An N-terminal histidine regulates Zn²⁺ inhibition on the murine GABA_A receptor β3 subunit

Abstract

1. Whole-cell currents were recorded from Xenopus laevis oocytes and human embryonic kidney cells expressing GABA_A receptor β3 subunit homomers to search for additional residues affecting Zn²⁺ inhibition. These residues would complement the previously identified histidine (H267), present just within the external portal of the ion channel, which modulates Zn²⁺ inhibition.

2. Zinc inhibited the pentobarbitone-gated current on β3^H267A homomers at pH 7.4, but this effect was abolished at pH 5.4. The Zn²⁺-sensitive spontaneous β3 subunit-mediated conductance was also insensitive to block by Zn²⁺ at pH 5.4.

3. Changing external pH enabled the titration of the Zn²⁺ sensitive binding site or signal transduction domain. The pK_a was estimated at 6.8±0.03 implying the involvement of histidine residues.

4. External histidine residues in the β3 receptor subunit were substituted with alanine, in addition to the background mutation, H267A, to assess their sensitivity to Zn²⁺ inhibition. The Zn²⁺ IC₅₀ was unaffected by either the H119A or H191A mutations.

5. The remaining histidine, H107, the only other candidate likely to participate in Zn²⁺ inhibition, was substituted with various residues. Most mutants were expressed at the cell surface but they disrupted functional expression of β3 homomers. However, H107G was functional and demonstrated a marked reduction in sensitivity to Zn²⁺.

6. GABA_A receptor β3 subunits form functional ion channels that can be inhibited by Zn²⁺. Two histidine residues are largely responsible for this effect, H267 in the pore lining region and H107 residing in the extracellular N-terminal domain.

Introduction

The divalent cation, Zn²⁺, is concentrated in specific neurones in the central nervous system and can be released following neural stimulation into the synaptic cleft where it can bind to numerous membrane proteins (Frederickson, 1989; Frederickson & Bush, 2001). In particular, Zn²⁺ can interact with a variety of ligand-gated and voltage-operated ion channels (Smart et al., 1994; Harrison & Gibbons, 1994). With respect to the γ-aminobutyric acid type A (GABA_A) and C (GABA_C) receptors, Zn²⁺ acts as a non-competitive or mixed inhibitor of GABA-activated Cl⁻ currents (Smart & Constanti, 1990; Dong & Werblin, 1995) and at least for the GABA_A receptor, the potency of inhibition is largely dependent upon the receptor subunit composition (Smart et al., 1994). There has been considerable recent interest in determining likely binding sites for Zn²⁺ on GABA_A receptors with histidines featuring prominently as the main targets on GABA_A receptor β3 (Wooltorton et al., 1997a), β1 (Horenstein & Akabas, 1998), α6 subunits (Fisher & Macdonald, 1998), and GABA_C ρ1 subunits (Wang et al., 1995). Despite these studies, three problems have hampered the search for Zn²⁺ binding sites on GABA_A receptors: these include the acknowledged heterogeneity of native GABA_A receptors and their differential sensitivity to inhibition by Zn²⁺ (Smart et al., 1994; Rabow et al., 1995); the difficulty of determining Zn²⁺ binding sites on recombinant GABA_A receptors composed of multiple subunits, e.g., α, β and γ subunits, which are generally accepted as the likeliest combination of subunits representing most native GABA_A receptors; and the prospect of there being multiple discrete binding sites for Zn²⁺ on a single GABA_A receptor.

In response to the preceding difficulties, two studies adopted a reductionist approach searching for Zn²⁺ binding sites by using homomeric GABA_A receptors comprising β3 (Wooltorton et al., 1997a) or β1 (Horenstein & Akabas, 1998) subunits. The β1 and β3 subunits form functional homomeric receptors that exhibit spontaneous gating, which is inhibited by picrotoxin or Zn²⁺ (Sigel et al., 1989; Krishek et al., 1996; Wooltorton et al., 1997a). They can also be activated or modulated by pentobarbitone, whilst GABA is, for some species, largely ineffective (Sigel et al., 1989; Sanna et al., 1995; Krishek et al., 1996; Wooltorton et al., 1997b; Davies et al., 1997). Site-directed mutagenesis studies of β1 and β3 subunit receptors revealed that a single histidine (H) residue (H267) located in the presumed external portal of the ion channel (second transmembrane domain, TM2) was responsible for the binding and/or inhibitory effect of Zn²⁺ on spontaneous or pentobarbitone-gated β subunit ion channels (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). The location of this ‘site' (if indeed it represents a binding site) was unexpected since it suggested Zn²⁺ penetrated, to a limited extent, into the anion-selective channel of the GABA_A receptor.

Although these experiments revealed a substantial reduction in the potency of Zn²⁺ as an inhibitor at homomeric β subunit GABA_A receptors, application of higher concentrations of Zn²⁺ (exceeding 1–2 mM) caused a complete inhibition of the current gated by β3 homomeric subunits suggesting that additional residues, possibly forming a lower affinity binding site, were still present (Wooltorton et al., 1997a). Similar experiments were also performed on α1βi (where i=1 or 3) heteromers which were not spontaneously-gated and possessed a sensitivity to GABA unlike their homomeric counterparts. The inclusion of the mutant βi^H267A subunits in place of wild-type βi subunits also revealed a considerable increase in the IC₅₀ for Zn²⁺, but again, increasing the Zn²⁺ concentration resulted in substantial inhibition of the GABA-gated current (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). Taken together, the results with homomeric and heteromeric GABA_A receptors indicate that additional amino acid residues are contributing to Zn²⁺ inhibition by participating in signal transduction or by forming an extra binding site(s) for Zn²⁺.

