sign in sign up
<<
Home , Bicuculline, Gabazine, GABA receptor agonist, GABA receptor antagonist, Hydrastine, Picrotoxin

The Journal of Neuroscience, July 1994, 14(7): 4299-4307

Pharmacology of Novel GABA Receptors Found on Rod Horizontal Cells of the White Perch Retina

Haohua Qian and John E. Dowling The Biological Laboratories, Harvard University, Cambridge, Massachusetts 02318

A novel type of GABA is present on rod-driven (H4) GABA is the main inhibitory in the vertebrate horizontal cells of the white perch retina (Qian and Dowling, CNS. Two GABA receptor types, GABA, and GABA,, have 1993a). These receptors have been tentatively termed GA- beenwell characterized. GABA, receptorsare -gatedchlo- BA, receptors. In this study, the pharmacological properties ride channels usually linked to fast synaptic inhibition. Such of these receptors were further investigated by applying receptorsare sensitiveand often can be allosterically several conformationally restricted GABA, receptor ago- modulated by both and (Olsen, nists, GABA, antagonists, and a GABA, to the H4 1982; Bormann, 1988; Sivilotti and Nistri, 199 1). GABA, re- horizontal cells. GABA analogs locked in a partially folded ceptors are coupled to potassium and/or calcium channels conformation had a variety of effects. lsonipecotic acid had through G-proteins, and their responsesordinarily have a slow no effect on these receptors, whereas activated time course (Gage, 1992). GABA, receptors can be selectively them but .with low potency (EC,, = 137 PM). THIP (4,5,6,7- activated by and antagonized by and tetrahydroisoxazolo[5,4-clpyridin-3-01) acted as a competi- 2-hydroxysaclofen (Hill and Bowery, 1981; Bormann, 1988; tive antagonist on these receptors with an inhibition constant Kerr et al., 1988; Sivilotti and Nistri, 199 1). of 82.5 PM. P4S (piperidine-4-sulfonic acid) activated the re- In addition to thesetwo well-known types ofGABA receptors, ceptors at high concentrations (> 1 mM), but at lower con- the literature suggestsanother GABA receptor type is present centrations it was a competitive antagonist with an inhibition in the vertebrate CNS. Johnston and colleagues(Johnston et al., constant of 80.9 PM. l4AA (-4-acetic acid), a GABA 1975; Drew et al., 1984) reported a bicuculline- and baclofen- analog with an extended conformation, potently inhibited the insensitive GABA binding site on rat cerebellar membranesthat GABA responses on H4 horizontal cells with an inhibition was antagonized by CACA (cis-aminocrotonic acid), a confor- constant of 1.67 PM. , which can assume both par- mationally restricted GABA analog. Such novel GABA binding tially folded and extended conformations, acted as a mixed sites were also reported in the catfish brain (Myers and Tun- agonist-antagonist. The GABA responses on H4 horizontal nicliff, 1988). Johnston (1986) proposedthat the receptors me- cells were resistant to several competitive GABA, receptor diating thesenovel responsesbe termed GABA, receptors. An antagonists including bicuculline, , and SR-95531, unusual GABA responsewas also reported in central visual but they were very sensitive to (I&,, = 237 nM). pathways by Arakawa and Okada (1988) and Sivilotti and Nistri The inhibition by picrotoxin was both competitive and non- (1989) who showed that such responsesare bicuculline- and competitive in nature. On the other hand, TBPS (tert-butyl- baclofen-insensitive, do not showdesensitization even following bicyclophosphorothionate), another GABA, receptor chan- sustainedexposure to GABA, and are dependent on extracel- nel blocker, had minimal effects on these receptors. The lular (Nistri and Sivilotti, 1989). specific GABA, agonist 3-APA (3-aminopropyl phosphonic More recently, a bicuculline-resistant GABA responsewas acid) acted as a competitive antagonist on H4 horizontal observed by Polenzani et al. (199 1) in Xenopusoocytes express- cells with an inhibition constant of 43 PM. The GABA re- ing mRNA derived from bovine retinas. Such responseswere sponses on the horizontal cells were also resistant to strych- sustainedand had a higher sensitivity to GABA than typical nine, a . The results provide fur- GABA, receptor responses.This GABA responsecould not be ther evidence that the GABA receptors on H4 horizontal cells modulated by barbiturates and benzodiazepines,and could not in perch are pharmacologically distinct from known GABA be activated by baclofen or suppressedby . At about receptors, supporting the designation of GABA, receptors the sametime, Cutting et al. (199 1) usingprimers derived from for them. a highly conservedregion of GABA, and genes, [Key words: GABA, receptor, retina, horizontal cells, phar- cloned a GABA receptor (~1) subunit from a human retinal macology, GABA] cDNA library that, when expressedin Xenopusoocytes, formed a homooligomericGABA receptor. Shimadaet al. (1992) showed that the responsesmediated by thesereceptors resemble closely Received Sept. 27, 1993; revised Dec. 27, 1993; accepted Jan. 13, 1994. the properties of the GABA responsesrecorded from oocytes We thank Drs. Gordon L. Fain and George B. Grant for reading the manuscript. This work was supported in part by NIH Grant EY00824, and a fellowship from expressingretinal mRNA, and they proposedthat the receptors Firht For Sight. Research Division of the National Societv to Prevent Blindness. made up of pl subunits are the GABA, receptors. Corresp&bence should be addressed to Dr. Haohua Qian, Department of Cellular and Developmental Biology, The Biological Laboratories, Harvard Uni- We have recorded GABA responsesfrom rod-driven (H4) versity, 16 Divinity Avenue, Cambridge, MA 02138. horizontal cells of the white perch retina that are sustainedand Copyright 0 1994 Society for Neuroscience 0270-6474/94/144299-09$05.00/O mediated by the opening of chloride channels(Qian and Dowl- 4300 Qian and Dowling - of GABA Receptors ing, 1992, 1993a). These responses are bicuculline insensitive All other agents were obtained from Research Biochemicals Inc. (Natick, and not modulated by either barbiturates or benzodiazepines. MA). Data analysis. McPherson (1989) has presented a detailed mathe- They are also baclofen insensitive and cannot be blocked by matical analysis of -receptor interactions. Briefly, the relation of either phaclofen or 2-hydroxysaclofen. Thus, these responses drug concentration and response of any receptor mediated response can resemble the putative GABA, responses recorded by Polenzani be described by the Hill equation: et al. (199 1) and by Shimada et al. (1992). -=R [Cl” More detailed studies on the receptors in Xenopus oocytes R rrlax [Cl8 + EC;,’ made up of the p 1 subunit has revealed a unique pharmacology where R is the response of the cell induced by the drug at concentration (Kusama et al., 1993; Woodward et al., 1993). For example, [cl, Lx is the maximum response of the cell, n is the Hill coefficient, most GABA, act as antagonists on Xenopus oocytes and EC,, is the concentration at which a half-maximum response is expressing the pl subunit. The detailed pharmacology of the induced. On white nerch H4 horizontal cells. 