An intracellular redox sensor for reactive oxygen species at the M3-M4 linker of GABA_Aρ1 receptors

Abstract

Background and Purpose

Reactive oxygen species (ROS) are normally involved in cell oxidative stress but also play a role as cellular messengers in redox signalling; for example, modulating the activity of neurotransmitter receptors and ion channels. However, the direct actions of ROS on GABA_A receptors were not previously demonstrated. In the present work, we studied the effects of ROS on GABA_Aρ1 receptor function.

Experimental Approach

GABA_Aρ1 receptors were expressed in oocytes and GABA-evoked responses electrophysiologically recorded in the presence or absence of ROS. Chemical protection of cysteines by selective sulfhydryl reagents and site-directed mutagenesis studies were used to identify protein residues involved in ROS actions.

Key Results

GABA_Aρ1 receptor-mediated responses were significantly enhanced in a concentration-dependent and reversible manner by H₂O₂. Potentiating effects were attenuated by a free radical scavenger, lipoic acid or an inhibitor of the Fenton reaction, deferoxamine. Each ρ1 subunit contains only three cysteine residues, two extracellular at the Cys-loop (C¹⁷⁷ and C¹⁹¹) and one intracellular (C³⁶⁴) at the M3-M4 linker. Mutant GABA_Aρ1 receptors in which C³⁶⁴ was exchanged by alanine were completely insensitive to modulation, implying that this site, rather than a cysteine in the Cys-loop, is essential for ROS modulation.

Conclusion and Implications

Our results show that the function of GABA_Aρ1 receptors is enhanced by ROS and that the intracellular C³⁶⁴ is the sensor for ROS actions.

Introduction

Reactive oxygen species (ROS) such as superoxide (O₂⁻·), hydroxyl radical (OH·) and hydrogen peroxide (H₂O₂), are highly active molecules, inducing oxidative stress in cells (Adam-Vizi, 2005; Rhee, 2006; Halliwell, 2011). ROS are involved in brain processes underlying normal aging and the development of neurodegenerative disorders (Parkinson's and Alzheimer's diseases, schizophrenia, amyotrophic lateral sclerosis and ischaemia-reperfusion injury) (Annunziato et al., 2002; Brennan and Kantorow, 2009; Do et al., 2009). Beside their role in pathological processes, some ROS act as cellular messengers in redox signalling (Bao et al., 2009; Toledano et al., 2010; Rice, 2011). ROS are primarily produced as by-products of the mitochondrial oxidative metabolism and act as local diffusible messengers that regulate neuron-glia and interneuronal communication (Giniatullin et al., 2005; Safiulina et al., 2006; Kishida and Klann, 2007). In the CNS, ROS can be also generated secondary to the activation of neuronal NMDA and AMPA receptors (Kishida and Klann, 2007). It was also suggested that ROS levels modulate neurotransmission and eventually cause changes in neuronal activity and induce synaptic plasticity (Colton et al., 1989; Bernard et al., 1997; Frantseva et al., 1998; Knapp and Klann, 2002; Garcia et al., 2011). Considering the production of ROS and their effects on many ligand-and voltage-gated ion channels (Aizenman et al., 1990; Vega-Saenz De Miera and Rudy, 1992; Li et al., 1998; Annunziato et al., 2002; Dirksen, 2002; Campanucci et al., 2008; Coddou et al., 2009; Rice, 2011), we hypothesized that ionotropic GABA receptors might also be targets for ROS actions. Previous work showed that diverse redox agents modulate the activity of native and cloned ionotropic GABA receptors (Amato et al., 1999; Pan et al., 2000; Calero and Calvo, 2008; Calero et al., 2011; Gasulla et al., 2012), but direct effects of ROS on GABA receptors had not been examined before. A number of studies have indicated that GABAergic neurotransmission is sensitive to ROS (Sah and Schwartz-Bloom, 1999; Sah et al., 2002; Takahashi et al., 2007; Saransaari and Oja, 2008; Tarasenko et al., 2012) and high levels of these agents are normally generated in the retina (Brennan and Kantorow, 2009). However, the molecular targets for ROS actions on the synaptic GABAergic machinery remained so far elusive.

