HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: 92/5/2909    most recent 01198.2003v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (18) Citing Articles via Google Scholar Google Scholar Articles by Lin, W. Articles by Gilbertson, T. A. Search for Related Content PubMed PubMed Citation Articles by Lin, W. Articles by Gilbertson, T. A.

Taste Receptor Cells Express pH-Sensitive Leak K⁺ Channels

¹Cell and Developmental Biology and ²The Rocky Mountain Taste and Smell Center, University of Colorado Health Sciences Center at Fitzsimons, Aurora 80045; ³Department of Biomedical Sciences, Colorado State University, Fort Collins, Colorado 80523; and ⁴Department of Biology and The Center for Integrated BioSystems, Utah State University, Logan, Utah 84322

Submitted 11 December 2003; accepted in final form 6 July 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Two-pore domain K⁺ channels encoded by genes KCNK1-17 (K2p1-17) play important roles in regulating cell excitability. We report here that rat taste receptor cells (TRCs) highly express TASK-2 (KCNK5; K2p5.1), and to a much lesser extent TALK-1 (KCNK16; K2p16.1) and TASK-1 (KCNK3; K2p3.1), and suggest potentially important roles for these channels in setting resting membrane potentials and in sour taste transduction. Whole cell recordings of isolated TRCs show that a leak K⁺ (K_leak) current in a subset of TRCs exhibited high sensitivity to acidic extracellular pH similar to reported properties of TASK-2 and TALK-1 channels. A drop in bath pH from 7.4 to 6 suppressed 90% of the current, resulting in membrane depolarization. K⁺ channel blockers, BaCl₂, but not tetraethylammonium (TEA), inhibited the current. Interestingly, resting potentials of these TRCs averaged –70 mV, which closely correlated with the amplitude of the pH-sensitive K_leak, suggesting a dominant role of this conductance in setting resting potentials. RT-PCR assays followed by sequencing of PCR products showed that TASK-1, TASK-2, and a functionally similar channel, TALK-1, were expressed in all three types of lingual taste buds. To verify expression of TASK channels, we labeled taste tissue with antibodies against TASK-1, TASK-2, and TASK-3. Strong labeling was seen in some TRCs with antibody against TASK-2 but not TASK-1 and TASK-3. Consistent with the immunocytochemical staining, quantitative real-time PCR assays showed that the message for TASK-2 was expressed at significantly higher levels (10–100 times greater) than was TASK-1, TALK-1, or TASK-3. Thus several K_2P channels, and in particular TASK-2, are expressed in rat TRCs, where they may contribute to the establishment of resting potentials and sour reception.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Taste stimuli interact with receptors or ion channels of taste receptor cells (TRCs), which guides the acquisition of nutrients and avoidance of toxins. Acids elicit sour taste, with the degree of sourness a function of proton concentration. Mechanisms of sour transduction involve multiple ion channels (Gilbertson et al. 1992, 1993; Kinnamon and Roper 1988; Kinnamon et al. 1988; Lin et al. 2002a; Miyamoto et al. 1988, 1998; Stevens et al. 2001; Ugawa et al. 1998), membrane proteins (Bigiani and Roper 1994; Okada et al. 1987, 1993), and intracellular molecules (Liu and Simon 2001; Lyall et al. 2001; Richter et al. 2003; Stewart et al. 1998). Thus sour taste coding seems to integrate signals from multiple pathways.

TRCs are excitatory, generating spontaneous and evoked action potentials. The excitability is regulated by potassium channels that set resting membrane potentials (V_rest) and regulate action potential frequency (Hille 2001). In TRCs, control of V_rest has been attributed to delayed rectifying K⁺ (K_DR) channels, leak K⁺ (K_leak) channels (Kolesnikov and Bobkov 2000; Miyamoto et al. 1991; Okada et al. 1986; Roper and McBride 1989), and inward rectifier K⁺ (K_ir) channels (Sun and Herness 1996). However, V_rest in most TRCs is from –36 to –69 mV (Miyamoto et al. 2000), potentials where K_DR and K_ir conduct little current (Chen et al. 1996). Recently, a leak K⁺ channel (K_leak) was found in mouse taste buds cells, conducting time- and voltage-independent currents and contributing to setting V_rest (Bigiani 2001). Its molecular identity has not been determined.

We previously reported that acids depolarize taste receptor cells by two different mechanisms: activation of an inward current, possibly mediated by acid-sensing ion channels (ASICs) (Lin et al. 2002a), and suppression of a steady-state leak conductance (Lin et al. 2002b). Preliminary studies revealed that the acid-suppressed conductance shared properties with the reported K_leak in mouse (Bigiani 2001) and with the cloned two-pore domain K⁺ (K_2P) channels of the TASK family (TWIK-related acid-sensitive K⁺ channel) (Duprat at al 1997; Girard et al. 2001; Kim et al. 1998, 2000; Leonoudakis et al. 1998; Rajan et al. 2000; Reyes et al. 1998), which are sensitive to extracellular pH (Millar et al. 2000) and contribute to the establishment of V_rest. Recently, an additional member of the K2P family, TALK-1 (TWIK-related alkaline pH activated K⁺ channel type 1) (Han et al. 2003), has been identified, whose properties closely match those of TASK-2. The voltage-independent activation and open-rectification of TASKs permit substantial current at both V_rest and depolarized potentials. Thus TASKs could provide potential mechanisms for sour taste transduction and the control of V_rest in TRCs. Since acid is present in foods commonly, and sour-sensitive TRCs respond broadly to stimuli with different modalities (Caicedo et al. 2002; Gilbertson et al. 2001; Sato and Beidler 1997), acid modification of V_rest via TASKs may modulate other taste sensations.

