Probing the Structural Basis of Zn2+ Regulation of the Epithelial Na+ Channel*

Background: Extracellular Zn2+ regulates epithelial Na+ channel (ENaC) activity. Results: Specific mutations selectively weaken either the stimulatory or inhibitory effect of Zn2+. Conclusion: External Zn2+ regulates ENaC by interacting with multiple extracellular sites within the γ-subunit. Significance: This report provides novel insights into the structural basis of Zn2+ regulation of ENaC. Extracellular Zn2+ activates the epithelial Na+ channel (ENaC) by relieving Na+ self-inhibition. However, a biphasic Zn2+ dose response was observed, suggesting that Zn2+ has dual effects on the channel (i.e. activating and inhibitory). To investigate the structural basis for this biphasic effect of Zn2+, we examined the effects of mutating the 10 extracellular His residues of mouse γENaC. Four mutations within the finger subdomain (γH193A, γH200A, γH202A, and γH239A) significantly reduced the maximal Zn2+ activation of the channel. Whereas γH193A, γH200A, and γH202A reduced the apparent affinity of the Zn2+ activating site, γH239A diminished Na+ self-inhibition and thus concealed the activating effects of Zn2+. Mutation of a His residue within the palm subdomain (γH88A) abolished the low-affinity Zn2+ inhibitory effect. Based on structural homology with acid-sensing ion channel 1, γAsp516 was predicted to be in close proximity to γHis88. Ala substitution of the residue (γD516A) blunted the inhibitory effect of Zn2+. Our results suggest that external Zn2+ regulates ENaC activity by binding to multiple extracellular sites within the γ-subunit, including (i) a high-affinity stimulatory site within the finger subdomain involving His193, His200, and His202 and (ii) a low-affinity Zn2+ inhibitory site within the palm subdomain that includes His88 and Asp516.

Sodium transport mediated by the epithelial Na ϩ channel (ENaC) 2 plays an important role in the maintenance of extracellular fluid volume and regulation of blood pressure and in the regulation of airway surface liquid volume. Altered channel activity induced by mutations of ENaC genes results in several disorders, including Liddle syndrome and pseudohypoaldosteronism type 1, and contributes to the pathogenesis of cystic fibrosis (1)(2)(3)(4). Specific ENaC variants are suggested to contribute to the genesis of essential hypertension in selected populations (5)(6)(7). ENaC activity is regulated by a variety of extracellular factors, including proteases, shear stress, pH, anions, nucleotides, and transition metals (8 -15). Although some metals are essential for normal physiological function, excessive environmental exposure can be toxic (16,17). For instance, particulate matter and airborne particles containing high amounts of transitional metals, including zinc and copper, contribute to pulmonary and cardiovascular toxicity (18,19). Although recent studies have suggested that ion channels are potential molecular targets of transition metals (20), the mechanisms that confer the harmful effects of heavy metals are poorly understood.
Zinc is the second most abundant transition metal in living organisms and is thought to complex with ϳ10% of the human proteome (21). It has catalytic, structural, or regulatory roles in proteins that are involved in diverse biological processes (22). For example, zinc regulates voltage-and ligand-gated ion channels and may function as a signaling ion in brain (23,24). Zinc may play a pathophysiological role in several disorders such as Alzheimer disease, cancer, diabetes, and depression (22).
Previous studies have demonstrated that low concentrations of extracellular Zn 2ϩ activate ENaC. Zn 2ϩ increases short-circuit current in amphibian epithelia expressing native Na ϩ channels (8). Zn 2ϩ activates ENaCs in heterologous expression systems by directly interacting with the channel and altering channel gating (25). Single channel recordings revealed that external Zn 2ϩ increased the number of observed channels within a patch without changing single channel conductance (10). The response of ENaC to extracellular Zn 2ϩ is biphasic. Low concentrations of Zn 2ϩ activate the channel, with a maximal response at 100 M Zn 2ϩ . Further increases in [Zn 2ϩ ] reduce channel activity. This bell-shaped dose-response relationship suggests that Zn 2ϩ enhances channel activity at low concentrations and is inhibitory at high concentrations (25). The stimulatory effect of Zn 2ϩ on ENaC has been attributed to relief of Na ϩ self-inhibition (25). Na ϩ self-inhibition is a rapid down-regulation of ENaC open probability when the extracellular Na ϩ concentration increases (9,26). However, the Zn 2ϩbinding sites within ENaC have not been identified, and the biphasic dose response to the heavy metal has not been explained.
