Zn2+ and H+ Are Coactivators of Acid-sensing Ion Channels*

Acid-sensing ion channels (ASICs) are cationic channels activated by extracellular protons. They are expressed in sensory neurons, where they are thought to be involved in pain perception associated with tissue acidosis. They are also expressed in brain. A number of brain regions, like the hippocampus, contain large amounts of chelatable vesicular Zn2+. This paper shows that Zn2+ potentiates the acid activation of homomeric and heteromeric ASIC2a-containing channels (i.e.ASIC2a, ASIC1a+2a, ASIC2a+3), but not of homomeric ASIC1a and ASIC3. The EC50 for Zn2+ potentiation is 120 and 111 μm for the ASIC2a and ASIC1a+2a current, respectively. Zn2+ shifts the pH dependence of activation of the ASIC1a+2a current from a pH0.5 of 5.5 to 6.0. Systematic mutagenesis of the 10 extracellular histidines of ASIC2a leads to the identification of two residues (His-162 and His-339) that are essential for the Zn2+ potentiating effect. Mutation of another histidine residue, His-72, abolishes the pH sensitivity of ASIC2a. This residue, which is located just after the first transmembrane domain, seems to be an essential component of the extracellular pH sensor of ASIC2a.

ASIC1a is present in brain and sensory neurons, whereas its splice variant ASIC1b is found only in sensory neurons. Both ASIC1a and ASIC1b mediate fast inactivating currents upon modest but rapid acidification of the external medium. ASIC2a is substantially expressed in the brain, whereas its variant ASIC2b is present in both brain and sensory neurons. ASIC2b has no activity on its own but can form functional heteromers with other ASIC subunits and particularly ASIC3 (7). ASIC3 is specifically found in the small nociceptive sensory neurons and generates a biphasic current with a fast inactivating phase, followed by a sustained component (10). All the ASIC subunits share the same topological organization with intracellular N and C termini and two putative transmembrane domains flanking a large cysteine-rich extracellular loop (16).
In sensory neurons, ASIC currents are thought to play an important role in nociception during a tissue acidosis, for instance in muscle and cardiac ischemia (18 -23) and in inflammation (24). It has been also proposed that some might participate in touch sensation (25,26). Their function in the central nervous system is less documented. An important role of ASICs in signal transduction associated with local pH variations during normal neuronal activity has been proposed (27,28). They might also be involved in pathological situations such as brain ischemia and epilepsy that produce significant extracellular acidification.
Some ASICs are expressed in brain regions that contain large amounts of chelatable Zn 2ϩ . For instance, presynaptic vesicles of glutamatergic hippocampal terminals contain Zn 2ϩ in up to millimolar concentrations (29,30). Zn 2ϩ corelease with the neurotransmitter results in a transient increase of the local synaptic Zn 2ϩ concentration up to 100 -300 M from resting levels below 500 nM (31)(32)(33)(34)(35), and has the potential to alter the behavior of various membrane channels and neurotransmitter receptors (35)(36)(37).
Here we show that Zn 2ϩ potentiates the activation of ASIC2a-containing channels. We have used site-directed mutagenesis to investigate the structural determinants of Zn 2ϩ coactivation in the extracellular loop of the ASIC2a subunit. We have identified two histidine residues, His-162 and His-339, that are essential for the Zn 2ϩ potentiating effect, but also another histidine residue, His-72, which causes a large shift in pH sensitivity of ASIC2a when mutated and could be an important component of the pH sensor.

Expression of ASIC Currents in Xenopus oocytes-Xenopus laevis
were purchased from CRBM (Montpellier, France). Pieces of the ovary were surgically removed, and individual oocytes were dissected in a saline solution (ND96) containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 2 mM MgCl 2 , and 5 mM HEPES (pH 7.4 with NaOH). Stage V and VI oocytes were treated for 2 h with collagenase (1 mg/ml, type Ia, Sigma) in ND96 to remove follicular cells. cRNA of rat ASIC1a, WT and mutated ASIC2a and ASIC3 were synthesized with the mCAP RNA capping kit from Stratagene and injected (0.15-2.5 ng/oocyte) using a pressure microinjector. The oocytes were kept at 19°C in the ND96 saline solution supplemented with penicillin (6 g/ml) and streptomycin (5 g/ml). Oocytes were studied within 1-3 days following injection. In a 0.3-ml perfusion chamber, a single oocyte was impaled with two standard glass microelectrodes (1-2.5-megohm resistance) filled with 3 M KCl and maintained under voltage clamp using a Dagan TEV 200 amplifier. Currents were sampled at 100 Hz and low pass-filtered at 500 Hz. Data acquisition and analysis were performed using pClamp software (Axon Instruments). All experiments were performed at room temperature (20 -22°C) in ND96 bathing solution. For measurements of ASIC currents, changes in extracellular pH were induced by rapid perfusion of the experimental chamber. Depending on the pH range, acidic solutions were buffered with HEPES (pH Ͼ 6), MES (pH between 6 and 5), or acetate (pH Ͻ 5).