100 UM GABA induces novel GABA receptors on the H4 horizontal cells of the white maximal responses isee Fig. IB). The dose-response curve for antagonists can be described by the perch retina is not known. These cells provide an excellent sys- equation tem in which to study the properties of these novel receptors in , since the GABA responses on H4 horizontal cells seem R [II” to be mediated solely by these novel receptors (Qian and Dowl- RO- = l - [I]” + IC;,’ ing, 1993a). Although similar GABA responses have been ob- where R, is the GABA response without antagonist, R is the response served in other retinal neurons, in all other cases reported so induced by GABA in the presence of antagonist, [A is the concentration of the antagonist, n is Hill coefficient, and IC,, is the concentration of far the responses are mixed with the responses of other GABA antagonist at which halfof GABA response is inhibited. To elicit sizable receptors and/or GABA transporters (Dong and Werblin 1993; responses, 10 FM GABA was used as the test concentration. Feigenspan et al., 1993; Lukasiewicz et al., 1993; Qian and For a competitive antagonist, the value of IC,, can be used to estimate Dowling, 1993b). In this study, the pharmacological properties the inhibition constant (K,) of the antagonist. However, the relation of of GABA receptors on H4 horizontal cells of the white perch K, and IC,, depends on the concentration of agonist used for the test, which can be expressed as follows: retina were characterized with agents known to interact with classic GABA, and GABA, receptors. K, = i”-, ~ R”)““IC,,, \ 1’mar / where R, is the test response (same as R, in Eq. 2), that is, the response Materials and Methods induced-by 10 PM GABA. For the GABA receptors on H4 horizontal Cell isolation. Solitary retinal cells were isolated from the white perch cells (with 10 UM GABA as test) K. = 0.286 IC,, (see Fig. 1B). retinausing methods described previously (Dowling et al., 1985). Brief- A more precise way to determine the inhibit;& constant okcompet- ly, eyes were enucleated, rinsed with 70% , and hemisected at itive antagonists is to measure the shift of EC,, value of the dose- the ora serata. The retina was isolated using fine forceps, and incubated response curve in the presence of a certain concentration of the antag- for 40 min in an enzymatic solution made up of Leibovitz’s L- 15 culture onist. K, can be determined as follows: media (GIBCO-Bethesda Research Labs Life Technologies Inc., Gaith- [II ersbug, MD) containing 25 U/ml papain (Worthington Biochemical Kj = I-- \ ’ (4) Corporation, Freehold, NJ) and a few crystals of I-cysteine. The tissue (E&L- ,) was washed five times with fresh L- 15 and triturated through a sterilized pipette. Aliquots of the supernatant containing the isolated cells were where EC,,,, is the EC,, for dose-response relationship of GABA in the placed in 35 mm petri dishes containing 3 ml of L- 15 culture medium presence of the antagonist at concentration [Zj. with no antibiotics added. The dishes were left in the dark at room For noncompetitive antagonists, the dose-response curve can be ex- temperature for up to 1 week. pressed as Electrical recordings. Before each experiment, the culture medium was replaced with Ringer’s solution, containing (mM) choline Cl (145), -=R M* [Cln CaClz (2.4), MgCl, (1.5), glucose (lo), and HEPES (10). The whole-cell R mrx [Cl,, + EC;,’ version of the patch-clamp recording technique was used to record membrane current from horizontal cells (Hamill et al., 198 1). Pipettes where M is the maximum response remaining in the presence of the with tip diameters of about 1 pm were pulled on a Narishige two-step antagonist. puller. Recording pipettes were filled with an intracellular solution con- The experimental data were fitted to the theoretical equations by the taining (mM) CsCl(124), CaCl, (l), EGTA (1 l), MgCl, (2), HEPES (10). commercial software program SIGMAPLOT (Jandel Scientific Inc., San A patch-clamp amplifier (Dagan Corporation, Minneapolis, MN), linked Rafael, CA). All data are expressed as means + SD. to an IBM-PC 486 computer, was used to measure membrane current recorded from the cell. The voltage-clamp procedure was controlled using the software ~CLAMP (Axon Instruments Inc., Burlingame, CA). Results Cells were held at -50 mV. Membrane potentials were corrected for Properties of GABA responseson H4 horizontal cells electrode tip potential as in Fenwick et al. (1982). Recordings were stored both on the computer and on video tape using a Vetter PCM recorder. Figure 1A showsthe responsesof a white perch H4 horizontal The resistance of a typical patch pipette was about 2 Ma, and series cell to a 30 set pulse of GABA at several different concentra- resistance during whole-cell recording was about 4 M% Consequently, tions. When the cells were held at -50 mV, GABA induced an a 100 pA current would produce an error of 0.4 mV in horizontal cell inward current. We have shown previously that this current is membrane voltage. Recordings were not compensated. carried by chloride ions (Qian and Dowling, 1993a).The current Drug delivery. A local superfusion system was used to deliver onto the cells (Qian et al., 1993). Drugs were added at various concen- responseshave a slow time course, and they show no sign of trations to Ringer’s solution and introduced into the culture dish through desensitization during the prolonged GABA application even a pipette (tip diameter - 200 Frn) via a gravity-fed superfusion system at the highest concentrations (30-100 PM). A dose-response at a rate of about 1 ml/min. The pipette was mounted on a microma- curve averaged from 15 cells is shown in Figure 1B. The con- nipulator and positioned about 500 pm from the cells. The perfusate was removed from the culture dish by a suction-operated siphon. For tinuous curve is a theoretical fit of the data to Equation 1 with experiments with antagonists, drugs were coapplied with GABA. GABA n = 1.83 and EC,, = 3.03 WM. Theseresults are in good agreement and 14AA were obtained from Sigma Chemical Co. (St. Louis, MO). with our earlier data (Qian and Dowling, 1993a). The Journal of Neuroscience, July 1994, U(7) 4301 B .

