Multiple GABA Receptor Subtypes Mediate Inhibition of Calcium Influx at Rat Retinal Bipolar Cell Terminals

The Journal of Neuroscience, April 1995, 75(4): 2668-2679

Zhuo-Hua Pan’ and Stuart A. Lipton* 1Department of Neurology, Children’s Hospital and Harvard Medical School, Boston, Massachusetts 02115 and *Department of Neurology, Children’s Hospital, Beth Israel Hospital, Brigham and Women’s Hospital, Massachusetts General Hospital, and Program in Neuroscience, Harvard Medical School, Boston, Massachusetts 02115

Inhibitory effects of GABA on K+-evoked Ca2+ influx into sensitiveto GABA (Tachibanaand Kaneko, 1987; Karschin and rat retinal bipolar cell terminals were studied using calcium Wbsle, 1990; Suzuki et al., 1990; Yeh et al., 1990; Heidelberger imaging methods. Application of high K+ evokes a sus- and Matthews, 1991). It is speculatedthat the input of GA- tained, reversible increase in [Ca2+li at bipolar cell termi- BAergic amacrinecells to bipolar cell terminalscreates negative nals, which occurs mainly via dihydropyridine-sensitive (L- feedback to generate transient responsesin the inner retina type) Ca2+ channels. There are at least two GABA receptor (Tachibanaand Kaneko, 1987; Maguire et al., 1989). subtypes coexisting at bipolar cell terminals: a convention- Recently, a novel bicuculline/baclofen-insensitiveGABA real GABA, receptor and a bicuculline/baclofen-insensitive ceptor has been reported on tectal neuronesof the frog (Nistri GABA receptor. Activation of either GABA receptor inhib- and Sivilotti, 1985; Sivilotti and Nistri, 1989) and retinal neu- ited the K+-evoked Ca*+ response. However, these two rons in several species(Feigenspan et al., 1993; Qian and Dowl- GABA receptor subtypes have distinct properties. GABA, ing, 1993; Dong et al., 1994; Lukasiewicz et al., 1994; Matthews receptors suppress the Ca*+ response only at relatively et al., 1994; Zhang and Slaughter, 1994). The newly cloned GA- high concentrations of agonist, and with fas’ kinetics and BAp subunits and GABAp-like responsesencoded by bovine a narrow dynamic range. In contrast, the bicucullinelbaclo- retinal mRNA, expressedin Xenopus oocytes, display similar fen-insensitive GABA receptors produce inhibition on the properties (Cutting et al., 1991; Polenzani et al., 1991; Shimada Caz+ response at a much lower concentration of agonist, et al., 1992; Kusamaet al., 1993b). This novel GABA receptor and with slow onset and a wider dynamic range. The phar- is commonly referred to as the GABA, receptor (Johnston, macologic profile of the bicucullinelbaclofen-insensitive 1986; Shimadaet al., 1992). Although a similar GABA receptor GABA receptor at bipolar cell terminals is most similar to has been found in several species,the GABA, receptor of rat the GABA, receptor reported by Feigenspan et al. (1993). bipolar cells has been reported to be insensitive to picrotoxin Unlike the GABA, receptors described in other species, it (Feigenspanet al., 1993). The implication of this discrepancy is is extremely insensitive to picrotoxin. Therefore, it may be not clear, since the pharmacologic properties and molecular appropriate to refer to this receptor as a picrotoxin-insen- identities of the putative GABA, receptor are still not fully un- sitive GABA, receptor. 3-Aminopropyl(methyl)phosphinic derstood. acid (3-APMPA) and 3-aminopropylphosphonic (3-APA), Very few studieshave examined selective antagonistsfor this two phosphate analogs of GABA, selectively antagonize newly describedGABA receptor or family of receptors.Several the picrotoxin-insensitive GABA, receptors but not the GA- phosphateanalogs of GABA, which are the GABA, receptor BA, receptors in this system. These results imply a func- agonistsor antagonists(Slaughter and Pan, 1992), and imidaz- tional role for multiple GABA receptors in regulating syn- ole-4-acetic acid (14AA), a GABA analog with an extendedcon- aptic transmission at bipolar cell terminals. formation (Kusamaet al., 1993; Qian and Dowling, 1994), have [Key words: GABA, GABA receptors, calcium, 3-amino- been reportedto display antagonistproperties to the bicuculline/ propyl(methyl)phosphinic acid, synaptic transmission, bi- baclofen-insensitiveGABA receptorsexpressed in Xenopus oo- polar cells, retina, rat] cytes (Kusamaet al., 1993a; Woodward et al., 1993). 14AA has also been reported to be an effective antagonistof the GABA, GABA is one of the major inhibitory neurotransmittersin the receptor in fish retinal horizontal cells (Qian and Dowling, CNS in general and in the retina in particular (Olsen and Venter, 1994). However, the properties of these compoundshave not 1986; Massey and Redburn, 1987; Tauck et al., 1988; Sivilotti been studied on the picrotoxin-insensitive type of GABA, re- and Nistri, 1991; Slaughter and Pan, 1992). There is abundant ceptorsfound in the rat retinal neurons. evidence that the terminals of retinal bipolar cells are highly Recent studieshave shown that GABA can act at multiple subtypesof GABA receptorson bipolar cells. Both GABA, and Received Aug. 24, 1994; revised Oct. 13, 1994; accepted Oct. 24, 1994. GABA, receptorshave been reported in retinal bipolar cells of We thank Dr. Malcolm M. Slaughter for helpful discussions and valuable comments on the manuscript. This work was supported in part by a Scholar goldfish, rat, tiger salamander,and white perch (Tachibanaand Award from the American Foundation for AIDS Research (to Z.-HI!) and by Kaneko, 1987; Feigenspanet al., 1993; Qian and Dowling, NIH Grant ROl EY05477 (to S.A.L.). 1993b; Lukasiewicz et al., 1994; Matthews et al., 1994). Pre- Correspondence should be addressed to Dr. Zhuo-Hua Pan at the above ad- dress. viously, a baclofen-sensitiveGABA, receptor has been reported Copyright 0 1995 Society for Neuroscience 0270-6474/95/152668-12$05.00/O in tiger salamanderbipolar cell terminals (Maguire et al., 1989). The Journal of Neuroscience, April 1995, 15(4) 2669