It is conceivable that residues affecting Zn²⁺ inhibition are not uniquely the preserve of the β subunits and may possibly be located on α subunits also, particularly since exchanging α1 for α3 isoforms in heteromeric γ-subunit-containing GABA_A receptors resulted in altered sensitivity to Zn²⁺ (White & Gurley, 1995). For this reason, and because it is clear that an additional residue(s) mediating Zn²⁺ inhibition exists on β subunits, this study utilized site-directed mutagenesis and expression of recombinant β subunit GABA_A receptors in combination with electrophysiology to locate those residues involved.

Methods

Vector construction and site-directed mutagenesis

The mouse GABA_A receptor β3 subunit cDNA was cloned into the vector pRK5. Site-directed mutagenesis was achieved using 27-mer oligonucleotides and a primer-directed polymerase chain reaction method (Quikchange kit, Stratagene). DNAs for transfection were made using the Plasmid Maxi Kit (Qiagen) and all mutant constructs were completely sequenced using the BigDye ready reaction mix (Perkin-Elmer/Applied Biosystems) and an ABI 310 automated DNA sequencer (Applied Biosystems). In order to introduce multiple amino acid residues at position H107 in the GABA_A receptor β3 subunit, we used a variation of this technique. The minimum degeneracy required to encode all amino acids at a given position in a degenerate oligonucleotide is 32-fold, i.e. the codon triplet NNS (where N=G, A, T or C and S=G+C). This triplet was then incorporated, at the equivalent position to H107, into two 27-mer degenerate oligonucleotides (H107X1 5′-AAGTCATTTGTCNNSGGAGTGACAGTG-3′ and H107X2 5′-CACTGTCACTCCSNNGACAAATGACTT-3′). PCR was performed using these primers for 18 cycles at 94°C for 1 min, 55°C for 1 min and 68°C for 15′ using 10 ng of the wild-type GABA_A receptor β3 subunit construct as the template. One potential disadvantage of this method might be that E. coli transformed with the resultant PCR product could harbour two different plasmid types because the sense and antisense strands could differ at the degenerate position. However, DNA sequencing of the resultant mutants demonstrated that very few (2 from 32) templates were mixed. We were therefore able to easily generate multiple amino acid substitutions at the target locus, 13 out of a possible 20 amino acid substitutions (single amino acid code: C, G, I, K, L, P, Q, R, S, T, V, W, Y) were generated in the first 32 templates analysed. This method therefore offers a quick and very versatile approach for producing numerous amino acid substitutions at the same position.

Cell preparation: Oocyte extraction and microinjection

Oocytes were aseptically removed from anaesthetized Xenopus laevis after immersion in 0.5% tricaine and then stored in modified Barth's medium (MBM) containing (mM): NaCl 110, KCl 1, NaHCO₃ 2.4, Tris-HCl 7.5, Ca(NO₃)₂ 0.33, CaCl₂ 0.41, MgSO₄ 0.82, and gentamycin 50 μg ml⁻¹ at pH 7.6. Stage IV and V oocytes were separated and centrifuged (700–1100 g for 6 min at 10°C) for nuclear microinjection with 10 nl of 1 mg ml⁻¹ DNA solution, encoding either the murine β3 wild-type, mutant β3^H267A, and further selected histidine mutant β3 GABA_A receptor subunits incorporating H267A as a background. The injected oocytes were incubated at 18°C for 24 h then subsequently stored at 10°C and fed with fresh MBM every 2–3 days.

HEK cells and transfection

Human embryonic kidney (HEK) cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% foetal calf serum, 2 mM glutamine, 100 u ml⁻¹ penicillin G and 100 mg ml⁻¹ streptomycin at 37°C in 95% air–5% CO₂ (Smart et al., 1991). Exponentially growing cells were subject to electroporation (400 V, infinite resistance, 125 μF; BioRad Gene Electropulser II) with plasmids containing wild-type or mutant GABA_A receptor subunit cDNAs together with a reporter plasmid expressing the S65T mutant of jellyfish green fluorescent protein (GFP; Heim et al., 1994).

Electrophysiology: intracellular recording

Membrane currents were recorded from cDNA-injected Xenopus oocytes using a two-electrode voltage clamp technique. Oocytes were superfused with an amphibian Ringer containing (mM): NaCl 110, KCl 2, HEPES 5 and CaCl₂ 1.8 (pH 7.4) at 8–10 ml min⁻¹ (bath volume, 0.5 ml). Voltage and current microelectrodes were filled with 0.6 M K₂SO₄ (1–2 MΩ). Currents were recorded using an Axoclamp 2B amplifier in conjunction with a Gould 2400S pen recorder.