0.8 .

0.6

Figure 1. A, GABA elicitedcurrent responses in an H4 horizontalcell. The concentrationof GABA is shownon the left of eachtrace (PM).The durationof GABA applicationis shownby the bar at the top. B, Dose-responsecurve for GABA inducedcurrents in H4 horizontalcells. The data are from 15 cells.For each cell, the responseswere expressed as a fraction of the maximumresponse induced by 100 PM GABA, R,,,. The continuouscurve is a mathematicalfit of data pointsto the Hill equationwith the parametersof n = 1.83,EC,, = 3.03 PM.

current in the cells, although the responsegenerated by 100 I.LM GABA, receptor agonists isoguvacine was only about one-third the amplitude of the re- Drug rationale. The GABA molecule exhibits a considerable sponsegenerated by 10 PM GABA. Interestingly, the responses conformational flexibility becauseeach single bond within the induced by isoguvacine returned to the baseline much faster molecule can rotate freely. Different conformations of GABA than did the responsesinduced by GABA. This has also been may interact with different types of receptors(Johnston, 1992). observed in responsesinduced by CACA and muscimol (Qian Previous studies have shown that GABA interacts best with and Dowling, 1993a), and may reflect the characteristicsof the GABA, receptors in a partially folded, planar conformation as binding of theseagents to these receptors. shown in Figure 2A (Krogsgaard-Larsenet al., 1985; Johnston, Isoguvacine can fully activate the receptor on H4 horizontal 1992). Indeed, compounds with such a restricted conformation cells, as shown by the dose-responsecurve for isoguvacine in are all potent GABA, receptor agonists,and several such com- Figure 3B. However, isoguvacine (EC,, = 137 WM) is much less pounds that were used in this study are shown in Figure 2A. potent on H4 horizontal cell receptors than. is GABA. This is an amino group locked GABA analog, contrasts with its action on classicGABA, receptors, which are whereas isoguvacine and P4S (piperidine-4-sulfonic acid) are more sensitiveto isoguvacine than to GABA (Krogsgaard-Lar- derivatives of isonipecotic acid. THIP (4,5,6,7-tetrahydroisox- sen et al., 1977). Isonipecotic acid has a very similar molecular azolo[5,4-clpyridin-3-01) is an amino group locked muscimol structure to isoguvacine (Fig. 2A) and also strongly activates analog. GABA, receptors(Bowery et al., 1978; Krogsgard-Larsenet al., GABA can also assumea fully extended conformation as 1985). However, isonipecoticacid (100 PM) elicited no responses shown in Figure 2B, and I4AA (imidazole-4-acetic acid) is an from H4 horizontal cells (Fig. 3A), nor did it have any effects analog of such conformation. I4AA is a GABA, receptor ago- on GABA induced responses(data not shown). nist, but it is not as potent as the compounds in a partially THIP (4,5,6,7-tetrahydroisoxazolo[5,4-clpyridin-3-01) is a folded conformation. Muscimol, on the other hand, can exist GABA, receptor agonist that is more potent than GABA in in both the fully extended and partially folded conformation activating GABA, receptors (Krogsgaard-Larsenet al., 1977; since the N-C bond is not locked. These two conformations of Alger and Nicoll, 1982). No responseswere elicited in the hor- muscimol are shown in Figure 2C. izontal cells when THIP was presentedalone. When THIP was Partiallyfolded analogs. Figure 3A showsthe effectsof GABA coapplied with GABA, it reducedthe GABA responsesas shown (10 PM), isoguvacine (100 PM), and isonipecotic acid (100 PM) in Figure 4A. The responsesare expressedas a fraction of the on an H4 horizontal cell isolated from the white perch retina. response induced by 10 FM GABA in Ringer’s, and can be Both GABA and isoguvacine induced a slow sustainedinward described by Equation 2, with parametersof n = 2.1 and IC,, 4302 Qian and Dowling l Pharmacology of GABA Receptors A . B . A. Isonipecotic Acid

GABA GABA

GABA

B.

C . 1.2 - - - GABA Isoguvacine . Iaoguvauille 1.0 - .- / / 0.8 - / I 0.6 - I I 0.4 - THIP

