Abstract
Acid-sensing ion channel (ASIC) subunits associate to form homomeric or heteromeric proton-gated ion channels in neurons throughout the nervous system. The ASIC1a subunit plays an important role in establishing the kinetics of proton-gated currents in the CNS, and activation of ASIC1a homomeric channels induces neuronal death after local acidosis that accompanies cerebral ischemia. The ASIC2b subunit is expressed in the brain in a pattern that overlaps ASIC1a, yet the contribution of ASIC2b has remained elusive. We find that coexpression of ASIC2b with ASIC1a in Xenopus oocytes results in novel proton-gated currents with properties distinct from ASIC1a homomeric channels. In particular, ASIC2b/1a heteromeric channels are inhibited by the nonselective potassium channel blockers tetraethylammonium and barium. In addition, steady-state desensitization is induced at more basic pH values, and Big Dynorphin sensitivity is enhanced in these unique heteromeric channels. Cultured hippocampal neurons show proton-gated currents consistent with ASIC2b contribution, and these currents are lacking in neurons from mice with an ACCN1 (ASIC2) gene disruption. Finally, we find that these ASIC2b/1a heteromeric channels contribute to acidosis-induced neuronal death. Together, our results show that ASIC2b confers unique properties to heteromeric channels in central neurons. Furthermore, these data indicate that ASIC2, like ASIC1, plays a role in acidosis-induced neuronal death and implicate the ASIC2b/1a subtype as a novel pharmacological target to prevent neuronal injury after stroke.
Introduction
The acid-sensing ion channels (ASICs) are voltage-insensitive cation channels activated by extracellular acidification (Waldmann et al., 1997; Waldmann, 2001). There are four ASIC genes (ACCN1-4) that encode at least six individual ASIC subunits with distinct and overlapping patterns of expression (Wemmie et al., 2006; Xiong et al., 2008). Individual ASIC subunits associate into homomeric channels composed of three identical ASIC subunits and heteromeric channels composed of different ASIC subunits (Bassilana et al., 1997; Jasti et al., 2007). The ASIC1a subunit is expressed in neurons throughout the brain and is required for high-affinity proton sensing in central neurons (Wemmie et al., 2002). ASIC1a is also required for normal fear-related behaviors, learning, memory, and seizure termination (Wemmie et al., 2002, 2003; Coryell et al., 2008; Ziemann et al., 2008, 2009). In addition, ASIC1a plays a causative role in neuronal damage induced by prolonged acidosis (Xiong et al., 2004; Gao et al., 2005; Sherwood and Askwith, 2009). Inhibiting ASIC1a activity is neuroprotective in mouse models of stroke, and ASIC1a has also been implicated in the neuronal pathogenesis of multiple sclerosis, Huntington's disease, and Parkinson's disease (Xiong et al., 2004; Friese et al., 2007; Pignataro et al., 2007; Arias et al., 2008; Wong et al., 2008). Together, these results indicate that ASIC1a plays important roles in neuronal signaling and death in the brain. Yet, the contribution of other ASIC subunits to these processes has not been fully elucidated.
Two ASIC2 subunits are abundantly expressed in central neurons in a pattern that overlaps with ASIC1a (Price et al., 1996; Lingueglia et al., 1997). ASIC2a and ASIC2b are transcript variants distinct in the first one-third of the coding region (Lingueglia et al., 1997). ASIC2a forms homomeric channels, but their role in vivo is uncertain as severe acidification is required for their activation. ASIC2a also forms heteromeric channels with ASIC1a (ASIC2a/1a) that display unique characteristics compared with ASIC1a and ASIC2a homomers, including differences in apparent proton sensitivity and inactivation kinetics (Bassilana et al., 1997; Baron et al., 2002a; Askwith et al., 2004; Chu et al., 2004). ASIC2a/1a channels are abundant in central neurons and are responsible for a subset of proton-gated currents (Baron et al., 2002a; Askwith et al., 2004; Gao et al., 2004; Wu et al., 2004; Chu et al., 2006; Jiang et al., 2009). The other ASIC2 subunit, ASIC2b, does not produce proton-gated current when expressed alone, and it is unknown whether ASIC2b/1a channels have unique properties (Coscoy et al., 1999; Hesselager et al., 2004). Thus, the role of ASIC2b in central neurons is unknown.
Here we show that coexpression of ASIC2b with ASIC1a results in ASICs that display novel properties that are distinct from ASIC1a homomeric and ASIC2a/1a heteromeric channels. We find proton-gated currents in cultured neurons with properties similar to ASIC2b/1a channels, and these currents contribute to acid-induced cell death. Together, these results directly implicate ASIC2 in neuronal injury and suggest that the unique ASIC2b/1a channel subtype plays a role in ASIC1a-dependent processes.
Materials and Methods
Recombinant DNA expression in Xenopus oocytes.
Unfertilized oocytes were harvested from female Xenopus laevis purchased from Xenopus I, using standard procedures (Sherwood and Askwith, 2009). One to three hours after isolation, oocyte nuclei were injected with the pMT3 expression plasmid containing mouse ASIC cDNA (at ∼5 ng of a 100 ng/μl stock) using a PV820 Pneumatic Picopump (World Precision Instruments). Oocytes were incubated at 18°C for 18–72 h before experiments were performed.
Two-electrode voltage clamp on Xenopus oocytes.
Whole-cell macroscopic current was measured using the two-electrode voltage-clamp technique at a holding potential of −60 mV unless otherwise noted (Sherwood and Askwith, 2009). Electrodes (∼2 MΩ) were pulled using a Sutter P-97 micropipette puller (Sutter Instrument Company) and filled with 3 m KCl. Data were acquired using an Oocyte Clamp OC-725 Amplifier (Warner Instruments), an Axon Digidata 1200 digitizer, and pCLAMP-8 software (Molecular Devices). Most experiments were done using Frog Ringer's solution (116 mm NaCl, 2 mm KCl, 5 mm HEPES, 5 mm MES, 2 mm CaCl2, 1 mm MgCl2) with a pH adjusted to the indicated levels using 1N NaOH or 1N HCl. Because ASIC2b/1a heteromeric channels display steady-state desensitization at pH 7.4, basal pH for all experiments was pH 7.9 (unless stated otherwise). Barium, tetraethylammonium (TEA), zinc, N, N, N′, N′-tetrakis(2-pyridylmethyl)ethylenediamine, and psalmotoxin 1 (PcTX1) experiments were done with Ringer's solution that also contained BaCl2, tetraethylammonium chloride, ZnCl2, TPEN, or purified PcTX1 peptide (Peptides International) at the concentrations indicated in the figures and figure legends. Ion permeability studies were done with modified Frog Ringer's solution containing 0.4 mm CaCl2, 1 mm MgCl2, 5 mm HEPES, 5 mm MES, and 116 mm NaCl, 116 mm LiCl, or 116 mm KCl, as indicated, with a pH adjusted to the indicated levels using 1 m NaOH, KOH, or LiOH as appropriate. Oocyte recordings were done in a modified RC-Z3 250 μl oocyte recording chamber (Warner Instruments). The solution exchange rate in the recording chamber was ∼1 ml/s. ASIC current properties (pH-dependent activation and steady-state desensitization) were assessed as described previously (Sherwood and Askwith, 2009). Proton-gated current evoked in experimental conditions was always flanked by saturating pH applications to evoke maximal current (pH 4.5 unless otherwise noted). For quantification, peak current amplitude in experimental conditions was normalized to the average of the flanking maximal current amplitudes to minimize the impact of potential tachyphylaxis of proton-gated current. Big Dynorphin was synthesized by EZBiolab. Purified PCTX1 was purchased from Peptides International.