ENaCs are typically composed of three homologous subunits, each of which contains short intracellular N and C ter-mini, two transmembrane domains, and a large extracellular region. The crystal structure of the chicken acid-sensing ion channel 1 (cASIC1; homologous to ENaC) reveals a trimeric channel complex with a large extracellular domain (ECD) connected to the transmembrane domain via a wrist region (27). The extracellular region of a subunit is organized into five subdomains, referred to as palm, ␤-ball, knuckle, finger, and thumb. Recent studies suggest that the extracellular regions of ENaC subunits have a similar overall design to that of cASIC1 and that the ECD may sense various extracellular signals (12, 14, 28 -31). However, the exact locations where most extracellular factors interact with ENaC and the mechanistic details of how the initial contacts lead to altered channel activity remain elusive.
The ECDs of ENaC subunits contain numerous residues, such as His, Cys, Glu, and Asp, which are capable of serving as Zn 2ϩ -binding ligands (32). We hypothesized that some of these residues mediate Zn 2ϩ interaction with ENaC. Several lines of evidence suggest that the ␥-subunit has a particularly important role in regulating channel gating in response to external factors (11,33,34). Mutations were introduced into His residues within the ␥-subunit ECD to identify sites that have a role in the regulation of ENaC by Zn 2ϩ . Three His residues were identified within the finger subdomain that participate in a high-affinity Zn 2ϩ activating site. Additionally, mutation of a palm subdomain His residue altered the low-affinity inhibitory response. These His residues are unique to the ␥-subunit, as they are not found at the homologous sites in the ␣and ␤-subunits.
All experiments were carried out at room temperature (20 -24°C). Two-electrode voltage clamp was performed using either an Axoclamp 900A amplifier (Molecular Devices, Sunnyvale, CA) or a TEV-200 voltage clamp amplifier (Dagan Corp., Minneapolis, MN) and a DigiData 1440A interface (Molecular Devices). Data acquisition and analyses were done with pCLAMP 8 or 9 (Molecular Devices). Oocytes were placed in a recording chamber from Warner Instruments (Hamden, CT) and perfused with bath solutions at a constant flow rate of 12-15 ml/min. Oocytes were continuously clamped at Ϫ60 or Ϫ100 mV as specified in the figure legends. Na ϩ Self-inhibition Response-Na ϩ self-inhibition responses were examined as described previously (36,37). Briefly, Na ϩ self-inhibition was determined by measuring the decrease in current from the peak (I peak ) to a steady state (I ss ) elicited by a rapid increase in extracellular Na ϩ concentration from 1 to 110 mM. The [Na ϩ ] jump was done by rapidly replacing a 1 mM Na ϩ bath solution (NaCl-1 containing 1 mM NaCl, 109 mM N-methyl-D-glucamine, 2 mM KCl, 2 mM CaCl 2 , and 10 mM HEPES, pH 7.4) with a 110 mM Na ϩ bath solution (NaCl-110 containing 110 mM NaCl, 2 mM KCl, 2 mM CaCl 2 , and 10 mM HEPES, pH 7.4). Rapid solution exchange was performed with a 16-channel Teflon valve perfusion system (AutoMate Scientific, Inc., Berkeley, CA). The magnitude of Na ϩ self-inhibition was represented by the ratio of amiloride-sensitive I ss and I peak .