GABA_A receptors are GABA-gated pentameric chloride channels, members of the Cys-loop receptor superfamily, that mediate most of the inhibitory neurotransmission in the CNS (Moss and Smart, 2001; Farrant and Nusser, 2005). GABA_A receptors are made up by combination of diverse functionally distinct subunits (α₁₋₆, β₁₋₃, γ₁₋₃, δ, ε, π, θ, ρ₁₋₃) that commonly form heterooligomeric complexes (e.g. GABA_Aα₁β₂γ₂ receptors) (Farrant and Nusser, 2005; receptor nomenclature follows Alexander et al., 2013). Most of these heteromeric GABA_A receptors are antagonized by bicuculline or picrotoxin and modulated by benzodiazepines and barbiturates (Moss and Smart, 2001). In contrast, GABA_Aρ receptors appear to be exclusively composed of ρ subunits which are widely distributed in the CNS, but are highly expressed only in the retina, densely and selectively concentrated in bipolar cells (Enz et al., 1995; Boué-Grabot et al., 1998; Wegelius et al., 1998). GABA_Aρ receptor-mediated responses are blocked by picrotoxin but are typically insensitive to bicuculline (Zhang et al., 2001). GABA_Aρ1 receptors display both high affinity for GABA and poor desensitization (Farrant and Nusser, 2005), and these distinctive properties allow them to mediate several modes of inhibitory signalling in the retina (Matthews, 1994; Zhang and Slaughter, 1995; Dong and Werblin, 1998; Hartveit, 1999; McCall et al., 2002; Lukasiewicz et al., 2004; Hull et al., 2006; Chávez et al., 2010).

In the present work, GABA_Aρ1 receptor activity was measured before, during and after ROS generation using an in vitro cell model. GABA_Aρ1 receptors were heterologously expressed in Xenopus laevis oocytes and GABA_Aρ1 receptor-mediated Cl⁻ currents electrophysiologically recorded. Our results showed that responses mediated by GABA_Aρ1 receptors were potentiated by ROS. Additionally, experiments involving the chemical modification of sulfhydryl groups and site-directed mutagenesis indicated that Cys³⁶⁴, located at the intracellular M3-M4 linker of the ρ₁ subunits, was essential for modulation by ROS.

Methods

All animal care and experimental procedures were carried out in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the CONICET-University of Buenos Aires Animal Care and Use Committee. A total of 50 frogs were used in the experiments described here.

RNA preparation, oocyte isolation and cell injection

Human cDNA encoding the ρ₁ GABA_A receptor subunit, cloned in the in vitro transcription-suitable vector pGEM, was used as a template to synthesize cRNAs in vitro. Site-directed mutagenesis was achieved by the PCR overlap extension method using QuickChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA). cRNA solutions (0.3–1 ng nL⁻¹) were prepared in Rnase-free H₂O and stored at −70°C. Xenopus laevis (Nasco, Modesto, CA, USA) oocytes at stages V and VI were used for expression of exogenous cRNAs. Isolation and maintenance of cells were carried out as previously described (Miledi and Woodward, 1989). Briefly, frogs were anaesthetized with 3-aminobenzoic-acid ethylester (∼1 mg mL⁻¹) and ovaries surgically removed. Ovaries were incubated with 400 U mL⁻¹ collagenase for 4 h at 23–24°C and isolated oocytes maintained in an incubator at 18°C in Barth's medium (in mM: 88 NaCl; 0.33 Ca(NO₃)₂; 0.41 CaCl₂; 1 KCl; 0.82 MgSO₄; 2.4 NaHCO₃; 10 HEPES and 0.1 mg mL⁻¹ gentamicin; pH adjusted to 7.4 with NaOH). After 1 day, each oocyte was manually microinjected (microinjector Drummond Sci. Co., Broomall, PA, USA) with 50 nL of a solution containing 5–50 ng of cRNA.

