Histidine 140 Plays a Key Role in the Inhibitory Modulation of the P2X 4 Nucleotide Receptor by Copper but Not Zinc*

To elucidate the role of extracellular histidines in the modulation of the rat P2X 4 receptor by trace metals, we generated single, double, and triple histidine mutants for residues 140, 241, and 286, replacing them with ala-nines. cDNAs for the wild-type and receptor mutants were expressed in Xenopus laevis oocytes and in human embryonic kidney 293 cells and examined by the two electrode and patch clamp techniques, respectively. Whereas copper inhibited concentration-dependently the ATP-gated currents in the wild-type and in the single or double H241A and H286A receptor mutants, all receptors containing H140A were insensitive to copper in both cell systems. The characteristic bell-shaped con-centration-response curve of zinc observed in the wild-type receptor became sigmoid in both oocytes and human All the out

The notion that trace metals such as zinc or copper are atypical brain messengers has attracted much attention in view of their emerging role in the modulation of brain excitability (1). The importance of trace metals in synaptic activity is highlighted by the description that both copper and zinc are stored in synaptic vesicles from where they are released by electrical depolarization, reaching a high micromolar concentration at the synaptic cleft (2)(3)(4). Trace metals are known to modulate a wide variety of brain ionotropic receptors such as glycine, N-methyl-D-aspartate, ␥-aminobutyric acid, nicotinic, and the novel nucleotide receptors (P2X) family (5)(6)(7)(8)(9). The P2X purinoceptors are homomeric or heteromeric membrane channels gated by extracellular ATP and related synthetic nucleotides; North (10) recently reviewed the principles of the molecular physiology of this family of receptors. Within the P2X receptor family, the P2X 4 is the most widely distributed in the central nervous system, including the cerebellum and the CA1 region of the hippocampus, where it has been proposed to play a role in glutamatergic synapses (11).
The P2X 4 receptor is an interesting model of an ionic channel differentially modulated by trace metals. Acuñ a-Castillo et al. (12) and Coddou et al. (13) reported that zinc potentiates the ATP-evoked currents whereas copper has an inhibitory effect on the activity of this receptor. Based on these findings, Acuñ a-Castillo et al. (12) proposed that trace metals modulate the activity of the P2X 4 receptor via two separate metal-binding sites. One of these sites has a preferential selectivity for copper and is characterized by a non-competitive inhibition of the ATP-gated channel activity. The second site shows preference for zinc and apparently leads to an increase in affinity for the binding of ATP. Copper and zinc also modulate other members of the P2X family of receptors, but their effects depend on the receptor subtype. For example, both zinc and copper facilitate the ATP-evoked responses on the P2X 2 receptor (14,15) whereas in the P2X 7 receptor, copper inhibits the ATP-gated currents, whereas zinc is virtually without effect (13). These findings further support the notion that the P2X purinoceptors may have separate metal-binding sites to modulate the channel activity.
This research was aimed at identifying structural determinants for the binding of copper and to assess the hypothesis that copper and zinc bind to separate, independent metalbinding sites in the receptor. Based on the notion that histidine residues are commonly found in consensus copper-binding motifs (16), as extensively studied in superoxide dismutase, where the imidazole ring is essential to coordinate copper (17), and both histidines and cysteines have been defined in zinc-binding motifs (18), we were challenged to clarify the role of the three histidines of the rat P2X 4 receptor in the modulatory action of trace metals. Interestingly, the three rat P2X 4 histidine residues are located extracellularly. To tentatively identify putative residues involved in the copper inhibitory modulation, we preliminarily used chemicals to selectively alkylate histidine and cysteine residues. Next, we performed site-directed mutagenesis to create single, double, and triple mutants where the histidine residues (His-140, His-241, and His-286) of the receptor were replaced by alanines. The cDNAs coding for the wildtype and mutated receptors were expressed in two cell systems; the modulation by copper or zinc was examined in either Xe-nopus laevis oocytes and HEK 1 293 cells. The present results allow us to conclude that only His-140 plays a key role in the modulator action of copper, because this particular amino acid cannot be replaced by alanine, threonine, or arginine and still conserve the inhibitory modulation by copper.