In goldhsh bipolar cell terminals, a baclofen-insensitive GA- saturate the fluorescence Ca 2+ indicator. Under our condition, 80 mM BA,-like receptor has also been reported (Heidelberger and Mat- K+ evoked an eouivalent or sometimes slightly smaller fluorescence signal than that eboked by 60 mM K+. However, 40 mM K’ evoked a thews, 1991; Matthews et al., 1994). The common effect of all consistentlv smaller fluorescence signal than 60 mM K+ (85% 2 3%. of these GABA receptors is, in one way or another, to regulate mean 2 SEM, n = 7). In order to-assure that the CaZ+ indicator was the Ca*+ response of the terminal. Although the inhibitory effect not saturated by [Ca*+],, 40 mM K+ was therefore used to evoke Ca2+ of GABA on Ca2+ influx into giant goldfish bipolar terminals responses throughout this work. Pharmacological agents were mixed with high K+ saline and applied by pneumatic (“puffer”) pipettes has been directly visualized with calcium imaging methods (Hei- placed about 3040 pm away from the bipolar cell terminals, unless delberger and Matthews, 1991; Matthews et al., 1994), a detailed otherwise specified. picture of the interaction of multiple GABA receptor subtypes Fluorescence measurement. Fluorescence measurements were per- in regulating Ca*+ responses at bipolar cell terminals, especially formed at room temperature on a confocal laser-scanning microscope in mammalian species, is not clear. Furthermore, the functional (NORAN, WI) coupled to an imaging system (Universal Imaging, PA). The video scan module (Odyssey) was mounted on an upright Nikon significance of the coexistence of GABA receptor subtypes at microscope fitted with a 40X, 0.75 NA water-immersion objective. Op- bipolar cell terminals remains unknown. tical excitation was accomplished using the 488 nm line of an argon In this study, by using calcium imaging, we demonstrated the laser. The emitted fluorescence passed through a 5 15 nm primary barrier colocalization of multiple GABA receptor subtypes at rat bipolar filter before it reached the photomultiplier tube. The laser intensity was minimized to prevent dye bleaching during the course of the experi- cell terminals. Particularly, we further characterized the phar- ments. During measurements, the fluorescence images were acquired in macologic properties of the GABA, receptor and the dynamic real-time mode (30 frames/set), and the fluorescence data at a defined course of inhibition resulting from multiple GABA receptor sub- bipolar cell terminal were stored and later exported to a SIGMAPLOT types at rat bipolar cell terminals. Previous studies have mainly graphics program. In some cases, the real-time video images were di- focused on the responses recorded from the bipolar cell soma. rectly recorded onto an optical memory disc recorder (OMDR, Pana- sonic) for later analysis. An advantage here is that the responses of synaptic terminals Data analysis. Since fluo-3 is a single-wavelength chromophore, the could be directly visualized. fluorescence is a function of the [Ca*+], as well as the dye concentration. A preliminary report of part of this work has appeared (Pan However, for a fixed region, the change of fluorescence directly reflects and Lipton, 1994b). the change of [Ca*+],. Therefore, data are not calibrated to absolute values of ]Ca2+],, but are displayed as a ratio of fluorescence change. Materials and Methods For the purpo&of clarity, we will use the term “calcium” instead of “fluorescence” throughout the uauer. A total of 826 biuolar cell ter- Bipolar cell isolation. Bipolar cells were dissociated from 3-4 week minals were studied h this report. Each result reported&in this paper postnatal Long-Evans rats. The dissociation and culture methods are was based on the observations obtained from at least four bipolar cell similar to those previously published, with minor modifications (Leifer terminals. et al., 1984; Lipton and Tauck, 1987). In brief, after decapitation of the Chemicals. 3-Aminopropylphosphinic acid (3-APPA), zrurzs-, and cis- animal, both retinas were removed and placed in a saline based upon 4-aminocrotonic acid (TACA-and^CACA) were purchased from Tocris Hanks’ balance salts (in mM): NaCl, 137; NaHCO,, 1; Na,HPO,, 0.34; Neuramin (Bristol, UK); 3-aminopropyl(methyl)phosphinic acid (3- KCl, 5.36; KH,PO,, 0.44; CaCl,, 1.25; MgSO,, 0.5; MgCl,, 0.5; HE- APMPA), 4,5,6,7-tetrahydroisoxazolo[5,4-clpyridin-3-01 (THIP), mus- PES-NaOH, 5; glucose, 22.2; with phenol red, 0.001% v/v; adjusted to cimol, SR95531, phaclofen, hydroxy-saclofen, and nimodipine from pH 7.2 with 0.3 N NaOH. The retinas were incubated for 50 min at Research Biochemicals Inc. (Natick. MA): GABA. (-)-bicuculline 37°C in 10 ml of an enzyme solution that consisted of the saline de- methobromide, picrotoxinin, strychnine,’ imidazole&cetic acid scribed above, supplemented with 0.2 mglml DL-cysteine, 0.2 mg/ml (14AA), piperidine-4-sulfonic acid (P4S), baclofen, and 3-aminopropyl- bovine serum albumin, and 3.5 U/ml papain, adjusted to pH 7.2 with phosphonic (3.APA) from Sigma (St. Louis, MO). 0.3 N NaOH. Following three rinses in saline, the retinas were me- chanically dissociated by gentle trituration with a glass pipette. The Results resulting cell suspension was plated onto poly-r-lysine-coated glass coverslips. The culture medium was comprised of Eagle’s essential me- Characterization of calcium responses at bipolar cell dium supplemented with 2 mu glutamine, 1 kg/ml gentamicin, 16 mM terminals glucose, and 5% rat serum. Cell cultures were incubated at 37°C in a As shown in Figure 1, A and B, a sustained increase in [Ca*+], 5% CO,/95% 0, humidified atmosphere for about 1 hr before dye loading. at bipolar cell terminals was elicited by applying high K+ (40 Dye loading. Prior to dye loading, the culture medium was replaced mM). There was no significant fading of the Ca*+ response dur- with Hanks’ solution. Then, cells were loaded with the membrane-per- ing >20 set of continuous measurement. After removing the meable fluorescence dye fluo-3/AM (Molecular Probes, OR) at a con- high K+, the [Ca*+], rapidly recovered to the basal level, dem- centration of 20 PM, supplemented with 0.0025% pluronic acid, in dark- onstrating the Ca 2+ buffering capacity of bipolar cell terminals ness at room temperature for 1 hr (Kao et al., 1989; Pan and Lipton, 1994a). After dye loading, the cells were washed with Hanks’ solution (Fig. 1B). to remove the remaining dye. Cells were kept in darkness for another To test whether the increase in [Ca*+], evoked by high K+ was hour. Before fluorescence measurements, a coverslip of cells was trans- due to influx of extracellular Ca2+, high K+ was applied in Ca2+- ferred to a glass-bottomed chamber on the stage of a confocal micro- free saline containing 5 mM EGTA. Under these conditions, no scope. The cells were continuously superfused at a rate of about 1 ml/ min with a recording solution. The recording solution is the same as increase in [Ca2+], was observed (data not illustrated). Thus, the Hanks’ solution described above except that it contains 2.5 mu influx of extracellular Ca*+ is essential for a depolarization- CaCl,. evoked increase in [Ca*+],. Bipolar cell ident$cation. The intact bipolar cells could be easily To determine the subtype of Ca*+ channel responsible for the identified based on their characteristic morphology (Tachibana and Ka- sustained CaZ+ response in bipolar cell terminals, we used ni- neko, 1987; Karschin and Wassle, 1990; Yeh et al., 1990). We did not make efforts to identify the subtype of bipolar cells, but we preferen- modipine, a specific L-type Ca2+ channel blocker. The Ca*+ re- tially selected bipolar cells with long axons. Based on previous reports, sponse was substantially suppressed when 4 PM nimodipine was virtually all of these dissociated bipolar cells are rod bipolar cells (Kar- coapplied with high K+ in the pneumatic pipette and included schin and Wlssle, 1990). in the bath solution (Fig. 1 C). Therefore, most of the Ca*+ influx High potassium application and drug delivery. Depolarization was evoked by high K+ saline, formulated by replacement of equimolar Na+ evoked by high K+ appears to occur via an L-type CaZ+ channel. in the recording solution. A number of K+ concentrations were tested Further characterization of the Ca*+ channels at rat bipolar cell to determine the proper K+ to evoke a strong Ca2+ signal but not to terminals was beyond the scope of this work. 2670 Pan and Lipton * GABA Modulation Calcwm Response