Using H⁺ sensitivity as a reporter in whole cell recordings, we characterized TASK-like currents in TRCs. We show that some TRCs possess a highly H⁺-sensitive K_leak conductance that controls V_rest. Immunocytochemical and molecular biological approaches show the presence of TASK-like channels in rat taste buds, and TASK-2 is the most highly expressed of these channels. Together, we provide strong evidence for the presence of a subset of K2P channels in TRCs and suggest potential roles in setting V_rest and in sour taste transduction. Preliminary results have been published in abstract form (Burks et al. 2003; Lin et al. 2002b).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Electrophysiological recordings

Adult Sprague-Dawley male rats were used in this study. Vallate papillae taste bud isolation, whole cell patch-clamp recordings, and data acquisition were as described previously (Lin et al. 2002a). The bath solution (Tyrode's) was comprised of (in mM) 140 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 HEPES, 10 glucose, and 10 sodium pyruvate (pH 7.4 with NaOH). Acidic solutions were obtained by adding 1 M citric acid or HCl to the bath solution to obtain the desired pH. K⁺ channel inhibitors BaCl₂ (5 mM) and 10 mM tetraethylammonium (TEA; Sigma Chemical, St. Louis, MO) were added to Tyrode's and bath-applied to taste cells. The intracellular pipette solution contained (in mM) 140 KCl, 1 CaCl₂, 2 MgCl₂, 10 HEPES, 11 EGTA, 1 ATP, and 0.4 GTP (pH 7.2 with KOH). To ensure that recordings were obtained from TRCs, we applied depolarizing voltage steps to induce voltage-gated K⁺ and/or Na⁺ current, since nonsensory epithelial cells and some glia-like taste cells (Akabas et al. 1990; Bigiani 2001) do not possess these currents. For steady-state measurements, holding current was recorded at various holding potentials and 10- or 20-mV hyperpolarizing voltage pulses were used to monitor membrane conductance. Statistical analyses and curve fittings were conducted using Origin 6.1.

Immunocytochemistry

Rats were anesthetized with sodium pentobarbital (40 mg/kg) or ketamine–xylazine (100-20 mg/kg) and perfused transcardially with 0.1 M PBS followed by buffered 4% paraformaldehyde. The tongue and positive control tissues of brain and kidney were removed and postfixed for 2 h before being transferred into PBS with 25% sucrose overnight. The tissues were frozen and cut with a cryostat into free-floating 30-µm-thick sections. Sections containing taste buds of foliate and vallate papilla were selected, rinsed, and incubated in blocking solution containing 2% normal goat or donkey serum, 0.3% Triton X-100, and 1% bovine serum albumin in PBS for 1.5 h. The sections were incubated with polyclonal antibodies against TASK-1, TASK-2 (Alomone Labs), or TASK-3 (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:500 to 1:50 dilutions in the blocking solution overnight at 4°C, followed by rinsing and incubation with the FITC-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 h at room temperature. Sections were washed and mounted on slides with Fluoremount-G (Fisher Biotech, Birmingham, AL). Positive control tissues consisted of brain or kidney, which reportedly expresses TASK channels (Millar et al. 2000; Reyes et al. 1998). Negative controls involved omitting the primary antibodies and preincubation of the primary antiserum with immuno-peptides provided by the company. Pictures were taken with an Olympus Fluoview laser scanning confocal microscope.

Western blotting

To verify whether the antibody against the TASK-2 binds to protein in taste tissue with corresponding molecular weight, Western blotting was performed using an anti-TASK-2 antibody. Taste tissue containing vallate and foliate papillae collected from three adult rats and brain tissue (cortical layer) were homogenized in a buffer containing: 50 mM HEPES, 100 mM sodium pyrophosphate, 100 mM sodium fluoride, 10 mM EDTA, 1% Triton-X 100, and protease inhibitors (Sigma Chemical) 2 µg/ml aprotinin, 5 µg/ml leupeptin, and 34 µg/ml PMSF. Homogenates were centrifuged at 15,000 rpm for 30 min. About 10 µl of taste tissue supernatant and 9 µl of brain supernatant containing 100 µg total protein each were separated by SDS-PAGE on Tris-HCl gels (10%; Bio-Rad) and transferred onto polyvinylidene difluoride membranes (Bio-Rad). The membranes were incubated with 5% nonfat dry milk in Tris-based saline for 1 h at room temperature followed by incubation with anti-TASK-2 antibody (1:300; Alomone Labs) overnight at 4°C on a shaker. After rinsing, the membrane was incubated with a goat anti-rabbit secondary antibody conjugated with horseradish peroxidase (1:50,000; Bio-Rad) for 1 h at room temperature. Signal was detected by enhanced chemiluminescence (ECL plus, Amersham Pharmacia). Immunoblot results shown are unenhanced scans of the Fuji medical X-ray film (Fuji Photo Film).

Isolation and purification of taste bud RNA

Taste buds were isolated from the fungiform, circumvallate, or foliate papillae of the male rat (Sprague-Dawley) tongue according to established procedures (Gilbertson and Fontenot 1998), washed to remove nonadherent cells, and immediately placed into 1.5-ml microfuge tubes with 200 µl RNAlater (Ambion, Austin, TX). The taste buds were centrifuged at 6,000 rpm for 7 min. The resulting pellet was resuspended in lysis buffer from the RNeasy Mini Kit from Qiagen (Valencia, CA), and RNA was extracted according to manufacturer's instructions, including DNase I treatment. For positive or negative controls, RNA was extracted from 100 mg of brain tissue (for TASK-3), kidney (for TASK-1, TASK-2, and TALK-1), pancreas (TALK-1), and liver (TALK-1) using Tri Reagent (MRC, Cincinnati, OH) according to the manufacturer's instructions.