EXPERIMENTAL PROCEDURES
Materials-ATP trisodium salt and penicillin-streptomycin were purchased from Sigma. Copper and zinc chlorides were obtained from Merck. The salts used to prepare the incubation media were purchased from Sigma-Aldrich or Merck. Diethyl pyrocarbonate (DEPC) and Nethylmaleimide (NEM) were purchased from Sigma. Dulbecco's modified Eagle's medium, Opti-MEM, and LipofectAMINE were purchased from Invitrogen. Triple-distilled water with minimal electroconductivity was used; the copper or zinc contamination was less than 0.1 M.
Plasmid DNA was isolated using the Qiagen plasmid DNA purification kit according to the manufacturer's instructions. All the mutations were confirmed by sequencing using an Applied Biosystems ABI PRISM 310 DNA sequencer.

Expression of P2X 4 Receptors and Electrophysiological Recordings from X. laevis Oocytes
A segment of the ovary was surgically removed under anesthesia from X. laevis frogs; oocytes were manually defolliculated and next incubated with collagenase as detailed previously (12). Oocytes were injected intranuclearly with 3-5 ng of cDNA coding for either the wild-type or H140A, H241A, and H286A P2X 4 receptor proteins. After 36 -48 h of incubation at 15°C in Barth's solution (NaCl, 88 mM; KCl, 1 mM; NaHCO 3 , 2.4 mM; HEPES, 10 mM; MgSO 4 , 0.82 mM; Ca(NO 3 ) 2 , 0.33 mM; CaCl 2 , 0.91 mM; pH 7.5) containing 10 IU/liter penicillin/10 mg streptomycin and 2 mM pyruvate. Oocytes were clamped at Ϫ70 mV using the two-electrode voltage-clamp configuration with an OC-725C clamper (Warner Instruments Corp). ATP-gated currents were recorded following a 15-s ATP application dissolved in the regular perfusion buffer. ATP and metal chloride salts dissolved in Barth's medium were perfused using a peristaltic pump operating at a constant flow of 2 ml/min. Non-injected oocytes did not evoke currents when exogenous ATP was applied. Each experimental protocol was performed in at least two separate batches of oocytes from different frogs; each experiment was replicated in at least four separate oocytes. The relative ATP potency was estimated from single concentration-response curves; the median effective concentration (EC 50 ) and the Hill coefficient (n H ) was graphically derived using Graph Pad software. The maximal current (I max , nanoAmpere) was obtained from each oocyte challenged with a saturating ATP concentration.

Chemical Alkylation of Histidine or Cysteine Residues
To ascertain whether certain amino acids are essential for trace metal modulation, we first assessed whether alkylation of histidine residues with DEPC altered the inhibitory action of copper or the facilitator action of zinc. After routine applications of 10 M ATP and testing of the inhibitory action of 10 M copper or the facilitator action of 10 M zinc, oocytes were treated for 5 min with 500 M DEPC to modify the histidine residues. After this treatment, oocytes were perfused for 20 -30 min to wash out the remaining DEPC prior to challenges with ATP or the metals. ATP concentration-response protocols (1-1000 M) of ATP alone or ATP plus copper (1 min pre-incubated) were performed in separate oocytes. The ATP concentration-response curve was normalized against 500 M ATP, a value close to the maximal current. The current gated by 500 M ATP was obtained prior to the treatment with DEPC. Likewise, similar protocols were performed in separate oocytes using 300 M NEM to selectively alkylate cysteine residues. As in the case of the DEPC-treated oocytes, oocytes were washed for 20 -30 min from the residual chemical prior to testing with ATP and the trace metals. Results were normalized, as in the case of the DEPC treatment, against a standard of ATP obtained prior to applying the alkylating agent.

