| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institute of Cell Biophysics, Russian Academy of Sciences, Pushchino, Moscow Region, 142290, Russia
Submitted 31 March 2003; accepted in final form 16 July 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
; Lindemann 1996). Evidence suggests that complex cell-to-cell communications may occur within the taste bud and that a single neurotransmitter may not mediate all of them. Particularly, efferent synapses between taste cells and nerve fibers have been elucidated in ultrastructural studies on the taste bud (Delay and Roper 1988; Yang et al. 2000a; Yoshie et al. 1996). True chemoreceptive cells are thought to establish synapses with afferent nerve endings to relay information on the quality and intensity of taste stimuli to the brain (Lindemann 1996). Afferent synapses have been demonstrated with confidence for type III cells (Royer and Kinnamon 1991; Yee et al. 2001), while their presence in the other cell subclasses is debatable (Pumplin and Getschman 2000; Yang et al. 2000a). Gustducin, the 
-subunit of gustatory G protein (McLaughlin et al. 1992), is expressed in a subset of type II cells (Boughter et al. 1997; Sbarbati et al. 1999; Smith et al. 1999; Yang et al. 2000b). The presence of -gustducin in these cells is a prerequisite for the normal bitter and sweet reception (Ruiz-Avila et al. 2001; Wong et al. 1996). Therefore in addition to type III cells, gustducin-positive cells of the type II may also be chemoreceptive; however, they do not form synapses with the afferent fibers but instead release signaling molecules to modulate, for instance, the activity of neighboring taste cells.
A variety of neuroactive substances, including acetylcholine, norepinephrine, serotonin, amino acids (e.g., glutamate and GABA), and peptides (e.g., substance P and calcitonin gene-related peptide) have been localized immunocytochemically to the taste bud (summarized in Ganchrow 2000; Nagai et al. 1996; Yamamoto et al. 1998). Although their niches in taste bud physiology are not known with certainty, there may be little doubt regarding the involvement of these compounds in cell-to-cell signaling. Consistently, a number of transmembrane receptors, including adrenergic, glutamate, cholecystokinin, leptin, and muscarinic receptors have been identified in taste cells (Chaudhari et al. 1996, 2000; Herness et al. 2002a,b; Kawai et al. 2000; Ogura 2002). In addition, taste cells express transporters known to mediate neurotransmitter, e.g., GABA and serotonin, uptake (Obata et al. 1997; Ren et al. 1999). Electrophysiological experiments and Ca2+ imaging data indicate that many of the identified putative neurotransmitters can control excitability of taste cells by either regulating voltage-gated K+ currents (serotonin, noradrenaline, cholestokinin, leptin), Ca2+ currents (serotonin, noradrenaline, cholestokinin), and Cl- currents (acetylcholine, serotonin, noradrenaline, cholestokinin), or mobilizing intracellular Ca2+ (acetylcholine, serotonin, noradrenaline, cholestokinin, leptin) (Delay et al. 1997; Ewald and Roper 1994; Herness and Chen 1997, 2000; Herness and Sun 1999; Herness et al. 2002a,b; Kawai et al. 2000; Obata et al. 1997; Ogura 2002; Ren et al. 1999). As a neurotransmitter, glutamate may act via both metabotropic and ionotropic receptors (Chaudhari et al. 1996, 2000; Toyono et al. 2002), coupled to the modulation of a resting current and/or mobilization of intracellular Ca2+ (Bigiani et al. 1997; Caicedo et al. 2000; Lin and Kinnamon 1999).
Purines (ATP, ADP, and adenosine) and pyrimidines (UTP and UDP) are widely recognized as neurotransmitters, co-transmitters, or neuromodulators acting in the peripheral nervous system and CNS (Burnstock 2001; Di Virgilio and Solini 2002; Dunn et al. 2001; Galligan et al. 2000; Ralevic and Burnstock 1998; Stojilkovic and Koshimizu 2001). Recently, we found functional evidence implicating P2Y-like receptors in ATP-dependent Ca2+ mobilization and ionic current modulation in mouse taste cells (Kim et al. 2000). These observations point to a putative role for ATP as one more neurotransmitter/neuromodulator operative in the taste bud. Here we further characterize the purinergic signaling system in mouse taste cells.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Taste buds and taste receptor cells (TRCs) were isolated from mouse circumvallate and foliate papillae and occasionally from fungiform papillae using the modified protocol (Behe et al. 1990). Animals (6-8 wk old) were anesthetized with CO2 gas and killed by upper cervical dislocation, and tongues were removed. To allow peeling the lingual epithelium off a tongue, it was injected between the epithelium and muscle layers with the 1 mg/ml collagenase B, 1 mg/ml dispase II (both from Boehringer Mannheim), and 0.5 mg/ml trypsin inhibitor (Sigma). The tongue was incubated in an oxygenated Ca2+-free solution for 30-35 min. The peeled epithelium was pinned serosal side up in a dish covered with Sylgard resin and incubated in the Ca2+-free solution for 10-30 min. The isolated epithelium was kept at room temperature in a bath solution for 4-6 h. Taste buds were removed by gentle suction with a firepolished pipette (with an opening of 110-130 µM) and then expelled into a photometric chamber. To isolate individual TRCs, the protocol was modified in that 0.2 mg/ml elastase (Boehringer Mannheim) was added to the enzyme cocktail, and taste buds were sucked by a pipette with an opening of 70-90 µM.
