Interaction of H⁺ and Zn²⁺ on recombinant and native rat neuronal GABA_A receptors

Abstract

1. The interaction of Zn²⁺ and H⁺ ions with GABA_A receptors was examined using Xenopus laevis oocytes expressing recombinant GABA_A receptors composed of subunits selected from α1, β1, γ2S and δ types, and by using cultured rat cerebellar granule neurones.

2. The potency of Zn²⁺ as a non-competitive antagonist of GABA-activated responses on α1β1 receptors was reduced by lowering the external pH from 7.4 to 5.4, increasing the Zn²⁺ IC₅₀ value from 1.2 to 58.3 μm. Zinc-induced inhibition was largely unaffected by alkaline pH up to pH 9.4.

3. For α1β1δ subunits, concentration-response curves for GABA were displaced laterally by Zn²⁺ in accordance with a novel mixed/competitive-type inhibition. The Zn²⁺ IC₅₀ at pH 7.4 was 16.3 μm. Acidification of Ringer solution resulted in a reduced antagonism by Zn²⁺ (IC₅₀, 49.0 μm) without affecting the type of inhibition. At pH 9.4, Zn²⁺ inhibition remained unaffected.

4. The addition of the γ2S subunit to the α1β1δ construct caused a marked reduction in the potency of Zn²⁺ (IC₅₀, 615 μm), comparable to that observed with α1β1γ2S receptors (IC₅₀ 639 μm). GABA concentration-response curves were depressed in a mixed/non-competitive fashion.

5. In cultured cerebellar granule neurones, Zn²⁺ inhibited responses to GABA in a concentration-dependent manner. Lowering external pH from 7.4 to 6.4 increased the IC₅₀ from 139 to 253 μm.

6. The type of inhibition exhibited by Zn²⁺ on cerebellar granule neurones, previously grown in high K⁺-containing culture media, was complex, with the GABA concentration-response curves shifting laterally with reduced slopes and similar maxima. The Zn²⁺-induced shift in the GABA EC₅₀ values was reduced by lowering the external pH from 7.4 to 6.4.

7. The interaction of H⁺ and Zn²⁺ ions on GABA_A receptors suggests that they share either a common regulatory pathway or coincident binding sites on the receptor protein. The apparent competitive mode of block induced by Zn²⁺ on α1β1δ receptors is shared by GABA_A receptors on cerebellar granule neurones, which are known to express δ-subunit-containing receptors. This novel mechanism is masked when a γ2 subunit is incorporated into the receptor complex, revealing further diversity in the response of native GABA_A receptors to endogenous cations.

The divalent cation Zn²⁺ is distributed widely throughout the central nervous system (CNS) and concentrated principally within neurones (Frederickson, 1989). Zinc-containing nerve fibres are mostly located in the telencephalon and Zn²⁺ ions have been located within synaptic vesicle-type structures in nerve terminals (Frederickson, 1989; Smart, Xie & Krishek, 1994). A number of different physiological roles for Zn²⁺ have been proposed, including: acting as a packaging or storage agent for neurotransmitters within synaptic vesicles (Frederickson, 1989); being released from nerve terminals following stimulation and interacting with postsynaptic excitatory and inhibitory amino acid receptors in the CNS (Peters, Koh & Choi, 1987; Westbrook & Mayer, 1987; Mayer & Vykicky, 1989; Mayer, Vyklicky & Westbrook, 1989; Smart & Constanti, 1990); exerting a neuromodulatory role, influencing γ-aminobutyric acid_A (GABA_A) receptor-mediated synaptic transmission in the hippocampus (Xie & Smart, 1991; Nakazawa, Inoue, Watano, Koizumi & Inoue, 1995); and a developmental role, influencing the structural and functional maturation of some classes of neurones, particularly in the cerebellum (Dvergsten, Fosmire, Ollerich & Sandstead, 1984).

An intriguing aspect of the functional role of Zn²⁺ in the CNS involves the interaction with the GABA_A receptor. Zinc inhibits GABA-activated responses on a variety of neuronal preparations (Smart et al. 1994; Harrison & Gibbons, 1994); however, not all GABA_A receptors, particularly those in adult brain slices, are sensitive to Zn²⁺ (Smart et al. 1994). A comparison of embryonic and adult rat sympathetic neurones revealed that Zn²⁺ inhibition of GABA_A responses was correlated with neuronal age, with younger neurones being more sensitive to Zn²⁺ (Smart & Constanti, 1990; Smart, 1992). By utilizing recombinant GABA_A receptors (see Sieghart, 1995 and Rabow, Russek & Farb, 1995, for review), later studies demonstrated that inhibition by Zn²⁺ was dependent on the receptor subunit composition, in particular the presence of the γ subunit (Draguhn, Verdoorn, Ewert, Seeburg & Sakmann, 1990; Smart, Moss Xie & Huganir, 1991). This principle may in part underlie the variation in Zn²⁺ sensitivity observed with native GABA_A receptors (Smart et al. 1994).

In comparison with Zn²⁺, protons are naturally occurring in all tissues and thus constitute part of the physiological milieu to which receptors are normally exposed. Lowering of external pH can occur during a number of pathological processes, including ischaemic episodes, epileptiform activity and anoxia (Chesler, 1990; Chesler & Kaila, 1992) and is also, but less dramatically, a feature of normal synaptic transmission (Chen & Chesler, 1992) and activation of GABA_A receptors per se (Kaila & Voipio, 1987; Kaila, 1994). By the nature of their size and charge, protons are capable of interacting with many types of voltage- and ligand-gated ion channels. GABA-activated responses mediated by invertebrate muscle (Takeuchi & Takeuchi, 1967; Smart & Constanti, 1982; Pasternack, Bountra, Voipio & Kaila, 1992) and vertebrate neuronal GABA_A receptors (Groul, Barker, Huang, MacDonald & Smith, 1980; Gallagher, Nakamura & Shinnock-Gallagher, 1983; Smart, 1992; Kaila, 1994; Krishek, Amato, Connolly, Moss & Smart, 1996; Pasternack, Smirnov & Kaila, 1996) are sensitive to changes in the external pH. Furthermore, recombinant GABA_A receptors are differentially modulated by external pH, depending on the subunit composition and may partly explain why native GABA_A receptors exhibit variable sensitivities to protons (Krishek et al. 1996).

