Serotonin Modulates Voltage-Dependent Calcium Current in Necturus Taste Cells

Rona J. Delay, Sue C. Kinnamon, Stephen D. Roper


Delay, Rona J., Sue C. Kinnamon, and Stephen D. Roper. Serotonin modulates voltage-dependent calcium current in Necturus taste cells. J. Neurophysiol. 77: 2515–2524, 1997. Necturus taste buds contain two primary cell types: taste receptor cells and basal cells. Merkel-like basal cells are a subset of basal cells that form chemical synapses with taste receptor cells and with innervating nerve fibers. Although Merkel-like basal cells cannot interact directly with taste stimuli, recent studies have shown that Merkel-like basal cells contain serotonin (5-HT), which may be released onto taste receptor cells in response to taste stimulation. With the use of whole cell voltage clamp, we examined whether focal applications of 5-HT to isolated taste receptor cells affected voltage-activated calcium current (I Ca). Two different effects were observed. 5-HT at 100 μM increased I Ca in 33% of taste receptor cells, whereas it decreased I Ca in 67%. Both responses used a 5-HT receptor subtype with a pharmacological profile similar to that of the 5-HT1A receptor, but the potentiation and inhibition of I Ca by 5-HT were mediated by two different second-messenger cascades. The results indicate that functional subtypes of taste receptor cells, earlier defined only by their sensitivity to taste stimuli, may also be defined by their response to the neurotransmitter 5-HT and suggest that 5-HT released by Merkel-like basal cells could modulate taste receptor function.


The neurotransmitter serotonin (5-hydroxytryptamine,5-HT) is found in only one cell type in Necturus taste buds, Merkel-like basal cells (Delay et al. 1993). Three to seven of these small, ovoid cells are found at the base of each taste bud, evenly spaced around the outer edge (Kim and Roper 1995; Kim et al. 1993). Merkel-like basal cells selectively take up 5-HT and release it on depolarization in a Ca2+-dependent fashion (Welton and Roper 1992), as would be expected if 5-HT were a neurotransmitter released by these cells. Merkel-like basal cells do not extend processes to the taste pore and cannot be directly activated by taste stimulation; however, they form synapses with taste receptor cells and with the innervating nerve fibers (Delay and Roper 1988). Thus Merkel-like basal cells are positioned to function as interneurons in the taste bud, and could modify or modulate synaptic transmission between taste cells and afferent nerves (Delay et al. 1993; Reutter 1971).

Changes in levels of circulating 5-HT can affect taste thresholds and satiety as assessed in behavioral studies (Alvarado et al. 1990; Montgomery and Burton 1986a,b; Yen and Fuller 1992). In rats, 5-HT is released by the hypothalamus in response to the presentation of food (Schwartz et al. 1990). Mutant mice lacking 5-HT2C/1C receptors overeat compared with wild-type littermates (Tecott et al. 1995). Furthermore, tricyclic antidepressants, such as imipramine, which are potent blockers of high-affinity 5-HT uptake mechanisms, often have side effects in taste perception (Deems et al. 1991; Finley 1994; Henkin 1994). These data suggest that 5-HT may play a role in taste transduction.

The cellular effects of 5-HT on taste transduction have only recently been studied (Ewald and Roper 1994; Roper and Ewald 1992). In semi-intact preparations, Ewald and Roper (1994) found that bath application of 5-HT hyperpolarizes taste receptor cells and increases their membrane resistance. In the present study we have investigated the role of 5-HT in peripheral taste transduction, examining whether 5-HT alters specific electrical properties of isolated taste receptor cells. We demonstrate that there are two functionally different subpopulations of taste receptor cells, on the basis of the modulation of their voltage-activated Ca2+ current (I Ca) by focally applied 5-HT. In one subset of taste receptor cells, 5-HT potentiates I Ca, whereas in the second subset 5-HT inhibits I Ca. The two responses, potentiation and inhibition of I Ca, are mediated by the same receptor subtype but different second-messenger systems. These results suggest that Merkel-like basal cells may modulate synaptic transmission between taste receptor cells and the innervating nerve fibers.