Photometry
Isolated taste buds and individual taste cells were plated onto a coverslip coated with Cell-Tak (BD Biosciences, Bedford, MA) inside an attached ellipsoidal resin chamber of 150 µl volume. Taste buds were loaded in the normal bath solution containing 5 µM fura-2 AM +0.02% pluronic (Molecular Probes, Eugene, OR) for 40 min. The buds (cells) were washed out twice and incubated in the dye-free bath solution for 30 min for complete de-esterification of the loaded dye. Fluorescence of fura-2-loaded cells was recorded at 510 nm using an inverted fluorescent microscope Axiovert 100S (Zeiss, Germany) equipped with a Fluar 40x objective (NA = 0.95) and a microscope ratio photometry system (RM-D model, Photon Technology International, South Brunswick, NJ) operating in the photon counting mode. A DeltaScan illuminator allowed for measuring an excitation spectrum that was useful for the detection of cells overloaded with fura-2 or cells wherein fura-2-AM has not been completely de-esterified. Episodically, fluorescent images of taste buds or isolated taste cells were acquired by an ICCD camera IC-200 (8 bits, PTI). By using a variable diaphragm, an optical signal from a cell group or an individual cell was collected. All chemicals were bath-applied using a hand made perfusion system driven by gravity that allowed the complete substitution of a bath solution for nearly 3 s. Experiments were carried out at room temperature of 22-25°C under red light illumination.
Solutions
The enzymes were dissolved in the following solution (in mM): 140 NaCl, 5 KCl, 0.3 MgCl2, 0.3 CaCl2, 10 HEPES-NaOH (pH 7.4), and 10 glucose. Ca2+-free solution contained (in mM) 140 NaCl, 5 KCl, 1 MgCl2, 1 EGTA, 0.5 EDTA, 10 HEPES-NaOH (pH 7.4), and 10 glucose. Normal bath solution contained (in mM) 135 NaCl, 5 NaHCO3, 5 KCl, 1 MgCl2, 1 CaCl2, 10 HEPES-NaOH (pH 7.4), and 10 glucose. When needed, the normal bath solution was modified in that 1) 1 mM CaCl2 was substituted for 5 mM CaCl2, or for 1 mM EGTA (free Ca2+ < 10 nM), or 1 mM EGTA + 0.6 mM CaCl2 (free Ca2+ 
100 nM); and 2) 140 mM NaCl and 10 mM HEPES-NaOH were substituted for 140 mM N-methyl-D-glucamine (NMG) chloride and 10 mM HEPES-NMGOH, respectively. H-89 was from Calbiochem; thapsigargin, U-73122, U-73343, 2-APB, 8BrcAMP, Sp-cAMPS, SQ 22536, FCCP, and IBMX were from Sigma-RBI; ATP, UTP, ADP, and buffers and salts were from Sigma-Aldrich. Caloxin was synthesized in the Institute of Bioorganic Chemistry (Moscow).
Data analysis
The intensity of fluorescence recorded from a cell at a given excitation, e.g., 340 nm, ranged typically between 105 and 4 x 105 photons/s, and nearly 104 photons/s out of this number were produced by background light. There were two likely sources of background light: glass fluorescence and a reflection of highly intensive excitation light from the recording camera and objective lenses with consequent passing through an imperfect dichroic mirror and emission filter. In general, the background light was invariable among experiments, allowing for the automatic correction of cell fluorescence for its intensity. Only cells that exhibited the stable 340/380-nm ratio of 1.1-1.5 at rest, corresponding to the 70- to 110-nM concentration of intracellular Ca2+, were studied. Corrected experimental 340/380 nm ratios were converted into Ca2+ concentrations using the apparent Kd = 224 nM for Ca2+-fura-2 binding and the equation originally introduced by Grynkiewicz and co-authors (Grynkiewicz et al. 1985). Because of nonconfocality of the microscope, the magnitude of Ca2+ responses was likely to be underestimated (Bernhard et al. 1996). Curves were fitted by using a Marquardt-Levenberg nonlinear least squares algorithm (SigmaPlot 4.0, SPSS, Chicago, IL).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
Since we used nonconfocal fluorescent microscope, it was not possible to resolve clearly individual cells in a taste bud based on their fluorescent images (Fig. 2Aa). Nevertheless, we were able to choose an appropriate region of interest (ROI) with fluorescence apparently characterizing Ca2+ responses of individual cells in a taste bud, owing to the nonuniform loading of cells with fura-2, (Fig. 2Aa, yellow curves). First, given the characteristic size of individual taste bud cells (Fig. 1), 15- to 20-µm ROIs were typically used. Second, for several ROIs chosen within the same taste bud, ATP exerted similar, if not identical, changes in fluorescence intensity (Fig. 2B). Third, ATP responses of taste buds and isolated taste bud cells were virtually identical in their shape and kinetics (Fig. 2, B-D). In the experiments described below, we generally used photomultiplier-based microscope photometry rather than the imaging approach because the former provided a much better signal-to-noise ratio and higher sensitivity, thus enabling us to study ATP responses in detail.