Muscimol Concentration (M) P4s

Figure 2. Molecular structures of GABA and several conformationally Figure 3. A, The effects of GABA (10 PM), isoguvacine (100 FM), and restricted analogs. A, GABA and its analogs in a partially folded con- isonipecotic acid (100 PM) on an H4 horizontal cell. B, Dose-response formation. Such a conformation is preferred by GABA, receptors, and curve for isoguvacine on H4 horizontal cells. Data were from seven all the compounds illustrated are potent GABA, receptor agonists. B, cells; for each cell the responses are expressed as a fraction of the max- GABA and its analog (14AA) in a fully extended conformation. 14AA imum response on that cell induced by 100 PM GABA, R,,,. The con- is a weak agonist on GABA, receptors. C, Muscimol can exist in both tinuous curve is a mathematical fit of the data points to the Hill equation a fully extended and partially folded conformation. with the parameters of n = 1.48, EC,, = 137 PM. The dashed line indicates the dose-response curve for GABA as shown in Figure 1B. The maximum response for isoguvacine (10 FM) was 100.1 f 27.4 pA, = 301 FM. For a competitive antagonist, the IC,, value allows a value close to the maximum response elicited by 100 PM GABA on for an estimation of the inhibition constant (K,), which was 85 the same group of cell (96.3 f 17.2 PA). KM for THIP (Eq. 3). That the inhibition by THIP is competitive is shown in Figure 4B. In the presence of 1 mM THIP, the GABA mM, P4S activated small membrane currents. Thus, P4S acts dose-response curve was shifted to the right on the abscissa as a at the GABA site. Similar observations have without affecting the maximum GABA response. The amount been made on oocytes expressing GABA p 1 subunits (Kusama shifted (from an EC,, of 3.03 PM in Ringer’s to an EC,, of 38.6 et al., 1993). When P4S was coapplied with GABA, P4S reduced ELM in the presence of 1 mM THIP) can be used to calculate the GABA-induced responses (Fig. 4A). The calculated IC,, for more precisely the inhibition constant, K,, according to Equa- P4S was 293 KM, which corresponds to an inhibition constant tion 4. The calculated K, for THIP was 85.2 PM, close to the K, (K,) of 82.8 PM. As was the case for THIP, 1 mM P4S shifted value estimated from the IC,, value. the GABA dose-response curve to the right on the abscissa P4S (piperidine-4-sulfonic acid), another specific and pow- without affecting the maximum response. That is, P4S also acts erful GABA, receptor agonist, is a derivative of isonipecotic as competitive antagonist. The calculated inhibition constant acid in which the carboxyl group is substituted by a sulfonate was 80.9 WM for P4S. group (Krogsgaard-Larsen et al., 1980). When P4S was pre- Extendedanalogs. 14AA (imidazole-4-acetic acid) induced no sented to H4 horizontal cells at concentrations less than 1 mM, responses in H4 horizontal cells at concentrations as high as 1 no responses were observed. At concentrations greater than 1 mM (data not shown). However, this agent acted as a powerful The Journal of Neuroscience, July 1994, 74(7) 4303

A. Muscimol GABA GABA + Muscimol

0 THIP

al 0.8 m E 0.6 ,a iz 0.4

2 0.2

s 0.0

1o-7 10-6 lo-” Drug b:n:e&:on ):j3 Figure 5. The responses of an H4 horizontal cell to 100 PM muscimol (left), IO PM GABA (center), and the coapplication of 10 PM GABA and 100 PM muscimol (right).Note the response induced by coappling GABA and muscimol was smaller than the response induced by GABA alone, indicating an antagonistic effect of muscimol on the GABA response.

bition constant (K,) was 1.92 FM. In the presenceof 100 PM 14AA, the GABA dose-responsecurve was shifted to the right on the abscissawithout affecting the maximum response(Fig. 4B). The amount shifted (from an EC,, of 3.03 FM in Ringer’s Ringer to an EC,, of 186 WM in the presenceof 100 FM 14AA) indicates 1 mM THIP an inhibition constant of 1.67 WM for 14AA.

l 1 mM P4S Analogs with mixed conformation. Muscimol is a potent GA- 100 pMyM I4AA BA, receptor agonist. Although muscimol can also induce some responsein H4 horizontal cells, it is lesseffective than is GABA. I In addition, the maximum responseto muscimol is only about 10-7 10-s 10-5 10-4 10-3 10-2 one-third of the maximum GABA responseregardless of con- GABA Concentration (M) centration (seeFig. 4 in Qian and Dowling, 1993a).This implies that muscimol acts as partial agonist on H4 horizontal cells. Figure 4. Effect of GABA analogs on GABA responses of H4 hori- Muscimol can also antagonize the GABA responseas shown in zontal cells. A, Inhibition curves for THIP (six cells), P4S (six cells), Figure 5. Figure 5 (left) showsa responseof a horizontal cell to and I4AA (five cells). For each cell the responses induced by 10 /LM 100PM muscimol, which wasabout 36% ofthe responseinduced GABA in the presence of the agents are expressed as a fraction of the in that cell by 10 PM GABA (Fig. 5, center). When 100 FM response induced by 10 FM GABA in Ringer’s (R,). The continuous curves are a mathematical fit of the data points to Equation 2, with n muscimol was coapplied with 10 PM GABA, the responsewas = 2.1 and IC,,, = 301 KM for THIP and n = 1.46 and IC,, = 6.72 ELM only about 60% of the size of the responseelicited by 10 FM for 14AA. The calculated curve for P4S (not shown) is similar to the GABA alone (Fig. 5, right). Similar inhibition of GABA re- curve for THIP with parameters of n = 1.7 and IC,, = 293 PM. The inhibition constants (K,) calculated from IC,,, values are 85.0, 82.8, dnd sponsesby muscimol was observed in four other cells. 1.90 PM for THIP, P4S, and 14AA, respectively. B, Dose-response curves for GABA on H4 horizontal cells in the presence of I mM THIP (six GABA, receptor antagonists cells), 1 mM P4S (five cells), and 100 PM 14AA (five cells). For each cell Bicuculline is the most common competitive GABA, receptor responses to 100 FM GABA in Ringer’s solution were treated as R,,,. antagonist, and its effectivenessis often usedas the criterion for The continuous curves are a mathematical fit of the data points to Equa- the involvement of GABA, receptors in a response(Hill and tion 1 with parameters of n = 2.2 and EC,, = 38.6 PM for the GABA response in the presence of THIP, and n = I .74 and EC,, = 186 PM in Bowery, 1981). The GABA responsesfrom H4 horizontal cells the presence of 14AA. The GABA response curve in the presence of are almost completely resistant to bicuculline. Even with a con- P4S (not shown) has parameters of n = 2.17 and EC,, = 44.1 FM. For centration of 500 PM bicuculline, 96% of GABA responsesre- comparison, the GABA dose-response curve in Ringer’s from Figure 1A is replotted as the dashed line. The inhibition constants calculated main (Qian and Dowling, 1993a). Hydrastine, a recently de- from EC,, values are 85.2, 80.9, and 1.67 FM for THIP, P4S, and 14AA, veloped and potent competitive GABA, receptor antagonist respectively. These data are presented as mean + SE. (Huang and Johnston, 1990) also did not substantially block the GABA responsesin H4 horizontal cells (Fig. 6A). When 100 PM of hydrastine was coapplied with 10FM GABA, the responses competitive antagonist on GABA-induced responsesin H4 hor- elicited were decreasedby lessthan 20% (0.19 +- 0.06, n = 4). izontal cells. Indeed, of all the analogstested in this study, I4AA A failure to block the GABA responsesin H4 horizontal cells was the most potent. Responsesinduced by 10 KM GABA in was also observed with SR-95531 (), another specific the presenceof 14AA are shown in Figure 4A. The contin,uous competitive GABA, receptor antagonist (Heaulme et al., 1986) curve is a mathematical fit of the data to Equation 2 with pa- (Fig. 6B). In the presenceof 100 hi SR-95531, 74% of the rameters of y1= 1.46 and IC,, = 6.72 WM. The calculated inhi- control responsesremained (0.74 ? 0.08, y2= 4). 4304 Qian and Dowling * Pharmacology of GABA Receptors A. ^o 1.2 -q 1.0 s B 0.8 d% 0.6