Patch clamp recording

Whole-cell membrane currents were recorded from single HEK A293 cells using a List EPC7 amplifier. Thin-walled borosilicate patch electrodes (resistance 2–7 MΩ) were filled with electrolyte containing (mM): KCl 140, MgCl₂ 2, CaCl₂ 1, HEPES 10, EGTA 11 and adenosine triphosphate 2, pH 7.2. Cells were continuously superfused with a Krebs solution containing (mM): NaCl 140, KCl 4.7, MgCl₂ 1.2, CaCl₂ 2.5, HEPES 10 and glucose 11, pH 7.4. For some of the low pH experiments, to check the stability of pH 5.4 Krebs, HEPES was replaced with MES. This exchange did not affect the pH sensitivity of Zn²⁺ inhibition. Recordings were performed 24–48 h after transfection by holding cells at −40 mV and filtering membrane currents at 10 kHz (−3dB, 6th pole Bessel, 36 dB per octave). Drugs and Krebs solution were rapidly applied (exchange rate 100 ms) to single cells using a modified U-tube (Wooltorton et al., 1997b). All drugs were constituted in Krebs solution and corrected to pH 5.4 or 7.4. Diethylpyrocarbonate was made fresh immediately prior to use and discarded after 30 min.

Analysis of ligand-modulated membrane conductances and currents

For Xenopus oocytes, membrane conductances were determined by applying hyperpolarizing voltage commands (1 s duration, −10 mV amplitude and 0.2 Hz frequency) from −25 mV holding potential in the absence and presence of a ligand. To construct equilibrium concentration–response relationships for GABA and pentobarbitone, the ligand-induced conductance change (ΔG) was calculated by subtracting the resting conductance from the conductance measured in the presence of each ligand. All the conductances, unless specified, were normalised (ΔG_N) to the maximum conductance change (ΔG_{N, max}) and subsequently fitted with the following equation:

where EC₅₀ represents the concentration of ligand ([A]) inducing 50% of the maximal conductance evoked by a saturating concentration of ligand and n_H is the Hill coefficient. The reductions in the resting membrane conductance by picrotoxinin and Zn²⁺ were used to construct antagonist concentration–inhibition relationships. The antagonist-sensitive conductance (equivalent to the β3 subunit-gated spontaneous membrane conductance; Wooltorton et al., 1997b) was defined as 100% after the addition of a saturating concentration of picrotoxinin (10 μM, Wooltorton et al., 1997a,b). The inhibition of this conductance by intermediate concentrations of antagonists were fitted with the equation:

where ΔG_N^/ and ΔG_N represent the normalized GABA-induced (at a given GABA concentration) or β3 subunit-gated conductance in the presence and absence of antagonist respectively. B represents the antagonist concentration and IC₅₀ defines the concentration of antagonist producing a 50% inhibition of the GABA-induced or β3 subunit-gated conductance. For the whole-cell recording from transfected HEK cells, peak currents induced by pentobarbitone in the absence and presence of Zn²⁺ were measured. Data were assessed for significance using unpaired t-tests or ANOVA with a Tukey post-hoc test. Only HEK cells transfected with β3 subunit cDNAs were sensitive to both Zn²⁺ and picrotoxinin at the concentrations used in this study. Moreover, spontaneously-gated β3 subunit channels were not observed in untransfected cells (Krishek et al., 1996).

Confocal microscopy

The expression patterns of mutant β3^H107X subunits (where X=substituting amino acid) were labelled with a murine bd17 antibody (Boehringer Mannheim, 5 μg ml⁻¹) and resolved with a secondary anti-mouse IgG conjugated with a tetramethylrhodamine analog (TRITC, excitation, 568 nm, emission 580 nm wavelengths). The expression of the receptor subunits was localized using a Leica DMRE fluorescence microscope with both 40X and 63X oil immersion objectives and confocal microscopy was performed with the Leica TCS SP multi-band spectrophotometer with the krypton laser line at 18°C. Images were processed off-line by using Corel Photopaint.

Results

Strategy for site-directed mutagenesis

Additional residues, possibly forming a presumed low affinity Zn²⁺ binding site, were assumed to exist on GABA_A receptors after a histidine residue, H267, located in the second transmembrane domain of the β subunits was substituted for alanine. This residue is critical for Zn²⁺ inhibition of both the spontaneous and pentobarbitone-gated currents that characterize β3 and β1 homomeric GABA_A receptors (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). Substitution of H267 greatly reduced the sensitivity of the receptor to Zn²⁺ (Figure 1A,B), causing a large rightward displacement of the Zn²⁺ inhibition curve, increasing the Zn²⁺ IC₅₀ by over 1000-fold (β3, 0.15±0.03 μM; β3^H267A, 199±13 μM; mean±s.e.mean, n=7 oocytes). Although very effective, this mutation did not abolish inhibition indicating the likely presence of an additional lower affinity Zn²⁺ binding site, or conceivably, Zn²⁺ binding was not completely disrupted around the environment of H267.

Figure 1.