Electrophysiological recordings

Two-electrode voltage-clamp recordings were performed 3–7 days after oocyte injection, with an Axoclamp 2B amplifier (Axon Instruments, Union City, CA, USA). Standard glass recording electrodes were made in a Narishige PB-7 puller (Narishige Scientific Instrument Laboratory, Tokyo, Japan) and filled with 3 M KCl. Pipette resistance values were approximately 1 MΩ. The holding potential was set to −70 mV and current traces acquired by a PC through a Labmaster TL-1 DMA interface (Scientific solutions Inc., Solon, OH, USA) using AXOTAPE software (Axon Instruments). Cells were placed in a chamber (volume 100 μL) continuously superfused (12 mL min⁻¹) with frog Ringer's solution (in mM: 115 NaCl; 2 KCl; 1.8 CaCl₂; 5 HEPES; pH 7.0). GABA and other drugs were applied through the perfusion system (Goutman et al., 2005). N-ethyl maleimide (NEM) was freshly prepared prior to each experiment in normal Ringer's. A stock solution of H₂O₂ (1 M) was stored at −20°C and its concentration was confirmed spectrophotometrically at 240 nm. pH was adjusted to 7.0 with NaOH (1 M) or HCl (1 M). All the experiments were carried out at room temperature (23–24°C) and were replicated in at least five different oocytes isolated from at least two different frogs.

Data analysis

Data are expressed as means ± SEM and were analysed with Prism v. 5.0 (Graphpad Software, Inc., San Diego, CA, USA). Concentration-response curves for GABA and concentration-effect curves for H₂O₂ were fit with a logistic equation of the following form: I_GABA/B = [Aⁿ/(Aⁿ + EC₅₀ⁿ)] × 100 where A is the agonist concentration, B the maximal response, EC₅₀ the concentration of agonist that elicits half-maximal responses and n the Hill coefficient. Percentage of potentiation was calculated as [(I_GABAρ1H2O2 × 100/I_{GABAρ1control}) − 100], where I_GABAρ1H2O2 indicates the current amplitude evoked at each particular GABA concentration in the presence of H₂O₂ and I_{GABAρ1control} the corresponding responses in the absence of modulator. Student's t-tests (two-tailed) were employed to evaluate significant differences between parameters.

Materials

The transcription kit mMessage mMachine was purchased from Ambion (Austin, TX, USA), QuickChange Site-Directed Mutagenesis Kit was from Stratagene and type I or type II collagenase from Worthington (Freehold, NJ, USA). The agonist and all the drug and salts, HEPES, 3-aminobenzoic-acid ethylester and Rnase-free H₂O were purchased from Sigma-Aldrich (St Louis, MO, USA).

Results

Functional modulation of GABA_Aρ1 receptors by H₂O₂

Application of GABA to oocytes expressing homomeric GABA_Aρ1 receptors induced large inward Cl⁻ currents displaying all features of the bicuculline resistant component of the retinal GABA receptor-mediated responses (Zhang et al., 2001; Hull et al., 2006). For example, in addition to their bicuculline insensitivity, they are antagonized by 1,2,5,6-tetrahydropyridin-4-yl)methylphosphinic acid (TPMPA) and picrotoxin, non-desensitizing blockers and display the same pharmacological profile for agonists. Figure 1 illustrates representative responses elicited by 0.3 μM GABA in the absence or presence of H₂O₂, recorded at −70 mV. Significant potentiation of the GABA-evoked responses was produced by H₂O₂ (500 μM). In order to characterize H₂O₂ effects, we mainly used two different procedures with equivalent results namely, H₂O₂ applied during the plateau of the GABA responses (Figure 1A), or co-applied with GABA (Figure 1B). Pre-incubation with H₂O₂ (up to 10 min) was also tested and gave the same results. H₂O₂ effects were reversible (Figure 1A) and a second application (not shown) produced similar results. Our previous work showed that several redox modulators of the GABA_Aρ1 receptors, such as ascorbic acid and glutathione, act rapidly and extracellularly (Calero and Calvo, 2008; Calero et al., 2011). In contrast, potentiation of the GABA_Aρ1 receptor activity by H₂O₂ showed a relatively slow onset (Figure 1A). Because H₂O₂ is membrane-permeable (Desagher et al., 1997), this slow time course of action might indicate an intracellular mechanism. No appreciable effects on oocyte properties, such as membrane potential, membrane resistance or current baseline under voltage-clamp, were observed during H₂O₂ applications (500 μM, up to 10 min) (Figure 1C).