Oocyte data analysis.
Data were analyzed using the Axon Clampfit 9.0 software. To measure the rate of channel desensitization, the decay phase of current was fitted to a single-exponential equation: I = k0 + k1 × e−t/τd, and the tau of inactivation (τinact.) was calculated. Acid-induced sustained current was measured at the plateau phase of decay at least 20 s after application of pH 5.0 and normalized to peak current amplitude. For pH-dependent activation and steady-state desensitization, the pH0.5 was calculated by fitting the data from individual oocytes using the following equation: I/IpH max = 1/{1 + (EC50/[H+])n} = 1/{1 + 10n (pH − pH0.5)}, where n is the Hill coefficient and EC50 and pH0.5 are the proton concentration and pH-yielding half of the saturating peak current amplitude (IpH max). The pHmax was determined by the application of pH 5.0 from a holding pH of 7.9 (unless otherwise noted). The EC50 for Big Dynorphin was calculated by fitting the data from individual oocytes using the following equation: I/Imax = 1/{1 + (EC50/[peptide])n }, where n is the Hill coefficient and EC50 is the peptide concentration inducing half of the saturating peptide effect (Imax). Relative cation permeabilities for ASICs expressed in oocytes were determined using the following equation: ΔErev = Erev(x) − Erev(Na+) = (RT/F)ln{Px[X]o/PNa+[Na+]o}, where Erev(x) and Erev(Na+) are the experimentally determined reversal potentials of cation “x” and Na+, respectively; [X]o and [Na+]o are the extracellular concentrations of cation x and Na+; and the constants R, T, and F have their standard meanings. Statistical analysis was done with the two-tailed Student's t test (paired or unpaired data as indicated) or ANOVA with post hoc Bonferroni's multiple comparison test as indicated in the figure legends. A p value <0.05 was considered significant.
Chinese hamster ovary cell transfection and calcium permeability experiments.
The whole-cell patch-clamp technique was used to measure calcium permeability of acid-evoked current in Chinese hamster ovary (CHO) cells transiently transfected with green fluorescent protein (GFP), mouse ASIC1a, and one of the following: mouse ASIC2b, mouse ASIC2a, or vector only (Hoagland et al., 2010). Briefy, trypsinized CHO cells (∼107 cells) were suspended in 0.4 ml of electroporation solution (120 mm KCl, 25 mm HEPES buffer, 10 mm K2HPO4, 10 mm KH2PO4, 2 mm MgCl2, 0.15 mm CaCl2, 5 mm EGTA, and 2 mm MgATP, pH 7.6) and mixed with 20 μg of total plasmid DNA containing ASIC expression vectors (or empty pcDNA3.1 vector as indicated) and pEGFP-C1 vector (Clontech) at a 1:1 ratio. Cells were electroporated with the Gene Pulser Xcell system (Bio-Rad) and plated at a density of 35 cells/mm2 onto 10 mm coverslips in a 35 mm culture dish. Cells were used for patch clamping 2–3 d after transfection. Transfected cells were identified by expression of GFP. Data were collected at 5 kHz using an Axopatch 200B amplifier, Digidata 1322A, and Clampex 9.0 (Molecular Devices). Extracellular solutions contained either 160 mm NaCl or 80 mm CaCl2 with 20 mm HEPES buffer and 10 mm glucose. The pH was adjusted with NaOH or Ca(OH)2 as appropriate. The intracellular pipette solution (pH 7.4) contained 160 mm NaCl, 10 mm HEPES, and 10 mm EGTA. The pipette solution was allowed to dialyze with cytosolic contents for at least 5 min after whole-cell access was gained before ion permeability was measured. The membrane potential was held constant at −50 mV between voltage-ramp protocols. ASIC current was evoked by application of acidic extracellular solutions (pH 5.3), and a voltage ramp (−150 to 50 mV) was initiated immediately, before the current decayed into the plateau phase. To isolate background conductance, an identical voltage ramp was performed before and after ASIC activation at holding pH 7.6. Mean background conductance was subtracted from conductance after acid application, and data were fitted to a linear equation to determine the reversal potential of ASIC current. Junction potentials of sodium and calcium containing extracellular solutions (−0.4 and +3.4 mV, respectively) were corrected post hoc. PNa/PCa was calculated from the change in reversal potentials when Na+ was replaced with Ca2+ in the extracellular solution using the following equations: ΔErev = Erev,Na − Erev,Ca = (RT/F) ln(PNa[Na+]o/4PCa′[Ca2+]o) and PCa′ = PCa/(1 + eEF/RT), where the universal gas constant (R), Faraday constant (F), and temperature (T) have their standard meanings. After the voltage ramp, zinc modulation (0.3 mm ZnCl2) of proton-gated current was measured to ensure appropriate heteromeric channels were expressed in the recorded CHO cells.
Culture of cortical and hippocampal neurons.
Primary hippocampal and cortical neuron cultures were prepared using previously published methods with some minor alterations (Askwith et al., 2004). Briefly, hippocampi and cortices were dissected from postnatal day 0–1 pups, freed from extraneous tissue, and cut into pieces. Dissected tissue was transferred into Leibovitz's L-15 medium containing 0.25 mg/ml bovine serum albumin and 0.38 mg/ml papain and incubated for 15 min at 37°C with 95% O2–5% CO2 gently blown over the surface of the medium. After incubation, the dissected tissue was washed three times with mouse M5-5 medium (Earle's minimal essential medium with 5% fetal bovine serum, 5% horse serum, 0.4 mm l-glutamine, 16.7 mm glucose, 5000 U/l penicillin, 50 mg/l streptomycin, 2.5 mg/l insulin, 16 nm selenite, and 1.4 mg/l transferrin) and triturated. Dissociated cells were then centrifuged for 4.5 min at 730 rpm, and M5–5 medium was aspirated. Cells were resuspended in supplemented Neurobasal-A medium (1% B27 supplement containing anti-oxidants, 1% B27supplement minus anti-oxidants, 0.5 mm l-glutamine, 0.5 mg/ml gentamycin, 2.5 mg/l insulin, 16 nm selenite, and 1.4 mg/l transferrin). Cells were plated in 24-well plates containing 10 mm poly-d-lysine-coated glass coverslips at a density of 5 × 104 cells per well. After 48–72 h, 10 μm cytosine β-d-arabinofuranoside was added. Neurons were maintained at 37°C with 5% CO2 for 14–21 d before experiments were performed.