Inhibition of spontaneous Cl⁻ currents through β3 subunit GABA_A receptors by Zn²⁺ and sequence alignments of external histidine residues for GABA_A and GABA_C receptor subunits. (A) Membrane currents recorded from wild-type β3 and mutant β3(H267A) subunits expressed in oocytes held at −40 mV. Membrane conductances in the presence and absence of Zn²⁺ were assessed using hyperpolarizing voltage commands (−10 mV, 1 s, 0.2 Hz). (B) Zinc concentration inhibition relationships for blocking the spontaneous Cl⁻ current transduced by β3 wild-type and β3^H267A mutant GABA_A receptors. The IC₅₀s were 0.15±0.01 μM (β3) and 199±13 μM (β3^H267A). Data points represent the mean±s.e. from seven cells. (C) Schematic diagram of the GABA_A receptor β3 subunit illustrating the transmembrane domains (TM) and external N- and C-termini and large intracellular loop between TM3 and TM4. The external histidine residues and single histidine in TM2 are highlighted. (D) sequence alignments of part of the N-terminal extracellular domain and TM1–TM3 for GABA_A receptor α1, α6, β1 and β3 subunits, and the GABA_C receptor subunit, ρ1. Histidines are depicted as bold characters and those previously demonstrated to affect Zn²⁺ modulation are also italicized.

There are numerous residues that could participate in forming a Zn²⁺ binding site or in contributing to the Zn²⁺ inhibitory signal transduction. These include, other external histidines, glutamates, aspartates and cysteines. The latter were discounted since if the two N-terminal cysteines participate in disulphide bridge formation (Barnard et al., 1987; Amato et al., 1999), then they would be unavailable to coordinate with Zn²⁺. Tentative identification of the residues likely to be involved was obtained by using H⁺ and diethylpyrocarbonate (DEPC) as probes.

Sensitivity of Zn²⁺ inhibition to H⁺ and diethylpyrocarbonate

The Zn²⁺ sensitivity of the mutant β3^H267A subunit homomer was assessed in the presence of low pH and also after exposing the receptor to DEPC. Both experimental protocols would ascertain whether histidine residues were important for the additional Zn²⁺ inhibitory effect since H⁺ will compete with Zn²⁺ for binding to imidazole groups on histidines and DEPC will covalently and irreversibly modify histidines (Miles, 1977; Lundblad & Noyes, 1984) preventing any potential Zn²⁺ binding. Using an external pH 7.4, pentobarbitone (1 μM–3 mM; PB), used since these homomers are insensitive to GABA, increased the membrane conductance for β3^H267A subunit expressing oocytes. Zinc (300 μM) laterally displaced the PB equilibrium concentration–response curve, increasing the EC₅₀ from 46±10 μM to 96±10 μM (P<0.05, n=5), and reduced the maximum response (Figure 2A). Reducing the external pH to 5.4 resulted in a small increase in the potency of PB. However, low pH ablated the residual Zn²⁺ inhibition of the PB-modulated conductance causing the PB concentration-response curves in the absence and presence of Zn²⁺ to be coincident (EC₅₀, PB control, 21±4 μM; +Zn²⁺, 19±4 μM; P>0.05, n=5, Figure 2B).

Figure 2.

Sensitivity of Zn²⁺ regulation of β3^H267A GABA_A receptor subunits expressed in oocytes to external pH. Concentration response curves were constructed for pentobarbitone (PB) modulated currents and normalized to the response evoked by 50 μM PB, in the absence and presence of 300 μM Zn²⁺ recorded from oocytes expressing β3^H267A homomers in Ringer at pH 7.4 (A) and 5.4 (B). Curves were fit to the data as described in the methods (n=3). (C) Zinc concentration inhibition relationship for spontaneous currents mediated by β3^H267A homomers in oocytes exposed to Ringer at pH 7.4 and 5.4. Data were obtained from n=5 cells.

β3 subunit homomers possess the property of spontaneous gating since both Zn²⁺ and picrotoxin can induce reductions in membrane Cl⁻ conductance that is evident only in receptor expressing cells (Wooltorton et al., 1997b). Similar to the PB conductance, the Zn²⁺ sensitivity (10 μM–2 mM) of the spontaneous conductance for β3^H267A subunit homomers was also abolished by reducing the external pH from 7.4 to 5.4 (Figure 2C) suggesting that H⁺ and Zn²⁺ are probably competing for similar amino acid residues. Furthermore, by increasing H⁺ concentration 100-fold and completely preventing inhibition by Zn²⁺, it is likely that such a pH change traversed the pK_a for those amino acid(s) involved in Zn²⁺ inhibition. As histidine is the only amino acid with a pK_a that would be affected by this pH change (pK_a 6.1) it became the primary candidate on β3 homomers.

Further evidence for the involvement of histidine(s) in the inhibitory effect of Zn²⁺ was obtained by using DEPC. Application of 1 mM PB to HEK cells expressing the β3^H267A receptor caused an inward current and associated rebound current characteristic of these homomers (Wooltorton et al., 1997b). Zinc (100 μM) reduced the PB-modulated currents in a reversible manner (Figure 3); however, subsequent exposure of the cell to 1 mM DEPC for 5 min, whilst not affecting the PB-induced response (Figure 3B), resulted in the complete and irreversible loss of the inhibitory action of Zn²⁺ (Figure 3).

Figure 3.

Modification of histidine residues in β3 subunits by diethylpyrocarbonate affects Zn²⁺ inhibition. (A) Membrane currents modulated by PB and Zn²⁺ in HEK cells expressing β3^H267A mutants in the absence and presence of diethylpyrocarbonate (DEPC). Drugs were applied for the duration indicated by the lines. (B) Bargraph of PB-modulated current for β3^H267A mutants following exposure to 1 mM PB; +100 μM Zn²⁺; after recovery; following continuous application of 1 mM DEPC with either PB or PB+Zn²⁺ (n=3).

