Zn2+ inhibition of recombinant GABAA receptors: an allosteric, state-dependent mechanism determined by the gamma-subunit

Abstract

1. The gamma-subunit in recombinant gamma-aminobutyric acid (GABAA) receptors reduces the sensitivity of GABA-triggered Cl- currents to inhibition by Zn2+ and transforms the apparent mechanism of antagonism from non-competitive to competitive. To investigate underlying receptor function we studied Zn2- effects on macroscopic and single-channel currents of recombinant alpha 1 beta 2 and alpha 1 beta 2 gamma 2 receptors expressed heterologously in HEK-293 cells using the patch-clamp technique and rapid solution changes. 2. Zn2+ present for > 60 s (constant) inhibited peak, GABA (5 microM)-triggered currents of alpha 1 beta 2 receptors in a concentration-dependent manner (inhibition equation parameters: concentration at half-amplitude (IC50) = 0.94 microM; slope related to Hill coefficient, S = 0.7) that was unaffected by GABA concentration. The gamma 2 subunit (alpha 1 beta 2 gamma 2 receptor) reduced Zn2+ sensitivity more than fiftyfold (IC50 = 51 microM, S = 0.86); increased GABA concentration (100 microM) antagonized inhibition by reducing apparent affinity (IC50 = 322 microM, S = 0.79). Zn2+ slowed macroscopic gating of alpha 1 beta 2 receptors by inducing a novel slow exponential component in the activation time course and suppressing a fast component of control desensitization. For alpha 1 beta 2 gamma 2 receptors, Zn2+ accelerated a fast component of apparent desensitization. 3. Zn2+ preincubations lasting up to 10 s markedly increased current depression and activation slowing of alpha 1 beta 2 receptors, but had little effect on currents from alpha 1 beta 2 gamma 2 receptors. 4. Steady-state fluctuation analysis of macroscopic alpha 1 beta 2 gamma 2 currents (n = 5) resulted in control (2 microM GABA) power density spectra that were fitted by a sum of two Lorentzian functions (relaxation times: 37 +/- 5.6 and 1.41 +/- 0.15 ms, means +/- S.E.M.). Zn2+ (200 microM) reduced the total power almost sixfold and accelerated the slow (23 +/- 2.8 ms, P < 0.05) without altering the fast (1.40 +/- 0.16 ms) relaxation time. The ratio (fast/slow) of Lorentzian areas was increased by Zn2+ (control, 3.39 +/- 0.55; Zn2+, 4.9 +/- 0.37, P < 0.05). 5. Zn2+ (500 microM) depression of previously activated current amplitudes (% control) for alpha 1 beta 2 gamma 2 receptors was independent of GABA concentration (5 microM, 13.2 +/- 0.72%; 100 microM, 12.2 +/- 2.9%, P < 0.8, n = 5). Both onset and offset inhibition time courses were biexponential. Onset rates were enhanced by Zn2+ concentration. Inhibition onset was also biexponential for preactivated alpha 1 beta 2 receptors with current depression more than fourfold less sensitive (5 microM GABA, IC50 = 3.8 microM, S = 0.84) relative to that in constant Zn2+. 6. The results lead us to propose a general model of Zn2+ inhibition of GABAA receptors in which Zn2+ binds to a single extracellular site, induces allosteric receptor inhibition involving two non-conducting states, site affinity is state-dependent, and the features of state dependence are determined by the gamma-subunit.

