Requirement of Multiple Protein Domains and Residues for Gating K ATP Channels by Intracellular pH*

ATP-sensitive K (cid:1) channels ( K ATP ) are regulated by pH in addition to ATP, ADP, and phospholipids. In the study we found evidence for the molecular basis of gating the cloned K ATP by intracellular protons. Systematic con- structions of chimerical Kir6.2-Kir1.1 channels indi-cated that full pH sensitivity required the N terminus, C terminus, and M2 region. Three amino acid residues were identified in these protein domains, which are Thr-71 in the N terminus, Cys-166 in the M2 region, and His-175 in the C terminus. Mutation of any of them to their counterpart residues in Kir1.1 was sufficient to completely eliminate the pH sensitivity. Creation of these residues rendered the mutant channels clear pH-dependent activation. Thus, critical players in gating K ATP by protons are demonstrated. The pH sensitivity enables the K ATP to regulate cell excitability in a num- ber of physiological and pathophysiological conditions when pH is low but ATP concentration is normal. ATP-sensitive K (cid:1) channel ( K ATP ) a unique member in the K (cid:1) channel family, which directly intermediary metabolism a the inward rectifying significantly larger than from the pcDNA3-injected oocytes, the strongly activated the exposure to 3 m M azide, and 3) 100 (cid:1) M Ba 2 (cid:1) inhibited the currents. When a mutant failed to express functional channels in (cid:3) 60 oocytes tested, another two injections of the same mutant from different colonies were followed in (cid:3) 60 cells each. If there was still a lack of expression, we believed that the mutation was too severe to produce any functional channels, and further experimentation was not attempted. Experiments were performed in a semi-closed recording chamber (BSC-HT, Medical Greenvale, NY) in which oocytes were placed on a supporting nylon mesh, and the perfusion solution bathed both the top and bottom surface of the oocytes. The perfusate and the superfusion gas entered the chamber from two inlets at one end and flowed out at the other end. There was a 3 (cid:4) 15-mm gap on the top cover of the chamber that served as the gas outlet and an access to the oocytes for recording microelectrodes. At base line, the chamber was ventilated with atmospheric air. Exposure of the oocytes to CO 2 was carried out by switching to a perfusate that had been bubbled for at least 30 min with a gas mixture containing CO 2 at various concentrations balanced with 21% O 2 and N 2 and superfused with the same gas 27–30). The high solubility of CO 2 resulted in a detectable change intra- or extracellular as fast as 10

experiments in which we used the Kir6.2 with a truncation of 36 amino acids at the C-terminal end, i.e. Kir6.2⌬C36 that expresses functional channels without the SUR subunit and retains fair ATP sensitivity (22). Several chimerical channels were generated based on peptide sequences of Kir6.2 and Kir1.1, a Kir channel that is inhibited by intracellular acidosis (23)(24)(25) and has been shown to express functional channels in its chimeras with Kir6.2 (26). These chimeras were studied in whole-cell recording using hypercapnia, a condition that does not cause channel rundown (21). Our results indicate that there are three separate protein domains in the Kir6.2 protein that are crucial for the pH sensitivity. We have identified critical amino acid residues within each of these protein domains. Replacement of any of them with their counterpart residues in Kir1.1 interrupts the channel sensitivity to CO 2 /pH.

MATERIALS AND METHODS
Frog oocytes were obtained from Xenopus laevis. The frogs were anesthetized by bathing them in 0.3% 3-aminobenzoic acid ethyl ester. A few lobes of ovaries were removed after a small abdominal incision (ϳ5 mm). Then the surgical incision was closed, and the frogs were allowed to recover from the anesthesia. Xenopus oocytes were treated with 2 mg/ml collagenase (Type I, Sigma) in OR2 solution (NaCl 82 mM, KCl 2 mM, MgCl 2 1 mM, and HEPES 5 mM, pH 7.4) for 90 min at room temperature. After 3 washes (10 min each) of the oocytes with the OR2 solution, cDNAs (25-50 ng in 50 nl of water) were injected into the oocytes. The oocytes were then incubated at 18°C in the ND-96 solution containing 96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM HEPES, and 2.5 mM sodium pyruvate with 100 mg/liter Geneticin added, pH 7.4.