External histidine residues in the β3^H267A subunit: candidates for coordinating Zn²⁺ inhibition

A review of the histidine residues postulated to be accessible to external Zn²⁺ in the β3 subunit revealed H107, H119 and H191 in the N-terminal extracellular domain in addition to H267 contained in the ion channel pore forming region, TM2 (Figure 1C). Notably, the histidine at position 107 is highly conserved throughout GABA_A, GABA_C, glycine and nicotinic acetylcholine receptor subunits. However, H119 is conserved only within the GABA_A receptor β subunit family and H191 is unique to the GABA_A receptor β3 subunit (Figure 1D). Both H107 and H119 were considered as primary candidates for coordinating or modulating Zn²⁺ inhibition because H191 was not present in the β1 subunit which also exhibited a similar low potency Zn²⁺ inhibition following substitution of H267. The external histidines were systematically substituted for alanines leaving the background mutation H267A intact since this mutation removed the high potency effect of Zn²⁺. The mutations were assessed individually and then, if appropriate, also in multiples.

Expression of β3^{H119A, H267A} in Xenopus oocytes resulted in functional receptors exhibiting both spontaneous and PB-modulated currents. The H119A substitution caused a small reduction in PB potency compared to the H267A mutant but not when compared to the wild-type (Figure 4, Table 1); however, the sensitivity to Zn²⁺ was not further reduced with the Zn²⁺ inhibition curves yielding similar IC₅₀s of approximately 200 μM (Figure 4B, Table 1). Similarly, oocytes injected with β3^{H191A, H267A} cDNA also resulted in functional receptor expression. As for H119A, the PB equilibrium concentration response curves displayed a small reduction in PB potency only when compared to the H267A mutant (Figure 4C, Table 1). The H191A substitution did not, however, reduce the potency of Zn²⁺ (Figure 4D, Table 1). The Hill coefficients for the H119A and H191A β3 subunit receptors were significantly increased compared to the β3 wild-type receptor when the H267A mutation was included, suggesting that within the Zn²⁺ inhibition curve for β3 wild-type subunits, there is more than one component to the inhibition. Thus other residues are important for Zn²⁺ inhibition and that the smaller, lower potency Zn²⁺ inhibitory component can only be clearly observed after substituting H267.

Figure 4.

Effect of mutating histidine residues H119 and H191 in GABA_A receptor β3^H267A homomers on Zn²⁺ inhibition. Concentration response curves for PB were established for β3, β3^H267A and either β3^{H119A, H267A} (A) or β3^{H191A, H267A} (C) GABA_A receptor constructs expressed in oocytes. Data were normalised to the maximum response for each construct and obtained from n=11 cells. Concentration inhibition relationships for Zn²⁺ antagonizing the spontaneous Cl⁻ current was determined for β3, β3^H267A and either β3^{H119A, H267A} (B) or β3^{H191A, H267A} (D) constructs. Data are mean±s.e.mean from n=8 cells.

Table 1.

PB and Zn²⁺ potencies on homomeric β subunit receptors

To ensure that H119 and H191 were not acting in tandem, a combined mutant receptor, β3^{3H119A, H191A, H267A}, was expressed in oocytes. However, despite the reduction in spontaneous gating and a rightward shift in the PB equilibrium concentration response curve compared to the β3^H267A mutant (Table 1), Zn²⁺ still inhibited PB-modulated responses. The maximum responses to PB (1 mM) were inhibited non-competitively by 60±5% (β3^H267A) compared to 58±10% (β3^{H119A, H191A, H267A}) in the presence of 200 μM Zn²⁺ (P>0.05, n=8). Clearly Zn²⁺ was still exerting a considerable inhibitory effect on the mutant β3 subunit despite ablating two of the three external histidine residues. To ascertain the pH sensitivity of the residual Zn²⁺ inhibition on the β3^{H119A, H191A, H267A} receptor, the external pH was varied between 5.4 and 8.4. Zinc inhibition was reduced by lowering the external pH to 5.4 with pH titration yielding a pK_a of 6.8±0.1 (Figure 5). These results implied that H107 is probably important for Zn²⁺ inhibition since this was the last external histidine.

Figure 5.

Zn²⁺ sensitivity of the β3^{H119A,H191A,H267A} GABA_A receptor is affected by external pH. In expressing oocytes, control responses to 100 μM PB were recorded in the presence of 200 μM Zn²⁺ over the external Ringer pH range 5.4 to 8.4 and presented as a pH titration. The reduced inhibitory effect of Zn²⁺ as the external pH increased was determined as a percentage of the control PB response at each pH in the absence of Zn²⁺. The curve was generated according to the inhibition model described in the methods. The pK_a determined from the curve fit was 6.8±0.1 (n=3).