Figure 1. Zn²⁺ inhibition of peak current

Aa and b, macroscopic currents triggered by GABA (5 μM, filled bar) in the presence of a range of Zn²⁺ concentrations (in μM; left of trace) for individual cells expressing α1β2 and α1β2γ2 subunits. Downward current deflection reflects receptor opening. GABA was applied using rapid solution changes. c, inhibition curves were constructed by plotting peak current as percentage of control versus Zn²⁺ concentration (generally n= 6). Smooth curves are least-squares fits of the inhibition equation (see Methods). Fitted parameters are: α1β2 (5 μM GABA), IC₅₀= 0.94 μM, S= 0.70; α1β2γ2 (5 μM GABA), IC₅₀= 51 μM, S= 0.86; α1β2γ2 (100 μM GABA), IC₅₀= 322 μM, S= 0.79. Ba, currents triggered by application (bars) of different GABA concentrations in the absence (CTL, open symbols) and presence (filled symbols) of Zn²⁺ concentrations from individual cells expressing indicated receptors. b, peak current (percentage of control) concentration-response relationships for GABA at a fixed Zn²⁺ concentration (n= 5). Continuous lines were drawn by hand.

Figure 2. Zn²⁺ effects on macroscopic gating

Aa, normalized α1β2 responses (5 μM GABA) replotted from Fig. 1B to illustrate the effects of Zn²⁺ on macroscopic current kinetics. Zn²⁺ response was rightward shifted to permit clear observation of activation time course. Inset, plot of control response and triexponential function fit (smooth line, shifted for clarity). Exponential components are indicated by their respective time constants (fast activation, τ_aF; fast, τ_dF, and slow, τ_dS densensitization). Time constants (τ) are the reciprocal of associated rates (k). Fitted parameters are: A_aF= 1.2 nA, k_aF= 108 s⁻¹, A_dF= -0.81 nA, k_dF= 33.3 s⁻¹, A_dS= -0.33 nA, k_dS= 2.48 s⁻¹, B = -0.42 nA. Ac, GABA concentration dependence of activation and desensitization rates from fitted triexponential functions (generally n= 5 cells). Ab, semilogarithmic plots of currents normalized to peak current. Control (○) activation is described by a single fast exponential (dashed line, time constant, τ_aF). Zn²⁺ (•) adds a second slow exponential component to activation (time constant, τ_aS). The sum of these exponentials (continuous line) describes the activation time course. Ba, normalized α1β2γ2 responses (100 μM GABA) replotted from Fig. 1B to illustrate the effects of Zn²⁺ on macroscopic current kinetics. Bb, GABA concentration dependence of activation and desensitization rates from fitted triexponential functions (generally n= 5 cells).

Figure 3. Effects of Zn²⁺ preincubation on current magnitude and kinetics

Aa, currents triggered by GABA (5 μM, filled bars) from a single cell expressing α1β2 receptors with differing periods of Zn²⁺ preincubation. An initial control response is displayed on the extreme left. Shown next on the same baseline are responses over a range of Zn²⁺ preincubation periods (time difference between Zn²⁺ (10 μM, open bar) and GABA applications). A final bracketing control response is shown on the extreme right. Inset shows four normalized responses aligned in time on an expanded time scale with indicated preincubation periods. Ab, plots of activation parameters derived from multiexponential fitting. Plotted parameters include fast (k_aF) and slow (k_aS) activation rates, and fraction of activation time course that is slow (% Slow). B, same protocol as in Aa applied to a single cell expressing α1β2γ2 receptors (5 μM GABA, 500 μM Zn²⁺). C, fractional inhibition of current magnitude versus Zn²⁺ preincubation in cells expressing α1β2 and α1β2γ2 receptors as indicated (n= 3-7 cells). Thick smooth curve is a biexponential function fit of α1β2 response (parameters: A_F= -0.31, k_F= 6.7 s⁻¹, A_S= -0.17, k_S= 0.72 s⁻¹, and B = 0.79).

Figure 4. Zn²⁺ effects on macroscopic current fluctuations of α1β2γ2 receptor

A, power density spectrum of steady-state macroscopic current activated by 2 μM GABA from a single cell expressing α1β2γ2 subunits. Smooth continuous curve represents the sum of two individual Lorentzian functions (dashed lines) with corner frequencies of 5.8 and 137 Hz (arrows). B, power density spectrum of steady-state macroscopic current activated by 2 μM GABA in the presence of 200 μM Zn²⁺ from the same cell. Corner frequencies are 9.6 and 136 Hz (arrows).

