Reactive Oxygen Species Potentiate the P2X₂ Receptor Activity through Intracellular Cys⁴³⁰

Abstract

P2X receptor channels (P2XRs) are allosterically modulated by several compounds, mainly acting at the ectodomain of the receptor. Like copper, mercury, a metal that induces oxidative stress in cells, also stimulates the activity of P2X₂R and inhibits the activity of P2X₄R. However, the mercury modulation is not related to the extracellular residues critical for copper modulation. To identify the site(s) for mercury action, we generated two chimeras using the full size P2X₂ subunit, termed P2X_2a, and a splice variant lacking a 69 residue segment in the C terminal, termed P2X_2b, as the donors for intracellular and transmembrane segments and the P2X₄ subunit as the donor for ectodomain segment of chimeras. The potentiating effect of mercury on ATP-induced current was preserved in Xenopus oocytes expressing P2X_4/2a chimera but was absent in oocytes expressing P2X_4/2b chimera. Site-directed mutagenesis experiments revealed that the Cys⁴³⁰ residue mediates effects of mercury on the P2X_2aR activity. Because mercury could act as an oxidative stress inducer, we also tested whether hydrogen peroxide (H₂O₂) and mitochondrial stress inducers myxothiazol and rotenone mimicked mercury effects. These experiments, done in both oocytes and human embryonic kidney HEK293 cells, revealed that these compounds potentiated the ATP-evoked P2X_2aR and P2X_4/2aR currents but not P2X_2bR and P2X_2a–C430A and P2X_2a–C430S mutant currents, whereas antioxidants dithiothreitrol and N-acetylcysteine prevented the H₂O₂ potentiation. Alkylation of Cys⁴³⁰ residue with methylmethane-thiosulfonate also abolished the mercury and H₂O₂ potentiation. Altogether, these results are consistent with the hypothesis that the Cys⁴³⁰ residue is an intracellular P2X_2aR redox sensor.

Introduction

Reactive oxygen species (ROS) are chemically unstable, highly active molecules that include oxygen ions, free radicals, and peroxides. Although historically associated with pathological processes, including the cell damage by oxidative stress, there is considerable evidence indicating that ROS are also cell signaling molecules (Dröge, 2002). Within the past years, it has also become evident that the properties of ionic channels in the CNS and peripheral tissues are modified by ROS. Several voltage- and ligand-gated ionic channels are modulated by ROS, including voltage-gated channels (Chiamvimonvat et al., 1995; Annunziato et al., 2002), ryanodine receptor channels (Aracena-Parks et al., 2006), and purinergic P2X receptor channels (P2XRs) (Mason et al., 2004), strongly suggesting that ROS are acting or may act as modulators of neurotransmission.

Mercury is a heavy metal associated with cell damage; this metal has high affinity for sulfhydryl groups, inactivating enzymes, cysteine residues, and sulfur-containing antioxidants. Excess mercury results in a decrease of the antioxidant defense and a subsequent increase in oxidative stress (Annunziato et al., 2002); so mercury may act as an oxidative stress inducer (Nieminen et al., 1990; Lund et al., 1993). In the brain, mercury also causes behavioral and/or cognitive disturbances (Carpenter, 1994). This could reflect its action as an stress inducer but also as a direct modulator of ion channels, as indicated by its ability to interact with voltage- and ligand-gated ionic channels (Kiss and Osipenko, 1994). Mercury also affects the function of P2XRs, a family of two transmembrane domain ATP-gated channels (Surprenant and North, 2009). It potentiates the ATP-evoked currents in cells expressing P2X₂R (Clyne et al., 2002; Lorca et al., 2005) and attenuates agonist-induced current in P2X₄R-expressing (Acuña-Castillo et al., 2000) and P2X₇R-expressing (our unpublished results) cells. Because modulation of P2XR activity induced by mercury shares the same pattern with copper, it was reasonable to hypothesize that both metals bind to a common allosteric site. However, this is not the case, as documented by Lorca et al. (2005) and Coddou et al. (2005).

Herein, we searched for a mercury binding site at the P2X_2aR and ascertained whether ROS mimicked the action of this metal. Experiments were done with rat P2X₂Rs, P2X₄R, and their chimeras and mutants (see Fig. 1A), expressed in Xenopus oocytes and HEK293 cells. Both forms of P2X₂Rs, the full size receptor, termed P2X_2aR, and the shorter receptor form missing a 69 residue sequence in the C terminus, termed P2X_2bR (Koshimizu et al., 1998), as well as P2X₄R were used. We also used two chimeras that contain the ectodomain of the P2X₄ subunit and the transmembrane and cytoplasmic domains of P2X_2a and P2X_2b subunits (He et al., 2003). This strategy helped us to locate the residue that binds mercury in the C-terminal Val³⁷⁰–Gnl⁴³⁸ sequence of the P2X_2aR. Using site-directed mutagenesis, we identified the Cys⁴³⁰ residue in this sequence as a putative target for the mercury-induced potentiation of the P2X_2aR. The same residue is also crucial for the potentiation induced by hydrogen peroxide (H₂O₂) and oxidative stress inducers, unmasking an intracellular redox sensor in this receptor.

Figure 1.