Rat Kir1.1 (ROMK1, GenBank TM accession number X72341) and mouse Kir6.2 (mBIR, GenBank TM accession number D50581) cDNAs were generously provided by Dr. S. Hebert at Yale University and Dr. S. Seino at Chiba University in Japan, respectively. The cDNAs were subcloned to a eukaryotic expression vector (pcDNA3.1, Invitrogen Inc., Carlsbad, CA) and used for Xenopus oocyte expression without cRNA synthesis. Chimerical constructs were prepared by the overlap extension at the junction of the interested domains using the polymerase chain reaction (Pfu DNA polymerase, Stratagene, La Jolla, CA). Sitespecific mutations were made using a site-directed mutagenesis kit (Stratagene). The orientation of the constructs and correct mutations were confirmed with DNA sequencing.
Whole-cell currents were studied on the oocytes 2-4 days after injection. Two-electrode voltage clamp was performed using an amplifier (Geneclamp 500, Axon Instruments Inc., Foster City, CA) at room temperature (ϳ24°C). The microelectrodes were filled with 3 M KCl. One of the electrodes (1.0 -2.0 megaohms) served for voltage measurements, and the other electrode (0.3-0.6 megaohms) was used for current recording. Current records were low-pass filtered (Bessel, 4-pole filter, 3 dB at 5 kHz), digitized at 5 kHz (12-bit resolution), and stored on computer disc for later analysis (pClamp 6, Axon Instruments) (21,(27)(28)(29)(30). The extracellular solution contained 90 mM KCl, 3 mM MgCl 2 , and 5 mM HEPES, pH 7.4. Under this condition, cells showed a membrane potential of ϳ0 mV, so that we studied the currents with a holding potential of 0 mV. Expressions of functional Kir channels were confirmed using one or two of the following methods. 1) The amplitude of the inward rectifying currents was significantly larger than that recorded from the pcDNA3-injected oocytes, 2) the currents were strongly activated with the exposure to 3 mM azide, and 3) 100 M Ba 2ϩ inhibited the currents. When a mutant failed to express functional channels in ϳ60 oocytes tested, another two injections of the same mutant from different colonies were followed in ϳ60 cells each. If there was still a lack of expression, we believed that the mutation was too severe to produce any functional channels, and further experimentation was not attempted.
Experiments were performed in a semi-closed recording chamber (BSC-HT, Medical System, Greenvale, NY) in which oocytes were placed on a supporting nylon mesh, and the perfusion solution bathed both the top and bottom surface of the oocytes. The perfusate and the superfusion gas entered the chamber from two inlets at one end and flowed out at the other end. There was a 3 ϫ 15-mm gap on the top cover of the chamber that served as the gas outlet and an access to the oocytes for recording microelectrodes. At base line, the chamber was ventilated with atmospheric air. Exposure of the oocytes to CO 2 was carried out by switching to a perfusate that had been bubbled for at least 30 min with a gas mixture containing CO 2 at various concentrations balanced with 21% O 2 and N 2 and superfused with the same gas (21,(27)(28)(29)(30). The high solubility of CO 2 resulted in a detectable change in intra-or extracellular acidification as fast as 10 s in these oocytes.
The following nomenclatures were used in the present study. Kir6.2 and Kir1.1 were divided into three segments in their peptide chains, i.e. 1) the N terminus, 2) C terminus, and 3) all the rest M1 through M2 regions. A chimera with both N and C termini from Kir6.2 and the rest from Kir1.1 was named SOS. If only the N terminus was from Kir6.2, it was called SOO. If the SOS carried the M1 region from Kir6.2, it was referred to SOSm1. To facilitate the comparison of amino acid residues between the SO chimeras and the wt channels, they were numbered by their original locations in the wt Kir1.1 or Kir6.2 protein rather than by their new positions in the sequences of the chimeras.
