Regulation of the Putative TRPV1t Salt Taste Receptor by Phosphatidylinositol 4, 5-Bisphosphate

Abstract

Regulation of the putative amiloride and benzamil (Bz)-insensitive TRPV1t salt taste receptor by phosphatidylinositol 4,5-bisphosphate (PIP₂) was studied by monitoring chorda tympani (CT) taste nerve responses to 0.1 M NaCl solutions containing Bz (5 × 10⁻⁶ M; a specific ENaC blocker) and resiniferatoxin (RTX; 0–10 × 10⁻⁶ M; a specific TRPV1 agonist) in Sprague-Dawley rats and in wildtype (WT) and TRPV1 knockout (KO) mice. In rats and WT mice, RTX elicited a biphasic effect on the NaCl + Bz CT response, increasing the CT response between 0.25 × 10⁻⁶ and 1 × 10⁻⁶ M. At concentrations >1 × 10⁻⁶ M, RTX inhibited the CT response. An increase in PIP₂ by topical lingual application of U73122 (a phospholipase C blocker) or diC8-PIP₂ (a short chain synthetic PIP₂) inhibited the control NaCl + Bz CT response and decreased its sensitivity to RTX. A decrease in PIP₂ by topical lingual application of phenylarsine oxide (a phosphoinositide 4 kinase blocker) enhanced the control NaCl + Bz CT response, increased its sensitivity to RTX stimulation, and inhibited the desensitization of the CT response at RTX concentrations >1 × 10⁻⁶ M. The ENaC-dependent NaCl CT responses were not altered by changes in PIP₂. An increase in PIP₂ enhanced CT responses to sweet (0.3 M sucrose) and bitter (0.01 M quinine) stimuli. RTX produced the same increase in the Bz-insensitive Na⁺response when present in salt solutions containing 0.1 M NaCl + Bz, 0.1 M monosodium glutamate + Bz, 0.1 M NaCl + Bz + 0.005 M SC45647, or 0.1 M NaCl + Bz + 0.01 M quinine. No effect of RTX was observed on CT responses in WT mice and rats in the presence of the TRPV1 blocker N-(3-methoxyphenyl)-4-chlorocinnamide (1 × 10⁻⁶ M) or in TRPV1 KO mice. We conclude that PIP₂ is a common intracellular effector for sweet, bitter, umami, and TRPV1t-dependent salt taste, although in the last case, PIP₂ seems to directly regulate the taste receptor protein itself, i.e., the TRPV1 ion channel or its taste receptor variant, TRPV1t.

INTRODUCTION

In rodents, chorda tympani (CT) taste nerve responses to NaCl are derived from at least two salt taste receptor/ion channels in fungiform taste receptor cells (TRCs). Apical Na⁺influx through the Na⁺-specific benzamil (Bz)-sensitive epithelial Na⁺channel (ENaC) accounts for 60–70% of the CT response to NaCl. The second receptor/ion channel is Bz-insensitive. It does not discriminate between Na⁺, K⁺, NH₄⁺, and Ca²⁺. Several studies suggest that a nonspecific cation channel TRPV1t, a variant of the pain receptor (transient receptor potential vanilloid-1; TRPV1) accounts for the remaining 30–40% of the Bz-insensitive NaCl CT response in rats and wildtype (WT) mice (Lyall et al. 2004, 2005a, 2009). The TRPV1t contribution to the NaCl CT response is increased by 100% or more by TRPV1t agonists: resiniferatoxin (RTX), capsaicin, alcohol, nicotine, cetylpyridinium chloride, naturally occurring Maillard peptides, Maillard reacted peptides conjugated with various sugar moieties, and by elevated temperature (DeSimone and Lyall 2006, 2008; Lyall et al. 2009). Some of the agonists that modulate the TRPV1t-dependent NaCl CT response in rodents also modulate human salt taste (Dewis et al. 2006; Katsumata et al. 2008; Rhyu et al. 2006). In humans, salt taste is predominantly amiloride-insensitive (Feldman et al. 2003; Ossebaard and Smith 1995). This suggests that TRPV1t may also play a role in human salt taste perception.

All of the TRPV1t modulators tested to date produce enhancement of the TRPV1t-dependent NaCl CT response at low concentrations but inhibit it at higher concentrations. The biphasic CT response profiles for the above modulators are observed over a wide range of agonist concentrations (Katsumata et al. 2008; Lyall et al. 2004, 2005a,b, 2007, 2009). At present, the intracellular effectors through which the above agonists modulate the sensitization and desensitization of TRPV1t and produce changes in the TRPV1t-dependent NaCl CT response are not well understood. We have recently shown that the sensitization and desensitization of the TRPV1t-dependent NaCl CT response to RTX is modulated by changes in TRC Ca²⁺([Ca²⁺]_i) and by altering the phosphorylation state of TRPV1t by protein kinase C epsilon (PKCε), a Ca²⁺-insensitive isoform of PKC or calcineurin, a Ca²⁺-dependent serine-threonine phosphatase (PP2B) (Lyall et al. 2009). Changes in [Ca²⁺]_i, PKCε and calcineurin also modulate the activity of native TRPV1 in dorsal root ganglion neurons and the cloned TRPV1 expressed in heterologous cells (Petrocellis and Marzo 2005). In addition to the above modulators, changes in membrane phosphatidylinositol 4, 5-bisphosphate (PIP₂) levels modulate TRPV1 activity (Chuang et al. 2001; Nilius et al. 2007; Prescott and Julius 2003). Accordingly, we hypothesize that PIP₂ also modulates the TRPV1t-dependent NaCl CT responses by regulating TRPV1t activity in TRCs.

In a mammalian cell, PIP₂ represents ∼1% of the total acidic membrane lipids and >99% of the doubly phosphorylated phosphoinositides (Golebiewska et al. 2006). PIP₂ has been shown to regulate a variety of ion channels. These include inwardly rectifying and voltage-gated K⁺channels, the two-pore (2-P) domain K⁺channels, voltage-gated Ca²⁺channels, cyclic nucleotide-gated channels, intracellular Ca²⁺release channels, Cl⁻ channels, TRP channels, and ENaC (Delmas et al. 2005; McLaughlin and Murray 2005; Nilius et al. 2007; Suh and Hille 2005; Yue et al. 2002). Among TRP channels, PIP₂ tonically inhibits only TRPV1, and the channel is released from inhibition by PIP₂ hydrolysis (Chuang et al. 2001; Nilius et al. 2007; Prescott and Julius 2003). All of the other TRP channels, namely TRPM4, TRPM5, TRPM7, TRPM8, and TRPV5, are activated by PIP₂ (Nilius et al. 2007).

Here, we studied the regulation of the native TRPV1t in TRCs by PIP₂. We monitored NaCl CT responses in the presence of Bz in anesthetized Sprague-Dawley rats and in wild-type (WT) and TRPV1 knockout (KO) mice in vivo. CT responses were monitored in the absence and presence of RTX, a specific TRPV1t modulator (Lyall et al. 2004, 2005a, 2009), whereas PIP₂ levels were either enhanced or diminished in TRC membranes in vivo. Changes in PIP₂ were induced by the topical lingual application of U73122, a nonspecific blocker of phospholipase Cs (PLCs) or phenylarsine oxide (PAO), a blocker of phosphoinositide (PI) 4 kinases. In addition, we increased PIP₂ in TRC membranes directly by the topical lingual application of diC8-PIP₂, a short chain synthetic PIP₂. We also studied the effects of PAO, U73122, and diC8-PIP₂ on CT responses to representative sweet (sucrose), bitter (quinine), and umami [monosodium glutamate (MSG) and MSG + IMP (inosine 5′-monophosphate)] stimuli and on the ENaC-dependent component of the NaCl CT response. The results summarized here suggest that both the open-state activity and the RTX-induced sensitization and desensitization of TRPV1t are modulated by PIP₂. An increase in PIP₂ inhibited TRPV1t activity and desensitized the channel toward activation by RTX. In contrast, a decrease in PIP₂ enhanced TRPV1t activity and sensitized the channel to further stimulation by RTX. Unlike TRPV1t, ENaC-dependent NaCl CT responses showed no sensitivity to changes in PIP₂. On the other hand, an increase in PIP₂ enhanced CT responses to sweet and bitter stimuli. We conclude that PIP₂ is a common intracellular effector that differentially regulates different taste qualities.