Importance of H107 in β3 subunits for the inhibition by Zn²⁺

The expression of our initial substitution of H107 for alanine, in the background mutation of H267A in β3 subunits (β3^{H107A, H267A}), did not result in the formation of functional β3 subunit receptors (n=8). Similarly, the single mutant β3^H107A subunits also failed to form functional receptors (n=10). To assess whether the expression of the H107A mutant was dependent upon the expression system, human embryonic kidney (HEK) cells were separately transfected with either β3^H107A or β3^{H107A, H267A} subunit cDNAs. On each occasion these cells failed to express functional receptors (n=5) although co-transfection of the reporter DNA encoding for GFP validated the transfection technique and expression competency of the HEK cells.

The highly conserved nature of H107 amongst members of the ligand-gated ion channel superfamily initially suggested an important role for this amino acid in subunit assembly, function, protein folding and/or transport to the cell surface. To enhance the prospect that one or more particular β3^H107 variant subunits may be expressed, a series of mutations were made of the form, β3^H107X, where X was A, R, K, G, I and L. Mutant β3 subunit expression was examined using confocal microscopy to assess their ability to access the cell surface membrane (Figure 6). All of these mutants demonstrated clear cell surface expression; however, transfection of HEK cells resulted in functional ion channels for only β3^H107G and to a lesser extent, with β3^H107I, although expression levels varied with cell morphology. Curiously, the small, rounded, intensely fluorescent (due to GFP co-expression) transfected HEK cells expressed robust PB-modulated currents (>100 pA) for the β3^{H107G, H267A} mutant homomers. In contrast, the flattened slightly fluorescent transfected HEK cells had poor almost unusable expression levels with PB-modulated currents less than 10pA. The β3^H107G mutant was examined for sensitivity to Zn²⁺ with the background mutation H267A. Membrane currents were activated by 1 mM PB and the co-application of 1 mM Zn²⁺ demonstrated only slight inhibition of either the primary inward current or rebound current characteristic of these β subunit homomers (Figure 7A; Wooltorton et al., 1997b). Inhibition curves for Zn²⁺ continuing up to 2 mM caused less than 10% inhibition in the PB-modulated current suggesting this residue is critically important for transducing or coordinating Zn²⁺ inhibition of β3 homomers (Figure 7B). If the background mutation, H267A, was omitted, expression levels of β3^H107G, even in small rounded HEK cells was quite marginal and difficult to use quantitatively (peak currents to 1 mM PB, <20 pA) but 1 mM Zn²⁺ effectively blocked the PB-induced currents (data not shown).

Figure 6.

Confocal microscopy of the cell surface expression of H107X mutants. The panels illustrate HEK cells expressing wild-type β3 homomers (A) and β3 receptor mutants: H107A (B), H107G (C), H107K (D) after exposure to bd17 antisera. Staining for the β3(H107X) mutants revealed their relative location between cell surface membrane and intracellular compartments. Scale bar represents 10 μm.

Figure 7.

Ablation of inhibition by Zn²⁺ on H107G β3 subunit mutants. (A) Membrane currents recorded from β3^{H107G, H267A} expressing HEK cells modulated by 1 mM PB in the absence and presence of 1 mM Zn²⁺. (B) Zn²⁺ inhibition concentration response curves for β3 wild-type, β3^H267A and β3^{H107G, H267A} subunit receptors. All data were fitted according to the methods (n=5).

To analyse whether the H107G mutation prevented Zn²⁺ inhibition of heteromeric GABA-sensitive receptors, the H107G substitution was incorporated into an α1β3^H267A subunit background forming α1β3^{H107G, H267A}. This receptor retained its sensitivity to Zn²⁺ inhibition typical of an α1β3^H267A receptor (IC₅₀s: 21±3 μM, α1β3^H267A; 24±6 μM, α1β3^{H107G, H267A}, n=6), suggesting that for αβ heteromers, H107 was not a critical determinant.

Discussion

The previous examination of H267 as a major determinant of the inhibitory effect of Zn²⁺ on the GABA_A receptor also revealed the likely existence of additional residues that can form another Zn²⁺ sensitive site or are involved in signal transduction. This became apparent from constructing Zn²⁺ inhibition curves for the spontaneous current mediated by wild-type and mutant β3 subunits. Although the H267A mutation caused a clear 1326-fold increase in the Zn²⁺ IC₅₀, it was also apparent that complete inhibition of the conductance gated by β3 subunit receptors was achieved at high Zn²⁺ concentrations. The use of homomeric β3 subunit ion channels simplified the strategy for site-directed mutagenesis since these channels are expected to form near symmetrical arrangements of amino acids unaffected by subunit-subunit interactions with α and γ subunits. Thus single mutations would be expected to have maximum impact on the Zn²⁺ inhibitory effect and/or binding site. To date, whether homomeric β subunits actually exist in vivo remains unclear, but if so, they would constitute a membrane conductance capable of supporting tonic inhibition by virtue of their ability to spontaneously gate (Sigel et al., 1989; Krishek et al., 1996). In addition, their sensitivity to external Zn²⁺ is within the range predicted to exist in the CNS following stimulation, increasing from the suggested basal levels of 1 pM–1 nM to approximately several micromolar (Frederickson & Bush, 2001).