|
Cells were stimulated by the bath application of ATP and episodically by UTP and ADP, which all elevated intracellular Ca2+ ([Ca2+]in) in a dose-dependent manner (Fig. 3). The response magnitude 
[Ca2+] versus nucleotide concentration C was fitted by the equation [Ca2+] = A/[1 + (EC50/C)n], where A and EC50 are the magnitude of the saturated response and half-maximal concentration, respectively. The analysis resulted in A = 235, 210, and 131 nM, and EC50 = 1.8, 11, and 2.2 µM, n = 0.78, 0.72, and 0.74 for ATP, UTP, and ADP, respectively (Fig. 3B). These findings suggest that ATP and UTP serve as full agonists of TRC purinoceptors, whereas ADP is a partial agonist.
|
With millimolar Ca2+ in the bath, a fast initial Ca2+ peak followed by a plateau was characteristic of the cellular responses to ATP (Fig. 4A, left), which solely peaked intracellular Ca2+ in the absence of external Ca2+ (Fig. 4A, right). This strongly argues that Ca2+ release forms the initial Ca2+ peak, while Ca2+ influx equilibrated by Ca2+ extrusion is responsible for the plateau persistent of ATP responses observed in the presence of external Ca2+ (Figs. 3A and 4A). Based on these observations and given that in our previous experiments, the specific P2X agonist 
,
-methylene-D-ATP never caused ATP-like responses (Kim et al. 2000), we inferred that TRC sensitivity to ATP is mediated by P2Y receptors coupled to G-proteins. The experiments described below were carried out to characterize intracellular signaling machinery driving ATP-dependent Ca2+ transients.
|
Phospholipase C and IP3 receptors mediate ATP responses
Intracellular events driven by ubiquitously expressed P2Y receptors have been intensely studied (reviewed in Burnstock 2001; Di Virgilio and Solini 2002; King et al. 1998; North and Barnard 1997). Coupling of the P2Y receptor to the activation of phospholipase C (PLC) is well documented, although the modulation of adenylate cyclase (AC) and ion channels may also occur (summarized in Burnstock 2001; Di Virgilio and Solini 2002). Based on the ability of ATP to release intracellular Ca2+ (Fig. 4A), there may be little doubt that, in TRCs, ATP responses are mediated by the phosphoinositide cascade. Indeed, preincubation of TRCs with the PLC inhibitor U-73122 (10 µM) for 1-3 min led to a dramatic (nearly 10-fold) decrease in TRC sensitivity to ATP (Fig. 4B; 8 cells), which was almost irreversible (data not shown). As a negative control, TRCs were treated with U-73343 (10 µM), a much less potent PLC inhibitor, which affected the ATP responses slightly, reducing their magnitude by 10-15% only (Fig. 4C; 3 cells).
In line with the effects of U-73122, the IP3 receptor inhibitor 2-APB (50 µM) decreased the ATP responses by 50-70% (9 cells), the effect being reversible in four cases (Fig. 5A) and only partly reversible in five cases (data not shown; see, however, Fig. 9A). Although designed for targeting IP3 receptors, 2-APB is also known as a blocker of a variety of Ca2+ permeable channels involved in G-protein signaling (e.g., Diver et al. 2001; Kukkonen et al. 2001), suggesting two likely mechanisms, namely, IP3 receptor inhibition and the direct blockage of Ca2+ influx, whereby the compound might affect ATP responses (Fig. 5A). In four experiments, 2-APB was applied when ATP responses reached a steady-state level, i.e., Ca2+ channels involved were open (Fig. 5B). In the presence of external ATP, the inhibitor caused a decrease in intracellular Ca2+ below the resting level (Fig. 5B, dash-dot line), presumably due to both Ca2+ channel inhibition by 2-APB and the activation of Ca2+ extrusion by ATP. With 100 nM Ca2+ in the bath highly attenuating Ca2+ influx, 2-APB effectively reduced Ca2+ release (Fig. 5C; 3 cells). This observation clearly indicates that, although 2-APB blocks Ca2+ influx, IP3 receptors are the major 2-APB target responsible for the inhibition of the Ca2+ transients. Together, the above results validate PLC and IP3 receptors as key effectors in the ATP transduction cascade.
|
|
Contribution of SOC channels to ATP responses
A number of Ca2+ permeable cation channels are widely recognized to mediate Ca2+ influx on cell stimulation (reviewed in Clapham et al. 2001; Eatock 2000; Lewis 1999; Wei et al. 1998). Among them, store operated channels (SOC), defined as channels that are activated in response to the depletion of Ca2+ stores, are universally involved in Ca2+ mobilization mediated by the phosphoinositide cascade (Lewis 1999; Stojilkovic et al. 2000; Venkatachalam et al. 2002). Particularly, evidence implicates SOC channels in the generation of responses of mudpuppy and mouse TRCs to bitter stimuli (Ogura et al. 2002). As noted above (Fig. 4B), the PLC inhibitor U73122
[GenBank]
eliminated both the peak and the plateau of ATP responses, demonstrating that PLC activation is a prerequisite for ATP-dependent Ca2+ influx. The PLC inhibition should have uncoupled P2Y receptors from the Ca2+ store discharge, thus preventing SOC channel activity, although the direct gating of Ca2+ channels by G-proteins or their activation via, for instance, the cAMP cascade would be still possible in the presence of U73122
[GenBank]
. This dissection favors SOC channels in mediating Ca2+ entry triggered by external ATP. Hence, we performed a number of experiments to identify SOC channels in mouse TRCs and to elucidate their contribution to ATP responses.
