Interaction between copper and zinc at GABA_A receptors in acutely isolated cerebellar Purkinje cells of the rat

Abstract

1. Nanomolar concentrations of Cu²⁺ induce a slowly reversible block of GABA_A receptor-mediated currents which can be removed by chelating substances.

2. The possible interaction of Cu²⁺ with the Zn²⁺ binding site on the GABA_A receptor complex was studied in acutely isolated Purkinje cells using whole-cell recording and a fast drug application system.

3. When Zn²⁺ was applied together with 2 μM GABA, the Zn²⁺-induced block of GABA-mediated currents was not additive to the Cu²⁺-induced block. In the presence of 0.1 μM Cu²⁺ in the bath solution the degree of inhibition of GABA-mediated responses by Zn²⁺ was strongly attenuated.

4. Preapplication of 100 μM Zn²⁺ during 10 s, terminated 1 s before exposure to 2 μM GABA did not affect the GABA current in Cu²⁺-free solution, but relieved its block by 0.1 μM Cu²⁺. This effect of Zn²⁺ was concentration-dependent with an EC₅₀ of 72 μM.

5. When the Cu²⁺-induced block was removed by histidine, preapplication of Zn²⁺ did not increase the GABA current, indicating that the relief of Cu²⁺ block by Zn²⁺ is the result of its ability to actively remove Cu²⁺ from the GABA receptor complex.

6. It is proposed that the inhibitory effects of Zn²⁺ and Cu²⁺ on GABA-induced currents result from an action of these metal ions at distinct, but conformationally linked sites on the GABA_A receptor protein. Under physiological conditions Zn²⁺ would liberate Cu²⁺ from the GABA_A receptor, thus facilitating Cu²⁺ turnover and its binding by other endogenous chelating molecules.

Introduction

Cu²⁺ and Zn²⁺ are essential nutrients with close physiological interactions. High dietary levels of Zn²⁺ depress, Zn²⁺-deficiency increases Cu²⁺ absorption (Cousins, 1985). Both Zn²⁺ and Cu²⁺ are cofactors in the metalloenzyme Cu²⁺, Zn²⁺-superoxide dismutase. Zn²⁺ competes with Cu²⁺ for the binding site, it decreases the ability of Cu²⁺ to transfer electrons in certain conditions. Both metals are present in remarkably high levels in the brain, largely bound to proteins; the exact levels of the free ions in the extracellular space are not known. They occur in presynaptic terminals and are released with synaptic activity (Assaf & Chung, 1984; Frederickson, 1989; Hartter & Barnea, 1988; Kardos et al., 1989). Zn²⁺ ions can modulate a number of ligand- and voltage-gated ion channels (Harrison & Gibbons, 1994; Xie & Smart, 1991). Both Zn²⁺ and Cu²⁺ block GABA_A receptor-mediated currents. The mechanisms of Zn²⁺ and Cu²⁺ modulation of GABA-mediated responses have been analysed separately in a number of preparations (Celentano et al., 1991; Draguhn et al., 1990; Kilic et al., 1993; Ma & Narahashi, 1993; Smart & Constanti, 1990; Smart et al., 1991; Trombley & Shepherd, 1996; White & Gurley, 1995; Yakushiji et al., 1987). Cu²⁺ and Zn²⁺ had a very similar action on GABA-mediated responses when studied in dorsal root ganglion cells. In this preparation Cu²⁺ antagonized the blocking action of Zn²⁺, suggesting that Cu²⁺ and Zn²⁺ may share a common site on the GABA_A receptor.

We have recently found copper blocking GABA-induced currents on Purkinje cells with very high affinity (Sharonova et al., 1998). The block appeared at Cu²⁺ concentrations of about 10 nM, half-maximal inhibition (IC₅₀) at 35 nM. It developed within about 1 min but washout took several minutes. We have now analysed the interaction between these two trace metals and suggest closely related but not identical sites of action at the GABA_A receptor.