RT-PCR

First-strand cDNA was synthesized using the OmniScript RT Kit (Qiagen). The maximum volume of taste RNA or 50 ng of brain RNA was used for the reaction, with the total volume being 20 µl. Reactions were also set up in which the reverse transcriptase enzyme was omitted as a control to detect genomic DNA contamination. After first-strand synthesis, 2 µl of cDNA was added to a PCR reaction mix [final concentration: 500 mM KCl, 100 mM Tris-HCl (pH 8.3), 2.0 mM Mg²⁺, 1x TaqMaster PCR enhancer (Eppendorf, Westbury, NY), 200 µM dNTPs, 500 nM forward and reverse primers, and 1.25 U Taq polymerase]. The following primer sequences were used for the three TASK channels and TALK-1 in the RT-PCR assays: TASK-1 (accession no. AF031384; rat), 5'-TGTTTTGGTTTGGTTCTCGT-3' (sense, nucleotides 1728–1747), 5'-GTGACCTGGACAAAGACACC-3' (antisense, 1868–1887); TASK-2 (accession no. AF319542; mouse), 5'-CAGCCATCTTCATCGTGTG-3' (sense, 557–575), 5'-ACTTCCAGCCATCTGTAGGG-3' (antisense, 896–915); TASK-3 (accession no. AF192366; rat), 5'-CGCATGAACACCTTCGTG-3' (sense, 481–498), 5'-GGACAACCACCCGTCTTG-3' (antisense, 890–907); and TALK-1 (accession no. AY404471; mouse), 5'-AAGGCAACTCCACCAATCCC-3' (sense, 251–270), 5'-AGAAGCCCTCACGGAAGC-3' (antisense, 593–610). Amplification by regular PCR included an initial 5-min denaturation step followed by 40 cycles of a three-step PCR: 30-s denaturation at 95°C, 30-s annealing at a predetermined optimal temperature (62°C for TASK-1 and TALK-1, 57°C for TASK-2, 59°C for TASK-3), and 45-s extension at 72°C, and concluding with a 7-min final extension step. Amplified sequences were visualized by electrophoresis in 2% agarose gels poured using 1x TAE buffer (40 mM Tris-Acetate, 1 mM EDTA) or by real-time technology. cDNA to be sequenced was either purified directly after PCR using the QIAquick PCR purification kit (Qiagen) or extracted from agarose gels using the QIAquick gel extraction kit. Sequences were determined by the dye-terminator method using an ABI (Foster City, CA) Model 3100 Automatic Sequencer.

qPCR

To quantify TASK-1, TASK-2, TASK-3, and TALK-1 mRNA levels among the different taste epithelia, we used a two-tube RT-PCR assay with the PCR step conducted in a real-time thermal cycler (SmartCycler, Cepheid, Sunnyvale CA). The procedures for first-strand synthesis are the same as described earlier, except the reaction was scaled 100 µl. Two microliters of cDNA was used for each qPCR reaction. The HotMaster Taq DNA polymerase kit (Eppendorf) was used, with the following final concentration: 1x reaction buffer, 3.5 mM Mg²⁺, 200 µM dNTPs, 300–900 nM sense and antisense primers, 300–900 nM fluorescent probes, and 1.25 U HotMaster Taq. Two-step PCR protocols were used to amplify TASK1 and TASK3 (15-s denaturation at 95°C and 60-s annealing and extension at 60°C) and TALK-1 (15-s denaturation at 95°C and 60-s annealing and extension at 62°C), while a three-step PCR protocol (15-s denaturation at 95°C, 30-s annealing at 57°C, and 30-s extension at 72°C) was used to amplify TASK2. Primers and probes were designed for the three TASK channels, TALK-1, and the housekeeping gene, GAPDH, using Oligo 6.0 Primer Analysis Software (Molecular Biology Insights, Cascade, CO). Primer and probe sequences for the qPCR assays are listed in Table 1. We used a TaqMan (ABI) detection system in which the primer pairs for channel-specific sequences were multiplexed with the primer pairs for GAPDH for comparison of expression levels in the three types of taste buds (Bustin 2000). Channel-specific probes were labeled at the 5'-end with carboxyfluorescein fluorescent dye (FAM) as the reporter fluorophore and Black Hole quencher-1 (BHQ-1) at the 3'-end as the quencher. The GAPDH probe was labeled with carboxy-X-rhodamine fluorescent dye (ROX) as the reporter fluorophore and Black Hole quencher-2 (BHQ-2) as the quencher. All probes were obtained from Integrated DNA Technologies (Coralville, IA). All qPCR assays were carried out in triplicate, and a minimum of three independent experiments was conducted.

–C_T

(1)

(2)

(3)

(4)

where, C_T is the cycle threshold for the K_2P channels or GAPDH determined empirically, and C_T^CAL is the cycle threshold for the calibrator, the most highly expressed channel in each assay. Mean relative expression values and SD were calculated from the three individual sets of pooled taste bud types. To determine if there were significant differences among the expression of K_2P channels in the three taste bud types, multiple pairwise comparisons were made using a one-way ANOVA followed by Bonferroni's posthoc test for significance (SPSS 10.0, SPSS, Chicago, IL).

To determine if the efficiencies of the target and reference (GAPDH) amplification were consistent across template dilutions, we evaluated the C_T values for each set of K_2P primers and GAPDH in three separate multiplexed reactions. For each of the PCR reactions, the absolute value of the slope of the log input versus C_T was <0.1, showing equal amplification efficiencies for the different starting template concentrations (cf. Fig. 6, inset). There was no effect on C_T values when the GAPDH primers were either limited or not limited in the reactions.

View larger version (19K): [in this window] [in a new window]   FIG. 6. Relative expression of TASK and TALK channel mRNA in 3 lingual taste bud types. Expression was determined relative to an internal calibrator as described in Eqs. 1–4. In all 3 taste bud types, TASK-2 expression was significantly greater than TASK-1, TALK-1, or TASK-3. Data are presented ±SD as the mean of 3 independent experiments each done in triplicate. Inset: relative PCR efficiency plots of the 3 TASK and TALK-1 channels and GAPDH plotting log total RNA input vs. CT. Absolute values of the slopes for all 4 assays were <0.1.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

Whole cell patch-clamp recordings were performed on freshly isolated vallate TRCs, and acidic stimuli were bath-applied. From a total 168 TRCs recorded, 19 cells responded to a pH drop from pH 7.4 to 5 (acidified with citric acid or HCl), with a slight change in current when held at –80 mV, close to the equilibrium potential of K⁺. When held at less negative potentials, such as –60 or –40 mV, which increases the driving force for K⁺, these TRCs responded to acid stimulation with a sizable sustained inward current accompanied by a significant reduction of the membrane conductance (Fig. 1A), leading to cell depolarization in current-clamp configuration (Fig. 1B). The majority of cells (120) responded to acid stimulation with a large rapidly activating and desensitizing inward current, which depolarized cells by increasing membrane conductance. ASICs have been proposed to mediate this response (Lin et al. 2002a). Some of these cells (3 of 9 tested) also possessed the acid-suppressive conductance with the same response profile as Fig. 1A. This could be observed when TRCs were held at 20 mV, where the ASIC-like current reached its reversal potential and was largely diminished (data not shown).