To enhance Ca2+ influx, external Ca2+ was elevated to 5 mM. Internal Ca2+ stores were depleted by the pretreatment of TRCs with thapsigargin (1 µM), an inhibitor of sarco(endo-)plasmic reticulum Ca2+-ATPase, in the presence of 100 nM Ca2+ in the bath (Fig. 6A). After 10-15 min of exposure to thapsigargin, the restoration of bath Ca2+ to 5 mM gave rise to a significant Ca2+ influx blockable by Gd3+, elevating intracellular Ca2+ to a level that exceeded the control one (Fig. 6A, dashed line) by 100-200%. Such a prominent increase in Ca2+ permeability in response to the depletion of Ca2+ stores was exhibited by all of 11 cells tested, the finding strongly arguing for the presence of Gd3+-sensitive SOC channels in the TRC plasma membrane. We envisaged that if external ATP emptied Ca2+ stores, then activation of SOC channels should also be observed. In a number of experiments (n = 13), TRCs were stimulated by ATP at the concentration of 0.01-1 mM with 100 nM Ca2+ in the bath. The rationale for such an experimental protocol was that a decrease of bath Ca2+ below 100 nM discharged Ca2+ stores, activating SOC channels per se (data not shown). In a typical experiment (n = 16), the drop in bath Ca2+ from 5 mM to 100 nM decreased intracellular Ca2+ but activated SOC channels negligibly (Fig. 6B). The bath application of ATP released reticular Ca2+ (Fig. 6B), thereby depleting Ca2+ stores. The expected activation of SOC channels indeed occurred as an increased Ca2+ flow was observed when the concentration of extracellular Ca2+ was restored to 5 mM (Fig. 6B; 11 cells). As with thapsigargin (Fig. 6A), the ATP-dependent Ca2+ influx was inhibited by Gd3+ ions (Fig. 6B). We assessed a value of Ca2+ influx triggered by the bath Ca2+ jump (Fig. 6, A and B) based on the following equation
 | (1) |
 | (2) |
|
|
With 5 mM Ca2+ in the bath, cytosolic Ca2+ was 143 ± 24 nM on average (n = 16). When bath Ca2+ was rapidly removed, intracellular Ca2+ fell to a lower level with the initial rate of 1.2-1.9 nM/s (Fig. 7A, left; n = 11) that is equivalent to 1,200-1,900 Ca2+ ions extruded per second from the TRC cytoplasm at intracellular Ca2+ of about 140 nM. Given that Ca2+ efflux mediated by Na+/Ca2+ exchangers and Ca2+-ATPases is characterized by the apparent dissociation constants of about 1 and 0.2 µM, respectively (DiPolo and Beauge 1999), we assumed that a rate of Ca2+ extrusion at low intracellular Ca2+ is proportional to its concentration. In other words, at the resting concentration of 68 ± 19 nM, which is characteristic of cytosolic Ca2+ with 100 nM Ca2+ in the bath (n = 21), the extrusion rate should be 143/68 times lower than that at 143 nM, or 0.76 ± 0.15 nM/s on average. Figure 7B summarizes the overall experiments and indicates that both thapsigargin and ATP increase significantly Ca2+ flow, i.e., Ca2+ permeability. Particularly, the pretreatment of TRCs with 10 µM ATP led to more than twofold increase in Ca2+ influx (Fig. 7B).
Although these results point to SOC channels as a major pathway for Ca2+ entry coupled to P2Y receptor activation, some contribution of receptor-operated Ca2+ channels gated by the PLC-controllable second messengers diacylglycerol (DAG) and phosphatidylinositol 4,5-bisphosphate (PIP2) may not be ruled out entirely. In diverse preparations, DAG has been found to regulate a variety of Ca2+ channels mediating Ca2+ entry either directly or by activating protein kinase C (summarized in Clapham et al. 2001; Hardie 2003; Petersen and Fedirko 2001). Because in our preliminary experiments, the DAG analog 1,2-dioctanoyl-sn-glycerol (10-100 µM) and the PKC activator phorbol-12 myristate-13-acetate (200 nM) only negligibly affected intracellular Ca2+ at rest (n = 7; data not shown), DAG/PKC-gated Ca2+ channels are apparently nonfunctional in ATP responsive TRCs. We did not study yet a role of PIP2 in the regulation of ion channels in TRCs. However, given that the hydrolysis of PIP2 generally inactivates channels rather than enhances their activity (e.g., Kobrinsky et al. 2000; Runnels et al. 2002; Suh and Hille 2002; Yue et al. 2002), it is highly unlikely that TRCs express Ca2+ channels, which are inhibited at rest due to the direct channel-PIP2 interaction and activated on TRC stimulation by ATP, owing to PIP2 degradation. Thus we believe that receptor-operated channels are not involved in ATP transduction.
Mechanisms of Ca2+ extrusion
Several basic processes shape intracellular Ca2+ signals, including Ca2+ extrusion by the plasma membrane Ca2+-ATPase (PMCA) and Na+/Ca2+ exchange and Ca2+ sequestering by the sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA) and mitochondrial Ca2+ uptake (Berridge et al. 2000; Duchen 2000; Penniston and Enyedi 1998; Philipson and Nicoll 2000). In both excitable and nonexcitable cells, the PMCA and Na+/Ca2+ exchange are mainly responsible for Ca2+ extrusion (Carafoli 1991; Penniston and Enyedi 1998). Generally, Na+/Ca2+ exchange is characterized by a low affinity but a high capacity for Ca2+, while the Ca2+-ATPase has a high affinity but a low capacity for the divalent cation (Penniston and Enyedi 1998; Philipson and Nicoll 2000). In many cells, Na+/Ca2+ exchange dominates at large intracellular Ca2+ loads, while the PMCA plays a housekeeping role, mediating the fine tuning of the intracellular Ca2+ level below the concentration of about 200 nM. (Carafoli 1991; Penniston and Enyedi 1998).