Data are presented as means Ϯ S.E. Analysis of variance or Student's t test was used. Differences of CO 2 and pH effects before versus during exposures were considered to be statistically significant if p Յ 0.05.

RESULTS
Differential Sensitivities of Kir6.2 and Kir1.1 Channels to pH-The Kir6.2⌬C36 was expressed in Xenopus oocytes. As reported previously (21,22), inward rectifying currents were recorded in the whole-cell configuration after Kir6.2⌬C36 cDNA injection. Exposure to CO 2 produced reversible and concentration-dependent activation of the inward rectifying currents (Fig. 1A). The effect was mediated by pH rather than molecular CO 2 , as intracellular, but not extracellular, acidification to the same levels seen during CO 2 exposures produced the same degrees of channel activation (22). In excised insideout patches, the Kir6.2⌬C36 currents were also stimulated by modest acidification on the cytosolic side of membranes in the absence of ATP and other cytosolic soluble factors. With strong acidification, however, the channel activation was followed by marked inhibition that appeared to be a result of channel rundown as reported previously (8,11,13,20,31). Similar effects were observed in the presence of 1 mM ATP in the internal solution, 2 indicating that the pH sensitivity is independent of ATP. Since the channel rundown occurs in excised patches, further studies were done using whole-cell recording and hypercapnic acidosis with 15% CO 2 , an experimental condition that we have well documented previously (21,(27)(28)(29)(30). In contrast to the Kir6.2, Kir1.1 was inhibited during hypercapnia and intracellular acidification (Fig. 1B), as demonstrated previously (23-25, 27, 32).
Essential Structures in the Kir6.2 Responsible for the pH Sensitivity-Our results suggest that the pH sensitivity in the Kir6.2 is an inherent property of the channel protein, similar to those described in several other Kir channels (23-25, 27, 30, 32-34). To identify the specific structures in the Kir6.2 protein that enable the channel to respond to acidic pH, chimerical Kir channels were constructed by recombination of protein domains of Kir6.2 and Kir1.1. We reasoned that the pH sensitivity relied on the integrity of the proton-sensing and channelgating mechanisms in the Kir6.2 protein. Interruption of the integrity will cause a loss of the pH sensitivity. Thus, we divided Kir6.2 and Kir1.1 into three segments in their peptide chains, i.e. 1) the N terminus, 2) C terminus, and 3) all the rest M1 through M2 regions. The chimera with both N and C termini from Kir6.2 and the rest from Kir1.1 is named SOS (S refers to six, and O to one). If the N terminus is the only part from Kir6.2, it is called SOO. Accordingly, the wild-type (wt) Kir6.2⌬C36 refers to SSS, and the wt Kir1.1 to OOO ( Fig. 2A).
Six chimerical channels were systematically constructed using these Kir6.2 and Kir1.1 protein domains ( Fig. 2A) . Series of voltage commands (from Ϫ140 to 120mV with 20-mV increments at a holding potential of 0 mV) were applied to the cell in a bath solution containing 90 mM K ϩ . Under this condition, clear inward rectifying currents were seen in the oocyte 3 days after the injection. These currents were strongly and reversibly inhibited when the cell was exposed to 15% CO 2 . B, similar experiments were performed on another Kir1.1-injected oocyte. Unlike the Kir6.2⌬C36, the Kir1.1 currents were strongly inhibited with the same concentration of CO 2 .
the recombinant channels showed inward rectifying currents of 1-3 A, which were significantly larger than those recorded from the vector-injected cells (0.5 Ϯ 0.1 A, n ϭ 9). The SSO and OSO were the only two that did not show evident channel expression. Azide (3 mM) treatment of the cells injected with SSO or OSO did not reveal any additional increase in the current amplitude. Thus, how SSO and OSO respond to acidosis is unclear. Although all other chimeras expressed detectable inward rectifying currents, none of them displayed a hypercapnic response as large as the wt Kir6.2⌬C36 ( Fig. 2A, Table  I). Clearly, these chimeras had caused interruptions of the pH-dependent gating mechanisms.
