Journal of Neurophysiology


Bigiani, A., R. J. Delay, N. Chaudhari, S. C. Kinnamon, and S. D. Roper. Responses to glutamate in rat taste cells. J. Neurophysiol. 77: 3048–3059, 1997. We studied taste transduction in sensory receptor cells. Specifically, we examined the actions of glutamate, a significant taste stimulus, on the membrane properties of taste cells by applying whole cell patch-clamp techniques to cells in rat taste buds isolated from foliate and vallate papillae. In 55 of 91 taste cells, bath-applied glutamate, at concentrations that elicit taste responses in the intact animal (10–20 mM), produced one of two different responses when the cell membrane was held near its presumed resting potential, −85 mV. “Sustained” glutamate responses were observed in the majority of taste cells (51 of 55) and consisted of an outward current (reduction of the maintained inward current). Sustained glutamate responses were voltage dependent, were decreased by membrane depolarization, and were accompanied by a reduction in membrane conductance. An analysis of the reversal potential of sustained responses in different ionic conditions and the effect of ion substitutions suggested that the currents were carried by cations. The data suggest that sustained responses are mediated by the closure of nonselective cation channels. Other taste cells (4 of 55) responded to glutamate with a transient inward current—so-called “transient” responses. Transient glutamate responses were voltage dependent and Na+ dependent, and appeared to be generated by nonspecific cation channels activated by glutamate. l(+)-2-amino-4-phosphonobutyric acid (l-AP4), a specific agonist of a metabotropic glutamate receptor (mGluR4) recently identified in rat taste cells and believed to be involved in taste transduction, mimicked the sustained glutamate responses. These findings indicate that glutamate, at concentrations at or slightly above threshold for taste in rats, produces two different membrane currents. The properties of these two responses suggest that there may be two different sets of nonspecific cation channels in taste cells, one closed by glutamate (sustained response) and the other opened (transient response). Our findings on the effect of l-AP4 suggest that the sustained response is the membrane mechanism mediating, at least in part, taste transduction for glutamate.


Amino acids, especially glutamate, are significant taste stimuli in vertebrates. Glutamate (occurring as the free amino acid, not bound in protein) is a key molecule in determining the flavor of many foods, including meat, cheese, fish, and vegetables (Fuke and Konosu 1991). Many investigators believe that the taste of glutamate is a separate and independent taste quality on par with sweet, sour, salty, and bitter (Faurion and Vayssettes 1990; Ninomiya and Funakoshi 1989; Schiffman and Gill 1987; Yamaguchi 1987, 1991). Furthermore, glutamate seems to alter the preference for other gustatory stimuli (e.g., Mosel and Kantrowitz 1952; Pritchard and Norgren 1991). Ingestion of glutamate may also be involved in important vital functions such as the regulation of food intake and the regulation of protein metabolism (Mori et al. 1991; Nijima 1991; Rogers and Blundell 1990; Viarouge et al. 1991, 1992). In addition to its natural occurrence in many foods, glutamate is added to a wide range of foods worldwide in the form of monosodium glutamate.

Despite its potency as a taste stimulus, there is a lack of understanding about how glutamate acts on taste receptor cells. It is likely that glutamate binds to apical membrane receptors on chemosensory cells of taste buds. Activated receptors could elicit a cellular response either through intracellular second-messenger cascades (i.e., G-protein-coupled or metabotropic receptors) or by being directly linked to ion channels (i.e., ionotropic receptors). Indeed, there is evidence for both types of glutamate receptors in taste buds. For example, ionic conductances in response to glutamate have been demonstrated in phospholipid bilayers into which epithelial membranes from vallate/foliate taste regions were incorporated, implying the presence of ionotropic glutamate receptors (Teeter et al. 1992). Isolated taste buds and taste cells from mice respond to applied glutamate, but the underlying receptors have not been well characterized (Hayashi et al. 1995; Lin et al. 1996; Sugimoto 1994, 1996). With the use of optical methods to record membrane potentials and intracellular Ca2+ in mouse taste cells, Hayashi et al. (1996) showed that there may be multiple receptors (metabotropic and ionotropic) for glutamate. Lastly, Chaudhari et al. (1996) identified specific metabotropic and ionotropic glutamate receptors in lingual epithelium from foliate and vallate papillae in rats. These workers demonstrated that one particular metabotropic glutamate receptor (mGluR4) was expressed in taste buds per se, and may be involved in taste transduction for glutamate. Yet, to date, little is known about the membrane mechanisms involved in the transduction of glutamate taste.