Methods

Preparation of Purkinje cells

Neurons were isolated from cerebella of 2–3 week-old male Wistar rats (Versuchstieranstalt Heinrich-Heine-Universität, Düsseldorf). Saggital slices were cut by vibratome perpendicularly to the cerebellar cortex surface and were incubated at room temperature for 1–8 h on a mesh near the bottom of a 150 ml beaker. The incubation solution was continuously bubbled with carbogene (5% CO₂+95% O₂). The solution had a following composition (in mM): NaCl 125; KCl 5; CaCl₂ 1.5; MgCl₂ 1.5; NaH₂PO₄ 1.28; NaHCO₃ 25; glucose 10; phenol red 0.01%. One at a time, slices were transferred to the recording chamber and neurons were dissociated using a vibrating tip of a fused glass pipette (Vorobjev, 1991). The solution for dissociation and recording had the following composition (in mM): NaCl 150; KCl 5; CaCl₂ 2.7; MgCl₂ 2.0; HEPES 10, pH adjusted to 7.4 with NaOH. In most experiments 100 nM Cu₂SO₄ was present in the solution.

Whole-cell recording

Voltage-clamp recording was obtained using the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). Glass recording patch pipettes were prepared from filament-containing borosilicate tubes using a two-stage puller. The electrodes, having resistances of 4–5 MΩ, were filled with recording solution of the following composition (in mM): KCl 140; CaCl₂ 0.5, MgCl₂ 4; HEPES 10; EGTA 5; ATP-Na 3 (pH adjusted to 7.3 with KOH). When high concentrations of GABA (10–50 μM) were used, recordings were performed with an intracellular solution containing low Cl⁻ concentration resulting in a reduced amplitude of agonist induced current, in order to minimize the error due to the voltage drop through the uncompensated series resistance. 125 mM KCl in the intracellular solution were replaced by potassium methanesulphonate. Recordings were carried out at room temperature (22–26°C) using an EPC-9 patch-clamp amplifier. Currents were filtered at 2 kHz, sampled at 1 kHz, and stored on a computer disk. Data were collected with commercially available software (TIDA for Windows). The holding potential was maintained at −70 mV.

Drug application

A fast perfusion technique was used to apply GABA and drugs (Vorobjev et al., 1996). Isolated Purkinje cells were first patch clamped and then lifted into the outflow of the control bath solution. The substances were applied through two different glass capillaries, 0.1 mm in diameter, and, for continuous perfusion, they were added to the bath solution. The delivery ports of the capillaries were positioned within 0.4 mm from the cell under study. For exposure the application system was moved so as to place the cell in the solution stream leaving one of the application capillaries. One capillary was used to apply GABA or GABA together with substances (coapplication). Another capillary was used for perfusion of the cell with substances applied without GABA. The flow through each tube was gravity-driven. For activation of GABA_A channels, in most experiments 2 μM GABA was applied for periods of 1 s, at 30 s intervals. Cu²⁺ (Cu₂SO₄) and Zn²⁺ (ZnCl₂) were dissolved in the perfusate and applied extracellularly. All reagents were obtained from Sigma, Deisenhofen, Germany.

Data analysis

To quantify the inhibitory effects of Zn²⁺ upon the current induced by a constant GABA concentration the following equation was used:

where I_max is the maximal degree of block of the GABA-mediated response achieved by the tested ligand, IC₅₀ is the concentration of the ligand producing a half-maximal block of GABA-mediated responses, and n is the Hill coefficient. The removal of Cu²⁺-induced block of GABA responses by preapplication of Zn²⁺ was fitted by the equation

where I_Cu is the amplitude of GABA responses blocked by Cu²⁺ and I₀ the amplitude of the control GABA response and EC₅₀ is the concentration of Zn²⁺ producing a half-maximal restoration of the GABA response. Other values have the meaning noted above. Data values are presented as mean±s.e.mean.