Trace Metal Concentration-response Curves
To examine the effect of copper, 0.1-600 M of the metal was preapplied for 1 min before a 15-s ATP pulse, at a concentration close to the receptor EC 50 , which in the wild-type receptor was 10 M ATP. Care was taken to achieve full reversal of the copper effect prior to a larger copper application. The median inhibitory concentration of copper (IC 50 ) was interpolated from each metal concentration-response curve. A single oocyte was tested with at least four increasing concentrations of copper. If the relative potency of ATP was modified in the mutant receptors, the ATP concentration was adjusted to a value close to its EC 50 ; 60 M ATP was used to evaluate the H140A mutant receptor, and 30 M ATP was used for the His-286 mutant. This protocol was also used to evaluate the potency of copper in the double or triple histidine mutants or the mutants that substituted His-140 by either threonine or arginine.
Parallel experiments were performed to assess whether the histidine mutants affected the facilitator action of zinc. In these protocols, the metal was applied 1 min prior to a challenge with 1 M ATP in the wild-type P2X 4 receptor or the H241A mutant. 3 M ATP was used to evaluate the facilitator action of zinc in the H140A and H286A mutants. These concentrations of ATP were chosen based on previous findings (12) when we reported that the largest zinc potentiation was attained with low ATP concentrations, which induce 5% of the maximal nucleotide-evoked current (EC 5 ). In all cases, ATP applications were spaced at regular 10-min intervals, minimizing desensitization and after attaining full reversibility of the zinc effect. The effect of zinc in the double or triple mutants was assessed likewise.

ATP Concentration-response Protocols
To study with more detail the mechanism of trace metal modulation, we next assessed the effect of the pre-application of copper or zinc in the ATP concentration-response curves in the wild-type and mutant P2X 4 receptors. Protocols were performed applying increasing concentrations of ATP (0.1-3,000 M) for 15 s. Curves were normalized against 500 M ATP for the wild-type receptor and H241A mutant receptor and against 1 mM ATP for the H140A and H286A mutants. The ATP EC 50 and its I max were derived from each ATP concentration-response curve.

Transfection of HEK 293 Cells and the Recording of the ATP-gated Currents from the Whole Cell Configuration
HEK 293 cells were transiently transfected with 1 g of cDNA coding for the wild-type, H140A, H241A, or H286A P2X 4 receptor protein. Cells grown to 50% confluence on 35-mm culture dishes were incubated with cDNA mixed with 8 l of LipofectAMINE in 1 ml of serum-free medium (Opti-MEM). After 3 h at 37°C, the medium was replaced with Dulbecco's modified Eagle's medium, and 48 h later the cells were transferred to the experimental chamber containing recording buffer, which contained the following (in mM): 150 NaCl, 1 CaCl 2 , 1 MgCl 2 , 10 glucose, 10 HEPES, pH 7.3. Whole-cell ATP-dependent currents were obtained from single HEK 293 cells using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Patch pipettes (2-4 megohm) were filled with the following (in mM): 150 CsCl, 10 tetraethyl ammonium chloride, 10 EGTA, 10 HEPES, pH 7.3, and 275-285 mOsm; pH was adjusted with CsOH. The junction potential (typically Ͻ5 mV) was compensated. Only cells with membrane potential more negative than Ϫ55 mV, access resistance Ͻ10 megohm (compensated at 80%), and input resistance Ͼ100 megohm were tested. The cells were discarded if the input or access resistance changed Ͼ0%. The ATP-dependent currents were recorded at a holding potential of Ϫ80 mV to augment the driving force of the ions that mediate the ATP-evoked currents and to avoid contamination of voltage-dependent currents. Responses were digitized at a frequency of 10 KHz and analyzed using the pCLAMP 8 system (Axon Instruments, Foster City, CA). All the experiments were carried out at room temperature.

Recording of ATP-evoked Currents: Effect of Trace Metals
ATP-dependent currents were evoked with 5-s ATP pulses applied by superfusion, using a peristaltic pump (Masterflex Pump; Cole Palmer Instruments). ATP concentration-response protocols were performed to derive the EC 50 , by applying 0.1 to 500 M pulses of ATP for 5 s; between ATP applications, the cells were washed for 5 min with drugfree buffer. To minimize desensitization because of the larger concentrations of ATP (30 M or more), these applications were spaced 10 min. The ATP EC 50 , I max , and n H were obtained from each ATP concentration-response plot by fitting the data using the Hill equation. Full concentration-response curves were derived from single cells; protocols were repeated using several batches of cells.
To study the modulation of the ATP-evoked currents by the metals, separate cell batches expressing the wild-type or the mutant receptors were pre-incubated with 1-300 M copper or 1-100 M zinc for 1 min before the addition of a challenge concentration of 10 M ATP, a value in the order of the nucleotide EC 50 . ATP applications were regularly spaced at 10-min intervals, the optimal time previously determined to attain full recovery of the ATP-evoked current, an indication of the reversibility of the metal effects. Metal concentration-response plots were fitted using Graph Pad software; the copper IC 50 was interpolated from each plot.