Inhibitory eflect qf GABA on calcium ivlflux at bipolar cell terminals involves at least two components

Application of GABA itself at concentrations of l-10 FM did not produce any increase in [Cal+], at bipolar cell terminals (data not shown). When high K’ and GABA were coapplied, as shown in Figure 2, GABA inhibited the K+-evoked Ca?’ response in a dose-dependent fashion. GABA produced a delayed, sustained inhibition on the Ca*+ response at submicromolar concentrations (Fig. 2A,B). GABA at concentrations of 1 pM or less, and usually at 2 pM, did not significantly suppress the initial peak of the Ca*+ response, but rather produced a delayed, sustained inhibition (Fig. 2A-D). This is demonstrated by compar- ing the Ca2+ responses evoked by high K+ in the presence of 1 or 2 FM GABA with those evoked by high K+ alone at the same terminals, as shown in Figure 2, C and D. In addition, GABA at a concentration of l-2 pM, but not at submicromolar concentrations, frequently reduced the rate of rise of the Ca*+ response (Fig. 2A-D). Bicuculline, however, prevented this effect (see Fig. 3A,B). GABA at 5 pM or higher concentrations always completely suppressed the Ca*+ response (Fig. 2E). Bicuculline 6B 40mMK+ (100 FM) partially eliminated the suppression of the early part 5- of the Ca2+ increase produced by 5 pM GABA, but did not block the delayed, sustained suppression (Fig. 2F). These results indicate that the suppression of the initial increase in [CaZ+], was, at least partially, due to the activation of bicuculline-sensitive GABA receptors. The delayed, sustained inhibitory effect of GARA at a concentration of 1 pM was not significantly affected by bicuculline (100 FM), as demonstrated by measuring Ca2+ responses at the same terminals in the presence and absence of bicuculline (Fig. 3A,B). This indicates that GABA at 51 FM does not significantly activate GABA, receptors during its delayed inhibition I I I I of Ca*+ response. However, bicuculline did increase the initial rise time of the CaZ+ response (Fig. 3A,B), consistent with the notion that the initial phase of GABA’s action was mediated by GABA, receptors, as mentioned above. Furthermore, the inhib- :Oli itory effect of GABA at 1 FM concentration was not significantly affected by coapplication or bath application of 100 pM 4 bicuculline plus 200 pM picrotoxinin (Fig. 3C). SR95531 (10 PM), a more potent GABA, receptor antagonist, and strychnine PM), 3 (1 a glycine receptor antagonist, did not block the inhibitory effect of GABA (Fig. 3D,E). Baclofen (100 FM) did not mimic GABA’s inhibitory effect (Fig. 3F), and GABA, receptor 2 antagonists, such as phaclofen (100 pm) and hydroxy-saclofen (100 PM), did not block the action of GABA (data not illustrat- 1 ed). ii Therefore, both conventional GABA, and bicuculline/baclo- 01 I I fen/picrotoxin-insensitive GABA receptors are involved in the 0 5 10 15 inhibition of the Ca*+ response at rat bipolar cell terminals. Time (s) These results are consistent with the earlier demonstration of a picrotoxin-resistant GABA, receptor in recordings from rat bi- Figure 1. Potassium-evoked increase in [Caz+], at bipolar cell termi- polar cell somas (Feigenspan et al., 1993). nals. A, Fluorescence, pseudocolor images of a representative bipolar cell before and during application of 40 mM K+. The horizontal scale The picrotoxin-resistant property of this receptor on rat bi- bar represents 5 pm. The fluorescence (left vertical bar) is an arbitrary polar cells is clearly different from that of GABA, receptors 256 point gray scale converted to pseudocolor. B, Time course of flu- reported in retinal neurons of other species. Therefore, we tested orescence measured at another bipolar cell terminal. High K’ (40 mM for a picrotoxin-sensitive but bicuculline-insensitive component. and the same in all figures) application is indicated by a dark bar above We compared the time course of K+-evoked Ca*+ responses at the trace. The vertical axis is the fluorescence normalized to that of the basal level. C, Nimodipine (4 FM) suppressed the majority of the Ca2+ a number of GABA concentrations in the presence of bicuculline response. In this case, 4 PM nimodipine was also included in the bath (100 pM) with those elicited in the presence of bicuculline (100 solution. pM) and picrotoxinin (200 PM). At relatively high GABA concentrations, for example 5 pM as shown in Figure 4, A and B, The Journal of Neuroscience, April 1995, 15(4) 2671

0.2 PM GABA + high K+ iB 0.5 /IM GABA + high K+

6- 5-

4- 3- 2- 1 -

7 -c high K+ 6 1 PM GABA + high K+

Figure 2. Dose-dependent GABA suppression of the K+-evoked Ca2+ response and demonstration of the in- 6 F high KC volvement of multiple GABA receptors. GABA was coapplied with high 5- 5 PM GABA + high Kc K+. A-D, GABA at concentrations of 100 PM bicuculline 1 or 2 FM did not significantly suppress 4 - the initial peak of Ca*+ response but produced a delayed, sustained inhibition and slightly reduced the rate of initial Ca2+ rise. E, Five micromolar GABA totally suppressed the Ca?+ response. F, When 100 p,M bicuculline was coapplied with 5 FM GABA and high K+, an initial Ca2+ increase was observed. In C-F, a second Ca2+ response evoked by high K+ alone was OL I I I 0’ I I I equivalent to the control (Fig. I). These 0 5 10 15 0 5 10 15 control Caz+ responses were measured at the same terminal about 1 min after Time (s) the initial measurements.