Whole-cell patch clamp on primary neurons.
To record ASIC-dependent current, neurons were perfused with extracellular solution at varying pH values. The extracellular solution contained 140 mm NaCl, 5.4 mm KCl, 10 mm HEPES buffer, 2 mm CaCl2, 1 mm MgCl2, 5.55 mm glucose, 10 mm MES buffer, 10 μm 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), 50 μm d-2-amino-5-phosphonovaleric acid, 30 μm bicuculline, and 500 nm tetrodotoxin. Certain extracellular solutions also contained 10 mm BaCl2, 0.3 mm ZnCl2, or 50 nm PcTX1 (purified peptide) as indicated. The pH was adjusted with 1N NaOH. Patch electrodes (2–4 MΩ) were fire polished with a microforge (Narishige). The intracellular pipette solution contained 121 mm KCl, 10 mm NaCl, 2 mm MgCl2, 5 mm EGTA, 10 mm HEPES, 2 mm Mg-ATP, and 300 μm Na3GTP, pH 7.25. The membrane potential was held constant at −70 mV. Data were collected at 5 kHz using an Axopatch 200B amplifier, Digidata 1322A, and Clampex 9 (Molecular Devices). Neurons were superfused continuously with the extracellular solution from gravity-fed perfusion pipes at a flow rate of about 1 ml/min. Perfusion pipes were placed 250–300 μm away from cells, and flow was directed toward the recorded cells to ensure fast solution exchange. For normalization, maximal acid-gated current was evoked by the exogenous application of pH 6.0 from a holding pH of 7.9. Typically, three to four applications of pH 6.0 solution were required before acid-evoked peak current amplitude stabilized and experimental conditions were tested. Peak current amplitude was normalized to the average peak current amplitude of flanking pH 6.0-evoked current.
RNAi design and analysis.
Anti-mouse ASIC2b small hairpin RNA (shRNA) was designed using Oligoengine2.0 software (Oligoengine). The mouse ASIC2b coding sequence was analyzed, and a 19-nucleotide (19-nt) sequence, tgaccgtgtgcaacaacaa, corresponding to the ASIC2b-specfic splice variant sequence (starting at nucleotide position 397) was selected. Two complimentary 60-mer DNA oligonucleotides containing BgIII and HINDIII restriction cloning sites, ASIC2b 19-nt sense and antisense sequence, and a hairpin-promoting element were purchased from Oligoengine. The 60-mer strands were annealed together and cloned in a pSUPERgfp/neo expression plasmid according to the manufacturer's instructions (Oligoengine). Anti-ASIC2b shRNA-expressing plasmid was prepared from bacteria using Midiprep kits (Qiagen). All inserts were sequenced by the Plant-Microbe Genomics facility at The Ohio State University (Columbus, OH). Either the anti-ASIC2b shRNA-expressing plasmid or pSUPERgfp/neo was transfected into CHO cells along with mouse ASIC2b, ASIC2a, or ASIC1a (at a 3:2 ratio). Cells were allowed to grow for 2 d after transfection and harvested for quantitative RNA analysis.
RNA was isolated from transfected CHO cells using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. RNA was treated with the DNA-free kit (Ambion) to destroy possible contaminating genomic DNA. cDNA was then made from DNase-treated RNA using reverse transcription and SuperScript III Reverse Transcriptase (Invitrogen), Oligo(dT)12–18, 10 mm dNTP Mix, and RNaseOUT Recombinant RNase Inhibitor (Invitrogen). For each sample, 50 ng of RNA was used in the cDNA synthesis reaction, and a “no-RT” reaction lacking reverse transcriptase was performed to produce the template for the no-RT control experiments. Real-time PCR was performed using the Power SYBR Green PCR Master Mix (Applied Biosystems) on a StepOnePlus Real Time PCR System using StepOne software (Applied Biosystems). Reactions were set up in triplicate. Per well, 20 μl reactions consisted of 10 μl of 2× Power SYBR Green Master Mix, 7.8 μl of DEPC-treated DNase, RNase-free water (Invitrogen), 0.6 μl of each forward and reverse primer at 10 μm concentration (Integrated DNA Technologies), and 1 μl of diluted cDNA (1 part cDNA to 5 parts total volume). Final concentrations of reaction components were 1× Power SYBR Green Master Mix and 300 nm of each primer. Primers used include the following: mouse β-actin forward 5′-tacagcttcaccaccacagc-3′ and reverse 5′-tctccagggaggaagaggat-3′; mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward 5′-acccagaagactgtggatgg-3′ and reverse 5′-ggatgcagggatgatgttct-3′; mouse ASIC2b forward 5′-tgtcaccgtgtgcaacaac-3′ and reverse 5′-cagcagcagtcctagccagt-3′; mouse ASIC2a forward 5′-aaaccgaagcagttcagcat-3′ and reverse 5′-gccccttgaacttgcagtag-3′; mouse ASIC1a forward 5′-ctgtaccatgctggggaact-3′ and reverse 5′-gctgcttttcatcagccatc-3′. Cycling conditions consisted of an initial activation at 95°C for 10 min, followed by 40 cycles of two-step PCR; 95°C for 15 s, 60°C for 1 min, followed by a melt curve analysis. Cycle threshold was taken at 0.2 ΔRn. Melt-curve analysis indicated single products. All reactions were performed in triplicate along with “no-cDNA” controls (in which water was added instead of cDNA). A no-RT control reaction was always done with each primer set on each batch of cDNA. Efficiency measurements were performed with all primer sets, and primers displayed efficiencies calculated to be within 0.1 of each other using the equation e = 10(−1/slope) − 1. Data were normalized to β-actin expression to determine the ΔCt according to the manufacturer's instructions. Relative RNA expression was determined using the ΔΔCt method. To asses the effectiveness of anti-ASIC2b shRNA, CHO cells were cotransfected with either anti-ASIC2b shRNA-expressing plasmid or empty pSUPER plasmid and one of the following: mASIC1a, mASIC2a, or mASIC2b plasmid. Compared with pSUPER control, the anti-ASIC2b shRNA-expressing plasmid reduced ASIC2b RNA levels to 35.9 ± 3.8% (p = 0.002 compared with GAPDH) but did not reduce GAPDH, mASIC1a, or mASIC2a RNA levels (GAPDH RNA, 113.6 ± 14.5% vs control; ASIC1a RNA, 140 ± 15%, p = 0.2 compared with GAPDH; ASIC2a RNA, 135.5 ± 5.9%, p = 0.08 compared with GAPDH; n = 4 transfections for all groups).
Transfection of primary neurons.