Expression of ASIC Currents in COS Cells-COS cells, at a density of 20,000 cells/35-mm diameter Petri dish, were transfected with a mix of pCI-CD8 and the following plasmids: pCI-rASIC1a, pCI-rASIC2a, or pCI-rASIC3 (1:5 ratio) using the DEAE-dextran method. Cells were used for electrophysiological measurements 1-3 days after transfection. Successfully transfected cells were recognized by their ability to fix CD8 antibody-coated beads (Dynal). Ion currents were recorded using the whole cell patch-clamp technique (38). Data were sampled at 500 Hz and low pass-filtered at 3 kHz using pClamp8 software (Axon Instruments). The statistical significance of differences between sets of data was estimated by single-sided Student's t test. The pipette solution contained (in mM): KCl 140, NaCl 5, MgCl 2 2, EGTA 5, HEPES 10 (pH 7.35); and the bath solution contained (in mM): NaCl 150, KCl 5, MgCl 2 2, CaCl 2 2, HEPES 10 (pH 7.45). MES or acetate were used instead of HEPES to buffer bath solution pH ranging from 6 to 5, and from 4.5 to 3, respectively. ZnCl 2 was added to the bath solution at concentration ranging from 1 M to 10 mM. Changes in extracellular pH were induced by shifting one out of six outlets of a microperfusion system in front of the cell. Experiments were carried out at room temperature (20 -22°C).
Site-directed Mutagenesis-Mutants were prepared by polymerase chain reaction using a modification of the method of gene splicing by overlap extension (39) or using the GeneEditor in vitro site-directed mutagenesis system from Promega. The amplified products were digested by EcoRI and XhoI and subcloned into the pBSK-SP6-globin vector specially designed for Xenopus oocyte expression (5). cRNA were synthetized from the NotI-digested vector using the mCAP RNA capping kit from Stratagene.

RESULTS
Effect of Zn 2ϩ on the Heteromeric ASIC1aϩ2a Current-The extensive colocalization of the ASIC1a and ASIC2a subunits in central neurons (3,15) led us to first study the effect of Zn 2ϩ on the heteromeric ASIC1aϩ2a channel expressed in Xenopus oocytes. The properties of the current recorded after coinjection of equal quantities of both ASIC1a and ASIC2a cRNA were in good agreement with those already described for a heteromeric ASIC1aϩ2a current (15). However, in the presence of 300 M Zn 2ϩ , the submaximal ASIC current induced by pH 6.3 was increased ϳ3-fold (Fig. 1A, b and c), whereas Zn 2ϩ alone at pH 7.4 could not activate any current (Fig. 1A, a). The same Zn 2ϩpotentiated ASIC current could be recorded when Zn 2ϩ was applied either before the acidic pH jump (tested up to 3 min; Fig. 1A, d) or simultaneously (Fig. 1A, c). When the ASIC current was fully inactivated (after 5 s at pH 6.3; Fig. 1A, e), it could not be reactivated by the addition of Zn 2ϩ . However, application of Zn 2ϩ in the course of the inactivation process was still able to increase the fraction of the ASIC current that was not inactivated (Fig. 1A, f). Based on these results, Zn 2ϩ was applied simultaneously with the pH drop in all subsequent experiments.