We have addressed this question by applying glutamate to taste cells from rat foliate and vallate papillae and recording responses with patch-clamp techniques. Our data indicate that glutamate elicits two kinds of membrane responses and that one of these responses is mimicked by l(+)-2-amino-4-phosphonobutyric acid (l-AP4), a specific ligand for mGluR4. An abstract of these data has been published (Bigiani et al. 1995).


Isolation of taste buds

Male Sprague-Dawley rat pups 10–30 days old and weighing20–50 g were used in this study. Foliate and vallate taste buds were isolated according to Gilbertson et al. (1993), adapted from Béhé et al. (1990). Briefly, rats were deeply anesthetized by CO2, followed by dislocation of the cervical vertebrae. Tongues were rapidly removed and placed in Tyrode solution. One milligram of collagenase A (Boehringer Mannheim, Indianapolis, IN), 2.5 mg of dispase (grade II; Boehringer Mannheim), and 1.0 mg of trypsin inhibitor (type I-S; Sigma Chemical, St. Louis, MO) in 1.0 ml of Tyrode solution were injected between the lingual epithelium and muscle layer. Tongues were incubated in Ca2+-free Tyrode solution at room temperature for 30–60 min. After incubation, the lingual epithelium could be peeled free from the underlying tissue with gentle dissection. The freed epithelium was pinned serosal side up in a Sylgard-lined petri dish and incubated in Ca2+-free Tyrode solution for 10 min to loosen the attachment of taste buds to papillae. Foliate and vallate taste buds were removed from the epithelium by gentle suction with a fire-polished pipette (tip diameter ∼50 μm) and plated on the bottom of a chamber that consisted of a standard glass slide onto which a silicon ring 1–2 mm thick and 15 mm ID was pressed. The glass slide was precoated with Cell-Tak (∼3 μg/cm2; Collaborative Research, Bedford, MA) to improve adherence of isolated taste buds to the bottom of the chamber. The chamber was placed on the stage of an upright Olympus microscope (model BHWI, equipped with a ×40 water-immersion objective, working distance 3.1 mm) and taste buds were viewed with Nomarski optics at ×750 (Fig. 1 A).

Fig. 1.

A: isolated taste bud from rat foliate papillae photographed with Nomarski differential interference contrast optics. Scale bar: 20 μm. B: voltage-gated Na+ and K+ currents recorded in whole cell configuration from taste cell in isolated foliate taste bud such as that shown in A. Cellwas held at −85 mV and stepped in 10-mV increments from −45 to −5 mV.

Recording techniques

In the majority of experiments, membrane ionic currents were monitored in taste cells with the use of either nystatin-perforated (Horn and Marty 1988; Korn et al. 1991) or whole cell (Hamill et al. 1981) patch-clamp configurations. Patch pipettes were made from soda lime glass capillaries (Baxter Scientific Products, McGaw Park, IL) on a two-stage vertical puller (model PB-7; Narishige, Tokyo, Japan). Pipette resistances typically were 4–8 MΩ when filled with standard intracellular (pipette) solution. Cell-attached seal resistances were in the range of 5–30 GΩ with both types of recording configurations.

Whole cell membrane currents were recorded at room temperature (20–22°C) with the use of an Axon-patch amplifier (model 1D; Axon Instruments, Foster City, CA). Signals were recorded and analyzed with the use of either a Macintosh Quadra 800 computer equipped with a MacADIOS II data acquisition board and SuperScope II software (GWI, Somerville, MA), or a 486-based computer equipped with Digidata 1200 data acquisition system and pClamp software (Axon Instruments). Signal filtering and digitization were adjusted according to the specific features of the membrane currents. For voltage-gated currents, signals were prefiltered at 5 kHz and digitally recorded at 50-μs intervals; for glutamate-induced currents, signals were usually prefiltered at 100–200 Hz and digitized at 3- to 10-ms intervals. In some recordings, signal averaging was used to improve signal-to-noise ratio. Leakage and capacitive currents were not subtracted from records.

The access resistance of the patch pipette tip was estimated by dividing the amplitude of the voltage steps by the peak of the capacitive transients (from which stray capacitance had been subtracted). In whole cell configuration, values typically ranged from ∼10 to 25 MΩ. In perforated-patch configuration after the high-resistance seal had been obtained, access resistance was monitored periodically to check the diffusion of nystatin into the cell membrane. Access to the cell interior was judged by the first appearance of a capacitive transient (e.g., Horn and Marty 1988). We usually began our experiments when the access resistance dropped to∼50 MΩ.