In a number of experiments, we assessed a relative contribution of different mechanisms to the fall phase of ATP responses, i.e., to the extrusion of Ca2+ ions from the TRC cytoplasm. To inhibit Na+/Ca2+ exchange, bath Na+ was substituted for impermeable N-methyl-D-glucamine (NMG) cations. It turned out that in control (Fig. 8A) and in the absence of external Na+ (Fig. 8B; n = 8), ATP responses exhibited virtually identical kinetics of the fall phases, as illustrated by the normalized responses superimposed in Fig. 8C. Therefore in contrast to photo- and olfactory receptor cells and muscle cells, where exclusively Na+/Ca2+ exchangers pump out Ca2+ ions entering into the cytoplasm during cell excitation (Blaustein and Lederer 1999; Cervetto et al. 1989; Reisert and Matthews 1998), this mechanism contributes negligibly to the extrusion of Ca2+ ions on TRC activation by ATP. Although we did not expect the reticular Ca2+-ATPase to mediate the fall phase of ATP responses to a high extent, some TRCs were examined in the presence of thapsigargin to determine a contribution of the SERCA to ATP response termination.
|
To distinguish the direct effects of thapsigargin on the fall phase caused by SERCA inhibition from the indirect influence due to SOC channel activation, the inhibitor was applied with 100 nM Ca2+ in the bath. Apparently, no change in response kinetics was observed in such experiments compared with control ATP responses (Fig. 8, D and F; n = 9). Meanwhile, the SERCA was indeed inhibited by thapsigargin within the 30- to 40-s period of preincubation, because significant Ca2+ influx occurred in the presence of 1 mM bath Ca2+ (n = 4), indicating the activation of SOC channels (Fig. 8E). Thus a relative contribution of the SERCA to the removal of Ca2+ mobilized during ATP response is negligible.
To disrupt mitochondrial function, TRCs were preincubated with 1 µM FCCP, a protonophore collapsing inner membrane potential. In all of four such experiments, TRCs treated with FCCP for nearly 5 min exhibited ATP responses with the falling phase, which kinetics was statistically indistinguishable from that of control responses (data not shown), suggesting that in ATP-responsive TRCs, mitochondria are only a secondary Ca2+ clearing system. Together, the above findings demonstrate, albeit indirectly, that the PMCA is a major Ca2+ pump extruding Ca2+ from ATP-sensitive TRCs.
Of the known inhibitors of PMCA activity, only caloxin 2A1 has been reported to inhibit specifically certain PMCA isoforms (Holmes et al. 2003). The others, such as Hg2+ and La3+ or orthovanadate, may affect intracellular Ca2+ via diverse mechanisms and for this reason they were not used by us. When added to the bath solution, caloxin 2A1 (2 mM) influenced neither a resting Ca2+ level in TRCs nor the kinetics of Ca2+ extrusion during ATP responses (n = 3; Fig. 8, G and H). Thus mouse TRCs do not express caloxin-sensitive PMCA isoforms.
Modulation of the P2Y cascade by cyclic AMP
Mounting evidence suggests reciprocal and synergistic links between cAMP and IP3/calcium pathways. For instance, Ca2+ both stimulates and inhibits distinctive isoforms of adenylyl cyclase (AC) (Hanoune and Defer 2001) and activates Ca2+-calmodulin-dependent phosphodiesterase (PDE) (Mehats et al. 2002), while many of agonists, which stimulate AC, suppress IP3 production (Fisher 1995). The ability of cAMP to inhibit distinctive PLC isoforms is apparently mediated by their phosphorylation. Particularly, cAMP-dependent protein kinase (PKA) has been found in vitro to phosphorylate PLC-2, the only PLC isoform identified so far in mammalian TRCs, thereby diminishing its activation by G subunits (Liu and Simon 1996). PKA-mediated phosphorylation of IP3 receptors, one more pivotal element of the phosphoinositide cascade, is also of physiological significance, because this covalent modification enhances IP3-induced Ca2+ release, irrespective of the receptor subtype (Thrower et al. 2001; Wojcikiewicz and Luo 1998). We therefore studied whether intracellular cAMP modulates TRC responsiveness to ATP.
A number of compounds are conventionally used to alter a level of intracellular cAMP, including forskolin, a classical AC activator, AC and PDE inhibitors, and membrane permeable cAMP analogs. We found that at the concentrations of 10-50 µM typically used for AC activation, forskolin directly blocks K+ channels in mouse TRCs (S. S. Kolesnikov, unpublished observations) as in the case of pheochromocytoma cells (Garber et al. 1990), Aplisia sensory neurons (Baxter and Byrne 1990), smooth muscle cells (Inoue et al. 1993), and rat TRCs (Herness et al. 1997). As an alternative to the forskolin stimulation, intracellular cAMP was elevated in the TRC cytoplasm by the extracellular application of the permeable cAMP derivatives 8BrcAMP (400 µM) and/or Sp-cAMPS (200 µM) in the presence of IBMX (100 µM), a PDE inhibitor. With these compounds in the bath, the magnitude of ATP responses was reduced by 10-30% compared with control responses (n = 18, data not shown). However, these effects of the cAMP derivatives on ATP responses might not be conclusive because spontaneous TRC desensitization was frequently observed at serial applications of ATP (Fig. 9A; 9 cells). In another experimental series, TRCs were treated with the permeable AC inhibitor SQ 22536 (50 µM) to decrease a cAMP concentration in the TRC cytoplasm. Of 12 TRCs tested, 9 cells displayed ATP responses increased by 20-60% in the presence of SQ (Fig. 9B). In three experiments, the SQ effect was reversible (Fig. 9B). In contrast, six TRCs still exhibited increased ATP responses after the withdrawal of SQ from the bath (Fig. 9C), presumably because the AC inhibitor was washed out slowly from the cell cytoplasm. Since ATP responses never increased spontaneously, these results demonstrate conclusively the dependence of TRC sensitivity on intracellular cAMP.