View larger version (21K): [in this window] [in a new window]   FIG. 1. The pH-sensitive Kleak conductance. A: a drop of bath pH from 7.4 to 5.5 (acidified with citric acid) induced sustained inward current accompanied with a significant reduction of membrane conductance in voltage-clamp configuration. Holding potential was –60 mV (n = 19). B: under current-clamp configuration (I = 0), acid stimulation resulted in membrane depolarization. C: current-voltage relationship of the pH-sensitive Kleak. Taste receptor cells (TRCs) were held at different voltages ranging from –100 to –30 mV, and pH 5 acid-induced currents were recorded and normalized to response amplitude obtained at –60 mV of the same cell. The plotted current-voltage curve was fitted to the Goldman-Hodgkin-Katz current equation, showing a reversal potential of –90 mV at a value close to EK (–94 mV), which suggests that acid blocked a background or leak K+ channel. Each point represented 3–5 cells (means ± SE). D: dose-dependent inhibition of K+ currents with decreasing extracellular pH. Control bath pH was 7.4. Cells were held at –60 mV and were challenged with acidic extracellular pH ranging from 7.18 to 4. Amplitudes of current responses induced by different pH stimulations collected from 3 cells were normalized to the peak response of pH 5 (means ± SE). Kleak channel showed sharp pH dependence.

The high H⁺ sensitivity is a hallmark feature of TASKs (Goldstein et al. 2001; Lesage and Lazdunski 2000; O'Connell et al. 2002). We therefore examined the pH dependence of the K_leak by holding TRCs at –60 mV and stimulating with acidic solutions ranging from pH 7.18 to 4. As shown in Fig. 1D, the K_leak was highly sensitive to extracellular pH. A drop of bath pH from pH 7.4 to below 7.18 inhibited the conductance and resulted in a sustained current, which was about 90% reached by pH 6 and saturated at pH 5. This pH sensitivity was similar to cloned TASK-2 tested at similar resting bath pH conditions (Morton et al. 2003), but markedly different from the ASIC-like current in TRCs (Lin et al. 2002a). Our data thus suggested that TASK-like channels might be present in TRCs and involved in sour taste transduction.

TASK channels exhibit distinct pharmacological characteristics. Unlike the voltage-gated K⁺ channels, they are insensitive to the "classical" K⁺ channel blocker TEA, but are blocked by Ba²⁺, a common inhibitor of two-pore domain channels (Ashmole et al. 2001; Duprat et al. 1997; Girard et al. 2001; Kim et al. 1998, 2000; Leonoudakis et al. 1998; Rajan et al. 2000; Reyes et al. 1998). The K_leak conductance in taste cells reported by Bigiani (2001) also is insensitive to TEA but blocked by Ba²⁺. We therefore examined effects of TEA and Ba²⁺ on the acid-sensitive K_leak in TRCs. As shown in Fig. 2B, TEA (10 mM) added to the bath solution did not inhibit either the resting conductance or the acid-induced current (n = 3). Instead, the acid responses increased slightly. However, when BaCl₂ (5 mM) was added to the bath, the resting conductance was greatly reduced, leading to an abolishment of acid responses (n = 4; Fig. 2C). Similar to Ba²⁺ block, bath application of quinidine (1 mM) suppressed the conductance and the acid-induced responses (data not shown). Additionally, we examined the effect of the Cl^– channel blocker NPPB (0.1 mM), which reportedly blocks an acid-sensitive Cl^– current (Miyamoto et al. 1998) and the volume-sensitive Cl^– current (Gilbertson 2002) in TRCs, and found no effect on the pH-sensitive current in these cells (data not shown).

View larger version (20K): [in this window] [in a new window]   FIG. 2. Pharmacological effects of K+ channel blockers. A: a TRC responded to pH 5 at a holding potential of –40 mV. B: bath application of the "classical" K+ channel blocker tetraethylammonium (TEA; 10 mM) suppressed neither steady-state conductance nor acid-induced response. C: BaCl2 (5 mM) applied to the bath reversibly blocked the steady-state conductance, inhibiting response to acid (n = 4). All 3 traces were recorded from the same cell.

View larger version (27K): [in this window] [in a new window]   FIG. 3. Correlation of the pH-sensitive Kleak with the resting membrane potential (Vrest). A: depolarizing voltage steps from –80 to 60 mV in 10-mV increments were applied to a cell expressing the pH-sensitive Kleak. Traces are shown for every 20 mV. Note the presence of both voltage-gated inward and outward currents, which indicated that the cell was a TRC. Vrest of the TRC was –76 mV. B: peak outward current at different potentials, including both leak and voltage-gated components is plotted. Leak currents were present at all potentials tested and could be observed easily at voltages between –40 to –60 mV, where there were no voltage-activated K+ currents under our recording conditions. C: amplitudes of leak current measured at –40 mV were plotted vs. Vrest of the same cells showing significant correlation (r = –0.77, SD = 34, P = 0.002, n = 13). D: cells expressing the pH-sensitive Kleak all had highly negative Vrest near EK. The Vrest closely correlated to the amount of conductance reduced by acid stimulation (r = –0.78, SD = 0.38, P = 0.0006, n = 15). The larger the pH-sensitive conductance, the more negative the Vrest, which suggested that the pH-sensitive Kleak was responsible for setting Vrest in these cells.