Our standard procedure for recording was to adjust the pipette potential for zero current flow before establishing a seal. This zero current potential served as reference for subsequent measurements. Because the pipette and the bath contained different solutions, a liquid junction potential developed at the tip of the pipette (for review, see Barry and Lynch 1991). This potential was measured as described by Neher (1992) and was 7 mV (pipette solution negative) with the pipette solution used in perforated-patch recording and 4–5 mV with the pipette solutions used in whole cell recordings. All data have been corrected for these liquid junction potentials.



Our standard bathing saline (Tyrode solution; adapted from Béhé et al. 1990) consisted of (in mM) 120 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 glucose, 10 sodium pyruvate, 20 sodium methanesulfonate, and 10 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES), buffered to pH 7.4 with NaOH. The Ca2+-free Tyrode solution used in cell isolation contained 2 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA) in place of CaCl2 and MgCl2. To test the involvement of Na+ to glutamate responses, an Na+-free solution was used [composition, in mM: 120 choline chloride, 5 KCl, 2 CaCl2, 1 MgCl2, 10 glucose, 10 pyruvic acid, 20 methanesulfonic acid, and 10 HEPES, buffered to pH 7.4 with tris(hydroxymethyl)aminomethane (Tris)].


We found it critical to maintain isoosmolarity during glutamate application because taste cells appeared very sensitive to small fluctuations in osmotic strength of the surrounding medium. Thus, for chemical stimulation of taste cells, sodium glutamate (10–20 mM) replaced corresponding concentrations of sodium methanesulfonate in Tyrode solution. Osmolarity of bathing and test solutions were carefully measured with an osmometer. The concentration of glutamate used in these experiments is at or slightly above the threshold for glutamate detection in most mammals, including rats (e.g., Yamamoto et al. 1991). In Na+-free experiments, glutamic acid instead of sodium glutamate was added, a corresponding amount of methanesulfonic acid was eliminated, and the bathing solution was buffered to pH 7.4 with Tris (see above). In experiments in which we tested the effect of the mGluR4 agonist l-AP4 (Tocris Cookson, Bristol, UK; 0.83 mM) or the epithelial sodium channel blocker amiloride (30 μM), drugs were added directly to Tyrode solution. In all cases test solutions were bath applied. Gravity-fed solutions were controlled by multisolenoid manifold valves (General Valve, Fairfield, NJ) and introduced through a common inlet into the recording chamber. Flow rates were typically 2–3 ml/min, thereby permitting extracellular solution changes in <10 s.


Two different pipette solutions were used, depending on the configuration of patch-clamp recording. For perforated-patch recording, nystatin was dissolved in dimethyl sulfoxide with constant stirring to yield a stock solution of 50 mg/ml. The nystatin stock solution was prepared fresh for each experiment. Nystatin was added to a pipette solution consisting of (in mM) 55 KCl, 75 K2SO4, 8 MgCl2, and 10 HEPES, buffered to pH 7.2 with KOH, to give a final nystatin concentration of 250 μg/ml. The tip of the patch pipette was filled with nystatin-free solution and backfilled with nystatin solution. For whole cell recording, the pipette solution contained (in mM) 140 KCl, 2 MgCl2, 1 CaCl2, 11 EGTA (pCa 8), 10 HEPES, buffered to pH 7.2 with KOH, 2 ATP, and 0.4 guanosine 5′-triphosphate. In some experiments KCl was replaced by equal concentration of CsCl. All chemicals were from Sigma Chemical except l-AP4 (Tocris Cookson).


After isolation, taste cells typically remain aggregated at their apical poles, resembling a taste bud (Fig. 1 A). However, partially disaggregated taste buds and sometimes isolated taste cells could be obtained during the isolation procedure (cf. Béhé et al. 1990). We noticed no difference between isolated cells or cells in isolated taste buds with respect to voltage-gated conductances, leak conductances, or responses to glutamate. As has been shown previously for rat, foliate and vallate taste cells possessed voltage-gated conductances, including an inward Na+ current and an outward delayed rectifier K+ current (Fig. 1 B; cf. Akabas et al. 1990; Béhé et al. 1990; Herness and Sun 1995). In the present study we recorded from a total of 91 taste cells from foliate and vallate taste buds. To be included in these analyses, recordings had to meet three criteria. 1) In the perforated-patch or whole cell configuration, cells had to possess voltage-gated Na+ and K+ currents, or voltage-gated K+ currents alone when depolarized from a holding potential at or near −85 mV (Fig. 1 B). This indicated that we were recording from taste receptor cells and not surrounding epithelial cells, which lack these currents (Akabas et al. 1990; Kinnamon and Roper 1988). 2) Cells had to show an elongate morphology typical for taste receptor cells (Royer and Kinnamon 1994). 3) Recordings had to remain stable for >5 min to ensure that changes in holding current could not be attributable to changes in seal resistance. The use of nystatin-perforated patches enhanced the stability of our recording and typically allowed us to record from a single cell for as long as 1 h. We noticed no other differences relevant to these studies between recordings in conventional whole cell mode versus perforated-patch-clamp configuration. Thus we pooled the results with the use of these two techniques, unless otherwise indicated. Taste cells had a zero current potential of −51 ± 11 (SD) mV (range −37 to −77 mV;n = 22). This is likely to be an underestimate of the resting potential in situ; therefore we held taste cells at −85 mV to approximate the true resting potential (Béhé et al. 1990; Chen et al. 1996; Doolin and Gilbertson 1996; Herness and Sun 1995) and to improve recording stability. At −85 mV, a maintained inward current (holding current) was observed in all cells. The mean value of this current was −21.1 ± 10.9 (SD) pA (n = 75). Cell input resistance at this holding potential was 3.6 ± 2.3 GΩ (n = 47) and cell capacitance was 11.7 ± 4.5 pF (n = 34).