The increase in the magnitude of ATP responses induced by SQ (Fig. 9, B and C), which quite possibly causes no effects at the 50 µM concentration other than AC inhibition, can be explained by that at reduced cAMP, PKA mediated phosphorylation of PLC uncouples P2Y receptors from Ca2+ mobilization to a lesser extent (Liu and Simon 1996). If so, the inhibition of PKA should also have sensitized TRCs and mimicked the SQ effect on ATP responses. Therefore in a number of experiments, TRCs were studied in the presence of the PKA inhibitor H-89. The compound caused diverse effects on TRCs, the most marked of which was the prominent recovery of responsiveness of the cells that exhibited nontypical subtle ATP responses with a very small initial peak (Fig. 10A). The pretreatment with 200 nM H-89 rendered such TRCs generative of fairly robust ATP responses on the PKA inhibition (Fig. 10A; n = 3). For cells exhibiting ATP responses with a relatively small peak in respect to a plateau (2 < peak/plateau < 4; see Fig. 4A), H-89 augmented the peak magnitude by 20-40% and increased the peak to plateau ratio to the value of 4-6 (n = 5; Fig. 10B). The PKA inhibitor affected slightly or negligibly robust ATP responses with the peak to plateau ratio of about 5-8 (n = 4; Fig. 10C). Together, these observations point to that cAMP-dependent phosphorylation controls a gain of the signaling cascade mediating ATP responses. Perhaps, individual variations in a resting cAMP level account for a scatter in ATP responsiveness observed among TRCs (Fig. 10).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
For G-protein signaling, prolonged exposure of cells to agonists triggers a chain of events that result in the deactivation of a G-protein cascade and a rapid attenuation in cell responsiveness, owing to a negative feedback regulation of a cascade gain (Bunemann et al. 1999; Burns and Baylor 2001; Fain et al. 2001; Ronnett and Moon 2002 for review). Although our experiments do not allow the dissection of ATP signaling in great detail, two processes, the deactivation of the phosphoinositide cascade and the removal of mobilized Ca2+ from the cell cytoplasm, are most likely to determine the fall phase of ATP responses by analogy with the well-studied systems (Figs. 3, 4, 5). By excluding the Na+/Ca2+ exchange, SERCA, and mitochondria, our findings (Fig. 8) point to PMCA as a major Ca2+ pump mediating Ca2+ clearance during an ATP response. Despite the strong indirect evidence (Fig. 8, D-F), this inference remains to be validated by the direct inhibition of PMCA. To our knowledge, the only known specific PMCA inhibitor is caloxin 2A1, a peptide inhibiting the activity of certain PMCA isoforms, i.e., PMCA1b and PMCA4, from outside, with EC50 of approximately 1 mM (Holmes et al. 2003). In our experiments (Fig. 8, G and H), this compound affected negligibly both resting Ca2+ and the kinetics of Ca2+ transients triggered by ATP, leading us to the conclusion that the caloxin-sensitive PMCA isoforms do not operate in mouse TRCs.
We found ATP and UTP to release Ca2+ in the TRC cytoplasm with nearly identical efficacies, while ADP appeared to be a partial agonist (Fig. 3B). Of the receptor subtypes, from P2Y1 to P2Y14, identified in different tissues (Abbracchio et al. 2003), P2Y2 and P2Y4 receptors exhibit the related sequences of agonist potencies (e.g., Bogdanov et al. 1998; King et al. 1998; Santiago-Perrez et al. 2001). This may be indicative of that just the P2Y2 and/or P2Y4 isoforms mediate TRC sensitivity to ATP, albeit the purinoceptors involved remain to be characterized on the molecular level. The P2Y receptors regulate a variety of cellular functions via G-proteins from the Gi/o and Gq/11 subfamilies, which couple them to PLC activation, adenylate cyclase stimulation/inhibition, or ion channel modulation (Communi et al. 1997, 2001; Cooper and Rodbell 1979; Currie and Fox 1996; Erb et al. 2001; Filippov et al. 1998; Ikeushi and Nishizaki 1996; Liu and Rosenberg 2001; Lu et al. 2000; Pucéat et al. 1998). Eight G-protein -subunits have been identified in TRCs: Gs, Gi-2, Gi-3, Gq, G14, G15, -transducin, and -gustducin (Margolskee 2002). With the likely exception of Gs, one or more of these G-proteins may play a role in ATP transduction. Note, however, that no other PLC isozymes have been identified so far in mammalian TRCs but PLC2 (Asano-Miyoshia et al. 2000; Huang et al. 1999; Rossler et al. 1998), knockout of which abolishes the perception of sweet and bitter compounds and amino acids (Zhang et al. 2003). From these results, a possible inference is that of the known PLC isozymes regulated by G-proteins, only the PLC2 isoform is expressed in TRCs, and by that, PLC2 is an obligatory element of the phosphoinositide cascade mediating ATP responses. Although the data obtained from co-expression and reconstruction experiments point to both -subunits of the Gq/11 type and -subunits of Gi proteins as likely regulators of PLC2, existing evidence suggests this PLC isoform to be specifically activated in vivo by Gi -subunits (reviewed in Rebecchi and Pentyala 2000; Singer et al. 1997). Thus a G-protein that presumably belongs to the Gi group expressed in TRCs, i.e., Gi-2, Gi-3, transducin, or gustducin, links a P2Y receptor to PLC2. Future experiments are necessary to define the nature of P2Y receptors and G-proteins involved in ATP signaling in TRCs.