Parallelism in the actions of mercury and copper on P2X_2aR and P2X₄R activation. A, Schematic representations of P2XRs used in this work. The P2X_4/2a and P2X_4/2b chimeras contain the extracellular Val⁶⁶–Tyr³¹⁵ domain of the P2X₄ subunit (black) and the transmembrane and intracellular domains of the P2X2 subunits (gray); both splice forms of P2X₂ subunits, the full size P2X_2a and the P2X_2b subunit lacking the amino acid sequence Val³⁷⁰–Gln⁴³⁸ in the C terminal, were used. B, C, Representative current traces from single oocytes expressing the homomeric P2X_2aR (B) and P2X₄R (C). ATP was repeatedly applied for 10 s (solid horizontal bars) at regular 10 min intervals in 30 μm (B) and 10 μm (C) concentrations. Each metal was preapplied in 10 μm concentrations for 1 min and coapplied with ATP. Horizontal dashed line represents the current response obtained with ATP alone, which served as control for these protocols. Notice that the effect of mercury on the peak amplitude of current was not completely reversible for the P2X_2aR after 10 and 20 min washouts.

Materials and Methods

Chemicals.

ATP trisodium salt, H₂O₂, dithiothreitrol (DTT), N-acetylcysteine (NAC), myxothiazol, rotenone, and penicillin–streptomycin were purchased from Sigma-Aldrich. Mercury, copper, and zinc chlorides were obtained from Merck. Methylmethane-thiosulfonate (MMTS) was obtained from Toronto Research Chemical. The salts used to prepare the incubation media were purchased from Sigma-Aldrich or Merck. Samples of the triple-distilled water used in buffer preparation were analyzed for electroconductivity; trace metal contamination was assessed by inductively coupled plasma optical emission spectrometry using an Optima 2000 DV ICP-Emission Spectrometer (PerkinElmer Life and Analytical Sciences). Analysis revealed that trace metal contamination was <0.01 μm. ATP, metals, and H₂O₂ solutions were prepared daily before usage.

P2XRs used in experiments.

All experiments were done with rat P2XRs (Fig. 1A). The difference between P2X_2aR and P2X_2bR splice forms is the lack of Val³⁷⁰–Gln⁴³⁸ sequence in the C terminal in the later subunit (Koshimizu et al., 1998). The P2X_4/2a and P2X_4/2b chimeras contain the ectodomain of the P2X₄R (Ile⁶⁶–Tyr³¹⁵) instead of the P2X₂ subunit sequence (He et al., 2003). The P2X_2a–C430A and P2X_2a–C430S mutants were generated by PCR using the proofreading Pfu polymerase (Promega), followed by DpnI digestion of the methylated parental plasmid. A 27-mer couple of primers were designed for the C430A mutant: sense, ggT gTC CAC ggA gCC Agg AgT gCA gTC ggC; antisense, Ctg TCC AgT CTC CTC ggC CTg CCT CCA TCT CTg C. For the C430S mutant, the following primers were designed: sense, gCT gTC CAg TCT CCT Cgg CCT AgC TCC ATC TCT gCT CTg ACT gA; antisense, TCA gTC AgA gCA gAg Atg gAg CTA ggC CgA ggA gAC Tgg ACA gC. To circumvent unwanted mutations, a region surrounding the targeted amino acid and presenting unique restriction sites was subcloned in the parental cDNA and then verified by automated sequencing.

Current measurements in oocytes.

Experiments with Xenopus oocytes were conducted in accordance with the principles and procedures of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the Animal Ethical Committee of the P. Catholic University of Chile. A segment of the ovary was surgically removed under anesthesia from Xenopus laevis frogs; oocytes were manually defolliculated and incubated next with collagenase as detailed by Acuña-Castillo et al. (2000). Oocytes were injected intranuclearly with 3–5 ng of cDNA coding for the different rat receptors used. After a 36–48 h incubation in Barth's solution [in mm: 88 NaCl, 1 KCl, 2.4 NaHCO₃, 10 HEPES, 0.82 MgSO₄, 0.33 Ca(NO₃)₂, and 0.91 CaCl₂, pH 7.5] supplemented with 10 IU/L penicillin/10 mg streptomycin and 2 mm pyruvate, oocytes were clamped at −70 mV using the two-electrode voltage-clamp configuration with an OC-725C clamper (Warner Instruments). ATP-gated currents were recorded after regular 10 s ATP applications repeated every 10 min for up to 10 μm ATP; the percentage of change of the currents under these conditions was 1.2% in oocytes (n = 217), and in HEK293 cells, the percentage of current change was 1.3% (n = 137) under the same conditions. For higher ATP concentrations (>100 μm), the pulses were spaced up to 25 min to avoid receptor desensitization. After metal incubation, the recovery of control currents was always assessed. Non-injected oocytes did not evoke currents when exogenous ATP was applied. ATP, metal chloride salts, and amino acids were dissolved in Barth's media and perfused using a pump operating at a constant flow of 2 ml/min.

Current measurements in HEK293 cells.