Glutamate-induced currents

We tested the effects on taste cells of glutamate applied at concentrations that were at or slightly above the taste threshold in rats. In 55 of 91 foliate and vallate taste cells, with either perforated-patch recording (nystatin in the patch pipette) or conventional whole cell recording, bath-applied glutamate (10–20 mM) affected the resting inward holding current in two different ways. That is, glutamate elicited one of two types of membrane responses, sustained or transient, as described below. The lack of sensitivity to glutamate in the remaining taste cells (36 of 91) might be due to 1) responses below the signal-to-noise ratio for our recording conditions; 2) disruption of transduction mechanisms by the enzymatic treatment; 3) desensitization of a transient glutamate response (see below); or 4) a frank absence of glutamate transduction in those cells. We observed no significant or consistent differences in membrane properties (input resistance, membrane capacitance, voltage-gated currents) between taste cells that responded to glutamate and those that did not (data not shown).


In the majority of taste cells that responded to bath-applied glutamate (51 of 55 cells), the responses consisted of a reduction in the inward holding current (that is, an outward current, Δi out). After a delay that is attributed to the time taken to perfuse the bath, Δi out slowly reached a peak and remained constant as long as glutamate was present (Fig. 2, bottom). Hereafter this response is named the “sustained response.”

Fig. 2.

“Sustained” glutamate responses in 2 different rat foliate taste cells, recorded with perforated-patch technique. Bath-applied glutamate reduced tonic inward current. Cells were voltage clamped at −87 mV. i h, Holding current.

For 20 mM glutamate, Δi out for sustained responses was 5.7 ± 5.5 (SD) pA (range 0.8–29 pA; n = 36) at a holding potential at or near −85 mV. Glutamate responses were reversible: holding current returned to baseline level when glutamate was removed from the bath. However, a rapid desensitization of an initial, rapid response cannot be ruled out. Rapid desensitization can only be revealed by a much more rapid application than is obtainable with bath perfusion (Franke et al. 1987; Gilbertson et al. 1991). Sustained glutamate responses were accompanied by a concomitant decrease in membrane conductance as monitored by measuring the cell input resistance throughout Δi out (Fig. 3). Similar results were obtained in four cells. This suggests that, on balance, ion channels in the taste cell membrane are closed by glutamate.

Fig. 3.

Change in input resistance during sustained glutamate responses, measured with perforated-patch technique in a foliate taste cell. Input resistance was monitored by application of a train of−40-mV steps (not shown) from a holding potential of −57 mV. In this cell, input resistance increased from 4.8 to 10.3 GΩ during glutamate response, suggesting that membrane conductance was reduced.

Sustained glutamate responses varied linearly with voltage: Δi out became smaller when the membrane was depolarized. Figure 4 A shows the responses elicited in a foliate taste cell held at different holding potentials.

Fig. 4.

A: voltage dependence of sustained glutamate responses. Responses were elicited by 20 mM glutamate applied to a cell that was voltage clamped at different holding potentials (V h). Perforated-patch configuration. Note the increase in noise as membrane was depolarized. B: current-voltage relationships for sustained glutamate responses, recorded with perforated-patch (○) and whole cell configuration (•). Apparent reversal potentials are −20 and −17 mV, respectively. Each point represents average (mean ± SD) of 2–9 cells, unless otherwise indicated. nor iGlu, glutamate-induced current normalized to that elicited at −85 mV.