Isolation of taste receptor cells

Mudpuppies (Necturus maculosus) were acquired from commercial vendors and maintained in 5% artificial seawater at 8–10°C. Animals were killed by decapitation after being anesthetized by submersion in ice water for 30 min. Isolated taste receptor cells were prepared by the method described in Kinnamon et al. (1988). Briefly, lingual tissue was removed from the tongue with blunt dissection and pinned out in a Sylgard-coated dish. The tissue was rinsed with amphibian physiological saline (APS) solution and treated with a 0.1% solution of collagenase (Worthington Biochemical) in APS plus 0.1% albumin with 5 mM glucose until the surrounding, nontaste epithelium could be gently peeled away from the taste buds. In this manner, taste buds were left on their pedestals of connective tissue. After a short period of time in calcium-free APS, taste receptor cells could be gently sucked into a fire-polished pipette and plated onto recording chambers coated with Cell-Tak (Collaborative Research). Isolated taste receptor cells were examined immediately after plating into the recording chambers. For experiments involving pertussis toxin (PTX), nontaste epithelium was stripped away as above and exposed taste receptor cells were treated for 12–24 h at 4°C in a PTX (500 ng/ml) APS (plus albumin and glucose). Cells with many different morphotypes were obtained from Necturus taste buds, but only those cells with an elongated or bipolar morphology, indicative of receptor cells, were chosen for electrophysiological recording. This selection criterion excluded nontaste epithelial cells, stem cells, and Merkel-like basal cells.

Gigaseal whole cell recording

Membrane currents were recorded with the use of the whole cell configuration of the patch-clamp technique (Hamill et al. 1981). Patch pipettes were made from hematocrit capillary tubes and coated with a soft dental wax to reduce electrode capacitance. The resistance of the pipettes was between 2 and 4 MΩ when filled with intracellular solution.

Whole cell currents were measured at room temperature with the use of an Axopatch 1D patch-clamp amplifier (Axon Instruments). Voltage pulses applied to the pipette were controlled by a computer system (11/73, Digital Equipment) equipped with a Cheshire data interface (Indec Systems). Pipette resistance was electronically canceled with the patch-clamp amplifier. Leak and linear capacitative currents were measured with the use of hyperpolarizing pulses (−25 mV) applied to the pipette from the holding potential. These currents were automatically subtracted from the records. Membrane capacitance was estimated by integrating the capacitative transient and dividing by the amplitude of the voltage step. Whole cell input resistances ranged from 1 to 10 GΩ. Cells were held at −80 mV (occasionally, −100 mV) and pulsed from −20 to +40 mV in 10 mV steps. The cells were not compensated for whole cell capacitance or series resistance with the patch clamp so the true series resistance could be monitored continuously and the experiment terminated if the resistance exceeded 15 MΩ.

I Ca was isolated from the other voltage-activated currents as follows. Na+ current was blocked with 1 μM tetrodotoxin in the bath. K+ currents were blocked with 10 mM tetraethylammonium bromide in the bath and cesium (Cs+) in the intracellular solution. Ba2+ was substituted for Ca2+ to avoid activating Ca2+-dependent currents or second-messenger systems. Because of Ca2+ current rundown, only a single drug was tested on each cell.

Perfusion system and solutions

The recording chamber consisted of either a 35-mm plastic petri dish or a Sylgard “O” ring on a standard 1.0-mm glass microscope slide. The bath perfusion system was gravity fed, with the solution changes controlled by multisolenoid manifold valves (General Valve). Focal application of 5-HT and other drugs was achieved by a modified U tube system, as described by Oxford and Wagoner (1989). The U tube consisted of a tiny loop of No. 10 polyethylene tubing shaped as a “U” with a small hole at the apex of the U. The tubing was connected to a series of solution-filled reservoirs. Solutions flowed from the reservoirs through the tube and into a vacuum trap. When the vacuum was discontinued, the test solution would flow gently but rapidly over a closely positioned cell. All test solutions applied in this manner contained the dye Fast Green to allow visual monitoring of the test solution. Fast Green had no effect on membrane currents in taste cells.