The potential for H⁺ and Zn²⁺ to interact at GABA_A receptors was suggested by the sensitivity of the Zn²⁺ binding site on invertebrate muscle GABA receptors to external pH. Increasing the concentration of H⁺ enhanced the GABA-induced conductance, whilst the antagonistic effect of Zn²⁺ was markedly reduced, suggesting that Zn²⁺ and H⁺ competed for a similar site on this GABA receptor (Smart & Constanti, 1982). The present study investigates the extent of this phenomenon by studying the pH dependence of the Zn²⁺-induced inhibition of GABA-activated responses mediated by recombinant and native neuronal GABA_A receptors. The results demonstrate that Zn²⁺ and H⁺ can interact at vertebrate GABA_A receptors, and this interaction is dependent on the GABA_A receptor subunit composition. Moreover, a novel modulatory role for Zn²⁺ was observed for GABA_A receptors composed of α1β1δ subunits and native GABA_A receptors in cerebellar granule neurones. These results lead to the possibility that both H⁺ and Zn²⁺ could be used as supplementary tools for probing the in vivo composition of native neuronal GABA_A receptors, in addition to revealing a potential physiological mechanism that could regulate GABA_A receptor function.

METHODS

Expression vector construction

Murine α1, β1, γ2S and δ subunit cDNAs were incorporated into the mammalian expression vector pGW1 (British Biotechnology) as described previously (Krishek, Xie, Blackstone, Huganir, Moss & Smart, 1994).

Cell preparation

Xenopus laevis oocytes

Xenopus laevis were anaesthetized by immersion in a 0.2% w/v solution of ethyl-m-aminobenzoate (Tricaine). Stage IV-VI Xenopus oocytes were surgically removed and microinjected with equimolar ratios of murine cDNAs corresponding to combinations of α1, β1, γ2S and δ GABA_A receptor subunits and subsequently cultured in a modified Barth's medium (MBM) comprising (mm): 110 NaCl, 1 KCl, 2.4 NaHCO₃, 7.5 Tris-HCl, 0.33 Ca(NO₃)₂, 0.41 CaCl₂, 0.82 MgSO₄; and 50 μg ml⁻¹ gentamicin; pH 7.6 as previously described (Krishek et al. 1994). All the Xenopus oocyte experiments were performed at room temperature (18–23°C). Cells were used after 2 days and accepted for study if they possessed membrane potentials of −30 to −60 mV and input resistances of 1–5 MΩ.

Cerebellar granule neurones

Sprague-Dawley postnatal day 4 (P4) rats were killed by decapitation and the cerebella removed and placed in ice-cold Hanks’ Ca²⁺/Mg²⁺-free balanced salt solution (HBSS). Meninges were removed and the cerebella were placed in the plating medium MEM 10/10 containing: MEM Earle's salts, 10% v/v horse serum, 10% v/v fetal calf serum, 600 mg l⁻¹ glucose, 2 mm glutamine and 100 u ml⁻¹ penicillin G and 100 μg ml⁻¹ streptomycin. The cerebella were subsequently taken up into a 5 ml syringe and expelled through a 210 μm Nylon mesh into a 10 ml centrifuge tube. After allowing the tissue to sediment for 5 min, the microexplants and dissociated cells were plated onto poly-l-lysine-coated glass coverslips and maintained at 37°C in a 95% air and 5% CO₂ atmosphere. After 18 h in vitro, the cultures were treated with 10 μm cytosine arabinoside for 24 h, then the media replaced with MEM 10 supplemented with 20 mm KCl (total concentration 25.37 mm). MEM 10 medium had a similar composition to MEM 10/10 but lacked the fetal calf serum. All media was replenished on days 5 and 7 after plating. Neurones were used for experimentation between days 6 and 11 in vitro, and possessed membrane potentials of −60.5 ± 12.5 mV (mean ± s.e.m.).

Electrophysiology

Intracellular recording

Membrane currents were recorded from Xenopus laevis oocytes, retaining their follicular cell envelope, using a two-electrode voltage-clamp method in conjunction with an Axoclamp-2A amplifier. Oocytes were superfused at 8–10 ml min⁻¹ (bath volume 0.5 ml) with an amphibian Ringer solution containing (mm): 110 NaCl, 2 KCl, 5 Hepes and 1.8 CaCl₂; pH 7.4. The microelectrodes were filled with 3 m KCl (voltage) and 0.6 m K₂SO₄ (current) and possessed resistances of 5–10 MΩ and 1–2 MΩ, respectively. All data were recorded on a Gould 2200 ink-jet pen recorder.

Whole-cell recording

Experiments on the cerebellar granule neurones were performed using a List EPC5 amplifier in the whole-cell recording mode. Patch electrodes (1–7 MΩ) were filled with a solution containing (mm): 140 KCl, 2 MgCl₂, 1 CaCl₂, 10 Hepes, 11 EGTA and 2.5 adenosine triphosphate. The cells were viewed under Nomarski differential interference contrast optics (Nikon Optiphot-2) and continuously superfused with a Krebs solution containing (mm): 140 NaCl, 4.7 KCl, 1.2 MgCl₂, 2.5 CaCl₂, 10 Hepes and 11 glucose. Membrane currents were filtered at 1.5 kHz (−3 dB, 8-pole Butterworth filter, 48 dB octave⁻¹) and recorded on a Gould WindoGraf 40–8474-00 pen recorder.