IThe search for the additional residues influencing the residual Zn²⁺ inhibition centred on histidines since lowering the Ringer pH appeared quite effective at abolishing all Zn²⁺-induced inhibition at concentrations up to 1 mM. The pH reduction from 7.4 to 5.4 would encompass the pK_a for histidine with pK_as for other amino acids (assuming no changes are imparted by the local protein microenvironment) lying outside this range. It was assumed that H⁺ and Zn²⁺ are possibly competing for similar binding sites involving histidine residues. At pH 5.4, raising the Zn²⁺ concentration to 2 mM failed to reveal any inhibition of PB-gated currents on β3^H267A constructs. In addition, prevention of Zn²⁺ inhibition by the histidine-modifying reagent, DEPC, provided additional evidence suggesting the involvement of histidine residues.

Evidence for the involvement of H107 in Zn²⁺ inhibition

The search for additional residues affecting Zn²⁺ inhibition on the GABA_A receptor β3 subunit resulted in the identification of histidine residue, H107. The importance of this residue was highlighted by the failure of separate mutation of residues H119 and H191 with H267 to diminish the residual Zn²⁺ inhibition. A number of H107 mutations ablated functional expression of β3 subunit ion channels. A role for H107 in subunit-subunit assembly was thought to be unlikely since H⁺, DEPC and presumably Zn²⁺ can all access this residue in the mature β3 subunit homomer thus affecting receptor function. This degree of accessibility would appear incompatible with H107 being involved in close binding proximity to the interface of another juxtaposed subunit, for example, forming part of an assembly box. Moreover, in the studies of GABA_A receptor assembly boxes to date, H107 has not been identified as a contributory amino acid (Taylor et al., 1999; Klausberger et al., 2000). Taken together with the confocal microscopy data, this suggested that H107 was not important for receptor assembly, protein folding or transport of receptors to the cell surface, but more likely to be interfering with ion channel gating as spontaneous gating, a feature associated with both β1 and β3 homomers, and also PB-gated currents were absent for β3^H107A receptors. This histidine is highly conserved across GABA_A, GABA_C and glycine receptor subunits and is also found in the comparable position in members of the nicotinic acetylcholine receptor family suggesting it plays a vital role that cannot simply be provided by substituting amino acids. Interestingly, more conservative substitutions of H107 to lysine or arginine (β3^H107R, β3^H107K), both residues with positively charged side chains, failed to result in functioning mutant β3 receptors despite cell surface expression.

The use of three different approaches to affect histidines, namely H⁺, DEPC and Zn²⁺ provided compelling evidence for the involvement of a histidine residue in Zn²⁺ inhibition for which H107 is the remaining candidate. Interestingly, the ability of Zn²⁺ to inhibit PB-modulated currents mediated by β3^H107G receptors suggested that in the absence of H267A, substitution of H107 alone has little individual effect on Zn²⁺ sensitivity in accordance with its participation with a presumed low affinity site or inhibitory signal transduction. However, given the clear disparity between the Zn²⁺ IC₅₀s for wild-type β3 and mutant β3^H267A receptors (1326-fold), we might expect a biphasic inhibition concentration curve for Zn²⁺ unless the low affinity Zn²⁺ site only becomes available after mutation of H267A.

It is unclear whether H107 constitutes a discrete binding site for Zn²⁺ on the N-terminus from the previously identified residue H267 in the pore-lining region, or whether both residues are participating in signal transduction. The tertiary structure of the β3 homomeric receptor is unknown and thus it may be possible that H107 and associated parts of the N-terminus are localised near the mouth of the ion channel such that H107 and H267 may participate in the formation of a single Zn²⁺ binding site. Although this cannot be discounted at present, the number of intervening amino acid residues (160) do not conform to known Zn²⁺ binding site consensus sequences such as those contained in metalloenzymes. For the latter, the number of intervening amino acid residues (X) between three coordinating histidines are classified into the short (1–3 residues) and long (18–123 residues) spacers (HX_1–3HX_18–123H) (Vallee & Falchuk, 1993). Furthermore, if H107 and H267 do form part of the same site/transduction mechanism, mutating H107 alone should have produced some effect on Zn²⁺ inhibition which was not observed, albeit by measuring quite small PB-modulated currents in HEK cells.

Potential Zn²⁺ binding domains on GABA_A, GABA_C and glycine receptors

A comparison of previous studies reveals some interesting facets to Zn²⁺-induced modulation of inhibitory amino acid receptors. For GABA_A, GABA_C and glycine receptor subunits, histidine residues have been identified as major elements of the inhibitory effect of Zn²⁺ on GABA- and glycine-activated responses. Homomeric receptors formed from either GABA_A receptor β1 or β3 subunits were dependent upon H267 in the pore-lining region, TM2, for Zn²⁺ inhibition (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). The position of this residue and the prospect of Zn²⁺ penetrating into the anion-selective ion channel suggested that, as for the nicotinic acetylcholine receptor, the ion channel gate and selectivity filter reside deep within the ion channel (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). In both studies, however, a residual effect of Zn²⁺ was evident suggesting alternative binding sites capable of modulating the β subunit homomers.

IAn alternative homomeric receptor formed from GABA_C ρ1 subunits is also sensitive to inhibition by Zn²⁺ (Calvo et al., 1994; Wang et al., 1994, 1995; Chang et al., 1995). Mutation of an N-terminal histidine residue, H156, to tyrosine resulted in abolition of Zn²⁺ inhibition up to 1 mM and only slight (<10%) inhibition at 10 mM, contrasting with the IC₅₀ for Zn²⁺ on wild-type ρ1 subunit receptors of between 10–50 μM (estimated from Figure 4B, Wang et al., 1995). Interestingly, even in this study, mutation of the highly conserved H163 (equivalent to H107 for the GABA_A receptor β3 subunit) to either asparagine or tyrosine failed to produce functional ion channels. Moreover, H156 was also sensitive to the substituting amino acid with asparagine failing to produce functional ion channels.