Zn 2ϩ (300 M) induced a left shift of the ASIC1aϩ2a halfmaximal activation from pH 5.5 to pH 6, and an increase in the Hill coefficient from 1.05 to 1.4 (Fig. 1B, typical experiment; Fig. 2A, mean results), whereas the maximal current induced at pH 4 was not modified. Fig. 2B shows the mean factor by which the ASIC1aϩ2a current amplitude was increased by 300 M Zn 2ϩ at different pH values. The potentiation by Zn 2ϩ depended on the extracellular pH, with greater effects between pH 6.9 and 6.3 than between pH 6 and pH 5. The relative Zn 2ϩ -induced increases in current amplitude reached 4.00 Ϯ 1.06-fold (n ϭ 6) the control current, for pH 6.6. Zn 2ϩ has no effect on the pH-dependent ASIC1aϩ2a current inactivation (Fig. 2C). Zn 2ϩ concentrations between 1 M and 10 mM were tested on ASIC1aϩ2a currents activated at pH 6.3 (Fig. 2D). The Zn 2ϩ concentration producing the half-maximal increase in current amplitude was 111 M with a Hill coefficient of 0.8. A Zn 2ϩ concentration of 300 M was used in all further experiments.
Effect of Zn 2ϩ on Homomeric ASIC1a and ASIC2a Currents-To determine which ASIC subunit was responsible for the Zn 2ϩ potentiation of the heteromeric ASIC1aϩ2a current, we have tested the effects of Zn 2ϩ on homomeric ASIC1a and ASIC2a currents expressed in Xenopus oocytes. ASIC2a requires rather low extracellular pH for activation (pH 0.5 ϭ 4.34, Fig. 3A) (6,15). When applied simultaneously with acid, 300 M Zn 2ϩ increased the amplitude of the ASIC2a current only in the first half of the activation curve, between pH 6.3 and pH 4.5 ( Fig. 3A). Even if the effect of Zn 2ϩ on the ASIC2a activation curve seems moderate, the relative increases in current amplitude reached 1.21 Ϯ 0.04-fold (n ϭ 11) to 7.00 Ϯ 1.89-fold (n ϭ 4) the control current, for pH 4.5 and 6.3, respectively (Fig. 3B). Zn 2ϩ coactivated the homomeric ASIC2a current induced by pH 5.7 with an EC 50 of 120 M and a Hill slope factor of 1.44 (Fig. 3B, inset).
In contrast to ASIC2a, ASIC1a current is not potentiated by Zn 2ϩ (Fig. 3C). Zn 2ϩ instead exerted a very small inhibition on the ASIC1a current, but this effect was not significant. Fig. 3D illustrates that the effects of 300 M Zn 2ϩ on the ASIC1a, ASIC2a, and ASIC1aϩ2a currents induced by pH 6 and 5 are similar after expression in COS cells (transfection) or Xenopus oocytes (cRNA injection).
Effect of Zn 2ϩ on Heteromeric ASIC2aϩ3 and Homomeric ASIC3 Currents-Since the Zn 2ϩ -induced potentiation of ASIC currents seemed to be linked to the presence of the ASIC2a subunit, we tested the effect of Zn 2ϩ on the heteromeric ASIC2aϩ3 channel expressed in Xenopus oocytes. Such an association has been characterized by Babinski et al. (17). The ASIC3 current was not potentiated but rather slightly inhibited by 300 M Zn 2ϩ (Fig. 4A). However, when ASIC2a was coexpressed with ASIC3, Zn 2ϩ potentiated both the peak and the plateau phases of the current (Fig. 4B). Fig. 4C illustrates the relative Zn 2ϩ -induced increase of ASIC3, ASIC2aϩ3, and ASIC1aϩ2a currents activated at pH 6, 5, and 4. The potentiation of the ASIC2aϩ3 current by Zn 2ϩ was similar to that of the ASIC1aϩ2a current. Zn 2ϩ did not reactivate the ASIC current when fully inactivated by pH 6.3 (e) but increased the not fully inactivated current (f). B, On the same oocyte, Zn 2ϩ (300 M) induced a left shift of the current activation curve associated with an increase in the Hill coefficient (E, control; q, ϩ Zn 2ϩ ). The current amplitude is expressed as a fraction of the current induced by pH 5 (I/I pH 5 ). Holding potential, Ϫ70 mV.