APS consisted of (in mM) 112 NaCl, 2 KCl, 8 CaCl2, and 5 N-2-hydroxyethylpiperazine-N′ethanesulphonic acid (HEPES), buffered to pH 7.2 with NaOH. The external bath solution contained 94 mM NaCl, 20 mM BaCl2, 2 mM KCl, 5 mM HEPES, 10 mM tetraethylammonium bromide, and 1 μM tetrodotoxin. The intracellular solution used for measuring Ca2+ current was composed of (in mM) 100 cesium acetate, 10 NaCl, 10 HEPES, 0.23 CaCl2, 1.0 ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid [(EGTA) or 5 bis-(o-aminophenoxy)-N,N,N′,N′-tetraacetic acid (BAPTA)], 2 MgCl2, 5 ATP, and 0.5 guanosine 5′-triphosphate (GTP), with free Ca2+ concentration calculated to be 10−7 M. The GTP and ATP were added fresh daily and the final pH was titrated to 7.2 with tris(hydroxymethyl)aminomethane-OH.

Suppliers were as follows: BAPTA, Molecular Probes; PTX, ketanserin tartate, and methysergide maleate, Research Biochemical; 1-(5-isoquinolinesulfonyl)-2-methylpiperazine HCl (H-7), N-[2-((p-bromocinnamyl)amino)ethyl]-5-isoquinolinesulfonamide HCl (H-89), and 12,13-dibutylate (PDBr), CalBioChem; and ATP, GTP, 1,2-dioctanoyl-sn-glycerol (DOG), and 8-(4-chlorophenyllthio)-adenosine 3′:5′-cyclic monophosphate (8cpt-cAMP), Sigma.


We recorded from elongate cells isolated from Necturus taste buds. These cells were identified as taste receptor cells on the basis of morphology and electrical excitability (Kinnamon and Roper 1987). Necturus taste receptor cells can be subdivided into dark cells and two types of light cells on the basis of morphological criteria (Cummings et al. 1987; Delay and Roper 1988; Farbman and Yonkers 1971; McPheeters et al. 1994). Although we did not attempt to identify taste receptor cell subtype because of experimental constraints, it is probable that the majority of cells examined in this study were dark cells, because dark cells tend to survive the isolation procedures better than the other cell types (McPheeters et al. 1994). Approximately 10% of the isolated taste receptor cells had no I Ca (see also Bigiani and Roper 1993; Bigiani et al. 1996; McPheeters et al. 1994) and 5-HT could not unmask or evoke an I Ca from these cells.

Ca2+ channels in Necturus taste receptor cells have been reported to be of a single type, with properties intermediate between those of N-type and L-type channels. They require strong depolarization for activation and exhibit slow, voltage-dependent inactivation. In the present study, I Ca was isolated from the other voltage-dependent currents and its modulation was studied in detail. Typical current traces for I Ca are shown in Fig. 1 A. These currents were evoked by stepping the voltage from −20 to +40 mV from a holding potential of −80 mV. The apparent peak of the inward current was +10 mV (Fig. 1 B), although the current varied somewhat from cell to cell (range = 200–6,000 pA). I Ca inactivates, even with barium (Ba2+) as the charge carrier (Kinnamon and Roper 1987; Kinnamon et al. 1989). This rundown of the calcium current is displayed in Fig. 1 C. We tried to minimize calcium current rundown by recording with perforated patches formed with nystatin rather than standard whole cell mode, but had unreliable results. With whole cell recording, calcium current rundown could be fitted with an exponential curve. In control cells(n = 9), an exponential fit to the first four to five data points and the last point (which would occur after washout in the test cells) produced a predicted rundown for I Ca that consistently matched the observed rundown (P > 0.96).

Fig. 1.

Calcium current activation and rundown over time. A: typical calcium current elicited from a taste cell after blocking the other voltage-activated currents. B: current-voltage (I-V) relationship for the cell in A. Voltage-activated calcium current (I Ca) peaks near 10 mV for this cell. C: peak of I Ca plotted against time. Dashed line: predicted rundown in I Ca over time for this cell.