200 PM picrotoxinin did not show any significant effect on the the GABA, receptor. 3-APMPA by itself, at concentrationsas time courseof the Ca2+response. At submicromolarGABA con- high as 5 mM, did not inhibit K+-evoked Ca2+responses (Fig. centrations,for example 0.2 and 0.5 pM, 200 PM picrotoxinin 5A), but the delayed, sustainedinhibitory effect producedby 1 partially blocked GABA’s inhibitory effect (Fig. 4C-F). How- PM GABA was completely blocked by 200 PM 3-APMPA (al- ever, our data cannot resolve whether picrotoxinin blocks an though inhibition of the initial Ca*+rise was still present) (Fig. additional componentof delayed GABA inhibition or represents 5B). This finding is consistentwith our observationsthat GABA a weak antagonistfor a single delayed component. Feigenspan at this concentration mainly activates GABA, receptors,and the et al. (1993) reported that picrotoxin blocks only a small per- only visible inhibition due to the activation of the GABA, re- centageof the GABA current in rat bipolar cells. Taken together, ceptor was on the very initial Ca” rise (compareFig. 3A,B). As these results indicate that the bicuculline/baclofen-insensitive shown in Figure X, the inhibitory effect of GABA at higher GABA receptor in rat bipolar cell terminals is extremely picro- concentrations(5 PM), which also activate GABA, receptors toxinin insensitive. (compare Fig. 2E,F), was not blocked by 200 FM 3-APMPA. Consistent with this interpretation, a combination of 100 p,M GABA, receptor antagonists bicuculline and 200 FM 3-APMPA totally blocked the inhibitory We found that micromolar 3-APMPA, a potent GABA, agonist, effect of high concentrationsof GABA (10 FM) on the Ca2+ was able to block the inhibition of Ca2+responses mediated by response(Fig. 5D). 2672 Pan and Lipton - GABA Modulation Calcium Response

B 1 ,uM GABA + high K+ 1 PM GABA + high Kf 100 PM bicuculline

D 1 PM GABA + high K+ 6 10 /LM SR95531

7 6- E 1 /IM GABA + high K+ 1 PM strychnine 100 PM baclofen + high K+ Figure 3. Demonstration that the de- 5- 6- layed, sustained inhibitory effect of I 5- pM GABA was not significantly affected by 100 PM bicuculline (A, B), 4- 100 FM bicuculline plus 200 PM picrotoxinin (C), 10 PM SR95531 (D), or 1 3- FM strychnine (E). One hundred micromolar baclofen did not mimic the 2- inhibitory effect produced by GABA (F). However, the initial rate of rise of 1 - [Ca?+], was faster than that in the absence of bicuculline (compare A and 0 I I B). In A and B, the measurements were 0 5 10 15 performed at the same bipolar cell terminals. Time (s)

To confirm further that 3-APMPA is a GABA,, but not GA- analogsof GABA, and 14AA (imidazole-4-aceticacid), a GABA BA,, antagonist in our preparation, we tested the effects of analog with an extendedconfirmation. We found that 3-APA by 3-APMPA on two other putative GABA, agonists,THIP (4,5,6,7- itself at a concentration of 500 p,M did not show any inhibition tetrahydroisoxazolo[5,4-clpyridin-3-01)and P4S (piperidine-4-s& on the Ca*+response (data not shown).Like 3-APMPA, 3-APA fonic acid). Thesetwo GABA analogshave beenreported to dis- selectively blocked the GABA, but not the GABA, effect on play agonistproperties only at GABA, receptors(Kusama et al., Ca2+responses, although it was lesspotent than 3-APMPA (data 1993a; Woodward et al., 1993; Qian and Dowling, 1994). Con- not illustrated). In addition, 3-APPA (200 PM) and I4AA (100 sistentwith thesereports, the inhibitory effects of THIP (100 FM) FM) displayed antagonistic effects on the GABA, receptor in and P4S (100 PM) were blocked by bicuculline (100 PM) (Fig. our preparation(data not shown). However, both of thesedrugs 6A-D). However, 3-APMPA, at concentrationsas high as 500 pM, are also GABA, receptor agonists at these concentrations. did not block this inhibition (Fig. 6E,F). Therefore, they had to be administeredwith GABA, antagonists We also tested 3-APA (3-aminopropylphosphonicacid) and to enable us to observe their inhibition on the GABA, receptor 3-APPA (3-aminopropylphosphinicacid), two other phosphate in our preparation(data not illustrated). The Journal of Neuroscience, April 1995, 75(4) 2673

100 pM bicuculline 100 pM bicuculline 200 uM Dicrotoxinin B 6 c * 5 /IM GABA + high K+ 6‘I 5 pM GABA + high K+

0 ' I / I I I 0 1 I I I I 0 2 4 6 8 10 0 2 4 6 8 10 12 I C 0.5 pM GABA + high K+ 1. ID 0.5pM GABA + high K+ 1 10 c

6 -

0 ’ I I 0 I I 0 5 10 15 0 5 10 15

76 - E 0.2 pM GABA + high K+ 0.2 PM GABA + high K+ Figure 4. High concentrations of picrotoxinin only slightly blocked the inhibitory effect produced by submicromolar concentrations of GABA. The same bipolar cell terminals were tested in the presence of 100 pM bicuculline (A, C, E) versus 100 FM bicuculline plus 200 PM picrotoxinin (B, D, F) at each GABA concentration. Bicuculline or bicuculline plus picrotoxinin were coapplied with GABA and high K+. Pi- I I crotoxinin (200 FM) did not signifi- 0 cantly block the inhibitory effect pro- 0 5 10 15 duced by 5 PM GABA (A, B), and only slightly blocked the mhlixtory effect by Time (s) 0.2 FM and 0.5 FM GABA (C-F).