Because none of the chimeras had a full CO 2 /pH sensitivity as did the wt Kir6.2⌬C36, it is possible that the essential structures for the pH sensitivity are not contained in these chimeras. To include them, we extended the C terminus to the entire M2 region of Kir6.2. This chimera SOSm2 expressed large inward rectifying currents (4.9 Ϯ 3.0 A, n ϭ 7), with its inward rectification more like Kir6.2 than Kir1.1 (Fig. 2B). Exposure of the SOSm2 to 15% CO 2 enhanced the inward rectifying currents by 123.9 Ϯ 15.1% (n ϭ 7), which were slightly smaller but not significantly different from the wt Kir6.2⌬C36 (p Ͼ 0.05) (Figs. 2B). Inclusion of the M1 sequence to the N terminus (SOSm1), however, did not produce any significant additional effect on the CO 2 sensitivity over the SOS (Fig. 2A).
To determine if two of the three protein domains are sufficient for the pH sensitivity, chimeras containing two of the M2, N terminus and C terminus from Kir6.2 and the remaining sequences from Kir1.1 were constructed. The OOSm2 and SOOm2 failed to yield any functional channels, whereas the SOS had pH sensitivity of only about one-third that of the SOSm2 (Fig. 2A). These results thus indicate that protein domains necessary for the full CO 2 /pH sensitivity consist of at least the N terminus, C terminus, and M2 region in Kir6.2.
Amino Acid Residue in the N Terminus of the Kir6.2 Protein-The N terminus has been previously shown to play an important part in pH sensitivity in several Kir channels. A lysine residue at near the M1 region (Lys-80 in Kir1.1, Lys67 in Kir4.1) is a critical player (24,27). This lysine residue is not found in Kir6.2. At the same location, the Kir6.2 has a threonine instead (Thr-71). Since Kir1.1 and Kir4.1 channels with this lysine residue are inhibited by acidic pH (25,28), it is possible that the lack of the positively charged residues renders the Kir6.2 an opposite pH sensitivity. To test this hypothesis, we performed site-directed mutagenesis experiments at this site. Mutation of the Thr-71 to lysine yielded fair inward rectifying currents in whole-cell recordings. When the T71K was exposed to 15% CO 2 , we found that the mutant channels became insensitive to hypercapnia (Fig. 3A). The same result was obtained in the SOSm2 carrying the T71K mutation (Fig. 3D). i.e. the N terminus, M1 through M2 segment, and C terminus. Chimerical channels were constructed using these protein domains. The fragment from Kir6.2 is named "S," and otherwise, "O." Responses of these chimeras to acidic pH were studied using 15% CO 2 . The wt Kir6.2⌬C36 or SSS was stimulated during hypercapnia, whereas the Kir1.1 was inhibited. The chimerical recombination caused severe interruption of the CO 2 /pH sensitivity, as none of the Kir6.2-Kir1.1 chimeras had a similar CO 2 sensitivity as the wt Kir6.2⌬C36. When the M2 region from Kir6.2 was included in the SOS, the chimera SOSm2 showed CO 2 sensitivity almost identical to the wt Kir6.2⌬C36. Similar construction of the M1 region (SOSm1), however, had no additional effect. Note that the OSO and SSO failed to express functional channels. I, current. B, response of the SOSm2 chimera to hypercapnia. The SOSm2 chimera contains the M2 region, N and C termini from Kir6.2 and the rest from Kir1.1. This chimera expressed functional Kir currents just like the wt Kir6.2⌬C36 rather than Kir1.1. CO 2 sensitivity was also like the wt Kir6.2⌬C36. The inward rectifying currents increased by 130% during a 5-min exposure to 15% CO 2 . Wash-out led to a good recovery.
