Functional expression of ionotropic purinergic receptors on mouse taste bud cells

Abstract

Neurotransmitter receptors on taste bud cells (TBCs) and taste nerve fibres are likely to contribute to taste transduction by mediating the interaction among TBCs and that between TBCs and taste nerve fibres. We investigated the functional expression of P2 receptor subtypes on TBCs of mouse fungiform papillae. Electrophysiological studies showed that 100 μm ATP applied to their basolateral membranes either depolarized or hyperpolarized a few cells per taste bud. Ca²⁺ imaging showed that similarly applied 1 μm ATP, 30 μm BzATP (a P2X₇ agonist), or 1 μm 2MeSATP (a P2Y₁ and P2Y₁₁ agonist) increased intracellular Ca²⁺ concentration, but 100 μm UTP (a P2Y₂ and P2Y₄ agonist) and α,β-meATP (a P2X agonist except for P2X₂, P2X₄ and P2X₇) did not. RT-PCR suggested the expression of P2X₂, P2X₄, P2X₇, P2Y₁, P2Y₁₃ and P2Y₁₄ among the seven P2X subtypes and seven P2Y subtypes examined. Immunohistostaining confirmed the expression of P2X₂. The exposure of the basolateral membranes to 3 mm ATP for 30 min caused the uptake of Lucifer Yellow CH in a few TBCs per taste bud. This was antagonized by 100 μm PPADS (a non-selective P2 blocker) and 1 μm KN-62 (a P2X₇ blocker). These results showed for the first time the functional expression of P2X₂ and P2X₇ on TBCs. The roles of P2 receptor subtypes in the taste transduction, and the renewal of TBCs, are discussed.

Heterogeneous taste bud cells (TBCs) and sensory nerve fibres form mammalian taste buds, and their interaction through chemical signalling is suggested to trigger not only taste sensation but also the turnover of TBCs. In fact, TBCs release serotonin (Huang et al. 2005) and ATP (Finger et al. 2005) in response to taste stimuli, and release ATP on depolarization through specific channels (Romanov et al. 2007). Recently, it has been shown that TBCs release ATP in response to taste stimulation via pannexin 1 hemichannels and that the released ATP causes the release of serotonin from other TBCs in mouse tongues (Huang et al. 2007). Also, TBCs and associated nerve fibres express a variety of receptors for neurotransmitters, including cholecystokinin (Herness et al. 2002b), serotonin (Kaya et al. 2004), adrenaline (Herness et al. 2002a), acetylcholine (Ogura, 2002) and ATP (Bo et al. 1999; Kim et al. 2000; Rong et al. 2000; Kataoka et al. 2004). As the first step to test for such chemical signalling, we used several methods to investigate the functional expression of P2 receptors, receptors for ATP, on TBCs in mouse fungiform papillae.

TBCs are classified into four cell types, type I to type IV. Type I cells are assumed to be supportive. Type II cells are taste receptor cells, but they have no chemical synapses with taste nerve fibres. Only type III cells form chemical synapses with taste nerve fibres (Seta & Toyoshima, 1995). Type II cells thus may function via novel type chemical synapses (Clapp et al. 2004), or paracrine signalling systems (Kaya et al. 2004) to transmit taste information to type III cells or to taste nerve endings.

On the other hand, these TBCs are short lived. Their lifetime is estimated to be about 9 days (Beidler & Smallman, 1965), although a subset of type II cells lives longer (Farbman, 1980; Cho et al. 1998). It is suggested that older TBCs are eliminated by apoptotic pathways (Zeng & Oakley, 1999; Zeng et al. 2000) and are replaced with newborn TBCs derived from type IV cells, precursor cells for all cell types. It appears that P2 receptor subtypes contribute to the apoptosis.

Mammalian P2 receptors consist of seven P2X subtypes (ion channels) and eight P2Y subtypes (G-protein-coupled receptors). The expression of P2X₂ and P2X₃ was shown on rat taste nerve fibres immunohistochemically (Bo et al. 1999), and that of P2X₃ was suggested electrophysiologically on trigeminal afferents (Rong et al. 2000). P2Y subtype expression was shown pharmacologically in mouse TBCs (Kim et al. 2000; Baryshnikov et al. 2003), and later identified as P2Y₁ in rat TBCs (Kataoka et al. 2004), and as P2Y₂ and P2Y₄, and probably P2Y₁ and P2Y₆, in mouse TBCs in foliate and circumvallate papillae (Bystrova et al. 2006). However, the expression of other P2 receptor subtypes, typically that of P2X subtypes, remains to be identified.

In the present study, we show the expression of P2X₂ and P2X₇, in addition to P2Y₁, on the basolateral membranes of mouse fungiform TBCs, with a variety of techniques including voltage- and current-clamping, reverse transcriptase-mediated polymerase chain reactions (RT-PCR), immunohistostaining, Ca²⁺ imaging, and dye uptake. Also we show the cell-type-dependent expression of P2 subtypes. The roles of these purinergic receptors are discussed in terms of TBC networks and TBC turnover.

Methods

Preparation of peeled epithelia

Mouse lingual epithelia were prepared as described in our previous papers (Ohtubo et al. 2001; Higure et al. 2003). In brief, ddY-strain mice were killed by decapitation after being anaesthetized with ether. We hypodermically injected an elastase solution (see ‘Solutions’ for the composition of this and following solutions) into the tongues, incubated them at 25°C for 10 min, and then peeled off the lingual epithelia with forceps. We mounted an epithelium on a recording platform with the basolateral membrane side of TBCs (the opposite side of the tongue surface) facing upward (and receptor membranes facing inside the recording platform), and placed the recording platform under a microscope equipped with a water-immersion objective. The basolateral membrane was always irrigated with either a physiological saline or stimulating solutions, to minimize the hydrolysis of ATP in the stimulating solutions. The time constant of the exchange of solutions estimated with Rhodamine B, a fluorescent dye, was 3.5 s. The receptor membrane was acclimated to the physiological saline. All experimental protocols were conducted in compliance with the Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences approved by the Council of the Physiological Society of Japan.