The increased noise in the transmembrane current when taste cells were depolarized is probably due to activation of voltage-gated conductances (e.g., Bigiani et al. 1996). The instability of the recordings when the membrane was held at depolarized values prevented us from recording true reversal potential (E rev) for sustained responses. An apparent E rev was extrapolated by a linear fit through normalized values of sustained glutamate responses at different holding potentials in different cells. These measurements were performed with perforated-patch recording as well as with whole cell recording. Figure 4 B shows the current-voltage relationship for glutamate responses from eight cells in perforated-patch conditions (○) and nine cells in whole cell conditions (•).

The extrapolated values for E rev were similar in the two recording conditions, approximately −20 mV with perforated-patch recording and −17 mV with whole cell recording. We also measured the E rev in whole cell configuration for glutamate responses with Cs+-containing pipettes to reduce any contribution from specific potassium conductance mechanisms. Data pooled from three cells yielded a value for E rev of about −18 mV, similar to that obtained with K+-containing pipettes.

An E rev in the range of −17 to −20 mV excludes the possibility that glutamate-induced sustained currents are selectively carried by Na+ or K+ ions. E rev for the sustained responses, recorded in the perforated-patch configuration (approximately −20 mV), was close to the chloride equilibrium potential (approximately −16 mV). However, in the whole cell recording configuration, E rev was approximately −17 mV, whereas the chloride equilibrium potential was 0 mV. This discrepancy suggests that Cl may not be involved in sustained responses. To explore this further, we tested the effects of 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid (DIDS; 500 μM), a known Cl conductance blocker in Necturus taste cells (Taylor and Roper 1994). Responses elicited by glutamate in the presence or absence of DIDS showed no differences (n = 3). Taken together, the data suggest that sustained glutamate responses are generated by the closure of nonselective cation channels.

The involvement of nonselective cation channels in glutamate responses was further tested by replacement of Na+ with choline in the bath (see Methods). Removing extracellular Na+ has been shown to suppress the outward current evoked by glutamate or l-AP4 through nonspecific cation channels in the retina (Thoreson and Miller 1993). Figure 5 shows the result of replacement of extracellular Na+, recorded in whole cell configuration.

Fig. 5.

Effect of replacing extracellular Na+ with choline on glutamate responses. In normal Tyrode solution, glutamate (10 mM) evoked a sustained outward current (top). Glutamate failed to evoke any response when sodium was replaced with choline (bottom). Traces were averaged to improve signal-to-noise ratio (top: 3 traces were averaged; bottom: 4 traces were averaged). Whole cell configuration.

Figure 5, top, shows the outward current evoked by glutamate in Tyrode solution. When choline was substituted for sodium, the inward holding current was reduced, as has been shown by others (Doolin and Gilbertson 1996; Gilbertson et al. 1993). More importantly for the present study, responses to glutamate were suppressed (Fig. 5, bottom). This was observed in all cells tested (n = 3).

Lastly, amiloride-sensitive Na+ channels, believed to play a role in the transduction of certain taste stimuli (Avenet and Lindemann 1988, 1991; Gilbertson et al. 1992, 1993; Heck et al. 1984; Ossebaard and Smith 1995), were not involved in glutamate responses. Sustained glutamate responses in foliate taste cells were unaffected by the presence of 30 μM amiloride in the bath (n = 5 cells). At this concentration, amiloride blocks the current through amiloride-sensitive Na+ channels in about a third of the foliate taste cells in rat (e.g., Doolin and Gilbertson 1996).

In summary, the data are consistent with sustained responses to glutamate being mediated by the closure of nonselective cation channels that are permeable to Na+ and K+. Most of the inward holding current through these channels should be carried by Na+ at normal resting potentials (i.e., near −85 mV).

In other tissues, currents through nonspecific cation channels can be suppressed by divalent cations such as Ca2+ (Almers et al. 1984; Decker and Dani 1990; Yau and Baylor 1989; Zufall and Firestein 1993). In the case of N-methyl-d-aspartate (NMDA) receptors, Ca2+ acts as both a blocker and a permeator (Ascher and Nowak 1988; Mayer and Westbrook 1987). We found that removing extracellular Ca2+ enhanced the inward holding current (data not shown). However, removing Ca2+ (i.e., ≤1 μM) did not significantly alter the outward current evoked by glutamate (Fig. 6) or unmask any glutamate responses in cells that were previously unresponsive (n = 5).

Fig. 6.

Effect of changing extracellular Ca2+ on sustained glutamate responses. Top: response elicited by glutamate, recorded in Tyrode solution {Ca2+ concentration ([Ca2+]) = 2 mM}. Bottom: response elicited by glutamate, recorded in low Ca2+ ([Ca2+] = 1 μM). Reducing [Ca2+] did not enhance outward current evoked by glutamate. Note increase in membrane noise when cell was bathed in nominally calcium-free solution (bottom). Whole cell configuration.