Analysis of GABA-activated membrane conductance

GABA-induced membrane conductance (ΔG_N) was calculated in Xenopus oocytes by subtraction from the resting membrane conductance and normalized to the response induced by 10 μm GABA at pH 7.4. In the absence and presence of GABA, membrane conductances were frequently determined by stepping the membrane potential from a holding potential range of −30 to −40 mV with a −10 mV voltage command step (1 s duration and 0.2 Hz frequency). These data were used to construct equilibrium concentration-response relationships for GABA. The data were fitted with a logistic state function of the form:

where ΔG and ΔG_max represent the normalized GABA-induced conductance at a given concentration and the maximum conductance induced by a saturating concentration of GABA, respectively. A represents the concentration of GABA and EC₅₀ defines the concentration of GABA inducing 50% of the maximum response. n_H is the Hill coefficient. A Levenberg-Marquardt non-linear least-squares fitting routine was used.

The Zn²⁺ concentration-inhibition curves were fitted with an antagonist inhibition model described by:

where ΔG_N′ and ΔG_N represent the normalized GABA-induced conductance at a given concentration in the presence and absence of antagonist, respectively. B represents the Zn²⁺ concentration and IC₅₀ defines the concentration of Zn²⁺ producing a 50% inhibition of the GABA response.

Drugs and solutions

Drugs and solutions were bath-applied in the Xenopus oocyte experiments. For rapid application to the neurones, solutions were applied close to (100–200 μm) the cells using a Y-tube manufactured from glass electrode tubing. The Y-tube was positioned at the periphery of the × 40 water immersion objective. Solutions were driven through the Y-tube in response to a vacuum. Drug application was achieved by isolating the vacuum using a solenoid valve, thereby allowing solution to flow out of the Y-tube over the recorded cells by gravity. A complete exchange of solution around the cells was effected within 50–80 ms. Throughout, Zn²⁺ was pre-applied for a minimum of 1 min prior to GABA application in Zn²⁺-containing Ringer (oocytes) or Krebs (cerebellar granule neurones) solutions. The zwitterionic nature of the GABA molecule is not substantially affected over the range of external pH from 5.4 to 9.5 (Krishek et al. 1996). Similarly, it is unlikely that the inhibition caused by Zn²⁺ of GABA-activated responses is dependent upon complexation, since the slopes of the GABA concentration-response curves are not increased by Zn²⁺ (Smart & Constanti, 1982).

RESULTS

Inhibitory effect of Zn²⁺ on α1β1 GABA_A receptors: modulation by external pH

Zinc is a potent antagonist at cultured neuronal GABA_A receptors (Westbrook & Mayer, 1987; Smart & Constanti, 1990) and also at invertebrate GABA receptors (Smart et al. 1994). Previous evidence suggests that the Zn²⁺ binding site on the lobster muscle GABA receptor involved at least one histidine residue that could also competitively bind H⁺. Accordingly, lowering the external pH reduced the effectiveness of Zn²⁺. If a similar binding site exists on selected vertebrate GABA_A receptors, then it should be possible to regulate the Zn²⁺-induced antagonism by changing the pH of the external solution.

For recombinant α1β1 GABA_A receptors expressed in Xenopus oocytes, the whole-cell membrane conductance activated by 10 μm GABA was inhibited by 51.07 ± 1.6% (mean ± s.e.m., n = 23 oocytes) with 1 μm Zn²⁺ at pH 7.4 (Fig. 1A). Lowering the pH of the Ringer solution over the range 7.4 to 5.4 resulted in a potentiation of the GABA-induced response; however, the antagonism induced by 1 μm Zn²⁺ was now effectively abolished (Fig. 1B). In contrast, increasing the external pH from 7.4 to 9.4 reduced the response to 10 μm GABA, but the antagonism induced by 1 μm Zn²⁺ was apparently unaffected (51.7 ± 3.25%; Fig. 1C).

Figure 1. Sensitivity of the Zn²⁺ inhibition of GABA-activated conductances in Xenopus oocytes expressing α1β1 GABA_A receptors.

Membrane currents and conductance changes evoked by bath-applied 10 μm GABA in the absence and presence of 1 μm zinc at pH 7.4 (A), 5.4 (B) and 9.4 (C), respectively. The GABA-activated responses in all oocytes increased at low pH and decreased at high pH. Records are selected from two oocytes. Current calibration of 100 nA corresponds to A and B, and 250 nA corresponds to C. GABA was applied for the duration indicated by the horizontal bars. Recovery responses were obtained after 3 min in Ringer solution at the respective external pH values following exposure to Zn²⁺. Membrane conductance was ascertained by regularly applying hyperpolarizing voltage-clamp commands (−10 mV amplitude, 1 s duration, frequency 0.2 Hz) from the holding potential of −40 mV.

The ability of H⁺ to reduce the inhibitory potency of Zn²⁺ was examined by constructing concentration-inhibition curves for Zn²⁺ antagonizing the 10 μm GABA-activated response. At pH 7.4, the Zn²⁺ IC₅₀ was 1.2 ± 0.4 μm (Fig. 2A, Table 1). Increasing the alkalinity of the Ringer solution to pH 9.4 produced little change in the inhibition curve or the IC₅₀ (1.4 ± 0.5 μm; Fig. 2A; Table 1); however, increasing the H⁺ concentration 100-fold to pH 5.4, caused a parallel displacement in the inhibition curve and a substantial increase in the Zn²⁺ IC₅₀ to 58.3 ± 9.4 μm (Fig. 2A; Table 1). For a complete block of the GABA-activated response, 1–2 mm Zn²⁺ was required at pH 5.4.

Figure 2. Concentration-response relationships for Zn²⁺ inhibition of GABA-activated conductances on α1β1 GABA_A receptors.