Effect of 5-HT and other amines on voltage-dependent Ca2+ current

In >90% of the taste receptor cells examined, 5-HT (100 μM) altered I Ca. The responses fell into two groups. In one subset of taste receptor cells, 5-HT decreased I Ca (Fig. 2). The inhibitory effect of 100 μM 5-HT on the peak of I Ca (+20 mV) is shown compared with wash in Fig. 2 A. (I Ca before addition of 5-HT was higher than that obtained after wash.) This decrease in I Ca in response to 5-HT was observed at other voltage steps (Fig. 2 B). The inhibitory effect of 5-HT was partly reversible (Fig. 2 C). To average the inhibitory effects of 100 μM 5-HT, the current before application of 5-HT was normalized to 1.0 and the change in I Ca was given by a fraction of this number (n = 11). At 100 μM,5-HT decreases averaged ∼45% of I Ca before rundown was taken into account (Figs. 1 C and 4). However, when corrections were made for rundown, the average decrease in I Ca was only 28% (Fig. 4).

Fig. 2.

Inhibitory effects of serotonin (5-HT) on I Ca of isolated taste receptor cells. A: change in peak of I Ca (+20 mV) by focal application of 100 μM 5-HT. Cell was held at −80 mV, then pulsed in 10 mV steps from −20 to +40 mV. B: I-V relationship for cell in A in the presence of 100 μM 5-HT and after wash. I Ca before application of 5-HT was greater than that shown for wash. C: effect of 100 μM 5-HT on I Ca plotted as peak of I Ca over time. Dashed line: predicted rundown in I Ca over time (see Fig. 1 C). D: combined data from taste receptor cells responding to 5-HT with a reduction in I Ca. Current before application of 5-HT was normalized to 1.0 and change in I Ca is given by a fraction of this number (n = 11).

Fig. 4.

Dose-response relations for the modulation of I Ca by 5-HT. Filled circles: reduction in I Ca. Open circles: increase in I Ca. Data are shown as % changes in I Ca in the presence of 5-HT. These values were calculated by subtracting I Ca during 5-HT application from projected (extrapolated) values in the absence of 5-HT (see Fig. 1 C and Fig. 2 C). Mean ± SE for each concentration and the number of observations for each point (values in parentheses) are shown. Curves were fit by the equation Y(% change) = Y max{1/[1 +(X mid/X S)]}, where Y max is the maximum response, X mid is the midpoint of the curve, s is slope, and X is concentration. The values for the points for the curve to fit the increase in I Ca were as follows: Y max = 56.7, X mid = 3.95, s = 2.15; whereas the values for the curve to fit the decrease in I Ca were Y max = 28, X mid = 9, s = 0.87.

In the second subset of taste receptor cells, 5-HT increased I Ca (Fig. 3). The peak uncompensated Ca2+ current traces before and in the presence of 100 μM 5-HT are shown in Fig. 3 A. A plot of the current versus voltage for othervoltage steps (Fig. 3 B) shows that 5-HT increased I Ca at the other voltage steps without a shift in the peak current. The time course of a typical experiment is shown is in Fig. 3 C. The increase in Ca2+ current averaged 55% (range = 15–300%, n = 12) at 100 μM 5-HT (Fig. 3 D).

Fig. 3.

Potentiation effects of 5-HT on I Ca. A: potentiation of peak I Ca by 100 μM 5-HT. Cell was held at −80 mV and pulsed in 10 mV steps from −20 to +40 mV. Shown is the effect of 5-HT on I Ca (+10 mV). B: I-V relationship for cell in A in the presence of 100 μM 5-HT and after wash. C: time course of effect of application of 10 μM 5-HT. 5-HT was applied during the solid bar shown above the graph. Dashed line: predicted rundown in I Ca over time. D: average increase in I Ca by 100 μM 5-HT in the 2nd subset of taste receptor cells (n = 11).