Wild-type hippocampal neurons were transfected with empty pSUPERgfp/neo vector (control) or anti-ASIC2b shRNA-containing pSUPERgfp/neo vector (see above) at the time of culture, before plating. Briefly, dissociated hippocampal cells were split into two groups from the same preparation (3–4 million cells each group), suspended in 100 μl of Nucleofector solution from the Basic Nucleofector kit (Amaxa Biosystems), and mixed with 3 μg of plasmid DNA. Hippocampal cells were electroporated with the Nucleofector II (program O-05; Amaxa) and plated on poly-d-lysine coverslips in 24-well dishes at a density of 500,000–1,000,000 cells per well in Neurobasal-A medium (see above). Cytosine β-d-arabinofuranoside (10 μm) was added as before. Neurons were used for patch clamping (single cells) or qRT-PCR analysis (whole coverslips) from 14 to 20 d in culture. Transfected cells were identified for patching by GFP fluorescence and selected for recording using a TE2000-S epifluorescent microscope (Nikon).
Acid-induced neuronal damage assays.
At 14–17 d in culture, cortical neurons were randomly divided into designated experimental groups (see Fig. 8). Neurobasal medium was removed before washing cells two times with ECF solution (140 mm NaCl, 5.4 mm KCl, 25 mm HEPES buffer, 20 mm glucose, 1.3 mm CaCl2, 1.0 mm MgCl2) at pH 7.6 (pH adjusted with 1N NaOH). Cells were then washed with pH 7.6 ECF inhibitor solution containing 10 μm Dizocilpine (MK-801), 20 μm CNQX, 5 μm nimodipine, and 500 nm tetrodotoxin. Specific acidosis interventions were performed as described in the presence or absence of 10 mm Ba2+ using ECF with inhibitors (Xiong et al., 2004; Gao et al., 2005; Sherwood and Askwith, 2009). Within each culture, three to twelve different wells were used for a given intervention. After acidosis interventions, cells were washed two times with pH 7.6 ECF without inhibitors, and fresh Neurobasal culture medium was returned to cultures. Neurons were allowed to recover for 18–24 h at 37°C with 5% CO2. Acid-induced damage was assessed after the recovery period using the Vybrant MTT cell proliferation assay kit (Invitrogen) as per the manufacturer's instructions, scaled for use in 24-well plate cultures. Briefly, at 24 h after intervention, cells were incubated in fresh Neurobasal culture medium containing 1.2 mm 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) for 4 h at 37°C with 5% CO2. After MTT treatment, medium was removed and cells were treated with DMSO (10 min at 37°C with 5% CO2) to solubilize formazan produced as a result of MTT metabolism. DMSO extract from each well (100 μl) was collected on a 96-well plate, and formazan content was determined by reading absorbance at 540 nm. To determine background absorbance, 100 μl samples from empty wells treated with Neurobasal-MTT medium and DMSO were also read. The effect of acid treatment on neurons was measured by subtracting average background absorbance from individual well data and normalizing individual well data to the average absorbance from the “No Acid” control group in the absence of barium. Data from individual wells were then averaged so that one neuronal culture yielded one n for a given acidosis intervention. “% Acid-Induced Damage” is the percentage decrease in absorbance of “Acid”-treated groups compared with No Acid groups in the presence or absence of barium.
Results
ASIC2b/1a channels are sensitive to barium and TEA
To determine whether ASIC2b affects the ASIC1a-dependent current, we analyzed proton-gated currents in Xenopus oocytes coexpressing mouse ASIC2b with mouse ASIC1a and compared them with currents elicited from oocytes expressing either subunit alone. Oocytes expressing ASIC1a showed proton-gated currents when the extracellular pH was rapidly decreased (Fig. 1A). No proton-gated currents were observed when ASIC2b was expressed alone, and expression of ASIC1a with ASIC2b resulted in proton-gated currents that appeared grossly similar to ASIC1a homomeric channels (Fig. 1A). To identify compounds that could distinguish ASIC2b/1a heteromeric channels from ASIC1a homomeric channels, we tested TEA. Previous studies showed that ASIC2a channels with a mutation within the pore (G430V) were inhibited by the classic nonspecific inhibitor of potassium channels, TEA (Adams et al., 1999). Since TEA inhibits mutant ASIC2a channels, we reasoned that ASIC2a- or ASIC2b-containing heteromeric channels may also be sensitive to TEA. We found that proton-gated currents in oocytes expressing ASIC1a alone were not inhibited by 10 mm TEA (Fig. 1B,C). However, proton-gated currents in cells coexpressing ASIC2b or ASIC2a with ASIC1a were inhibited by TEA (Fig. 1B,C). Both ASIC2b/1a- and ASIC2a/1a-expressing cells showed equivalent concentration-dependent TEA inhibition at low millimolar concentrations, whereas ASIC1a showed inhibition only at higher concentrations of TEA (50 mm) (Fig. 1D). Although TEA does not distinguish between ASIC2a- or ASIC2b-containing heteromeric channels, it does indicate that ASIC2b/1a heteromeric channels produce proton-activated currents on the plasma membrane that display pharmacologically distinct properties from ASIC1a homomeric channels.
To further determine whether ASIC2b/1a currents could be distinguished from ASIC1a or ASIC2a/1a currents, we tested barium, another nonspecific inhibitor of potassium channels (Hurst et al., 1995; Harris et al., 1998; Jiang and MacKinnon, 2000). ASIC1a was unaffected by 10 mm barium, whereas current from ASIC2b/1a heteromeric channels was appreciably inhibited (Fig. 1B,C). Interestingly, barium had a significantly greater effect on ASIC2b/1a current compared with either ASIC1a homomeric or ASIC2a/1a heteromeric current (percentage inhibition was 2.2 ± 2.0% for ASIC1a, 8.4 ± 1.2% for ASIC2a/1a, and 23.9 ± 3.1% for ASIC2b/1a; p < 0.01 compared with ASIC1a, p < 0.05 compared with ASIC2a/1a, one-way ANOVA) (Fig. 1C). In addition, the concentration dependence of barium inhibition was significantly lower for ASIC2b/1a compared with ASIC1a or ASIC2a/1a, although all channels were sensitive to barium at high concentrations (Fig. 1E).
Extracellular TEA and barium inhibition of potassium channels occurs through a pore-blocking mechanism (Eaton and Brodwick, 1980; Taglialatela et al., 1991). Such ion channel pore blockers often show voltage-dependent inhibition, with inhibition decreasing as current nears the reversal potential, and no inhibition with outward current. TEA and barium inhibition of ASIC2b/1a currents was also voltage dependent as the effect of either compound was abolished at positive holding potentials (Fig. 1F,G). These results are consistent with a pore-blocking mechanism for ASIC2b/1a inhibition similar to that observed with extracellular TEA and barium inhibition of other ion channels.