METHODS

CT taste nerve recordings

We generated concentration–response relationships between RTX concentration and the magnitude of the TRPV1t-dependent NaCl CT response in Sprague-Dawley rats, WT mice, and TRPV1 KO mice under various experimental conditions that are expected to alter membrane PIP₂ levels in TRCs. Animals were housed in the Virginia Commonwealth University animal facility in accordance with institutional guidelines. All animal protocols were approved by the Institutional Animal Care and Use Committee of Virginia Commonwealth University. Female Sprague-Dawley rats (150–200 g) were anesthetized by intraperitoneal injection of pentobarbital sodium (60 mg/kg), and supplemental pentobarbital sodium (20 mg/kg) was administered as necessary to maintain surgical anesthesia. The animal's corneal reflex and toe-pinch reflex were used to monitor the depth of surgical anesthesia. Body temperatures were maintained at 37°C with a Deltaphase Isothermal PAD (Model 39 DP, Braintree Scientific, Braintree, MA). The left CT taste nerve was exposed laterally as it exited the tympanic bulla, severed, desheathed, and wrapped around a 32-gauge platinum/iridium wire electrode. Stimulus solutions maintained at room temperature were injected into a Lucite chamber affixed by vacuum to a 28-mm² patch of anterior dorsal lingual surface (Ye et al. 1993). For stimulation or rinsing, 3-ml aliquots were injected at a rate of 1 ml/s into the perfusion chamber. The CT responses were recorded under zero lingual current-clamp and analyzed as described previously (Katsumata et al. 2008; Lyall et al. 2009).

WT mice (C57BL/6J) and homozygous TRPV1 KO mice (30–40 g; B6. 129S4-Trpv1^tmijul, The Jackson Laboratory, Bar Harbor, ME) were anesthetized by intraperitoneal injection of pentobarbital sodium (30 mg/kg), and supplemental pentobarbital sodium (10 mg/kg) was administered as necessary to maintain surgical anesthesia. The rest of the procedure was the same as in rats. At the end of each experiment, animals were killed by an intraperitoneal overdose of pentobarbital sodium (∼195 mg/kg body weight for rats and 150 mg/kg weight for mice).

The composition of the various stimulating solutions used in the CT experiments is shown in Table 1. The various drugs and their concentrations used in this study and their specific targets are listed in Table 2. The anterior lingual surface was stimulated with a rinse solution (R; pH 6) and with a salt solution (pH 6) containing 0 to 10 × 10⁻⁶ M RTX. RTX specifically modulates the TRPV1t-dependent NaCl CT response. In previous studies, the RTX-induced increase in the TRPV1t-dependent NaCl CT response varied with pH. The relationship between pH and the magnitude of the CT response was bell shaped. The maximum increase in the CT response was observed around pH 6 (Lyall et al. 2004). RTX versus TRPV1t-dependent NaCl CT concentration–response curves were obtained before and after treating the anterior tongue with various drugs listed in Table 2. In our previous studies (DeSimone et al. 2001), in the presence of the cetylpyridinium chloride (a TRPV1t agonist weaker than RTX) concentration that produced a maximum enhancement in the TRPV1t-dependent NaCl CT response, the CT response was a saturating function of NaCl concentration with a K_m of 0.185 ± 0.035 M. In this study, all experiments were done at 0.1 M NaCl to enable us to observe both maximum enhancement and inhibition of the TRPV1t-dependent NaCl CT response in the presence of agonists and antagonists, respectively.

Table 1.

Composition of stimulating solutions for CT experiments

Solution	Composition	pH
Rinse (R)	0.01 M KCI + 0.2 M mannitol + 0.01 M HEPES	6
Salt stimuli (N)	0.01 M KCl + 0.1 M NaCl + 0.01 M HEPES	6
R + RTX	R + 0.1 × 10−6 M to 10 × 10−6 M RTX	6
N + RTX	N + 0.1 × 10−6 M to 10 × 10−6 M RTX	6
Control-1	0.3 M NH4Cl
Control-2	0.3 M NaCl
Sucrose	0.3 or 0.5 M
KCl	0–0.5 M
SC45647	0.003 or 0.005 M
Quinine	0.01 M
HCl	0.03 M
R + mannitol	R + 0.210 M
R + mannitol + SC45647	R + 0.205 M + 0.005
N + SC45647	N + 0.005 + 0.005 M
N + SC45647 + RTX	N + 0.005 M + 1 × 10−6 M
N + quinine	N + 0.01 M
N + quinine + RTX	N + 0.01 M + 1 × 10−6 M
MSG + mannitol	0.1 M + 0.01 M
MSG + mannitol + RTX	0.1 M + 0.01 M + 1 × 10−6 M

CT, chorda tympani; RTX, resiniferatoxin.

Table 2.

Summary of the effect of various drugs that modulate PIP₂ levels in TRCs and alter the Bz-insensitive NaCl CT response

Drug	M	Intracellular Target	Number of Animals (N)
Bz	5 × 10−6	↓ ENaC	12
SB-366791	1 × 10−6	↓ TRPV1t	6
U73122	10 × 10−6 to 250 × 10−6	↓ PLC and ↑ PIP2	3
U73343	250 × 10−6	↔ PLC	3
PAO	5 × 10−6 to 50 × 10−6	↓ PI4 kinases and ↓ PIP2	8
diC8-PIP2	115 × 10−6 or 250 × 10−6	↑ PIP2	12
MSG + IMP	0.1 M + 1 × 10−3	↑ T1R1 + T1R3	3
↑ GPCRs (T2Rs, T1R1 + T1R3, T1R2 + T1R3)		↑ PLCβ2 and ↑ TRPM5	3

The GPCR associated with sweet taste (T1R2 + T1R3) was activated by adding sucrose or SC45647 to the salt mixture, the GPCR associated with umami taste (T1R1 + T1R3) was activated measuring responses to MSG and MSG + IMP, and GPCR associated with bitter taste (T2R) was activated by adding quinine to the salt mixture. Benzamil and SB were added to the salt stimuli and produced their effects on the Bz-insensitive NaCl CT responses immediately. However, all other drugs were dissolved directly in 3 ml of dimethyl sulfoxide (DMSO) and were topically applied to the tongue for at least 30 min. In some experiments, a stock solution of diC8-PIP₂ in DMSO (0.584 × 10⁻³ M) was added to 0.525 M sucrose or 0.00525 M SC45647 solution to yield a stimulating solution containing 250 × 10⁻⁶ M diC8-PIP₂ + 0.3 M sucrose or 250 × 10⁻⁶ M diC8-PIP₂ + 0.003M SC45647. Equivalent amount of DMSO was added to the rinse solution. DMSO by itself has no effect on the CT responses to salt stimuli (Lyall et al. 1999). Unless stated otherwise, all drugs were purchased from Sigma. ↑, increase; ↓, decrease; ↔, no change; Bz, benzamil (blocks Na⁺ entry through apical epithelial Na⁺ channel; ENaC); SB-366791, N-(3-methoxyphenyl)-4-chlorocinnamide (blocks Na⁺ entry through TRPV1t); U73122, 1-[6-[((17β)-3-methoxyestra-1,3,5[10]-trien-17-y)amino]hexyl]-1H-pyrrole-2,5-dione (a nonspecific blocker of phospholipase Cs (PLCs)) and blocks the hydrolysis of PIP₂ to inositol-1,4,5-trisphosphate (IP₃) + diacylglycerol (DAG) by PLC (Chuang et al. 2001; Liu et al. 2005; Prescott and Julius 2003); U73343, 1-[6-[((17β)-3-methoxyestra-1,3,5[10]-trien-17-yl)amino]hexyl]-2,5-pyrrolidine-dione (an inactive analog of U73122); PAO, phenylarsine oxide (a trivalent arsenic, reacts with vicinal thiol groups in proteins inhibits all isoforms of PI4 kinases and blocks the resynthesis of PIP₂ from PI (Liu and Quin 2005)); diC8-PIP₂, dioctanoyl-PIP₂ (Echelon, Salt Lake City, UT); a short-chain synthetic PIP₂ that increases PIP₂ levels in cell membranes (Liu et al. 2005); IMP, inosine-5′-monophosphate (specifically enhances CT responses to MSG); GPCRs, G-protein–coupled receptors.