Both effects of 5-HT were dose dependent. The dose-response relationship for these effects is shown in Fig. 4. The apparent half-maximal concentration (EC50) for the increase in I Ca was ∼4 μM; the EC50 of the decrease was ∼9 μM. At lower concentrations of 5-HT (1–10 μM), the fraction of cells showing an increase of I Ca was sharply increased from that observed with 100 μM 5-HT. We attribute part of this change to our inability to detect reliably the small decreases in response that would be expected on the basis of the dose-response results. Small responses to 5-HT would be hard to distinguish from the rundown of the Ca2+ current. Another possible explanation is that 5-HT receptors mediating the potentiation of I Ca had a higher affinity for 5-HT than did 5-HT receptors mediating the reduction in I Ca.

Dopamine and epinephrine were also tested to see whether either of these catecholamines could modulate I Ca. Neither dopamine nor epinephrine was observed to modulate I Ca, even at concentrations as high as 200 μM (data not shown).

5-HT receptor subtypes modulating ICa in taste receptor cells

The observation that taste receptor cells could be subdivided into two different groups on the basis of their response to 5-HT suggested that these different responses could be mediated by different 5-HT receptor subtypes. Seven major 5-HT receptor subtypes have been classified (Bard et al. 1993; Humphrey et al. 1993; Matthes et al. 1993; Ruat et al. 1993a,b) on the basis of molecular cloning and intracellular coupling systems. The 5-HT receptors (except for 5-HT3) can be grouped into the following families: receptors that inhibit adenylyl cyclase (AC; 5-HT1A, 5-HT1B, and 5-HT1D receptors, although the 5-HT1A receptor has also been reported to activate AC) (Barbaccia et al. 1983; Hoyer and Schoeffter 1991; Shenker et al. 1985); receptors that activate phospholipase C (5-HT1C and 5-HT2) (Kaneko et al. 1992; Pritchett et al. 1988); and receptors that stimulate AC (5-HT4, 5-HT5, 5-HT6, and 5-HT7) (Monsma et al. 1993; Ruat et al. 1993a,b; Tsou et al. 1994). A series of experiments was undertaken to characterize the 5-HT receptor subtype(s) in taste receptor cells.

The effects of one agonist and three different antagonists were examined. The characteristic effects of these agents according to published reports are summarized in Table 1. Neither ketanserin, a selective antagonist of 5-HT1C, 5-HT2, 5-HT5, and 5-HT7 receptors, nor 3-tropanyl-indole-3-carboxylate methiodide (ICS 205, 930), a selective antagonist of 5-HT3 and 5-HT4 receptors, was able to block the effects of 5-HT on I Ca in taste receptor cells (data not shown). The 5-HT agonist 5-methoxytryptamine (5-MeOT) has been shown to activate several of the 5-HT receptor subtypes (see Table 1) with a potency comparable with that of 5-HT; however, 5-MeOT was much less potent than 5-HT at modulating I Ca in taste receptor cells (n = 22, data not shown). These data suggest that the 5-HT1C, 5-HT1D, 5-HT4, 5-HT5, 5-HT6, and 5-HT7 receptor subtypes were not involved in modulating I Ca in taste receptor cells. Methysergide, an antagonist of 5-HT1, 5-HT2, 5-HT5, and 5-HT7 receptor subtypes, was tested to determine whether it could block the serotonergic modulation of I Ca. Even at concentrations as low as 0.1 μM, methysergide blocked both actions of 5-HT (n = 15). By the process of elimination, these data suggest that a 5-HT1A or a 5-HT1B receptor is likely to mediate the modulation of I Ca by 5-HT.

View this table:
Table 1.

Comparison of pharmacological profiles and effector pathways for selected agonists/antagonists of 5-HT receptor subtypes