ASIC2b/1a channels are inhibited by PcTX1
PcTX1 is a tarantula venom peptide that inhibits ASIC1a homomeric channels and has been used to prevent activation in vitro and in vivo (Escoubas et al., 2000; Xiong et al., 2004; Mazzuca et al., 2007; Pignataro et al., 2007; Coryell et al., 2008, 2009; Ziemann et al., 2008, 2009). PcTX1 is thought to be an ASIC1a-specific inhibitor and does not inhibit heteromeric channels, such as ASIC2a/1a (Escoubas et al., 2000, 2003; Chen et al., 2006; Salinas et al., 2006). To determine whether PcTX1 acts on ASIC2b/1a channels, oocytes expressing ASIC1a, ASIC2b/1a, or ASIC2a/1a were treated with PcTX1 before activation. As expected, ASIC1a homomers were inhibited by 0.2 μm PcTX1 whereas ASIC2a/1a heteromers were not (Fig. 2A,C). Surprisingly, oocytes expressing ASIC2b/1a were also sensitive to PcTX1 (Fig. 2B). In fact, PcTX1 inhibited ASIC2b/1a current by 82.0 ± 3.2%, which was significantly greater than its effect on ASIC1a current (52.6 ± 5.7%; p < 0.001, one-way ANOVA) (Fig. 2D). However, the IC50 for PcTX1 inhibition of ASIC1a homomeric and ASIC2b/1a heteromeric channels was not statistically different, suggesting both channels have equivalent affinities for the peptide (Fig. 2E). Because PcTX1 inhibits ASIC1a by enhancing steady-state desensitization and ASIC2b/1a channels showed greater inhibition with 0.2 μm PcTX1, we reasoned that ASIC2b/1a channels may be more prone to undergo steady-state desensitization compared with ASIC1a (Chen et al., 2005).
ASIC2b enhances steady-state desensitization
Steady-state desensitization of ASIC is induced when the channels are exposed to mild decreases in pH that are not sufficient for maximal channel activation (Babini et al., 2002). PcTX1 acts by increasing the apparent proton sensitivity of steady-state desensitization and causing ASIC1a to desensitize at basal pH 7.4, which normally does not induce desensitization (Chen et al., 2005). Thus, the increased PcTX1-mediated inhibition of ASIC2b/1a could be indicative of intrinsic differences in steady-state desensitization. Steady-state desensitization is experimentally observed as a decrease in pH 5.0-evoked peak current amplitude directly after exposure (2 min) to a mildly acidic “conditioning” pH (Fig. 3A). For example, when oocytes expressing ASIC1a alone are conditioned at pH 7.4, steady-state desensitization of pH 5.0-evoked current is not induced (I/Imax = 98.6 ± 9.7%; p = 0.89 compared with conditioning pH 7.9). But exposing cells to pH 7.2 induces steady-state desensitization and reduces pH 5.0-evoked current to 62.4 ± 8.8% of maximal current (p = 0.01 compared with conditioning pH 7.9) (Fig. 3B). In oocytes expressing ASIC2b/1a, enhanced steady-state desensitization was observed such that pH 5.0-evoked current was reduced even at a conditioning pH of 7.4 (78.6 ± 4.0%; p < 0.001 compared with conditioning pH 7.9) (Fig. 3A). A more detailed analysis revealed that the pH50 of steady-state desensitization was 7.28 ± 0.02 for ASIC2b/1a (n = 10) and 7.18 ± 0.1 for ASIC1a (n = 4; p = 0.01) (Fig. 3B). Thus, the presence of ASIC2b caused a shift in the pH dependence of steady-state desensitization such that less acidic pH values induced desensitization.
We also examined other intrinsic properties of ASIC2b/1a heteromeric channels. Acid-dependent sustained current and inactivation rate after activation were not different in cells expressing ASIC2b/1a versus ASIC1a alone [τinactivation: ASIC1a, 3.16 ± 0.18 s (n = 9); ASIC2b/1a, 2.60 ± 0.43 s (n = 11; p = 0.283); percentage sustained current: ASIC1a, 0.44 ± 0.11% (n = 11); ASIC2b/1a, 1.16 ± 0.64% (n = 7; p = 0.084)]. The apparent proton sensitivity of activation was not different [pH50act. = 6.54 ± 0.04 for ASIC1a (n = 4), pH50act. = 6.42 ± 0.06 for ASIC2b/1a (n = 5; p = 0.16)] (Fig. 3C). These results are consistent with past reports assessing coexpression of ASIC2b with ASIC1a (Hesselager et al., 2004). Next, we examined monovalent cation permeability of ASIC1a and ASIC2b/1a channels (Fig. 3D,E). The calculated ion permeability ratios for sodium and potassium were similar for ASIC1a and ASIC2b/1a (PK+/PNa+ = 0.113 ± 0.034 for ASIC1a, 0.148 ± 0.050 for ASIC2b/1a; p = 0.62). Yet, cells expressing ASIC2b/1a showed a reduced permeability for lithium (PLi+/PNa+ = 1.022 ± 0.105 for ASIC1a, 0.667 ± 0.082 for ASIC2b/1a; p = 0.04). This provides additional evidence that ASIC2b forms novel heteromeric channels with ASIC1a that have unique pore features and is consistent with our results showing these channels are inhibited by both TEA and barium through a pore-blocking mechanism.
ASIC2b/1a is permeable to calcium ions
Currently, ASIC1a and human ASIC1b homomers are the only ASICs known to be appreciably permeable to calcium ions (Waldmann et al., 1997; Waldmann and Lazdunski, 1998; Bassler et al., 2001; Hoagland et al., 2010). The unique calcium permeability of ASIC1a has significant implications for its role in normal physiology, as well as mediating neuronal death in a variety of pathological conditions (Voilley, 2004; Xiong et al., 2004; Wemmie et al., 2006; Xiong et al., 2006; Xu and Duan, 2009). Ca2+ permeability was measured in CHO cells transiently transfected with ASIC1a, ASIC2b/1a, or ASIC2a/1a (Hoagland et al., 2010). As expected, current from all three channels showed a linear current–voltage relationship with a reversal potential near zero in extracellular solutions containing 160 mm sodium (Fig. 4A–C). When extracellular sodium was replaced with calcium, a large negative shift in the reversal potential was observed for ASIC2a/1a channels, whereas only a minor shift was observed for ASIC1a and ASIC2b/1a channels (Fig. 4A–C). This result suggested that ASIC2b/1a heteromeric channels are calcium permeable like ASIC1a homomeric channels. The relative permeability of Na+ compared with Ca2+ (PNa/PCa) was calculated from the change in reversal potentials between sodium- and calcium-containing solutions (Fig. 4D). ASIC2a/1a heteromeric channels had a PNa/PCa of 25.5 ± 5.6 (n = 6), supporting previous studies showing that ASIC2a/1a channels posses little calcium permeability (Bassilana et al., 1997; Yermolaieva et al., 2004). The PNa/PCa was 1.8 ± 0.9 (n = 8) for ASIC1a and 4.1 ± 1.4 (n = 6) for ASIC2b/1a, indicating both channels are permeable to calcium.
