Stomatin Modulates Gating of Acid-sensing Ion Channels*

Acid-sensing ion channels (ASICs) are H+-gated members of the degenerin/epithelial Na+ channel (DEG/ENaC) family in vertebrate neurons. Several ASICs are expressed in sensory neurons, where they play a role in responses to nociceptive, taste, and mechanical stimuli; others are expressed in central neurons, where they participate in synaptic plasticity and some forms of learning. Stomatin is an integral membrane protein found in lipid/protein-rich microdomains, and it is believed to regulate the function of ion channels and transporters. In Caenorhabditis elegans, stomatin homologs interact with DEG/ENaC channels, which together are necessary for normal mechanosensation in the worm. Therefore, we asked whether stomatin interacts with and modulates the function of ASICs. We found that stomatin co-immunoprecipitated and co-localized with ASIC proteins in heterologous cells. Moreover, stomatin altered the function of ASIC channels. Stomatin potently reduced acid-evoked currents generated by ASIC3 without changing steady state protein levels or the amount of ASIC3 expressed at the cell surface. In contrast, stomatin accelerated the desensitization rate of ASIC2 and heteromeric ASICs, whereas current amplitude was unaffected. These data suggest that stomatin binds to and alters the gating of ASICs. Our findings indicate that modulation of DEG/ENaC channels by stomatin-like proteins is evolutionarily conserved and may have important implications for mammalian nociception and mechanosensation.

Acid-sensing ion channels (ASICs) 1 are non-voltage gated Na ϩ channels found in the central and peripheral nervous systems of mammals (1)(2)(3)(4)(5)(6)(7) and other vertebrates (8,9). The expression of ASIC1, -2, and -3 genes generates H ϩ -gated channels as either homomultimers when expressed individually or as heteromultimers when co-expressed in combination. In the central nervous system, ASIC1 and -2 are expressed in regions of synaptic activity where they play a role in synaptic plasticity and learning (10 -13). Other ASICs are expressed in peripheral sensory neurons, including specialized sensory nerve terminals (14 -16), where they may function as transducers of sensory stimuli. Because ASICs are activated by protons in pH ranges that occur during tissue ischemia or injury, they are candidates to sense acidosis associated with painful conditions such as cardiac ischemia (17,18), skin (16, 19 -21) and muscle inflammation (22), and malignancy (23). In other sensory neurons they may function as mechanotransducers; targeted deletion of ASIC2 or ASIC3 results in mice with altered mechanosensitivity in the skin (14,16,20).
ASICs are members of the larger degenerin/epithelial Na ϩ channel (DEG/ENaC) family of ion channels. Two members in the nematode Caenorhabditis elegans, MEC-4 and MEC-10, are required for normal touch sensation (24 -26) and generate small constitutive currents when expressed in Xenopus oocytes (27). Additional mec genes that encode intracellular or extracellular proteins are also necessary for touch sensation in the worm. Genetic and direct interaction studies suggest a model of mechanotransduction whereby these accessory proteins function to link MEC-4 and MEC-10 channels to the intracellular cytoskeleton and the extracellular matrix; deformation of these components by mechanical stimuli is thought to open the channel (28,29). One of these accessory proteins in C. elegans, MEC-2, is an integral membrane protein with a short cytoplasmic NH 2 terminus, a hydrophobic midportion that forms a hairpin turn within the cell membrane, and a longer cytoplasmic COOH terminus (30). mec-2 mutations abolish touch sensation in the worm (31,32), and genetic studies suggest that it interacts with the channel subunits (MEC-4 and MEC-10) and cytoplasmic microtubules (MEC-7 and MEC-12), forming a link between the channel and the underlying cytoskeleton (30,33). Heterologous expression studies confirmed that MEC-2 interacts with MEC-4 and MEC-10 and found that MEC-2 potentiates current (27).