Second-messenger pathway for the potentiation of ICa by 5-HT

Both the 5-HT1A and the 5-HT1B receptors couple to pathways that regulate intracellular adenosine 3′,5′-cyclic monophosphate (cAMP). To investigate whether cAMP was involved in the serotonergic modulation of I Ca in taste receptor cells, we focally applied 8cpt-cAMP, a membrane permeant analogue of cAMP. The majority of cells (16 of 22) responded with an increase in I Ca (Fig. 5). Inhibition of I Ca was never observed with 8cpt-cAMP. This result suggests that the increase in I Ca could be mediated by upregulation of cAMP. It also rules out the 5-HT1B receptor because that receptor has only been reported to decrease cAMP. If elevation of cAMP stimulates I Ca, its action on the Ca2+ channel may be via a cAMP-dependent kinase that would phosphorylate the channel, or it may directly activate the channels. To test whether cAMP-dependent kinases were involved in the modulation of I Ca by 5-HT, we treated isolated cells with H-89 (20 μM) for 30 min before recording. H-89 is a selective inhibitor of cyclic-nucleotide-dependent kinases at this concentration (Chijiwas et al. 1990). In cells treated with H-89, 5-HT either reduced I Ca (8 of 13; Fig. 6 A) or had no effect (Fig. 6 B); that is, 5-HT never potentiatedI Ca in H-89-treated cells. These results are consistent with5-HT-mediated stimulation of adenylyl cyclase, resulting in activation of a cAMP-dependent kinase and phosphorylation of Ca2+ channels leading to potentiation of I Ca. The data also suggest that the inhibitory effect of 5-HT on I Ca is not mediated through cAMP.

Fig. 5.

Membrane-permeable analogue of cAMP, 8cpt-cAMP, potentiates I Ca in taste receptor cells. Example of potentiation of I Ca by 8cpt-cAMP is shown. Dark line above graph: time course of application of 8cpt-cAMP. In 70% of tested cells, 8cpt-cAMP increased I Ca. In the other 30% of taste receptor cells, no response to 8cpt-cAMP was observed.

Fig. 6.

H-89, a selective inhibitor of protein kinase A, affects modulation of I Ca by 5-HT. Cells were treated ≥30 min before recording with 20 μM H-89 in APS. H-89 was included in all bath solutions. A: even after treatment with H-89, taste receptor cells showed a reduction in I Ca in response to 5-HT. B: in some of the taste receptor cells treated with H-89, 5-HT did not elicit any change in I Ca.

Second-messenger pathway for inhibition of ICa by 5-HT

In neurons of the dorsal raphe and the hippocampus,5-HT exerts an inhibitory effect via a PTX-sensitive G protein (Bockaert et al. 1992; Frazer et al. 1990; Penington et al. 1991). We tested for this possibility by treating isolated taste buds for 12–24 h with PTX (500 ng/ml) in APS at 4°C. Of the PTX-treated cells, a few had no response to 5-HT (6 of 22; Fig. 7 A), whereas the majority of cells responded to 5-HT with a potentiation of I Ca (16 of 22, Fig. 7 B). These results suggest that the inhibitory action of 5-HT on I Ca in taste receptor cells is mediated by a PTX-sensitive G protein. In the PTX-treated cells, 5-HT never caused inhibition of I Ca. It was also noted that although the magnitude of I Ca varied from cell to cell, in general I Ca was greater in the PTX-treated taste receptor cells than in the controls.

Fig. 7.

Effect of pertussis toxin (PTX) on the ability of 5-HT to modulate I Ca. Intact taste buds were treated with 500 ng/ml PTX in APS for 12–24 h at 4°C. Control taste buds were treated the same way, except no PTX was added to the APS. A: 30% of the PTX-treated taste receptor cells showed no response to 5-HT. B: 70% of PTX-treated taste receptor cells responded with an enhancement in I Ca. PTX treatment completely blocked the reduction in I Ca normally observed in response to 5-HT.

Some 5-HT receptors activate phospholipase C (PLC),thereby hydrolyzing phosphoinositol 4,5-bisphosphate(PIP2) to form inositol 1 4,5-trisphosphate (IP3) and diacylglycerol (DAG). DAG subsequently activates protein kinase C (PKC). We tested whether PKC was involved in the inhibitory response by applying agents that mimic DAG, DOG, or PDBr (Hockberger et al. 1989). Both DOG and PDBr potentiated I Ca (data not shown), suggesting that activation of PLC is not the pathway that mediates the inhibition of I Ca.