Figure 1.

Potentiating effects of H₂O₂ on responses mediated by GABA_Aρ1 receptors expressed in Xenopus laevis oocytes. Representative traces of GABA_Aρ1 receptors mediated Cl⁻ currents elicited by GABA (0.3 μM) applications (indicated as bars) in the absence (control) or presence of H₂O₂ (500 μM). H₂O₂ was either applied during the GABA-evoked responses (A) or co-applied with GABA (B), flanked by control responses to GABA. (C) Lack of effect of H₂O₂ on a representative baseline current recorded from a non-transfected oocyte. For this and the subsequent figures, the oocytes were voltage-clamped at −70 mV. Scale bars indicate current amplitude and time.

Concentration-response curves for GABA were also performed in the absence (control) or presence of H₂O₂ (Figure 2A). H₂O₂ (500 μM) produced a leftward shift in GABA_EC50 without significantly affecting the maximal responses to GABA and nH (EC_{50 GABA} = 0.76 ± 0.03 μM, nH = 2.4 ± 0.3, n = 6; EC_{50 GABA+H2O2} = 0.64 ± 0.03 μM, nH = 2.5 ± 0.4, n = 6; P < 0.005). In order to determine the concentration range for effective H₂O₂ modulation, we tested increasing concentrations of H₂O₂. A concentration-effect curve (Figure 2B) was fitted to a sigmoid equation (see Methods). No saturation was observed at the maximal H₂O₂ concentration tested (2 mM), but higher concentrations of H₂O₂ significantly increased leak currents and were toxic to the oocyte membrane. Similar to effects displayed by other GABA_Aρ1 receptor redox modulators previously studied (Calero and Calvo, 2008; Calero et al., 2011), the degree of potentiation exerted by H₂O₂ on GABA_Aρ1 receptor responses decreased as GABA concentration increased (Figure 2C). For example, the amplitude of currents evoked by 0.3 μM GABA was enhanced by 74 ± 3% (n = 10), whereas potentiation of currents evoked by 30 μM GABA was only 1.3 ± 0.7% (n = 5). For GABA concentrations lower than 3 μM, potentiation induced by H₂O₂ was always significant (P < 0.05). Current-voltage relationships (I-V curves) were carried out, in the presence or absence of H₂O₂. In the presence of H₂O₂ (500 μM), a significant change in slope without alteration in linearity of the I-V relationship or reversal potential (between −120 and +40 mV) was observed (Figure 2D). Therefore, H₂O₂ effects were voltage-independent and not due to variations in intracellular Cl⁻ levels.

Figure 2.

Analysis of H₂O₂ effects on GABA_Aρ1 receptors. (A) C-R curves for GABA in the absence (control) or presence of H₂O₂ (500 μM). Response amplitudes were expressed as fraction of 30 μM GABA-evoked currents (maximal response). (B) Potentiation of GABA_Aρ1 receptors responses (0.3 μM GABA) by increasing concentrations of H₂O₂. (C) GABA concentration dependence of the induced H₂O₂ (500 μM) potentiation of GABA_Aρ1 receptors responses. (D) I-V relationship for GABA_Aρ1 receptors responses evoked by GABA (0.3 μM) in the absence or presence of H₂O₂ (500 μM).