Typically, stimulus solutions remained on the tongue for 1–2 min. The volume of fluid in the chamber at any time is ∼0.03 ml (Ye et al. 1993). With a 3-ml rinse solution, this volume is displaced in <1 s.

Control stimuli consisting of 0.3 M NH₄Cl and 0.3 M NaCl (Table 1) applied at the beginning and at the end of experiment were used to assess preparation stability (Fig. 4B and Supplementary Fig. S2).¹ The preparation was considered stable only if the difference between the magnitude of the responses to control stimuli at the beginning and at the end of the experiment was <10% (Lyall et al. 2009). The following stimulus series were used in the CT experiments: R → 0.3 M NH₄Cl → R → 0.3 M NaCl → R → N → R → (N + Bz) → R → (N + Bz + RTX) → R. The R → (N + Bz + RTX) → R step was repeated for each RTX concentration between 0.1 × 10⁻⁶ M and 10 × 10⁻⁶ M (Fig. 1 and Supplementary Fig. S1). At the end of the RTX concentration series, the control stimuli were again applied (R → 0.3 M NH₄Cl → R→ 0.3 M NaCl → R).

Fig. 4.

Effect of inhibiting TRPV1t activity on CT responses to umami taste stimuli. A: representative CT recording in a TRPV1 KO mouse in which the tongue was 1st rinsed with a rinse solution (R) and then with N, N + Bz, monosodium glutamate (MSG), MSG + Bz, MSG + Bz + inosine 5′-monophosphate (IMP), and R + IMP. The arrows indicate the time period when the tongue was stimulated with different solutions. B: representative CT recording in which the rat tongue was 1st rinsed with a rinse solution, R, and then with N, N + Bz, N + Bz + SB, MSG, MSG + Bz, and MSG + Bz + SB (Table 1). The control responses to 0.3 M NH₄Cl are shown at the beginning and end of the experiment and were nearly identical. The arrows indicate the time period when the tongue was stimulated with different solutions. C: the 1st part of the figure shows the summary of the effects of RTX on the CT responses in the presence of N + Bz and MSG + Bz. The 2nd part of the figure shows the summary of the effects of IMP on the CT responses in the presence of MSG + Bz and MSG + Bz + SB. No statistical difference (P > 0.05; paired) was observed between the RTX-induced increase in the CT response to N + Bz [δ RTX_{N + Bz} = (N + Bz + RTX) − (N + Bz)] or MSG + Bz [δ RTX_{MSG + Bz} = (MSG + Bz + RTX) − (MSG + Bz)]. No statistical difference (P > 0.05; paired) was observed between IMP-induced increase in the CT response to MSG + Bz [δ IMP_{MSG + Bz} = (MSG + Bz + IMP) − (MSG + Bz)] or MSG + Bz + SB [δ IMP_{MSG + Bz + SB} = (MSG + Bz + SB + IMP) − (MSG + Bz + SB)].

Fig. 1.

Effects of phenylarsine oxide (PAO) and phosphatidylinositol 4,5-bisphosphate (PIP₂) on N + benzamil (Bz) chorda tympani (CT) responses in the absence and presence of resiniferatoxin (RTX). Figure shows 2 representative CT recordings in which the rat tongue was 1st rinsed with a rinse solution, R (Table 1), and then with N + Bz solutions containing 0–10 × 10⁻⁶ M RTX (Table 1) before (Control) and after topical lingual application of (A) 15 × 10⁻⁶ M PAO (Post-PAO) or (B) 250 × 10⁻⁶ M diC8-PIP₂ (Post-PIP₂). The arrows indicate the time period when the tongue was stimulated with different solutions.

The data were digitized and analyzed off-line. In CT experiments, both transient (phasic) and tonic (steady-state) parts of the NaCl CT response were quantified. We also quantified the transient (phasic) response to the application of rinse (R) to a tongue already superfused with R. To quantify the phasic part of the CT response, the height of the stimulus-induced maximum CT response relative to baseline response was divided by the mean steady-state (tonic) response to 0.3 M NH₄Cl. To quantify the tonic (steady-state) part of a response, the area under the response versus time curve was taken over the final 30 s of the response. To normalize, this area was divided by the area under the 0.3 M NH₄Cl response curve over the final 30 s of the tonic response period. The normalized data were reported as means ± SE of the number of animals. In some figures, raw CT recordings are shown in the same animal before and after topical lingual application of a drug with an arbitrary Y scale that reflects the output of the integrator.

For clarity, the points on the graphs of the mean normalized tonic responses versus the logarithm of the RTX concentration were connected by smooth curves. The curves were generated using a fitting function that models the characteristic biphasic property of the RTX concentration versus the magnitude of the CT response. The biphasic property has been observed with every agonist of the TRPV1t-dependent NaCl CT response thus far examined (Katsumata et al. 2008; Lyall et al. 2004, 2005b, 2007, 2009). The fitting function used was

$$R=\frac{r+bh\left(x\right)}{1+h\left(x\right)+j\left(x\right)}$$ (1)

where

$$h\left(x\right)={10}^{n\left(x-a\right)}$$ (2)

and

$$j\left(x\right)={10}^{n\left(x-a\right)+m\left(x-d\right)}$$ (3)

Here R is the response, x is the logarithm of the RTX concentration expressed in moles per liter, and a, b, d, m, n, and r are parameters chosen by least squares criteria (Katsumata et al. 2008; Lyall et al. 2009).

Detection of PLCβ₁, PLCβ₂, PLCβ₃, and PLCγ₁ in anterior and posterior taste fields

To confirm that the effects of U73122 on TRPV1t-dependent NaCl CT responses are caused by its effect on PLCs, the PCR technique was used to detect the presence of several PLC isoforms (PLCβ₁,PLCβ₂, PLCβ₃, and PLCγ₁) in TRCs. Anterior lingual epithelium containing fungiform papillae and lingual epithelium from circumvallate papilla were obtained after collagenase treatment (Lyall et al. 2004). RNA was prepared using the RNeasy Protect Kit (Qiagen, Valencia, CA), and the cDNA was generated using the M-MLV Reverse Transcriptase Kit (Invitrogen, Carlsbad, CA) according to the manufacturers' protocols. PCR screening of the fungiform and circumvallate cDNAs for the presence of rat PLCβ₁,PLCβ₂, PLCβ₃, and PLCγ₁ was performed with Taq DNA Poymerase (Roche, Indianapolis, IN) using the primer pairs shown in Table 3. The reaction annealing temperatures and PCR product sizes are also listed in Table 3. As negative controls, the reactions were run with water replacing the cDNA template.

Table 3.

Primer pairs

Intracellular Effector (GenBank No.)	Direction	Sequence	Product Size (Annealing Temperature)
PLCβ1 (NM_001077641)	Forward	CACCTCTGAAGCAAGAGCAGGTC	360 bp (55.5°C)
	Reverse	GTTGTCATGGTGAATCCATGGG
PLCβ2 (NM_053478)	Forward	CTCGCTTTGGGAAGTTTGC	526 bp (55.5°C)
	Reverse	GCATTGACTGTCATCGGGT
PLCβ3 (NM_033350)	Forward	GACATGTTTGGTCTCCCAGTTGAC	306 bp (55.5°C)
	Reverse	GGTCATCTGGGATGTAGTCAGAAGC
PLCγ1 (NM_013187)	Forward	CCAAGGACCTGAAGAACATGCTG	627 bp (57.5°C)
	Reverse	GTGTGCCCATGGTAAATGACTGG

PCR reaction products were separated using 1.2% agarose gel at 100 V applied potential for 2.5 h.