To determine which, if any, cloned TASKs were present in TRCs, three commercially available antibodies against TASK-1, TASK-2, and TASK-3 were used in immunocytochemical experiments. No antibodies were available for TALK-1 channels. The anti-TASK-1 antibody at 1:100 or 1:50 final dilutions failed to label TRCs, although a few nerve fibers adjacent to taste buds were immunopositive (Fig. 4, C and D). In contrast, anti-TASK-1 labeled control tissues, including a subset of neurons in the motor trigeminal nucleus of the brain stem and cerebellar granule neurons (Fig. 4, A and B) (Karschin et al. 2001; Millar et al. 2000). The anti-TASK-3 antibody apparently did not react with TRCs either, although the antibody labeled nerve fibers, and possibly nontaste cells that were adjacent to TRCs (Fig. 4, H–J). In contrast, anti-TASK-2 strongly labeled a small subset of TRCs from both foliate and vallate papillae (Fig. 4, E–G). Interestingly, the antibody also labeled many TRCs weakly. The reactivity was seen in whole cells and not restricted to the apical compartment, consistent with previous reports that the Ba²⁺-sensitive K_leak is located primarily at the basolateral membrane of TRCs (Miyamoto et al. 2000). Positively labeled TRCs generally had an elongated shape and round or oval nuclei. Western blotting of protein from circumvallate taste buds probed with anti-TASK-2 antibody confirmed expression of TASK-2 (Fig. 4K). These results were consistent with electrophysiological data and suggested heterogeneous expression of TASK-2 in rat TRCs.

View larger version (74K): [in this window] [in a new window]   FIG. 4. Expression of TASK channel proteins in TRCs. Antibodies against TASK-1, -2, and -3 were applied to control tissues and to foliate and vallate papillae containing TRCs. Photomicrographs were images obtained using a laser-scanning confocal microscope. A and B: anti-TASK-1 antibody at 1:100 final dilution reacted positively to neurons in motor trigeminal nucleus of brain stem and granule layer of cerebellum, respectively. C: anti-TASK-1 antibody did not label TRCs in vallate papillae. D: transmission image of C, showing presence of many taste buds. E: low magnification image showing presence of immunoreactive TRCs to anti-TASK2 antibody (1:100) in many taste buds of a foliate papilla. F and G: high magnification showing labeled individual TRCs in foliate and vallate papillae. Immunoreactive TRCs were elongate and often had round- or oval-shaped nuclei. H: antibody against TASK-3 at working dilutions of 1:500 to 1:50 apparently did not react with TRCs. I and J: positive reaction for the anti-TASK-3 antibody is present in adjacent nerve fibers and a few nontaste cells. Negative controls consisted of omitting the primary antibody and preabsorbing the primary antibody with antigen peptide overnight before application to taste tissues (data not shown). Scale: 20 µm. K: Western blot showing an anti-TAKS-2 antibody labeled an 45-kD band indicating the presence of TASK-2 protein in rat foliate and vallate taste buds (taste). Negative control was rat cortex (brain).

To verify expression of TASK channels in rat taste buds, a series of RT-PCR assays were performed on mRNA isolated from fungiform, foliate, and vallate taste buds. PCR products for both TASK-1 and TASK-2 messages could be found in all three lingual taste bud types in a minimum of three independent experiments (Fig. 5). To further confirm their identity, the PCR products were sequenced, and the resulting sequences were compared with published sequences using BLAST (http://www.ncbi.nlm.nih.gov/BLAST/). Over the regions sequenced (461 bp for TASK-1 and 302 bp for TASK-2), there was 100% identity between our PCR product for TASK-1 and TASK-1 from rat cerebellar granule cells and 95% homology between rat taste bud TASK-2 and TASK-2 from mouse kidney (data not shown). RT-PCR analysis for TASK-3 expression in rat taste buds produced more equivocal results. In most assays, no bands representing the appropriately sized PCR product were found consistently in any taste bud type, although bands in the positive control lanes (e.g., from brain) were found (Fig. 5). Rarely, however, a band was found for TASK-3 in one or more of the lanes containing one of the three taste bud types (data not shown). Considering the immunocytochemical results for TASK-3, it is possible this may have represented contamination of our taste buds with surrounding nontaste cells on these rare occasions.

View larger version (50K): [in this window] [in a new window]   FIG. 5. RT-PCR reveals the presence of several K2P channels in mRNA from the 3 lingual taste bud types. Primers for TASK-1 and TASK-2 amplify ethidium bromide–stained PCR products of expected sizes (TASK-1: 333 bp; TASK-2: 359 bp), but TASK-3 in most cases was negative (expected product: 427 bp). Positive controls (rat kidney or brain RNA) are shown for TASK-1, TASK-2, and TASK-3 with each set of primers. Negative (–) control lanes represent those in which cDNA was omitted from the PCR reaction. This top gel was overexposed to showe the presence of TASK-1 in rat taste buds. Inset: RT-PCR assay revealing expression of TALK-1 in 3 lingual taste buds and pancreas (expected product size: 360 bp), but not in liver and kidney. Negative (–) control represents the omission of template in the PCR reaction.

Since the immunocytochemical data indicated that TASK-2 expression was greater than that of TASK-1 (or TASK-3), a series of multiplexed Taqman-style quantitative real-time PCR reactions were run on pooled cDNA isolated from the three lingual taste buds of several rats. In a single tube, primer sets for one of the three TASK channels or TALK-1 and a dual-labeled fluorogenic probe specific for a region within the K_2P primer boundaries was multiplexed with primers and a probe for the housekeeping gene GAPDH. Each of the four K_2P channels was analyzed in this manner to determine their expression relative to GAPDH by calculating the C_T values for each replicate as described. To compare expression among the various K_2P channels and taste bud types, the relative expression of each K_2P channel within each taste bud type was determined with respect to an internal calibrator (i.e., the most highly expressed K_2P channel). As shown in Fig. 6, relative TASK-2 expression in all three taste bud types is conservatively 10–100 times more highly expressed than TASK-1 and TALK-1, which, in turn, is from 2 to 80 times more highly expressed than TASK-3. Expression of TASK-2 was significantly higher than TASK-1, TALK-1, or TASK-3 within a taste bud type (e.g., TASK-2 > TASK-1 and TASK-2 > TASK-3; P < 0.05); however, TASK-1 expression was not significantly different from TASK-3 expression. There were no significant differences for TASK-1, TASK-2, or TALK-1 expression within each of the three taste bud types as determined by ANOVA. Because of its exceedingly low expression, differences in TASK-3 expression among the three taste bud types were not statistically analyzed.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
In this study, we employed multiple techniques to identify a background or leak K⁺ channel and to examine its potential function in setting V_rest and transducing sour taste in TRCs of rat vallate papillae. Our results from whole cell patch-clamp recordings showed that a K_leak current in a subset of TRCs exhibited high sensitivity to acidic extracellular pH. Ba²⁺, a common blocker for two-pore domain leak K⁺ channels, but not the "classical" K⁺ channel blocker TEA, inhibited the conductance. The amplitude of pH-sensitive K_leak correlated significantly with V_rest, suggesting that K_leak was the major determinant for setting V_rest in these cells. These three important features suggested the presence of TASKs or TASK-like K_2P channels in TRCs. Consistent with this, we found strong TASK-2 immunoreactivity in a subset of TRCs and verified TASK-2 protein expression with Western blotting. Both RT-PCR and qPCR assays were consistent with the expression of TASK-2 and, to a lesser extent, TASK-1 and TALK-1 in rat taste buds. TASK-3 expression seems minimal and may be indicative of contamination from nontaste cells. Quantitative assays showed that TASK-2 was expressed at significantly higher levels than TASK-1 and TALK-1 in all taste bud types. Together, our data provide strong evidence that several K_2P channels (predominantly TASK-2) are present in rat TRCs and may be involved in transducing sour taste, setting resting potentials, and regulating cell excitability.