Electrophysiology

Electrophysiological studies under cell-attached and whole-cell-clamp conditions were carried out according to our previous papers (Ohtubo et al. 2001; Higure et al. 2003).

In brief, electrical responses were recorded with electrodes (∼5 MΩ) filled with a KCl electrode solution, amplified with a voltage-clamp amplifier (Axopatch 200B, Axon Instruments, Union City, CA, USA), filtered at 10 kHz, digitized with an A/D converter (Digidata 1200, Axon Instruments), and stored using pCLAMP data acquisition and analysis software (ver. 9.0, Axon Instruments) on a personal computer. The liquid junction potential of ∼5 mV between recording and reference electrodes was neglected.

Ca²⁺ imaging

After mounting on a recording platform, we soaked the basolateral membrane side of the peeled epithelium in a fura-2 AM (acetoxymethyl ester form of the Ca²⁺-selective fluorescent dye, fura-2) solution for 20 min at 37°C, washed with the physiological saline for 10 min, and then placed under a fluorescence microscope equipped with a water-immersion objective. The basolateral membrane was irrigated similarly in the whole-cell-clamp experiments, except that the Ca²⁺ concentration was decreased to 1 mm (lowered-Ca²⁺ saline) to prevent the decrease of Ca²⁺ responses during the repeated application of ATP. The receptor membrane was acclimated to the physiological saline.

Fura-2-stained TBCs and non-taste lingual epithelial cells were excited at 340 and 380 nm with a spectroscope-type high-speed wavelength changer (C7773, Hamamatsu Photonics, Hamamatsu, Japan). Images of fura-2 fluorescence were acquired every 2.5 s with an intensified CCD camera through a water-immersion objective (Fluor-40X, Olympus, Tokyo, Japan), stored in a computer, and analysed with AQUACOSMOS software (version 2.0, Hamamatsu Photonics). Averaged Ca²⁺ levels over respective cell areas were sequentially plotted as a ratio of F₃₄₀/F₃₈₀. Whenever the ratio obtained from given TBCs or non-taste lingual epithelial cells increased by 0.2 or more, in a time-locked manner with the stimulus, we concluded that these cells displayed Ca²⁺ responses to ATP or its analogues.

RT-PCR

The expression of P2 receptor subtypes on TBCs and non-taste lingual epithelial cells was tested with RT-PCR experiments using oligonucleotide primers specific for 14 different P2 receptor subtypes (Table 1). In order to obtain the total mRNA of TBCs, we liberated taste buds by puffing a Ca²⁺- and Mg²⁺-nominally-free solution onto the cleft between taste buds and surrounding epithelial cells in the physiological saline (Fig. 1). The liberated taste buds were collected with another micropipette, transferred into Isogen (Wako Pure Chemical Industries, Ltd, Osaka, Japan), homogenized and stored at −80°C. Lingual epithelia containing no taste buds among fungiform papillae (non-taste lingual epithelia) were immediately homogenized in Isogen, and stored at −80°C. Total mouse brain mRNA similarly extracted and stored was used as a control to test the primer sets.

Table 1.

Summary of gene-specific primer sets employed for varying ATP receptors

Subtypes	Primer sequence 5′-sense-3′ 5′-antisense-3′	Expected product size	Position	Reference
P2X1	5′-CATTGTGCAGAGAACCCAGAA-3′	776 bp	530–550	(Coutinho-Silva et al. 1999)
	5′-ATGTCCTCCGCATACTTGAAC-3′		1285–1305
P2X2	5′-ACGTTCATGAACAAAAACAAG-3′	360 bp	1355–1375	(Ren et al. 2003)
	5′-TCAAAGTTGGGCCAAACCTTTGG-3′		1692–1714
P2X3	5′-CTGTATATCAGACTTCTTCACCTACGA-3′	596 bp	194–220	(Ren et al. 2003)
	5′-TTATGTCCTTGTCGGTGAGGTTAG-3′		766–789
P2X4	5′-GAGAATGACGCTGGTGTGCC-3′	356 bp	533–552	(Dai et al. 2001)
	5′-TTGGTGAGTGTGCGTTGCTC-3′		869–888
P2X5	5′-TCCACCAATCTCTACTGC-3′	400 bp	743–760	(Dai et al. 2001)
	5′-CCAGGTCACAGAAGAAAG-3′		1125–1142
P2X6	5′-TACGTACTAACAGACGCA-3′	254 bp	7–24	(Dai et al. 2001)
	5′-ATATCAGGGTTCTTTGGG-3′		243–260
P2X7	5′-AAGTCTCTGCCTGGTGTC-3′	401 bp	499–516	(Dai et al. 2001)
	5′-GGCATATCTGAAGTTGTAGC-3′		880–899
P2Y1	5′-TGGCGTGGTGTACCCTCTCAAGTC-3′	558 bp	1101–1124	(Shariatmadari et al. 2003)
	5′-CGGGACAGTCTCCTTCTGAATGTA-3′		1635–1658
P2Y2	5′-CTGGAACCCTGGAATAGCAC-3′	513 bp	489–508	(Fischer et al. 2001)
	5′-GCTGGTGGTGACGAAGTAGA-3′		982–1001
P2Y4	5′-AGCCCAAGTTCTGGAGATGGTG-3′	492 bp	341–362	(Suarez-Huerta et al. 2001)
	5′-GGTGGTTCCATTGGCATTGG-3′		813–832
P2Y6	5′-CACCTGTGATTTGGCAACTG-3′	336 bp	1538–1557	BC027331
	5′-TCTTGGCAAATGGATGTGAA-3′		1854–1873
P2Y12	5′-CACCTCAGCCAATACCACCT-3′	453 bp	445–464	NM_027571
	5′-AACATGAAGGCCCAGATGAC-3′		878–897
P2Y13	5′-GAAGAGAGGCACATGCAACA-3′	345 bp	1656–1675	NM_028808
	5′-TTACTAATGCCAGGCCAACC-3′		1981–2000
P2Y14	5′-CAGTGCATGGAGCTCAAAAA-3′	347 bp	895–914	NM_133200
	5′-GCAGCCGAGAGTAGCAGAGT-3′		1222–1241
β-actin	5′-CACCCTGTGCTGCTCACC-3′	328 bp	381–398	NM_007393
	5′-GCACGATTTCCCTCTCAG-3′		691–708
α-gustducin	5′-CTCTTCCAGGAGAAAGTGGC –3′	248 bp	924–943	XM_144196
	5′-TCAGAAGAGCCCACAGTCTT –3′		1153–1172