Statistical Analysis
Curve fitting was performed using the sigmoidal equations proper of concentration-response curves (GraphPad Prism 3.0; San Diego, CA). ATP and zinc EC 50 , I max , and n H , as well as the copper IC 50 were derived per oocyte in each experimental protocol (13). Maximal amplitudes were also derived. Non-parametric analysis and parametric tests were performed as appropriate (Kruskal-Wallis, Friedman, and Quade (see Ref. 20 for review) and Dunnetts's tests for multiple comparisons with a single control); significance was set at p Ͻ 0.05.

Chemical Modification of Histidine Residues with DEPC Impairs the Inhibitory Action of Copper
Copper non-competitively inhibited the currents evoked by ATP in wild-type receptors expressed in oocytes (Fig. 1A). DEPC treatment, which is expected to alkylate-exposed histidine residues, decreased the ATP I max (44.5 Ϯ 7.8%, p Ͻ 0.01) and abolished the inhibitory action of copper (Fig. 1B). In contrast, NEM treatment, which preferentially modifies the SH group of cysteines, significantly increased the ATP I max (166.3 Ϯ 24.5%, p Ͻ 0.001). Moreover, NEM treatment did not abolish the non-competitive copper inhibition (Fig. 1C), suggesting that histidines, rather than cysteines, are primarily involved in the action of copper.

Functional Characteristics of Wild-type and Mutated P2X 4 Receptors Expressed in Oocytes
Both in the wild-type and the receptor mutants, the 15-s ATP pulse application resulted in currents with a distinct rapid peak and a slower component. The P2X 4 receptor mutants resulted in functional channels; their ATP EC 50 and I max values are summarized in Table I. Compared with the wild-type, the ATP EC 50 of the H140A and H286A mutants was 6-and 4-fold larger than the wild-type receptor; mutant H241A conserved the wild-type EC 50 (Table I). On the other hand, the H286A mutant I max was reduced by 67% (p Ͻ 0.01), whereas that of the H140A and H241A mutants was unaltered. The double and the triple histidine mutants with the H140A and H286A substitution also had a lower I max and a larger EC 50 than the wild-type (Table I).

Characterization of the Trace Metal Modulation in Wildtype and Mutated P2X 4 Receptors Expressed in Oocytes
Copper Modulation-Copper decreased in a reversible and concentration-dependent manner the current elicited by a concentration of ATP close to the EC 50 of each receptor (Fig. 2). The copper IC 50 derived from these experiments was 8.9 Ϯ 1.5 M (n ϭ 7) for the wild-type receptor, a value that was not significantly different from 8.7 Ϯ 2.4 (n ϭ 4) and 7.1 Ϯ 2.7 M (n ϭ 6) for the H241A and H286A mutants, respectively. However, the H140A mutant was largely resistant to the modulator action of copper; 600 M copper inhibited only 33.8 Ϯ 8.5% (n ϭ 4; see Fig. 2), suggesting that His-140 is essential for the copper-induced inhibition.
Zinc Modulation-Zinc facilitated in a biphasic manner the ATP-gated currents; a maximum was reached at 10 M zinc, which amounted to a 7.1 Ϯ 0.7-fold (n ϭ 7) increase over the current elicited by an ATP EC 5 , the preferred concentration used to evaluate the zinc modulation. Larger metal concentrations elicited smaller effects, resulting in a bell-shaped curve (recordings and inset in Fig. 3). The H241A and H286A mutants also showed a similar biphasic zinc curve; the maximal potentiation of the currents was 8.9 Ϯ 0.6-and 8.0 Ϯ 2.4-fold for the H241A and H286A mutants, respectively (n ϭ 4; see Fig. 3). In contrast, the zinc curve for the H140A mutant was sigmoid, reaching a maximal potentiation at 30 M zinc, which resulted in a 27.4 Ϯ 7.7-fold potentiation (p Ͻ 0.001, n ϭ 4; see Fig. 3). This potentiation was 4-fold larger than for the wildtype receptor.