Measurement of PIP₂ in isolated anterior lingual epithelium using thin-layer chromatography

Rat lingual epithelium containing fungiform papillae was isolated as described before (Lyall et al. 2004). Isolated lingual epithelia from eight rats were pooled and labeled for 4 h with ³²P. After chemical treatment with PAO (10 × 10⁻⁶ M or 20 × 10⁻⁶ M), U73122 (50 × 10⁻⁶ M or 100 × 10⁻⁶ M), or U73343 (100 × 10⁻⁶ M) for 10 min, the reaction was stopped by the addition of 1.8 ml chloroform-methanol-HCl (100:200:2, vol/vol/vol). The phases were separated by addition of 0.6 ml of chloroform and 0.6 ml of 2 M KC1. The lower organic phase was collected, and the aqueous phase was washed again with 2 ml of chloroform to yield a residual organic phase. The combined organic phases were dried under nitrogen; resuspended in 50 μl of chloroform-methanol (9:l) containing 15 μg of unlabeled PIP₂; and applied to silica gel-H plates impregnated with 1% potassium oxalate for thin layer chromatography. The lipids were separated using chloroform-methanol-4 N NH₄OH (45:35:10, vol/vol/vol), air-dried, and sprayed with 1,2-dichlorofluorescein (0.1%) in isopropyl alcohol. The lipids were visualized under ultraviolet light at 357 nm. The spots corresponding to PIP₂ were scraped and counted in a liquid scintillation counter (Murthy and Makhlouf 1991).

Statistical methods

In CT studies, Student's t-test was used to analyze the differences between sets of data. Because we are comparing the normalized CT responses before and after the topical application of TRPV1t modulators in the same CT preparation, a paired t-test was used to evaluate statistical significance. In experiments in which repeated measurements to sequential drug applications were obtained, the data were analyzed using two-way ANOVA and all statistical significant values were Bonferroni-adjusted (Lyall et al. 2009). For all other comparisons, Student's t-test was used to analyze the differences between sets of data.

RESULTS

Regulation of TRPV1t activity by PIP₂

To study whether TRPV1t activity is modulated by PIP₂, RTX versus N + Bz CT concentration–response relations were obtained in separate groups of rats before and after topical lingual application of PAO, U73122, U73343, or diC8-PIP₂, drugs that are expected to alter PIP₂ levels in cell membranes (Table 2).

Effects of PAO and RTX on TRPV1t activity

Consistent with previous studies (Lyall et al. 2004, 2005a, 2009), RTX initially produced an increase in the N + Bz CT response between 0.25 × 10⁻⁶ M and 1 × 10⁻⁶ M (Fig. 1A, Control). Above 1 × 10⁻⁶ M RTX, the CT response was diminished. In four rats, a bell-shaped concentration–response relationship was observed between RTX concentration and the magnitude of the tonic N + Bz CT response (Fig. 2A, ○). The maximum increase in the CT response occurred at 1 × 10⁻⁶ M RTX. At 10 × 10⁻⁶ M RTX, the CT response was inhibited close to the rinse baseline. RTX also enhanced the phasic N + Bz CT response in a biphasic manner. Maximum enhancement and inhibition in the phasic response was also observed at 1 × 10⁻⁶ M and 10 × 10⁻⁶ M RTX, respectively (Supplementary Fig. S1D). These results show that both phasic and tonic components produce similar RTX concentration versus N + Bz CT response relations. The phasic CT response has two components. One component is derived from the transient mechanical rinse artifact that is observed every time the lingual surface is superfused with a rinse or test solution (Supplementary Fig. S1B). The second component constitutes the chemical response of the nerve to a taste stimuli applied to the tongue. The phasic response is sensitive to the rate at which the taste stimulus is superfused in the lingual chamber (Lyall et al. 2001). In contrast, tonic component of the CT response is stable and is not affected by the rate of superfusion and the mechanical rinse artifact. Accordingly, in this study, most of the data analysis was performed on the tonic component of the CT response.

Fig. 2.

Summary of the effects of PAO, PIP₂, U73122, and U73343 on N + Bz CT responses in the absence and presence of RTX. A: in 4 animals, normalized tonic N + Bz + RTX (0–10 × 10⁻⁶ M) CT responses were compared before and after PAO treatment. Significant difference was found between RTX interactions with the N + Bz CT response post-PAO treatment relative to control (P < 0.0002 for both; 2-way ANOVA). B: in 5 animals, normalized tonic N + Bz + RTX (0–10 × 10⁻⁶ M) CT responses to RTX were compared before and after diC8-PIP₂ treatment. Significant difference was found between RTX interactions with the N + Bz CT response post-PIP₂ treatment relative to control (P = 0.0025 for both; 2-way ANOVA). C: in 3 animals, normalized tonic N + Bz + RTX (0–10 × 10⁻⁶ M) CT responses were compared before and after U73122 treatment. Significant difference was found between RTX interactions with the N + Bz CT response post-U73122 treatment relative to control (P = 0.0054 for both; 2-way ANOVA). D: in 3 animals, normalized tonic N + Bz + RTX (0–10 × 10⁻⁶ M) CT responses were compared before and after U73343 treatment. No significant differences were found for RTX interactions with the N + Bz CT response post-U73343 treatment relative to control concentration (P > 0.05 for both, 2-way ANOVA).

CT responses were also monitored after topical lingual application of PAO at 5 × 10⁻⁶, 15 × 10⁻⁶, or 50 × 10⁻⁶ M. At 15 × 10⁻⁶ M, PAO enhanced the magnitude of the control N + Bz CT response and the CT response in the presence of RTX (Fig. 1A, Post-PAO). In a separate set of rats, 5 × 10⁻⁶ M PAO induced no change in the N + Bz CT responses in the absence or presence of RTX relative to control (data not shown). Increasing the PAO concentration to 50 × 10⁻⁶ M produced qualitatively similar effects on the CT response as observed with 15 × 10⁻⁶ M PAO (data not shown). Because the magnitudes of the CT responses were not different at 15 × 10⁻⁶ M and 50 × 10⁻⁶ M PAO, the data were pooled for analysis. In four animals, PAO enhanced the magnitude of the tonic N + Bz CT response in the absence and presence of RTX (Fig. 2A, ●). No attenuation in the CT response was observed between 1 × 10⁻⁶ and 10 × 10⁻⁶ M RTX relative to control. These results suggest that blocking the resynthesis of PIP₂ from PI by PAO increased the open-state activity of native TRPV1t and sensitized it to further stimulation with RTX. It also eliminated the desensitization of the channel at RTX concentrations >1 × 10⁻⁶ M.

Effect of PIP₂ and RTX on TRPV1t activity

PIP₂ levels in TRC membranes were increased by the topical lingual application of either 115 × 10⁻⁶ M or 250 × 10⁻⁶ M diC8-PIP₂. Under control conditions (Fig. 1B, Control), RTX produced a biphasic effect on the N + Bz CT response as before. No significant effects were observed on the CT responses to N, n + Bz, or n + Bz + RTX relative to control with 115 × 10⁻⁶ M diC8-PIP₂ treatment (data not shown). After 250 × 10⁻⁶ M diC8-PIP₂ treatment, the N + Bz CT response was decreased in the absence of RTX to near baseline. In addition, diC8-PIP₂ attenuated the effect of RTX on the N + Bz CT response (Fig. 1B, Post-PIP₂) relative to control. In five rats, diC8-PIP₂ inhibited the basal level of the tonic N + Bz CT response and attenuated the magnitude of the neural response in the presence of RTX (Fig. 2B).

Effects of U73122, U73343, and RTX on TRPV1t activity

U73122, a nonspecific inhibitor of PLCs, blocks the hydrolysis of PIP₂ to IP₃ + DAG and is expected to increase PIP₂ levels in cell membrane (Pochynyuk et al. 2008). In three animals, under control conditions (Fig. 2C, ○) RTX produced a biphasic effect on the tonic N + Bz CT responses as before. CT responses were also monitored after topical lingual application of U73122 at 10 × 10⁻⁶, 50 × 10⁻⁶, 150 × 10⁻⁶, or 250 × 10⁻⁶ M. U73122 at 10 × 10⁻⁶ and 50 × 10⁻⁶ M produced no changes in the CT response to N + Bz in the absence or presence of RTX relative to control (data not shown). U73122 at 150 × 10⁻⁶ or 250 × 10⁻⁶ M produced similar changes in N + Bz and N + Bz + RTX CT responses as shown in Fig. 1B. Because the magnitudes of the CT responses were not different at 150 × 10⁻⁶ or 250 × 10⁻⁶ M U73122, the data were pooled for analysis. In three animals, after exposure to U73122, the N + Bz CT response was decreased in the absence of RTX to near baseline. In addition, U73122 attenuated the effect of RTX on the N + Bz CT response relative to control (Fig. 2C, ●). In three additional rats, the inactive analog, U73343, at 250 × 10⁻⁶ M produced no effect on the N + Bz CT response in the absence or presence of RTX (Fig. 2D). These results show that the open-state TRPV1t activity and its sensitivity to further stimulation by RTX are enhanced at lower cell membrane PIP₂ levels. However, increasing cell membrane PIP₂ levels by inhibiting PLCs or by direct application of a synthetic PIP₂, inhibited open-state TRPV1t activity and its sensitivity to further stimulation by RTX.