Involvement of Extracellular Histidines in the Coactivation of
ASIC2a by Zn 2ϩ -ASIC2a appears as the major subunit conferring Zn 2ϩ sensitivity to ASICs. We used site-directed mutagenesis to investigate the structural determinants of Zn 2ϩ coactivation in the extracellular loop of the ASIC2a subunit. Amino acids found at Zn 2ϩ binding sites in proteins include histidine, cysteine, and occasionally aspartate or glutamate (40). When charged water-soluble sulfhydryl reactive reagents like MTSET, a cationic methanethiosulfonate, were applied extracellularly, they had no effect on the ASIC2a current, making unlikely that a cysteine residue interact with extracellularly applied Zn 2ϩ (data not shown). On the other hand, diethylpyrocarbonate that reacts with several amino acid side chains, including the imidazole group side of histidine residues (41), suppressed the Zn 2ϩ activation of the ASIC2a current (Fig. 5B) and of the ASIC1aϩ2a current (data not shown), suggesting a possible involvement of histidine residues in the Zn 2ϩ coactivating effect. We systematically replaced the 10 histidine residues in the extracellular loop of ASIC2a by alanine and checked the Zn 2ϩ sensitivity of the mutants (Table I). All mutants were still functional with no significant modifications of their pH 0.5 except the H72A mutant that had lost its ability to be activated by acidic pH down to pH 3. The expected potentiation of the ASIC2a current by zinc (up to 2 times at pH 5.5) was observed for the wild-type channel and for several mutants, but two of them, H162A and H339A, displayed no apparent potentiation, with little if any modification of their properties (Table I and Fig. 5, C and D). Each of these mutants lacked the potentiating effect of 300 M Zn 2ϩ , which could be as high as an 8-fold increase compared with the wild-type current amplitude at pH 6 ( Fig. 5E). The absence of effect is not due to a shift in the pH dependence since the pH 0.5 of the two mutants (H162A, pH 0.5 ϭ 4.71; H339A, pH 0.5 ϭ 4.74) was not significantly modified compared with the wild-type channel (pH 0.5 ϭ 4.72) ( Table I).
Effect of Zn 2ϩ on Heteromeric ASIC Currents Involving Mu-tated ASIC2a Subunits-We coexpressed wild type ASIC2a or H162A and H339A mutants with ASIC1a. The pH sensitivity of the ASIC1aϩ2a heteromeric currents was similar whether the ASIC2a subunit was mutated or not (Fig. 6B). Surprisingly, the ASIC1aϩASIC2a H162A current remained highly sensitive to zinc, whereas the ASIC1aϩASIC2a H339A current was practically insensitive (Fig. 6A). Even if the pH dependence found for the heteromeric ASIC1aϩASIC2a H339A current (pH 0.5 ϭ 5.7, Fig. 6B) was intermediate between that of homomeric ASIC1a (pH 0.5 ϭ 6.4) and ASIC2a (pH 0.5 ϭ 4.7), this did not completely exclude a mixture of the two different homomeric currents. To confirm the heteromeric association of the ASIC1a and ASIC2a H339A subunits, we have used the G430V mutant of ASIC1a, which displays a low basal amiloride-sensitive current and is not activated by acidification (15). Co-expression of this mutant with ASIC2a H339A induced a proton-activated current with a pH 0.5 significantly different from that of the ASIC2a H339A channel alone, indicating that a heteromultimeric channel was indeed formed by association of both subunits (Fig. 6C). This confirmed that the ASIC2a H339A mutation was responsible for the lack of Zn 2ϩ coactivation on the heteromeric channel formed with ASIC1a. We coexpressed wild type or H162A and H339A mutants of ASIC2a with ASIC3. The pH sensitivity of the heteromeric ASIC2aϩ3 current was similar whether the ASIC2a subunit was mutated or not (Fig. 6E). When coexpressed with ASIC3, the H162A and the H339A mutated forms of ASIC2a decreased the effect of Zn 2ϩ on both transient and sustained current. The decrease in zinc effect was more pronounced for the H339A mutation than for the H162A mutation (Fig. 6D), as observed previously with the ASIC1aϩ2a current.
