INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The sense of taste provides organisms with critical information for regulating food intake. Sweet, umami, and salty tastes signal the presence of nutritionally important carbohydrates, proteins, and ions, whereas bitter and sour tastes generally reflect potentially toxic substances, spoiled meat, or immature fruits that may be harmful. Organic and inorganic acids, such as citric acid and HCl, elicit sour taste. Because the sourness of an acid is generally proportional to proton concentration, protons are thought to be the primary sour-taste stimulus. The anion does influence the sourness of acids, however, because organic acids elicit a stronger sour taste than inorganic acids at the same pH.

A number of sour-taste transduction mechanisms have been proposed in various species. In mudpuppy (Necturus maculosus) taste receptor cells, apically localized K⁺ channels are blocked directly by protons, leading to depolarization of the taste cells (Kinnamon and Roper 1988; Kinnamon et al. 1988). Additionally, acids can alter the coupling of gap junctions between mudpuppy taste cells, which also may contribute to acid taste transduction (Bigiani and Roper 1994). In frog taste cells, proton-gated Ca²⁺ channels and a proton transporter are believed to mediate acid transduction (Miyamoto et al. 1988; Okada et al. 1987, 1993). However, little is known whether these mechanisms operate in taste cells from mammalian species.

In mammals, several mechanisms likely contribute to sour taste, reflecting the various effects of protons on ion channels and receptors (Hille 1992). In hamster fungiform taste cells, protons have been shown to permeate amiloride-sensitive Na⁺ channels (called also epithelial Na⁺ channel, ENaC) and lead to receptor cell depolarization (Gilbertson et al. 1992, 1993). Although amiloride (30-300 µM) significantly reduced the aversiveness of citric acid in behavioral studies (Gilbertson and Gilbertson 1994), amiloride at 10 µM has little inhibition on the acid responses recorded from acid-best nerve fibers (Hettinger and Frank 1990). Because the amiloride inhibition constant for ENaC is about 0.1-0.2 µM in taste cells, additional transduction mechanisms for sour taste must exist. One sour-taste transduction mechanism that has been proposed involves acid modulation of a hyperpolarization-activated cation channel (HCN). Acids are proposed to lower the threshold for activation of this conductance, eliciting an inward current at resting potentials (Stevens et al. 2001). A Cl conductance also has been proposed to be involved in sour transduction because responses to acids are sensitive to the Cl channel blocker 5-nitro-2-[3-phenylpylamino]-benzoic acid (NPPB) (Miyamoto et al. 1998). Another mechanism that has been proposed is pH tracking by taste cells. Acid stimulation decreases intracellular pH in taste cells (Lyall et al. 1997; Stewart et al. 1998) and afferent nerve responses correlate with intracellular pH levels (Lyall et al. 2001), suggesting that changes in intracellular pH may be involved in the transduction mechanism. In addition, acids can either increase or decrease intracellular Ca²⁺ levels. Amiloride partially suppresses acid-induced [Ca²⁺]_i increases but has no effect on acid-induced [Ca²⁺]_i decreases, which may be mediated by G-protein-coupled pathway (Liu and Simon 2001).

Recent molecular studies suggest that acid-sensing ion channels (ASICs) also may be involved in acid transduction. The ASIC family is a branch of the ENaC/DEG supergene family, which is believed to play an important role in the pain perception accompanying tissue acidosis (Waldmann et al. 1999). Several subunits, such as, ASIC1, -2a, -2b, and -4 are present in taste receptor cells (Brand 2000; Buffington et al. 2002; Liu and Simon 2001; Ugawa et al. 1998). According to Ugawa et al. (1998), the first cloned ASIC subunit from rat taste receptor cells is identical with the cloned MDEG1/ASIC2a (Waldmann et al. 1996, 1999), known also as BNC1 (Price et al. 1996) and BNaC1 (Garcia-Anoveros et al. 1997). When expressed heterologously in Xenopus oocytes, the taste MDEG1/ASIC2a produced a large, rapidly activating cation current that was partially amiloride sensitive. In addition, ASIC2a is co-localized with a modulatory subunit MDEG2/ASIC2b in single taste cells, and these two subunits can form heteromeric channels in heterologous expression systems (Shimada and Ugawa 2001). It is not known if all the ASIC subunits expressed in taste cells are co-localized and participate in forming ASIC channels. ASIC subunits can form homomeric and heteromeric channels, and assembly with modulatory subunits (such as ASIC2b and -4) may further diversify the channel kinetics, proton and amiloride sensitivities. To determine the contribution of ASICs to sour-taste transduction, it is important to examine whether taste cells respond to acidic stimuli with current profiles typical of ASICs.