g 0.4 p: 4 0.2 2 0 0.0

1 o-Q 1 o-6 1 o-' 1 o-6 1 o-5 Picrotoxin Concentration (M)

- - Ringer

l Picrotoxin / .--- / lx- / 0.6 1’

Figure 6. Effect of 100 FM hydrastine (A) and 100 PM SR-95531 (B) on the GABA responses of H4 horizontal cells. Arrows point to GABA responses obtained in the presence of the agent. The bar above the traces 1 ...... r ...... I ...... ‘.I . . . ..CLJ shows the duration of GABA application. 1 o-' 1 o-6 1 o-5 1 o-4 1 o-3 GABA Concentration (M) Picrotoxin is the most commonly usednoncompetitive GA- BA, receptor antagonist. It behavesas a chloride channelblock- Figure 7. Effect of picrotoxin on GABA responsesof H4 horizontal er on these receptors, reducing the amplitude of GABA,-me- cells. A, Inhibition curve for picrotoxin. The continuous curve is a diated responseswithout affecting EC,, values (Akaike et al., mathematical fit of data points to Equation 2 with parameters of n = 1985; Inoue and Akaike, 1988). The GABA responsesof H4 1.11 and IC,, = 237 nM. Data are from four cells. B, Dose-response curves of GABA on H4 horizontal cells in the presence of 1 PM pic- horizontal cells were very sensitiveto picrotoxin. An inhibition rotoxin. Data are from four cells. The dose-response curve for GABA curve for picrotoxin on H4 horizontal cell responsesis shown is shifted to the right and is vertically compressed in the presence of in Figure 7A. The data can be fitted by Equation 2 with an IC,, picrotoxin, indicating that picrotoxin inhibits the responses by both value of 237 FM, which is an order of magnitude lower than the competitive and noncompetitive mechanisms. The continuous curve is IC,, value reported for typical GABA, receptors (Inoue and a mathematical fit of the data points to Equation 4 with parameters of n - 1.63 and EC,, = 6.97 FM. See Figure 4 for more details. Akaike, 1988; Lehoullier and Ticku, 1989). The inhibition by picrotoxin of GABA responsesin H4 cells is both competitive and noncompetitive. Dose-responsecurves GABA responsesof oocytes expressingbovine retinal mRNA of the GABA responseelicited in the absenceand presenceof (Woodward et al., 1992). 1 PM picrotoxin are shown in Figure 7B. In the presenceof 1 TBPS (tert-butylbicyclophosphorothionate) is a potent GA- FM of picrotoxin, about 20% of the GABA responseremained BA, receptor channelblocker that is thought to bind to the same (Fig. 7A). Twenty percent of the responsein Figure 7B corre- site on GABA, receptors as picrotoxin (Akaike et al., 1985; spondsto a GABA concentration of 8 FM, a value very close to Tehrani et al., 1986; Ticku 1986; Renterghemet al., 1987; Inoue that used in the experiments shown in Figure 7A (i.e., 10 FM). and Akaike, 1988). The effectsof TBPS on the GABA responses The GABA dose-responsecurve was shifted to the right in the elicited from white perch H4 horizontal cellsare shown in Figure presenceof picrotoxin, indicating the competitive nature of the 8. The responsesare expressedas the fraction of the response effect; on the other hand, the maximum responsewas also re- induced by 10 PM GABA in Ringer’s. No significant inhibition duced, indicating a noncompetitive component to the blockade. of the GABA-induced responseswas observed in the presence A similar phenomenon has been observed for picrotoxin on the of TBPS at concentrations as high as 10 PM. Even in the presence The Journal of Neuroscience, July 1994, M(7) 4305