Potentiation of GABA_Aρ1 receptors by H₂O₂ is mediated by an intracellular cysteine residue

Many ionic channels sensitive to redox modulation can be chemically modified through oxidation of cysteine residues. We have previously shown that reducing and oxidizing thiol agents are also effective modulators of the GABA_Aρ1 receptor function (Calero and Calvo, 2008). The reversible effect of H₂O₂ on GABA_Aρ1 receptors is consistent with a direct modulatory action and thiol groups located at the ρ₁ subunits are the most reactive candidates to be oxidized by this agent. In order to elucidate if cysteines are involved in this modulation, we examined the effect of DTT, a membrane permeable reagent which reversibly reduces the thiol groups (Lauriault and O'Brien, 1991). Pre-incubation with DTT (2 mM for 120 s) enhanced the potentiating effects of H₂O₂ on GABA_Aρ1 receptor responses (% P_control = 72.3 ± 5.5%; % P_DTT 2 mM = 129.2 ± 17.5%; n = 6; P < 0.03) (Figure 3A). We further examined if H₂O₂ actions on GABA_Aρ1 receptor-mediated responses were affected by the presence of a reagent that irreversibly modifies cysteine thiol groups. NEM is a membrane-permeable, irreversible alkylating reagent which selectively forms covalent bonds with free sulfhydryl groups, preventing any further chemical reaction at these sites (at pH = 7) (Means and Feeney, 1971). The concentration of NEM was kept as low as possible and incubation periods as short as possible to prevent non-specific effects (Calero et al., 2011). Pre-incubation with NEM (30 μM for 120 s) completely prevented the potentiating effects of H₂O₂ on GABA_Aρ1 receptor responses (% P_control = 68.1 ± 5.1%; % P_{NEM 30 μM} = 2.9 ± 3.2%; n = 8; P < 0.0001) (Figure 3B). These results suggest that H₂O₂ modulates GABA_Aρ1 receptor function by interacting with one or more cysteines.

Figure 3.

Cysteine thiols are involved in H₂O₂ modulation of responses mediated by GABA_Aρ1 receptors. Potentiation of GABA_Aρ1 receptors responses (0.3 μM GABA) by H₂O₂ (500 μM) was (A) enhanced when cysteine thiols were reduced with DTT (2 mM); *P < 0.03, n = 6 and (B) prevented when sulfhydryl residues were alkylated by NEM (30 μM); ***P < 0.0001, n = 8.

Each ρ1 subunit contains only three cysteine residues, two extracellular located in the N-terminal domain that form the characteristic Cys-loop (C¹⁷⁷ and C¹⁹¹) of the GABA_Aρ1 receptors and one intracellular (C³⁶⁴), located at the M3-M4 linker (Zhang et al., 2001). Mutations of the cysteines at the Cys-loop are known to render receptors non-functional (Amin et al., 1994). To determine whether C³⁶⁴ was involved in the modulation of GABA_Aρ1 receptors by H₂O₂, we performed site-directed mutagenesis, replacing this amino acid residue by alanine, a small neutral amino acid insensitive to redox modulation. Mutant GABA_Aρ1C364A receptors expressed in oocytes responded to GABA (EC_{50 GABAAρ1C364A} = 0.52 ± 0.03 μM, n = 7; EC_{50 GABAAρ1wt} = 0.70 ± 0.04 μM, n = 6) and showed a similar pharmacological profile. Interestingly, in sharp contrast with wild-type receptors, the mutant GABA_Aρ1C364A receptors were largely insensitive to H₂O₂ applications and there was no potentiation of 0.3 μM GABA responses by 500 μM H₂O₂ was as follows: % P_{GABAAρ1C364A} = 1.1 ± 0.9 versus % P_GABAAρ1wt = 72.3 ± 9.3; n = 5; P < 0.0001 (Figure 4A). Because H₂O₂ potentiation observed in wild-type receptors could be due to the polar environment of C³⁶⁴,and because GABA_Aρ1C364A receptors showed changes in GABA affinity, compared with wild-type receptors, we performed a more conservative mutation, replacing this cysteine by serine, an amino acid of similar polarity and size. Mutant GABA_Aρ1C364S receptors expressed in oocytes showed typical responses to GABA and EC₅₀ values for GABA were not significantly different from the EC₅₀ values obtained for wild-type receptors (EC_{50 GABAAρ1C364S} = 0.69 ± 0.04 μM, n = 14; EC_{50 GABAAρ1wt} = 0.71 ± 0.04 μM, n = 7; n.s.). However, H₂O₂ (500 μM) failed to potentiate GABA responses elicited from this mutant C364S GABA_Aρ1 receptor. Potentiation of 0.3 μM GABA responses by 500 μM H₂O₂ was as follows: % P_{GABAAρ1C364S} = 2.5 ± 1.9 versus % P_GABAAρ1wt = 72.6 ± 6.2, n = 8; P < 0.0001 (Figure 4B). To further analyse if the insensitivity of GABA_Aρ1C364S receptor to H₂O₂ depended on GABA concentration, we carried out concentration-response curves for GABA either in the absence (control) or the presence of H₂O₂ (Figure 4C). H₂O₂ (500 μM) did not affect GABA EC₅₀ in the mutant receptors (EC_{50 GABAAρ1C364S} = 0.71 ± 0.02 μM, nH = 2.6 ± 0.2; EC_{50 GABAAρ1C364S+H2O2} = 0.67 ± 0.02 μM, nH = 2.7 ± 0.3; n = 5; n.s).