Systematic mutagenesis was subsequently carried out on the Thr-71 by replacing it with acidic, nonpolar, and other polarneutral residues. There is a methionine at this site in Kir2.1, Kir2.3, Kir5.1, and Kir7.1. Thus we constructed the T71M mutant. The substitution of the Thr-71 with such a nonpolar residue greatly reduced the pH sensitivity (Fig. 3D). Replacement of the Thr-71 with an aspartate did not produce a functional channel. When it was mutated to serine, the mutant T71S showed an increase in the current amplitude by Ͼ100% with hypercapnia (Fig. 3D). Since both threonine and serine are the potential substrate of phosphorylation, we mutated the Thr-71 to another polar-neutral residue, asparagine. Currents of the T71N mutant were enhanced by 160.3 Ϯ 19.4% (n ϭ 5) by 15% CO 2 (Fig. 3, B and D), suggesting that a polar-neutral residue at this site is necessary for maintaining the pH sensitivity in Kir6.2.
Creation of the threonine residue in the OSS produced clear inward rectifying currents that increased reversibly by 72.0 Ϯ 18.2% (n ϭ 5) in response to 15% CO 2 (Fig. 3, C and D), an increase that was significantly larger than the OSS (p Ͻ 0.01). Construction of this threonine in OOSm2 failed to produce functional expression. These results therefore suggest that the Thr-71 is the determinant residue in the N terminus for the pH sensitivity of Kir6.2.
Residues Involved in the M2 Membrane-spanning Region-The M2 region of several Kir channels was aligned in Fig. 4A. Amino acid residues in this region are highly homologous between Kir1.1 and Kir6.2. It is known that the area on the  -168). Phenylalanine has a side chain much larger than that of leucine and valine, seen at its counterpart positions. To elucidate whether the size of residues controls the pH sensitivity, we site-specifically mutated these residues in Kir6.2 to those found in Kir1.1. The F168L and I162F mutants were constructed using both Kir6.2⌬C36 and SOSm2 as the template. Of the four mutants studied, the SOSm2-I162F was the only one expressing inward rectifying currents (Table I), which remained, although smaller, to be stimulated during hypercapnia by 74.4 Ϯ 14.6% (n ϭ 8) (Fig. 4B). Subsequently, we studied two other sites in which a cysteine residue exists in each of Kir6.2 and Kir1.1. Mutations of Cys-175 to leucine in Kir1.1 and SOS did not produce channels that were stimulated by CO 2 . The other cysteine mutation, however, was remarkable. The Kir6.2-C166A mutant showed large baseline currents (12.5 Ϯ 4.6 A, n ϭ 4) and a much weaker inward rectification than the wt Kir6.2⌬C36 (Fig. 4C). At highly negative membrane potentials, the inward rectifying currents also became weaker, suggesting that this site contributes to the voltage-dependent rectification. More interestingly, substitution of this single residue with alanine (Kir6.2-C166A) completely abolished the CO 2 sensitivity of Kir6.2 (Ϫ1.9 Ϯ 1.1%, n ϭ 4) (Fig. 4, B and C). The same effect was also observed in the SOSm2-based mutant SOSm2-C166A (Ϫ13.4 Ϯ 8.0%, n ϭ 8).