RTX and SB-366791 specifically modulate TRPV1t activity without altering CT responses to sweet, bitter, and umami taste stimuli

To study whether TRPV1t inhibition or activation specifically modulates the N + Bz CT response or has nonspecific effects on other taste qualities, further experiments were performed to investigate the relationship between TRPV1t activity and CT responses to sweet, bitter, and umami stimuli. RTX induced a biphasic response in both tonic (Fig. 3A, ○) and phasic (Supplementary Fig. S1, A and C) N + Bz CT responses in WT mice. The maximum increase in the tonic and phasic responses was obtained at 1 × 10⁻⁶ M RTX. The responses were reduced to the rinse artifact level at 10 × 10⁻⁶ M RTX. In TRPV1 KO mice, both the tonic (Fig. 3A, ●) and phasic (Supplementary Fig. S1, B and C) NaCl CT responses were inhibited to baseline in the presence of Bz. This indicates that TRPV1 KO mice elicit no CT response to N + Bz. The tonic and phasic responses to n + Bz remained at baseline in the presence of 1 × 10⁻⁶ and 10 × 10⁻⁶ M RTX. Taken together, the above results indicate that both phasic and tonic components of the Bz-insensitive NaCl CT response are derived from Na⁺influx through TRPV1t channels.

Fig. 3.

Effects of RTX and SB-366791 on CT responses to N + Bz, sucrose, and quinine. A: effects of RTX on the normalized tonic N + Bz CT responses from 3 wild-type (WT) and 3 knock-out (KO) mice. In KO mice, RTX did not elicit a CT response above baseline between 0 and 10 × 10⁻⁶ M (P > 0.05 with respect to 0; paired 2-sample t-test). B: effects of Bz, RTX, and SB-366791 (SB) on the normalized tonic N + Bz CT responses from 3 rats. In the presence of N + Bz, the CT response was significantly decreased from N (P = 0.0001). In the presence of N + Bz + SB-366791, it further decreased to baseline (P > 0.05 with respect to 0; paired 2-sample t-test). RTX elicited no tonic CT response above baseline in the presence of Bz + SB. C: effects of Bz + SB-366791 (1 × 10⁻⁶ M) on the tonic CT responses to 0.5 M sucrose and 0.01 M quinine. No statistical difference was observed between the tonic CT response to sucrose and quinine in the absence and presence of Bz + SB (P > 0.05; paired; n = 3).

SB-366791 is a specific blocker of TRPV1 (Gunthorpe et al. 2004) and TRPV1t (Lyall et al. 2004). SB-366791 (1 × 10⁻⁶ M) inhibited the rat CT response to N + Bz to rinse baseline (Fig. 3B). In the continuous presence of SB-366791, RTX at 1 × 10⁻⁶ M induced no increase in the neural response above baseline (Fig. 3B). SB-366791 produced similar results in WT mice as shown in Fig. 3B (data not shown). These results suggest that RTX failed to modulate the N + Bz CT response in rats and WT mice in the presence of SB-366791, a TRPV1/TRPV1t inhibitor, or in KO mice in which the native TRPV1/TRPV1t activity is silenced genetically.

Consistent with previous studies (Lyall et al. 2004), RTX (1 × 10⁻⁶ or 10 × 10⁻⁶ M) (data not shown) or Bz + SB-366791 (1 × 10⁻⁶ M) (Fig. 3C) had no effect on the tonic CT responses to 0.3 M sucrose or 0.01 M quinine. In an earlier study (Lyall et al. 2004), steady-state tonic CT responses to 0.02 M HCl were also not altered in the presence of 1 × 10⁻⁶ M or 10 × 10⁻⁶ M RTX. RTX also does not alter the ENaC-dependent NaCl CT response (Lyall et al. 2004).

TRPV1 KO mice elicited CT responses to MSG + Bz (Fig. 4A). Consistent with this, rats elicited CT responses to MSG + Bz + SB-366791 (Fig. 4B). In rats and WT mice, in the presence of Bz + SB-366791, there is no contribution of Na⁺to glutamate CT response. An equivalent concentration of MSG elicited a smaller CT response relative to NaCl in TRPV1 KO mice (Fig. 4A), WT mice (data not shown), and rats (Fig. 4B). This is because Na⁺and Cl⁻ are relatively permeable across the paracellular shunt (tight junctions) between adjacent TRCs in the taste bud. In contrast, glutamate, a larger anion than Cl⁻, has a lower permeability across the paracellular shunt. Thus, in MSG solutions, Na⁺has a greater permeability across the paracellular shunt relative to glutamate. This generates a diffusion potential across the lingual epithelium that is positive in the basolateral compartment. This, in turn, hyperpolarizes the receptor potential and results in a lower CT response to MSG relative to NaCl (Ye et al. 1991). CT responses to N + Bz and MSG + Bz were monitored in the absence or presence of 0 and 1 × 10⁻⁶ M RTX (Supplementary Fig. S2A). In three rats, RTX increased the tonic CT response in the presence of MSG + Bz by the same magnitude (δ RTX) as in N + Bz (Fig. 4C; P > 0.05, paired). These results suggest that, although the TRPV1t-dependent NaCl CT response is anion dependent, the RTX-induced increase in Na⁺ flux through TRPV1t does not depend on the accompanying anion.

IMP increased the CT response to MSG in the presence of Bz or Bz + SB by the same magnitude (Fig. 4C; P > 0.05, paired). These results suggest that genetically silencing TRPV1t in TRPV1 KO mice or inhibiting its activity pharmacologically with SB-366791 does not affect CT responses to glutamate and glutamate + IMP. Taken together, the above results suggest that RTX is a specific modulator of TRPV1t. At the concentrations used in this study, RTX had no effect on sweet, bitter, or umami stimuli.

Changes in membrane PIP₂ levels do not alter ENaC activity in TRCs

Topical lingual application of PAO (15 × 10⁻⁶ M) enhanced the tonic CT response to 0.1 and 0.3 M NaCl (Fig. 5A, filled bars) relative to control (Fig. 5A, open bars). All of the increase in the 0.1 M NaCl CT response could be accounted for by the increase in the NaCl + Bz CT response (P < 0.0001; paired). PAO did not induce a significant change in the ENaC-dependent NaCl CT response relative to control (Fig. 5A, Bz-sensitive; P > 0.05, paired).

Fig. 5.

A and B: effect of PAO and diC8-PIP₂ on TRPV1t-dependent and ENaC-dependent NaCl CT responses. A: in 4 rats, the post-PAO values for 0.1 M NaCl (P = 0.0068; paired), 0.3 M NaCl (P = 0.0001), and 0.1 M NaCl + Bz (P = 0.0001; paired) were significantly enhanced relative to control. No effect of PAO was observed on the ENaC-dependent NaCl CT response (Bz-sensitive; P > 0.05; paired). B: in 6 rats, the post-PIP₂ values for 0.1 M NaCl, 0.3 M NaCl, and 0.1 M NaCl + Bz were significantly attenuated relative to control (*P < 0.0001; paired). No effect of diC8-PIP₂ was observed on the ENaC-dependent NaCl CT response (Bz-sensitive; P > 0.05; paired). C and D: effect of topical lingual application of U73122 on CT responses to sweet and bitter stimuli. Two representative CT responses are shown in which the rat tongue was 1st rinsed with a rinse solution, R (Table 1), and then with (C) 0.01 M quinine or (D) 0.3 M sucrose (Table 1) before (Control) and after topical lingual application of 250 × 10⁻⁶ M U73122 (Post-U73122). The arrows indicate the time period when the tongue was stimulated with different solutions. In 3 rats, topical lingual application of U73122 inhibited the CT response to sucrose and quinine to rinse baseline relative to control (P > 0.05 with respect to 0; paired).