In this study, we have used the whole cell configuration of the patch-clamp technique (Hamill et al. 1981) and imaging with the Na⁺-sensitive dye SBFI-AM to examine responses of single taste cells from rat vallate papillae to acidic stimuli. Specifically, we addressed whether taste cells respond to acids with currents similar to those mediated by ASICs in neurons and heterologous systems. Our results demonstrate that acids induce a large depolarizing current in most taste cells and that the current shares general properties with currents mediated by ASICs. Further, these currents are accompanied by increases in intracellular Na⁺ levels. Therefore ASICs or ASIC-like channels may play an important role in sour-taste transduction. Preliminary results have been published in abstract form (Lin and Kinnamon 1999; Lin et al. 2000).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Isolation of taste buds

Taste buds were freshly isolated from male Sprague-Dawley adult rat vallate papillae. Briefly, rats were killed with CO₂ inhalation followed by cervical dislocation. The tongue was removed and placed in cold Tyrode's solution. An enzyme mixture containing 3 mg dispase, 0.7 mg collagenase B (Boehringer Mannheim, Indianapolis, IN), and 1 mg trypsin inhibitor (type I-S; Sigma Chemical, St. Louis, MO) in 1.0 ml of Tyrode's was injected beneath the epithelium surrounding vallate papillae. The tongue was then incubated in Ca²⁺- and Mg²⁺-free oxygenated Tyrode's for 50 min at room temperature, or until the epithelium could be gently separated from the underlying muscle and connective tissue. The stripped lingual epithelium containing vallate taste buds was pinned serosal side up in a silicone elastomer (Sylgard, Dow Corning, Midland, MI)-coated petri dish and incubated in Ca²⁺- and Mg²⁺-free Tyrode's for 10-20 min. Taste buds were removed by gentle suction with a glass pipette and plated onto Cell-Tak (cell and tissue adhesive; Collaborative Research, Bedford, MA)-coated glass slide chambers.

Solutions and chemicals

The extracellular recording solution (Tyrode's) was composed of (in mM) 140 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid buffer (HEPES), 10 glucose, and 10 sodium pyruvate (pH 7.4 with NaOH). The Ca²⁺- and Mg²⁺-free Tyrode's for isolating taste buds contained 2 mM bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA, Molecular Probes, Eugene, OR). To maintain constant ionic concentration, the pH values of stimuli were adjusted from a starting pH of 7.4 with either 1 M citric acid or HCl for acidic solutions or 1 M NaOH for basic solutions. A Cl channel blocker, NPPB (Calbiochem, San Diego, CA), was dissolved in DMSO and diluted 500-fold with Tyrode's to 100 µM. An epithelial Na⁺ channel blocker, amiloride (Sigma Chemical), was used at final concentration of 100 or 500 µM. These compounds were bath-applied to taste cells. Low Cl Tyrode's contained 140 mM Na⁺ gluconate instead of NaCl. The standard 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). In low-Cl pipette solution, KCl (130 mM) was replaced with equal molar K gluconate. In the experiments where gramicidin was used, the bath solution with various Na⁺ levels was made by replacing NaCl with N-methyl-D-glucamine (pH 7.4 with HCl). Solutions were gravity-fed with an ~0.15 ml/s flow rate to a recording chamber (1.2 cm diam and 0.2 cm high).

Whole cell recordings

The whole cell voltage-clamp technique was used to record membrane currents (Hamill et al. 1981). The glass pipettes for recording were pulled from microhematocrit capillary tubes (Scientific Products, McGaw Park, IL) with a two-stage vertical puller (model PB-7; Narishige, Tokyo) or a Flaming/Brown micropipette puller (Model P-97; Sutter Instrument, Novato, CA). Pipette resistance was 3-6 M when filled with normal pipette solution and 4-8 M when filled with the low-Cl pipette solution. Membrane currents were low-pass filtered at 2 kHz and recorded with an Axopatch patch-clamp amplifier (Model 200B, Axon Instruments, Foster City, CA). Voltage-activated Na⁺ and K⁺ currents were generated by applying depolarizing voltage steps from a holding potential of 80 mV; these were used to distinguish taste cells from nonsensory epithelial cells. Hyperpolarizing voltage pulses (20 mV) were used to monitor membrane conductance. All voltage commands were generated by an Indec laboratory computer system (Sunnyvale, CA). Steady membrane currents were recorded on a strip chart recorder (Linear) as well as on magnetic tape recording system for later analysis.