Involvement of Extracellular Histidines in the Activation of ASIC2a by Protons-The ASIC2a H72A mutant is not activated by increasing the external H ϩ concentration (Table I). This can reflect a loss of the capability of the channel to sense extracellular pH. However, this could also be due to a disrup- The current amplitude is expressed as a fraction of the current induced by pH 5 (I/I pH 5 ). Zn 2ϩ (300 M) induced a left shift of the mean activation curve, the pH 0.5 being shifted to 6.0 (q) from a control value at 5.5 (E), associated with an increase of the Hill coefficient reaching 1.4 from the control value of 1.05. B, relative effect of 300 M Zn 2ϩ for each pH value. The ratio between the current amplitude in presence and in absence of Zn 2ϩ (I Zn 2ϩ /I control ) was calculated for each pH value and plotted as mean Ϯ S.E. (n ϭ 4 -11). *, p Ͻ 0.05; **, p Ͻ 0.005, significantly different from pH 4 ratio. C, effect of Zn 2ϩ on the pH-dependent inactivation of heteromeric ASIC1aϩ2a current. The current was induced by pH 4 from a resting pH value ranging from 8 to 5, normalized to the current induced from pH 7.4 (I/I pH 7.4 ), and expressed as mean Ϯ S.E. as a function of resting pH value (n ϭ 3-21). E, control; q, ϩ300 M Zn 2ϩ . D, dose-response curve of Zn 2ϩ potentiation of ASIC1aϩ2a current. Currents were activated by pH 6.3 in absence and presence of Zn 2ϩ concentrations ranging from 1 M to 10 mM. The current amplitude increase (␦I) was normalized to its maximal value (␦I/ ␦I max ) and plotted as mean Ϯ S.E. as a function of Zn 2ϩ concentration (n ϭ 3-12). Inset, original current traces showing the effect of 100 M and 300 M Zn 2ϩ . Holding potential, Ϫ70 mV. tion of the necessary association with other subunits to form a homomeric channel. Analysis of the pH sensor of ASICs is complicated by the absence of other modes of activation, such as capsaicin and/or temperature for the VR1 channel (42). To analyze in more detail the effect of the H72A mutation on the ASIC2a properties, we have used a previously described gainof-function mutant of ASIC2a that displays a medium to high amiloride-sensitive basal current and retains the property to be activated by external protons (5,6). This mutant corresponds to the change of one amino acid, Gly-430, located in a region preceding the second transmembrane domain of ASIC2a. This residue plays a key role in the gating of the channel (6, 43). The G430T mutant has a large basal current and a pH 0.5 of 6.6 for protons activation (Fig. 7B and Ref. 6). The H72A mutation has been introduced in this mutant, and the properties of the double mutant have been analyzed. Interestingly, the double mutant was functional (Fig. 7A), demonstrating that the H72A mutation is not a loss-of-function mutation and that subunit association was not disrupted. Like the single gain-of-function mutant (6), the double mutant showed a large constitutive current, which remained activated by low pH with no apparent inactivation (Fig. 7A), but the pH dependence was largely modified, shifted toward more acidic pH by almost 2 orders of magnitude (pH 0.5 ϭ 4.9; Fig. 7B). This drastic modification of the pH sensitivity of ASIC2a by the H72A mutation could explain the lack of activation of the single H72A mutant because shifting the pH 0.5 of the wild-type ASIC2a channel (pH 0.5 ϭ 4.7) to more acidic values by several pH units would lead to a channel virtually insensitive to proton activation.

DISCUSSION
Coactivation of ASIC Currents by Zn 2ϩ -Zn 2ϩ is known to exert a variety of inhibitory effects on ion channels, but stimulatory effects are rare. For instance, potentiation by Zn 2ϩ has been described for the activation of purinergic P2X 2 , P2X 3 , and P2X 4 channels by ATP (44 -46).