Figure 5. Zn²⁺ inhibition of currents previously activated by GABA

Aa, left panel, current response to Zn²⁺ application (500 μM, 3 s, open bar) from a cell expressing α1β2γ2 receptors preactivated by low [GABA] (5 μM, filled bar). Zn²⁺ concentrations (in μM) are given adjacent to responses. Inhibition is determined at the end of the Zn²⁺ application. Middle panel, same protocol and cell at higher GABA concentration (100 μM) and over a range of Zn²⁺ concentrations. Vertical scaling was adjusted to nearly normalize current magnitude at the start of Zn²⁺ application. Right panel, current in Zn²⁺, as percentage of control, versus Zn²⁺ concentration at indicated GABA concentration (n= 5). The smooth curve is a fit of the inhibition equation (IC₅₀= 49 μM, S= 0.89). Ab, left, time courses of inhibition onset, normalized and replotted from Aa (symbols identify responses). Inset, onset time course normalized to unity and plotted on semilogarithmic axes (500 μM Zn²⁺). Time course is well fitted by a triexponential function (smooth line; parameters: A_onF= 0.38, k_onF= 21 s⁻¹, A_onS= 0.15, k_onS= 6.5 s⁻¹, A_dS= 0.32, k_dS= 1.01 s⁻¹, B = 0.05). The periods dominated by each exponential component are indicated by dashed lines labelled with associated time constants (reciprocal of respective rates: fast (τ_onF) and slow (τ_onS) inhibition onset and slow desensitization (τ_dS)). Right, concentration-response relationships for onset parameters derived from triexponential fitting (fast (k_onF) and slow (k_onS) onset rates, and fraction of onset that is fast [ % Fast = 100A_onF/(A_onF+ A_onS)]). Asterisks mark k_onSs that are significantly different (P < 0.05, paired t test) from low Zn²⁺ (10 μM). Ac, time courses of inhibition recovery or offset replotted from Aa (symbols identify traces), shifted in magnitude to align current magnitude at the end of the Zn²⁺ application (open bar). Smooth lines are triexponential function fits (parameters are: 50 μM Zn²⁺, A_offF= 0.19 nA, k_offF= 15.4 s⁻¹, A_offS= 0.36 nA, k_offS= 2.9 s⁻¹, A_dS= -0.32 nA, k_dS= 0.47 s⁻¹, B = -1.26 nA; 500 μM Zn²⁺, A_offF= 0.60 nA, k_offF= 12.2 s⁻¹, A_offS= 0.63 nA, k_offS= 1.58 s⁻¹, A_dS= -0.93 nA, k_dS= 0.61 s⁻¹, B = -1.20 nA). Exponential components are labelled by associated time constants (reciprocal of respective rates: fast (τ_offF) and slow (τ_offS) inhibition onset and slow desensitization (τ_dS)). Inset, normalized offset time courses replotted on an expanded timescale (interrupted line, 50 μM Zn²⁺; thick line, 500 μM Zn²⁺). Right, concentration-response relationships for offset parameters (fast (k_offF) and slow (k_onS) offset rates, and fraction of offset that is fast (% Fast)) (generally n= 5). B, left, same protocol as in Aa applied to a cell expressing α1β2 receptors. Inset, onset time course normalized to unity and plotted on semilogarithmic axes (20 μM Zn²⁺). As with α1β2γ2 receptors, time course is triexponential where each exponential component is indicated and labelled as in Ab. Right, current magnitude (% of control) versus Zn²⁺ concentration (generally n= 5) and fitted by inhibition equation (smooth curve, IC₅₀= 3.8 μM, S= 0.84).