ASIC2b/1a and modulation with extracellular zinc
Previous work has shown that extracellular zinc ions modulate ASIC current in a subunit-dependent manner (Baron et al., 2001; Chu et al., 2004). Specifically, Zn2+ inhibits ASIC1a and ASIC2a/1a channels at nanomolar concentrations but potentiates ASIC2a and ASIC2a/1a heteromers at micromolar concentrations (Baron et al., 2001; Chu et al., 2004). Thus, Zn2+ has dual effects on some ASICs. We tested the effects of 0.3 mm Zn2+ on oocytes expressing ASIC1a, ASIC2b/1a, and ASIC2a/1a (Fig. 5A). Both ASIC1a and ASIC2b/1a showed reversible inhibition with zinc, whereas ASIC2a/1a was potentiated in the presence of 0.3 mm Zn2+. Quantification revealed a 51.00 ± 17.90% (n = 8) increase in pH 4.5-evoked current in the presence of 0.3 mm Zn2+ for ASIC2a/1a, a 41.20 ± 6.38% decrease for ASIC1a (n = 7), and a 19.68 ± 3.93% decrease for ASIC2b/1a (n = 8) (Fig. 5B). Thus, ASIC2b/1a heteromeric channels are not modulated by zinc in the same manner as ASIC2a/1a channels. Furthermore, the presence of ASIC2b attenuates zinc modulation as cells expressing ASIC2b with ASIC1a are less affected than cells expressing ASIC1a alone (Fig. 5B) (p = 0.01, t test). We also tested the effects of the zinc chelator TPEN as nominal concentrations of zinc can inhibit ASIC1a and ASIC2a/1a heteromeric channels (Chu et al., 2004). TPEN significantly enhanced current for all three channels suggesting ASIC2b/1a, like ASIC1a and ASIC2a/1a, is inhibited by low nanomolar concentrations of zinc (Fig. 5C).
ASIC2b enhances dynorphin modulation
The opioid peptide Big Dynorphin limits steady-state desensitization of ASIC1a homomers and can enhance acidosis-induced neuronal death (Sherwood and Askwith, 2009). Because ASIC2b affects both PcTX1 inhibition and the pH dependence of steady-state desensitization, we tested Big Dynorphin modulation of ASIC2b/1a heteromeric channels. As we reported previously, Big Dynorphin inhibited steady-state desensitization of ASIC1a homomeric channels. Concentrations as low as 10 nm were effective, and the maximal response was observed with 1 μm peptide (Fig. 6A,D). Cells expressing ASIC2b/1a showed enhanced sensitivity to Big Dynorphin compared with ASIC1a alone (Fig. 6B). In fact, 1 nm Big Dynorphin was sufficient for modulation of ASIC2b/1a current [current was 16.7 ± 4.7% of maximum for control and 48.8 ± 11.2% for Big Dynorphin (1 nm); p = 0.04], and a maximal effect was observed with 100 nm peptide (Fig. 6B,D). This effect is specific to the presence of the ASIC2b subunit as ASIC2a/1a heteromeric channels required higher concentrations of Big Dynorphin for modulation (Fig. 6C,D).
The ASIC2 gene confers barium sensitivity to proton-gated current in hippocampal neurons
Our results in a heterologous expression system indicated that ASIC2b/1a channels are both PcTX1 sensitive and inhibited by low millimolar concentration of barium. To determine whether ASIC2b/1a contributes to proton-gated currents in neurons, we assessed barium- and PcTX1-mediated inhibition of acid-evoked currents in wild-type (ASIC2+/+) and ASIC2 knock-out (ASIC2−/−) neurons that lack both ASIC2 subunits (Price et al., 2000). On average, whole-cell proton-gated currents in wild-type neurons were sensitive to both PcTX1 and barium (Fig. 7A,B). In contrast, proton-gated currents in ASIC2−/− neurons were inhibited by PcTX1 but not by barium, consistent with a lack of ASIC2b/1a heteromeric channels (Fig. 7A,B). The enhanced PcTX1 inhibition in ASIC2−/− neurons compared with wild-type neurons can be explained by a loss of the PcTX1-insensitive ASIC2a/1a heteromeric channels as well (Baron et al., 2002a; Jiang et al., 2009). Thus, on average, hippocampal neurons display characteristics suggesting ASIC2b/1a heteromeric channels contribute to proton-gated currents.
We also compared PcTX1 and barium sensitivity in individual neurons. The sensitivity of proton-gated currents to PcTX1 and barium was highly variable between individual wild-type neurons (Fig. 7C,E). Some neurons showed pH 6.0-evoked currents robustly inhibited by PcTX1 and little affected by barium, such as ASIC1a homomeric channels (Fig. 7C). Many neurons displayed pH 6.0-evoked currents that were robustly inhibited by both PcTX1 and barium, similar to ASIC2b/1a heteromeric channels (Fig. 7C,E). Other neurons showed currents that were predominantly insensitive to both PcTX1 and barium, such as ASIC2a/1a heteromeric channels (Fig. 7C,E). Yet, in the ASIC2−/− neurons, this variability was abolished, and all neurons showed PcTX1-sensitive, barium-insensitive currents (Fig. 7E). This is consistent with a prevalence of ASIC1a homomeric channels in ASIC2−/− neurons. Together, these data suggest that wild-type neurons express a diverse population of ASICs and that ASIC2b/1a channels make a substantial contribution to proton-gated current in cultured neurons.
In an effort to further define the contribution of ASIC2b to proton-gated current in central neurons, we transfected wild-type hippocampal neurons with a shRNA expression plasmid designed to induce RNA interference of ASIC2b mRNA. RT-PCR analysis showed that transfection of neurons with the anti-ASIC2b shRNA plasmid produced an average reduction of ASIC2b mRNA by 31.4 ± 5.92% in whole cultures compared with control transfections (p < 0.01) (Fig. 8A). Barium sensitivity, PcTX1 sensitivity, and steady-state desensitization were assessed in individual neurons transfected with the anti-ASIC2b shRNA versus control (Fig. 8B). We found that sensitivity to these conditions was again highly variable between individual neurons, consistent with the fact that hippocampal neurons express a heterogeneous compliment of ASICs. In both anti-ASIC2b and empty vector-transfected neurons, we recorded cells that were inhibited by PcTX1 and Zn2+ but showed little inhibition with Ba2+ or pH 7.4 conditioning (Fig. 8B, top). Other cells showed considerable inhibition with pH 7.4 conditioning, Ba2+, and PcTX1 (Fig. 8B, middle). Others, still, were not affected by pH 7.4 conditioning, Ba2+, or PcTX1 (Fig. 8B, bottom). Together, these data suggest individual neurons have proton-gated current profiles matching ASIC1a homomeric channels, ASIC2b/1a heteromers, and ASIC2a/1a heteromers, respectively. However, we found a significant reduction of the inhibitory effects of pH 7.4-condtioning and barium treatments in antiASIC2b shRNA-transfected cells compared with control cells (Fig. 8C). Based on these data, we calculated the relative percentage of neurons expressing predominantly ASIC1a-like homomers (>20% PcTX1 inhibition, <20% Ba2+ inhibition), ASIC2b/1a-like heteromers (>20% PcTX1 inhibition, >20 Ba2+ inhibition), and ASIC2a/1a-like heteromers (<20% PcTX1 inhibition, <20% Ba2+ inhibition) (Fig. 8D). We found that in control neurons, 45% of cells predominantly displayed ASIC2b/1a-like profiles, and this number was reduced to 10% in anti-ASIC2b shRNA plasmid-transfected cells (Fig. 8D). Together, these data demonstrate a considerable presence of ASIC2b/1a heteromeric channels in central neurons.