Topical lingual application of diC8-PIP₂ (250 × 10⁻⁶ M) decreased the magnitude of the tonic CT response to 0.1 and 0.3 M NaCl (Figs. 5B, filled bars) relative to control (Figs. 5B, open bars). All of the decrease in the 0.1 M NaCl CT response could be accounted for by the decrease in the NaCl + Bz CT response (P < 0.0001, paired). diC8-PIP₂ did not induce a significant change in the ENaC-dependent NaCl CT response relative to control (Figs. 5B, Bz-sensitive; P > 0.05, paired). Topical lingual application of U73122 (150 × 10⁻⁶ M) produced qualitatively similar changes in NaCl and NaCl + Bz as shown in Fig. 5B (data not shown). The inactive analog, U73343, had no effect on the CT responses to NaCl and NaCl + Bz relative to control (data not shown). These results suggest that in fungiform TRCs, ENaC activity is not sensitive to changes in PIP₂.

Changes in membrane PIP₂ levels modulate CT responses to sweet and bitter stimuli

We further studied whether the concentration of U73122 (a nonspecific blocker of PLCs) that blocks the N + Bz CT response also blocks CT responses to 0.01 M quinine and 0.5 M sucrose. U73122 at 150 × 10⁻⁶ M inhibited the tonic CT response to quinine (Fig. 5C) and sucrose (Fig. 5D) to near baseline relative to control. In three rats, after U73122 treatment, the tonic CT responses to sucrose and quinine were not statistically different from baseline (P > 0.05; paired; data not shown). In a separate set of rats, topical lingual application of 50 × 10⁻⁶ M U73122 did not produce any effect on the CT responses to sucrose or quinine relative to control (data not shown). These results show that the U73122 concentration that inhibited the N + Bz CT response is also an effective blocker of CT responses to sweet and bitter taste stimuli.

In type II TRCs, PLCβ₂ is involved in the transduction of sweet, bitter, and umami taste (Chandrashekar et al. 2006). PIP₂ is a substrate for PLCβ₂. We hypothesize that an increase in exogenous PIP₂ levels will augment endogenous membrane PIP₂. This will result in an increase in the concentration of substrate for PLCβ₂, which becomes activated in the presence of sweet and bitter stimuli. At 250 × 10⁻⁶ M, diC8-PIP₂ enhanced the CT response to quinine (Fig. 6A) and sucrose (Fig. 6B) relative to control. The normalized tonic CT responses from three rats are summarized in Fig. 6C. The results show that the concentration of diC8-PIP₂ that blocks the N + Bz CT response also modulates the neural response to sweet and bitter stimuli.

Fig. 6.

Effect of diC8-PIP₂ on CT responses to sweet and bitter stimuli. Two representative rat CT responses are shown in which the rat tongue was 1st rinsed with a rinse solution, R (Table 1), and then with 0.01 M quinine (A) or 0.5 M sucrose (B) (Table 1) before (Control) and after topical lingual application of 250 × 10⁻⁶ M diC8-PIP₂ (Post-PIP₂). The arrows indicate the time period when the tongue was stimulated with different solutions. C: summary of the effects of diC8-PIP₂ on the CT responses to sucrose and quinine. Topical lingual application of diC8-PIP₂ significantly enhanced the CT response to sucrose (P < 0.0001) and quinine (P < 0.05) relative to control (paired). D: a representative CT response is shown in which rat tongue was 1st stimulated with a rinse solution, R (Table 1), and then with 0.3 M sucrose, 0.003 M SC45647, 0.3 M sucrose + 250 × 10⁻⁶ M diC8-PIP₂, and 0.003 M SC45647 + 250 × 10⁻⁶ M diC8-PIP₂. The arrows indicate the time period when the tongue was stimulated with different solutions. E: changes in ³²P-PIP₂ induced by PAO and U73122. Treating the anterior lingual epithelium with 50 × 10⁻⁶ M (P = 0.0005) or 100 × 10⁻⁶ M (P = 0.003) U73122 increased ³²P-PIP₂ relative to control. Treating the anterior lingual epithelium with 10 × 10⁻⁶ M PAO did not alter ³²P-PIP₂ relative to control (P > 0.05). However, at 20 × 10⁻⁶ M, PAO decreased ³²P-PIP₂ relative to control (P = 0.01). At 100 × 10⁻⁶, M U73343 had no effect on ³²P-PIP₂ (P > 0.05; paired).

In mixtures containing 0.3 M sucrose + 250 × 10⁻⁶ M diC8-PIP₂ or 0.003 M SC45647 + 250 × 10⁻⁶ M diC8-PIP₂, PIP₂ produced an increase in the sweet response when applied along with the sweet stimulus (Fig. 6D). This is consistent with the fact that PIP₂ is lipid soluble and is incorporated into cell membranes very rapidly. However, in most of our studies, we topically applied synthetic PIP₂ and other drugs (PAO and U73122) for 30 min to ensure that the drugs reach their intracellular targets and produce maximum effects at the concentrations used.

Effect of PAO and U73112 on PIP₂ levels in the isolated anterior lingual epithelium

Treating the isolated anterior lingual epithelium containing fungiform taste papillae in vitro with 50 × 10⁻⁶ or 100 × 10⁻⁶ M U73122 increased the ³²P-PIP₂ counts per minute (cpm) by fourfold relative to untreated tissue (Fig. 6E; P = 0.0005 and 0.0003, respectively, paired). No increase in ³²P was observed with U73343 (P > 0.05, paired). In contrast, PAO produced a dose-dependent decrease in PIP₂ relative to control. At 20 × 10⁻⁶ M, PAO decrease PIP₂ by 51.5% relative to control (P = 0.01, paired). These results suggest that PAO and U73122 produce their effects by directly altering PIP₂ levels in taste cells.

Effect of G protein–coupled receptor and PLCβ₂ activation on TRPV1t activity

U73122 not only inhibited the N + Bz CT response (Fig. 2C) but also blocked CT responses to bitter and sweet taste stimuli (Fig. 5, C and D). Therefore in the next series of experiments, we studied whether RTX-sensitivity of TRPV1t is altered by the activation of GPCR-induced decrease in PIP₂ via PLCβ₂ in sweet and bitter sensing type II TRCs. RTX (1 × 10⁻⁶ M) produced the same increase in the CT response in N + Bz, N + Bz + 0.005 M SC45647 (Supplementary Fig. S2B), or N + Bz + 0.01 M quinine (Supplementary Fig. S2C). In three rats, no statistical difference was observed between the RTX-induced increase in the TRPV1t-dependent NaCl CT response in the absence or presence of SC45647 or quinine (P > 0.05, paired; data not shown). Taken together, the results shown in Supplementary Fig. S2, A–C, suggest that in sweet, bitter, and umami-sensing TRCs, GPCR-induced activation of PLCβ₂, and the resulting decrease in PIP₂ did not alter the RTX sensitivity of the N + Bz CT response. These studies suggest, but do not prove, that TRPV1t activity is independent of TRCs expressing GPCRs for bitter, sweet and umami taste.

Relationship between TRPV1t activity and CT responses to non-Na⁺salts

TRPV1t is a nonspecific cation channel (Lyall et al. 2004, 2005b, 2009). We hypothesize that inhibition or activation of TRPV1t would attenuate or enhance, respectively, CT responses to cations other than Na⁺. Consistent with this, in TRPV1 KO mice, the tonic CT response (R_m) to increasing KCl concentrations (Fig. 7A) was significantly smaller relative to its value in WT mice (Table 4). The concentration of KCl at which the CT response was half-maximum (K) was not different between WT and KO mice. These results suggest that in fungiform TRCs, part of the CT response to KCl is derived via K⁺entry through TRPV1t. We have previously shown that TRPV1t agonists and antagonists modulate a significant part of the CT responses to NH₄Cl (DeSimone et al. 2001) and CaCl₂ (Lyall et al. 2009). However, TRPV1t is not permeable to all cations. In TRPV1 KO mice, CT responses to normalized 0.03 M HCl were not different from WT mice (Fig. 7B). These results suggest that CT responses to HCl do not depend on H⁺entry through TRPV1t or TRPV1 (Liu and Simon 2001; Lyall et al. 2004). Taken together, these results suggest that TRPV1t plays a significant role in the elicitation of CT responses to Na⁺, K⁺, NH₄⁺, and Ca²⁺.