Zn 2ϩ Dose Responses-A 1 M ZnCl 2 stock solution (super pure, Ͼ99.999%, Sigma-Aldrich) was prepared in water and diluted to the desired concentrations in NaCl-110 bath solution. To examine Zn 2ϩ dose responses, oocytes expressing mENaCs were continuously clamped at Ϫ60 mV while Zn 2ϩ was applied at increasing concentrations in the range of 0.1 M to 5 mM. Due to the limited solubility of ZnCl 2 in NaCl-110 bath solution, Zn 2ϩ concentrations Ͼ5 mM were not tested. At the end of an experiment, 10 M amiloride was added to the bath to determine the amiloride-sensitive current. The changes in amiloride-sensitive currents in response to Zn 2ϩ applications were used to analyze the dose-response relationship.
Because the amiloride-sensitive current demonstrated a biphasic dose-response relationship with [Zn 2ϩ ], we assumed that ENaCs possess a high-affinity activating site and a lowaffinity inhibitory site for Zn 2ϩ . Thus, the normalized currents in the presence of increasing concentrations of Zn 2ϩ were fitted with Equation 1, where I R is the relative current at certain [Zn 2ϩ ] (i.e. current in the presence of Zn 2ϩ normalized to current in the absence of Zn 2ϩ ); I max and I min are the maximal and minimal normalized currents, respectively; C is the Zn 2ϩ concentration; and EC 50 and IC 50 are the concentrations with 50% effects for the activating and inhibitory phases, respectively.
Statistical Analysis-Data are presented as means Ϯ S.E. To minimize bias from batch-to-batch variation in oocytes, the significance in the differences between mutant and WT channels was always analyzed from data obtained in the same batches of oocytes. Student's t tests were used for comparison, and the significance levels are specified below and in the figure legends. Curve fittings were performed with OriginPro 8.5 (OriginLab, Northampton, MA).
The Na ϩ self-inhibition response reflects a decrease in current from a peak (I peak ) to a steady state (I ss ), which is elicited by a rapid increase in the extracellular Na ϩ concentration from 1 to 110 mM (Fig. 1). As currents approached a steady-state level, increasing concentrations of Zn 2ϩ were applied, and changes in current were measured. Increasing extracellular [Zn 2ϩ ] in the range of ϳ0.1-100 M led to a dose-dependent increase in current. The current fell when extracellular [Zn 2ϩ ] was increased above 100 M. The maximal current in the presence of Zn 2ϩ (I znm ) was essentially identical to the I peak observed following the increase in extracellular [Na ϩ ] (I znm /I peak ϭ 1.00 Ϯ 0.01, n ϭ 32), consistent with the notion that Zn 2ϩ relieves ENaC from inhibition by external Na ϩ . Fig. 1B illustrates the Zn 2ϩ dose-response relationship. Mean amiloride-sensitive currents in the presence of increasing extracellular [Zn 2ϩ ] were normalized to the base-line current. Similar to our previous report (25), we observed a bell-shaped dose response to Zn 2ϩ . This relationship fits a model of ENaC possessing a high-affinity Zn 2ϩ activating site and a low-affinity Zn 2ϩ inhibitory site (R 2 ϭ 0.997, Equation 1, see "Experimental Procedures") with an estimated EC 50 and IC 50 of 2.1 Ϯ 0.1 M (n ϭ 6) and 2.1 Ϯ 0.1 mM (n ϭ 6), respectively. The maximal relative current observed at 100 M Zn 2ϩ was 1.68 Ϯ 0.07 (n ϭ 6).
Ala Substitutions of Selected His Residues in the Extracellular Region of ␥mENaC Alter the Effect of Zn 2ϩ -There are a total of 10 His residues within the ECD of the ␥-subunit. Four residues are located within the finger subdomain, three in the ␤-ball subdomain, two in the palm subdomain, and one in the thumb subdomain (Fig. 2). We individually mutated these His residues to Ala and expressed the mutant ␥-subunits together with WT ␣and ␤-subunits in oocytes. We sequentially examined Na ϩ self-inhibition and the response to Zn 2ϩ (Figs. 3 and 4) to determine whether a mutation selectively disrupted the effect of Zn 2ϩ on channel activity or secondarily altered the Zn 2ϩ effect by modifying the Na ϩ self-inhibition response.