HEK293 cells were routinely maintained in DMEM containing 10% (v/v) fetal bovine serum (Biofluids) and 100 μg/ml gentamicin (Invitrogen). Cells were plated on 25 mm poly-l-lysine (0.01% w/v; Sigma-Aldrich)-coated coverslips at a density of 1 × 10⁵ cells per 35 mm dish. The transient transfection was conducted 24 h after plating the cells using 2 μg of DNA and 5 μl of LipofectAMINE 2000 reagent (Invitrogen) in 2 ml of serum-free Opti-MEM. After 4.5 h of incubation, the transfection mixture was replaced with normal culture medium. The experiments were performed 24–48 h after transfection. Electrophysiological experiments were performed on cells at room temperature using whole-cell patch-clamp recording techniques. The currents were recorded using an Axopatch 200B patch-clamp amplifier (Molecular Devices) and were filtered at 2 kHz using a low-pass Bessel filter. Patch electrodes, fabricated from borosilicate glass (type 1B150F-3; World Precision Instruments), using a Flaming Brown horizontal puller (P-87; Sutter Instruments), were heat polished to a final tip resistance of 2–4 MΩ. All current records were captured and stored using the pClamp 9 software packages in conjunction with the Digidata 1322A analog-to-digital converter (Molecular Devices). Patch electrodes were filled with a solution containing the following (in mm): 142 NaCl, 1 MgCl₂, 10 EGTA, and 10 HEPES, pH adjusted to 7.35 with 10 m NaOH. The osmolarity of the internal solutions was 306 mOsm. The bath solution contained the following (in mm): 142 NaCl, 3 KCl, 1 MgCl₂, 2 CaCl₂, 10 glucose, and 10 HEPES, pH adjusted to 7.35 with 10 m NaOH. The osmolarity of this solution was 295–305 mOsm. ATP, H₂O₂, and metal solutions were daily prepared in bath buffer and applied using a fast gravity-driven microperfusion system (BPS-8; ALA Scientific Instruments). Stock solutions of the oxidative stress inductors were prepared in either ethanol or dimethylsulfoxide, and aliquots were stored at −20°C. The current responses were recorded from single cells clamped at −70 mV. Concentration–response data were collected from recordings of a range of ATP concentration applied to single cells with a washout interval of 5 min between each application and normalized to the highest current amplitude.

Data analysis.

Concentration–response curves were performed applying ATP for 10 s in the 1–1000 μm dose range. Curves were normalized against the concentration of ATP that evoked the maximal response. For metal or H₂O₂ modulation experiments, at least five control ATP applications were performed; the average of all the control currents was used as the standard response (100%). The ATP median effective concentration (EC₅₀) was interpolated from each concentration–response curve. Likewise, the maximal ATP current (I_max) was obtained from each ATP concentration–response curve. Each protocol was performed in at least two separate batches of oocytes from different frogs; each experiment was repeated in at least four separate oocytes and HEK293 cells. Curve fitting was performed with GraphPad software. EC₅₀, IC₅₀, and I_max were obtained from each concentration–response curve. Receptor deactivation was expressed as the time between 10 and 90% decrease of the current amplitude after ATP removal, termed 10–90% deactivation time. Statistical analyses included Kruskal–Wallis and Mann–Whitney tests (GraphPad software), because we determined previously that nonparametric statistical tests are better fitted for our analysis as described (Acuña-Castillo et al., 2000).

Results

Metal modulation of P2X_2aR and P2X₄R

Extracellularly applied zinc, copper, and mercury (all in 10 μm concentrations) potentiated the ATP-gated current in oocytes expressing P2X_2aR; representative traces are shown in Figure 1B. The potentiation of response by mercury was still present 20 min after the washout of metal (35 ± 12%) (Fig. 1B, last trace, n = 6), whereas a 10-min washout period was sufficient to restore normal response in oocytes treated with copper and zinc. In contrast to the P2X_2aR, mercury and copper inhibited ATP-induced P2X₄R currents; again, the effect of mercury was slower to reverse compared with copper or zinc, although it attained control currents within 20–30 min of washout (Fig. 1C, n = 5). Like in the P2X_2aR-expressing oocytes, zinc potentiated the responses evoked by ATP (Fig. 1C). In our experimental conditions, the ATP potency was threefold to fourfold higher for the P2X₄R than the P2X_2aR and P2X_2bR (Table 1) (supplemental Fig. 1, available at www.jneurosci.org as supplemental material). We also observed that mercury markedly influenced the rates of P2X_2aR deactivation, increasing the 10–90% deactivation time from 4.4 ± 0.2 to 17.2 ± 2.8 s (n = 5; p < 0.05) (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). Such effect on the channel deactivation was also observed in the presence of copper, although the magnitude of the increase was only twofold (supplemental Fig. 2, available at www.jneurosci.org as supplemental material).

Table 1.