GABA, receptor agonists effect of 200 pM CACA was totally blocked by 200 FM A number of GABA analogs, such as TACA (trans-aminocro- 3-APMPA (Fig. 7F), indicating that CACA at this concentration tonic acid), CACA (cis-aminocrotonic acid,) and muscimol, was selective for the GABA, receptor. However, we found that have been reported to activate the GABA, receptor (Feigenspan, the effect of 500 pM CACA was not blocked by 3-APMPA (500 1993; Qian and Dowling, 1993; Dong et al., 1994; Lukasiewicz pM) but was inhibited by a combination of 3-APMPA (500 FM) et al., 1994; Matthew et al., 1994; Zhang and Slaughter, 1994). and bicuculline (100 PM) (Fig. 7G,H). This finding can be best Consistent with these reports, we found that TACA exhibited a rationalized by the notion that at such a high concentration, our potency similar to GABA (Fig. 7A). Muscimol was also effec- preparation of CACA activates both GABA, and GABA, re- tive but less potent than GABA. As shown in Figure 7, B and ceptors. c, muscimol at a concentration of 5 FM, but not 2 PM, mimicked the effect of submicromolar or low micromolar GABA (Fig. 2B- Distinct efects of GABA, versus GABA, receptors on calcium D). CACA had a similar effect but required much higher con- responses centrations. In the presence of 100 FM bicuculline, 100 FM The selective antagonism of 3-APMPA and 3-APA on the GA- CACA produced incomplete inhibition of slow onset (Fig. 70). BA, receptor made it possible to separate the effects of the GA- The concentration of CACA had to be raised to 200 FM to BA, receptor from that of the GABA, receptor, and therefore to totally suppress the Cal+ response (Fig. 7E). Furthermore, the compare the time courses of their inhibitory effects on Ca2+ 2674 Pan and Lipton * GABA Modulation Calcium Response

200 /JM 3-APMPA

0 5 10 15 0 5 10 15

5 ,uM GABA + high K+ D 10 PM GABA + high Kf Yc 200 pM 3-APMPA 10 200 pM 3-APMPA 6 1 a - 100 PM bicuculline Figure 5. 3-APMPA selectively blocks the GABAc receptor.A, 3- 5 APMPA (5 mM) did not inhibit the 6 - Ca?+ response to K’. B, 3-APMPA 4 (200 FM) completely blocked the delayed, sustained inhibition of the Ca?+ response by I pM GABA (but did not prevent the slight reduction in rise time of the Ca?+ response). C, 3-APMPA (200 FM) did not block the inhibitory FM D, effect by 5 GABA. The com- 0 5 10 15 -0 5 10 15 bination of 200 FM 3-APMPA and 100 PM bicuculline completely blocked the inhibitory effect by 10 pM GABA. Time (s) responses. Paired measurements of K+-evoked CaZ+ responses is the first calcium imaging study carried out in mammalian in the presence of bicuculline/picrotoxinin versus 3-APMPA retinal bipolar cell terminals. Moreover, we were able to char- were performed on the samebipolar cell terminal for a rangeof acterize the dynamics of the Ca2+ responseat bipolar terminals, GABA concentrations (Fig. 8). Several distinct properties of which have been shown to directly correlate with the releaseof GABA, and the GABA, receptors were observed. First, an in- neurotransmitters(Tachibana et al., 1993; Gersdorff and Mat- hibitory effect on the GABA, receptor was observed at GABA thews, 1994). concentrationsas low as 0.2 pM (Fig. 4E). However, GABA concentrationsas high as 20 FM were neededto fully suppress Properties of calcium responsesat bipolar cell terminals the Ca*+ response by activating this receptor (Fig. 8G). In con- Our studiesshow that high K+ evokes a sustainedCa* + response trast, the inhibitory effect of the GABA, receptor was virtually at bipolar cell terminals.The increasein [Ca*‘1, was due at least absent with l-2 pM GABA (Fig. 8B, except for a slight reduc- in part to Ca2+ influx upon K+ depolarization (we cannot exclude tion in the initial rise time of the Ca?+response;compare Figs. the possibility of Ca*+-induced intracellular Ca2+release). We 3A,B; 5B). Although GABA is more potent at the GABA, re- also demonstratedthat the influx of Ca*+ occurs mainly via di- ceptor, the onset of the inhibition is slower. As shownin Figure hydropyridine-sensitive (L-type) Ca’+ channels, since nimodi- 8E, even at concentrationsof 10 FM, there was still an initial pine, a specific L-type Ca*+ channel antagonist, eliminated most Cal+ response.This contrastswith the effect of the GABA, re- of the K+-evoked Ca2+response. This result is consistentwith ceptor, where 25 pM GABA always completely suppressedthe the sustainednature of the Ca2+response. It is interestingto note early part of Ca’+ response(Fig. 8D). Finally, when the GABA, that in mousebipolar cells only a transient(T-type) Caz+channel component was blocked by 3-APMPA, the Ca2+ response in the wasobserved, while in goldfish bipolar cells a sustained(L-type) presenceof relatively low concentrationsof GABA (5-10 pM) Ca*+ channel accountedfor most of the Ca*+influx (Kaneko et usually displayed a prominent rebound. The recovery of Ca*+ al., 1989; Heidelberger and Matthews, 1992; Tachibanaet al., responsesis not likely due to GABA, receptor desensitization. 1993). The reason is that GABA at higher concentrations (20 pM) was found to eliminatethis recovery (Fig. 8H), even when millimolar Pharmacology of the GABA, receptor of rat bipolar cell concentrationsof 3-APMPA were used (data not shown). The terminals latter makesincomplete blockade of GABA, receptorsunlikely. Our results show that both GABA, and GABA, receptorsare involved in regulating the Ca*+ response of bipolar cell termi- Discussion nals. The pharmacologicprofile of the GABA, receptor of bi- In this study, using a high-resolutionconfocal imaging system, polar cell terminals to TACA and CACA is consistentwith the we demonstratethe inhibitory effect of GABA on K+-evoked findings of Feigenspanet al. (1993) using patch-clamp record- Ca’+ responsesof rat bipolar cell terminals. Calcium imaging ing. This profile is similar to the so-calledGABA, receptor in methodshave previously been usedto study Ca*+ responsesat fish and amphibian retinal cells (Qian and Dowling, 1993a, giant goldfish bipolar cell terminals (Heidelbergeret al., 1991, 1994; Dong et al., 1994; Lukasiewicz et al., 1994; Matthews et 1992; Tachibanaet al., 1993). However, to our knowledge, this al., 1994; Zhang and Slaughter, 1994). The Journal of Neuroscience, April 1995, X(4) 2675

A 100 yM P4S + high Kf

1 -m

O ’ , I I 0 I / 0 5 10 15 0 5 10 5 8 5 , C 100pM THIP + high KC D 100 PM P4S + high KC 7- 100 PM bicuculline 100 PM bicuculline 6-

5 -

7 E 100 PM THIP + high KC ’6 - F 100 ,uM P4S + high K+ 6 - 500 PM 3-APMPA 500 PM 3-APMPA 5 - 5-

Figure 6. 3-APMPA does not antagonize the inhibitory effects of THIP or P4S. A and B, One hundred micromolar THIP or 100 p.M P4S suppressed K+- evoked Ca’+ responses. C and D, One hundred micromolar bicuculline completely blocked the inhibitory effects of I I I THIP (100 FM) and P4S (100 FM). o ’ E and F, Five hundred micromolar 0 5 10 15 3-APMPA did not antagonize the inhibition produced by THIP (100 FM) Time (s) or P4S (100 FM).