Several K⁺ channels are reportedly present in TRCs of various species and perform multiple functions, such as controlling resting potentials and the firing rate of action potentials, maintaining K⁺ homeostasis, and participating in taste transduction for salty, sour, sweet, bitter, and fatty acid stimuli (Bigiani 2001, 2002; DeSimone et al. 2001; Gilbertson et al. 2000; Herness and Gilbertson 1999; Kinnamon and Margolskee 1996; Lindemann 1996, 2001; Margolskee 2002; Miyamoto et al. 2000). The subset of K_2P channels examined in this study in TRCs share several important features with the leak K⁺ channels reported by Bigiani (2001) in taste bud cells in that they are voltage-independent, K⁺-selective, and sensitive to Ba²⁺. While Bigiani (2001) emphasized its possible glia-like function in maintaining K⁺ homeostasis in mouse taste bud cells, some of his recordings were from cells that did not express voltage-gated currents, suggesting they may not be TRCs. We concentrated only on TRCs and examined their roles in control of V_rest and in sour taste transduction. We showed that the K_leak current was blocked by small extracellular acidification and was not voltage-dependent, both of which are properties that are diagnostic of TASKs.

The TASK group consists of five subtypes, of which three mammalian representatives, TASK-1, -2, and -3, are known to be sensitive to low extracellular pH (Duprat et al. 1997; Kim et al. 1998, 2000; Leonoudakis et al. 1998; Rajan et al. 2000; Reyes et al. 1998). The TASK-4 subunit, also called TALK-2, is active only at alkaline pH (Decher et al. 2001; Girard et al. 2001) and therefore is not considered a likely candidate for the pH-sensitive K_leak in rat TRCs. Except for TASK-5, which does not form a functional homomeric channel (Kim and Gnatenco 2001), other TASKs conduct K⁺ currents that possess all the characteristics of mammalian background or leak conductance, i.e., they generate essentially instantaneous, noninactivating, voltage-insensitive currents that have an open or weakly outwardly rectifying current-voltage relationship in asymmetric K⁺ gradients as predicted by the Goldman-Hodgkin-Katz constant field equation. Recently, identification of an additional member of the K_2P family, TALK-1, was reported with qualitatively similar properties to TASK-2 (Han et al. 2003). Presently, much less is known about the physiological and pharmacological properties of rat TALK-1; however, its inhibition by decreases in extracellular pH in the physiological range also makes it a candidate for the types of acid responses we report.

The activity of TASK-like channels (TASKs and TALK-1) is strongly dependent on external pH and also is regulated tightly by many transmitters, neuropeptides, and other factors (Maingret et al. 2001; Niemeyer et al. 2001; Patel et al. 1999; Sirois et al. 2000; Talley et al. 2000; Washburn et al. 2002). Thus TASKs present a constellation of functional properties that is unique among all K⁺ channels cloned to date. Moreover, our data showing the inhibition of acid responses by Ba²⁺ are consistent with TASK-2 and TALK-1 channels, but not TASK-1 or TASK-3 (O'Connell et al. 2002). TASK-2 is present primarily in epithelial tissues such as lung, colon, kidney, intestine, and stomach (Reyes et al. 1998); the latter two also seem to express TALK-1 in rat (Kang and Kim 2004). Functions of TASKs have been shown in controlling resting potentials, cell volume, cell excitability and chemoreception for pH/or pCO₂ and anesthetics (Bayliss et al. 2001; Han et al. 2002; Millar et al. 2000; Niemiyer et al. 2001; Sirois et al. 2000; Talley et al. 2000).

Because of the functional similarity between TASK-2 and TALK-1, additional RT-PCR and qPCR assays using primers for TALK-1 were carried out in taste buds and both positive (pancreas) and negative (liver, kidney) control tissues. Expression of TALK-1 was found in all three lingual papillae, and the possibility remains that TALK-1 may be contributing to the acid-induced currents we report here. However, quantitative assays for TALK-1 mRNA expression showed that it is 10–100 times less prevalent than TASK-2 (cf. Fig. 6). Thus, of those K_2P channels with properties similar to those reported in this study, TASK-2 seems the most likely candidate based on its high relative expression, and we have focused our discussion on this channel accordingly.

The hallmark features of TASKs coupled with the high expression of TASK-2 in mammalian TRCs (Fig. 6) suggest that TASK-2 may play important roles in TRCs. First, TASK-2 could be a sour taste transducer because of the high extracellular pH sensitivity. The presence of a hyperpolarizing conductance produced by TASK-2 at resting potentials provides the electrical basis for H⁺ block-induced depolarization. Sour taste transduction can occur at both the apical and basolateral membrane of TRCs. However, tight junctions between taste bud cells and between epithelial cells separate the apical and basolateral membranes so that only limited amounts of H⁺ normally reach the basolateral membrane of TRCs. The membrane localization of TASK-2 is not known. Since positive immunoreactivity was present in the cell body and blockage of a basolateral background K⁺ conductance by Ba²⁺ led to depolarization and increase in the input resistance (Miyamoto et al. 2000), it is likely that TASK-2 is expressed at the basolateral membrane. Regardless of its location, 90% of the TASK-mediated pH-sensitive current could be obtained by a slight reduction of extracellular pH from 7.4 to 6. Therefore TASKs may be especially important in sensing weak acidic stimuli.