Characterization of P2XRs expressed in oocytes and HEK293 cells

Receptor	EC50 (μm)	nH	Imax	n
Xenopus laevis oocytes
P2X2a	33.3 ± 2.5	1.4 ± 0.1	12.1 ± 2.9	30
P2X2b	59.2 ± 9.1	1.4 ± 0.3	15.6 ± 1.9	6
P2X2a–C430A	293 ± 70*	1.2 ± 0.3	1.2 ± 0.5*	8
P2X2a–C430S	45.0 ± 6.3	1.4 ± 0.3	11.8 ± 0.8	4
P2X4/2a	4.8 ± 0.6	0.9 ± 0.2	6.7 ± 1.5	9
P2X4/2b	4.6 ± 2.4	1.2 ± 0.1	14.0 ± 1.5	15
P2X4	11.0 ± 2.6	1.4 ± 0.2	4.7 ± 0.6	12
HEK293 cells
P2X2a	10.2 ± 1.3	1.2 ± 0.2	4.6 ± 0.3	8
P2X2b	7.6 ± 1.3	1.1 ± 0.3	4.2 ± 0.9	4
P2X2a–C430S	18.1 ± 3.3	1.5 ± 0.6	5.4 ± 0.9	6

The EC₅₀ and n_H values were derived from the Hill equation. The maximal current (I_max) values (in microamperes for Xenopus and nanoamperes for HEK cells) were obtained applying a saturating concentration of ATP. *p < 0.01, Kruskal–Wallis test compared with the P2X_2aR.

Potentiating effects of mercury in P2X_4/2a but not P2X_4/2b chimera-expressing oocytes

Because mercury has opposite effects on P2X₂R and P2X₄R activities, in additional studies, we used P2X_4/2 chimeras (Fig. 1A). The potentiating effect of mercury was preserved in P2X_4/2a-expressing cells, as indicated by the ability of 10 μm mercury to enhance the 1 μm ATP-induced current by 3.7 ± 0.8-fold (Fig. 2A). This potentiation reflected a leftward shift in the sensitivity of chimeras for ATP, reducing the EC₅₀ from 4.8 ± 0.6 to 2.4 ± 0.6 μm (n = 5–9; p < 0.05) (Fig. 2A), without significantly altering the maximal ATP response. Zinc (10 μm) also potentiated the 1 μm ATP-evoked currents (2.6 ± 0.3-fold) (Fig. 2A), whereas 10 μm copper inhibited the ATP-evoked currents by 50% (p < 0.05) (Fig. 2A). In addition, copper did not significantly modify the ATP EC₅₀ (3.3 ± 1.6 μm; n = 5), but it diminished the maximal ATP response by 40% (data not shown). Mercury also modified the receptor deactivation, as can be deduced from representative tracings shown in Figure 2A; the 10–90% deactivation time augmented ninefold in the presence of mercury, from 12.6 ± 1.1 to 104.7 ± 11.5 s (n = 5; p < 0.01) (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). Copper and zinc did not modify significantly the deactivation constant for this chimera (supplemental Fig. 2, available at www.jneurosci.org as supplemental material).

Figure 2.

Effects of metal cations on P2X_4/2a (A) and P2X_4/2b (B) chimeras expressed in oocytes. Left column, The P2X_4/2a chimera is mercury sensitive (top), whereas the P2X_4/2b chimera is mercury resistant (bottom). Representative current traces from controls (black) and oocytes exposed to mercury 1 min before and during ATP application (gray). Middle column, Summary of effects of mercury, copper, and zinc, on the 1 μm ATP challenge; C refers to the control current in the absence of metals. Data shown are mean ± SEM values from four to six recordings per treatment; *p < 0.05, Kruskal–Wallis test. Right column, Concentration–response curves for ATP alone (filled circles) and ATP plus 10 μm mercury (open squares). The metal displaced leftward the nucleotide concentration–response curve for the P2X_4/2a but not P2X_4/2b chimera, reducing the ATP EC₅₀ value from 4.8 ± 0.6 to 2.4 ± 0.6 μm (p < 0.05, Mann–Whitney test; n = 5–9 per dose). Mercury was preapplied for 1 min and then coapplied with ATP.

The P2X_4/2b chimera was resistant to mercury (Fig. 2B), and 10 μm of the metal did not displace the ATP concentration–response curve (Fig. 2B) nor modified receptor deactivation. In contrast, 10 μm zinc potentiated the receptor activity by 2.6 ± 0.3-fold, whereas 10 μm copper significantly inhibited ATP-evoked currents by 55 ± 12% (Fig. 2B). As with the P2X_4/2a chimera, the copper and zinc pattern was consistent with the extracellular domain of the chimera corresponding to the P2X₄R phenotype. The ATP EC₅₀ was 4.6 ± 2.4 μm (Table 1). These results indicate that the mercury-sensitive residue is located in the Val³⁷⁰–Gln⁴³⁸ sequence of P2X₂R; this sequence contains a single cysteine at position 430.