Stomatin is the mammalian homolog of MEC-2, sharing 65% identity and 85% similarity. In a rare autosomal dominant hemolytic anemia called stomatocytosis, stomatin is absent in red blood cells, which have increased cationic leak current and swell and lyse prematurely (34 -38). However, the role of stomatin in the pathogenesis of the disease is unclear (39,40). Similar to MEC-2, stomatin interacts with the cytoskeleton in red blood cells (38) and co-localizes with actin in epithelial cells (41). Stomatin has also been shown to bind to the glucose transporter GLUT-1 and decrease the rate of glucose uptake (42). In addition, stomatin is also expressed in sensory neurons (43), and it co-localizes with ASIC-related DEG/ENaC subunits in rat mechanosensory neurons of whisker follicles (44). We therefore tested the hypothesis that stomatin modulates the activity of ASICs.
Heterologous Expression of cDNA-For co-immunoprecipitation and co-localization studies, COS-7 cells were transfected by electroporation using 10 7 cells and 20 g of plasmid DNA. mASIC1a, -2a, or -3 (10 g) was transfected alone or with m-stomatin-FLAG or empty vector (10 g). COS-7 cells were maintained in culture with Dulbecco's modified Eagle's medium with 10% fetal calf serum in a humidified atmosphere of 5% CO 2 in air at 37°C. Cells were studied 24 to 72 h after transfection.
For electrophysiological experiments, cDNAs were expressed using 6 g/1.5 ml cationic lipid (TransFast; Promega) in CHO cells grown in 35-mm dishes. mASIC subunits were expressed alone or with m/hstomatin at the described concentrations. DsRed was co-transfected to keep the total amount of cDNA constant. cDNA for green fluorescent protein (0.33 g/0.4 ml) was also expressed to facilitate detection of transfected cells by epifluorescence. We have found that Ͼ90% of green cells and Ͻ5% of non-green cells have an acid-activated current. Cells were cultured in F12 medium with 10% fetal bovine serum and 1% penicillin-streptomycin at 37°C and were studied 48 -72 h after transfection.
Biotinylation of Cell Surface Proteins-48 h following transfection by electroporation, COS-7 cells were washed with cold PBS three times on ice, labeled with 0.25 mg/ml sulfo-N-hydroxysuccinimidobiotin (Pierce) in PBS for 20 min on ice, washed twice with TBS, and lysed as described above. Labeled cell surface proteins were isolated by incubating 500 g of cell lysate with immobilized neutrAvidin beads (Pierce) for 2 h at 4°C. In parallel, ASIC3 was immunoprecipitated from 500 g of the same cell lysate by incubation with anti-ASIC3 3.2 antibodies at 4°C for 2 h followed by incubation with protein A-Sepharose as described above. The bound proteins were eluted with sample buffer for 20 min at 60°C, run on SDS-PAGE, and analyzed by Western blot with anti-ASIC2a or anti-ASIC3 antibodies as described above.
Electrophysiology-Whole cell patch clamp recordings (at Ϫ70 mV) from CHO cells were performed with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) and acquired and analyzed with Pulse/ Pulsefit 8.30 (HEKA Electronics, Lambrecht, Germany) and Igor Pro 4.04 (WaveMetrics, Lake Oswego, OR) software. Experiments were performed at room temperature. Currents were filtered at 5 kHz and sampled at 2 or 0.2 kHz. Series resistance was compensated by at least 50%. Capacitive currents were compensated for and recorded for normalization of peak current amplitudes (reported as current densities). Micropipettes (2-5 megaohms) were filled with an internal solution containing 100 mM KCl, 10 mM EGTA, 40 mM HEPES, and 5 mM MgCl 2 , pH 7.4, with KOH. External solution contained 120 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 10 mM HEPES, and 10 mM MES. pH was adjusted with tetramethylammonium hydroxide, and osmolarity was adjusted with tetramethylammonium chloride. Extracellular solutions were changed within 20 ms using a computer-driven solenoid valve system (17). Acidic pH solutions were applied to cells from a resting pH of 7.4. Data are means Ϯ S.E. Kinetics of desensitization were fit with a single exponential equation, and time constants () were reported. Statistical differences were assessed by two-tailed Student's t test. h-stomatin cDNA was used for the electrophysiological co-expression experiments with ASIC3. However, m-stomatin co-expressed at 0.0625 g of cDNA/1 g of ASIC3 cDNA (1:8 ratio) caused identical inhibition of the current (data not shown). Thus, m-stomatin was used for the other co-expression experiments.