Figure 4.

Intracellular cysteine C³⁶⁴ mediates H₂O₂ modulation of responses mediated by GABA_Aρ1 receptors. Mutation of the C³⁶⁴ affects the potentiating actions of H₂O₂. (A) H₂O₂ (500 μM) failed to potentiate GABA_Aρ_1C364A.; ***P < 0.0001, n = 5 and (B) GABA_Aρ_1C364S receptor responses (0.3 μM GABA); ***P < 0.0001, n = 8. (C) Concentration-response curves for GABA_Aρ_1C364S receptors in the absence (control) or presence of H₂O₂ (500 μM). Response amplitudes were expressed as fraction of 30 μM GABA-evoked currents (maximal response).

In summary, these results indicated that the intracellular cysteine C³⁶⁴, located at the intracellular M3-M4 linker of the ρ₁ subunits, acted as a specific target for the action of ROS and that its chemical modification by ROS potentiated the function of GABA_Aρ1 receptors.

Hydroxyl radicals are involved in the potentiation of GABA_Aρ1 receptor function by H₂O₂

In the presence of low concentrations of Fe²⁺, H₂O₂ generates hydroxyl radicals (OH·) via Fenton reaction (Sah et al., 2002). In order to determine if H₂O₂ can act indirectly through the production of hydroxyl radicals, we studied the effect of H₂O₂ on GABA responses in the presence of lipoic acid (LA; a free radical scavenger), deferoxamine (DFX), an iron chelator that inhibits the Fenton reaction, and iron (II) sulfate, a source of Fe²⁺ to enhance the Fenton reaction. The potentiation of GABA_Aρ1 receptor current responses elicited by 500 μM H₂O₂ during applications of 0.3 μM GABA were decreased in the presence of 200 μM LA (control = 82.1 ± 7.3%, n = 9; lipoic acid = 30.5 ± 5.6%, n = 6; P < 0.001) (Figure 5A) or 100 μM DFX (control = 68.7 ± 10.9%, n = 3; DFX = 32.2 ± 7.4%, n = 6; P < 0.002) (Figure 5B) and increased after pre-incubation with 100 μM FeSO₄ for 2 min (control = 71.3 ± 4.1%, n = 6; FeSO₄ = 104.4 ± 8.1%, n = 6; P < 0.005) (Figure 5C). Potentiation of the GABA_Aρ1 receptor-mediated current was reversible in the presence of iron as it was in the presence of H₂O₂ alone (not shown). None of these agents produced significant changes in the current baseline or modulated GABA_Aρ1 receptor responses, when applied alone. These results suggest that hydroxyl radicals contributed to the potentiation of GABA_Aρ1 receptor function by H₂O₂.