Results

Blocking actions of Cu²⁺ and Zn²⁺

We studied the effect of Zn²⁺ on GABA-induced current in control and in the presence of Cu²⁺ in the bathing medium. In the absence of copper in the outer solutions, Zn²⁺, applied together with GABA, suppressed the responses to GABA in a concentration-dependent manner (Figure 1A). The Zn²⁺ induced block was fitted by the logistic equation with the following parameters: IC₅₀ 36±5 μM with an I_max of 0.66±0.05 and a Hill slope of 1±0.2 (n=5). The presence of 100 nM Cu²⁺ in the bath solution – which itself reduced GABA induced currents to about 40% of control – markedly changed the concentration-response curve for inhibition of GABA induced currents by Zn²⁺ (Figure 1B). Figure 1C presents the analysis of six experiments carried out in Cu²⁺-free solutions and in the presence of 0.1 μM Cu²⁺ in the bath solution. Inhibition of GABA-mediated responses by Zn²⁺ was strongly attenuated in the presence of Cu²⁺. Moreover, at Zn²⁺ concentrations of 10 and 30 μM the amplitude of GABA-mediated currents was larger than in the presence of Cu²⁺ alone. Therefore, the dose-response curve for Zn²⁺ in the presence of Cu²⁺ was biphasic. Although the complex nature of the dose-response curve makes a quantitative analysis difficult, inspection of Figure 1C shows that the apparent IC₅₀ for inhibition of the GABA mediated response by Zn²⁺ is increased in the presence of Cu²⁺. Besides this, after coapplication of GABA and 100 μM Zn²⁺ in the presence of 0.1 μM Cu²⁺ the subsequent response to GABA was slightly enhanced (Figure 1B).

Figure 1.

Copper and zinc block GABA mediated currents. (A) illustrates the block of currents evoked by sequential application of 2 μM GABA+0, 3, 10, 30, 100 μM Zn²⁺ in the absence of Cu²⁺ in the bath solution. (B) shows results from another cell to which the same set of solutions was applied, however in the presence of 100 nM Cu²⁺ in the bath solution. Applications were made every 30 s and are marked by thin lines (GABA) and hatched bars (Zn²⁺). The response following trace 5 (coapplication of GABA and 100 μM Zn²⁺) displays still a partial removal of Cu²⁺-block by the preceding Zn²⁺-application. (C) In Cu²⁺-free bath solution Zn²⁺ blocks GABA mediated currents in a concentration-dependent manner (open circles). Data points are fitted in equation 1 with I_max=0.66±0.05, IC₅₀=36±5 μM; n=1±02. In the presence of 100 nM Cu²⁺ the concentration-response curve is biphasic (filled circles). Data points are normalized to the response elicited by 2 μM GABA in Cu²⁺-free solution. Each data point averages the responses of six cells.

Relief of Cu²⁺ inhibition by Zn²⁺

The increase of GABA current during coapplication with Zn²⁺ suggests a partial rescue from Cu²⁺ block by Zn²⁺. We therefore examined the effect of Zn²⁺-preapplication on Cu²⁺-induced inhibition of GABA responses. Preapplication of Zn²⁺ for 10 s was terminated 1 s before GABA exposure. Under control conditions (Cu²⁺-free solution) preapplication of Zn²⁺ (10–300 μM) for 10 s terminated 1 s before GABA exposure did not affect the GABA current (data not shown). The one second (1 s) interval between applications of Zn²⁺ and GABA was essential to minimize the direct blocking effect of Zn²⁺ on GABA responses (Figure 2B). When GABA currents were suppressed in the presence of 0.1 μM Cu²⁺ in the bath solution preapplication of 100 μM Zn²⁺ increased the amplitude of GABA-mediated response, in contrast to the inhibitory effect of high Zn²⁺ concentrations applied together with GABA (Figure 3B). In this set of experiments Zn²⁺ containing solutions were Cu²⁺-free.

Figure 2.

Properties of zinc inhibition of GABA-induced currents in Purkinje cells. (A) The inhibitory effect of Zn²⁺ does not depend on GABA concentration. Responses to application of different concentrations of GABA and GABA+100 μM Zn²⁺ obtained from a single cell. (B) Rapid reversal of Zn²⁺ inhibition. Responses to 50 μM GABA in control, after preapplication of 100 μM Zn²⁺ and superposition of two traces. Interval between GABA applications was 90 s, duration of GABA exposure 1 s, zinc application was for 10 s (arrow) terminating at the beginning of trace 2. Recovery from the zinc effect occurs within less than 1 s.