Fig. 7.

CT responses to KCl and HCl in WT and TRPV1 KO mice. A: the tonic KCl + Bz CT responses from 3 WT and 3 KO are summarized. B: representative CT responses to 0.3 M NH₄Cl and 0.03 M HCl in WT and TRPV1 KO mice. The arrows indicate the time period when the tongue was stimulated with different solutions. In 3 WT and 3 KO mice, the mean ratio between tonic HCl CT and tonic NH₄Cl CT response was 3.32 and 3.12, respectively. C: detection of phospholipase Cs (PLC)β₁, PLCβ₂, PLCβ₃, and PLCγ₁ in anterior and posterior taste fields. Using specific primer pairs (Table 3) in cDNAs made from rat anterior lingual epithelium containing fungiform taste papillae (FF) and lingual epithelium containing circumvallate papilla (CV), a single band of appropriate size was detected for PLCβ₁, PLCβ₂, PLCβ₃, and PLCγ₁. The negative controls (−) were run without cDNA template. Lanes 1 and 8 shows the molecular weight (MW) markers.

Table 4.

Fitted parameters for KCI CT responses in WT and TRPV1 KO mice

Mice	Rm	K	n
Wildtype	0.146 ± 0.010	0.278 ± 25	2.0 ± 0.3
TRPV1 knockout	0.108 ± 0.003	0.279 ± 9	2.5 ± 0.2
P	<0.0001	0.93	0.012

Values are means ± SE. The mean normalized CT response was plotted as a function of the KCl concentration (0.01–1 M) for both WT and TRPV1 KO mice. The data were fit to the Hill equation

$$R=\frac{R_m{c}^n}{K^n+c^n}$$ (4)

where R is the fitted normalized mean tonic CT response, R_m, is its maximum value, c is the KCl concentration in moles per liter, K is the KCl concentration at which R is half-maximal, and n is a positive number. The fitted parameters, R_m, K, and n were obtained using least squares fit criteria. The difference in each of the three parameters between WT and KO mice was tested for statistical significance using Student's t-test.

See Table 2 for abbreviations.

Detection of PLCβ₁, PLCβ₂, PLCβ₃, and PLCγ₁ in anterior and posterior taste fields

In cDNA made from anterior lingual epithelium containing fungiform papillae and lingual epithelium containing circumvallate papilla, using specific primer pairs shown in Table 3, we were able to detect the message for several PLC isoforms (PLCβ₁, PLCβ₂, PLCβ₃, and PLCγ₁; Fig. 7C). Most of the above PLC isoforms have previously been shown to be localized only in TRCs (Chandrashekar et al. 2006; Hacker et al. 2008; Toyono et al. 2005). These results suggest that U73122 increases membrane PIP₂ levels by inhibiting one or more PLC isoforms expressed in TRCs.

DISCUSSION

The results presented in this paper show that PIP₂ modulates TRPV1t activity in a manner similar to that of native TRPV1 in dorsal root ganglion neurons or the cloned TRPV1 channel expressed in heterologous cells. Below we will discuss the physiological significance of PIP₂ regulation of the putative TRPV1t salt taste receptor and the potential role of PIP₂ in other taste qualities.

An increase in PIP₂ inhibits and a decrease in PIP₂ enhances TRPV1t activity

Direct application of diC8-PIP₂ (Figs. 1B and 2B) or blocking PLC activity with U73122 (Fig. 2C) not only inhibited the control N + Bz CT response but also attenuated the RTX-induced increase in the neural response. Because both treatments produced equivalent results, it suggests that open-state TRPV1t activity and its further activation by RTX are inhibited by an increase in cell membrane PIP₂ levels. Consistent with this, inhibiting PLCs with U73122 increased PIP₂ levels in isolated lingual epithelium containing fungiform taste buds above control values (Fig. 6E). A decrease in PIP₂ levels induced by blocking its resynthesis by inhibiting PI4 kinases with PAO (Fig. 6E) significantly enhanced the open-state TRPV1t activity and its further activation by RTX (Figs. 1A and 2A). These results are consistent with observations that lowering of plasma membrane PIP₂ levels through antibody sequestration or PLC-mediated hydrolysis mimic the potentiating effects of nona-peptide bradykinin or nerve growth factor on cells expressing TRPV1 (Chuang et al. 2001). It is suggested that endogenous PIP₂ inhibits TRPV1 and PLC catalyzes the hydrolysis of PIP₂ to yield IP₃ and DAG and releases TRPV1 from PIP₂-mediated inhibition (Chuang et al. 2001). DAG activates PKC, and sensitizes TRPV1 to further activation by agonists (Petrocellis and Marzo 2005). Consistent with this, phorbol 12-myristate 13-acetate (PMA), a specific activator of PKC, increased the sensitivity of TRPV1t-dependent NaCl CT responses to RTX. Inhibiting PKC activity by RO31-8220 or topical lingual application of pseudo-substrate inhibitor peptide, a specific blocker of PKCε, abolished the PMA-induced sensitization of the TRPV1t-dependent NaCl CT responses to RTX (Lyall et al. 2009).

In dorsal root ganglion neurons that express native TRPV1 or cloned TRPV1 expressed in HEK293 cells, no inward currents are observed in the absence of TRPV1 agonists at normal physiological pH and temperature. It is suggested that interactions between several signaling effectors (PIP₂, [Ca²⁺]_i, ATP) and the phosphorylation state of TRPV1 determine whether TRPV1 channel is tonically active or inactive. ATP and calmodulin interact with the TRPV1 protein at its N-terminal ankyrin repeat domains. These domains contain long finger like folds between ankyrin repeats that can bind both ATP and calmodulin. PIP₂ binds to the C terminus of TRPV1 protein. The binding of ATP to the N terminus or PIP₂ binding to the C terminus of the TRPV1 protein prevents desensitization during repeated applications of capsaicin. Calmodulin binding to the N terminus seems to play an opposing role and is necessary for desensitization (Vennekens et al. 2008).

Our results further suggest that PIP₂ and RTX interact with TRPV1t at two different sites. Consistent with this, in TRPV1, deletion of a very short region in the C terminus (residues 777–792, including 4 basic residues, in rats) eliminated PIP₂ inhibition (Prescott and Julius 2003). At present, it is not clear if this region forms a PIP₂-binding pocket. In contrast, residues Y511 and T550 in the third and fourth transmembrane domain are responsible for binding to capsaicin. RTX binding requires the presence of methionine in position 547 (Petrocellis and Marzo 2005).

It has been suggested that, besides PIP₂, phosphatidyl inositol 4 phosphate may also have a role in maintaining TRPV1 activity (Lukacs et al. 2007). Depending on the experimental conditions, both inhibitory and activating effects of PIP₂ have been observed on TRPV1 (Vennekens et al. 2008). In excised patches, capsaicin-activated currents were inhibited by l-polylysine, a PIP₂ scavenger, and potentiated by PIP₂ (Stein et al. 2006). In the presence of extracellular Ca²⁺, high capsaicin concentrations desensitize TRPV1 by activating PLC, inducing depletion of both PIP₂ and its precursor phosphoinositol 4-phosphate. In these studies, U73122 and dialysis of PIP₂ and phosphoinositol 4-phosphate inhibited the desensitization of TRPV1. In addition, the conversion of PIP₂ to phosphoinositol 4-phosphate by PIP₂-5 phosphatase did not inhibit TRPV1 at high capsaicin concentrations. These studies suggest that phosphoinositol 4-phosphate plays a significant role in maintaining channel activity. In contrast, currents induced by low capsaicin concentrations and moderate heat were potentiated by a decrease in PIP₂ and inhibited by an increase in PIP₂ (Lukacs et al. 2007). It is likely that the sensitization and desensitization of TRPV1 may be regulated by interactions of several intracellular effectors. This includes Ca²⁺, calmodulin and ATP interactions with the N terminus and PIP₂ binding to the C terminus of the channel protein (Lishko et al. 2007). Further studies are needed to determine whether, along with PIP₂, phosphoinositol 4-phosphate also plays a role in the regulation of TRPV1t in the absence and presence of agonists.