Figure 5.

Hydroxyl radicals are involved in H₂O₂ potentiation of GABA_Aρ1 receptors responses. Potentiation of GABA_Aρ1 receptor responses by H₂O₂ (500 μM) was decreased in the presence of (A) the free radical scavenger LA (200 μM); **P < 0.001, n = 6 or (B) the iron chelator deferoxamine (100 μM); **P < 0.002, n = 6 and was enhanced (C) after pre-incubation with FeSO₄ (100 μM); **P < 0.005, n = 6.

Discussion and conclusions

The present findings are the first to demonstrate the existence of a putative intracellular redox sensor at an ionotropic GABA receptor. We showed here the potentiation of the homomeric GABA_Aρ1 receptor function by ROS and identified the intracellular C³⁶⁴ residue, located at the M3-M4 cytoplasmic linker of the ρ1 subunits, as the target for ROS actions.

GABA_Aρ1 receptors can be considered a simple and suitable model for studying the sensitivity of ionotropic GABA receptors to ROS for several reasons. GABA_Aρ1 receptors are key players in synaptic inhibition in the retina, a tissue that produces high levels of ROS. In addition, homomeric ρ receptors present less structural complexity compared with classic heteromeric GABA_Aαβγ receptors. In fact, ρ1 subunits contain only two molecular sites contributing potential reactive cysteine residues (the extracellular Cys-loop and the single intracellular C³⁶⁴), whereas other GABA_A receptor subunit subtypes contain many intracellular cysteine residues as potential targets for ROS modulation (Sedelnikova et al., 2005; Lo et al., 2008).

Mechanisms underlying the potentiation of GABA_Aρ1 receptors by ROS

The sensitivity of GABA_Aρ1 receptors to ROS was unknown, but earlier work indicated that ROS were capable of modulating GABAergic neurotransmission, presumably via both presynaptic and postsynaptic mechanisms (Colton et al., 1986; Sah et al., 2002; Safiulina et al., 2006; Takahashi et al., 2007; Saransaari and Oja, 2008; Yowtak et al., 2011; Tarasenko et al., 2012). Sah et al. showed that exposure of hippocampal slices to H₂O₂ concomitantly altered GABA_A receptor binding characteristics and increased GABA_A receptor-mediated Cl⁻ influx in CA1 pyramidal cells (Sah and Schwartz-Bloom, 1999; Sah et al., 2002). These results raised the question of how the postsynaptic ROS effects were exerted at the GABAergic synapses. Were they direct, acting through redox-sensitive sites on the GABA receptors or indirect, acting by peroxidation of membrane lipids located near to the Cl⁻ channel (Sah et al., 2002)? We have demonstrated here that ROS were capable of inducing functional changes on GABA_Aρ1 receptors, a GABA_A receptor subtype highly expressed in retinal bipolar cells. These changes were similar to that reported for classic GABA_A receptors from different brain regions (Sah et al., 2002). However, we also provided experimental evidence indicating that GABA_Aρ1 receptors are directly modulated by ROS. Potentiation of GABA_Aρ1 receptors by H₂O₂ was reversible, concentration-dependent, voltage-independent and strongly dependent on the GABA concentration. H₂O₂ effects were partially prevented in the presence of the free radical scavenger LA or by DFX, an inhibitor of the Fenton reaction. In contrast, potentiation of GABA_Aρ1 receptors by H₂O₂ was enhanced if the Fenton reaction was amplified by using iron (II) sulfate. These data suggested that H₂O₂ acted as a precursor for the generation of hydroxyl radicals that eventually exerted their effects on the GABA receptors. H₂O₂-induced potentiation of GABA_Aρ1 receptors persisted in the presence of DFX, thus H₂O₂ might also be acting directly on the receptor to produce these modulatory effects, without involving products of the Fenton reaction.