Ala substitutions at 4 of the 10 extracellular ␥-subunit His residues (␥H193A, ␥H200A, ␥H202A, and ␥H239A) significantly reduced maximal ENaC activation by Zn 2ϩ (Figs. 3 and 4). Interestingly, all four of the extracellular His residues whose mutations reduced Zn 2ϩ activation are within the predicted finger subdomain of the channel. In addition to suppressing Zn 2ϩ activation, ␥H239A also diminished Na ϩ self-inhibition, similar to what was previously observed when ␥His 239 was substituted with other residues (37,38). As the ␥H239A mutant was relieved of Na ϩ self-inhibition prior to the addition of Zn 2ϩ , it is not possible to determine whether this residue is involved in Zn 2ϩ binding to the channel. On the other hand, ␥H193A, ␥H200A, and ␥H202A did not significantly alter the Na ϩ self-inhibition response, suggesting that these three His residues in the finger subdomain have specific roles in Zn 2ϩ activation of ENaC.
Although ␥H88A did not affect the maximal Zn 2ϩ activation (p Ͼ 0.05 versus WT) or the Na ϩ self-inhibition response, it converted the Zn 2ϩ dose response from bell-shaped to monophasic by selectively eliminating the inhibitory effect of high [Zn 2ϩ ] (Figs. 3 and 4). These results are consistent with a mutation-induced loss of a low-affinity Zn 2ϩ -binding site that mediates channel inhibition at high [Zn 2ϩ ]. One mutation (␥H283A) moderately increased the maximal Zn 2ϩ activation (p Ͻ 0.01 versus WT) (Fig. 4). Collectively, these observations suggest that ␥His 88 within the palm subdomain and ␥His 193 ,     . Mutation-induced changes in the magnitude of Na ؉ self-inhibition and the maximal Zn 2؉ activation of mENaC. I ss /I peak values are inversely proportional to the magnitude of Na ϩ self-inhibition. I znm values represent the maximal relative currents in the presence of Zn 2ϩ . They were typically observed with 100 M Zn 2ϩ except for the ␥H88A mutant, which had I znm at 1 mM. Note that I znm is the observed value and different from the I max ( Table 1) that was derived from a best fit of the dose-response data. Black bars indicate that the values of mutants were significantly different (p Ͻ 0.01) from those of WT (n ϭ 5-10), obtained in the same batch of oocytes. The dashed lines show the mean values pooled from all oocytes expressing WT mENaC obtained in the analyses of these mutants (I ss /I peak ϭ 0.60 Ϯ 0.01, I znm ϭ 1.73 Ϯ 0.03, n ϭ 32).  (Table 1).
although it essentially eliminated the decrease in channel activity observed with higher [Zn 2ϩ ]. In contrast, ␥H193A, ␥H200A, and ␥H202A reduced the sensitivity of mENaC to activation by extracellular Zn 2ϩ and did not affect the inhibitory effect of high [Zn 2ϩ ], leading to increased EC 50 values and no change in the IC 50 relative to WT. The results imply that ␥H193A, ␥H200A, and ␥H202A selectively reduce the apparent affinity of the potentiating site for Zn 2ϩ while not interfering with the low-affinity Zn 2ϩ inhibitory site.
We also examined whether substitutions at multiple His residues led to further increases in the EC 50 for Zn 2ϩ . The Zn 2ϩ dose-response relationships of ␥H200A/␥H202A and ␥H193A/ ␥H200A/␥H202A were similar to those found in single mutants (Fig. 6). Thus, all three His residues within the finger subdomain are required for Zn 2ϩ -dependent channel activation. The lack of an additive effect of multiple His substitutions suggests that these His residues participate in a common Zn 2ϩbinding site.