1.2 - m. 1.0 - 1.2 0 -? p: 0.6 ? 1.0 V !?5 0.8 ii g 0.6 T PI P : 0.4 w 0.2 2

3 0.0 I I . . ..I IJ 1 o-7 10-e 1 o-5 lo-' ...... I . . ."..I . . -.....I 1 o-5 1 o-4 1 o-3 1 o-2 TBPS Concentration (M) SAPA Concentration (Id) Figure 8. Effect of TBPS on GABA responsesof H4 horizontalcells. Data were averaged from five cells. No significant reduction of the GABA response was observed in the presence ofTBPS at concentrations as high as 10 PM. Even at 100 PM, 30% of the GABA response remained. B. 1.2 of 100 PM TBPS, about 30% of the GABA-induced response Ringer remained. 1 .o 3-APA ,--- / T GABA, receptor agonists / 0.8 / Phosphateanalogs of GABA act as GABA, receptor agonists. / An example is 3-aminopropyl phosphonic acid (3-APA), *which 0.6 / is a selective GABA, receptor agonist(Slaughter and Pan, 1992). / When 3-APA wascoapplied with GABA on H4 horizontal cells, 0.4 / it inhibited the GABA-induced responses(Fig. 9A). The IC,, / T for 3-APA was 165 FM, which correspondsto an inhibition 0.2 constant of 47 PM. The inhibition by 3-APA on H4 horizontal cell GABA responsesis also competitive, as shown in Figure 0.0 1 ...... I ...... I . .‘.- 9B. In the presenceof 500 PM 3-APA, the dose-responsecurve for GABA was shifted to the right without a change in the lo-' 1 o+ 1 o-5 1 o-4 1o-3 maximum response. The calculated inhibition constant for GABA Concentration (M) 3-APA was 43.1 PM. Figure 9. Effect of 3-APA, a selective GABA, agonist, on GABA re- Glycine receptor antagonists sponses of H4 horizontal cells. A, Inhibition curve for 3-APA. Data are from five cells. The continuous curve is a mathematical fit of the data Portions of the pl subunit are as homologousto the subunits points to Equation 2 with n = 1.8 1 and IC,, = 165 PM, and they indicate of glycine receptorsas to the subunitsof GABA, receptors(Cut- an inhibition constant (K,) of 47 PM. B. Dose-response curves for GABA ting et al., 1991). Since H4 horizontal cells of perch possess on H4 horizontal cells in the presence of 500 FM 3-APA. Data are from glycine receptors (Zhou et al., 1993) an obvious question is five cells. The continuous curve is a mathematical fit of the data points to Equation 1 with parameters of n = 1.27 and EC,, = 38.2 KM. The whether the GABA responsesobserved here could be mediated calculated inhibition constant is 43.1 FM. See Figure 4 for more details. through glycine receptors. To test this, , a competitive glycine receptor antagonist, was coapplied with GABA. Strych- nine was relatively ineffective in blocking the GABA responses. be termed GABA, receptors (Shimada et al., 1992; Woodward In the presenceof 500 PM strychnine, GABA responseselicited et al., 1993). with 10 PM GABA were 60% of control responses(0.60 f 0.11, Becausethe GABA receptorsobserved in H4 horizontal cells n = 4). are linked to a chloride channel, it might be argued that the receptorsmediating theseresponses are a GABA, receptor sub- Discussion type. However, the GABA responseselicited from H4 horizon- The GABA responseselicited from H4 horizontal cells of the tal cells of the white perch differ physiologically and pharma- white perch are clearly distinct from classicGABA, and GABA, cologically from classicGABA, receptor responsesin every way receptor-mediated responses.On the other hand, the responses examined so far. Physiologically, GABA responsesfrom H4 observed in H4 horizontal cells are very similar to the GABA horizontal cells are sustained, showing no sign of desensitiza- responsesobtained from oocytes expressingbovine retina mRNA tion, and they have a slow time course of recovery. Pharma- and mRNA coding for the human pl GABA channel subunit cologically, the GABA responsesfrom H4 horizontal cells can- (Kusamaet al., 1993; Woodward et al., 1993). Thus, we support not be blocked by bicuculline, are more sensitiveto GABA than the view that the receptors mediating the GABA responsesin are typical GABA, receptors, and they cannot be modulated by H4 horizontal cells are unique receptors and that they should either barbiturates or benzodiazepines (Qian and Dowling, 4306 Qian and Dowling * Pharmacology of GABA Receptors