Figure 1. Collection of fungiform taste buds.

Each taste bud in peeled lingual epithelium (A) is liberated by puffing the Ca²⁺- and Mg²⁺-nominally-free solution onto the cleft between taste buds and surrounding epithelial cells in the physiological saline (B), and brought to another micropipette (C) that sucked the taste bud. Perigemmal cells surrounding the taste bud remained in the epithelium (D). Scale bar, 10 μm.

Taste buds were obtained from 21 mice, non-taste lingual epithelia from seven mice, and brains from two mice. Total mRNA was extracted as a water-soluble fraction, precipitated in 37.5% isopropanol supplemented with Dr. GenTLE precipitation carrier (Takara BIO, Tokyo, Japan), washed by ethanol, and then treated with DNase (Takara BIO, Japan) following the manufacturer's protocol. RT-PCR was carried out with a QIAGEN OneStep RT-PCR kit (QIAGEN GmbH, Hilden, Germany).

Reverse transcription was performed at 55°C for 30 min. PCR cycles consisted of an initial step of 94°C for 7.5 min and 45 subsequent cycles of 94°C for 30 s (denaturation), 50–62°C for 1 min (annealing), 72°C for 3 min (extension) and final extension step of 72°C for 10 min. Annealing temperatures were: 50°C for P2X₆ and P2X₇, 54°C for P2X₃, P2X₄, P2X₅, P2Y₁, P2Y₆ and α-gustducin, 58°C for P2X₂, P2Y₁₂, P2Y₁₄ and β-actin, 62°C for P2X₁, P2Y₂, P2Y₄ and P2Y₁₃. PCR products were analysed by 1.5% agarose gel electrophoresis, stained with ethidium bromide (0.5 μg ml⁻¹), and visualized by UV illumination.

Optimum annealing temperatures were obtained by preliminary PCR experiments on brain mRNA; we tested a range of annealing temperatures from 50°C to 62°C in 4°C steps, and took an annealing temperature that yielded a clear single band with a correct size on agarose gels for each primer set. Controls for DNA contamination and PCR carryover were performed by exclusion of RNA from the RT-PCR reaction. Amplification of housekeeping genes, β-actin and of α-gustducin, a G-protein that specifically expressed in TBCs, was used as a positive control.

Immunohistochemistry

Immunohistostaining of P2 receptor subtypes and cell marker proteins was observed in optical slices obtained with confocal laser scanning microscopy (TCS-SL, Leica Microsystems CMS GmbH, Wetzlar, Germany). For serotonin immunohistostaining, mice were injected intraperitoneally with 5-hydroxytryptophan (80 mg kg⁻¹) 1 h before being killed as described in a previous paper (Clapp et al. 2004). We fixed peeled epithelia overnight with a fixation solution, washed with six 10-min changes of a phosphate-buffered solution (PBS), and incubated the epithelia in 10 mm citrate buffer (pH 6.0) for 20 min at 85°C. After a brief wash with the PBS, the epithelia were incubated in a blocking solution for 4 h.

The epithelia were then probed using indirect immunofluorescence. Two antibodies to P2X₂ (1 : 1000, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA; 1 : 300, Alomone Laboratories, Israel), an antibody to P2X₄ (1 : 300, Calbiochem, Darmstadt, Germany), and two antibodies to P2X₇ (1 : 50, Santa Cruz Biotechnology; 1 : 200, Alomone Laboratories) served as the primary antibodies. Alexa 488-conjugated donkey anti-rabbit IgG (1 : 400, Molecular Probes, Eugene, OR, USA), Alexa 555-conjugated donkey anti-mouse IgG (1 : 400, Molecular Probes), and Alexa 633-conjugated donkey anti-goat IgG (1 : 200, Molecular Probes) served as secondary antibodies. Each antibody was diluted with the blocking solution to the concentration indicated, respectively, above. The specificities of these secondary antibodies were determined in the absence of respective primary antibodies. For negative control, 1 μg of Alomone's P2X₂ antibody was pre-incubated with 1 μg of its corresponding antigen peptide for 1 h at room temperature according to the manufacturer's instructions.

The cell types of TBCs were identified with cell-type-specific marker proteins: IP₃R3 (Clapp et al. 2001) and PLCβ2 (DeFazio et al. 2006) for type II cells, SNAP-25 and serotonin for type III cells (Yang et al. 2000; Clapp et al. 2004; DeFazio et al. 2006). We thus took anti-SNAP-25 mouse monoclonal antibody (1 : 1000, Sigma, St Louis, MO, USA), anti-SNAP-25 goat polyclonal antibody (1 : 10, Santa Cruz Biotechnology), anti-serotonin mouse monoclonal antibody (1 : 50, Dako A/S, Denmark), anti-IP₃R3 mouse monoclonal antibody (1 : 50, BD, Franklin Lakes, NJ, USA), and anti-PLCβ2 rabbit antibody (1 : 100, Santa Cruz Biotechnology) as primary antibodies, and the respective antibodies mentioned above as the secondary antibody. Although electron microscopy is essential for the precise identifications of cell types, it is known that identifications with antibodies against IP₃R3 (Clapp et al. 2004) and SNAP-25 (Yang et al. 2000) agree with electron microscopic identifications.