Mechanisms of Trace Metal Modulation: ATP Concentration-
response Curves-In contrast to the wild-type receptor, 10 M copper did not modify the ATP concentration-response curve in the H140A mutant, whereas it non-competitively inhibited the H241A and the H286A receptor mutants (see Fig. 4 and Table  II). 10 M zinc displaced the ATP-concentration-response curve of the H140A mutant to the left (Fig. 4) and additionally increased the I max from 104 Ϯ 8.3% in the wild-type to 205 Ϯ 21.8% (p Ͻ 0.001; see Table II). Zinc also increased the ATP I max of the H286A mutant to 238.2 Ϯ 58.2% (p Ͻ 0.001; see Table II and Fig. 4). The effect of zinc in the H241A mutant was identical to that of the wild-type receptor.
Trace Metal Modulation in the Double and Triple Histidine Mutants-The mutants containing the H140A substitution were resistant to the inhibitory action of up to 100 M copper. In contrast, the mutant containing the other two histidines conserved the inhibition by copper and behaved similar to that of the wild-type receptor (Fig 5A).
Zinc potentiated the magnitude of the ATP current in the H140A/H241A/H286A receptor, reaching a plateau at 100 M; the maximal potentiation was 32.9 Ϯ 5.5-fold, a value significantly larger than that observed in the wild-type receptor (p Ͻ 0.01; see Fig. 5B). In the double receptor mutants H140A/ H241A and H140A/H286A, the maximal zinc potentiation amounted to 12.0 Ϯ 6.0% and 11.7 Ϯ 4.7-fold, respectively (p Ͻ 0.01 each, n ϭ 5 each), values which are two-three times smaller than those attained with the H140A mutant. In contrast, the H241A/H286A mutant had a biphasic zinc curve, similar to that observed in the wild-type receptor (Fig. 5B); the maximal potentiation reached only 6.2 Ϯ 1.6-fold and served as a negative control, because the maximal potentiation in the wild-type receptor reached 7.1 Ϯ 0.7-fold.
Substitution of His-140 by Other Amino Acids-The copper IC 50 in the H140T mutant was 78.5 Ϯ 31.2 M (n ϭ 4), a value 8-fold higher than in the wild-type receptor (n ϭ 6, p Ͻ 0.01; see Fig. 6). The zinc curve was biphasic; its maximal effect reached 17.3 Ϯ 2.9-fold, a value 2.5-fold higher than the wild-type receptor (p Ͻ 0.01), but slightly smaller than the maximal potentiation obtained in the H140A mutant (27.4 Ϯ 7.7-fold increase; see Fig. 6).
The currents elicited by the H140R mutant were 100-fold smaller than the wild-type and were long lasting, even after ATP removal (Fig. 6, tracings). The estimated ATP EC 50 was 32.2 Ϯ 7.4 M (n ϭ 3; see Table I); furthermore, the Hill coefficient is about 3-fold less than the wild-type receptor. The weak currents attained with this mutant precluded a detailed analysis of its interaction with trace metals; in the few successful recordings, the currents were insensitive to the action of up to 100 M copper (Fig. 6); zinc was not evaluated, because the currents elicited by 1 M ATP were almost undetectable.

Patch Clamp Technique Confirms Trace Metal Modulation of the P2X 4 Receptor in HEK 293 Cells ATP Potency in Wild-type and Single Histidine Receptor Mutants-
The ATP EC 50 of the wild-type receptor was similar to that obtained in oocytes (Table I). In contrast to the oocytes, the ATP EC 50 values of the three single histidine mutants and the wild-type receptor showed no difference in their affinity for ATP (Table I). The H286A mutant had a lowered I max , which reached only 24% of that attained by the wild-type value (p Ͻ 0.001) as in the oocytes.
Trace Metal Modulation-Copper inhibited in a reversible and concentration-dependent manner the ATP-evoked currents; its IC 50 was 9.2 Ϯ 4.5 M (n ϭ 5) in the wild-type receptor (see Fig. 7, A and B and Table III). The H140A mutant was resistant to the inhibitory action of up to 300 M copper; in contrast, this concentration showed a minor, but consistent, current increase (see Fig. 7B and Table III). The copper IC 50 in the H241A and H286A was 9.3 Ϯ 3.0 and 11.0 Ϯ 4.5 M, respectively (n ϭ 4, each). In the wild-type receptor zinc showed a clear tendency to a biphasic curve, although it was not as marked as that in oocytes (Fig. 7C). In contrast, zinc caused a linear potentiation in the H140A mutant; the potentiation attained with 100 M zinc was 2.2-fold larger than in the wild-type receptor (p Ͻ 0.05; see Fig. 7C). Mutation of the other histidines had no effect on the zinc-induced potentiation (see Table III and Fig. 7A).