1993a). Furthermore, the responses are less sensitive to mus- and that are inhibited by high concentrations of isoguvacine cimol than to GABA itself (Qian and Dowling, 1993a), and (Drew and Johnston, 1992). Such properties are very similar to muscimol acts as a partial agonist on H4 horizontal cells. In the GABA responses from H4 horizontal cells and the responses addition, as we show here, none of the GABA analogs in a of oocytes expressing bovine retinal mRNA and the p 1 subunit. partially folded conformation (isoguvacine, isonipecotic acid, The one difference between the results of Johnston et al. and THIP, and P4S) potently activates the GABA responses on H4 the more recent results involves the effectiveness of &-amino horizontal cells. All of these compounds strongly activate GA- crotonic acid (CACA), a GABA analog conformationally con- BA, receptors, but they either have no effect, act as partial strained around one of its C-C bonds. Whereas the binding sites agonists, or act as antagonists on the GABA receptors of H4 described by Johnston et al. were more sensitive to CACA than horizontal cells. On the other hand, 14AA, a GABA analog in to its more extended tram isomer (TACA, truns-aminocrotonic a fully extended conformation, strongly inhibits the GABA re- acid), the GABA responses elicited from the perch horizontal sponses on H4 horizontal cells. Indeed, the inhibition constant cells and from oocytes expressing pl subunits or retinal mRNA (K,) of 14AA on GABA, receptors is comparable to the EC,, are more sensitive to TACA than to CACA (Kusama et al., value of GABA on GABA, receptors, suggesting that the fully 1993; Woodward et al., 1993; H. Qian and J. E. Dowling, un- extended GABA conformation is preferred for GABA,. receptor published observation). This may indicate that GABA, recep- binding. Thus, our results indicate that a different conformation tors are somewhat heterogeneous, similar to GABA, and GA- of GABA is required to interact with the receptors on H4 hor- BA, receptors. Indeed, a second p subunit (~2) has now been izontal cells than with GABA, receptors. This notion is sup- described and more may exist (Cutting et al., 1992). Also, it ported also by the observations that competitive antagonists of should be noted that bicuculline-resistant GABA responses have GABA, receptors, including bicuculline, hydrastine, and SR- been reported in several invertebrate preparations. These re- 95531, have little effect on the GABA responses of H4 hori- ceptors may be variant members of the GABA, receptor family. zontal cells. Furthermore, the H4 horizontal cell receptors are It is also possible that the low potency of CACA reflects a more sensitive to picrotoxin than are classic GABA, receptors, discrepancy between receptor binding studies and physiological and picrotoxin acts as both a competitive and noncompetitive studies. antagonist on these receptors. These receptors are also relatively insensitive to TBPS, another GABA, receptor chloride . These latter observations indicate that the chloride References channels linked to the GABA receptors on H4 horizontal cells Akaike N, Hattori K, Oomura Y, Carpenter D (1985) Bicuculline and also differ from the channels linked to classic GABA, receptors. picrotoxin block y-aminobutyric acid-gated Cl- conductance by dif- There is also molecular evidence that GABA, receptors are ferent mechanisms. Experientia 41:70-7 1. distinct from GABA, receptors. Although the pl subunit was Alger BE, Nicoll RA (1982) Pharmacological evidence for two kinds initially thought to be a GABA, receptor subunit, portions of of GABA receptor on rat hippocampal pyramidal cells studied in vitro. J Physiol (Lond) 328: 125-14 1. it are as homologous to glycine receptor subunits as to GABA, Arakawa T, Okada Y (1988) Excitatory and inhibitory action ofGABA receptor subunits (Cutting et al., 199 1). More importantly, the on synaptic transmission on the guinea pig superior colliculus slices. GABA responses of oocytes coexpressing both the pl subunits Eur J Pharmacol 158:2 17-224. and individual GABA, subunits are indistinguishable from the Bormann J (1988) Electrophysiology of GABA, and GABA, receptor responses of oocytes expressing the p 1 subunit alone, indicating subtypes. Trends Neurosci 1 1: 112-l 16. Bowery NC, Collins JF, Hudson AL, Neal MJ (1978) Isoguvacine, that the GABA, and pl subunits do not aggregate (Shimada et isonipecotic acid, muscimol and N-methyl isoguvacine on the GABA al., 1992). In oocytes expressing total retinal mRNA, two dis- receptor in rat sympathetic ganglia. Experientia 34: 1193-l 195. tinct types of GABA responses were observed, a classic GABA, Burt DR, Kamatchi CL (199 1) GABA, receptor subtypes: from phar- receptor response and a GABA, receptor-mediated response macology to molecular biology. FASEB J 5:29 16-2923. Cutting CR, Lu L, O’Hara BF, Kasch LM, Montrose-Rafizadeh C, (Polenzani et al., 1991). This also indicates that GABA, and Donovan DM, Shimada S, Antonarakis SE, Guggino WB, Uhl CR, GABA,. subunits aggregate independently of one another. GA- Kazazian HH (199 1) Cloning of the y-aminobutyric acid (GABA) BA, subunits readily aggregate in oocytes to form heteroligo- pl cDNA: a GABA subunit highly expressed in the retina. Proc Nat1 merit receptors with the properties of classic GABA, receptors Acad Sci USA 88:2673-2677. (Seeburg et al., 1990; Burt and Kamatchi, 199 l), whereas pl Cutting CR, Cunistin S, Zoghbi H, O’Hara B, Seldin MF, Uhl CR (1992) Identification of a putative GABA p2 receptor subunit cDNA subunits aggregate in oocytes to form homooligomeric GABA, and colocalization of the genes encoding p2 and pl to human chro- receptors (Shimada et al., 1992). The similarity of the GABA mosome 6q 14-q2 1 and mouse chromosome 4. Genomics 12:80 l- responses of H4 horizontal cells to those of oocytes expressing 806. only the pl subunit suggests that the functional receptors on the Dong CJ, Picaud SA, Werblin FR (1994) GABA transporters and GABAc-like receptors on catfish cone- but not rod-driven horizontal H4 horizontal cells could be homooligomeric. cells. J Neurosci:in press. Woodward et al. (1993) have recently proposed that the ter- Dowling JE, Pak MW, Lasater EM (1985) White perch horizontal cells minology p-like GABA receptors be used for these receptors, in culture: methods, morphology and process growth. Brain Res 360: rather than GABA,. receptors. We do not feel that this is ap- 331-338. propriate terminology. As noted above, the p subunit was first Drew CA, Johnston CAR (1992) Bicuculline- and baclofen-insensitive y-aminobutyric acid binding to rat cerebellar membranes. J Neuro- thought to be a GABA, subunit and is often included in listings them 58:1087-1091. of GABA, receptor subunits (Burt and Kamatchi, 199 1; Cutting Drew CA, Johnston CAR, Weatherby RP (1984) Bicuculline-insen- et al., 1991); thus, the terminology “p-like GABA receptor” sitive GABA receptors: studies on the binding of (-)-baclofen to rat suggests that these receptors are a subtype of GABA, receptor. cerebellar membrane. Neurosci Lett 52:3 17-32 1. Feigenspan A, Wassle H, Bormann J (1993) Pharmacology of GABA We prefer the term GABA, receptor, which was first introduced receptor Cll channels in rat retinal bipolar cells. Nature 361:159- by Johnston (Johnston et al., 1986) to describe receptor-me- 162. diated responsesthat are bicuculline- and baclofen-insensitive, Fenwick EM, Marty A, Neher E (1982) A patch-clamp study of bovine The Journal of Neuroscience, July 1994, 74(7) 4307