The ASIC2 gene is required for barium-mediated protection from acidosis
Prolonged acidosis induces toxicity in central neurons in an ASIC1a-dependent manner, and treatment with PcTX1 prevents ASIC1a-dependent neuronal death (Xiong et al., 2004; Gao et al., 2005; Sherwood and Askwith, 2009). ASIC-mediated neuronal death is thought to be linked to activity of the calcium-permeable ASIC1a homomeric channels. Since our data indicate that ASIC2b/1a channels are also calcium permeable, we reasoned that ASIC2b/1a channels may also contribute to acid-induced neuronal death. Therefore, we assessed whether the presence of 10 mm barium affects acidosis-induced cortical neuron death (Fig. 9A). Barium inhibits a variety of potassium channels and thus can alter intrinsic excitability of neurons and affect excitotoxic death. However, these experiments were performed in the presence of excitotoxicity inhibitors to specifically isolate acidosis and ASIC-mediated pathways. In wild-type neurons, acidosis caused significant cellular damage as indicated by MTT cleavage (acid treatment was 34.2 ± 9.5% reduced MTT cleavage vs control; n = 7) (Fig. 9B). In the presence of barium, the effect of acidosis on wild-type neurons was robustly reduced (acid treatment plus barium was 3.6 ± 6.4% increased MTT cleavage vs barium control; n = 7; p = 0.006) (Fig. 9B). This assay was also performed on neurons lacking ASIC2 subunits (ASIC2−/−). In the absence of barium, ASIC2 knock-out neurons showed an amount of damage after prolonged acidosis that was similar to wild-type neurons (acid treatment was 43.0 ± 7.0% reduced MTT cleavage vs control; n = 7) (Fig. 9B). However, barium failed to prevent acidosis effects in ASIC2 −/− neurons (acid treatment plus barium was 30.2 ± 8.1% reduced MTT cleavage vs barium control; n = 7; p = 0.25), suggesting that barium-mediated neuroprotection from acidosis required the presence of ASIC2 subunits. Because only the ASIC2b subunit conferred significant barium sensitivity, these results suggest the barium-dependent protection observed in wild-type cells is mediated by ASIC2b/1a heteromers. Together, these results indicate that in wild-type neurons, ASIC2 subunits contribute to acidosis-induced neuronal death and pharmacological blockers of ASIC2b/1a heteromeric channels can limit acidosis-induced toxicity.
Discussion
ASIC2b mRNA is abundant in neurons of the CNS in a pattern that overlaps with ASIC1a and the other ASIC2 transcript variant, ASIC2a (Price et al., 1996; Lingueglia et al., 1997). Identifying a function for ASIC2b in central neurons has been complicated by the fact that (1) ASIC2b fails to produce proton-gated currents when expressed alone, (2) ASIC2b/2a heteromeric channels require extremely acidic pHs for activation (below pH 4.0) that likely do not occur in the CNS (Ugawa et al., 2003), and (3) unique properties of ASIC2b/1a heteromeric channels and their existence in CNS neurons were unclear. Here, we provide the first detailed description of the unique properties of ASIC2b/1a heteromeric channels. We find that ASIC2b/1a channels are indistinguishable from ASIC1a homomeric channels in apparent proton sensitivity of activation and inactivation kinetics. Like ASIC1a homomeric channels, they are also PcTX1 sensitive and calcium permeable. Yet, ASIC2b/1a channels differ from ASIC1a homomeric channels in that they are sensitive to barium and TEA, undergo steady-state desensitization at more basic pH values (pH 7.4), and are modulated by lower concentrations of Big Dynorphin. It should be noted that dynorphin modulation was measured at different pH values for each channel subtype because each channel displays intrinsic differences in the pH dependence of steady-state desensitization. Thus, changes in apparent Big Dynorphin affinity could be caused by ASIC subunit-dependent changes in dynorphin affinity or could be the result of a pH dependence for Big Dynorphin action. In other heteromeric ASIC combinations, such as ASIC2b/2a and ASIC2b/3, ASIC2b enhances acid-induced sustained current (Ugawa et al., 2003; Deval et al., 2004). Notably, this property was not different between ASIC2b/1a heteromers and ASIC1a homomeric channels. The fact that ASIC2b confers distinct properties to different heteromeric channels suggests that the specific contribution of ASIC2b is dependent on the identity of the other subunit present within the heteromeric channel.
Heteromultimeric channels composed of ASIC1a and ASIC2a differ from ASIC1a homomeric channels by requiring more acidic pHs for activation and steady-state desensitization as well as inactivating at a distinctly different rate (Bassilana et al., 1997; Benson et al., 1999; Baron et al., 2002a; Askwith et al., 2004; Hesselager et al., 2004). In addition, ASIC2a/1a heteromeric channels are insensitive to PcTX1, require greater concentrations of Big Dynorphin for modulation, and are potentiated by extracellular zinc (Escoubas et al., 2000; Chen et al., 2005, 2006; Sherwood and Askwith, 2009). Thus, ASIC2a alters the kinetics of ASIC1a channels in a manner distinct from ASIC2b. In fact, for properties such as steady-state desensitization, dynorphin sensitivity, low-affinity zinc modulation, and PcTX1 inhibition, ASIC2a and ASIC2b have opposite effects. However, some properties conferred by ASIC2 subunits are similar. TEA inhibited both ASIC2a/1a and ASIC2b/1a heteromeric channels equally and had little effect on ASIC1a homomeric channels. Inhibition by barium was significantly more pronounced with ASIC2b/1a heteromeric channels compared with ASIC2a/1a heteromeric channels. The difference in the role of ASIC2b relative to ASIC2a must result from a difference in their amino acid sequences. ASIC2b and ASIC2a differ in the first 237 (ASIC2b) and 186 (ASIC2a) amino acids of the protein (Lingueglia et al., 1997). The differential region includes the intracellular N terminus, transmembrane domain 1 (TM1), and part of the extracellular domain, some of which appears to contribute to the extracellular vestibule of the pore (Jasti et al., 2007; Gonzales et al., 2009). Our data suggest that TEA and barium block the pore of ASIC2b/1a channels in a mechanism similar to inhibition of voltage-gated potassium channels (Heginbotham and MacKinnon, 1992; Hurst et al., 1995; Harris et al., 1998). Thus, it is tempting to speculate that amino acids near transmembrane domain 2 (that are conserved between ASIC2a and ASIC2b) may determine TEA block, but substantial barium inhibition requires the TM1 domain of ASIC2b. Additionally, residues N terminal to TM1 (that are distinct between ASIC2a and ASIC2b) have also been shown to affect ASIC ion permeability and could also affect calcium permeability as well as TEA or barium pore block (Coscoy et al., 1999).