DISCUSSION
The present results identify His-140 as a key residue involved in the inhibitory modulator role of copper in the P2X 4 receptor. Consistent with our hypothesis that the rat P2X 4 receptor has separate extracellular allosteric copper-and zincbinding sites, the present results show that the H140A mutation does not eliminate the positive modulator role of zinc; on the contrary, it markedly augments it. The combined use of oocytes and HEK 293 cells allowed us to confirm P2X 4 receptor trace metal modulation in two different cell expression model  Zinc displaced leftward the ATP concentration-response in wild-type and H241A receptors and increased the maximal response in H140A and H286A receptors. Copper inhibited in a non-competitive fashion the ATP concentration-response curve in wild-type, H241A, and H286A and had no effect on the H140A receptor mutant. Curves represent the mean of four-six separate experiments in each case. Symbols represent the mean values, and bars represent the S.E. systems using complementary experimental approaches. Overall, both methodologies yielded remarkably consistent data: the potency of ATP in the wild-type P2X 4 receptor, the n H of the ATP concentration-response plots, as well as the modulator characteristics of the trace metals, are similar in both cell expression systems, independent of the recording techniques used.
The potency of ATP in the H140A and/or H286A mutant receptors was lower in oocytes than in HEK 293 cells. This discrepancy does not deter our interpretations, because in both cell expression systems His-140 is essential for the inhibitory copper modulation. Consistently, the His-241 mutation has no effect on the properties of the receptor, including the modulator role of trace metals, confirming the specific contribution of His-140 in the inhibitory action of copper. Additionally, the I max of H286A mutant expressed in either system was reduced to a similar extent, suggesting that His-286 could play a role either in the expression or in the assembly of the subunits that compose the channel or modify the ATP binding or the channel conductance, in view of its proximity to the alleged ATP phosphate chain binding site (21).
Histidines are known to form part of consensus copper-binding motifs (16). In view of the lack of a linear consensus copperbinding motif in the primary sequence of the P2X 4 receptor, we hypothesize that the copper coordination site might be formed by amino acids distant in the primary structure of the protein and that come in close proximity upon protein folding. DEPC was used to specifically alkylate the imidazole group of histidines (22); this treatment resulted in the loss of the copper inhibition and a reduction in the P2X 4 ATP-evoked currents of the receptor. We are aware that there are controversial results reported in the literature with respect to the use of DEPC, because some authors have successfully modified histidine residues (23)(24)(25), whereas others have failed to detect significant changes (26,27). Considering that the rat P2X 4 receptor has only three histidine residues, it became challenging to establish which of these imidazole-bearing residues is responsible for the negative modulator role of copper, as well as the decreased channel activity. Site-directed mutagenesis revealed the critical role of His-140 but not His-241 or His-286 in the modulation by copper. Consistent with this interpretation, the copper concentration-response curve of the H241A/H286A mutant could be superimposed to that obtained in the wild-type receptor.
To clarify the specific requirement of His-140 on the binding of copper, we replaced this residue by threonine and arginine. The results obtained with these mutants confirmed the requirement of this particular imidazole at the copper-binding site to maintain the maximal copper effect. These results further stress the stringent spatial and chemical requirements of His-140 for the action of copper but not zinc. The finding that the substitution of His-140 by a threonine residue did not abrogate the effect of copper suggests that other amino acids participate in the full inhibitory effect of copper. Computerized models of the copper-binding site of the P2X 4 receptor revealed that Thr-123, Ser-124, Ser-141, and Thr-146 are found adjacent to His-140 in the receptor three-dimensional structure. 2 It is therefore plausible that these amino acids may participate in the binding of copper, as threonine has been reported to form part of the copper-binding site in plastocyanin (28). The H140R mutant dramatically altered the receptor conformation and properties precluding its further evaluation. The mechanism of the copper coordination and the identification of adjacent amino acid residues that participate in the copper coordination complex remain as open issues.