Lucifer Yellow CH uptake

We examined the uptake of Lucifer Yellow CH (LY) through P2X₇ by irrigating the basolateral membranes of TBCs mounted on the recording platform with a LY solution or a low divalent cation LY solution in the presence and absence of 3 mm ATP for 30 min at 37°C. The flow rate was 0.5 ml min⁻¹. Also, 100 μm pyridoxalphosphate-6-azophenyl-2′,4′-disulphonic acid (PPADS) or 1 μm 1-(N,O-bis[5-isoquinolinesulphonyl]-N-methyl-l-tyrosyl)-4-phenylpiperazine (KN-62) was added to the low divalent cation LY solution in several experiments.

Solutions

All solutions were prepared with deionized water and the components were expressed in millimolar concentrations, unless otherwise noted. The KCl electrode solution comprised: 150 KCl, 5 MgCl₂, 10 EGTA, 5 Na₂ATP, 0.3 Na₃GTP, and 10 Hepes-KOH, pH 7.2. The physiological saline comprised: 150 NaCl, 5 KCl, 2 CaCl₂, 0.5 MgCl₂, 10 glucose, and 5 Hepes-NaOH, pH 7.4. The elastase solution comprised: 1 mg ml⁻¹ elastase dissolved in the physiological saline. The low Ca²⁺ saline comprised: 150 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 10 glucose, and 5 Hepes-NaOH, pH 7.4. The Ca²⁺-nominally-free solution comprised: 150 NaCl, 5 KCl, 2 MgCl₂, 10 glucose, and 5 Hepes-NaOH, pH 7.4. The Ca²⁺- and Mg²⁺-nominally-free solution comprised: 140 NaCl, 5 KCl, 2 EGTA, 10 glucose and 10 Hepes, pH 7.4. The PBS comprised: 137 NaCl, 2.67 KCl, 8.09 Na₂HPO₄, and 1.47 KH₂PO₄. The low divalent cation saline comprised: 155 NaCl, 5 KCl, 0.5 CaCl₂, 10 glucose, and 5 Hepes-NaOH, pH 7.4. The fixation solution was the mixture of fresh 4% paraformaldehyde and Zambonic solution, both of which were prepared with the phosphate buffer. The blocking solution comprised: 3% normal donkey serum dissolved in the PBS and supplemented with 1% bovine serum albumin and 0.3% Triton X. Fura-2 AM solution comprised: 10 μm fura-2 AM dissolved in the physiological saline supplemented with 0.25% pluronic F-127. The LY solution was prepared by dissolving 0.5 mg ml⁻¹ LY in the physiological saline. The low divalent cation LY solution was prepared by dissolving 0.5 mg ml⁻¹ LY in the low divalent cation saline.

Results

Electrophysiological studies

A subset of TBCs responded to 100 μm ATP applied to their basolateral membranes in cell-attached and whole-cell electrophysiological recording conditions. In cell-attached recording, 3 cells out of 10 increased their firing frequency, and one cell decreased its frequency (Fig. 2). In whole-cell current-clamp mode, three cells out of 12 gave generator potentials, and increased firing frequency (Figs 3 and 4). The generator potentials often comprised discrete subthreshold depolarizing events which appeared to sum and lead to spiking (Fig. 3). Under the whole-cell voltage-clamp mode, 3 cells out of 12 increased membrane conductance, and generated an inward current (Fig. 5). The estimated reversal potential was ∼−10 mV. In contrast, one cell decreased inward current magnitude by decreasing membrane conductance. The estimated reversal potential was ∼−32 mV.

Figure 2.

Increasing (A) and decreasing (B) firing frequencies of different taste bud cells (TBCs) recorded under the cell-attached mode in response to 100 μm ATP. Bars over traces show the duration of the application of ATP to the basolateral membranes of TBCs, in this and following figures.

Figure 3.

Oscillating, depolarizing receptor potential with action potentials in response to 100 μm ATP (A), time-expanded sections labelled with numbered arrowheads (B and C), obtained under whole-cell current-clamp conditions. Holding currents, −20 pA.

Figure 4.

Non-oscillating depolarizing receptor potential with increasing firing frequency in response to 100 μm ATP (A), and time-expanded sections labelled with numbered arrowheads (B and C), obtained under zero-current whole-cell current-clamp mode.

Figure 5. Inward (A) and outward (B) currents of different TBCs, elicited by 100 μM ATP applied to the basolateral membrane of TBCs recorded under a whole-cell voltage-clamp mode, and their respective current–voltage relations (C and D).

C and D, plotted data are before (^), during (▴), and after (□) the application of ATP, as respective symbols label on current noises in A and B show. The reversal potential of ∼−10 mV for A was estimated by extrapolating the relations obtained with test potentials of −80 mV to −60 mV in 10-mV steps from a holding potential of −70 mV (C). That of −32 mV for B was similarly estimated with test potentials of −80 mV to −60 mV in 5-mV steps from the holding potential (D). A in this figure and traces in Fig. 3 were obtained from the same cell.

Ca²⁺ imaging

The averaged intracellular calcium concentration ([Ca²⁺]_i) of single taste buds was increased in response to the application of 1 μm ATP to their basolateral membrane side (Fig. 6). Ca²⁺ responses consisted of a phasic and a following tonic component. Also, Ca²⁺ responses occurred in non-taste lingual epithelial cells including perigemmal cells, cells surrounding TBCs, and epithelial cells lining lingual epithelia among fungiform papillae.

Figure 6. ATP-induced Ca²⁺ imaging of a single taste bud (A–C) and single TBCs (D).