A different homomeric receptor formed from glycine receptor α1 subunits also displays sensitivity to Zn²⁺ (Bloomenthal et al., 1994; Laube et al., 1995). Zinc has a dual regulatory role causing potentiation of glycine-gated currents at low Zn²⁺ concentrations (<10 μM) and then inhibition at higher concentrations (>10 μM). Recent site-directed mutagenesis experiments have noted that two histidines in the glycine receptor α1 subunit, H107 and H109, forming part of a recognized Zn²⁺ binding motif conserved in selected metalloenzymes (Vallee & Falchuk, 1993), ‘H-phenylalanine (F)-H', are involved in the inhibitory effects of Zn²⁺ (Harvey et al., 1999). Substitution of either or both residues to alanine is sufficient to reduce Zn²⁺ inhibition, an effect that also occurred when the flanking residue, T112, was substituted (Laube et al., 2000), suggesting localized involvement of these three residues in a Zn²⁺ binding site. The potentiating effect of Zn²⁺ is more complex and only H109A mutation appeared to abolish this effect; however, several other point mutations can also abolish Zn²⁺ induced potentiation (Laube et al., 1995; Lynch et al., 1998) suggesting that the H109A mutation may be affecting this aspect of Zn²⁺ action by allosteric mechanisms (Harvey et al., 1999). The relative positions of these histidines in the N-terminal portion of the glycine receptor α1 subunit are of interest. Histidine 107 is unique to the glycine receptor α1 subunit and only 5 residues C-terminal to the relative position of the Zn²⁺-sensitive H156 in the ρ1 subunit. Histidine 109 is the highly conserved histidine found throughout GABA_A, GABA_C and glycine receptor subunits (and is equivalent to H107 in the GABA_A receptor β3 subunit). Curiously, mutation of H109 to alanine was tolerated in glycine receptor α1 subunits allowing the formation of functional ion channels.

Examination of heteromeric GABA_A receptors revealed that for α1β1 and α1β3 constructs, mutation of H267 caused a profound reduction in the Zn²⁺ sensitivity of these receptors (Wooltorton et al., 1997a; Horenstein & Akabas, 1998). Abolition of Zn²⁺-induced inhibition was not evident since alternative sites clearly exist on the β subunits and potentially also on the α1 subunit. Interestingly, H107 in the β3 subunit, whilst clearly vital for Zn²⁺ inhibition on the β3 homomers does not appear to play any such role in the α1β3 heteromeric GABA_A receptor. This suggests that this residue can have markedly different roles depending upon the co-expressed subunits and may indicate that the homomeric and heteromeric forms of the GABA_A receptor have quite different quarternary structures markedly affecting the function, in this case, of H107. Whether this is simply a matter of access to H107 or a more complex change in signal transduction mechanisms between the receptors will require crystallographic structural data.

Previous studies have also identified some differences in Zn²⁺ inhibition of recombinant GABA_A receptors comprising different α subunits. Notably, receptors expressed with α6βγ subunits are more sensitive to inhibition than α1βγ counterparts (Knoflach et al., 1996; Fisher et al., 1997). These GABA_A receptor α subunits differ in the TM2–TM3 region of 12 amino acids at only two locations where histidine and serine residues in α6 replace asparagine and alanine residues in α1 respectively. Substitution of H273 to asparagine in the α6 subunit and co-expression with β3 and γ2L subunits, resulted in expressed GABA_A receptors with similar Zn²⁺ sensitivity to α1β3γ2L receptors. In comparison, conversion of N273 to histidine in the GABA_A receptor α1 subunit increased the Zn²⁺ inhibition of GABA-activated currents recorded from α1^N273Hβ3γ2L receptors to levels similar to that of α6β3γ2L receptors (Fisher & Macdonald, 1998). Thus this external histidine residue in α6 subunits appears to be responsible for the enhanced sensitivity to Zn²⁺ typified by α6 subunit-containing GABA_A receptors.

In conclusion, Zn²⁺ clearly has complex actions on inhibitory amino acid neurotransmitter receptors, but overall, histidine appears to be a favoured amino acid involved in the inhibitory action of this divalent cation. Moreover, in contrast to zinc-containing metalloenzymes (Vallee & Falchuk, 1993), a unique consensus motif does not appear, from the available evidence, to be important for coordinating Zn²⁺ inhibition between these highly homologous neurotransmitter receptors, since residues on the N-terminal, within TM2 and between TM2 and TM3 domains are clearly vital to support receptor regulation by Zn²⁺.

Abbreviations

DEPC

diethylpyrocarbonate

GABA_A

γ-aminobutyric acid type A

GABA_C

γ-aminobutyric acid type C

GFP

green fluorescent protein

ΔG

conductance change

HEK

human embryonic kidney

H

histidine

MBM

modified Barth's medium

n_H

Hill coefficient

PB

pentobarbitone

TM2

second transmembrane domain