A second type of response to glutamate was observed in a small percentage of taste cells (4 of 55 cells). This response consisted of an increase (Δi in) in the holding current at −85 mV. After a perfusion delay, Δi in rapidly reached a peak and then declined (Fig. 7). Hereafter this transient response is named the “transient response.” On rare occasions, glutamate elicited responses that appeared to be a combination of an early transient inward current (transient response) followed by a prolonged outward current (sustained response), suggesting that in some cells both responses occurred (n = 5 cells).

Fig. 7.

“Transient” glutamate response in a foliate taste cell voltage clamped at −87 mV (perforated-patch recording). Note that response declines during maintained presence of glutamate.

The low incidence of transient glutamate responses prevented their detailed characterization. Nevertheless, we were able to make the following observations. As was the case for sustained glutamate responses, transient glutamate responses did not seem to be affected by the recording configurations (perforated-patch vs. whole cell recordings). In contrast to sustained glutamate responses, however, transient responses decayed during maintained chemostimulation (compare Figs. 7 and 2). The amplitudes of transient responses decreased on membrane depolarization (Fig. 8).

Fig. 8.

Voltage dependence of glutamate transient responses. Responses were elicited by 20 mM glutamate in a foliate taste cell voltage clamped at different holding potentials. Whole cell configuration.

The increased noise in the current traces during transient glutamate responses suggested that the taste stimulus opened channels. If so, transient current responses elicited by glutamate could be produced by an influx of Na+ and/or Ca2+, by an outward efflux of Cl, or by a combination of these conductance changes. We tested the involvement of Na+ by replacing this ion with choline in the bathing solution. In Na+-free solution, transient glutamate responses were abolished (Fig. 9), as one might expect if responses were generated through nonspecific cation channels. No further characterizations were attempted.

Fig. 9.

Sodium dependence of transient glutamate responses. Glutamate at 20 mM was bath applied to a foliate taste cell voltage clamped at −84 mV (whole cell configuration). Top: glutamate evoked an inward current in Tyrode solution. Middle: after extracellular sodium was replaced with choline, glutamate failed to induce responses. Bottom: return to Tyrode solution restored transient glutamate response.

L-AP4-induced currents

Recent findings have shown that a metabotropic glutamate receptor, mGluR4, is expressed selectively in rat taste budsfrom foliate and vallate papillae, especially in young rats (Chaudhari et al. 1996). These workers postulated that mGluR4 or a closely related receptor is responsible, in part, for transducing glutamate taste (Chaudhari et al. 1996). We therefore tested whether l-AP4, an agonist for mGluR4, evoked responses in rat taste cells, and if so, how these compared with glutamate responses. We chose a concentration of l-AP4 of 0.83 mM because it was in the concentration range used by other researchers to test l-AP4 involvement in glutamate taste {e.g., 0.1–10 mM in conditioned taste aversion experiments: Chaudhari et al. 1996; 1 mM in intracellular Ca2+ ([Ca2+]i) measurement experiments: Hayashi et al. 1996}. Taste cells consistently responded to bath-applied l-AP4 with a reduction in the steady holding current (that is, Δi out), similar to the sustained responses induced by glutamate. This was observed in all cells tested (n = 3). In one case l-AP4 elicited a response that suggested that Δi out was preceded by a transient component, i.e., a biphasic response (Fig. 10 A). The mean amplitude of Δi out elicited by l-AP4 was 1.8 ± 0.3 (SD) pA (n = 3) at a holding potential of −85 mV. The response was maintained at its maximum as long as l-AP4 was present in the bath and became smaller at depolarized potentials (Fig. 10). In the same taste cells, glutamate (20 mM) elicited sustained responses with a the mean amplitude of 5.7 ± 3.2 (SD) pA (n = 3).

Fig. 10.

A: responses to l(+)-2-amino-4-phosphonobutyric acid (l-AP4), an agonist of metabotropic glutamate receptor mGluR4, recorded in a foliate taste cell voltage clamped at different holding potentials. Perforated-patch recording. B: current-voltage relationships for l-AP4 responses (□) and sustained glutamate responses (○) recorded in same cell. Responses to glutamate (20 mM) for this cell are shown in Fig. 4.

The apparent E rev for Δi out elicited by l-AP4 was similar to that observed for the sustained response to glutamate: Fig. 10 B shows the current-voltage plots for the l-AP4 (Δi out) and glutamate responses recorded in the same taste cell. During Δi out elicited by l-AP4, the membrane conductance of taste cells decreased (Fig. 11).

Fig. 11.