Systematic mutations of the Cys-166 were thereafter performed using Kir6.2⌬C36 and SOSm2. The mutant channels became CO 2 -insensitive when this residue was replaced with a negative or a polar-neutral residue, i.e. Kir6.2-C166E, Kir6.2-FIG. 3. Site-directed mutagenesis of threonine 71 in the Kir6.2. A, when the Thr-71 was mutated to a lysine residue, channel sensitivity to 15% CO 2 was totally eliminated. Instead, the currents were slightly inhibited. B, a replacement to asparagine yielded a channel that had a full CO 2 sensitivity as the wt Kir6.2⌬C36. C, Construction of the threonine residue in OSS led to a channel that showed a higher CO 2 sensitivity over the OSS. D, comparison of the CO 2 sensitivity of Thr-71 mutations in Kir6.2⌬C36. I, current; Kir6.2, Kir6.2⌬C36. C166S, SOSm2-C166E, and SOSm2-C166S (Fig. 4B). Mutation to a positive residue (Kir6.2-C166K and SOSm2-C166K) did not yield functional expression. In contrast, the mutant channels remained stimulated by hypercapnic acidosis, when a valine or histidine was the replacement (Table I). Indeed, the Kir6.2-C166V showed CO 2 sensitivity almost identical to the SOSm2 (Fig. 4B).
Reversal mutations of this cysteine residue, however, did not show any dramatic effect. The Kir1.1-A177C was still inhibited by CO 2 , and the SOS-A177C remained pH-insensitive (Fig. 4B). Therefore, it is possible that other adjacent residues within the M2 are also involved. To identify these residues, we performed further mutagenesis studies. We found that the formation of disulfide bond with Cys-175 was not a reason, because combined mutations of these two residues still showed an inhibited phenotype in SOS (SOS-C175L/A177C, -20.5 Ϯ 3.5%, n ϭ 4). Subsequently, we created mutants to include Ser-172 and/or Phe-173. The SOS-based mutants S172A/F173I/C175L/A177C (SOS-AILC) and S172A/C175L/A177C (SOS-ALC) expressed functional currents with clear inward rectification. Both of them were strongly stimulated by 15% CO 2 to a degree that statistically was not different from the SOSm2 and the wt Kir6.2⌬C36 channels (Fig. 4D). In contrast, the F173I/C175L/ A177C (SOS-ILC) had no significant effect. Thus, these results suggest that the molecular determinant for the pH-dependent activation is likely to be the short motif (ϳ 6 amino acids) at the intracellular end of the M2 centered by the Cys-166.
His-175 in the C Terminus-Our previous studies show that His-175 in the C terminus is critical for the pH sensitivity of Kir6.2 (21). This was confirmed in our current studies (Table I).
In addition, we found that mutation of this histidine residue in SOSm2 greatly reduced the current response to 15% CO 2 (6.5 Ϯ 0.7%, n ϭ 4) (Fig. 5A). Although construction of a histidine residue at the same site in SOOm2 did not produce functional expression, creation of such a residue in the SOO (SOO-K186H) gave rise to functional channels. During CO 2 exposure, the SOO-K186H currents increased by 44.4 Ϯ 12.9% (n ϭ 5), which were almost identical to the response of the SOS currents (p Ͼ   FIG. 4. Involvement of M2 residues in the CO 2 sensitivity. A, the M2 regions are aligned in five Kir channels. The residues in bold were studied in Kir6.2 and Kir1.1. Those boxed are essential for the CO 2 sensitivity. B, effects of site-specific mutations of the M2 residues on the CO 2 sensitivity. Note that the SOSm2, Kir6.2-C166V, SOS-ALC, and SOS-AILC all have similar CO 2 sensitivity, whereas others are either less sensitive to or inhibited by 15% CO 2 . C, membrane currents recorded from an oocyte injected with Kir6.2-C166A. During CO 2 (15%) exposure, the current amplitude showed very little difference from the base-line control. D, construction of three residues (S172A/ C175L/A177C) in the SOS (SOS-ALC) rendered the pH sensitivity as large as the SOSm2 and the wt Kir6.2⌬C36. I, current; Kir6.2, Kir6.2⌬C36. 0.05, Table I, Fig. 2A) but significantly greater than the SOO currents (p Ͻ 0.05) (Fig. 5B). Thus, the His-175 is likely to be the determinant player in the C terminus. DISCUSSION In the present studies, we have demonstrated several protein domains and amino acid residues in the Kir6.2 necessary for the pH sensitivity of the K ATP channels. Interruption of any of them causes a complete loss of the pH sensitivity in the channels; construction of these sites or residues yields fair pH sensitivity. Thus, gating of the K ATP channels by intracellular protons requires complex interactions of multiple protein domains.