Figure 3.

Brief Zn²⁺ exposure reduces copper induced block of GABA mediated currents. (C) Real-time-scheme of drug exposures: 100 nM Cu²⁺ was present throughout, the times of exposure to Zn²⁺, GABA, GABA+histidine and GABA+Zn²⁺ are indicated. The time for the two second traces illustrated in A and B is also given. The 10 s preapplication of 100 μM Zn²⁺ (at vertical arrows in A and B) which was terminated 1 s before GABA application removed the Cu²⁺ block. Two μM GABA (lines above traces) was applied for 1 s with 30 s intervals. (A) Sequential traces illustrate that the Zn²⁺ preapplication enhances the amplitude of GABA-induced responses. This effect is occluded by 50 μM histidine (open bars) applied together with GABA. (B) Illustrates sequential traces in another cell. Zn²⁺ preapplication enhanced the GABA response. Zn²⁺ applied together with GABA (filled bar above trace) markedly reduced this enhancement (third trace). Without Zn²⁺ preapplication Zn²⁺ did not alter the response significantly (sixth trace).

The copper-induced block of GABA current can be easily and reversibly removed by short exposure to a metal chelator such as histidine (Sharonova et al., 1998). In order to determine whether Zn²⁺ augments the GABA responses by removing Cu²⁺ from the receptor molecule, we compared the effects of Zn²⁺ and histidine on GABA responses suppressed by Cu²⁺. To avoid the formation of histidine Zn²⁺ was preapplied during 10 s as before. Figure 3A demonstrates that preapplication of 100 μM Zn²⁺ (at vertical arrows) increases the GABA current suppressed by Cu²⁺. However, when the Cu²⁺ block was removed by coapplication of GABA with 50 μM histidine, preapplication of Zn²⁺ failed to further augment the response to GABA. Thus the potentiating effect of Zn²⁺ is occluded by histidine-chelation of Cu²⁺. Figure 3B illustrates that a GABA response relieved from copper block by preapplied Zn²⁺ can be blocked by Zn²⁺ when coapplied with GABA. The further traces are sequential GABA responses showing that, without Zn²⁺ preapplication (without removal of copper block), the GABA response is only slightly reduced by 100 μM Zn²⁺. Thus, the effect of Zn²⁺ on the GABA response in the presence of 0.1 μM Cu²⁺ depends on the way of Zn²⁺ administration. Preapplication of Zn²⁺ during 10 s increased subsequent responses to GABA in a dose-dependent way with an EC₅₀ of 72±13 μM and a Hill slope of 1.35±0.3 (five cells, Figure 4).

Figure 4.

Zn²⁺ reduces Cu²⁺-block in a concentration dependent manner. (A) Real-time-scheme of experiments; upper line: Cu²⁺, lower lines Zn 10−300 nM for 10 s and GABA 2 μM for 1 s. data points in the dose response curve are normalized to control GABA responses and represent the mean responses of five cells. The data are fitted by equation 2 with EC₅₀=72±13 μM, n=1.35±0.3.

Zn-induced block at different GABA concentrations

We have previously shown that an increasing GABA concentration decreased the blocking effect of Cu²⁺ (Sharonova et al., 1998). In contrast, the Zn²⁺-induced block of GABA responses could not be surmounted by increasing the GABA concentration (Figure 2A). The block induced by 100 μM Zn²⁺ was about the same at 2 μM GABA (to 0.55±0.05) and at 50 μM GABA (to 0.49±0.07) (n=4 cells). Thus, Zn²⁺ antagonizes GABA responses in cerebellar Purkinje cells in a non-competitive way.

Discussion

Do Cu²⁺ and Zn²⁺ bind to a common site?