A, left panel, normalized zinc concentration-inhibition curves were constructed for GABA at external pH 5.4, 7.4 and 9.4. For this and all subsequent zinc inhibition curves, the GABA-induced conductances were normalized to the response evoked by 10 μm GABA in the absence of zinc. All data were fitted according to the antagonist inhibition model (see Methods) and IC₅₀ values determined (Table 1; n = 9). The points represent the means ± s.e.m. A, right panel and B, indicate GABA equilibrium concentration conductance curves constructed in the absence and presence of 1 μm zinc at pH 7.4 and 9.4, or 50 μm zinc at pH 5.4 and normalized to the response induced by 10 μm GABA in the absence of zinc at pH 7.4. The data were fitted with the logistic equation (see Methods; n = 18) and EC₅₀ values and Hill coefficients determined (Tables 2 and 3).

Table 1.

Analysis of the Zn²⁺ inhibition plots for GABA in Xenopus laevis oocytes injected with murine GABA_A receptor cDNAs

	IC50 (μm)
Subunit composition	pH 7.4	pH 5.4	pH 9.4
α1β1	1.2 ± 0.4	58.3 ± 9.4	1.4 ± 0.5
α1β1δ	16.3 ± 3.5	49.0 ± 10.2	16.4 ± 4.5
α1β1γ2S	639.9 ± 34.9	–	–
α1β1γ2Sδ	615.4 ± 84.3	–	–

The IC₅₀ values were obtained from fitting the antagonist inhibition model to the normalized curves (see Methods) for the inhibition of the response to 10 μm GABA. All values are means ± s.e.m. (n = 24).

The effect of external pH on the steady-state inhibition of the GABA-induced conductance by Zn²⁺ was studied using GABA equilibrium concentration-response curves. To compare the inhibition induced by Zn²⁺ at each pH, equipotent Zn²⁺ concentrations approximating the IC₅₀ values were selected for study. At pH 7.4, the GABA concentration-response curve was depressed at all GABA concentrations by 1 μm Zn²⁺, consistent with a non-competitive type of inhibition (Fig. 2A). The GABA EC₅₀ and Hill coefficient were unaffected by Zn²⁺ (Tables 2 and 3). Lowering the external pH from 7.4 to 5.4 enhanced the GABA-induced conductance and laterally displaced the concentration-response curve with a corresponding reduction in the EC₅₀ to 3.2 ± 0.18 μm. The inhibitory effect of Zn²⁺ was markedly reduced with 50 μm now required to produce an equivalent depression in the GABA concentration-response curve compared with 1 μm Zn²⁺ at pH 7.4 (Fig. 2B). The antagonism produced by 50 μm Zn²⁺ at pH 5.4 was of a mixed/non-competitive type with a depression in the maximum response and small lateral curve displacement. There was no change in the Hill coefficient, but the EC₅₀ for GABA was increased to 7.4 ± 0.5 μm (Tables 2 and 3). In contrast, GABA concentration-response curves at pH 9.4 were depressed compared with those at pH 7.4. The inhibition by Zn²⁺ at GABA concentrations > 10 μm was unaffected by the increased external pH; however, at GABA concentrations < 10 μm, inhibition was reduced and eventually abolished at GABA concentrations < 2 μm (Fig. 2B). Unusually, the type of inhibition observed at pH 9.4 was consistent with uncompetitive antagonism, with the inhibition increasing with the GABA concentration (Fig. 2B). The Hill coefficient for the GABA concentration-response curves in the presence of 1 μm Zn²⁺ was unaffected, whereas the EC₅₀ was reduced from 2.5 to 1 μm (Tables 2 and 3).

Table 2.

Analysis of the equilibrium concentration-response curves for GABA in the absence and presence of Zn²⁺ in Xenopus laevis oocytes injected with GABA_A receptor cDNAs

	Hill coefficient
	pH 7.4		pH 5.4		pH 9.4
Subunit composition	Control	+Zn2+	Control	+Zn2+	Control	+Zn2+
α1β1	0.95 ± 0.09	1.2 ± 0.1(1)*	1.1 ± 0.06	1.1 ± 0.07 (50)	1.2 ± 0.14	1.2 ± 0.2 (1)
α1β1δ	0.85 ± 0.05	0.8 ± 0.06 (10)	0.86 ± 0.05	0.75 ± 0.03 (50)*	0.85 ± 0.15	0.7 ± 0.12 (10)
		0.6 ± 0.03 (100)*
α1β1γ2S	1.3 ± 0.4	1.2 ± 0.03 (5)	–	–	–	–
		1.5 ± 0.9 (300)
α1β1γ2Sδ	1.4 ± 0.06	1.5 ± 0.07 (10)	–	–	–	–
		1.4 ± 0.1 (300)

The Hill coefficient (n_H) data were determined from fitting the logistic model to the normalized curves (see Methods) with all values given as means ± s.e.m. (n = 50). Numbers in parentheses indicate the Zn²⁺ concentration (μm)

^*represents a significant change from control, P < 0.05.

Table 3.

Analysis of the equilibrium concentration-response curves for GABA in the absence and presence of Zn²⁺ recorded from GABA_A receptor constructs

	EC50 (μm)
	pH 7.4		pH 5.4		pH 9.4
Subunit composition	Control	+Zn2+	Control	+Zn2+	Control	+Zn2+
α1β1	5.87 ± 0.7	6.6 ± 0.5 (1)	3.2 ± 0.18	7.4 ± 0.5 (50)*	2.5 ± 0.3	1.0 ± 0.1 (1)*
α1β1δ	4.9 ± 0.41	28.4 ± 3.5 (10)*	4.9 ± 0.4	15.6 ± 1.2 (50)*	4.3 ± 0.34	47.8 ± 5.8 (10)*
		361 ± 53.1 (100)*
α1β1γ2S	9.7 ± 0.2	10.06 ± 0.21 (5)	–	–	–	–
		6 ± 0.3 (300)*
α1β1γ2Sδ	27.7 ± 0.96	27.7 ± 0.95 (10)	–	–	–	–
		16.2 ± 1.02 (300)*

The EC₅₀ data were determined from fitting the logistic model to the normalized curves (see Methods) with all values given as means ± s.e.m. (n = 50). Numbers in parentheses indicate the Zn²⁺ concentration (μm)

^*represents a significant change from control, P < 0.05.