Multiple Protein Domains in Proton Sensing-Several Kir channels are pH-sensitive, such as Kir1.1, Kir1.2, Kir2.3, Kir2.4, Kir4.1, Kir4.2, and the heteromeric Kir4.1-Kir5.1 (24,25,28,29,(32)(33)(34)(35)(36)(37). Although pH sensitivities vary among individual members, there is one common feature in these Kir channels; they all are inhibited by intra-and/or extracellular protons. Unlike these Kir members, the major effect of acidosis on the cloned K ATP channels is stimulatory (21). We have previously shown that the pH sensitivity in the K ATP is an inherent property of the channel that depends on the Kir6 subunit but not the SUR1 (21). We believe that the localization of the pH-sensing mechanisms to the Kir6 subunit is important, because the finding suggests a possibility of dissecting the molecular mechanisms underlying pH sensing using available molecular genetic techniques. In fact, our current studies have provided information about the pH-sensing mechanisms in the Kir6.2 protein. We have found that the proton-dependent Kir6.2 gating requires at least three protein domains, i.e. the N terminus, C terminus, and M2 region. Replacement of any of these protein domains with that in Kir1.1 severely interrupts the channel sensitivity to hypercapnic acidosis.
A similar requirement of multiple protein domains for proton sensing has also been found in other Kir channels. In Kir1.1, several residues in the N and C termini are known to be involved in the pH sensitivity (23,27,38,39). In Kir2.3, proton sensing relies on an interaction of both N and C termini at near the membrane-spanning regions, whereas proton sensors are likely to be located in more distal areas of the C terminus (34). These previous observations as well as our current findings indicate that the channel gating by even simple molecules such as protons attributes to movements of several parts of the channel proteins. Thus, the control of the movements may allow complex modulations of the channels.
Potential Proton-sensing Site-The complex movements of these protein domains appear to start with protonation of certain amino acid residues in the channel protein. Titratable amino acids have been the focus of studies that are aimed at demonstrating the proton sensors. In Kir1.1, a lysine residue at the N terminus and several histidine residues in the C terminus have been believed to be the proton sensors (23,24,40). In Kir2.3, the C-terminal histidine residues seem to play the same role as their counterparts in Kir1.1, whereas the N-terminal lysine residue does not exist (34). In Kir6.2, we have previously demonstrated a histidine residue (His-175), mutation of which alone is sufficient to eliminate completely the acid-induced channel activation (21). The finding has been proven by our current data. Our results from the present study further suggest that this histidine residue is the proton binding site for the following reasons. First, the His-175 is the only titratable residue found in all three protein domains shown to be crucial for pH sensitivity of Kir6.2. Second, removal of this residue eliminates the pH-dependent activation in both the wt Kir6.2⌬C36 and SOSm2. Third, creation of a histidine residue at the same site in SOO yields substantial pH sensitivity in the mutant channel. Thereby, it is very likely that this histidine is the protonation site that initiates channel activation after proton binding.
Putative Gating Mechanisms-How are other protein domains and amino acid residues involved in the channel sensitivity to hypercapnic acidosis? The lack of titratable residues within these protein domains indicates that they are not the proton sensors. Thus, a possible explanation for their involvement in the pH sensitivity is that they act as a unit in gating the channels by intracellular protons. Recent studies in CNG, Kir, and Kv channels have shown that protein domains at near the membrane-spanning sequences play an important role in channel gating (26,34,40,41,42,44,49). During the gating process, these protein domains interact with each other, leading to a change in protein conformation as well as channel activity. Because of the similarity in their locations, the N-and C-terminal domains of Kir6.2 may also take part in the gating process. Indeed, certain physical interactions of these intracellular termini have been demonstrated (49). Therefore, interruption of any of these protein domains may interfere the N-C-terminal interaction and in turn jeopardize the pHdependent channel gating.