Second, TASK-2 could be a major determinant for setting V_rest and controlling TRC excitability in some TRCs. K_2P family members like TASKs and TALK-1 conduct instantaneous current at all potentials. This is distinct from the delayed rectifier K⁺ channel, which opens only at depolarizing potentials, and the K_ir channel, which opens mostly at hyperpolarizing potentials. At < –60 mV, K_ir rarely conducts significant outward current due to a rapid and highly voltage-dependent block by intracellular polyamines and also by Mg²⁺ (Ruppersberg 2000). The K_ir channel has been proposed for inducing the negative V_rest (Sun and Herness 1996). However, in our study, the slightly outwardly rectifying current-voltage relationship of the pH-sensitive K_leak and sizable leak current recorded at –40 mV suggests that K_ir may not be a major contributor to V_rest in these TRCs. Instead, our results that V_rest in these TRCs is correlated significantly with the leak current at –40 mV and to the pH-sensitive K_leak suggests that TASK-2 or TALK-1 is a major determinant for V_rest in these cells. However, this by no means excludes possible contributions of K_ir and other stationary conductances in control of V_rest. The V_rest in TRCs varies over wide ranges (Miyamoto et al. 2000). The fact that TASK-2 is only weakly expressed in many TRCs and that V_rest is not identical to E_K, even in cells showing strong pH-sensitive K_leak, suggests that a number of other conductances may also contribute to setting V_rest of TRCs.

Acids or sour taste are known to modify other taste sensations (Frank et al. 1983; Sakurai et al. 2000). The broad tuning of TRCs to different taste qualities makes it possible that the acid modification could occur at peripheral receptor cells levels. However, mechanisms underlying these modifications are poorly understood. This is in part due to the fact that H⁺ ions can interact with many ion channels, transport proteins, and intracellular signaling components, and in part due to the functional heterogeneity of TRCs. Acid modification may occur by direct interaction of H⁺ on ion channels that function as taste transducers, such as the epithelial Na⁺ channel (ENaCs). Protons modulate its activity and salt sensation by interaction with either the extracellular (Zhang et al. 1999) or intracellular sides of the channel (Lyall et al. 2002), or the H⁺ modification may be indirectly altering activities of ion channels and proteins that influence cell excitability. The degree of modification depends on acid concentrations and the H⁺ sensitivity of these components.

Among many potential pH-sensitive molecular targets that could change cell excitability, K_2P channels are obvious candidates, since some channels in this family regulate membrane potential and input resistance, key determinants of cell excitability. Daily complex foods consist of mixtures of taste qualities and many are slightly acidified to stimulate appetite. By blockage of K_2P channels, acids that are subthreshold for sour taste sensation may modify TRCs membrane properties, thus influencing sensations produced by other taste stimuli. Immunocytochemical studies showed that the expression level of TASK-2 in TRCs was heterogeneous, with strong expression in some but weak expression in many other TRCs. Depending on the number of TASK-2 channels, concentrations of H⁺ ions and V_rest, the shift in membrane potential induced by H⁺-dependent blockage of TASK-like channels may have diverse outcomes. Shifts that bring membrane potentials closer to threshold potentials for firing of Na⁺- and/or Ca²⁺-dependent action potentials can enhance the excitability, whereas stronger depolarization may induce firing of action potentials or may actually suppress the firing rates if prolonged depolarization occurs and deactivates voltage-gated channels. Moreover, as activities of TASKs also are regulated tightly by neurotransmitters and neuropeptides, they also present a substrate for efferent modulation and cell-to-cell regulation of TRCs (Finger et al. 1990). Therefore K_2P channels, such as the highly expressed TASK-2 channel, may provide unique molecular substrates for dynamic modulation of cell excitability.

	GRANTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
This study was supported by National Institute of Health Grants DC-00766 to S. C. Kinnamon; DK-59611, DC-02507, DC-00353, and DC-00347 to T. A. Gilbertson; and DC-00443 to W. Lin.

	ACKNOWLEDGMENTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
We thank Dr. Diego Restrepo for support and Dr. Tatsuya Ogura for assistance with data analysis. The authors thank L. Qiao for assistance with the Western blot shown in Fig. 4K.

	FOOTNOTES

 
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Address for reprint requests and other correspondence: T. A. Gilbertson, Dept. of Biology and Center for Integrated BioSystems, Utah State Univ., Logan, UT 84322-5305 (E-mail: tag{at}biology.usu.edu).

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION GRANTS ACKNOWLEDGMENTS REFERENCES

 
Akabas M, Dodd J, and al-Awqati Q. Identification of electrophysiologically distinct subpopulations of rat taste cells. J Membr Biol 114: 71–78, 1990.[CrossRef][ISI][Medline]

Applied Biosystems. Relative quantitation of gene expression. User bulletin 2, 1997. http://docs.appliedbiosystems.com/pebiodocs/04303859.pdf

Ashmole I, Goodwin PA, and Stanfield PR. TASK-5, a novel member of the tandem pore K⁺ channel family. Pfluegers 442: 828–833, 2001.

Bayliss DA, Talley EM, Sirois JE, and Lei Q. TASK-1 is a highly modulated pH-sensitive 'leak' K⁺ channel expressed in brainstem respiratory neurons. Respir Physiol 129: 159–174, 2001.[CrossRef][ISI][Medline]

Bigiani A. Mouse taste cells with glialike membrane properties. J Neurophysiol 85: 1552–1560, 2001.[Abstract/Free Full Text]

Bigiani A. Electrophysiology of Necturus taste cells. Prog Neurobiol 66: 123–159, 2002.[CrossRef][ISI][Medline]

Bigiani A and Roper SD. Reduction of electrical coupling between Necturus taste receptor cells, a possible role in acid taste. Neurosci Lett 176: 212–216, 1994.[CrossRef][ISI][Medline]

Burks CA, Hansen DR, Rao S, Lin W, Kinnamon SC, and Gilbertson TA. Rat taste buds express multiple members of the KCNK family of two-pore domain potassium channels. Chem Sense 28:558, 2003.