Due to their reactivity, the cysteine residues in the receptor protein were good candidates for sensing ROS. Chemical protection studies, using the selective membrane-permeable sulfhydryl reagent NEM, and site-directed mutagenesis experiments, where C³⁶⁴ was replaced by alanine or serine, indicated that this particular intracellular residue was essential for ROS effects. The slow onset of potentiation (illustrated in Figure 1) was also consistent with an intracellular mechanism of action. External H₂O₂ concentrations below the range of those normally used in previous studies (Vega-Saenz De Miera and Rudy, 1992; Rice, 2011) had significant effects on GABA_Aρ1 receptors. The effective intracellular concentrations sensed by the GABA_Aρ1 receptors are expected to be lower than bath concentrations, due to the high reducing power of the oocyte cytoplasm. The intracellular antioxidant network that maintains redox balance of amphibian oocytes is composed of many enzymic activities and metabolites, including superoxide dismutase, catalase, ascorbic acid and glutathione (Ferrari et al., 2008). We found that the ROS effects on GABA_Aρ1 receptors were completely washed out in the absence of supplementary reducing agents, probably because the intracellular environment caused potentiation to cease after H₂O₂ applications were stopped. This also suggests that C³⁶⁴ may undergo a reversible chemical modification producing a transient conformational change in the receptor that, in the absence of ROS, rapidly relaxed to a lower energy state. One possible interpretation is that oxidation of the thiol group of C³⁶⁴ by ROS induced protein structural rearrangements that affected GABA binding. The leftward shift produced in concentration-response curves for GABA in the presence of H₂O₂ is compatible with this hypothesis. In addition, because H₂O₂ treatment did not change the reversal potential of the I-V curves is unlikely that, in our experiments, ROS actions were due to a change in the intracellular Cl⁻ levels.

Potential physiological relevance of the modulation of ionotropic GABA receptors by ROS

ROS production in neurons can affect many targets, including several neurotransmitter receptors (Rice, 2011). The modulation of nicotinic cholinergic and purinergic receptors by ROS was exerted through specific intracellular cysteines (Campanucci et al., 2008; Coddou et al., 2009) and such actions could be involved in neuropathological events (Campanucci et al., 2008; 2010). Concerning GABA_A receptors, is quite remarkable that all ρ subunits display a conserved single intracellular cysteine residue at the M3-M4 linker (C³⁶⁴ in ρ₂ and C³⁷⁹ in ρ₃). Moreover, most of the GABA_A receptor subunits contain also one or more cysteine residues at their intracellular loops. Particularly, the M3-M4 intracellular loop is known to interact directly with several cellular regulatory proteins which can be involved in GABA_A receptor oligomerization, assembly, forward trafficking and clustering (Boué-Grabot et al., 2004; Lo et al., 2008). Thus, the importance that these intracellular cysteines might have for GABA_A receptor function during endogenous ROS generation in neurons deserves to be further studied. Given that H₂O₂ is generated normally during cell activity, whereas hydroxyl radicals are typically generated under pathological conditions, it will be important to establish whether redox modulation of GABA receptors is physiological or pathophysiological. It will be also interesting to examine whether oxidation of these intracellular cysteines by ROS might represent a common mechanism for regulating the activity of diverse GABA_A receptor subtypes and other members of the Cys-loop receptor superfamily.

As GABA_Aρ1 receptors provide significant inhibitory drive to the synaptic terminals of retinal bipolar cells, including tonic, reciprocal and lateral inhibition (Zhang and Slaughter, 1995; Lukasiewicz et al., 2004; Hull et al., 2006; Chávez et al., 2010), modulation of the GABA_Aρ1 receptor activity by ROS could eventually shape ganglion cell responses via control of glutamate release at these terminals. Nevertheless, whether or not ROS modulation of ionotropic GABA receptors represents a physiologically relevant mechanism for controlling the activity of retinal neuronal circuits, will need to be assessed by using both retinal slices and in vivo models.

Glossary

DFX

deferoxamine

LA

lipoic acid

NEM

N-ethyl-maleimide

ROS

reactive oxygen species

Conflict of interest

None.