In addition to increasing the EC 50 values of Zn 2ϩ activation, ␥H193A, ␥H200A, ␥H202A, ␥H200A/␥H202A, and ␥H193A/ ␥H200A/␥H202A also reduced the magnitude of the increase in current in response to Zn 2ϩ (I max ) ( Table 1). The reduced I max observed in the mutants could be explained either by a reduction in the efficacy of channel activation by Zn 2ϩ or by the presence of an inhibitory effect that limits the extent of channel activation. To distinguish these two possibilities, we examined channels with His substitutions at Zn 2ϩ activating sites (␥H200A/␥H202A) and the Zn 2ϩ inhibitory site (␥H88A). As shown in Fig. 6E, ␥H88A/␥H200A/␥H202A reduced Zn 2ϩ sensitivity by 8-fold compared with ␥H88A channels (EC 50 ϭ 20.4 Ϯ 3.0 M (n ϭ 6) versus 2.6 Ϯ 0.3 M (n ϭ 10); p Ͻ 0.001). However, both mutants had similar I max values (1.81 Ϯ 0.11 (n ϭ 6) versus 1.79 Ϯ 0.09 (n ϭ 10); p Ͼ 0.05), demonstrating that mutations of the Zn 2ϩ activating sites within the finger subdomain do not alter the magnitude of the maximal Zn 2ϩ activation when the low-affinity Zn 2ϩ inhibitory site is not  Table 1.

TABLE 1 Fitting parameters for Zn 2؉ dose responses
Parameters were from best fits of the dose-response data by nonlinear curve fitting, as described under "Experimental Procedures." Values are means Ϯ S.E., averaged from individual oocytes expressing either WT or mutant channels. Parameters for WT channels were pooled from nine batches of oocytes that were used in the experiments and intended for reference. Student's t tests were done between the WT and mutant channels in the same batch of oocytes. All oocytes were injected with cRNAs for the WT mENaC ␣and ␤-subunits together with either the WT or mutant mENaC ␥-subunit.  [3][4][5]. As Zn 2ϩ activates ENaC through relief of Na ϩ self-inhibition, it is unclear whether ␥His 239 is a Zn 2ϩbinding site similar to the other finger subdomain His residues or whether the ␥H239A mutation simply prevents channel inhibition by external Na ϩ . To distinguish between these two possibilities, additional mutations were added to restore Na ϩ self-inhibition in the setting of the ␥H239A mutant. Previously, we found that the G481A mutation within the ␣-subunit greatly enhanced the Na ϩ self-inhibition response (39). Channels bearing both the ␣G481A and ␥H239A mutations exhibited a robust Na ϩ self-inhibition response and a biphasic response to Zn 2ϩ (Fig. 7). Both ␣G481A and ␣G481A/␥H239A channels had I max values that were significantly greater than the WT values (3.4 Ϯ 0.2 (n ϭ 6) and 3.7 Ϯ 0.2 (n ϭ 6), respectively), presumably as a result of the enhanced Na ϩ self-inhibition. The EC 50 for Zn 2ϩ activation of ␣G481A/␥H239A was similar to that of WT (4.2 Ϯ 0.5 M (n ϭ 6) versus WT; p Ͼ 0.05), as was the IC 50 of the Zn 2ϩ inhibitory response (1.2 Ϯ 0.0 mM (n ϭ 6) versus WT; p Ͼ 0.05). Thus, ␥H239A did not weaken the apparent Zn 2ϩ affinity of either the high-or low-affinity site.