In this study, we determined that 3-APMPA and 3-APA, two GABA,-like receptor expressed by GABAp subunits, which is phosphate analogs of GABA, selectively antagonized the GA- sensitive to picrotoxin, was actually cloned from the human ret- BA, receptor but not the GABA, receptor. Our results are con- ina (IC,, = 0.4 pM for 1 pM GABA) (Cutting et al., 1991, 1992; sistent with studies performed on the bicuculline/baclofen-insen- Shimada et al., 1992). The GABA,-like receptor expressed in sitive GABA receptor encoded by bovine retinal mRNA injected Xenopus oocytes with mRNA extracted from bovine retina is into Xenopus oocytes (Woodward et al., 1993). More recently, also sensitive to picrotoxin (IC,, = 0.5 pM for 1 pM GABA) the antagonist nature of 3-APA at the GABA, receptor also has (Polenzani et al., 1991; Woodward et al., 1992). Therefore, it is been reported in fish horizontal cells (Qian and Dowling, 1994). unlikely that the picrotoxin insensitivity is a unique property of The selective antagonist properties of 3-APMPA and 3-APA at the GABA, receptor in mammals. Whether this receptor is the GABA, receptor may be useful in the study of the functional unique to retinal bipolar cells in mammals is uncertain because role of GABA, receptor in retina. rat bipolar cells are the only mammalian retinal neurons that One unique property of the pharmacology of the rat bipolar have been studied so far in this manner. Further studies will be cell GABA, receptor is its picrotoxin insensitivity. The GABA, required to understand this apparent discrepancy and the under- receptor in our preparation is extremely resistant to picrotoxin. lying molecular mechanism. This clearly differentiates this GABA, receptor from that re- Bicuculline sensitivity has been used as a major criteria to ported in retinal horizontal and bipolar cells of fish and amphib- distinguish GABA, versus GABA, receptors (Johnston, 1986; ians (Qian and Dowling, 1993a, 1994; Dong et al., 1994; Lu- Shimada et al., 1992). With the similarity of the pharmacologic kasiewicz et al., 1994; Matthews et al., 1994). Interestingly, the profile of the GABA, receptor described here in rat to that of 2676 Pan and Lipton - GABA Modulation Calcium Response

1 ,LLMTACA + high K+ 100 PM bicuculline 7 6- 6 5- 5 4- 4 3- 2- 2 1 -l&n- 1 I IJ o 01 I I I 0 5 10 15 0 5 10 15 6 7 D 100 PM CACA + high K+ 100 uM bicuculline

5 4 3 2 1 3J ‘L 1 1-Q 0 ’ I I I 0 1 I I I 0 5 10 15 0 5 10 15 200 ,uM CACA + high Kt 200 /IM CACA + high K+ Figure 7. Effects of TACA, musci- ; -E 100 PM bicuculline 200 /..LM 3-APMPA mol, and CACA on K+-evoked Ca*+ 6- responses. A, One millimolar TACA, in 5- the presence of 100 FM bicuculline, produced an inhibitory effect on Ca2+ 4- responses similar to GABA. B, two 3- millimolar muscimol, in the presence 2 of 100 FM bicuculline, did not produce 2- substantial inhibition of the Ca2+ re- 1 - 1 sponse. C, Five millimolar muscimol, 0 0 1 t I I in the presence of 100 FM bicuculline, 0 5 10 15 0 5 10 15 produced a delayed inhibition of the Ca2+ response. D, One hundred milli- 500 PM CACA + high K+ molar CACA, in the presence of 100 ; -G 500 PM 3-APMPA FM bicuculline, had a somewhat vari- 6- able inhibitory effect on the Ca2+ response. E, Two hundred millimolar 5- CACA, in the presence of 100 pM bi- 4- 6 cuculline, completely suppressed the t 3 Ca2+ response. F, The effect of 200 p,M CACA was blocked by 200 pM 2 3-APMPA. G, The effect of 500 FM 1 I- CACA was not blocked by 500 pM O ’ I 1 I 3-APMPA. H, The effect of 500 FM CACA was antagonized by the com- 0 5 10 5 0 5 10 15 bination of 500 FM 3-APMPA and 100 PM bicuculline. Time (s)

GABA, receptorsdescribed in other species,with the exception fish bipolar cell terminals, was reported to be sensitiveto sub- of its extreme picrotoxinin resistance,it may be appropriateto micromolar CACA and could directly modulate Ca*+ channels refer to this receptor as a picrotoxin-insensitive GABA, receptor. (Heidelberger and Matthews, 1991; Matthews et al., 1994). Alternatively, a new nomenclature,such as GABA,, could be However, our calcium imaging methodsdid not reveal this novel adopted, but we discouragethis until the underlying molecular GABA,-like receptor in our preparation, since CACA was in- compositionof this family of GABA receptor subunitsis known. effective at low concentrations.We also failed to detect a baclofen-sensitiveGABA, receptor in our preparation,although such Coexisting multiple receptors at bipolur cell terminals a receptor has also been reported to modulateCa2+ channels at Our results demonstratethat at least two types of GABA recep- tiger salamanderbipolar cell terminals (Maguire et al., 1989) tors, a GABA, and a picrotoxin-insensitive GABA, receptor, and goldfish ganglion cells (Bindokas and Ishida, 1991), and to coexist in rat bipolar cell terminals,consistent with the findings regulate retinal signal processing (Pan and Slaughter, 1991; of Feigenspanet al. (1993) on bipolar cell somas.Coexistence Slaughter and Pan, 1992). It is not clear if this discrepancyis of GABA, and GABA, receptors at bipolar cell terminals has due to speciesvariation or methodologicaldifferences. also been reported in tiger salamanderand white perch (Qian Distinct properties of GABA, and GABA, receptors and Dowling, 1993b; Lukasiewicz et al., 1994). Moreover, a In this study, we directly demonstratethat both GABA, and the novel baclofen-insensitiveGABA,-like receptor,found in gold- GABA, receptorsare involved in regulating the voltage-depen- The Journal of Neuroscience, April 1995, 15(4) 2677

100 PM bicuculline 40 ,wM picrotoxinin 200 /&I 3-APMPA 8 I . 1 8 I- I i’[[GABA+highK+ 1 :, 2 /IM GABA + high KC

2 2 - 1 LJ- 1- 0 ’ / I , I I 0’ I , I I I 0 5 10 15 20 25 0 5 10 15 20 25 8 7 C 5 PM GABA + high K+ i CD 5 PM GABA + high KC 6 - 6 c 5 - 4- 3 - 2 - 13 0 ’ I I I I I 0 5 10 15 20 25 0 5 10 15 20 25