Change in input resistance during l-AP4 responses in a foliate taste cell. Input resistance was measured as in Fig. 3. In this cell, input resistance increased from 4.5 to 5.7 GΩ during l-AP4 response, suggesting that membrane conductance was reduced. Perforated-patch recording.


Our goal in this study was to determine how glutamate, an important taste stimulus for vertebrates, affects membrane properties of rat taste cells from foliate and vallate papillae. Foliate and vallate taste cells were studied because prior reports showed that a certain metabotropic glutamate receptor, mGluR4, is expressed in these cells and may underlie glutamate taste transduction (Chaudhari et al. 1996). Earlier data on nerve recordings also have implicated foliate and vallate taste buds as key sites for glutamate taste reception (Ninomiya et al. 1991). Our main findings are that glutamate, at concentrations at slightly above threshold for taste in rats, induces two different responses in approximately two-thirds of these taste cells (sustained and transient; Figs. 2 and 7). Our results suggest that both types of glutamate responses may be mediated by nonspecific cation channels on which the amino acid acts in different ways. In the case of sustained responses, our data indicate that glutamate and l-AP4 close cation channels, thereby reducing a steady inward current at rest (about −85 mV). Under voltage clamp, this produces an outward current response that would correspond to a hyperpolarization under physiological conditions. In contrast, transient responses seem to be mediated by glutamate opening cation channels and producing a transient inward current. Under physiological conditions this would be a depolarizing response. The properties of these two responses suggest that there may be two different sets of nonspecific cation channels in taste cell membranes. It is tempting to speculate that these responses occur in two different subpopulations of taste cells. However, it is also plausible that both responses coexist in some taste cells but that one response may obscure the other. The occasional occurrence of biphasic responses (transient inward current followed by maintained outward current) in fact supports this.

Although our data are generally consistent with a role for nonselective cation conductances in glutamate responses, especially the sustained outward currents, we cannot unequivocally rule out an involvement of Cl conductances. DIDS, a potent and specific blocker of chloride conductance channels in Necturus taste cells (Taylor and Roper 1994), did not alter responses elicited by glutamate. However, DIDS may not be as effective a chloride channel blocker in rat taste cells (S. Wladkowsky, W. Lin, M. McPheeters, S. C. Kinnamon, and S. Mierson, in preparation). Nonetheless, our findings were corroborated by the lack of change in the E rev for glutamate responses when the Cl equilibrium potential was altered (see Fig. 4 B). These data need to be substantiated by more detailed investigations such as extensive ion substitution experiments. Alternative explanations for the mechanisms of glutamate responses in taste cells include an electrogenic glutamate transporter or a sodium-dependent glutamate transporter with properties of a ligand-gated chloride channel (see, for example, Brew and Atwell 1987; Eliasof and Werblin 1993; Fairman et al. 1995; Wadiche et al. 1995). However, the properties of none of these transporters are completely compatible with the glutamate responses observed in the present report.

The present findings may or may not apply to taste cells in other regions of the tongue, such as fungiform papillae. A recent study by Doolin and Gilbertson (1996) emphasizes this point. These workers showed that the occurrence of functional amiloride-sensitive sodium channels varies markedly among fungiform, foliate, and vallate taste papillae in rat tongue. Thus general principles regarding the mode of action of glutamate on taste cells await further studies on these and other papillae.

Hayashi et al. (1996) recently also reported that glutamate had two different effects on taste cells. These workers used dye imaging techniques on taste cells from mice to measure [Ca]i and transmembrane potentials. They reported that glutamate, NMDA, and l-AP4 increased [Ca]i in some cells and decreased it in others. The change in [Ca]i was not always in the same direction for glutamate, NMDA, and l-AP4. The observed increases and decreases in [Ca]i would be consistent with a membrane depolarization and hyperpolarization, respectively, although when a voltage-sensitive fluorescent dye was used, Hayashi et al. only observed depolarizing responses in mouse taste cells. Previously, Teeter et al. (1992) reported that glutamate and NMDA elicited reversible conductance increases in lipid bilayers to which mouse vallate and foliate epithelial membrane vesicles had been fused. The channels appeared to be somewhat selective for Na+, but were permeable to K+ and to divalent ions as well. The findings presented in our report are generally consistent with the above studies on the dual effects of glutamate on taste cells.