The N terminus proximal to the M1 region is known to be critical for the pH sensitivity in several Kir channels (24,27,29,36,38,39,43). A lysine residue (Lys-80 in Kir1.1 or Lys-67 in Kir4.1) is particularly important in the pH-dependent channel gating. Mutation of this lysine leads to a complete loss of pH sensitivity in homomeric Kir1.1, Kir1.2, Kir4.1, Kir4.2 (24,27,43) and the heteromeric Kir4.1-Kir5.1 (27,29,36). Since lysine is potentially titratable, this residue has been believed to be a protonation site (24,40). Unlike lysine, threonine found at the same position in Kir6.2 is not titratable at all, which appears to play a similar role as does Lys-80 in Kir1.1. Our systematic mutagenesis analysis indicates that the polar-neutral nature of the threonine rather than its phosphorylation potential is the key, as the similar pH sensitivity can be produced with other polar-neutral residues at this location. Mutation to a positive residue totally abolishes the pH sensitivity of Kir6.2, whereas switches to a nonpolar or an acidic residue greatly reduce the pH sensitivity or prevents the channel from expression. Thus, a polar-neutral residue is required at this site for the pH-dependent activation in Kir6.2. Such a requirement suggests that this site is involved in the interaction with other residue(s) or protein domain, probably by forming hydrogen bonds or the maintenance of a critical conformation for the channel gating.
The M2 region lines the inner pore of Kir channels, especially the area on the cytosolic side where the Cys-166 is located. The M2 region is known to be responsible for rectification, single channel conductance and gating (45)(46)(47). Indeed, some residues in the M2 region have been shown to be involved in multiple channel biophysical properties. We have previously found that mutations of Asn-171 affect rectification, single channel conductance, and pH sensitivity of Kir1.1 (27). As for the Cys-166, we have noticed that mutations of this residue change rectification. Perhaps the most intriguing effect of the Cys-166 mutation is the elimination of the pH sensitivity. Our data indicate that the Cys-166 is an important player in a small motif of 5-6 residues. Construction of this motif rather than the cysteine alone in the SOS produces the full pH sensitivity. Thus, the M2 is involved in the channel gating by intracellular protons in the Kir6.2.
It is known that cysteine is titratable with pK 8.5. The Cys-166 titration, however, is not necessary for the activation of Kir6.2 because the C166V mutant showed similar pH sensitivity as the wt channel. Since the Cys-166 has been reported to affect Kir6.2 gating by ATP (48), it is possible that this site is involved in the channel gating by protons as well as via its interaction with other residues, especially Thr-71. Based on the structural information of the bacterial KcsA channel, the site of Cys-166 is located in the immediate vicinity of the Thr-71 (50), allowing a direct contact to each other. Such an interaction between these two sites may favor a special conformation of the M2 and/or M1 membranespanning domain, maintaining the channel to the open or closed state. It is possible that the interaction of these two residues is affected by protonation or deprotonation of the His-175. Consequently, the channel may be switched to another state.
In conclusion, the cloned K ATP channels are regulated by intracellular protons. The pH sensitivity requires three separate protein domains, i.e. the N terminus, C terminus, and M2 region. Three amino acid residues have been identified in each of these protein domains. Mutation of any of these residues completely eliminates the pH sensitivity similar to the mutation of the protein domains. Constructions of these residues allows substantial pH sensitivity. Therefore, K ATP gating by intracellular protons requires interactions of multiple protein domains and residues in the Kir6.2 channel protein.