Scheme 1

Scheme 2

Figure 6. Responses from gating models

Aa, responses from α1β2γ2 gating model (see text) to Zn²⁺ application (open bar, concentrations adjacent to response in μM) of currents preactivated by GABA (filled bar). Inset, current in Zn²⁺, as percentage of control, versus Zn²⁺ concentration and the fit by the inhibition equation (smooth curve, IC₅₀= 94 μM, S= 0.77). Ab, left, normalized time courses of inhibition onset replotted from Aa (symbols identify responses). Middle, onset time course normalized to unity and plotted on semilogarithmic axes (500 μM Zn²⁺). Time course is triexponential with each component indicated by dashed straight lines labelled with associated time constants (reciprocal of respective rates: fast (τ_onF) and slow (τ_onS) inhibition onset and slow desensitization (τ_dS)). Superimposed line is fitted by a triexponential function (parameters: A_onF= 0.47, k_onF= 18 s⁻¹, A_onS= 0.16, k_onS= 2.5 s⁻¹, A_dS= 0.17, k_dS= 0.22 s⁻¹, B = 0.04). Right, concentration- response relationships for fitted parameters from gating model (filled symbols) and empirical responses replotted from Fig. 5Ab (open symbols). Ac, time courses of inhibition offset replotted from Aa, shifted in magnitude to align currents at the end of the Zn²⁺ application (open bar). Time courses are well fitted by a triexponential with each component labelled by the associated time constant (reciprocal of respective rates: fast (τ_offF) and slow (τ_offS) inhibition onset and slow desensitization (τ_dS)). Middle, normalized offset time courses replotted on an expanded time scale. Right, concentration-response relationships for offset parameters from gating model (filled symbols) and empirical responses replotted from Fig. 5Ac (open symbols). Ad, α1β2γ2 model responses triggered at low (5 μM) and high (100 μM) GABA concentrations (bar) in control (□) and in constant Zn²⁺ (300 μM, ▪) with vertical scaling adjusted to normalize for peak control currents. Dominant model gating pathways are shown below responses where rate magnitudes are related to the emboldening and length of transition arrows and ZX* represents Zn²⁺-bound states (ZX, ZX_S and ZX_F) associated with state X (X = R, O, or D). B, α1β2 model responses at 5 μM GABA (bar) in control (□) and in the presence of Zn²⁺ (10 μM, ▪) with dominant model gating pathways shown adjacently. Below, semilogarithmic plot of onset time courses from above normalized to peak current magnitude. In control, activation is described by a single fast exponential (dashed line, time constant, τ_aF). Zn²⁺ adds a second slow exponential component to activation (time constant, τ_aS). The sum of the fitted triexponentials (superimposed line; A_onF= 0.31, k_onF= 19 s⁻¹, A_onS= 0.11, k_onS= 3.4 s⁻¹, A_dS= -0.31, k_dS= 0.54 s⁻¹, B = 0.87) describes the activation time course (see text).

Figure 7. Amino acid sequences

Above, cartoon of the general primary structure of GABA_A receptor subunits showing amino and carboxy termini, and four transmembrane spanning regions (shaded). The extracellular amino domain contains a cysteine loop (C-C). Lines indicate approximate location of amino acid sequences shown below. Region of protein involved in co-ordinating Zn²⁺ binding is putatively extracellular. Below, comparison of amino acid sequences of GABA_A receptor subunits with homologous residues in the neighborhood of a ρ1 histidine (156) on extracellular amino terminus that is critical to Zn²⁺ inhibition of GABA_C receptors (Wang et al. 1995). Amino acids are in three letter code and numbers refer to the codons of the predicted amino acid sequence of the full-length cDNA where the canonical methionine is codon 1. Arrow marks the location of the critical histidine (emboldened) of ρ1. Box encloses adjoining location on GABA_A subunits that contains a histidine (emboldened) residue on the α1-, a neutral leucine on β2-, and a positively charged arginine on the γ2-subunit.