A, time-lapse images. Frame numbers are identical to numerals on B except the frame labelled with 0, a differential interference contrast (DIC) image of a taste bud examined in this figure. Dashed lines, the outline of the taste bud. The outline of typical TBCs (a and b) and non-taste epithelial cells (c and d) are shown with continuous lines. B, averaged fluorescence ratios of the whole taste bud. C, frame 5 overlaid on frame 0 in A. D, fluorescence ratios of labelled cells in A. Horizontal scale, seconds. Vertical scale, a ratio of F₃₄₀/F₃₈₀.

A subset of TBCs, mostly on the peripheral part of taste buds, substantially increased [Ca²⁺]_i upon ATP application. Their response patterns were similar to that of the whole taste bud, though some cells yielded little tonic component. Also, the onset timing of Ca²⁺ responses differed among these cells, even though they were simultaneously exposed to ATP.

Successive applications of ATP decreased the response magnitude, suggesting that responding P2 receptors were desensitized during stimulation (Baryshnikov et al. 2003). In order to reduce the desensitization, we decreased the Ca²⁺ concentration in the physiological saline to 1 mm (lowered-Ca²⁺ saline). Although the lowered-Ca²⁺ saline partially reversed the desensitization, Ca²⁺ responses were still decreased with successive applications, which was remarkable in a few cells (Fig. 7A). In the following Ca²⁺ imaging, we used the lowered-Ca²⁺ saline unless otherwise noted.

Figure 7.

Fluorescence ratios of single TBCs (A–G) and a single non-taste lingual epithelial cell (H) in response to ATP or ATP agonists under indicated conditions.

The threshold concentration of ATP was between 0.01 μm and 0.1 μm, and the magnitude of the Ca²⁺ responses increased with increasing ATP concentration (Fig. 7B). The number of ATP-sensitive TBCs responding within a taste bud increased with increasing ATP concentration, and reached a saturation level of ∼5 cells per taste bud at ATP concentration higher than 100 μm (Table 2).

Table 2.

The numbers of responding taste bud cells in single taste buds

Agonist	Mean ±s.d.	n
ATP (1 μm)	2.0 ± 1.6	30
ATP (100 μm)	4.7 ± 0.7	12
ATP (3 mm)	5.3 ± 2.6	6
2MeSATP (1 μm)	2.1 ± 0.6	10
UTP (100 μm)	0.0 ± 0.0	11
BzATP (30 μm)	2.2 ± 0.6	3
α,β-meATP (100 μm)	0.0 ± 0.0	3

The irrigation of basolateral membranes with the Ca²⁺-nominally-free solution suppressed the tonic component magnitude, and slightly reduced the phasic one (Fig. 7C), suggesting that Ca²⁺ influx from extracellular space causes the tonic component, and that Ca²⁺ released from intracellular stores, and desensitization during stimulation, cause the phasic component.

A variety of ATP agonists also elicited Ca²⁺ responses (Fig. 7D–G and Table 2). That is, the Ca²⁺ response to 1 μm 2-methylthio ATP (2MeSATP), a P2Y₁ and P2Y₁₁ agonist (Ralevic & Burnstock, 1998), and that to 10 μm (data not shown) and 30 μm 3′-O-(4-benzoyl)benzoyl ATP (BzATP), a P2X₇ subtype agonist (Chessell et al. 1998) suggest the functional expression of the respective receptor subtypes. The failure to respond to 100 μm UTP, a P2Y₂ and P2Y₄ subtype agonist (Ralevic & Burnstock, 1998), suggests the absence of these receptor subtypes. Also the failure to respond to 10 and 100 μmα,β-methylene adenosine 5′-triphosphate (α,β-meATP), a P2X agonist except P2X₂, P2X₄ (North, 2002) and P2X₇ (Chessell et al. 1997), suggests the absence of other receptor subtypes.

Non-taste lingual epithelial cells also responded to 1 μm 2MeSATP, and 10 μm and 30 μm BzATP (data not shown). In addition, they responded to 100 μm UTP (Fig. 7H), suggesting that they functionally expressed P2Y₂ or P2Y₄ subtypes in addition to the P2 subtypes of TBCs mentioned above.

RT-PCR

The pharmacological classification of P2 receptor subtypes is difficult because of the lack of specific ligands for respective receptor subtypes. Instead, we classified them, except P2Y₁₁, by taking advantage of RT-PCR on TBCs excised from lingual epithelia with special care in eliminating non-taste lingual epithelial cells (see Methods). RT-PCR reactions with primers for P2 receptor subtypes yielded cDNA fragments of the correct size for 6 out of 14 tested primers sets (Fig. 8 and Table 1). These cDNA fragments showed that expressed P2 subtypes were P2X₂, P2X₄, P2X₇, P2Y₁, P2Y₁₃ and P2Y₁₄. RT-PCR reactions on non-taste lingual epithelial cells yielded cDNA fragments of the correct size for 11 out of 14 tested primers sets; P2X₂, P2X₄, P2X₅, P2X₇, P2Y₁, P2Y₂, P2Y₄, P2Y₆, P2Y₁₂, P2Y₁₃ and P2Y₁₄ (ECs in Fig. 8 and Table 1).

Figure 8. PCR-amplified products from cDNA reverse-transcribed from mRNA isolated from TBCs and non-taste lingual epithelial cells (ECs).

Oligonucleotide primers specific for 14 different P2 receptor subtypes were run in separate reactions and the products run on agarose gel electrophoresis. The expected product size of each reaction is listed below each lane. Oligonucleotide primers specific for β-actin and α-gustducin are similarly examined as positive controls. H₂O used in RT-PCR experiments is also examined as a negative control.