10 PM GABA + high Kf 6 - 5- 4- 3- 2- l-B4

Ei 7 -! 20 fi GABA + high K+ 20 PM GABA + high Kf Figure 8. Comparison of effects of GABA on GABA, receptors and GA- 6 6- BA, receptors. Inhibitory effects on 5 5- K+-evoked Ca2+ responses produced 4 4- by GABA acting at GABA, receptors were isolated by coapplying 100 pM 3 3- bicuculline and 40 p,M picrotoxinin (A, 2 2- C, E, G). Effects produced by GABA lylll acting at GABA, receptors were iso- I I I I i lated by coapplying 200 p,M 3-APMPA 0 ’ (B, D, F, H). Each GABA concentra- 0 5 10 15 20 25 tion was tested under both sets of conditions on the same bipolar cell termi- Time (s) nal dent Ca2+response at bipolar cell terminals.However, thesetwo suppressionof the early portion of the Ca*+ responseat 5 p.M receptorsdisplay distinct affinities and activation kinetics. The GABA (Fig. 80). This concentration is not much greater than GABA, receptor has a high affinity for GABA. The inhibitory the l-2 p,~ neededto observe a barely detectable inhibitory effect producedby this receptor wasobserved at submicromolar effect, consistingof a reduction in the rate of rise of the initial concentrations.In contrast, activation of GABA, receptors re- Ca2+response (Figs. 5B, 8B). quired relatively high concentrationsof GABA. The high poten- Fast activation kinetics is a property of most ionotropic li- cy of GABA at the GABA, receptor is consistentwith previous gand-gated ion channels. Although the GABA, receptor has electrophysiological studies(Polenzani et al., 1991; Kusamaet been reported to gate a chloride conductance(Feigenspan et al., al., 1993a; Qian and Dowling, 1993a; Matthews et al., 1994). 1993;Qian andDowling, 1993a),slow activation of this receptor While the GABA, receptor has a much higher affinity, it has not been previously described.The physiological implica- clearly displays a delayed onset and slow time course of Ca*+ tions of the slow onset of activity of the picrotoxin-insensitive inhibition. This delay is still prominent at GABA concentrations GABA, receptordescribed here are not yet clear. Further studies approximately 50-loo-fold greaterthan the thresholdconcentra- are neededto explain its phenomenonand its underlying mech- tion. This is in contrast to the GABA, receptor, which mediates anism. 2678 Pan and Lipton * GABA Modulation Calcium Response

Functional implications of the coexistence of multiple GABA BA,-like receptors on catfish cone- but not rod-driven horizontal receptors in bipolar cell terminals cells. J Neurosci 14:2648-2658. Feigenspan A, Wassle H, Bormann J (1993) Pharmacology of GABA It has been postulated that negative feedback from GABAergic receptor channels in rat retinal bipolar cells. Nature 361: 1599162. amacrine cells to bipolar cell terminals is the major mechanism Gersdorff H, Matthews G (1994) Dynamics of synaptic vesicle fusion generating transient responses in the inner retina. What, then, is and membrane retrieval in synaptic terminals. Nature 367:735-739. Heidelberger R, Matthews G (1991) Inhibition of calcium influx and the role of the GABA, receptor, and the implications of colo- calcium current by y-aminobutyric acid in single synaptic terminals. calization of GABA, and GABA, receptors at bipolar cell ter- Proc Natl Acad Sci USA 88:7135-7139. minals? Based on the distinct properties of GABA acting at these Heidelberger R, Matthews G (1992) Calcium influx and calcium cur- two GABA receptor subtypes, one apparent explanation for the rent in single synaptic terminals of goldfish retinal bipolar neurons. colocalization of multiple GABA receptors is fine tuning of the J Physiol (Lond) 447:235-256. Johnston GAR (1986) Multiplicity of GABA receptors. In: Receptor Ca?+ level in the bipolar cell terminals. The fast activation of biochemistry and methodology, Vol 5, Benzodiazepine/GABA recep- the GABA, receptor could generate fast, transient inhibition. tors and chloride channels (Olsen RW, Venter JC, eds), pp 57771. However, this effect would lack a wide dynamic range due to New York: Liss. its narrow activation parameters. On the other hand, the GABA, Kaneko A, Pinto LH, Tachibana M (1989) Transient calcium current of retinal bipolar cells of the mouse. J Physiol (Lond) 410:613-629. receptor, with its slow activation kinetics and broad activation Kao JPY, Harootunian AT, Tsien RY (1989) Photochemically generated range, could well generate slow, sustained, and graded re- cytosolic calcium pulses and their detection by fluo-3. J Biol Chem sponses. Of course, this is based on the assumption that the 264:B 179-8 184. synaptic concentration of GABA is regulated at micromolar lev- Karschin A, Wassle H (1990) Voltage- and transmitter-gated currents els. The concentration range of GABA at bipolar synaptic ter- in isolated rod bipolar cells of rat retina. J Neurophysiol 63:860-876. Kusama T, Spivak CE, Whiting P, Dawson VL, Schaeffer JC, Uhl GR minals is not clear. Lukasiewicz and Werblin (I 994) have re- (1993a) Pharmacology of GABA pl and GABA,,B receptors exported that light-evoked inhibitory synaptic inputs to bipolar pressed in Xenop~l.7 oocytes and COS cells. Br J Pharmacol 109:200- cells in tiger salamander are not signihcantly antagonized by 206. bicuculline. This result may imply that the light-evoked GABA Kusama T, Wang T-L, Guggino WB, Cutting GR, Uhl GR (199313) concentration at bipolar cell terminals is in the low micromolar GABA p2 receptor pharmacological profile: GABA recognition site similarities to pl. Eur J Pharmacol 245:83-84. range and, therefore, not high enough to activate GABA, recep- Leifer D, Lipton SA, Barnstable CJ, Masland RH (1984) Monoclonal tors. An alternative interpretation for their results is also quite antibody to Thy-l enhances regeneration of processes by rat retinal possible, as proposed by these authors (Lukasiewicz and Wer- ganglion cells in culture. Science 224:303-306. blin, 1994). They postulated that transmission from bipolar cells Lipton SA, Tauck DL (1987) Voltage-dependent conductances of sol- to ganglion cells is primarily modulated by GABA, receptors in itary ganglion cells dissociated from the rat retina. J Physiol (Lond) 385:361-391. the tiger salamander retina. Further studies, especially using se- Lukasiewicz PD, Werblin FS (1994) A novel GABA receptor modu- lective GABA, antagonists, will be necessary to answer this lates synaptic transmission from bipolar to ganglion and amacrine question. cells in the tiger salamander retina. J Neurosci 14: 1213-1223. The bipolar cells we studied are probably rod bipolar cells Lukasiewicz PD, Maple BR, Werblin FS (1994) A novel GABA re- based on previous reports (Karschin and Wassle, 1990). There- ceptor on bipolar cell terminals in the tiger salamander retina. J Neu- rosci 14:1202-1212. fore, it is not clear whether there is a similar colocalization of Maguire G, Maple B, Lukasiewicz P, Werblin F (1989) y-aminobutyr- GABA, and GABA, receptors at terminals other than rod bi- ate type B receptor modulation of L-type calcium channel current at polar cells. In tiger salamander retinal slice, it has been reported bipolar cell terminals in the retina of the tiger salamander. Proc Nat1 that there is no significant difference in GABA receptor phar- Acad Sci USA 86: 10144-10147. macology between ON and OFF bipolar cells (Lukasiewicz et Massey SC, Redburn DA (1987) Transmitter circuits in the vertebrate retina. Prog Neurobiol 28:55-96. al., 1994). However, in the same preparation, Zhang and Slaugh- Matthews G, Ayoub GS, Heidelberger R (1994) Presynaptic inhibition ter (1994) recently reported that activation of GABA, receptors by GABA is mediated via two distinct GABA receptors with novel preferably blocked the ON response in third order retinal neu- pharmacology. J Neurosci 14:1079-1090. rons, raising the possibility of a differential role of the GABA, Nistri A, Sivilotti L (1985) An unusual effect of y-aminobutyric acid on synaptic transmission of frog tectal neurones in vitro. Br J Phar- receptor in the ON and OFF pathways in the retina. Further macol 85:917-921. studies will be necessary to define further the physiological in- Olsen RW, Venter JC (1986) Benzodiazepine/GABA receptors and fluence of the GABA, receptor at the retinal bipolar-ganglion chloride channels: structural and functional properties. In: Receptor cell synapse. biochemistry and methodology, Vol 5. New York: Liss. Pan Z-H, Lipton SA (1994a) Effect of NMDA receptor stimulation versus voltage-dependent calcium channel stimulation on calcium References transient, and compartments in mammalian retinal ganglion cells. In- Bindokas VP, Ishida AT (1991) (-)-Baclofen and y-aminobutyric acid vest Ophthalmol Vis Sci [Suppl] 35:1909. inhibit calcium currents in isolated retinal ganglion cells. Proc Nat1 Pan Z-H, Lipton SA (1994b) Multiple GABA receptor subtypes me- Acad Sci USA 88:10759910763. diate inhibition of calcium influx at rat retinal bipolar cell terminals. Cutting GR, Lu L, O’Hara BE Kasch LM, Montrose-Rafizadeh C, Don- Sot Neurosci Abstr 20:217. ovan DM, Shimada S, Antonarakis SE, Guggino WB, Uhl GR, Ka- Pan Z-H, Slaughter MM (1991) Control of retinal information coding zazian HH Jr (1991) Clonine of the Y-aminobutvric acid (GABA) by GABA, receptor. J Neurosci 11: 18 10-l 82 1. pl cDNA: a GkBA’receptor&~bunit highly expressed in the retina. Polenzani L, Woodward RM, Miledi R (1991) Expression of mam- Proc Nat1 Acad Sci USA 88:2673-2677. malian y-aminobutyric acid receptors with distinct pharmacology in Cutting GR, Curristin S, Zoghbi H, O’Hara B, Seldin ME Uhl GR Xenopus oocytes. Proc Nat1 Acad Sci USA 88:4318-4322. (1992) Identification of a putative y-aminobutyric acid (GABA) re- Qian H, Dowling JE (1993a) Novel GABA responses from rod-driven ceptor subunit p2 and colocalization of the genes encoding p2 retinal horizontal cells. Nature 361:162-l 64. (GABRR2) and pl (GABRRI) to human chromosome 6q14-q21 and Qian H, Dowling JE (1993b) GABA responses on retinal bipolar cells. mouse chromosome 4. Genomics 12:BOl-806. Biol Bull 185:312. Dong C-J, Picaud SA, Werblin FS (1994) GABA transporters and GA- Qian H, Dowling JE (1994) Pharmacology of novel GABA receptors The Journal of Neuroscience, April 1995, 15(4) 2679