Role of the Cys⁴³⁰ residue in the mercury-induced P2X_2a receptor potentiation

To ascertain the role of this residue in the P2X_2aR sensitivity to mercury, we generated the P2X_2a–C430A and P2X_2a–C430S mutants of this receptor. The ATP potency for the P2X_2a–C430A mutant was reduced 10-fold compared with the wild type, and the receptor desensitization was markedly increased (Fig. 3A, Table 1). In contrast, the ATP EC₅₀ for the P2X_2a–C430S mutant and the rate of receptor desensitization were comparable with the wild-type receptor (supplemental Fig. 1, available at www.jneurosci.org as supplemental material) (Table 1) (n = 4). In both mutants, the potentiating effect of 10 μm mercury was significantly reduced but not completely abolished. In oocytes expressing the P2X_2a–C430A mutant, 10 μm mercury potentiated the ATP-evoked currents only 3.3 ± 0.7-fold compared with 12.3 ± 2.6-fold in the wild-type P2X_2aR (p < 0.05; n = 5–6) (Fig. 3A,B), whereas zinc and copper potentiation was not significantly altered (Fig. 3B). The effect of mercury in this mutant was totally reversible in contrast to the wild-type receptor (compare tracings of Figs. 3A, 1A). In cells expressing the P2X_2a–C430S mutant, 10 μm mercury potentiated by 2.2 ± 0.9-fold the ATP-evoked currents (n = 4). As with the P2X_2a–C430A mutant, copper and zinc potentiation was not changed in this mutant (Fig. 3B). Mercury also slowed the rate of P2X_2aR and P2X_4/2a chimera deactivation, whereas the rates of P2X_2bR and P2X_4/2b chimera were not significantly affected (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). Such effect of mercury probably reflects increased sensitivity of receptors to ATP, which in turn delays dissociation of ATP from its binding during washout periods.

Figure 3.

The mercury potentiation of ATP-induced current is significantly reduced in the C430 and C430S P2X₂R mutants expressed in oocytes. A, Representative current responses to 10 s ATP applications (horizontal bars) using for each receptor an ATP approximate EC₁₅ concentration in controls (C), mercury-treated cells, and washout cells (W). Mercury was applied in 10 μm concentration 1 min before and during ATP application. B, Comparison of the effects of metals on the peak amplitude of the ATP-evoked currents. Data shown are mean ± SEM values; *p < 0.05 compared with the P2X_2aR, Kruskal–Wallis test; n = 4–6. WT, Wild type.

Interaction of mercury with ectodomain P2X_2aR residues

In additional work, we analyzed the Cys⁴³⁰ residue-independent effects of mercury on potentiation of the ATP-evoked currents. Experiments with P2X_4/2a–C430A and P2X_4/2b chimeras revealed their insensitivity to mercury, in contrast to P2X_2b, P2X_2a–C430A, P2X_2a–C430S, and P2X_4/2a receptors that showed a residual mercury-dependent potentiation. These results, shown in supplemental Figure 3 (available at www.jneurosci.org as supplemental material), indicate that the extracellular domain of the P2X₂R contains additional residues that detect mercury directly or indirectly.

Effects of hydrogen peroxide on P2X₂R activity

The ATP-evoked P2X_2aR currents were potentiated by H₂O₂ in a concentration-dependent manner (Fig. 4A,B); 0.3 mm H₂O₂ increased by 2.8 ± 0.3-fold the ATP-gated currents (Fig. 4B). The P2X_4/2a chimera was also potentiated by 1 mm H₂O₂, reaching a significant 1.9 ± 0.2-fold increase of the currents (n = 4). After 10 min washout, the currents remained increased 0.3 ± 0.1-fold. In contrast, P2X_2bR, P2X_4/2b chimera, and P2X_2a–C430A and P2X_2a–C430S mutants were totally resistant to H₂O₂ (Fig. 4A,B). These results indicate that H₂O₂ mimics the Cys⁴³⁰ residue-dependent effects of mercury but not the minor extracellular-dependent action of this metal.

Figure 4.

H₂O₂ potentiates the ATP-gated currents in the P2X_2aR expressed in oocytes in a Cys⁴³⁰-dependent manner. A, Representative current responses of the P2X_2aR (top) and the C430A mutant (bottom). H₂O₂ (1 mm) was applied for 10 min before the ATP pulse. C, Control; W, washout; WT, wild type. B, Concentration-dependent effects of H₂O₂ on ATP-induced current in oocytes expressing wild-type and mutant P2X₂Rs. C, Effects of chemical modification of cysteine residues in the P2X_2aR by the membrane-permeable MMTS on H₂O₂ or mercury-dependent potentiation of ATP-induced currents. MMTS was applied for 3 min before ATP application. D, Summary of the effects elicited by H₂O₂, DTT, and the joint application of both agents on P2X_2aR (top) and P2X₄R (bottom) currents. In B–D, ATP was applied in approximate EC₁₅ concentration for 10 s, and data shown are mean ± SEM values. *p < 0.05, Kruskal–Wallis test; n = 4–6 per dose/treatment.

To confirm the role of cysteines in the H₂O₂-induced potentiation of P2X_2aR current, we chemically modified receptor cysteines using a plasma membrane-permeable MMTS. This reagent has an extremely rapid reactivity with thiol groups to form alkylthiosulfonated cysteines under normal physiological conditions (Valentine and Paglia, 1981). Oocyte treatment with 1 mm MMTS abolished the H₂O₂ potentiation in the wild-type P2X_2aR challenged with 10 μm ATP (Fig. 4C). MMTS treatment also abolished the mercury-induced potentiation (Fig. 4C). MMTS alone elicited a transient and reversible 20% current increase, but this effect reversed after the first washout (data not shown).

Conversely, the application of 1 mm DTT did not modify the ATP-evoked currents (Fig. 4D). The coapplication of H₂O₂ and DTT abolished the effect of H₂O₂ (Fig. 4D). Contrary to the P2X_2aR, 1 mm H₂O₂ inhibited by 43.8 ± 0.1% (n = 4) the P2X₄R activity (Fig. 4D). DTT application did not modify the receptor activity, but, when this agent was coapplied with H₂O₂, the inhibition was completely obliterated (Fig. 4D).