Na+ imaging

Intracellular Na⁺ levels in taste cells of isolated taste buds were measured using the membrane-permeable Na⁺-sensitive dye SBFI-AM (Molecular Probes). Taste buds were loaded with 10 µM SBFI-AM in the presence of a dispersing reagent, pluronic F-127 (final concentration <0.02%; Molecular Probes) for 40-60 min and subsequently were washed with Tyrode's for 20 min to allow hydrolysis of the AM ester. Images were acquired with an intensified CCD camera (IC100-ICCD, Paultek Imaging) through an oil-immersion objective lens (Fluor ×40, 1.3 NA, Nikon) of an inverted microscope (Diaphot TMD, Nikon). SBFI was excited at 350 nm, and the emitted fluorescence was collected with a 510-580-nm band-pass filter. Gramicidin (10 µM), a Na⁺ ionophore, was bath applied to taste cells to equilibrate [Na⁺]_i with external Na⁺ concentration. The video images were captured with an image processing board (Image Lightning 2000, Axon Instruments) and stored at 0.5-2 Hz with a PC computer using Axon Imaging Workbench software (Axon instruments). Na⁺ levels were obtained by measuring the fluorescence intensity averaged over cell area with time.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Types of acid responses

Single taste cells in isolated taste buds were voltage-clamped at 80 mV. Bath application of citric acid (pH 5) induced a response in most cells tested (148 of 168). Several different types of responses were observed, as illustrated in Fig. 1. The most common type of response (40% of responding cells) consisted of a single peak with a large, rapidly activating inward current that showed varying rates of desensitization (Fig. 1A). The second most common type of response (30% of responding cells) consisted of two peaks of inward current with the second peak usually showing slower desensitization (Fig. 1B). Often these peaks were not separated clearly in time. Because both single and double peak desensitizing currents could be recorded from a single cell during repeated stimulation, it is possible that single peak currents could represent two peaks that have merged. The third type of response consisted of an inward current with relatively slow desensitization kinetics (11% of responding cells; Fig. 1C). Several cells (36%) exhibited an additional OFF response; i.e., an inward current that was induced on acid removal (Fig. 1, A and B). These three types of responses along with the OFF response were associated with an increase in membrane conductance. They resemble responses mediated by cloned ASIC channels and sensory neurons that express ASIC channels (Bevan and Yeats 1991; Krishtal and Pidoplichko 1981; Waldmann et al. 1999). In addition, two other types of responses were recorded; all were associated with a decrease in membrane conductance. One consisted of a nondesensitizing inward current (13% of responding cells; Fig. 1D). The other was a slowly activating outward current, which followed the dominant inward current in most responses. This current persisted long after the acid was removed and could be seen easily due to the elevated baseline (Fig. 1, A and B). In some taste cells (5%), the outward current was the only response observed (Fig. 1E). These cells tended to be "leaky;" i.e., they had a low input resistance during whole cell recording. The responses associated with a decrease in membrane conductance were not examined further.

View larger version (16K): [in this window] [in a new window]   Fig. 1. Citric acid-induced currents profiles in vallate taste cells. A: a large inward current consisted of 1 peak and desensitized in response to prolonged citric acid stimulation. Note that an OFF response of inward current (arrow head) developed on the removal of the stimulus, and a delayed outward current (double arrow) followed the inward currents, similar to E. B: a typical response to pH 5 consisting of 2 peaks (a single arrow denotes 2nd peak). The time for activation and half inactivation for the 1st peak were much shorter than the 2nd peak. C: citric acid (pH 5) evoked a response showing slow desensitization kinetics. D: a sustained response accompanied by a decrease in conductance. Holding potential = 60 mV. E: citric acid (pH 5) evoked a slowly activating outward current accompanied by a decrease in membrane conductance. Note that the outward current was often found associated with the dominant inward current (Fig. 1, A and B). Except as mentioned all recordings were obtained at a holding potential of 80 mV.

Activation and desensitization kinetics

Current magnitude and activation kinetics were analyzed for all cells that showed relatively fast activating inward currents in response to acid stimulation. The mean current amplitude for single peaks was 59.7 ± 4.6 (SE) pA (n =60). In cells with two inward peaks, the mean current amplitude of the first peak was 63.8 ± 7.5 pA and the second peak was 44.6 ± 4.5 pA (n = 44). The activation time from response onset to peak current was 6.0 ± 0.4 s for cells exhibiting a single peak (Fig. 1A) and 5.9 ± 0.6 s and 17.4 ± 1.2 s for cells exhibiting two peaks, respectively. To analyze desensitization kinetics, we chose to analyze responses in those cells that showed little or no outward current because this current overlapped in time with desensitization. In addition, we analyzed only cells that had either a single peak or two peaks that were largely separated in time. Both single peaks and double peaks of inward current desensitized during stimulation. Time for half-desensitization of single peaks was 9.9 ± 1.0 s (n = 18). For double peaks, the first peak desensitized much faster than the second peak (half-desensitization time 3.0 ± 0.5 s compared with 22.9 ± 4.1 s; n = 7). Some cells responded to citric acid with only relatively sustained inward currents (Fig. 1C). The average amplitude for these currents was 30.0 ± 4.1 pA with an activation time of 9.5 ± 0.9 s and a half desensitization time of 32.5 ± 2.7 s (n = 16). Thus citric acid induced both relatively transient and sustained inward currents in taste cells. Similar transient and sustained responses to acids have been demonstrated in sensory neurons of trigeminal ganglia, dorsal root ganglia, and heterologous cells expressing cloned ASIC subunits (Bevan and Yeats 1991; Krishtal and Pidoplichko 1981; Liu and Simon 1999).