found on rod horizontal cells of the white perch retina. J Neurosci Dihydropyridine-sensitive calcium current mediates neurotransmitter 14:4299-4307. release from bipolar cells of the goldfish retina. J Neurosci 13:2898- Shimada S, Cutting G, Uhl GR (1992) y-aminobutyric acid A or C 2909. receptor? y-aminobutyric acid pl receptor RNA induces bicuculline-, Tauck DL, Frosch MP, Lipton SA (1988) Characterization of GABA- barbiturate-, and benzodiazepine-insensitive y-aminobutyric acid re- and glycine-induced currents of solitary rodent retinal ganglion cells sponses in Xenopus oocytes. Mol Pharmacol 41:683-687. in culture. Neuroscience 27:193-203. Sivilotti L, Nistri A (1989) Pharmacology of a novel effect of y-ami- Woodward RM, Polenzani L, Miledi R (1992) Characterization of bi- nobutyric acid on synaptic transmission of frog tectal neurones in cuculline/baclofen-insensitive y-aminobutyric acid receptors ex- vitro. Eur J Pharmacol 164:205-212. pressed in Xenopus oocytes. I. Effects of Cll channel inhibitors. Mol Sivilotti L, Nistri A (199 1) GABA receptor mechanisms in the central Pharmacol 42: 165-l 73. nervous system. Prog Neurobiol 36:35-92. Woodward RM, Polenzani L, Miledi R (1993) Characterization of bi- Slaughter MM, Pan Z-H (1992) The physiology of GABA, receptors cuculline/baclofen-insensitive (p-like) y-aminobutyric acid receptors in the vertebrate retina. Prog Brain Res 9:47-60. expressed in Xenopus oocytes. II. Pharmacology of y-aminobutyric Suzuki S, Tachibana M, Kaneko A (1990) Effects of glycine and acid, and y-aminobutyric acid, receptor agonists and antagonists. GABA on isolated bioolar cells of the mouse retina. J Phvsiol_ (Land). Mol Pharmacol 43:609-625. 421:645-662. I Yeh HH, Lee MB, Cheun JE (1990) Properties of GABA-activated Tachibana M, Kaneko A (1987) y-Aminobutyric acid exerts a local whole-cell currents in bipolar cells of the rat retina. Vis Neurosci inhibitory action on the axon terminal of bipolar cells: evidence for 4:349-357. negative feedback from amacrine cells. Proc Nat1 Acad Sci USA 84: Zhang J, Slaughter MM (1994) The GABAp receptor preferentially 3501-3505. suppresses ON responses in the amphibian retina. Invest Ophthalmol Tachibana M, Okada T, Arimura T Kobayashi K, Piccolino M (1993) Vis Sci [Suppl] 35:1364.