Stomatin
Interacts with ASICs-We tested whether stomatin associates with ASICs using two strategies. First, we asked whether ASICs and stomatin co-immunoprecipitate. ASIC1a, ASIC2a, or ASIC3 was co-expressed with stomatin (containing a FLAG epitope on the carboxyl terminus) in COS-7 cells. When stomatin was immunoprecipitated, all three ASICs were detected by immunoblotting (Fig. 1). However, we did not detect ASIC protein when expressed without stomatin. Thus, stomatin co-immunoprecipitates with each of the ASIC subunits.
Second, we tested whether stomatin and ASIC proteins colocalize by examining their immunostaining patterns. When transfected alone, stomatin localized in a punctate pattern around the nucleus and at the cell membrane ( Fig. 2A) in agreement with previous studies (43,48). ASIC1a, ASIC2a, and ASIC3 expressed alone showed a predominantly reticular pattern of staining. Co-expressing stomatin with individual ASIC proteins altered its cellular localization so it co-localized with the ASICs, showing a less punctate and more reticular pattern of staining (Fig. 2B). On the other hand, stomatin did not change its expression pattern or co-localize when co-expressed with the unrelated ion channel CFTR. Together, the co-immunoprecipitation and co-localization results suggest that stomatin associates with all three ASIC subunits.
Stomatin Modulates ASIC Current-The association of ASIC proteins with stomatin suggested that stomatin might alter ASIC channel function. To test this, we measured acid-evoked currents in CHO cells co-expressing stomatin with ASIC1a, ASIC2a, or ASIC3. When co-expressed with ASIC3, stomatin potently decreased current amplitude (Fig. 3, A and B). A 1:0.125 expression ratio (1 g of ASIC3:0.125 g of stomatin cDNA) eliminated the current, which was tested by applying pH 5 and 4 solutions. At a 1:0.0625 ratio (1 g of ASIC3:0.0625 g of stomatin), most cells produced small currents with properties that were otherwise similar to those from ASIC3 expressed alone. ASIC3 currents desensitized rapidly in the continued presence of acidic solution (see current returning to base line in Fig. 3A). This rate was unchanged by co-expression with stomatin at a 1:0.0625 ratio (Fig. 3C), and sensitivity to pH activation was not altered (Fig. 3D). However, given this low stomatin to ASIC3 cDNA expression ratio (1:16), these residual currents may reflect ASIC3 channels not bound to stomatin. As a negative control, we tested protein interacting with C kinase-1 (PICK1), a PDZ protein that binds to ASIC1a and ASIC2a but not ASIC3 (47,49). At a 1:1 cDNA ratio, PICK1 had no effect on ASIC3 current amplitude (Fig. 3B). Thus, the data indicate that stomatin potently inhibits the ASIC3 acid-evoked current.
Stomatin had a different effect upon ASIC2a. When co-expressed at a 1:1 cDNA ratio, stomatin did not decrease current amplitude; in fact, there was an insignificant trend toward increased current amplitude (Fig. 4, A and B). However, stomatin increased the rate of ASIC2a desensitization, which decreased the time constant (Fig. 4, A and C). Stomatin had little effect on pH sensitivity (Fig. 4D).  1. ASIC1a, ASIC2a, and ASIC3 co-immunoprecipitate with stomatin. COS-7 cells were transfected with ASIC1a, ASIC2a, or ASIC3 and stomatin-FLAG (Stomatin-FL) alone or together at 1:1 ratios and immunoprecipitated with anti-FLAG antibody. The immunoprecipitates (IP) were resolved by SDS-PAGE and probed with either ASIC1a, ASIC2a, or ASIC3 antibodies as indicated. 10 g of total lysate from the same samples were also run on SDS-PAGE and probed with the indicated antibodies.