pH dependence of inward currents

ASIC-mediated currents all show strong pH dependence. We therefore examined citric acid-induced currents at different pH values. As shown in Fig. 2A, a pH drop in the bath solution from 7.4 to 6.5 was necessary for activation of the inward current (the 0-current pH was pH 6.7). The size of the current increased dramatically, and the activation time shortened with increasing extracellular H⁺ concentration and was not saturated at pH 4. A dose response curve summarizing data from four to seven cells is presented in Fig. 2B.

View larger version (11K): [in this window] [in a new window]   Fig. 2. The pH dependence of acid-induced currents. A: actual current traces evoked by pH drops from a resting pH of 7.4. Note the much larger current amplitude and shorter activation time at lower pH values. All traces are from the same cell. B: the dose response curve for citric acid. All acid responses were normalized to the peak amplitude of the current induced by pH 5. Each point represents the mean ± SE of 4 to 7 cells.

A well-known phenomenon in sour taste is that organic acids, such as citric acid, taste sourer than inorganic acids at the same pH values. Heterologously expressed MDEG1/ASIC2a from rat vallate taste cells also shows bigger responses to citric acid than to HCl (Ugawa et al. 1998). We reasoned that the acid-induced inward current in taste cells should show similar features if ASICs are involved in sour-taste transduction. Responses to HCl and citric acid at the same pH were shown in Fig. 3. Although responses to HCl were strongly pH dependent and had a similar pH threshold, current amplitudes induced by HCl at pH values of 5.5, 5, and 4.5 were significantly lower than those induced by citric acid at the same pH (t-test, P < 0.05; n = 8, 11, and 7 respectively). Our results are thus consistent with many behavioral and psychophysical studies showing that organic acids taste sourer than inorganic acids at the same pH.

View larger version (19K): [in this window] [in a new window]   Fig. 3. Comparison of responses to citric acid and HCl at the same pH. A: current response to citric acid and HCl at pH 5 recorded from the same cell. Note that citric acid induced a larger and more rapidly activating response than HCl. B: each bar represents the mean ± SE of currents induced by citric acid () and HCl () at pH 4.5, 5, and 5.5; n = 7-11. The values were statistically significant at P < 0.05 (paired t-test).

ASIC-mediated acid responses depend not only on the stimulating pH but also on the conditioning pH. In psychophysical studies, adaptation with acidic solutions causes a decrease in the sour-taste intensity of acid stimuli (Ganzevles and Kroeze 1987). To test the effect of adaptation on acid-induced currents in taste cells, we incubated cells at various pH levels for 1 min before applying citric acid at pH 5. These data are shown in Fig. 4A and summarized in B. Relative to the normal resting condition of pH 7.4, larger transient currents were observed when cells were adapted to pH 8, suggesting that some current is desensitized at the physiological pH of 7.4. Although a drop in pH from 7.4 to 6.5 rarely produced a significant inward current (e.g., Fig. 2A), incubation at pH 6.5 substantially reduced subsequent responses to pH 5. The data demonstrate acid-induced responses were correlated negatively to the conditioning pH level, indicating strong self-adaptation. As this phenomenon also was observed in all heterologously expressed ASICs (Waldmann et al. 1999), our results support the notion that ASICs may mediate the proton-evoked inward currents in taste cells.

View larger version (14K): [in this window] [in a new window]   Fig. 4. Effects of adapting bath pH on acid responses. A: citric acid (pH 5)-induced inward current in cells preincubated at various adapting pH values. The cell was held at 80 mV. B: histogram showing acid responses from all cells (n = 4-6) were normalized to the peak amplitude of the response induced by a pH drop from 8 to 5. Note that adapting with pH values 7 significantly decreased the current amplitude induced citric acid at pH 5 (P < 0.05, paired t-test).