Expression of α1, β1 and δ GABA_A receptor subunits: a novel mode of Zn²⁺ antagonism

The level of inhibition of GABA responses by Zn²⁺ is reduced by the presence of γ subunits and also affected by exchanging α1 for α2 or α3 subunits (Draguhn et al. 1990; Smart et al. 1991; White & Gurley, 1995). However, the influence of the δ subunit on Zn²⁺ inhibition has not been quantitatively established. Oocytes expressing α1, β1 and δ subunits were assumed to co-assemble into α1β1δ GABA_A receptors since the latter exhibited a distinctive sensitivity to H⁺ (Krishek et al. 1996). Zinc inhibition curves for α1β1δ subunit receptors indicated that increasing the external pH from 7.4 to 9.4 had little effect on the antagonism of responses to 10 μm GABA (IC₅₀ values were 16.3 ± 3.5 μm (pH 7.4) and 16.4 ± 4.5 μm (pH 9.4), n = 8; Fig. 3A; Table 1); however, increasing the H⁺ concentration (pH 5.4) caused the inhibition curve for Zn²⁺ to be laterally displaced with an increased IC₅₀ of 49.0 ± 10.2 μm (n = 4; Fig. 3A; Table 1).

Figure 3. Modulation by external pH of Zn²⁺ inhibition of GABA-induced responses recorded from oocytes expressing GABA_A receptor α1β1δ subunits.

A, left panel, zinc concentration-inhibition relationships constructed at external pH 5.4, 7.4 and 9.4, and fitted according to the antagonist inhibition model and IC₅₀ values determined (Table 1; n = 9). A, right panel, GABA equilibrium concentration-response curves constructed in the absence and presence of 10 and 100 μm Zn²⁺ at pH 7.4. B, GABA equilibrium concentration-response curves constructed in the absence and presence of 50 μm zinc (left, pH 5.4) and 10 μm zinc (right, pH 9.4). The EC₅₀ values and Hill coefficients for GABA were determined from the logistic model (Tables 2 and 3; n = 20).

This inhibitory profile of Zn²⁺ on α1β1δ constructs initially appeared similar to that observed with α1β1 subunit-containing receptors; however, analysis of GABA equilibrium concentration-response curves revealed that Zn²⁺ acted as an apparent competitive inhibitor on α1β1δ receptors, a feature not observed with αβ or αβγ GABA_A receptor constructs (Fig. 3A). At pH 7.4, the GABA EC₅₀ was increased by 10 μm Zn²⁺, consistent with lateral/competitive displacement of the GABA concentration-response curve (Tables 2 and 3). The Hill coefficient was unaltered by 10 μm Zn²⁺. Interestingly, at a higher Zn²⁺ concentration (100 μm) the GABA response curve was further displaced but with a reduced slope despite the maximum response remaining unaffected. This suggests that the mechanism of Zn²⁺ antagonism is not purely competitive (Fig. 3A).

On lowering the external pH from 7.4 to 5.4, the GABA response was enhanced as expected (Krishek et al. 1996) but to ensure an equipotent effect of Zn²⁺ at pH 5.4 compared with that observed at pH 7.4, 50 μm Zn²⁺ was now required. The inhibitory potency of Zn²⁺, whilst remaining mixed/competitive, was reduced approximately 5-fold as estimated from the displaced concentration-response curves (Fig. 3B). The Hill coefficient was reduced and the EC₅₀ for GABA was markedly increased in the presence of Zn²⁺ (Fig. 3B; Tables 2 and 3). At pH 9.4, the antagonism induced by an equipotent concentration of Zn²⁺ (approximate IC₅₀ concentration) was slightly increased from that observed at pH 7.4 (Fig. 3B), but still the mechanism of action was apparently mixed/competitive without changing the Hill coefficient whilst increasing the EC₅₀ for GABA (Fig. 3B; Tables 2 and 3).

Influence of the γ2S subunit in determining the sensitivity of the GABA_A receptor to inhibition by Zn²⁺

As the δ and γ2 subunits differentially influence the sensitivity of α1β1 subunit-containing GABA_A receptors with respect to H⁺ and Zn²⁺, the effect of co-expressing these subunits, within the same hetero-oligomeric structure was assessed by adding γ2S to GABA_A receptors composed of α1β1 and α1β1δ subunits. For both α1β1γ2S and α1β1γ2Sδ subunit combinations, Zn²⁺ inhibition curves revealed that the GABA-activated responses were only inhibited by relatively high Zn²⁺ concentrations (> 10–100 μm) with estimated IC₅₀ values of 639.9 ± 34.9 μm (α1β1γ2S) and 615.4 ± 84.3 μm (α1β1γ2Sδ; Fig. 4A and B; Table 1), respectively. Complete inhibition of the 10 μm GABA responses were not achieved over the range of 0.1 μm to 1 mm Zn²⁺. Whether the δ subunit differentially influenced the level and type of inhibition by Zn²⁺ of GABA responses transduced by α1β1γ2S GABA_A receptors was assessed using GABA concentration-response curves. Interestingly, for both α1β1γ2S and α1β1γ2Sδ receptor constructs, low concentrations of Zn²⁺ (5–10 μm) were essentially ineffective, but at a higher Zn²⁺ concentration (300 μm), the maximum conductance induced by saturating concentrations of GABA (500 μm) was reduced in a mixed/non-competitive manner, with little change in the Hill coefficients and moderate reductions in the GABA EC₅₀ values (Fig. 4A and B; Tables 2 and 3).