These results suggested that mRNAs for P2X₂, P2X₄, P2X₇, P2Y₁, P2Y₁₃ and P2Y₁₄ subtypes occurred both in TBCs and in non-taste lingual epithelial cells, and that mRNAs for P2X₅, P2Y₂, P2Y₄, P2Y₆ and P2Y₁₂ subtypes occurred only in non-taste lingual epithelial cells. The detection of P2Y₂ and P2Y₄ subtypes on non-taste lingual epithelial cells agreed with their Ca²⁺ responses to UTP. Neither TBCs nor non-taste lingual epithelial cells had mRNAs for P2X₁, P2X₃ and P2X₆ though the primer sets for them yielded cDNA fragments of the correct sizes when tested on mouse brains (see Methods).

Immunohistostaining

In order to confirm the translation of mRNA for P2 receptor subtypes in TBCs, we immunostained their product proteins, except those for P2Y₁, P2Y₁₃ and P2Y₁₄, because P2Y₁ had already been identified in rat TBCs (Kataoka et al. 2004), and the antibodies against P2Y₁₃ and P2Y₁₄ were unavailable. In addition, we examined the dependence of their translation on cell types with cell-type-specific marker antibodies.

It was shown that IP₃R3 (Clapp et al. 2001) and PLCβ2 (Clapp et al. 2004) were type II cell markers, and SNAP-25 and serotonin were type III markers (Yang et al. 2000). However, single-cell RT-PCR showed that 5 TBCs out of 11 expressing PCR products for SNAP-25 expressed those for IP₃R3 in mouse circumvallate papillae (DeFazio et al. 2006). That is, the single-cell RT-PCR study suggests the coexistence of mRNA for IP₃R3 and that for SNAP-25 in the same cells in circumvallate papillae. If this is the case in fungiform papillae, and if fungiform TBCs translate both mRNAs, IP₃R3 would not be a cell-type-specific marker any more. Thus, we tested the overlap of immunoreactivity for these cell-type-specific marker proteins (Fig. 9A) before examining the cell-type-dependent expression of P2 receptor subtypes.

Figure 9. Immunohistochemical confocal images of P2X₂ and cell-type-specific markers.

A, immunoreactivity for PLCβ2 (blue), IP₃R3 (green), SNAP-25 (red), and their overlay. B, P2X₂-immunoreactivity (green) within a taste bud outlined by a dashed circle on each focal plane apart from its taste pore as indicated distance, and overlaid on respective DIC images. Note that no perigemmal cells (cells surrounding the taste bud) are P2X₂ immunoreactive. C, immunoreactivity for P2X₂ (green), SNAP-25 (red), and their overlay. P2X₂ immunoreactivity in the upper row is determined with a Santa Cruz's antibody, and other P2X₂ immunoreactivity in this figure is determined with Alomone's antibody (see text). Both P2X₂ antibodies show that P2X₂ occurs on a type III cell and several SNAP-25-immunoreacitve nerve endings. D, immunoreactivity for P2X₂ (green), serotonin (red), SNAP-25 (blue), and their overlay. White regions, the coexpression of these three antigens; yellow regions, that of P2X₂ and serotonin; violet regions, that of serotonin and SNAP-25. E, immunoreactivity for P2X₂ (green), IP₃R3 (red), and their overlay. F, P2X₂ immunoreactivity with a control antigen-blocking peptide and its DIC image. G, no primary antibody control for P2X₂ and its DIC image. H, P2X₂ immunoreactivity in lingual epithelial cells between filiform papillae, and its overlay on a DIC image. I, uptake of Lucifer Yellow CH (LY) in the presence of 3 mm ATP. Note that there may be ATP-independent LY uptake (see text). Ia, overlay of fluorescence micrograph of LY (green) at the soma level onto the DIC image of the same TBC. Ib, overlay of Ia onto IP₃R3 (red) and SNAP-25 (blue) images. Ic, 3-D reconstruction of the LY-stained cells, a type II cell and a type III cell. The type II cell is stained with LY. Upper white ring shows the maximum outline of the taste bud, and lower one shows that of the taste duct of the taste bud. Another three cells outside the taste bud are stained with LY: two near the upper white ring and one near the lower ring. Id, top view of Ic. Scale bars, 10 μm.

We optically and horizontally sliced each taste bud along their longitudinal axes with laser confocal microscopy, and counted the number of immunoreactive cell bodies. The confocal microscopy showed no overlaps between SNAP-25-immunoreactive cells (107 cells) and IP₃R3-immunoreactive cells (520 cells) in 56 taste buds. We similarly examined 12 taste buds for the overlap of IP₃R3 and PLCβ2 immunoreactivity. The immunoreactivity for both IP₃R3 and PLCβ2 occurred in 154 cells, that for IP₃R3 alone in 3 cells, and that for PLCβ2 alone in 11 cells. These results agreed with previous studies (Yang et al. 2000), and showed that IP₃R3 was a type II cell marker and SNAP-25 a type III cell marker in mouse fungiform TBCs. Type I and IV cells were identified as non-immunoreactive cells.

P2X₂ immunoreactivity occurred within taste buds but not in perigemmal cells (Fig. 9B). Within taste buds, some P2X₂-immunoreactive cells were immunoreactive to SNAP-25 (Fig. 9C) and serotonin (Fig. 9D) but not immunoreactive to IP₃R3 (Fig. 9E). Triple-immuno-histostaining showed that SNAP-25-immunoreactivity outlined serotonin-immunoreactive TBCs (a subset of type III cells), where P2X₂- and serotonin-immunoreactivity coexpressed on the fraction of the cell surface. These results showed that the subset of type III cells express P2X₂.

The antigen peptide for Alomone's P2X₂ antibody suppressed the immunoreactivity (Fig. 9F). Also, controls in which the primary antibody was omitted showed no specific labelling (Fig. 9G). Perigemmal cells were non-immunoreactive to P2X₂ (Fig. 9B), but a few non-taste lingual epithelial cells were immunoreactive to P2X₂ (Fig. 9H).