Bustin SA. Absolute quantification of mRNA using real-time reverse transcription polymerase chain reaction assays. J Mol Endocrinol 25: 169–193, 2000.[Abstract]

Caicedo A, Kim KN, and Roper SD. Individual mouse taste cells respond to multiple chemical stimuli. J Physiol 544: 501–509, 2002.[Abstract/Free Full Text]

Chen Y, Sun XD, and Herness S. Characteristics of action potentials and their underlying outward currents in rat taste receptor cells. J Neurophysiol 75: 820–831, 1996.[Abstract/Free Full Text]

Decher N, Maier M, Dittrich W, Gassenhuber J, Bruggemann A, Busch AE, and Steinmeyer K. Characterization of TASK-4, a novel member of the pH-sensitive, two-pore domain potassium channel family. FEBS Lett 492: 84–89, 2001.[CrossRef][ISI][Medline]

DeSimone JA, Lyall V, Heck GL, and Feldman GM. Acid detection by taste receptor cells. Respir Physiol 129: 231–245, 2001.[CrossRef][ISI][Medline]

Duprat F, Lesage F, Fink M, Reyes R, Heurteaux C, and Lazdunski M. TASK, a human background K⁺ channel to sense external pH variations near physiological pH. EMBO J 16: 5464–5471, 1997.[CrossRef][ISI][Medline]

Finger TE, Womble M, Kinnamon JC, and Ueda T. Synapsin I-like immunoreactivity in nerve fibers associated with lingual taste buds of the rat. J Comp Neurol 292: 283–290, 1990.[CrossRef][ISI][Medline]

Frank ME, Contreras RJ, and Hettinger TP. Nerve fibers sensitive to ionic taste stimuli in chorda tympani of the rat. J Neurophysiol 50: 941–960, 1983.[Abstract/Free Full Text]

Gilbertson TA. Hypoosmotic stimuli activate a chloride conductance in rat taste cells. Chem Senses 27: 383–394, 2002.[Abstract/Free Full Text]

Gilbertson TA, Avenet P, Kinnamon SC, and Roper SD. Proton currents through amiloride-sensitive Na channels in hamster taste cells. Role in acid transduction. J Gen Physiol 100: 803–824, 1992.[Abstract/Free Full Text]

Gilbertson TA, Boughter JD Jr, Zhang H, and Smith DV. Distribution of gustatory sensitivities in rat taste cells: whole-cell responses to apical chemical stimulation. J Neurosci 21: 4931–4941, 2001.[Abstract/Free Full Text]

Gilbertson TA, Damak S, and Margolskee RF. The molecular physiology of taste transduction. Curr Opin Neurobiol 10: 519–527, 2000.[CrossRef][ISI][Medline]

Gilbertson TA and Fontenot DT. Distribution of amiloride-sensitive sodium channels in the oral cavity of the hamster. Chem Sense 23: 495–499, 1998.

Gilbertson TA, Roper SD, and Kinnamon SC. Proton currents through amiloride-sensitive Na⁺ channels in isolated hamster taste cells: enhancement by vasopressin and cAMP. Neuron 10: 931–942, 1993.[CrossRef][ISI][Medline]

Girard C, Duprat F, Terrenoire C, Tinel N, Fosset M, Romey G, Lazdunski M, and Lesage F. Genomic and functional characteristics of novel human pancreatic 2P domain K⁺ channels. Biochem Biophys Res Commun 282: 249–256, 2001.[CrossRef][ISI][Medline]

Goldstein SA, Bockenhauer D, O'Kelly I, and Zilberberg N. Potassium leak channels and the KCNK family of two-P-domain subunits. Nat Rev Neurosci 2: 175–184, 2001.[ISI][Medline]

Han J, Kang D, and Kim D. Functional properties of four splice variants of a human pancreatic tandem-pore K⁺ channel, TALK-1. Am J Physiol 285: C529–C538, 2003.

Han J, Truell J, Gnatenco C, and Kim D. Characterization of four types of background potassium channels in rat cerebellar granule neurons. J Physiol 542: 431–444, 2002.[Abstract/Free Full Text]

Herness MS and Gilbertson TA. Cellular mechanisms of taste transduction. Annu Rev Physiol 61: 873–900, 1999.[CrossRef][ISI][Medline]

Hille B. Ion Channels of Excitable Membranes. Sunderland, MA: Sinauer Associates, 2001.

Kang D and Kim D. Single-channel properties and pH sensitivity of two-pore domain K⁺ channels of the TALK family. Biochem Biophys Res Comm 315: 836–844, 2004.[CrossRef][ISI][Medline]

Karschin C, Wischmeyer E, Preisig-Muller R, Rajan S, Derst C, Grzeschik KH, Daut J, and Karschin A. Expression pattern in brain of TASK-1, TASK-3, and a tandem pore domain K⁺ channel subunit, TASK-5, associated with the central auditory nervous system. Mol Cell Neurosci 18: 632–48, 2001.[CrossRef][ISI][Medline]

Kim D. Fatty acid-sensitive two-pore domain K⁺ channels. Trends Pharmacol Sci 24: 648–654, 2003.[CrossRef][Medline]

Kim D, Fujita A, Horio Y, and Kurachi Y. Cloning and functional expression of a novel cardiac two-pore background K⁺ channel (cTBAK-1). Circ Res 82: 513–518, 1998.[Abstract/Free Full Text]

Kim D and Gnatenco C. TASK-5, a new member of the tandem-pore K⁺ channel family. Biochem Biophys Res Commun 284: 923–930, 2001.[CrossRef][ISI][Medline]

Kim Y, Bang H, and Kim D. TASK-3, a new member of the tandem pore K⁺ channel family. J Biol Chem