The H140A mutant shows a dramatic change in the zinc concentration-response curve from a bell-shaped curve to a sigmoid. Two possible mechanisms may account for this novel finding. Either the trace metal-binding sites are allosterically coupled in such a way that interactions among them might occur, or the receptor has two allosteric metal-binding sites. One of the sites binds copper or zinc and causes inhibition (with higher copper affinity than zinc affinity), and the other site binds zinc but not copper and causes potentiation. Large zinc concentrations interact at the copper-preferring inhibitory metal-binding site. The lack of the inhibitory copper-binding site in the H140A mutant resulted in an augmented and sigmoid zinc concentration-response curve. Altogether, the present results are consonant with our working hypothesis that the copper-inhibitory and zinc-facilitator binding sites must be distinct.
The biphasic nature of the zinc modulation has also been observed in other ligand-gated ion channels. For example, the neuronal nicotinic receptors can be either facilitated or inhibited by zinc, depending on the combination of receptor subunits (24). Recombinant homomeric glycine receptors expressed in HEK cells also show a biphasic action of zinc, an effect postulated to occur by metal binding at two separate sites on the ␣ subunit and involving two histidine residues for its inhibitory 2 P. Bull, personal communication. modulation but not for its potentiating effect (23). The P2X 2 receptor, another member of the P2X receptor family, is potentiated by zinc (14,15); two histidines have been proposed to coordinate zinc in its extracellular domain (29). Because copper also positively modulates the P2X 2 receptor, it is not unlikely that the P2X 2 receptor possesses only a non-selective facilitator trace metal-binding site or a very low affinity inhibitory metal binding-site. In most ionotropic receptors zinc modulates via a facilitator action; however, the N-methyl-D-aspartate receptor is inhibited by zinc. Interestingly, the inhibitory action of zinc depends on the subunit conformation and apparently on critical histidine (30) and/or cysteine residues (31). In this particular glutamate receptor, zinc fails to show a biphasic nature, stressing that the modulator role of this metal varies between receptors and is not an intrinsic metal property.
Although this investigation focused on the modulator role of copper and zinc, we have not ignored that the selectivity of the P2X 4 receptor metal-binding sites might not be exclusive for these trace metals and may even encompass pH modulation. In the human P2X 4 receptor, the pH sensor is related to His-286, because the replacement of this residue by an alanine abolishes the pH-associated effect (27). The H286A mutant studied here showed a lower channel conductance, making it likely that this amino acid could be acting as a pH sensor in both rat and human receptors. We also do not exclude that these trace metal-binding sites are the sites of action of other metals, including lead, cadmium, and mercury, known to produce brain toxicity. In this regard, Acuñ a-Castillo et al. (12) showed that mercury inhibited whereas cadmium facilitated the ATPevoked currents in the P2X 4 receptor; other investigators have shown similar metal preferences as summarized by North and Surprenant (32). These mutants are interesting tools to investigate the selectivity of the inhibitory copper-preferring site of the P2X 4 receptor with respect to heavy metals and brain toxicity.
Trace metals such as copper and zinc may contribute to brain excitability through their modulator role on P2X receptors. Zinc is well characterized as a modulator of glutamate neurotransmission, particularly via N-methyl-D-aspartate receptors (1,7). Zinc and copper may also interact with voltage-gated ionic channels and may be important contributors to brain excitability (33).
In conclusion, the present results unequivocally identify His-140 as a key amino acid in the inhibitory metal-binding site of the P2X 4 receptor and demonstrate that this receptor has two separate and distinct trace metal binding-sites, explaining the differential modulation exerted by copper and zinc. As to whether the neuromodulator role of trace metals is operant in vivo remains a challenging question. The proposal that trace metals are novel atypical brain transmitters (1) underscores the notion that trace metals may have important consequences in synaptic activity that may be related to profound changes in brain functioning.