, while GABA applied to IPL activated GABA receptors at the axon terminal. Data points of Fig. 2B represent the reversal potentials of a BC obtained by applying GABA to OPL (abscissa) and to IPL (ordinate). Although the data points representing rod BCs and those representing ON-type cone BCs make separate clusters because the reversal potentials are different, all data points fall close to the line with a unit slope (0.93 for rod BC and 0.91 for the ON-type cone BC). We were unable to ®nd evidence that [Cl 2]i is different between the dendrite and the axon terminal. It may re¯ect the rapid diffusion of Cl 2, since we observed ¯uorescent images of the axon terminal and the dendrite within 2±3 s after membrane rupture at the soma. We found that [Cl 2]i in rod BCs was higher than [Cl 2]i in cone BCs. In the hippocampal neuron [2] and retinal ganglion cells [5,6], it has been shown that the polarity of the GABA-induced response reverses from depolarization in neonate to hyperpolarization in adult, suggesting that [Cl 2]i decreases during development. Probably, switch of the Cl 2 transporter from NKCC to KCC2 is responsible for the developmental change of [Cl 2]i [11,14,15]. NKCC maintains [Cl 2]i high [14], while KCC2 functions to keep [Cl 2]i low [11,15]. Higher [Cl 2]i of rod BCs (23.4 mM) than that of OFF-type cone BCs (7.9 mM) is in parallel with the type of the Cl 2 transporters expressed in these BCs [20]. Although the immunohistochemical data suggest high [Cl 2]i of ON-type cone BCs, it was 12.6 mM. A possible interpretation is that expression of NKCC in rod BCs is stronger than that in ON-type cone BCs. Since the resting potential of BCs was approximately 245 mV, GABAergic input to BCs is depolarizing in rod BCs, while that is hyperpolarizing in cone BCs. The present observation is in favor of the idea that horizontal cells exert GABAergic feed-forward action to rod BCs. Formation of the surround response in ON-type cone BCs is still unanswered by the feed-forward input from horizontal cells. It may be accounted for rather by the feedback from cone horizontal cells to cones [1,10,13]. Cone horizontal cells are GABAergic [12], and release GABA upon depolarization [17]. Cone photoreceptors have high sensitivity to GABA at the axon terminal [19]. Surround illumination reduces the amount of GABA release from horizontal cells resulting in the reduction of GABA-induced feedback to cone photoreceptors. This, in turn, increases glutamate release from cones and glutamate hyperpolarizes ON-type BC. It is unlikely that such a feedback system exists between horizontal cells and rods [4]. Furthermore, rods have an extremely low sensitivity to GABA at their axon terminal [19]. All these observations are in favor of supporting the notion that receptive ®eld surround is formed by different mechanisms between ON-type cone BCs and rod BCs; feedback from horizontal cells to cones in ON-type cone BCs,
164
H. Satoh et al. / Neuroscience Letters 310 (2001) 161±164
and feedforward action from horizontal cell to BCs in rod BCs. This work was supported by grant in aid to A. Kaneko (Nos. 11480247, 12053260) from the Ministry of Education, Science and Culture and to H. Satoh from Keio University Grant-in-Aid for Encouragement of Young Medical Scientists, and from the Promotion and Mutual Aid Corporation for Private Schools of Japan. We thank W.R. Taylor for providing software, S. Miyoshi and A. Koizumi for their technical advice. [1] Baylor, D.A., Fuortes, M.G.F. and O'Bryan, P.M., Receptive ®elds of cones in the retina of the turtle, J. Physiol., 214 (1971) 265±294. [2] Ben-Ari, Y., Cherubini, E., Corradetti, R. and Gaiarsa, J.L., Giant synaptic potentials in immature rat CA3 hippocampal neurones, J. Physiol., 416 (1989) 303±325. [3] Euler, T. and WaÈssle, H., Immunocytochemical identi®cation of cone bipolar cells in the rat retina, J. Comp. Neurol., 361 (1995) 461±478. [4] Fain, G.L., Gerschenfeld, H.M. and Quandt, F.N., Calcium spikes in toad rods, J. Physiol., 303 (1980) 495±513. [5] Fischer, K.F., Lukasiewicz, P.D. and Wong, R.O.L., Agedependent and cell class-speci®c modulation of retinal ganglion cell bursting activity by GABA, J. Neurosci., 18 (1998) 3767±3778. [6] Huang, B. and Redburn, D.A., GABA-induced increases in [Ca 21]i in retinal neurons of postnatal rabbits, Vis. Neurosci., 13 (1996) 441±447. [7] Kaneko, A., Physiological and morphological identi®cation of horizontal, bipolar and amacrine cells in gold®sh retina, J. Physiol., 207 (1970) 623±633. [8] Kaneko, A. and Tachibana, M., Effects of g-aminobutyric acid on isolated cone photoreceptors of the turtle retina, J. Physiol., 373 (1986) 443±461. [9] Kyrozis, A. and Reichling, D.B., Perforated-patch recording with gramicidin avoids artifactual changes in intracellular
[10] [11]
[12] [13] [14]
[15]
[16] [17] [18] [19]
[20]
chloride concentration, J. Neurosci. Methods, 57 (1995) 27± 35. Lasansky, A., Synaptic action mediating cone responses to annular illumination in the retina of the larval tiger salamander, J. Physiol., 310 (1981) 205±214. Lu, J., Karadsheh, M. and Delpire, E., Developmental regulation of the neuronal-speci®c isoform of K-Cl cotransporter KCC2 in postnatal rat brains, J. Neurobiol., 39 (1999) 558± 568. Marc, R.E., The structure of vertebrate retinas, In Toyoda, J., Murakami, M., Kaneko, A., Saito, T. (Eds.), The Retinal Basis of Vision, Elsevier, Tokyo, 1999, pp. 3±19. Piccolino, M. and Gerschenfeld, H.M., Characteristics and ionic processes involved in feedback spikes of turtle cones, Proc. R. Soc. Lond. B, 206 (1980) 439±463. Plotkin, M.D., Snyder, E.Y., Hebert, S.C. and Delpire, E., Expression of the Na-K-2Cl cotransporter is developmentally regulated in postnatal rat brains: a possible mechanism underlying GABA's excitatory role in immature brain, J. Neurobiol., 33 (1997) 781±795. Rivera, C., Voipio, J., Payne, J.A., Ruusuvuori, E., Lahtinen, H., Lamsa, K., Pirvola, U., Saarma, M. and Kaila, K., The K 1 / Cl 2 co-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation, Nature, 397 (1999) 251±255. Satoh, H., Aoki, K., Watanabe, S.-I. and Kaneko, A., L-type calcium channels in the axon terminal of mouse bipolar cells, NeuroReport, 9 (1998) 2161±2165. Schwartz, E.A., Depolarization without calcium can release g-aminobutyric acid from a retinal neuron, Science, 238 (1987) 350±355. Suzuki, S., Tachibana, M. and Kaneko, A., Effects of glycine and GABA on isolated bipolar cells of the mouse retina, J. Physiol., 421 (1990) 645±662. Tachibana, M. and Kaneko, A., g-aminobutyric acid acts at axon terminals of turtle photoreceptors: difference in sensitivity among cell types, Proc. Natl. Acad. Sci., 81 (1984) 7961±7964. Vardi, N., Zhang, L.L., Payne, J.A. and Sterling, P., Evidence that different cation chloride cotransporters in retinal neurons allow opposite responses to GABA, J. Neurosci., 20 (2000) 7657±7663.