The inhibitory action of Zn²⁺ on GABA-activated current is well documented (Celentano et al., 1991; Draguhn et al., 1990; Kilic et al., 1993; Smart et al., 1991; White & Gurley, 1995) as a direct interaction of Zn²⁺ with the GABA_A receptor/ion channel complex. Studies on recombinant GABA_A receptors have demonstrated that the action of Zn²⁺ requires presence of the γ subunit. A Zn²⁺ binding site seems to be present on channels that contain only α1 or β2 or a combination of these subunits. Zn²⁺ can block the current evoked by GABA in these channels with high potency (IC₅₀ about 1 μM). When either the α or the α and β subunits were co-expressed with the γ subunit, GABA_A receptors were rather insensitive to Zn²⁺ (Draguhn et al., 1990; Smart et al., 1991). Purkinje cells seem to contain only α1, β2, β3 and γ2 subunit mRNAs (Laurie et al., 1992; Persohn et al., 1992). In spite of the presence of the γ2 subunit in GABA_A receptors on Purkinje cells, their sensitivity to Zn²⁺ was not as low (IC₅₀ 36 μM) as might have been expected from studies on recombinant GABA_A receptors (Draguhn et al., 1990; White & Gurley, 1995). This may result from the influence of some other factors determining the sensitivity of the GABA_A receptor to Zn²⁺ (Berger et al., 1998).

The mechanism of Zn²⁺ antagonism varies in different cell types. Zn²⁺ noncompetitively inhibits GABA-activated current from neurons of rat superior cervical ganglion (Smart & Constanti, 1990), hippocampus (Legendre & Westbrook, 1991; Mayer & Vyklicky, 1989), cerebellar granule cells (Kilic et al., 1993), from dorsal root ganglion neurons (Ma & Narahashi, 1993) and from recombinant GABA_A receptors composed from α and β subunits (Draguhn et al., 1990; Smart et al., 1991). Zn²⁺ inhibition of some recombinant GABA_A receptors is surmountable by GABA, indicating competitive antagonism (White & Gurley, 1995). In spinal cord neurons the Zn²⁺ block was partially surmountable when the agonist concentration was raised (Celentano et al., 1991). A general model of Zn²⁺ inhibition of GABA_A receptors proposes that Zn²⁺ binds to a single site that allosterically induces two unconducting states; the site affinity is state dependent and controlled by the γ-subunit (Gingrich & Burkat, 1998). The inability to overcome Zn²⁺ inhibition at high GABA concentration in our study is consistent with a noncompetitive mechanism.

While the mechanisms of Zn²⁺ modulation of GABA mediated responses have been widely studied, information regarding the modulation of GABA-induced currents by Cu²⁺ is limited (Yakushiji et al., 1987; Trombley & Shepherd, 1996). Recently, we have described the properties of Cu²⁺-induced inhibition of GABA receptor currents in cerebellar Purkinje cells (Sharonova et al., 1998). The blocks of GABA responses by Zn²⁺ and Cu²⁺ ions in these neurons differ in several respects. Unlike Zn²⁺ inhibition, the effect of Cu²⁺ decreased with increasing GABA concentration, suggesting a competitive mode of interaction. In addition, Cu²⁺ acts with a higher apparent affinity than Zn²⁺ (IC₅₀ 0.035 μM for Cu²⁺ versus 36 μM for Zn²⁺). On the other hand, we have found an obvious interaction between the blocking effects of these metal ions. First, suppression of GABA responses by Zn²⁺ is diminished in the presence of Cu²⁺. Second, preapplication of Zn²⁺ can completely relief the block of GABA_A receptors induced by Cu²⁺. When Cu²⁺-induced block was removed by histidine, preapplication of Zn²⁺ did not further increase the GABA-current, suggesting that the interaction of Zn²⁺ with the GABA_A receptor results in the removal of Cu²⁺ from its binding site during this brief exposure. The decrease of the relative degree of Zn²⁺ inhibition in the presence of Cu²⁺ is explained by a model assuming that these two metal ions compete for the same binding site on the GABA_A receptor. However, the relief of Cu²⁺ block by Zn²⁺ preapplication is not consistent with a passive substitution of Cu²⁺ by Zn²⁺ because, as we have found previously (Sharonova et al., 1998), Cu²⁺, in contrast to Zn²⁺ ions dissociate too slowly (τ⩽3 min) from the GABA_A receptor to allow their unbinding from the protein just for the period of Zn²⁺ application (maximally 10 s); the copper binding sites remain occupied and unavailable for zinc during this time.