No immunoreactivity for P2X₄ was observed. P2X₇ immunoreactivity on TBCs was unclear. Although an antibody obtained from Santa Cruz immunostained a few TBCs immunoreactive to IP₃R3, its antigen peptide, the sequence of 264 amino acid residues, was unavailable. In order to perform absorption experiments, we tested another P2X₇ antibody, the antigen peptide for which was available. However, this P2X₇ antibody hardly immunohistostained any TBCs. Therefore, we tested the expression of P2X₇ by monitoring the uptake of Lucifer Yellow CH, as shown in the next section.

Uptake of Lucifer Yellow CH

In the LY solution, a few TBCs showed uptake of Lucifer Yellow CH (LY), a 457 Da fluorescent dye, in response to the application of 3 mm ATP on their basolateral membranes, while other TBCs showed LY uptake independent of ATP stimulation (Figs 9I and 10). The number of LY-stained cells per taste bud was 2.1 ± 0.9 (mean ±s.d., n= 16) including a mean of ATP-independently stained cells of 1.0 ± 0.8 cells per taste bud (n= 8). The difference gives ∼1 cell per taste bud showing ATP-dependent LY uptake. A few non-taste lingual epithelial cells also stained with LY both ATP dependently and ATP independently.

Figure 10. Numbers of LY-stained TBCs per taste bud under different conditions.

The basolateral membranes of TBCs were irrigated with 0 mm or 3 mm ATP for 30 min with or without a P2 blocker, PPADS and a P2X₇-selective blocker, KN-62. Numerals in parentheses show the numbers of taste buds examined, and horizontal bars, s.d.*P < 0.05 (two-tailed t test). **P < 0.003 (one-way ANOVA followed by Scheffé's multiple comparison).

The number of ATP-dependently stained cells per taste bud increased to approximately three in the low divalent cation LY solution on the basolateral membrane side, whereas that of ATP-independently stained cells remained unchanged (Fig. 10). The addition of either 100 μm PPADS, a non-selective blocker against P2 receptors, or 1 μm KN-62, a selective blocker against P2X₇, to the 3 mm ATP stimulating solution decreased the number of stained TBCs, and the number was comparable to that of ATP-independently stained ones. These results suggest that the ATP-dependent uptake results from the activation of P2X₇, showing that a few TBCs, the cell types of which are unidentified, express P2X₇. Also it is shown that ATP-independent uptake did not result from the activation of purinergic receptors by spontaneously released ATP. Both ATP-independently stained TBCs and ATP-dependently stained TBCs excluded Trypan Blue, showing that these TBCs were vital and that their cell membranes were intact. Further study is needed to identify the cell type that expressed P2X₇ and to elucidate the ATP-independent uptake mechanism.

Discussion

The present study showed for the first time that TBCs functionally expressed P2X₂ and P2X₇ in mice. Although RT-PCR studies suggested the expression of P2X₄, P2Y₁, P2Y₁₃ and P2Y₁₄, and though Ca²⁺ responses to 2MeSATP suggested the expression of P2Y₁ or P2Y₁₁, further studies are needed to confirm their expressions.

Electrophysiological studies suggested that ATP increased the open probability of one channel type with a reversal potential of ∼−10 mV, and decreased that of another channel type with a reversal potential of −32 mV (Fig. 5C and D). None of the ions involved in the present experimental conditions had equilibrium potentials equal to these reversal potentials. Thus it is likely that the ATP-opened channel is a non-selective cation channel which agrees with the facts that P2X₂, P2X₄ and P2X₇ open non-selective cation channels. The ATP-closed channel could be a non-selective cation channel as well, but more permeable to K⁺ than the ATP-opened channels.

ATP is hydrolysed into adenosine by endogenous ATPase, and adenosine depolarizes diaphragm muscle fibres (Esau, 1994) and carotid body cells (Xu et al. 2006), or hyperpolarized neural cells (Pan et al. 1995). However, we believe that ATP, not adenosine, produced the present results, because the basolateral membranes were directly exposed to running solutions of either the physiological saline or stimulating ATP solution, which washed away adenosine produced by the hydrolysis and continuously supplied ATP. Also, the present results showed that both ATP and non-hydrolysed ATP agonists elicited Ca²⁺ responses, and that ATP caused LY uptake which was blocked by ATP receptor antagonists. We thus conclude that ATP was a primary agonist for the present results, though we do not exclude the functional expression of receptors for adenosine on TBCs.

In a previous study, we showed with a voltage-sensitive dye that a mixture of taste substances depolarized a group of TBCs and hyperpolarized another group within single taste buds of mouse fungiform papillae (Ohtubo et al. 2001). We suggested that some responses of TBCs were secondary, and that they were responses to transmitters released from TBCs responding to taste substances. If one of the released neurotransmitters was ATP (Finger et al. 2005), the present study suggests that it would activate both the depolarizing and hyperpolarizing groups of TBCs, as we reported in the previous paper.

The threshold concentration of Ca²⁺ responses was between 0.01 μm and 0.1 μm, suggesting that the responsible P2 receptor subtypes have low EC₅₀ values. Our RT-PCR study suggested the expression of P2X₂, P2X₄, P2X₇, P2Y₁, P2Y₁₃ and P2Y₁₄. Published EC₅₀ values for these mouse P2 receptor subtypes include: P2X₂, 123 μm (Zhong et al. 2000); P2X₄, 2.3 μm (Jones et al. 2000); P2X₇, 298 μm (Chessell et al. 1998). Although the EC₅₀ of mouse P2Y₁ is unavailable, that of human P2Y₁ is 304 nm (Palmer et al. 1998). EC₅₀ values for P2Y₁₃ and P2Y₁₄ are unknown in any species. Within this limitation, we assume that P2X₄ and P2Y₁ are responsible for Ca²⁺ responses induced by 1 μm ATP, and that these two, plus P2X₂ and P2X₇ contribute to the generation of the electrophysiological responses to 100 μm ATP.