Effects of H₂O₂ and divalent metals on wild-type and mutant P2X₂Rs expressed in HEK293 cells

To exclude the cell-specific response to H₂O₂ and metals, in additional experiments, we used HEK293 cells as an expression system. In cells expressing the P2X_2aR, 1 mm H₂O₂ increased by 1.9 ± 0.3-fold the ATP-evoked currents (Fig. 5A,C, n = 4). The H₂O₂ potentiation had a slow onset and washout, as can be shown in a representative 10 min recording (Fig. 5B). Additionally, 10 μm Hg²⁺ increased by 5.9 ± 1.3-fold the ATP evoked-currents, and 10 μm Cu²⁺ induced a 4.5 ± 0.5-fold increase of the currents (Fig. 5C, n = 4–5). In parallel to experiments with oocytes, the P2X_2bR and the P2X_2a–C430S mutant expressed in HEK293 cells were also completely resistant to H₂O₂ (Fig. 5B,C, n = 4–5). Mercury increased the ATP-gated currents in P2X_2bR and P2X_2a–C430S mutant by 2.5 ± 0.5- and 1.9 ± 0.2-fold, respectively, the values significantly lower to that obtained in the P2X_2aR (p < 0.05) (Fig. 5C, n = 3–5). Copper induced a 5.7 ± 1.5- and 4.2 ± 0.9-fold increase of the ATP-evoked currents in cells expressing P2X_2bR and P2X_2a–C430S mutant, respectively, values that do not differ from that obtained in the P2X_2aR, confirming the extracellular nature of the copper-induced potentiation (Fig. 5C, n = 4).

Figure 5.

Potentiating effects of H₂O₂ on P2X_2aR but not P2X_2bR expressed in HEK293 cells. A, Representative recordings from single cells expressing P2X_2aR, P2X_2bR, and P2X_2a–C430S mutant, showing the effect of 1 mm H₂O₂ on the ATP-evoked currents. Notice that H₂O₂ potentiation is observed only in cells expressing the P2X_2aR. C, Control; W, washout. B, Effects of H₂O₂ on 3 μm ATP-stimulated current during repetitive agonist applications for 10 min in cells expressing P2X_2aR (top) and P2X_2bR (bottom). C, Summary of the effects of H₂O₂, Hg²⁺, and Cu²⁺ on ATP (approximate EC₁₅ concentration) induced peak current in P2X_2aR, P2X_2bR, and P2X_2a–C430S mutant-expressing cells. Data shown are mean ± SEM values. *p < 0.05, Kruskal–Wallis test; n = 4–7 per group.

Effects of oxidative stress inducers

We next tested the effect of myxothiazol and rotenone, two oxidative stress inducers (Thierbach and Reichenbach, 1981; Dawson et al., 1993; Gonzalez-Flecha and Demple, 1995), on the P2X_2aR and P2X_2bR activity. In HEK293 cells expressing the P2X_2aR, 100 nm myxothiazol increased the ATP-evoked currents by 1.95 ± 0.14-fold (Fig. 6A,B, n = 10), a similar value to that observed in H₂O₂ treatment. Myxothiazol also potentiated the ATP-evoked currents in oocytes expressing the P2X_2aR, with a maximal potentiation of 1.95 ± 0.18-fold (Fig. 6B, n = 5). The effect of myxothiazol was biphasic; higher concentrations did not potentiate but rather inhibited the receptor activity (data not shown). In contrast, the P2X_2bR was insensitive to myxothiazol; the ATP-evoked currents were not affected by this agent when the receptor was expressed in both HEK293 cells and oocytes (Fig. 6B, n = 3–5). Rotenone also potentiated the activity of the P2X_2aR; 600 nm induced a 1.52 ± 0.08- and 1.75 ± 0.38-fold increase of the ATP-evoked currents in HEK293 cells and oocytes, respectively (Fig. 6B, n = 5–6). Again, this effect was not observed in HEK293 cells nor oocytes expressing the P2X_2bR (Fig. 6B, n = 6).

Figure 6.

Effects of oxidative stress inducers on ATP-induced currents in P2X_2aR and P2X_2bR expressed in oocytes and HEK293 cells. A, Representative recordings from a HEK293 cell expressing the P2X_2aR during repetitive application of 3 μm ATP in the presence and absence of 100 nm myxothiazol. B, Summary of the effects of myxothiazol or 600 nm rotenone on the P2X_2aR (gray bars) and the P2X_2bR (white bars). In oocytes, these compounds were applied for 10 min before ATP application. Data shown are mean ± SEM values. *p < 0.05, Mann–Whitney test; n = 4–10. C, NAC prevents H₂O₂ or rotenone (Rot)-dependent potentiation of 3 μm ATP-induced currents in HEK293 cells expressing the P2X_2aR; representative traces (left) and mean ± SEM values (right). n = 4–5.

NAC prevents H₂O₂ and rotenone potentiation

We next tested the effect of NAC treatment on H₂O₂ or rotenone-induced potentiation. For this purpose, 100 μm NAC was added to the intracellular patch pipette solution. Under these conditions, neither 1 mm H₂O₂ nor 600 nm rotenone increased the ATP-evoked currents in HEK293 cells expressing the P2X_2aR (Fig. 6C).