Whereas functional features of receptor-operated and capacitative Ca2+ entry are well characterized, the molecular identity of ion channels involved is not established with confidence. Based on ion permeability and gating and regulatory properties, certain members of the TRP family of cation channels are believed to function as either receptor-operated channels or SOC channels (Clapham et al. 2001; Hardie 2003; Petersen and Fedirko 2001 for recent reviews). Recently, TRP channels, TRPM5, were identified in taste cells (Perez et al. 2002; Zhang et al. 2003). When heterologously expressed in oocytes, TRPM5 enhanced capacitative Ca2+ entry, suggesting that in TRCs, they operate as SOC channels (Perez et al. 2002). In HEK-293 cells transfected with TRPM5, these channels functioned in the receptor-operated mode and independently of store depletion (Zhang et al. 2003). As a possible reason for this discrepancy, in oocytes and in HEK293 cells, TRPM5 might be coupled to distinct signaling cascades and/or levels of channel expression might be different. For example in pontine neurons, TRPC3 channels mediate Ca2+ entry in response to activation of PLC-coupled receptors rather than Ca2+ store depletion (Li et al. 1999). In the expression systems, TRPC3 channels exhibited a receptor-operated phenotype at high levels of expression, and they operated exclusively as SOC channels at a low level of expression (Vazquez et al. 2001; Venkatachalam et al. 2001). Whatever the case, it remains to be elucidated if TRPM5 are receptor-operated or SOC channels in TRCs and whether they are involved in purinergic signaling.
The question about a source of extracellular ATP in a taste bud is of physiological importance. As a possibility, efferent nerve fibers or specialized taste bud cells may release ATP to modulate taste transduction machinery in targeted TRCs. Particularly, if P2Y receptors are indeed coupled to PLC2 via -subunits of the Gi-proteins, -subunits, which are liberated on TRC stimulation by ATP, may either inhibit AC (Gi-2, Gi-3) or activate PDE (transducin, gustducin). In each case, both a degradation of PIP2 to DAG and IP3 and a decrease in cytoplasmic cAMP would simultaneously occur. Therefore in TRCs expressing P2Y receptors coupled to PLC and AC, extracellular ATP may determine resting levels of several second messengers such as PIP2, DAG, Ca2+, and cAMP, and by that, the nucleotide may effectively modulate taste transduction machinery in a dose-dependent fashion.
It should be noted in conclusion that taste buds were isolated from circumvallate, foliate, and fungiform papilla and that ATP-responsive TRCs were present in a great abundance in virtually every bud tested. In preliminary experiments, we found ATP responsive TRCs in the rat circumvallate papillae. Collectively, our findings point to that ATP may be a ubiquitous neurotransmitter/neuromodulator operating in taste buds of rodents, if not all mammalians. It is of interest to phenotype taste bud cells involved in ATP signaling, some releasing ATP and some being a target.
|
DISCLOSURES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests: S. S. Kolesnikov (E-mail: staskolesnikov{at}yahoo.com).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESREFERENCES |
|---|
Asano-Miyoshia M, Abe K, and Emori Y. Co-expression of calcium signaling components in vertebrate taste bud cells. Neurosci Lett 283: 61-64, 2000.[ISI][Medline]
Baxter DA and Byrne JH. Reduction of voltage activated K+ currents by forskolin is not mediated via cAMP in pleural sensory neurons of Aplysia. J Neurophysiol 64: 1474-1483, 1990.
Behe P, DeSimone JA, Avenet P, and Lindemann B. Membrane currents in taste cells of the rat fungiform papilla: evidence for two types of Ca currents and inhibition of K currents by saccharin. J Gen Physiol 96: 1061-1084, 1990.
Bernhard SJ, Naim M, Zehavi U, and Lindemann B. Change in IP3 and cytosolic Ca2+ in response to sugars and non-sugar sweeteners in transduction of sweet taste in the rat. J Physiol 490: 325-336, 1996.[ISI]
Berridge MJ, Lipp P, and Bootman MD. The versatility and universality of calcium signaling. Nature Rev Mol Cell Biol 1: 11-21, 2000.[ISI][Medline]
Bigiani A, Delay RJ, Chaudhari N, Kinnamon SC, and Roper SD. Responses to glutamate in rat taste cells. J Neurophysiol 77: 3048-3059, 1997.
Blaustein MP and Lederer WJ. Sodium/calcium exchange: its physiological implications. Physiol Rev 79: 763-854, 1999.