We show that both homomeric ASIC2a channels and heteromeric ASIC2a-containing channels are potentiated when Zn 2ϩ

TABLE I
Alanine substitution of ASIC2a extracellular histidines and effect on zinc coactivation activity Schematic representation of the ASIC2a subunit showing the mutated histidines in the extracellular loop and the two membrane-associated domains (M1 and M2) is shown on the left. The pH for halfmaximal activation (pH 0.5 ) and the effect of co-application of Zn 2ϩ with acidic pH (pH 5.5) calculated as the ratio between the current amplitude in presence and in absence of 300 M Zn 2ϩ (I Zn 2ϩ /I Ctr ) were determined for each mutant at a holding potential of Ϫ70 mV. Means Ϯ S.E. are shown (n ϭ 3-10). **, p Ͻ 0.005, significantly different from wild type ASIC2a (unpaired t test). NA, nonactivated by acidic pH.  (unpaired t test). B, pH dependence of heteromeric channels containing WT or mutated ASIC2a subunits. The pH 0.5 of channels containing WT (q), H162A (f), and H339A (OE) ASIC2a subunits are 5.8, 6.0, and 5.7, respectively. Current amplitude was expressed as a fraction of the current induced by pH 3.0 (I/I max ). Each point corresponds to mean Ϯ S.E. of 8 -12 oocytes. C, pH dependence of homomeric ASIC2a H339A (OE), ASIC1a G430V (q), and heteromeric ASIC2a H339AϩASIC1a G430V (f) channels was established. Each point represents the mean Ϯ S.E. of 8 -12 oocytes. Currents were recorded at a holding potential of Ϫ70 mV. The ASIC1a G430V mutant is not activated by increasing the H ϩ concentration, as described previously (15). The pH 0.5 of the heteromeric channel is largely shifted toward more alkaline pH compared with the ASIC2a H339A mutant alone, demonstrating an association of this mutant with ASIC1a. D, Zn 2ϩ effect on heteromeric channels containing ASIC3 and the mentioned ASIC2a subunit at different pH. Data correspond to mean Ϯ S.E. (n ϭ 6 -14). *, p Ͻ 0.05; **, p Ͻ 0.005, significantly different from WT channel (unpaired t test). E, the H162A and H339A mutations in ASIC2a were without effect on the pH dependence of the ASIC2aϩ3 peak current. pH 0.5 values were, respectively, 4.3, 4.1, and 4.3 for channels containing WT (f), H162A (OE), and H339A (q) ASIC2a subunits. Each point corresponds to mean Ϯ S.E. of 6 -9 oocytes. Current amplitude was expressed as a fraction of the current induced by pH 3 (I/I max ). Holding potential for A-E, Ϫ70 mV.
is coapplied with acidic pH. Zn 2ϩ alone could not activate the ASIC1aϩ2a current, in good agreement with the results obtained by Adams et al. (47) on ASIC2a current. The potentiating effect of Zn 2ϩ on either homomeric or heteromeric ASIC2acontaining channels only appeared between pH 6.9 and pH 5, independent of the pH dependence of the current, suggesting the involvement of titrable residues that would less efficiently chelate Zn 2ϩ as pH decreases.
Physiological Significance of ASIC Current Coactivation by Zn 2ϩ -The Zn 2ϩ concentrations that potentiate the ASIC2acontaining channels (EC 50 ϭ 111 M for ASIC1aϩ2a current and EC 50 ϭ 120 M for ASIC2a) are compatible with the physiological concentration of synaptically released Zn 2ϩ (100 -300 M) (31)(32)(33)(34)(35). Our results suggest that ASICs might be a physiological target for Zn 2ϩ . Some ASICs that are expressed in the CNS (e.g. ASIC1aϩ2a, pH 0.5 ϭ 5.5, Fig. 2A) require rather acidic pH for activation. The coactivation by Zn 2ϩ shifts the pH dependence of the ASIC1aϩ2a channel closer to physiological pH values and strongly augments the activity of the heteromeric ASIC1aϩ2a and the homomeric ASIC2a channel at pH values just below neutral. The native ASIC-like current of rat primary cultured hippocampal neurons are indeed also coactivated by Zn 2ϩ (data not shown). 2 The potentiation by Zn 2ϩ of the hippocampal ASIC current could participate in the modulation of neuronal excitability by increasing the membrane depolarization induced by small pH changes.
Extracellular Histidines of ASIC2a Are Involved in Zn 2ϩ Effect-The effect of zinc is immediate and does not necessitate any pre-application, suggesting a direct interaction with a zinc-binding site located in the extracellular domain of the channel. The side chain of histidine has been involved in the zincbinding sites of numerous metalloproteins (40). Alanine substitution of each of the 10 histidine residues present in the extracellular loop of ASIC2a demonstrates that His-162 and His-339 are involved in the Zn 2ϩ effect. Replacement of His-162 or His-339 completely abolished the sensitivity of ASIC2a to 300 M Zn 2ϩ . The simplest interpretation of our results is that His-162 and His-339 are part of the Zn 2ϩ binding site on ASIC2a. However, it should be kept in mind that the mutagenesis alone does not provide a definite demonstration of such direct participation. All ASICs containing the ASIC2a subunit, i.e. homomeric ASIC2a and heteromeric ASIC1aϩ2a or ASIC2aϩ3, are sensitive to Zn 2ϩ . The data presented here using co-expression of mutated ASIC2a subunits with ASIC1a or ASIC3 definitely demonstrate that Zn 2ϩ sensitivity in heteromeric channels is carried by the ASIC2a subunit. A comparison between sequences of ASIC1a, ASIC2a, and ASIC3 shows that, although His-162 is highly conserved in all these subunits, His-339 is specific of ASIC2a (Fig. 8). Because ASIC2a is the main subunit responsible for the Zn 2ϩ sensitivity of ASICs, it is then tempting to assign to His-339 a predominant role in this property. However, the selective replacement of His-162 by alanine can also completely abolish the Zn 2ϩ sensitivity of the homomeric ASIC2a channel, whereas it has moderate or no effect on heteromeric channels. This could reflect some difference between the Zn 2ϩ binding sites of homomeric and heteromeric channels comprising the ASIC2a subunit.