Discussion

To understand the structural determinants of trace metal modulation in P2XR physiology, within the past years, we searched for critical residues linked to metal coordination in these receptors. The extracellular His¹²⁰ and His²¹³ are key residues for copper modulation in the P2X_2aR (Clyne et al., 2002; Lorca et al., 2005), whereas His¹⁴⁰ and Asp¹³⁸ are required for copper modulation in the P2X₄R (Coddou et al., 2003, 2007). However, the respective alanine mutants were all sensitive to mercury modulation, although they were completely copper resistant (Coddou et al., 2005; Lorca et al., 2005). These observations fostered the hypothesis that the site of mercury was different from that of copper in both the P2X_2aR and the P2X₄R. In this study, the use of the P2X_4/2 chimeras generated by He et al. (2003) was instrumental to identify the mercury site of action in the P2X_2aR. Comparing the sensitivity to trace metals and mercury in the P2X_4/2a versus the P2X_4/2b chimera, we deduced that the site of mercury was contained within the Val³⁷⁰–Gln⁴³⁸ sequence of the P2X_4/2a variant, located in the intracellular C terminal.

Based on the high chemical reactivity of sulfhydryl groups with mercury (Webb, 1966), we hypothesized that the Cys⁴³⁰ residue was critical for mercury and H₂O₂ modulation. Site-directed mutagenesis of this residue obliterated the mercury and H₂O₂-induced potentiation of the P2X_2aR activity. In addition, the uncharged and the plasma membrane-permeable thiol reagent MMTS (Valentine and Paglia, 1981) blunted both the effect of H₂O₂ and mercury, reinforcing the involvement of Cys⁴³⁰ in the modulation by these chemicals. H₂O₂ and mercury must be transported into the cell to reach their intracellular site of action. H₂O₂ crosses the oocyte cell membrane through aquaporins (Henzler and Steudle, 2000). In addition, an H₂O₂ gradient could be established across biological membranes, favoring its diffusion into the cell (Antunes and Cadenas, 2000; Seaver and Imlay, 2001). Mercury freely crosses the oocyte plasma membrane as a dichloride salt, an ionic species that rapidly diffuses the plasma membrane under our experimental conditions (Gutknecht, 1981). Mercury chloride has a high permeability rate because of its apolar nature and equilibrates with the oocyte cytoplasm within a few seconds (Webb, 1966; Huang and Narahashi, 1996).

According to our hypothesis, the oxidation state of Cys⁴³⁰ may vary depending on the availability of cell antioxidants, including glutathione, ascorbic acid, β-carotene, and vitamin E, among others. Redox couples such as glutathione/glutathione disulfide and NADH/NAD⁺ (nicotinamide adenine dinucleotide, reduced/nicotinamide adenine dinucleotide) will set the cell redox balance (Dröge, 2002); during a shift in the cell oxidation state, reflected in the likely oxidation of Cys⁴³⁰, the P2XR function should increase. In this context, the results obtained with NAC are consistent with the involvement of endogenous glutathione/glutathione disulfide. NAC is a precursor of glutathione (Dröge, 2002), so its addition to cell cytoplasm will favor antioxidant mechanisms and consequently prevent ROS potentiation of P2X_2aR activity. Consequently, in the presence of oxidative stress inducers myxothiazol or rotenone, we consistently observed an increase in the P2X_2aR activity. It is known that these compounds increase ROS production by interfering at mitochondrial complexes (Dawson et al., 1993; Gonzalez-Flecha and Demple, 1995; Starkov and Fiskum, 2001). Mercury can also raise ROS levels acting in the mitochondria (Nieminen et al., 1990). These indirect effects are the most likely mechanism of action of these agents and are summarized in Figure 7. H₂O₂ could also act indirectly, by the formation of the free radical OH⁻, through the Fenton reaction.

Figure 7.

Possible mechanisms of P2X_2aR modulation by ROS, mercury, and stress inducers. The Cys⁴³⁰ residue could be oxidized by indirect and/or direct effects of H₂O₂ and mercury on receptors. Top, Indirect effects of H₂O₂, mercury and the oxidative stress inducers myxothiazol and rotenone could be mediated by ROS and free radicals (FR), including the hydroxyl radical OH⁻ produced by mitochondria. H₂O₂ could also generate OH⁻ through the Fenton reaction. Bottom, H₂O₂ or Hg²⁺ could also oxidize directly the SH group of the Cys⁴³⁰ residue; mercury forming a mercaptide product CH₂–S–Hg²⁺ and H₂O₂ inducing the oxidation of the thiol group to the corresponding sulfonic acid derivative CH₂–SO₂⁻. The mercaptide could be rapidly formed because of the extreme affinity of mercury for thiols, whereas the sulfonic acid reaction might occur more slowly as a result of the formation of several intermediate reactive oxygen species that finally stabilize as the sulfonic acid derivative of Cys⁴³⁰. Endogenous antioxidants such as glutathione (GSH) could partially revert these reactions, resulting in the reduction of the Cys⁴³⁰ residue and formation of glutathione/glutathione disulfide (GS-SG). The exogenous antioxidant DTT can mimic the action of glutathione (data not shown).