Bogdanov YD, Wildman SS, Clements MP, King BF, and Burnstock G. Molecular cloning and characterization of rat P2Y4 nucleotide receptor. Br J Pharmacol 124: 428-430, 1998.[ISI][Medline]
Boughter JD, Pumplin DW, Yu C, Christy RC, and Smith DV. Differential expression of alpha-gustducin in taste bud populations of the rat and hamster. J Neurosci 17: 2852-2858, 1997.
Brown P and Dale N. Spike-independent release of ATP from Xenopus spinal neurons evoked by activation of glutamate receptors. J Physiol 540: 851-860, 2002.
Bunemann M, Lee KB, Pals-Rylaarsdam R, Roseberry AG, and Hosey MM. Desensitization of G-protein coupled receptors in the cardiovascular system. Annu Rev Physiol 61: 169-192, 1999.[ISI][Medline]
Burns ME and Baylor DA. Activation, deactivation, and adaptation in vertebrate photoreceptor cells. Annu Rev Neurosci 24: 779-805, 2001.[ISI][Medline]
Burnstock G. Purine-mediated signalling in pain and visceral perception. Trends Pharmacol Sci 22: 182-188, 2001.[Medline]
Caicedo A, Jafri MS, and Roper SD. In situ Ca2+ imaging reveals neurotransmitter receptors for glutamate in taste receptor cells. J Neurosci 20: 7978-7985, 2000.
Carafoli E. Calcium pump of the plasma membrane. Physiol Rev 71: 129-153, 1991.
Cervetto L, Lagnado L, Perry RJ, Robinson DW, and McNaughton PA. Extrusion of calcium from rod outer segments by both sodium and potassium gradients. Nature 337: 740-743, 1989.[Medline]
Chaudhari N, Landin AM, and Roper SD. A novel metabotropic glutamate receptor is a taste receptor for monosodium L-glutamate. Nat Neurosci 3: 113-119, 2000.[ISI][Medline]
Chaudhari N, Yang H, Lamp C, Delay E, Cartford C, Than T, and Roper S. The taste of monosodium glutamate: membrane receptors in taste buds. J Neurosci 16: 3817-3826, 1996.
Clapham DE, Loren W, Runnels LW, and Strübing C. The TRP ion channel family. Nat Rev Neurosci 2: 387-396, 2001.[ISI][Medline]
Communi D, Govaerts C, Parmentier M, and Boeynaems JM. Cloning of a human purinergic P2Y receptor coupled to phospholipase C and adenylyl cyclase. J Biol Chem 272: 31969-31973, 1997.
Communi D, Suarez Gonzales N, Detheux M, Brezillon S, Lannoy V, Parmentier M, and Boeynaems J-M. Identification of a novel human ADP receptor coupled to Gi. J Biol Chem 276: 41479-41485, 2001.
Cooper DMF and Rodbell M. ADP is a potent inhibitor of human platelet plasma membrane adenylate cyclase. Nature 282: 517-518, 1979.[Medline]
Currie KPM and Fox AP. ATP serves as a negative feedback inhibitor of voltage-gated Ca2+ channel currents in cultured bovine adrenal chromaffin cells. Neuron 16: 1027-1036, 1996.[ISI][Medline]
Delay RJ, Kinnamon SC, and Roper SD. Serotonin modulates voltage dependent calcium current in Necturus taste cells. J Neurophysiol 77: 2515-2524, 1997.
Delay RJ and Roper SD. Ultrastructure of taste cells and synapses in the mudpuppy Necturus maculosus. J Comp Neurol 277: 268-280, 1988.[ISI][Medline]
Di Virgilio F and Solini A. P2 receptors: new potential players in atherosclerosis. Br J Pharmacol 135: 831-842, 2002.[ISI]
DiPolo R and Beauge L. Metabolic pathways in the regulation of invertebrate and vertebrate Na+/Ca2+ exchange. BBA 1422: 57-71, 1999.[Medline]
Diver JM, Sage SO, and Rosado JA. The inositol trisphosphate receptor antagonist 2-aminoethoxydiphenylborate (2-APB) blocks Ca2+ entry channels in human platelets: cautions for its use in studying Ca2+ influx. Cell Calcium 30: 323-329, 2001.[ISI][Medline]
Duchen MR. Mitochondria and calcium: from cell signalling to cell death. J Physiol 529: 57-68, 2000.
Dunn PM, Zhong Y, and Burnstock G. P2X receptors in peripheral neurons. Prog Neurobiol 65: 107-134, 2001.[ISI][Medline]
Eatock RA. Adaptation in hair cells. Annu Rev Neurosci 23: 285-314, 2000.[ISI][Medline]
Erb L, Liu J, Ockerhausen J, Komng Q, Garrad RC, Griffin K, Nneal C, Krugh B, Santiago-Perez LI, Gonzales FA, Gresham HD, Turner JT, and Weisman GA. An RGD sequence in the P2Y2 receptor interacts with avb3 integrins and is required for Go-mediated signal transduction. J Cell Biol 153: 491-502, 2001.
Ewald DA and Roper SD. Cholinergic responses of taste cells in Necturus taste buds. Soc Neurosci Abstr
mean ± SE) vs. ATP (8 to 43 cells), UTP (3 to 9 cells), and ADP (4 to 8 cells) concentration. A response magnitude was calculated as the difference between a Ca2+ concentration in the peak and that at rest immediately before the ATP application. In B, the continuous thick and thin lines and the dotted line correspond to the Hill equation 