His-72 Is Involved in the pH Sensor of ASIC2a-Titrable histidine residues are major determinants of pH modulation in many ion channels (48,49), whereas glutamate residues have also been shown to be important for acid sensing as for the capsaicin receptor VR1 (42). The His-72 residue adjacent to the first transmembrane domain of ASIC2a (Table I) drastically changes its pH sensitivity since the ASIC2a H72A channel has become insensitive to acidic pH. However, it is not the unique determinant of the pH sensitivity because the ASIC2a G430TϩH72A double mutant remains activable by low pH (Fig. 7). Other positions have been shown previously to be involved in the pH dependence of ASIC2a like the Gly-430 residue situated just before the second transmembrane segment (6) and the region preceding the first transmembrane domain (50). The His-72 residue of ASIC2a is conserved in ASIC1a and ASIC3 (Fig. 8). These positions have been mutated in both ASIC1a and ASIC3, but in that case no modification of the pH dependence of the mutant channels was observed, with pH 0.5 ϭ 6.5 and 6.1 for the ASIC1a H73A and ASIC3 H73A peak currents, respectively, compared with pH 0.5 values of 6.4 for ASIC1a and 6.3 for ASIC3 (data not shown). There is a precedent for such a situation with two-P domain K ϩ channels (48,51); His-98 has been shown to be important for the pH dependence of TASK-3 near physiological pH, but TWIK-1, which also contains a histidine residue at the equivalent position, does not display a pH dependence in the same pH range.
In summary, the present study has revealed interesting structural features of acid-sensing ion channels, with special 2 A. Baron, R. Waldmann, and M. Lazdunski, manuscript in preparation.
FIG. 7. The H72A mutation modifies the pH dependence of homomeric ASIC2a current. A, the G430T mutation is associated with a basal amiloride-sensitive current (5), but the mutated channel is still activated by acidic pH with modified pH sensitivity compared with wild-type channel (6). Introduction of the H72A mutation in the G430T ASIC2a mutant does not eliminate the basal activity and the activation by acidic pH. The H72A mutation is therefore not loss-of-function, but it drastically alters the pH dependence of ASIC2a. The bar above the current recordings represents the pH pulse, value being indicated for each trace. The zero current base line is indicated by a dotted line. B, effect of the ASIC2a H72A mutation on the pH dependence of the ASIC2a G430T gain-of-function current. The pH 0.5 of the G430T mutant is 6.6 (OE). The double mutant H72AϩG430T has a pH 0.5 of 4.9 (f), although the H72A simple mutant is not activated by pH down to pH 3.0. Current amplitude was expressed as a fraction of the current induced by pH 3.0 (I/I max ). Each point represents the mean Ϯ S.E. of 5-8 oocytes. Holding potential for A and B, Ϫ70 mV.
FIG. 8. Alignments of rat acid-sensing ion channels in the regions surrounding the histidine residues important for Zn 2؉ regulation and pH sensing. Asterisks indicate the position of His-72, His-162, and His-339 in ASIC2a. Identical and similar residues are printed white on black/dark gray and black on gray, respectively. The Clustal program was used to generate the alignments. Accession numbers for rat ASIC1a, ASIC2a, and ASIC3 are U94403, U53211, and AF013598, respectively. emphasis on ASIC2a, which is coactivated both by external Zn 2ϩ and extracellular acidification. Several histidines have been identified as candidate electron donors to the Zn 2ϩ coordination site, and another histidine residue is a candidate for the pH sensor. However, there are clearly other residues directly or indirectly involved in pH sensing that will also need to be identified.