Sensitivity to amiloride

Cloned ASICs belong to ENaC/DEG superfamily, where sensitivity to the diuretic amiloride is a typical pharmacological feature. Because the amiloride-sensitive ENaC also reportedly plays a role in sour taste in hamster fungiform taste cells (Gilbertson et al. 1993), we tested the amiloride sensitivity of acid-evoked currents to determine whether ENaC or ASIC subunits may mediate the response in rat vallate taste cells. Bath application of 30 µM amiloride, which is sufficient to block all the current mediated by ENaCs (inhibition constant, IC₅₀: 0.2 µM) (Gilbertson et al. 1993), had no significant effect on the inward current induced by citric acid at pH 5. At higher concentrations, amiloride (100-500 µM) blocked ~40% of the transient (peak) current (Fig. 5; 16 of 18 cells), and 28% of the relatively sustained current (7 of 10 cells). The effect of amiloride was reversible. This result is in agreement with previous findings showing that the sustained current mediated by ASIC subunits is less sensitive to amiloride than the transient current (de Weille et al. 1998; Waldmann et al. 1997a). Interestingly, amiloride also blocked 56% of the OFF response (3 of 3 cells). These results suggested that ASIC subunits, rather than the ENaC, likely mediate the acid-induced inward current in rat vallate taste cells.

View larger version (17K): [in this window] [in a new window]   Fig. 5. Inhibition by amiloride. Bath application of amiloride (100 µM) partially reduced the acid-induced current. Top: current in the absence of amiloride. Bottom: 20 s during the acid response was removed (break).

Current-voltage relationship of acid-induced currents

ASICs are proton-gated Na⁺ channels with a high Na⁺ to K⁺ permeability. We next determined the reversal potential of the acid-induced currents. Figure 6A shows the citric acid induced responses at different holding potentials. Responses at more negative potentials had larger amplitudes as well as more rapid activation and desensitization. At holding potentials positive to 40 mV, only one peak inward current in response to citric acid at pH 5 could be recorded. Therefore when two peaks were recorded at more negative holding potentials, the current-voltage relationship was based on the first peak amplitude. The extrapolated reversal potential was ~45 mV, suggesting that Na⁺ contributes significantly to the response (Fig. 6B). Although our data suggest that acid-induced currents are predominantly mediated by a Na⁺ conductance, acid-induced currents in frog taste cells reportedly are mediated by a Ca²⁺ conductance (Okada et al. 1994). Thus we also examined the effect of changes in extracellular Ca²⁺ on acid-evoked responses. Increasing the Ca²⁺ concentration from 1 to 5 or 10 mM Ca²⁺ slightly decreased the current amplitude, suggesting that Ca²⁺ is not the primary current carrier for acid responses (data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 6. Current-voltage relationship of acid-induced currents. A: acid responses at various holding potentials; all recordings are from the same cell. B: the I-V relationship of the acid responses. The extrapolated reversal potential was ~45 mV, which is near the Na+ equilibrium potential.

Na+ imaging experiments

The results above suggested that Na⁺ is the principal current carrier for the acid-induced inward current. To demonstrate directly that Na⁺ influx occurs during acid stimulation, we measured intracellular Na⁺ levels using the Na⁺-sensitive dye SBFI-AM. As showed in Fig. 7A, a drop of the bath pH from 7.4 to 4 increased the fluorescence intensity. The elevation of intracellular Na⁺ level was pH dependent, as citric acid at pH 4 elicited significantly bigger responses than at pH 5 (n = 40; paired t-test, P < 0.001; Fig. 7C). Also, citric acid elicited slightly larger responses than HCl at the same pH (Fig. 7D; n = 27); however, these differences were not statistically significant. To determine if changes in the fluorescence intensity of SBFI loaded in taste cells were resulted from the Na⁺ influx, the Na⁺ ionophore gramicidin (10 µM) was applied to the bath solution containing various levels of Na⁺. As shown in Fig. 7B, the SBFI fluorescence intensity was altered in a Na⁺-dependent manner. Higher extracellular Na⁺ concentrations resulted in stronger fluorescence, indicating that changes in the fluorescence intensity of SBFI are due to changes in the intracellular Na⁺ concentration of rat taste cells. These results are consistent with those obtained from our patch-clamp recordings, suggesting that acid-induced Na⁺ influx plays an important role in rat sour-taste transduction.

View larger version (28K): [in this window] [in a new window]   Fig. 7. Acid-induced increases in intracellular Na+ levels. A: 2 representative cells in the same isolated taste bud increased fluorescence intensity in response to citric acid (pH 5), indicating increases in intracellular Na+ concentration. B: a Na+ ionophore, gramicidin, was applied to the bath solutions containing various concentration of Na+. Note that changes in the extracellular Na+ concentrations altered the SBFI fluorescence intensity in a [Na+]o-dependent manner. C: the acid-induced increase in intracellular Na+ was pH dependent. Responses to citric acid pH 4 were significantly larger than those to pH 5 (P < 0.001, n = 39). D. At pH 5, HCl also induced an increase in intracellular Na+, but the response was slightly smaller than those to citric acid. However the difference was not statistically significant (P = 0.45, n = 26).