In contrast to ASIC2a and ASIC3, stomatin did not alter the ASIC1a current. At a 1:1 or 1:10 cDNA ratio (ASIC1a:stomatin), stomatin did not change current amplitude (Fig. 5, A and  B), desensitization kinetics (Fig. 5C), or pH sensitivity (Fig.  5D). Thus, although stomatin interacted with each of the ASIC subunits, it had differential effects on ASIC channel function, depending on the subunit composition.
ASIC subunits can function as homomultimers or they can heteromultimerize with one another (5, 46, 50 -53). Therefore, we asked whether stomatin alters currents generated by channels composed of more than one ASIC subunit. Because stomatin produced the most dramatic effect on ASIC3, we focused on combinations that included this subunit. The fast desensitization kinetics (Fig. 6, A and C) confirm that co-expression of ASIC subunits resulted in the formation of heteromultimeric channels because heteromeric channels desensitize at a faster rate than homomeric channels (46,53,54). When we co-expressed ASIC3 with ASIC1a or ASIC2a, stomatin did not alter current amplitude (Fig. 6, A and B) or pH sensitivity (Fig. 6D). However, stomatin produced small but significant increases in the desensitization rates of both heteromultimeric channels (Fig. 6C).
Stomatin Does Not Alter ASIC Cell Surface Expression-Stomatin could decrease ASIC3 current amplitude by altering its expression at the cell surface or altering its gating. To distinguish between these possibilities, we quantitated ASIC3 expressed at the cell surface by biotinylating cell surface proteins. Following biotinylation, we precipitated proteins with avidin and detected ASIC3 by Western blot analysis. Stomatin did not alter the amount of biotinylated ASIC3, indicating that stomatin did not change ASIC3 surface expression (Fig. 7). Moreover, stomatin co-expression did not alter the total amount of ASIC3 protein (Fig. 1). These results suggest that stomatin reduced ASIC3 current by altering its gating rather than its expression at the cell surface. DISCUSSION The data indicate that stomatin binds to ASIC1a, -2a, and -3 subunits. Moreover, stomatin altered the function of specific ASIC channels. However, the functional effects of stomatin were strikingly different depending on the subunit. Stomatin had the most prominent effect on ASIC3, potently reducing acid-evoked current. In contrast, stomatin did not significantly decrease current amplitude generated by other ASIC channels; rather, it accelerated the rate of desensitization of ASIC2a homomeric channels as well as ASIC1aϩ3 and ASIC2aϩ3 heteromeric channels. Although stomatin interacted with ASIC1a protein, it did not alter ASIC1 current. Thus, binding alone is not sufficient to alter function. Our data provide important clues for understanding how stomatin modulates the function of ASICs. Stomatin reduced ASIC3 current but did not alter ASIC3 protein level or surface expression. Thus, the data suggest that stomatin inhibits ASIC3 by altering its gating. Also consistent with this mechanism, stomatin increased the desensitization rate of ASIC2a homomeric channels and ASIC1aϩ3 and ASIC2aϩ3 heteromeric channels. Thus, stomatin is an accessory subunit that interacts with ASIC channels at the cell surface to modulate gating.