chromaffin cells and of their sensitivity to acetylcholine. J Physiol Olsen RW (1982) Drug interactions at the GABA receptor-ionophore (Lond) 331:577-597. complex. Annu Rev Pharmacol Toxicol 22:245-277. Gage PW (1992) Activation and modulation of neuronal K+ channels Polenzani L, Woodward RM, Miledi R (1991) Expression of mam- by GABA. Trends Neurosci 15:46-5 1. malian y-aminobutyric acid receptors with distinct pharmacology in Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ (1981) Im- Xenopus oocytes. Proc Nat1 Acad Sci USA 88:43 18-4322. proved patch-clamp techniques for high resolution current recordings Qian H, Dowling JE (1992) A novel GABA response from rod-driven from cells and cell free membrane patches. Pfluegers Arch 391:85- horizontal cells of the white perch retina. Sot Neurosci Abstr 18:40 1. 100. Oian H. Dowlina JE (1993a) Novel GABA responses from rod-driven Heaulme M, Chambon J-P, Leyris R, Molimard J-C, Wermuth CG, . retinal horizontal cells. Nature 36 1: 162-I 64.. Biziere K (1986) Biochemical characterization of the interaction of Oian.-~ H. Dowline JE (1993b) GABA responses on retinal bipolar cells. three pyridazinyl-GABA derivatives with the GABA, receptor site. Biol Bull lSS:?i2. ~ ’ Brain Res 384:224-23 1. Qian H, Malchow RP, Ripps H (1993) Gap-junctional properties of Hill DR, Bowery NG (198 1) ‘H-baclofen and )H-GABA bind to bi- electrically coupled skate horizontal cells in culture. Vis Neurosci 10: cuculline-insensitive GABA, sites in rat brain. Nature 290: 149-152. 287-295. Huane JH. Johnston GA t 1990) (+)-Hvdrastine. a uotent competitive Renterghem CV, Bilbe G, Moss S, Smart TG, Constanti A, Brown DA, ant&o&t at mammalian GAB&receptors. Br J Pharmacol99:727. Barnard EA (1987) GABA receptors induced in Xenopus oocytes Inoue M, Akaike N (1988) Blockade of y-aminobutyric acid-gated by chick brain mRNA: evaluation of TBPS as a use-dependent chan- chloride current in frog sensory neurons by picrotoxin. Neurosci Res nel-blocker. Mol Brain Res 2:2 l-3 1. 5:380-394. Seeburg PH, Wisden W, Verdoom TA, Pritchett DB, Werner P, Herb Johnston GAR (1986) Multiplicity of GABA receptors. In: Receptor A, Ltiddens, Sprengel R, Sakmann B (1990) The GABA, receptor biochemistry and methodology, Vol 5, /GABA re- family: molecular and functional diversity. Cold Spring Harbor Symp ceptors and chloride channels (Olsen RW, Venter JC, eds), pp 570- Quant Biol 55:29-40. 571. New York: Liss. Shimada S, Cutting G, Uhl GR (1992) y-Aminobutyric acid A or C Johnston GAR (1992) GABA, agonists as targets for drug develop- receptor? y-Aminobutyric acid pI receptor RNA induces bicuculline-, ment. Clin Exp Pharmacol Physiol 19:73-78. -, and benzodiazepine-insensitive r-aminobutyric acid re- Johnston GAR, Curtis DR, Beart PM, Game CJA, McCulloch RM, sponses in Xenopus oocytes.-Mol Pharmacol 41:683-687. Twitchin B (1975) Cis and tvuns-4-aminocrotonic acid as GABA Sivilotti L. Nistri A (1989) Pharmacoloav of a novel effect of y-am- agonists of restricted conformation. J Neurochem 24: 157-160. inobutyric acid on the frog optic tectum in vitro. Em J Pharmacol Kerr DIB, Ong J, Johnston GAR, Abbenante J, Prager RH (1988) 164:205-212. 2-Hydroxysaclofen: an improved antagonist at central and peripheral Sivilotti L, Nistri A (199 1) GABA receptor mechanisms in the central GABA, receptors. Neurosci Lett 92:92-96. nervous system. Prog Neurobiol 36:35-92. Krogsgaard-Larsen P, Johnston GAR, Lodge D, Curtis DR (1977) A Slaughter MM, Pan ZH (1992) The physiology of GABA, receptors new class of GABA agonist. Nature 268:53-55. in the vertebrate retina. Prog Brain Res 90:47-60. Krogsgaard-Larsen P, Schousboe EFA, Curtis DR, Lodge D m(1980) Tehrani MHJ, Vaidyanathaswamy R, Verkade JG, Barnes EM Jr (I 986) Piperidine-4-sulphonic acid, a new specific GABA agonist. J Neu- Interaction of t-butylbicyclophosphorothionate with y-aminobutyric rochem 341165-759. acid-gated chloride channels in cultured cerebral neurons. J Neuro- Krogsgaard-Larsen P, Falch E, Hjeds H (1985) Heterocyclic analogues them 46: 1542-l 548. of GABA: chemistry, molecular pharmacology and therapeutic as- Ticku MK (1986) bind sites on the benzodiazepine/GABA pects. Prog Med Chem 22:67-l 20. receptor. In: Receptor biochemistry and methodology, Vol 5, Ben- Kusama T, Spivak CE, Dawson V, Schaeffer JC, Uhl G (1993) Phar- zodiazepine/GABA receptors and chloride channels (Olsen RW, Ven- macology of GABA p 1 and GABA a/p receptors expressed in Xenopus ter JC, eds), pp 195-207. New York: Liss. oocytes and COS cells. Br J Pharmacol 109:200-206. Woodward RM, Polenzani L, Miledi R (1992) Characterization o Lehoullier PF, Ticku MK (1989) The pharmacological properties of bicuculline/baclofen-insensitive y-aminobutyric acid receptors ex. GABA receptor-coupled chloride channels using 3bC!-influx in cul- messed in Xenoaus oocvtes I. Effects of Cl- channel inhibitors. MO tured suinal cord neurons. Brain Res 487:205-214. Pharmacol 42: 165-173: Lukasiewicz PD, Maple BR, Werblin FS (1993) A novel GABA re- Woodward RM, Polenzani L, Miledi R (1993) Characterization o ceptor on bipolar cell terminal in the tiger salamander retina. J Neu- bicuculline/baclofen-insensitive (p-like) y-aminobutyric acid recep rosci 14: 1202-l 2 12. tors expressed in Xenopus oocytes. II. Pharmacology of y-aminobu McPherson GA (1989) A mathematical approach to receptor char- tyric acid, and y-aminobutyric acid, receptor agonists and antago. acterization. In: Receptor pharmacology and function (Williams M, nists. Mol Pharmacol 43:609-625. Glennon RA. Timmermans P. eds). DD 47-84. New York: Dekker. Zhou ZJ, Fain GL, Dowling JE (1993) The excitatory and inhibiton Myers JM, Tunnicliff G (1988) Bicucuihne-insensitive GABA binding receptors on horizontal cells isolated from the white perch’ to catfish neuronal membranes. Neurochem Int 12: 125-l 29. retina. J Neurophysiol 70:8-19.