Effects of VR1 and Cl channel antagonists

The results above imply that ASICs may mediate responses to sour stimuli in taste cells. However, the vanilloid receptor, VR1, which can be activated by capsaicin, heat, and protons, has been reported to be present in taste cells (Liu and Simon 2001). Therefore we examined the possible role of VR1 in acid transduction. Bath application of capsaicin (1-20 µM) induced a very small inward current in a few cells (data not shown). Nonetheless, the capsaicin-induced current never mimicked those induced by citric acid. Further, the VR1 competitive antagonist capsazepine (10 µM) did not suppress the acid-evoked current (data not shown). Therefore VR1 apparently does not contribute significantly to the acid-induced desensitizing currents in rat vallate taste cells.

An NPPB-sensitive Cl channel also has been proposed to mediate sour taste in rat fungiform taste cells (Miyamoto et al. 1998). To investigate whether a Cl conductance contributes to acid-evoked currents in vallate taste cells, we applied the Cl channel antagonist NPPB (100 µM) prior to and during acid stimulation. As shown in Fig. 8A, NPPB partially inhibited the inward currents evoked by citric acid. This was surprising, given the acid-induced current reversed at a positive potential ~45 mV. To determine if the conductance that was blocked by NPPB, we replaced 130 mM KCl with equimolar K⁺ gluconate in the pipette solution, which shifted the Cl equilibrium potential (E_Cl) from 2 mV to 58 mV. The amplitude of acid-induced inward currents and the current-voltage relationship did not change significantly under these conditions. The reversal potential of citric-acid-induced inward currents remained positive despite the negative shift in E_Cl (Fig. 8, B and C). Furthermore, NPPB suppressed acid-evoked currents at E_Cl, which suggests that NPPB may suppress conductance(s) other than Cl in rat vallate taste cells. Our results are consistent with the observation in sensory neurons that changing extracellular Cl does not alter either acid-evoked transient or sustained current components (Benson et al. 1999).

View larger version (21K): [in this window] [in a new window]   Fig. 8. Effect of 5-nitro-2-[3-phenylpylamino]-benzoic acid (NPPB) and Cl replacement on acid-induced inward currents. A: the Cl channel blocker NPPB (100 µM) partially inhibited the inward current. B: current-voltage relationships of normalized acid-induced responses. These currents were recorded with pipettes containing 130 mM K+ gluconate (ECl = 58 mV). Note that the extrapolated acid-induced reversal potential was ~40 mV. Each point represented an average of 3-7 cells. C: summary of NPPB effects. Cells were held either at 80 mV to maximize the possible Cl current when the Cl equilibrium potential (ECl) was 2 mV or held at 58 mV to eliminate the Cl conductance when ECl = 58 mV. There was no significant difference between the two groups and the effect of NPPB persisted even when the cells were held at the Cl equilibrium potential (ECl) where there is no net Cl current.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study examined responses to sour (acid) stimuli in isolated taste receptor cells of rat vallate papilla using whole cell voltage-clamp recording and Na⁺ imaging. Based on previous reports showing that acids permeate tight junctions and interact with the basolateral membrane (Stewart et al. 1998), we applied acidic stimuli by rapid bath perfusion to taste cells without attempting to stimulate the apical membrane selectively. Our data show that acids are potent taste stimuli for vallate taste cells. Responses to acids were dominated by a large, rapidly activating inward current that showed desensitization with prolonged stimulation. Current amplitudes were pH dependent and also were affected by the adapting pH of the bath solution. The current reversed near the Na⁺ equilibrium potential and was accompanied by increases in intracellular Na⁺ levels, suggesting that Na⁺ is the predominant current carrier. Further, the current was partially suppressed by amiloride. These results provide physiological evidence for the presence of functional ASIC or ASIC-like channels in vallate taste cells. The large amplitude of the current and its existence in a high percentage of taste cells imply a prominent role of the pathway in sour-taste transduction.