ATP-evoked Ca²⁺ responses in mouse TBCs were inhibited by the PLC inhibitor U73122, suggesting they were mediated solely by the activation of P2Y receptors (Baryshnikov et al. 2003). However, the concentration of ATP used for the inhibition experiments was 10 μm (Figs 4 and 5 of their paper), which was much lower than the EC₅₀ of P2X₂, 123 μm (Zhong et al. 2000). Therefore, there were no conflict between the previous result and our present result on the expression of P2X₂. On the other hand, the concentration of 10 μm was close to the EC₅₀ of P2X₄, 2.3 μm (Jones et al. 2000). Our present result that P2X₄ immunoreactivity was undetectable in taste buds agrees with the previous result (Baryshnikov et al. 2003), suggesting that P2X₄ is less important in TBCs.

ATP caused LY uptake in a few TBCs per taste bud, and PPADS and KN-62 blocked the uptake. P2X₇ is shown to form large pore channels in response to prolonged exposure to ATP, and these large pore channels are permeable to LY (North, 2002; Faria et al. 2005). In addition, it was shown that the ATP stimulation in solutions containing low concentrations of extracellular divalent cations created a much larger aqueous pore (Chessell et al. 1997). These results agree with the present results in that the number of TBCs stained with LY increased in the low divalent cation solution. ATP-independent uptake of LY may be caused by carrier-mediated mechanisms as shown in skates and rats (Ballatori et al. 1999).

P2X₇ is involved in apoptosis of various cells (Wang et al. 2004; Bulanova et al. 2005). Mammalian TBCs are short-lived and a rapid turnover of TBCs occurs in each taste bud. For example, the lifetime of rat TBCs was ∼9 days (Beidler & Smallman, 1965) though a subset of type II cells expressing α-gustducin was shown to live longer (Farbman, 1980; Cho et al. 1998). It is suggested that old TBCs of mice employ p53, a tumour-suppressor protein linked to apoptosis, Bax, a bcl-2 family death factor, and a protease, caspase-2, as part of their apoptotic pathway (Zeng & Oakley, 1999; Zeng et al. 2000). In cultured mesangial cells of the renal glomerulus, the activation of P2X₇ in response to ATP increased p53 in protein levels and caused apoptotic DNA fragmentation (Schulze-Lohoff et al. 1998). Therefore, we assume that P2X₇ may contribute to the renewal of TBCs.

The present RT-PCR and Ca²⁺-imaging experiments showed no expression of P2Y₂ nor P2Y₄ receptor subtypes on TBCs in fungiform papillae, but they were expressed on non-taste lingual epithelial cells among fungiform papillae. In circumvallate and foliate papillae, on the other hand, immunohistostaining showed that isolated TBCs and isolated taste buds expressed P2Y₂ and P2Y₄ (Bystrova et al. 2006). This discrepancy may result from the differences among taste papillae and their innervation; the chorda tympani, a branch of cranial nerve VII, innervates TBCs in fungiform papillae and some TBCs in foliate papillae, while the glossopharyngeal nerve, cranial nerve IX, innervates the remainder of TBCs in foliate papillae and all TBCs in circumvallate papillae. The glossopharyngeal nerve may affect the translation of P2Y₂ and P2Y₄. Also it is equally possible that the epithelium in the two regions of the tongue is specified differently during development.

The present results suggest that these P2 receptor subtypes link the constituents of taste buds. It was shown that taste nerve fibres expressed P2X₂ and P2X₃ (Bo et al. 1999) and that taste stimulation released ATP from TBCs (Finger et al. 2005). Recent studies showed that type II cells released ATP through voltage-gated pannexin hemichannels in response to a variety of taste substances (Huang et al. 2007; Romanov et al. 2007) and released ATP caused serotonin release from other TBCs, probably type III cells (Huang et al. 2007). It is thus likely that P2X₂ found on the subset of type III cells in the present study mediates the signal transduction between TBCs. Also, P2Y₁ functionally expressed on rat type II and type III cells (Kataoka et al. 2004) may contribute to the signal transduction.

Therefore, we assume the following hypothetical roles of P2 subtypes within taste buds (Fig. 11). Type II cells release ATP as a paracrine signal (white arrows) in response to taste stimulation (arrowhead), and that released ATP stimulates P2X₂, P2X₃, or their heteromultimer on taste nerve fibres, and P2X₂, P2X₇, or P2Y₁ (shaded boxes) on type III cells. Released serotonin stimulates serotonin receptors (triangle) on taste nerve endings. Self-(circular arrow) or paracrine-stimulation of P2X₇ (shaded box) on TBCs triggers the apoptotic pathway, though intensive ATP release is essential for the high EC₅₀ of P2X₇. Also, this self- or paracrine-stimulation of P2X₂ or P2Y₁ on TBCs contributes to TBC–TBC interactions.

Figure 11. Hypothetical TBC networks linked with P2 receptor subtypes.

Taste stimulation (arrowhead) releases ATP from type II cells as a paracrine signal (white arrows). Released ATP stimulates P2 receptor subtypes (shaded boxes) on type III cells or taste nerve fibres. Stimulated type III cells release serotonin (5-HT), which activates serotonin receptors (triangle) on taste nerve endings. Self-stimulation (circular arrow) or paracrine stimulation of P2X₇ leads the type II and type III cells to apoptosis. Note that the present results showed P2X₂ was expressed on type III but not on type II cells. The cell type that expresses P2X₇ remains to be investigated, though it is shown in type II and III cells.

When P2X₄, P2Y₁₃ and P2Y₁₄ (not shown in the figure) were functionally expressed on TBCs, they would contribute to these TBC–TBC interactions as well as other P2 subtypes expressed on TBCs. The role of P2 receptor subtypes on non-immunoreactive TBCs, type I or IV cells, may be to regulate their supportive functions.