What is the mechanism whereby stomatin alters ASIC gat-ing? The capacity of stomatin to bind to ASICs and accelerate channel desensitization is reminiscent of the role of accessory ␤ subunits to facilitate inactivation of voltage-gated channels (55). In a similar manner, stomatin may function as an accessory subunit that directly modulates gating. Alternatively, stomatin could alter ASIC gating indirectly through the interaction of other proteins. Stomatin is highly expressed in mammalian lipid rafts (detergent-resistant, cholesterol-and sphingomyelin-rich regions of the membrane), which are important for sequestering proteins into complexes and localizing signal transduction processes (41,56,57). Recent experiments in C. elegans demonstrate that the stomatin-related proteins UNC-1 and UNC-24, along with the DEG/ENaC subunit UNC-8, are located in lipid rafts. Moreover, genetic disruption of UNC-1 abolished UNC-8 expression in lipid rafts (58). Stomatin might localize ASICs to specific membrane microdomains to facilitate interactions with other regulatory and signaling proteins. ASIC activity can be modulated by several endogenous ligands and signaling molecules (45, 59 -63). However, less is understood regarding intracellular proteins that directly bind to and affect ASIC function. Yeast two-hybrid screening recently identified two proteins that interact with ASICs. In contrast to stomatin, both proteins potentiate ASIC currents. PICK1 interacts with ASIC1 and ASIC2 (47,49) and potentiates current via ASIC2a when stimulated by protein kinase C (64). Channel-interacting PDZ domain-containing protein (CIPP) binds to the COOH terminus of ASIC3 and increases current (65). In addition, the ASIC1a and -2a heteromeric current is potentiated by co-expression of CFTR (66).
Earlier studies suggested that ASICs might share a conserved function with their homologs in C. elegans, the degenerins. Mutations of the genes encoding DEG channel subunits MEC-4 and MEC-10 disrupt light touch in specialized mechanosensory neurons (24 -26). Similarly, targeted deletion of ASIC2 and ASIC3 genes produced altered single fiber sensory nerve responses to cutaneous mechanical stimuli in mice (14,16). Our data identifies another conserved feature; both DEG and ASIC channels are modulated by stomatin-related proteins. Similar to our data, the stomatin homolog MEC-2 co-immunoprecipitated with DEG channels, modulated current, and did not alter channel surface expression when coexpressed in Xenopus oocytes (27). However, it is interesting that MEC-2 and stomatin produced opposite effects on current. MEC-2 (and to a lesser extent stomatin) increased constitutive current generated by DEG channels (27), whereas stomatin FIG. 6. Stomatin alters heteromeric ASIC currents. CHO cells were transfected with ASIC3 and ASIC1a or ASIC2a (0.5 g of cDNA each) and DsRed or stomatin (1 g of cDNA). A, representative currents evoked by application of pH 6 to CHO cells expressing indicated ASIC subunits with or without stomatin (stom). B, mean current density of pH 5-evoked currents (n ϭ 11-15). pA/pF, picoampere/picofarad. C, the time constants of desensitization of pH 6-evoked currents (n ϭ 12-23; *, p Ͻ 0.05 compared with respective ASIC heteromer without stomatin). D, pH dose-response data (n Ն 8). Currents were normalized to amplitudes of currents evoked by pH 5 .   FIG. 7. Stomatin does not alter cell surface expression of ASIC3. COS-7 cells were transfected with ASIC3 alone or with stomatin. After 48 h, cells were incubated with sulfo-N-hydroxysuccinimidobiotin at 4°C and solubilized in radioimmune precipitation assay buffer. Biotinylated protein was isolated from 500 g of cell lysate by incubation with neutrAvidin beads. ASIC3 protein was immunoprecipitated from 500 g of the same cell lysates by ASIC3-specific antibody as indicated. Biotinylated proteins and immunoprecipitated proteins were separated by SDS-PAGE and blotted with ASIC3-specific antibody. inhibited ASIC3 such that H ϩ no longer activated the channel. Given the purported role of DEG channels and MEC-2 in mechanosensation in C. elegans, it is intriguing to speculate that the interaction of stomatin with ASIC3 changes the channel from a pH sensor to a mechanosensor.
Modulation of ASICs has important implications. As H ϩgated channels, they are poised to sense acidosis associated with such painful conditions as inflammation, ischemia, or malignancy; thus modulation of ASICs by stomatin might be an important means to regulate nociception. ASICs also function in sensation of mechanical stimuli. Thus, stomatin might play an important role in mammalian mechanosensation, analogous to the purported role of their homologs in C. elegans.