On the other hand, Cu²⁺-induced block can easily be removed by exposure to histidine for 10 s (Sharonova et al., 1998). To explain the fast recovery of the GABA response from Cu²⁺ block we have to assume that histidine chelates not only free Cu²⁺, but also Cu²⁺ bound on the protein, and removes it from this site. When we preapplied Zn²⁺, we observed a similar recovery from Cu²⁺ block, which was much faster than passive dissociation of Cu²⁺ from the GABA_A receptor. Zn²⁺ seems to accelerate Cu²⁺ dissociation. The concentration response curve (Figure 4) shows that no removal of block occurs with only 10 μM Zn²⁺. The occupation of the Cu²⁺ binding site can be qualified as quasistationary; no turnover of Cu²⁺ occurs as the medium does not contain Cu²⁺ which could occupy vacant sites. The relief of Cu²⁺ block is only seen with higher Zn²⁺ concentrations. Since Zn²⁺ cannot interact with the binding site occupied by Cu²⁺, it has to bind to another site to facilitate Cu²⁺ dissociation. Thus, our results are consistent with the existence of different, but conformationally linked binding sites for Zn²⁺ and Cu²⁺ ions on the GABA_A receptor molecule. We propose that the binding of Zn²⁺ lowers the affinity for Cu²⁺ and causes subsequent dissociation of Cu²⁺ from its binding site as a result of negative cooperativity between the binding sites for Cu²⁺ and Zn²⁺.

The existence of distinct sites for zinc and copper cations on the GABA_A receptor is supported by a study (Fisher & Macdonald, 1998) demonstrating that the His residue located in the M2–M3 extracellular domain of the α6 subtype (rat α6 H273) plays an important role in determining the sensitivity of recombinant GABA_A receptors to zinc, but not to copper. The α6-subtype confers a relatively low sensitivity to copper, and replacement of the H273 with Asn did not affect inhibition by copper. Studies on rat α1 and α6 chimeras also suggest that the extracellular N-terminal domain of the α1 subunit contributes to a regulatory site(s) for divalent cations, conferring high sensitivity to inhibition by copper and cadmium. These findings are in keeping with our suggestion that inhibitory effects of Cu²⁺ and Zn²⁺ on GABA-induced currents result from interaction of these metal ions with distinct, but allosterically connected binding sites on the GABA_A-receptor.

Functional significance

Release of Cu²⁺ and Zn²⁺ into the extracellular space from cortical and hypothalamic synaptosomes was observed upon depolarization (Hartter & Garnea, 1988; Kardos et al., 1989). According to the estimate made by Kardos et al. (1989), the concentration of Cu²⁺ in the synaptic cleft is in the range of 100–250 μM. Most of the extracellular Cu²⁺ in the brain is not free copper, it is bound to protein. But under reduced amounts of Cu-binding proteins such as caerulopasmin or the PrPc (Brown et al., 1997) increased free copper levels and their pathogenic consequences are to be expected. Synaptically released Zn²⁺ may, depending on the free copper level or the occupancy of proteins with copper, have potentiating or inhibiting effects on gabaergic transmission. In physiological conditions Zn²⁺ would liberate Cu²⁺ from GABA_A receptors, thus facilitating Cu²⁺ turnover and its binding by other endogenous molecules which are more effective in chelating free Cu²⁺ than Cu²⁺ bound to protein. In this way many copper chelating substances may participate in the regulation of GABAergic transmission.

Furthermore, this mechanism may contribute to the beneficial effects of chelating agents and Zn²⁺ when they are used for treating Wilson's disease, an inherited disorder related with copper imbalance. Disturbances in copper metabolism result in its accumulation in the liver, the basal ganglia and cause hepatolenticular degeneration. Zinc is the treatment of choice for maintenance therapy because of its high efficacy and lack of toxicity, it blocks copper absorbtion (Brewer & Yuzbasiyan Gurkan, 1992). Our results raise the possibility that interaction with copper binding to neuronal proteins including the GABA_A-receptor may be partially responsible for the therapeutic effect of zinc in Wilson's disease.