We tested for the involvement of protein kinases, generally, in the inhibitory effect of 5-HT with the use of a broad-spectrum kinase inhibitor (Kawamoto and Hidaka 1984). H-7 is a nonselective blocker of protein kinases, particularly protein kinase A, PKC, and protein kinase G. Inclusion of H-7 in all bath solutions (50 μM) and in the patch pipette (200 μM) blocked all of the inhibitory effects of 5-HT on I Ca (n = 11). Although not investigated systematically, treatment with either H-89 or H-7 seemed to decrease the total amount of I Ca, suggesting that phosphorylation is required for calcium channel function in taste receptor cells. Collectively, these results suggest that the inhibition of I Ca by5-HT involves a PTX-sensitive G protein and a kinase sensitive to H-7 but not H-89; however, the full nature of this pathway is not yet known.


In this study we demonstrated that 5-HT modulates I Ca in taste receptor cells. On the basis of their response to 5-HT, taste receptor cells can be divided into two categories: cells in which 5-HT potentiates I Ca, and those in which 5-HT inhibits I Ca. Whether these responses to 5-HT define two morphologically distinct types of taste receptor cells, or different functional states of the same cell type, is unknown. Electrophysiologically distinct groups of taste receptor cells were recently reported by Bigiani and Roper (1993) and Bigiani et al. (1996). McPheeters et al. (1994) reported that both dark cells and a small subset of light cells exhibited I Ca. The question of which taste receptor cell type(s) 5-HT is acting on remains to be answered.

Potentiation of ICa by 5-HT

Data from the experiments with 8cpt-cAMP, H-89, and PTX suggest that 5-HT potentiation of I Ca in taste receptor cells is G protein-mediated and PTX insensitive. That is, a G protein-coupled 5-HT receptor activates AC, thereby raising cAMP, which stimulates a cAMP-dependent kinase.

5-HT receptors have been divided into seven different groups of receptors on the basis of molecular cloning and intracellular coupling systems (Bard et al. 1993; Humphrey et al. 1993; Matthes et al. 1993; Ruat et al. 1993a,b). In the present study, the results obtained by applying agonists and antagonists seem to rule out all but the 5-HT1A and 5-HT1B receptors as modulating I Ca in Necturus taste cells. Furthermore, results obtained by analyzing the effector systems apparently rule out 5-HT1B receptors mediating the potentiation of I Ca. 5-HT1B receptors decrease cAMP by inhibiting AC. The potentiation of I Ca by 5-HT in taste receptor cells operates via an increase in cAMP, not a decrease, so 5-HT1B receptors cannot be mediating this response. Therefore it seems that 5-HT1A would have to be mediating this response, according to the pharmacological profile. The comparison of the effector system for known 5-HT1A receptors with that regulating the potentiation of I Ca by 5-HT in taste cells should support this hypothesis. Unfortunately, there is some confusion about the effector system for 5-HT1A receptors. In some tissues 5-HT1A receptors activate AC (Barbaccia et al. 1983; Hoyer and Schoeffter 1991; Lucas et al. 1993; Markstein et al. 1986; Shenker et al. 1985; Taiwo et al. 1992). In other tissues 5-HT1A receptors inhibit AC (Clarke et al. 1987; DeVivo and Maayani 1988). Thus several reviews of 5-HT receptor subtypes have 5HT1A linked with inhibition and activation of AC (DeVivo and Maayani 1988; Frazer et al. 1990; Hoyer and Schoeffter 1991). Expression of the cloned 5-HT1A receptor in different cell types has shown that 5-HT1A receptors can activate several different second-messenger systems in the same cell (Fargin et al. 1989, 1991; Liu and Albert 1991; Raymond et al. 1992, i.e., Boess and Martin 1994), although activation of AC has not yet been reported. Thus it seems possible that a 5-HT1A receptor modulates the potentiation of I Ca in taste receptor cells, as our results suggest.