Oocytes
Region Surrounding ␥His 88 Is a Determinant of the Low-affinity Inhibitory Site for Zn 2ϩ -Next, the importance of the palm subdomain ␥His 88 and surrounding region was examined to explore the nature of the low-affinity Zn 2ϩ inhibitory site. As shown in Figs. 3 and 5, mutation of ␥His 88 to Ala prevented ENaC inhibition by high concentrations of Zn 2ϩ . Mutation of ␥His 88 to either Arg or Asp also rendered the Zn 2ϩ dose responses monophasic and did not significantly alter the Na ϩ self-inhibition response (Fig. 8, A and B). These results suggest that a His residue at this position is an essential determinant of the Zn 2ϩ inhibitory site.
Zn 2ϩ is often coordinated by multiple neighboring amino acid residues in metal-binding proteins (32). To define additional residues that contribute to the low-affinity Zn 2ϩ inhibitory site, we searched for additional potential Zn 2ϩ -binding residues using sequence alignments of ENaC and ASIC and the resolved cASIC1 structure (40). ␥His 88 aligns with Ala 82 in cASIC1, which is located immediately after the first ␤-strand of ECD (␤1) within the palm subdomain (Fig. 8C). Multiple regions within the same subunit (strand ␤1 and loops ␤1-␤2, ␤11-␤12, and ␤5-␤6) and an adjacent subunit (N-terminal end of helix ␣4 and loop ␣5-␤10) are in close proximity to Ala 82 in cASIC1. Eleven residues in these regions are within 6 Å of the Ala 82 (Fig. 8C). Among these residues, only Ala 413 and Met 364 have homologous residues in mENaC that are often found within a Zn 2ϩ coordination shell (␥Asp 516 and ␤Glu 444 , respectively). We examined whether these sites are involved in ENaC regulation by Zn 2ϩ . As shown in Fig. 8 (D and E), ␥D516A and ␥D516R largely prevented ENaC inhibition by high [Zn 2ϩ ], similar to what we observed with the ␥H88A mutant (Fig. 5). Mutation of the other residue that was predicted to be in close proximity to ␥His 88 (␤E444A) did not significantly alter the Zn 2ϩ effects on ENaC (data not shown). The proximity of ␥Asp 516 to ␥His 88 and their similar mutation-induced changes in the Zn 2ϩ response suggest that the two residues share a functional role in constituting the low-affinity Zn 2ϩ inhibitory site.
The residues corresponding to ␥His 88 and ␥Asp 516 in cASIC1 (Ala 82 and Ala 413 ) are located at a subunit interface (Fig. 8C). According to the suggested counterclockwise arrangement (viewed from above) of the ENaC ␣-, ␤-, and ␥-subunits (14,28), ␥His 88 (Ala 82 in cASIC1) and ␥Asp 516 (Ala 413 in cASIC1) are expected to interact with residues within the thumb subdomain of ␤ENaC. There are five negatively charged ␤ENaC residues within loops ␤9-␣4 and ␣5-␤10 of the thumb subdomain: ␤Glu 354 , ␤Asp 365 , ␤Glu 436 , ␤Glu 444 , and ␤Glu 448 . Given their locations within flexible loops, they might be in close proximity to ␥His 88 and ␥Asp 516 and contribute to the low-affinity inhibitory site. Mutations of these ␤-subunit residues to Ala did not alter the channel's response to Zn 2ϩ (data not shown). Thus, it is unlikely that these residues within the thumb domain of ␤ENaC are involved in Zn 2ϩ regulation of the channel.

DISCUSSION
The major goal of this study was to probe the structural basis of the regulation of ENaC by Zn 2ϩ . The biphasic dose response of Zn 2ϩ on the purinergic receptor P2X is believed to originate from Zn 2ϩ binding at two distinct sites within the receptor: a high-affinity activating site and a low-affinity inhibitory site (41). Because the ENaC/degenerin family is structurally similar to P2X (40), we hypothesized that the bell-shaped Zn 2ϩ doseresponse relationship with ENaC activity is due to the presence of a high-affinity activating site and a low-affinity inhibitory site within the ECD. We found that a minimal two-site model sufficiently describes the bell-shaped dose response of mouse ENaC to external Zn 2ϩ . The presence of a high-affinity activating site and a low-affinity inhibitory site is supported by our observations that Zn 2ϩ -mediated activation or inhibition can be selectively eliminated by targeted mutations of specific residues in the activating site or the inhibitory site, respectively. On the basis of these observations, we propose that ␥His 193 , ␥His 200 , and ␥His 202 within the finger subdomain contribute to a high-affinity Zn 2ϩ -binding site and that ␥His 88 and ␥Asp 516 in the palm subdomain are key determinants of a low-affinity Zn 2ϩ -binding site (Fig. 9).