A central question is which of the two types of responses to glutamate, if either, represents the actual taste transduction mechanism in situ. Glutamate has a variety of modulatory effects on a diverse range of cell types. Thus there is necessarily some uncertainty regarding how the present observations correlate with taste transduction per se. On the basis of data presented here and those of Teeter et al. (1992) and Hayashi et al. (1996), one cannot directly deduce how glutamate produces a receptor potential during taste stimulation. An important caveat in all the above experiments is that glutamate and its agonists have been tested on isolated taste buds or in lipid bilayers with the use of membranes isolated from lingual epithelium. Chemostimulation was not limited to the apical chemosensitive tips of taste receptor cells, as would be the case in situ. Access of glutamate to potentially nonsensory regions on taste receptor cells (i.e., basolateral membrane) or inclusion of membranes from nontaste cells into bilayers may confound the results. An additional caveat is that most studies on glutamate taste have tested relatively high concentrations (mM range) of this amino acid and its agonists, NMDA and l-AP4. This raises the concern for isolated cells or membranes that the agents may not be as selective as when they are applied (at much lower concentrations) in other tissues. However, it should be emphasized that glutamate as a taste stimulus is only effective at millimolar concentrations; lower concentrations are below taste threshold and were found to be ineffective in our patch-clamp recordings. Nonetheless, our findings tend to implicate sustained outward currents in glutamate receptor potentials. This conclusion is based, in part, on the important observation in behavioral experiments on rats that l-AP4—but not NMDA, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, or kainate—mimics the taste of glutamate (Chaudhari et al. 1996). Our studies here show that the dominant effect of l-AP4 is a sustained outward current response in rat taste receptor cells. Furthermore, recent studies by Lin and Kinnamon (1996) on rat fungiform taste cells indicate that the sustained outward currents elicited by glutamate are potentiated by 5′ nucleotides, an important and unique feature of glutamate taste (Kumazawa et al. 1991; Kuninaka et al. 1964; Yamaguchi et al. 1971; Yoshii et al. 1986). Collectively, these findings suggest that glutamate taste is transduced by a metabotropic receptor that reduces a tonic inward cationic current in taste receptor cells, produces a sustained outward current, and thereby elicits a hyperpolarizing receptor potential. Clearly, at this stage, this conclusion is a working hypothesis subject to further testing.

The sustained outward currents elicited by concentrations of glutamate (20 mM) and l-AP4 (0.83 mM) that are near behavioral threshold (Chaudhari et al. 1996) were small (average 5.7 and 1.8 pA, respectively). Nonetheless, one might expect sizable hyperpolarizing receptor potentials to these stimuli under physiological conditions (i.e., in current clamp). This is because taste cells have a high membrane resistance near their resting potential (i.e., −70 to −90 mV; Béhé et al. 1990; Bigiani et al. 1996; Chen et al. 1996; Doolin and Gilbertson 1996; Herness and Sun 1995). In our experiments, the input resistance of taste cells (an estimation of the cell membrane resistance) was ∼3.6 GΩ at a holding potential of about −85 mV. Therefore responses to glutamate and l-AP4 at the above concentrations should hyperpolarize taste cells by ∼20 and 7 mV, respectively.

Hyperpolarizing receptor potentials in taste cells, if they occur, are in contrast with the present dogma of taste transduction. It is usually assumed that all taste cells depolarize in response to chemical stimuli. Indeed, it is this assumption that led Hayashi et al. (1996) to interpret their findings to suggest that glutamate receptors for taste must be ionotropic (NMDA-like), not metabotropic. Only ionotropic responses consistently depolarized taste cells and led to an increase in [Ca]i in the experiments by Hayashi et al. This interpretation differs from the one presented here. However, activation of l-AP4 (metabotropic glutamate) receptors in the retina has been shown to decrease a resting inward (depolarizing) current, thereby hyperpolarizing cells (De la Villa et al. 1995; Nawy and Jahr 1990; Thoreson and Miller 1993; Yamashita and Wässle 1991). Furthermore, it is well established that decreased membrane conductance and hyperpolarizing receptor potentials underlie sensory transduction in vertebrate photoreceptors. Taste receptor cell hyperpolarization could induce action potential discharge in sensory afferent fibers if synapses between taste cell and sensory fiber were tonic inhibitory ones. Alternatively, the afferent discharge from the taste buds may represent an integrated signal derived by information processing within taste organs, including inhibitory and excitatory synapses (see Roper 1989, 1992). This remains speculation, however, and a more definitive explanation for glutamate transduction mechanisms in taste buds awaits further experimentation.


We thank F. Vaccari from the Università di Modena for excellent technical assistance.

This study was supported by National Institute of Deafness and Other Communications Disorders Grants DC-00374 and DC-00244 to S. Roper, by the Italian Ministero della Università e della Ricerca Scientifica e Tecnologica (University of Modena grant “Ricerca avanzata”) to A. Bigiani.


  • Address for reprint requests: S. D. Roper, Dept. of Physiology and Biophysics, University of Miami School of Medicine, PO Box 016430, Miami, FL 33101.


View Abstract