Inhibition of ICa by 5-HT

The second-messenger pathway mediating the inhibitory actions of 5-HT is different from that mediating the stimulatory action, but it was not fully resolved. 5-HT receptors that inhibit AC, decreasing intracellular cAMP, are probably not mediating this effect, because the experiments with H-7, a nonspecific protein kinase inhibitor, indicate that activation of a protein kinase is involved. Activators of PKC increased I Ca, suggesting that the reduction of I Ca by 5-HT is not mediated by the phospholipase C-activated PKC pathway. The involvement of a protein kinase also rules out a direct activator, such as that recently demonstrated in the rat dorsal raphe neurons by Penington et al. (1991), where 5-HT activated a G protein directly coupled to the Ca2+ channel. Thus, in Necturus taste receptor cells, the second-messenger pathway responsible for the ability of 5-HT to reduce I Ca appears to involve a 5-HT receptor coupled to a PTX-sensitive G protein that leads to activation of an unknown protein kinase.

The results obtained by applying selective agonists and antagonists seem to rule out all but the 5-HT1A and 5-HT1B receptors mediating the inhibition of I Ca. The 5-HT1B receptor has only been shown to inhibit AC activity, decreasing cAMP. The results with H-89, which did not block the inhibition of I Ca by 5-HT, and with H-7, a nonselective kinase blocker that did block all the effects of 5-HT, appear to rule out downregulation of cAMP as the pathway mediating the inhibition of I Ca, because they suggest that a non-cAMP-dependent protein kinase is required. Downregulation of cAMP is not normally coupled with activation of a protein kinase, so it does not seem likely that a 5-HT1B receptor could be mediating the inhibition of I Ca. Rather, some other second-messenger pathway that activates a non-cAMP-dependent protein kinase may be involved. Of the cloned 5-HT receptors, 5-HT1A receptors are the only receptors known to activate several second-messenger pathways. It is therefore plausible that a 5-HT1A receptor mediates both the inhibition and the potentiation of I Ca.

A point of interest is the dose-response relationships for 5-HT in taste receptor cells. At the lower concentrations (1–10 μM), the majority of the taste receptor cells (60–70%) showed an enhancement of I Ca in the presence of 5-HT, whereas at 100 μM, 5-HT predominantly inhibited (66%) the taste receptor cells. This shift in the effect of 5-HT could be due to the differing properties of the underlying receptors mediating the response. That is, 5-HT receptors that inhibit I Ca might have a lower affinity for 5-HT than the receptors that potentiate I Ca. Alternatively, the difference might be due to the regulatory state of the second-messenger pathway coupling the receptor to the Ca2+ channels. The changes in the proportion of cells that exhibits the inhibition versus potentiation of I Ca suggest that some of the taste receptor cells express either two receptors or, if there is only one receptor subtype, that it is coupled to two different pathways. In other tissues 5-HT has been shown to exert more than one effect, mediated by different second-messenger pathways and/or different 5-HT receptor subtypes (Andrade and Nicoll 1987; Corsi et al. 1991; Sanchez-Armass et al. 1991; Sumner et al. 1989). Either possibility, that of two different receptors or one receptor linked to two different second-messenger pathways, seems equally plausible in Necturus taste cells.

The biological significance of dual serotonergic modulation of I Ca in taste receptor cells is a matter of speculation at present. Ewald and Roper (1994) reported that, when taste receptor cells are constantly stimulated, bath application of 5-HT slowly hyperpolarizes the cells and increases their input resistance. Merkel-like basal cells, which contain 5-HT, may be stimulated to release the monoamine during chemostimulation of taste buds, and this could modify synaptic transmission at receptor cell synapses. Alternatively, efferent input to Merkel-like basal cells might stimulate 5-HT release. In the presence of 5-HT, one subset of taste receptor cells might release more neurotransmitter, whereas a second subset of taste receptor cells would release less. Such an effect could modulate the ability to discriminate certain taste stimuli.


The authors thank Drs. Albertino Bigiani, Vincent Dionne, and Heather Eisthen for critical reading of the manuscript.

This work was supported by National Institute of Deafness and Other Communications Disorders Grants 5R01 DC-00374 and 5P01 DC-00244 to S. D. Roper, and DC-00766 and DC-00244 to S. C. Kinnamon.


  • Address for reprint requests: R. J. Delay, Boston University Marine Program, Marine Biological Laboratory, Woods Hole, MA 02543.


View Abstract