Our results indicate that ASIC2b/1a heteromeric channels contribute to endogenous proton-gated currents in cultured neurons. On average, whole-cell proton-gated current in hippocampal neurons was inhibited by barium, and a relatively large percentage of individual neurons showed proton-gated currents that were both PcTX1 and barium sensitive (consistent with ASIC2b/1a). Furthermore, barium inhibition was eliminated in the ASIC2 knock-out, which lacks ASIC2b, and significantly reduced in neurons transfected with anti-ASIC2b shRNA plasmids. Thus, central neurons possess proton-gated currents consistent with three ASIC subtypes: ASIC2b/1a heteromeric channels, ASIC2a/1a heteromeric channels, and ASIC1a homomeric channels. Many processes in vivo and in vitro are attributed to homomeric ASIC1a channels because of their unique sensitivity to PcTX1 (Escoubas et al., 2000; Chen et al., 2005, 2006). We show, for the first time, that heteromeric ASIC2b/1a channels are also inhibited by PcTX1. This suggests ASIC2b/1a could contribute to processes previously thought to be specific to ASIC1a, such as fear-related learning and memory, unconditioned fear behavior, pain processing and neuronal damage after stroke (Xiong et al., 2004; Dwyer et al., 2009; Gregoire and Matricon, 2009; Xu and Duan, 2009; Ziemann et al., 2009). Yet, a specific role for ASIC2 subunits in behavior and central neuronal damage has not been reported. We speculate that ASIC2b might provide a unique contribution to ASIC1a-dependent currents by facilitating interactions with other proteins. ASIC2b binds PICK-1 through a C-terminal PDZ domain and mediates PKC regulation of the peripherally expressed ASIC subunit ASIC3 (Deval et al., 2004). Although ASIC1a and ASIC2a can also bind PICK-1, the presence of ASIC2b could offer a unique contribution to PICK-1 or PKC action in the heteromeric channel (Baron et al., 2002b; Duggan et al., 2002; Hruska-Hageman et al., 2002). Additionally, the unique intracellular N terminus of ASIC2b contains potential phosphorylation sites that could play a role in channel regulation. Finally, ASIC2b could play a role in ASIC1a localization. ASIC1a subunits are expressed on the cell body, dendrites, and dendritic spines of neurons (Wemmie et al., 2003). Synaptically localized ASIC1a affects dendritic spine density, and the C-terminal PDZ protein-binding motif of ASIC2a facilitates ASIC1a localization to dendritic spines via an interaction with PSD-95 (Zha et al., 2006, 2009). It is possible that ASIC2b similarly mediates localization of ASIC2b/1a heteromeric channels to dendritic spines as ASIC2b would possess the same PDZ protein-binding motif. However, unlike ASIC2a/1a heteromers, ASIC2b/1a channels are calcium permeable like ASIC1a homomeric channels. Thus, ASIC2b potentially offers a mechanism to regulate ASIC1a activity without diminishing downstream calcium-dependent signaling events, including those associated with dendritic spine density and neuronal death (Xiong et al., 2006; Zha et al., 2006).
Our data suggest that ASIC2 subunits (ASIC2b, in particular) directly contribute to acidosis-induced neuronal death. We find that barium has a strong effect on acidosis-induced neuronal death in wild-type neurons (and not ASIC2 knock-out neurons), suggesting that activation of ASIC2b/1a heteromeric channels mediates neuronal death. The fact that barium so profoundly inhibited acid-induced neuronal death was surprising given that barium-resistant ASIC1a homomeric channels, which are known to mediate acid-induced death, are also present in these neurons (Duan et al., 2011). It is possible that our results reflect an inherent difference in sensitivity to acid of neurons that predominantly express ASIC2b/1a heteromeric channels versus ASIC1a homomeric channels. In support of this idea, certain hippocampal neuron populations are selectively vulnerable to seizure-induced damage compared with others, which may be linked to specific ASIC subtypes expressed (Weng et al. 2010). We were also surprised to find no inherent difference in acidosis-induced neuronal death between wild-type and ASIC2−/− neurons. This could be attributable to a combined effect of losing both ASIC2a and ASIC2b, which have opposing effects on the characteristics of proton-gated currents. Indeed, previous reports have speculated that the presence of ASIC2 (specifically ASIC2a) attenuates acidosis-induced neuronal death (Yermolaieva et al., 2004; Chu et al., 2006). In addition, ASIC2 has been shown to be protective in mouse models of light-induced retinal degeneration (Ettaiche et al., 2004). However, other studies have suggested that the ASIC2 gene contributes to damage (Peng et al., 2004). Our results are the first to directly show ASIC2 contributes to acidosis-induced neuronal death. This may be important for other neurological disorders. Single nucleotide polymorphisms in the ACCN1 (ASIC2) gene may contribute to multiple sclerosis in humans, a condition that may also involve ASIC1a-mediated neuronal death (Bernardinelli et al., 2007; Friese et al., 2007; Vergo et al., 2011). The fact that the ASIC2b/1a channels are inhibited by compounds that do not affect ASIC1a homomeric channels indicates that ASIC2b/1a channels can be specifically targeted for therapeutic advantage. As specific ASIC subtypes appear to dominate proton-gated currents in subpopulations of neurons (Weng et al., 2010), such therapeutic compounds could provide effective inhibition of acidosis-induced neuronal death in populations of neurons that would otherwise be unaffected by ASIC1a homomer-specific channel inhibitors.
Footnotes
-
This work was supported by an NIH RO1 grant to C.C.A., and an NIH P30 grant to C. Beattie, as well as an American Heart Association Predoctoral Student Fellowship to T.W.S. We thank K. Mykytyn and E. (Hoagland) Frey for helpful discussions and comments on this manuscript. We are grateful to M. Welsh, J. Wemmie, and M. Price for supplying cDNA clones and transgenic animals.
- Correspondence should be addressed to C. C. Askwith, 4197 Graves Hall, 333 West 10th Avenue, Department of Neuroscience, The Ohio State University, Columbus, OH 43210. Askwith.1{at}osu.edu