Alternatively, mercury or H₂O₂ could directly oxidize the Cys⁴³⁰ residue, leading to the formation of a stable mercaptide or sulfonic acid derivative, a reaction that should be irreversible in the case of mercury. The mercury-induced potentiation of the ATP-evoked currents was slowly reversed after washout. Although the extent of recovery varied between individual cells, we consistently observed an ∼70% decrease in the current amplitude despite prolonged washout of this metal. This finding could be an indication that both direct and indirect mechanisms of the Cys⁴³⁰ residue oxidation might occur. Conversely, endogenous antioxidant defenses, such as glutathione/glutathione disulfide, may reverse the status of the Cys⁴³⁰ residue to its reduced state. Receptor recycling and trafficking of new synthesized receptors might also contribute in part to the recovery of basal receptor activity. Mirzoian and Luetje (2002) also discussed that the effect of mercury chloride on nicotinic receptors reversed slowly, persisting for several minutes after washout. In addition, the effects of mercury are described as slowly reversible or, in other cases, irreversible in several ionic channels and receptors (Gallagher et al., 1995; Huang and Narahashi, 1996; Mirzoian and Luetje, 2002). Interestingly, the effect of mercury on the P2X_2a–C430A and P2X_2a–C430S mutants was easily reversible, in contrast to its long-lasting effect on the wild-type receptor.

Several reasons might be invoked for the potentiation of the P2X_2aR activity by mercury and H₂O₂, but the most likely explanation is an increase in the receptor affinity for ATP, a notion derived from the fact that mercury shifted leftward the ATP concentration–response curve. The alkylthiosulfonation of the Cys⁴³⁰–P2X_2aR residue, in contrast to treatment with mercury or hydrogen peroxide, did not augment the channel activity. This could be explained because MMTS reaction does not lead to a charged Cys⁴³⁰ derivative and is not a redox reaction, like in the case of H₂O₂ or mercury (Fig. 7). Additionally, we observed a residual mercury potentiation that was independent of Cys⁴³⁰ as it can be deduced from Figures 3 and 5C and supplemental Figure 3 (available at www.jneurosci.org as supplemental material). This could reflect that mercury also binds to extracellular receptor residue(s) of the P2X_2aR. Additional work is required to identify this (or these) residue(s).

In contrast to our results, Mason et al. (2004) reported that P2X₂R-mediated currents were inhibited by H₂O₂ in HEK293 cells, with a median effective concentration close to 1 mm. Moreover, H₂O₂ mimicked the hypoxia-induced P2X₂R attenuation, establishing a link between receptor activity and ROS production. This finding prompted us to extend work with oocytes using HEK293 cells as an expression system. These experiments excluded the possibility that host cells influence effects of ROS and mercury. Several other reasons might explain these differences. We never exceeded 1 mm H₂O₂ because larger concentrations resulted in cell damage and membrane dissolution. In experiments with both cell types, we always observed a recovery of the response, even in long-lasting protocols (Fig. 5). We can rule out nonspecific effects because 1 mm H₂O₂ potentiates the P2X_2aR but not the P2X_2bR activity. In the work of Mason et al. (2004), it was also observed a transient increase followed by sustained inhibition of the P2X₂R-mediated currents after application of myxothiazol and rotenone. It is most likely that this transient increase corresponds to the potentiation that we now report, whereas the subsequent inhibition could be related to the cell death. These stress inducers were initially thought to decrease ROS production, but there is growing evidence that these agents do not diminish but rather increase ROS generation (Starkov and Fiskum, 2001).

Besides the diverse effects of ROS on voltage- and ligand-gated ionic channels (see Introduction), it has been proposed that H₂O₂ plays a role as an endogenous signal in physiological and pathophysiological conditions. In vascular tissues, for example, H₂O₂ might be one of the endothelium-derived hyperpolarizing factor(s) (Shimokawa and Matoba, 2004) and could participate in the intracellular signaling of the angiotensin II receptors augmenting NADPH oxidase activity (Touyz and Schiffrin, 2001), both of them important in regulation of the vascular tone. Moreover, several reports indicate that transduction pathways associated with protein kinase C or the arachidonic acid cascade can also regulate ROS production (Severson and Hee-Cheong, 1989). In the brain, H₂O₂ has been linked to long-term potentiation through protein kinase C or extracellular signal-regulated kinase activity or inhibiting phosphatases in the consolidation phase of long-term potentiation (Knapp and Klann, 2002).

In summary, we identified the Cys⁴³⁰ residue located in the P2X_2aR intracellular C terminal as a target for mercury and H₂O₂-induced potentiation of ATP-evoked currents. Because both agents decreased the P2X₄R activity, it is unlikely that the potentiation effects observed are attributable to nonspecific chemical reactions with other thiol groups of adjacent proteins or membrane components. The identification of Cys⁴³⁰ as a redox site in the P2X_2aR allowed us to hypothesize that this intracellular residue senses the cell redox state, setting a novel property for the P2XRs. The present findings will allow the identification of other members of the P2XR family that may also have a redox sensor, unraveling the implications of intracellular cysteine residues on receptor physiology. Altogether, these findings highlight that P2XRs are not only sensors of extracellular ATP and allosteric modulators but may also respond to intracellular stimuli that vary depending on the metabolic state of the cell.