Figure 4. Regulation of α1β1γ2S and α1β1γ2Sδ GABA_A receptor function by Zn²⁺.

A and B, left panels, zinc concentration-inhibition relationships constructed for 10 μm GABA-activated responses recorded from α1β1γ2S and α1β1γ2Sδ GABA_A receptors, respectively. All data were fitted according to the antagonist inhibition model and IC₅₀ values determined (Table 1, n = 6). A and B, right panels, GABA equilibrium concentration-response curves in the absence and presence of 5 and 300 μm or 10 and 300 μm zinc recorded from oocytes expressing α1β1γ2S and α1β1γ2Sδ GABA_A receptors, respectively. The EC₅₀ values and Hill coefficients for GABA were determined from the logistic model (Tables 2 and 3, n = 12).

Cerebellar granule neurones: emulation of α1β1δ receptors with regard to Zn²⁺ antagonism

The shift from non-competitive to mixed/competitive inhibition of the GABA-induced conductance by Zn²⁺ on co-expressing the δ subunit with α1β1 subunit-containing GABA_A receptors is intriguing. The δ subunit mRNA is expressed most prominently in the granule cell layer of the cerebellum (Laurie, Wisden & Seeburg, 1992) and could participate in forming αβδ complexes. These neurones were therefore selected to determine whether native GABA_A receptors also exhibit a mixed/competitive Zn²⁺ block in accordance with α1β1δ receptors.

The whole-cell membrane current activated by 10 μm GABA was inhibited reversibly by 10–500 μm Zn²⁺ at pH 7.4. Lowering the pH of the Krebs solution resulted in a small alleviation of the Zn²⁺ block (Fig. 5A). The reduced inhibitory potency of Zn²⁺ at low pH was manifest from the concentration-inhibition curves for the antagonism of the 10 μm GABA-activated current. Acidification of the Krebs solution to pH 6.4 caused a small parallel displacement in the inhibition curve and increased the IC₅₀ to 253.7 ± 33.9 μm from 139.0 ± 17.2 μm at pH 7.4 (Fig. 5B).

Figure 5. Modulation of GABA-activated currents in cultured cerebellar granule neurones by H⁺ and Zn²⁺ ions.

A, whole-cell membrane currents activated by 10 μm GABA before and after the application of 10–500 μm Zn²⁺ at external pH values of 7.4 (upper panel) and 6.4 (lower panel). Recovery responses were obtained in control Krebs solution 5 min after Zn²⁺ application at the respective external pH. Records are from two neurones. The 100 pA calibration bar applies to the upper records. The holding potential was −40 mV. B, zinc concentration-inhibition curves for the antagonism of the 10 μm GABA-activated response at external pH values of 7.4 (^) or 6.4 (•) constructed from fifteen cells. The curves were generated from the inhibition model (see Methods), but since the inhibition saturated at approximately 20% of the control GABA-activated response, a constant was incorporated and the IC₅₀ values were determined.

To determine the mode of Zn²⁺-induced inhibition in these cells the whole-cell GABA-induced current was measured in the absence and presence of Zn²⁺. At either pH 7.4 or 6.4, an approximate 10-fold increase in the GABA concentration was required in the presence of 300 μm Zn²⁺ to induce currents of similar magnitude to the controls (Figs 6A and 7A). In addition, GABA concentration-response curves revealed that the maximum normalized current (I_N,max) induced by saturating concentrations of GABA was not significantly reduced by 300 μm Zn²⁺ (I_N,max control, 2.1 ± 0.04; +Zn²⁺, 2.0 ± 0.06 at pH 7.4; and control, 1.7 ± 0.07; +Zn²⁺, 1.66 ± 0.05 at pH 6.4 (Figs 6B and 7B)); however, the GABA EC₅₀ was increased by 300 μm Zn²⁺ from 10.76 ± 0.67 to 71.51 ± 10.11 μm (pH 7.4) and 18.05 ± 2.27 to 102.63 ± 12.20 μm (pH 6.4; Figs 6B and 7B). This lateral competitive-type shift in the control curve in the presence of Zn²⁺ with no apparent reduction in the maximum response was associated with a reduced Hill coefficient from 1.0 ± 0.05 to 0.7 ± 0.05 (pH 7.4) and 0.9 ± 0.06 to 0.7 ± 0.04 (pH 6.4), suggesting a more complex mode of antagonism for Zn²⁺ consistent with ‘mixed inhibition’ (Smart & Constanti, 1982).

Figure 6. Inhibition by Zn²⁺ of GABA-activated responses at pH 7.4 on cerebellar granule neurones.

A, whole-cell currents recorded at −40 mV holding potential for different concentrations of GABA (1 μm to 2.5 mm) in the absence (upper panel) and presence of 300 μm Zn²⁺ (lower panel). B, analysis of GABA concentration-response curves at pH 7.4 in the absence (^) and presence (•) of 300 μm Zn²⁺. The continuous line curve fit was generated by the agonist logistic model (see Methods). The dotted lines illustrate the corresponding curves for data obtained at pH 6.4 (see Fig. 7).

Figure 7. GABA-activated responses at pH 6.4 on cerebellar granule neurones: inhibition by Zn²⁺.

A, whole-cell currents recorded for 1 μm to 2.5 mm GABA in the absence (upper panel) and presence of 300 μm Zn²⁺ (lower panel). Holding potential, −40 mV. B, GABA concentration-response curves are plotted at pH 6.4 in the absence (^) and presence (•) of 300 μm Zn²⁺. The continuous line curve fit was provided by the logistic model (see Methods). For comparison, the dotted lines plot the corresponding curves for data obtained at pH 7.4 (see Fig. 6).