Comparison of acid-induced currents in taste cells with currents in ASIC-expressing cells

Our results are consistent with molecular studies showing the presence of multiple subunits of MDEG1/ASIC2a, MDEG2/ASIC2b, ASIC1, and ASI4 in taste receptor cells (Brand 2000; Buffington et al. 2002; Liu and Simon 2001; Ugawa et al. 1998; Shimada and Ugawa 2001) and physiological studies showing that ASICs function as acid-sensing Na⁺ channels in other sensory neurons. In general, the dominant acid-induced current in taste cells shares characteristics with proton-activated currents obtained from sensory neurons of dorsal root ganglia and trigeminal ganglia (Bevan and Yeats 1991; Krishtal and Pidoplichko 1981; Liu and Simon 2001) and from heterologous cells that express cloned ASICs (Waldmann et al. 1999). DRG neurons, when studied under similar recording conditions, i.e., the neurons were held at 80 mV and bathed in pH 7.4 saline that consisted primarily of NaCl, responded to acids with a large transient, rapidly activating inward current. When challenged with a stimulus of pH <6, ~45% of the neurons also showed a second phase of inward current with a much slower time course (similar to Fig. 1B). This second phase often developed after the first phase had decayed substantially, and reached a peak after ~5-20 s. The current then desensitized to various degrees (20-60%), leaving a substantial sustained component. These proton-activated currents have reversal potentials close to the Na⁺ equilibrium potential, suggesting that Na⁺ is the principal current carrier (Bevan and Yeats 1991; Krishtal and Pidoplichko 1981).

The recently identified ASIC subunits, expressed in these sensory neurons, have been proposed to mediate acid-evoked currents and also may be responsible for the heterogeneity of the currents (Chen et al. 1998; Lingueglia et al. 1997; Waldmann et al. 1997a,b). ASIC subunits exhibit similar or distinct patterns of activation kinetics, pH dependence, and tissue specificity when expressed heterologously. Homomeric ASIC1a or ASIC2a channels all conduct an acid-evoked transient Na⁺ current, but the kinetics of ASIC2a is slower (de Weille and Bassilana 2001). Expression of homomeric ASIC1b or ASIC3 results in both transient and sustained currents. These homomeric channels do not closely mimic the native proton-activated Na⁺ channels in neurons. For example, the pH sensitivity of the relatively sustained current in the cloned ASIC 3 required much lower pH to activate than currents reported in DRG neurons. In heterologous cells, ASIC subunits can assemble with other functional subunits or with modulatory subunits, such as ASIC2b and ASIC4 (Grunder et al. 2000) to form functional heteromultimeric channels. These heteromultimeric channels often differ from their homomeric counterparts in current kinetics and in sensitivity to pH, amiloride, and toxins (de Weille and Bassilana 2001; Escoubas et al. 2000; Waldemann et al. 1999). In addition, functional ASICs can be modulated by neuropeptides such as FF and FMRFamide, which can potentiate the sustained component of acid-evoked currents in these homomeric channels (Askwith et al. 2000). Thus different combinations of subunits in vivo and peptide modulation may further diversify responses to protons.

Although the acid-induced current in many taste cells showed general properties of ASIC-mediated responses, it does not closely resemble properties of any particular subunit-formed channels. One example is the pH sensitivity. The acid-sensitive current in taste cells is activated at pH 6.5. In comparison, acid activates homomeric ASIC1a channels at pH 7, with pH_0.5 = 6.5; ASIC2a channels at pH 5.5-6, with pH_0.5 = 4.7-4, with heteromeric ASIC1a and 2a channels in between (Bassilana et al. 1997; de Weille and Bassilana 2001). Another example is amiloride sensitivity. The acid-activated current in taste cells is partially suppressed by 100 µM amiloride. The amiloride-sensitivity of ASIC channels is subunit dependent, varying from ASIC1a (IC₅₀: 14 µM), to ASIC1a and 2a (IC₅₀: 36.2 µM), to ASIC2a (IC₅₀: 72.5 µM), with maximum suppression <50% of the peak current (de Weille and Bassilana 2001). In this regard, the acid-activated current in taste cells is most similar to that of ASIC2a. However, the acid-activated current in taste cells often possesses a second peak and a relatively sustained component (e.g., Fig. 1). The true nature of these secondary components is not known currently. It is possible that the second peak component is resulted from the activation of intracellular Ca²⁺- or Na⁺-mediated signaling pathways because acids induce changes in intracellular Ca²⁺ (Liu and Simon 2001) and Na⁺ levels. However, given the fact that taste cells express multiple ASIC subunits, it is also likely that taste cell ASIC channels consist of a complex of homomeric and heteromeric channels, which may include channels assembled with modulatory subunits. Also it is possible that ASIC subunits in taste cells assemble with as yet unidentified subunits or that they may be associated with accessory proteins or peptides that modify channel activity. Finally, the acid-induced current may represent the contribution of other acid-modulated channels with similar kinetics but unrelated to ASICs, such as HCN1 and 4 (Stevens et al. 2001). Recordings from taste cells in which specific ASIC subunits have been deleted will be necessary to confirm the role of the ASIC subunits in sour transduction.