The following observations suggest that ␥His 193 , ␥His 200 , and ␥His 202 , presumably within the finger subdomain, contribute to a high-affinity Zn 2ϩ activating site. (i) Mutation of these residues increased the EC 50 of Zn 2ϩ activation and had no effect on the IC 50 ( Fig. 5 and Table 1), suggesting that these residues are specifically involved in Zn 2ϩ binding to an activating site. (ii) Although mutations of ␥His 193 , ␥His 200 , and ␥His 202 reduced the maximal ENaC current elicited by Zn 2ϩ , maximal Zn 2ϩ activation was not affected when the Zn 2ϩ inhibitory site was also mutated (␥H88A/␥H200A/␥H202A). Thus, ␥H200A/␥H202A appears to primarily affect the apparent affinity of the activating site and does not alter the efficacy of channel activation. (iii) ␥His 193 , ␥His 200 , and ␥His 202 appear to share a similar functional role in mediating Zn 2ϩ effects on the channel. Individual mutation of any of the three His residues induced a similar phenotype, including an increase in EC 50 , a decrease in maximal Zn 2ϩ activation, and a lack of effect on Na ϩ self-inhibition. Moreover, mutation-induced changes in EC 50 were not additive in ␥H193A/␥H200A/␥H202A channels (Fig. 6D). (iv) Their proximity in the sequence makes it structurally possible for them to contribute to a common binding site. Interestingly, the His 200 -Val 201 -His 202 sequence matches a common Zn 2ϩ -binding motif, HXH, where X can be any residue (42). This motif is also present in other channels (43,44).
Channels with substitutions of all three His residues are still activated by Zn 2ϩ , suggesting that there are additional sites within the channel where Zn 2ϩ binds and activates ENaC. It is notable that His residues are not present at sites in the ␣and ␤-subunits that correspond to ␥His 193 , ␥His 200 , and ␥His 202 (Fig. 9). Further studies are needed to identify additional sites involved in Zn 2ϩ activation of ENaC.
Our results demonstrate that ␥His 88 within the palm subdomain is a key determinant for the low-affinity inhibitory Zn 2ϩ -binding site. Interestingly, like the three finger subdomain His residues, ␥His 88 is also not conserved among ENaC subunits (Fig. 9). Taking advantage of the structural information of ASIC1, a negatively charged residue, ␥Asp 516 , was identified that lies in close proximity to ␥His 88 and is likely involved in the inhibitory effect of Zn 2ϩ as well (Fig. 9). However, it is also possible that mutations of ␥His 88 and ␥Asp 516 may diminish the inhibitory component in the biphasic Zn 2ϩ dose response by disrupting conformational changes induced by Zn 2ϩ binding, instead of abolishing Zn 2ϩ binding. Nevertheless, our observations suggest that this region (loops ␤1-␤2 and ␤11-␤12), located at the middle of the palm subdomain, contributes to a general inhibitory site for different channel regulators (Fig. 9). Indeed, a putative Cl Ϫ -binding site and a Cu 2ϩbinding site have been identified near this region in human ENaCs (14,28). The corresponding region in ASIC1 has been implicated in H ϩ sensing and desensitization (27,(45)(46)(47). This solvent-accessible region appears to be an excellent target for the development of therapeutic agents that modulate activities of these related channels.