DISCUSSION

Zn²⁺-induced inhibition is dependent on the GABA_A receptor subunit composition: role of α and γ subunits

Using recombinant GABA_A receptors expressed in Xenopus oocytes, the potency of Zn²⁺ as a non-competitive antagonist on α1β1 subunit constructs was clearly reduced by the co-expression of the γ2S subunit in accordance with previous studies on human embryonic kidney cells (Draguhn et al. 1990; Smart et al. 1991). The addition of the γ2 subunit causes the most profound change in the sensitivity of the GABA_A receptor to inhibition by Zn²⁺, although swapping other subunits in the ternary αβγ receptor complex can also be effective. Notably, exchanging α1 for either α2 or α3 subunits in the αxβ3γ2L (x = 1–6) receptor complex resulted in a 5.8- to 7.5-fold shift in the Zn²⁺ IC₅₀ (White & Gurley, 1995). The relative insensitivity to Zn²⁺ of the γ subunit-containing receptors was demonstrated by the inability of Zn²⁺ (up to 100 μm) to completely block the responses to GABA, saturating at less than 50% inhibition. Hetero-oligomeric αxβ2γ2 GABA_A receptors incorporating either the α4 or α6 subunits are also less sensitive to inhibition by Zn²⁺ compared with α1β1 constructs, but more sensitive when compared with α1β1γ2S receptors with IC₅₀ values of 37 (α4β2γ2) and 150 μm (α6β2γ2; Knoflach et al. 1996). A similar shift to a lower Zn²⁺ sensitivity following the exchange of α subunits in the rat GABA_A receptor complex has also been reported for α1β3γ2L (IC₅₀, 245 μm) compared with α6β3γ2L (IC₅₀, 47 μm) constructs (Saxena & Macdonald, 1996). The variability in the IC₅₀ values reported for Zn²⁺ on recombinant GABA_A receptors may be a consequence of changing the receptor subunit composition and species. However, the IC₅₀ values alone cannot be used as the sole identifying criterion for the Zn²⁺ inhibition of GABA_A receptors since it is clear that the apparent inhibitory efficacy of Zn²⁺ also varies, e.g. White & Gurley (1995) reported Zn²⁺ (10 μm) antagonism of responses to GABA that saturated at 10–25% inhibition on human α1β₃γ_2L constructs, despite the estimated IC₅₀ of 1.2 μm from a shallow Zn²⁺ concentration-inhibition plot. This contrasts with the small enhancement of GABA-activated responses (17.6%) by 10 μm Zn²⁺ on rat α1β1γ2L subunits (Saxena & Macdonald, 1996).

The δ subunit confers a novel Zn²⁺-sensitive profile on the GABA_A receptor

In comparison with the influence of the α and γ subunits in determining the degree of Zn²⁺ inhibition, the co-expression of the δ subunit resulted in a different type of antagonism of GABA-activated responses. The Zn²⁺ sensitivity of α1β1δ subunits (IC₅₀ 16.3 μm) was reduced compared with α1β1 (1.2 μm) receptors but considerably more sensitive than α1β1γ2S constructs (> 600 μm). This could be due to a changed quaternary structure of the receptor complex, which may reveal different binding sites for Zn²⁺ on the α1β1δ GABA_A receptor, or alternatively, the affinity of Zn²⁺ for binding site(s) that are common to both α1β1 and α1β1δ constructs, may be reduced in the latter. Interestingly, the type of inhibition caused by Zn²⁺ changed from non-competitive, which is typical of αβ subunit receptors (Smart et al. 1994), to mixed/competitive. Competitive-type behaviour could be interpreted as both Zn²⁺ and GABA competing for the same binding site(s) within the α1β1δ GABA_A receptor complex. This is considered unlikely since in functional competition studies, Zn²⁺ appeared to bind to a novel site on the GABA_A receptor (Celentano, Gyenes, Gibbs & Farb, 1991; Smart, 1992). An alternative explanation would involve Zn²⁺ antagonizing via an allosteric interaction with the GABA_A receptor affecting agonist binding and/or ion channel gating. Interestingly, for bullfrog dorsal root ganglion cells, the GABA-induced response was inhibited by high concentrations of Zn²⁺ (1–3 mm) in a competitive manner analogous to the Zn²⁺ inhibition seen in this study with α1β1δ GABA_A receptors (Yakushiji, Tokutomi, Akaike & Carpenter, 1987). Whether these amphibian ganglia contain δ subunits remains to be seen, but using in situ hybridization it appears that the rodent counterparts probably do not (Persohn, Malherbe & Richards, 1991). GABA-activated currents recorded from rat α6 β3δ constructs expressed in L929 fibroblasts can also be inhibited by Zn²⁺ with an IC₅₀ of 4.8 μm (Saxena & Macdonald, 1996). Responses to GABA were slow to recover from the effects of Zn²⁺, which was not observed in the present study using Zn²⁺ concentrations up to 2.5 mm with α1β1δ subunits.

The addition of the γ2S subunit to the construct α1β1δ, forming α1β1γ2Sδ, appeared to diminish the effectiveness of Zn²⁺, increasing the IC₅₀ compared with that for α1β1 or α1β1δ constructs and causing inhibition to revert to non-competitive. The similarity in the Zn²⁺ IC₅₀ values for the α1β1γ2S and α1β1γ2Sδ receptors could indicate a similar Zn²⁺ binding site(s) within these constructs; however, an alternative interpretation is that co-expression of α1, β1, γ2S and δ subunits only resulted in the formation of discrete α1β1γ2S and α1β1δ subunit receptors. This seems unlikely since as α1β1δ subunit assemblies are functionally expressed and exhibit distinctive properties (Saxena & Macdonald, 1994; Krishek et al. 1996), we would detect their greater sensitivity to Zn²⁺ in a mixed population of α1β1γ2S and α1β1δ receptors. Furthermore, expression of α1, β1, γ2S and δ subunits also results in receptors exhibiting distinctive single channel properties and sensitivity to H⁺ (Saxena & Macdonald, 1994; Krishek et al. 1996). Thus both of the δ subunit-containing receptors studied are likely to form functional receptor assemblies.