Acid-evoked responses in gustatory afferent neurons

Taste receptor cells in the rat vallate papilla are innervated by the glossopharyngeal nerve (GL). Approximately half of the GL nerve fibers respond best to acid stimuli. The response threshold for these acid-best fibers is 1 mM HCl, and responses increase markedly without saturation with increasing concentration up to 30 mM HCl. In addition, a significant number of quinine (bitter) or sucrose-best (sweet) fibers respond to acid with moderate or weak increases in firing rates (Frank 1991; Hanamori et al. 1988). Our results are in good agreement with the results obtained from nerve recordings. First, although most taste cells responded to citric acid (pH 5), the responses varied in magnitude, from a few picoampere up to 140 pA in some cells. Second, both transient and sustained responses were recorded from taste cells as well as nerve fibers. Third, large OFF responses were common in recordings from taste cells as well as nerve fibers (DeSimone et al. 1995). Our data also are consistent with other important features of sour taste, such as larger responses to organic acids than inorganic acids at the same pH, and self-adaptation. These are well known phenomena that have been demonstrated in many psychophysical and animal behavioral studies (Ganzevles and Kroeze 1987), although underlying mechanisms is not known.

Contribution of other pathways to sour-taste transduction

As indicated in the INTRODUCTION, a number of mechanisms for acid transduction have been proposed, and sour taste likely results from the integration of multiple pathways. In addition to activation of ASICs, other mechanisms in mammals include proton permeation of the amiloride-sensitive Na⁺ channel ENaC (Gilbertson et al. 1992, 1993), proton modulation of HCN channels (Stevens et al. 2001), and intracellular acidification (Lyall et al. 1997, 2001). Under our recording conditions, ENaC should not have contributed to acid responses because proton permeation of ENaC requires the absence of extracellular Na⁺. Also, 10 mM HEPES was present in the intracellular recording solution, which should buffer large changes in intracellular pH. However, intracellular pH does normally track extracellular pH (Lyall et al. 1997), and recent data suggest that changes in intracellular pH correlate with afferent nerve discharges (Lyall et al. 2001). A possible source of the intracellular acidification is ASICs because in the absence of Na⁺, the channels become permeable to protons (Waldmann et al. 1997a,b). This could occur in vivo if a substantial proportion of ASICs is present on the apical membrane where Na⁺ concentrations are usually low. Indeed, we have found in preliminary experiments that replacement of Na⁺ with a nonpermeable cation results in a shift of the reversal potential of acid-induced currents to more positive potentials (data not shown). Detailed studies were not undertaken due to the rapid deterioration of the preparation under these conditions.

Another mechanism that has been suggested is an NPPB-sensitive Cl conductance. Although NPPB blocked a significant portion of acid responses, our data suggest that the NPPB-sensitive acid responses were not mediated by a Cl conductance because varying the Cl concentration in either extracellular or intracellular recording solutions failed to change the amplitude or reversal potential of the acid-induced response. One possible explanation of the NPPB effect in our study is a reduction of the ionic strength of the stimulating solution because NPPB caused some precipitation of solutes when added to acidic solutions.

Finally, we observed taste cell responses that did not resemble current profiles typical of ASICs (e.g., Fig. 1, D and E). It is likely that these responses are transduced by mechanisms other than ASICs, although we did not investigate these in detail. It is not surprising that multiple mechanisms contribute to sour-taste transduction, given that taste cells are a heterogeneous population of cells and different cells express different compliments of ion channels (Akabas et al. 1990).

Are ASICs located apically or basolaterally?

In vivo taste stimuli interact primarily with the apical membrane due to the presence of tight junctions between taste cells and between surrounding perigemmal cells. However, small inorganic ions such as H⁺, Na⁺, K⁺, and Cl may diffuse through the tight junctions resulting in an increase in the concentration of these ions in the basolateral region (Elliott and Simon 1990; Heck et al. 1984; Simon et al. 1993; Ye et al. 1991, 1994). Because our experiments were performed on cells in isolated taste buds, it is possible that most of the acid-induced responses are due to the stimulation of channels or receptors located at the basolateral membrane. Acids have been shown to stimulate trigeminal nerve fibers (Bryant and Moore 1995). Because trigeminal nerve endings are located deep within the oral epithelium, the fact that acids are effective stimuli to the nerve implies that acids must penetrate the epithelium in concentrations sufficient to cause an adequate change in intraepithelial pH in the vicinity of the nerve endings. Results from chorda tymani nerve recording showed that HCl responses were not influenced by voltage perturbations applied across the intact lingual epithelium, suggesting that acid responses are transduced primarily by channels on the basolateral membrane of taste cells (Stewart et al. 1998). Interestingly, ASIC2a immunoreactivity was present on both apical and basolateral membranes of vallate taste cells (Ugawa et al. 1998), indicating that both apical and basolateral channels may participate in taste transduction. Further experiments will be required to determine the relative contribution of apical and basolateral channels to sour transduction.