The activating effect of Zn 2ϩ on ENaC depends on the presence of Na ϩ self-inhibition (Figs. 3-5) (25). External Zn 2ϩ may disrupt either Na ϩ binding to its receptor site or Na ϩinduced conformational changes. Our observations that mutations at the implicated Zn 2ϩ -binding sites (␥H193A, ␥H200A, ␥H202A, and ␥H88A) did not alter the Na ϩ self-inhibition response suggest that the binding sites for Zn 2ϩ and Na ϩ are distinct. The lack of effect on the apparent affinity for Zn 2ϩ activation of a mutation of a critical site for Na ϩ self-inhibition (␥H239A) also supports the notion of distinct Zn 2ϩ -and Na ϩbinding sites. Therefore, it is likely that Zn 2ϩ binding interferes with Na ϩ -induced motions within the ECD. Similar to Zn 2ϩ , several extracellular factors such as proteases, Cl Ϫ , H ϩ , and small molecules are thought to regulate ENaC gating in part by altering the Na ϩ self-inhibition response (13,15,28,36,48). It would be of great interest to determine whether other extracellular regulators affect Na ϩ self-inhibition in a similar manner to Zn 2ϩ .
Recent studies suggest that ENaCs may belong to ligandgated channels, like other members of the ENaC/degenerin family (25,26,49). Multiple external regulators of ENaC, including transition metals, Cl Ϫ , H ϩ , and other small molecules, exert their effects on ENaC gating by binding to regions within the ECD that are structurally distant from the channel pore (13)(14)(15)28). These gating regulators are expected to remotely influence the channel gate through a series of conformational changes. The identification of these allosteric regulatory sites will advance our understanding regarding mechanisms by which the binding of allosteric regulators alters channel gating.
In summary, this study provides novel insights into the structural basis for Zn 2ϩ regulation of ENaC. Our results suggest that external Zn 2ϩ regulates ENaC activity by binding to mul- FIGURE 9. Finger and palm subdomains of ␥ENaC contain key sites for Zn 2؉ interaction with the channel. A, a structural model of cASIC1. Three ASIC1 subunits are rendered as colored ribbons from coordinates in the Protein Data Bank (code 3HGC) (40). The red circle identifies the finger subdomain of ASIC1 to show the likely locations of ␥His 193 , ␥His 200 , and ␥His 202 that are proposed to contribute to the high-affinity activating Zn 2ϩ -binding site. Their exact homologous sites within the ASIC1 structure are not known due to a low level of sequence homology in the finger subdomains. The sequence alignments of mouse, rat, and human ␣-, ␤-, and ␥-subunits and cASIC1 were performed with Vector NTI 11.0 (Invitrogen). Rat and human ENaC sequences are omitted for clarity. The three His residues are shown in red. The red square shows a subunit interface between the palm subdomain of the green subunit and the thumb subdomain of the blue subunit, where ␥His 88 and ␥Asp 516 are proposed to reside in a mENaC structure and contribute to the low-affinity inhibitory Zn 2ϩ -binding site. The green ASIC1 subunit was arbitrarily considered to correspond to ␥mENaC. The sequences of loops ␤1-␤2 and ␤11-␤12 of ␣␤␥mENaC and cASIC1 are shown with identical residues highlighted in yellow and ␥His 88 and ␥Asp 516 in red. B, enlargement of the area enclosed by the red square in A. Three ␤-strands (␤1, ␤11, and ␤12) are labeled as in the literature (40). The ␥His 88 -homologous Ala 82 and ␥Asp 516 -homologous Ala 413 residues were mutated to His and Asp, respectively, in the ASIC1 model. The labeled blue sphere shows Cl Ϫ bound within the thumb subdomain of the blue subunit. Structural models were drawn using PyMOL 1.3 (50). tiple extracellular sites, including a high-affinity activating site in the finger subdomain and a low-affinity inhibitory site in the palm subdomain of the ␥-subunit. These findings advance our understanding of the regulation of ENaC gating by extracellular factors.