Modulation of the Zn²⁺-induced inhibition by H⁺ ions

The prospect of Zn²⁺ and H⁺ sharing binding sites on the GABA_A receptor followed the marked reduction in Zn²⁺ antagonism on lowering the external Ringer solution pH around invertebrate muscle GABA receptors. Subsequent titration indicated that H⁺ probably binds to similar amino acids that co-ordinated the Zn²⁺ binding site (Smart & Constanti, 1982). Given the variety of recombinant GABA_A receptor complexes that can be expressed and their differential sensitivity to H⁺ (Krishek et al. 1996), it was conceivable that H⁺ could interact with Zn²⁺ binding sites on various neuronal GABA_A receptors. The results with α1β1 and α1β1δ receptor subunit combinations indicated that increasing the H⁺ concentration reduced the inhibitory potency of Zn²⁺, whereas increasing the external pH subtly changed the mode of antagonism exerted by Zn²⁺ on α1β1 receptors from non-competitive to uncompetitive. The Zn²⁺ binding site(s) might vary between different receptor complexes, but a parsimonious explanation is that Zn²⁺ binds to similar sites and exerts an allosteric effect, the nature and extent of which is dependent upon receptor subunit composition. In accordance with a similar binding site for Zn²⁺ on GABA_A receptors, the α1β1γ2S construct, which is less sensitive to inhibition by Zn²⁺ (Draguhn et al. 1990; Smart et al. 1991) is also less sensitive to external pH (Krishek et al. 1996).

Interaction of Zn²⁺ and H⁺ ions at native GABA_A receptors

The modulation of Zn²⁺ inhibition by H⁺ on recombinant GABA_A receptors should also occur at selected neuronal GABA_A receptors. Cerebellar granule neurones were chosen to investigate the H⁺-Zn²⁺ interaction since these cells express δ subunit mRNA (Laurie et al. 1992). The sensitivity of Zn²⁺ inhibition to pH on granule neurones was quite similar to the profile observed with α1β1δ constructs and unlike that seen with α1β1 or α1β1γ2S constructs. This might support the existence of αβδ subunit receptors, although granule neurones could express up to ten different receptor subunits (Wisden, Korpi & Bahn, 1996). An interesting difference between GABA_A receptors on granule neurones and α1β1δ constructs, is that inhibition caused by Zn²⁺ on the latter proceeds to 100%, whilst for the former it saturates at 80%. The presence of saturation could be due to GABA_A receptor heterogeneity and may be caused by αiβjγ2 or αiβjγ2δ subunit constructs (where i = 1, 6; j = 1–3), which have less sensitivity to Zn²⁺. The data are in accordance with granule neurones possessing a significant population of αiβjδ subunit receptors. Moreover, under the depolarizing culture conditions caused by supplementing the media with KCl, δ subunit mRNA is predicted to increase 10-fold in the first 4 days in vitro (Gault & Siegel, 1997). It is reasonable to expect some or all of the mRNA to be translated into δ subunit protein that is ultimately assembled into functional receptors. Evidence to support this notion comes from using cerebellar granule neurones maintained in a low K⁺-containing culture medium where GABA-activated currents are antagonized in a non-competitive manner by 30 μm Zn²⁺ (Kilic, Moran & Cherubini, 1993). This action of Zn²⁺ may be due to the non-depolarizing culture conditions (Kilic et al. 1993), where the δ subunit transcripts are expected to be insufficient to produce significant levels of δ subunit protein within the receptor complex (Gault & Siegel, 1997).

In comparison with the cerebellum, using cultured hippocampal neurones, the 5 μm GABA-induced conductance increase was reduced by decreasing the external pH from 8.4 to 7.4 (Pasternack et al. 1996). This profile was reversed in the presence of 50 μm Zn²⁺ such that low pH (7.4) now caused an enhanced GABA response compared with that at pH 8.4. One interpretation of this result is that on exposing hippocampal neurones to alkaline pH, the inhibitory effect of Zn²⁺ is enhanced. Thus acidifying the Krebs solution alleviated the Zn²⁺ block and apparently enhanced the GABA-activated current. This interaction was also studied for neuronal GABA_A receptors in cultured rat sympathetic neurones (Smart, 1992) but external pH had little effect on the inhibitory action of Zn²⁺, possibly due to the expression of different isoforms of GABA_A receptor not included in the present study. Overall, H⁺ and Zn²⁺ can clearly interact at some native GABA_A receptors, but the full interpretation of these effects will only become apparent when the subunit compositions of individual native neuronal GABA_A receptors are known.

Physiological role(s) for Zn²⁺ and H⁺ ions interacting with GABA receptors

The ability of H⁺ and Zn²⁺ to interact at the GABA_A receptor raises the prospect of regulation by endogenous ions. The hippocampus and neocortex are relatively rich in zinc-containing nerve fibres, with other areas, including the cerebellum, exhibiting weaker but detectable levels (Frederickson, 1989). Therefore, changes in external pH near where zinc-containing nerve fibres impinge on inhibitory synapses might be expected to have a regulatory role. In the cerebellum, the likely presence of δ subunit-containing GABA_A receptors will influence the type of regulation by Zn²⁺, dependent upon accompanying α, β and/or γ subunits. For example, α1β1δ subunits are inhibited in a competitive manner and lowering the external pH causes a 5-fold change in sensitivity to Zn²⁺; however, for αβ constructs, the non-competitive inhibition and sensitivity to Zn²⁺ suggest that an equivalent acidification of extracellular pH would cause a profound 50-fold reduction in sensitivity to Zn²⁺. Thus the susceptibility of neuronal GABA_A receptors to Zn²⁺ inhibition under circumstances of lowered pH will be dependent on receptor subunit composition. This type of endogenous regulation of GABA_A receptors may have implications for neuronal excitability in both healthy and diseased neural tissue.