In contrast to TRPV1t, ENaC activity was insensitive to changes in PIP₂ (Fig. 5, A and B). In A6 cells, an increase in PIP₂ was shown to increase ENaC activity by direct interaction with β- or γ-ENaC subunits (Yue et al. 2002). In another study, inhibiting PLC with U73122 in A6 cells enhanced ENaC activity by increasing the mean open time of the channel (Pochynyuk et al. 2008). These results suggest that PIP₂ regulation of ENaC may be tissue and cell specific.

An increase in PIP₂ enhanced the responses to representative sweet and bitter stimuli (Fig. 6, A–C). PIP₂ is a substrate of PLC. An increase in substrate concentration is expected to increase PLC activity. Sweet and bitter sensing TRCs express PLCβ₂. Inhibiting PLCβ₂ activity by U73122 inhibited the tonic CT responses to representative sweet and bitter stimuli (Fig. 5, C and D). Thus overall, our data suggest that, in TRCs, an increase in PIP₂ specifically inhibits TRPV1t. It increases PLCβ₂ activity and does not alter ENaC activity in TRCs.

Relationship between TRPV1t activity and CT responses to salty, sweet, sour, bitter, and umami stimuli

In TRPV1 KO mice, KCl concentration versus the magnitude of the CT response profiles gave significantly smaller R_m values relative to WT mice (Fig. 7A; Table 3). In an earlier study (DeSimone et al. 2001), cetylpridinium chloride, a compound that modulated the TRPV1t-dependent NaCl CT response, also altered responses to KCl and NH₄Cl in a biphasic manner. The concentration of cetylpridinium chloride that completely inhibited the N + Bz CT response to baseline induced an ∼30% decrease in the CT response to 0.1 M KCl. These results suggest that a significant part of the KCl CT response is derived from K⁺entry through TRPV1t. RTX also produced a biphasic response on the CT response to CaCl₂, and SB-366791, a specific TRPV1t/TRPV1 blocker, inhibited a significant part of the CaCl₂ CT response (Lyall et al. 2009). These results suggest that a significant part of the CT response to CaCl₂ also depends on Ca²⁺influx through TRPV1t. These data further strengthen the hypothesis that TRPV1t is a nonselective cation channel that is permeable to Na⁺, K⁺, Ca²⁺, and NH₄⁺ and plays a significant role in eliciting CT responses to the above cations (Lyall et al. 2004).

However, TRPV1t is not permeable to H⁺(Fig. 7B). This conclusion is supported by the observations that, in the absence of an agonist, TRPV1t-dependent NaCl CT responses show no pH sensitivity (Lyall et al. 2002). In addition, HCl CT responses are not altered in the presence of 1 × 10⁻⁶ or 10 × 10⁻⁶ M RTX (Lyall et al. 2004). Thus, although a decrease in pH is a sufficient signal to activate TRPV1, a decrease in pH by itself does not modulate the activity of TRPV1t. However, the agonist-induced increase in the TRPV1t-dependent NaCl CT responses showed pH sensitivity, indicating that agonist-induced activation of the channel is pH dependent (Katsumata et al. 2008; Lyall et al. 2004, 2005a).

Genetically silencing TRPV1 (Fig. 4A) or inhibiting its activity pharmacological (Fig. 4, B and C) did not alter CT responses to MSG or MSG + IMP. In sweet, bitter, or umami sensing type II TRCs, activation of specific GPCRs is linked to the activation of PLCβ₂, hydrolysis of PIP₂ into IP₃ + DAG, and a decrease in membrane PIP₂ levels. However, the RTX-induced increase in the TRPV1t-dependent NaCl CT response was not significantly affected in mixtures containing N + Bz + quinine or N + Bz + SC45647 (Supplementary Fig. S2). In contrast, a decrease in membrane PIP₂ levels increased the TRPV1t-dependent NaCl CT response in the presence of RTX (Figs. 1A and 2A). These results suggest that TRPV1t activity is most likely not dependent on TRCs expressing GPCRs and PLCβ₂. Consistent with this, CT responses to sucrose and quinine were not altered in the presence of 1 × 10⁻⁶ or 10 × 10⁻⁶ M RTX (Lyall et al. 2004). Our data suggest that the observed effects of U73122 on the TRPV1t-dependent NaCl CT response (Fig. 2C) are most likely caused by the inhibition of one or more PLC isoforms (Fig. 7B) expressed in a subset of TRCs that are not involved in sweet, bitter, and umami taste transduction.

Relationship between TRPV1t activity and salt taste perception

In WT mice and rats, the TRPV1t-dependent NaCl CT response contributes ∼30% of the total NaCl CT response (Fig. 5, A and B). Using salt solutions containing amiloride, NaCl detection thresholds were found not to differ between TRPV1 KO mice and WT mice (Ruiz et al. 2006; Treesukosol et al. 2007). These results suggested that, in TRPV1 KO mice, additional salt taste receptors, present in taste fields other than the fungiform TRCs, must contribute to salty taste. During stimulation of the tongue with salt solutions containing agonists, the contribution of TRPV1t to the total NaCl CT response can be increase by 100% or more (DeSimone and Lyall 2006, 2008; Lyall et al. 2009). In humans, salt taste is predominantly amiloride insensitive and may contribute as much as 80% to human salt taste perception (Ossebaard and Smith 1995). Recent studies suggest that the activation of TRPV1t, especially with nonpungent agonists, such as Maillard-reacted peptides, produced increased salt taste sensitivity in humans (Katsumata et al. 2008) and in animal models (Rhyu et al. 2009). Human subjects show significant increase in salt taste perception when NaCl was presented in mixtures containing low concentrations of Maillard-reacted peptides. At high peptide concentrations of the peptide, the same human subjects reported a significant decrease in salt taste perception (Katsumata et al. 2008). This suggests that TRPV1t may also play a role in human salt taste perception.

In Dahl salt-sensitive rats, maintained on a 0.32% Na⁺diet containing a concentration of Maillard-reacted peptide conjugated with galacturonic acid that produces maximum activation of the TRPV1t-dependent NaCl CT response, lowered Na⁺balance and inhibited the spontaneous rise in blood pressure observed in rats maintained on the salt diet alone (Masilamani et al. 2009). In contrast, Dahl salt-sensitive rats given Maillard-reacted peptide conjugated with xylose that produced maximum inhibition of the TRPV1t-dependent NaCl CT response increased the rate at which the blood pressure rose spontaneously relative to rats maintained on the salt diet without the peptide (Masilamani et al. 2009). Thus under the conditions in which TRPV1t is upregulated and contributes significantly to salt taste perception, modulating its activity affects changes in salt taste sensitivity in a predictable manner.

A major issue with using TRPV1t agonists as salt taste enhancers is that they increase TRPV1t-dependent NaCl CT responses within a narrow range of concentrations. At high concentrations, all of the compounds tested to date behave as antagonists of TRPV1t and decrease NaCl CT responses. The results presented in this paper and in an earlier publication (Lyall et al. 2009) suggest that the TRPV1t agonist concentration range can be greatly extended by inhibiting the desensitization of TRPV1t channel by lowering cell Ca²⁺, preventing channel dephosphorylation, and lowering cell membrane PIP₂ levels. The same experimental manipulations also eliminate the salt taste–suppressing effect of most TRPV1t modulators at high concentrations. Further understanding of how TRPV1t is regulated in TRCs may facilitate the use of TRPV1t agonists as salt taste enhancers and blood pressure–lowering drugs in the future.

Based on the comparative effects of PIP₂ modulators on the CT response to salty, sweet, bitter, and umami stimuli summarized in Table 5, our data suggest that TRPV1t, a nonspecific cation channel, accounts for all of the Bz-insensitive CT response to Na⁺in WT mice and rats. TRPV1t also contributes significantly to the CT response to K⁺, Ca²⁺, and NH₄⁺ salts but is not permeable to H⁺. RTX is a specific agonist of TRPV1t and produces biphasic effects on the CT responses to Na⁺, K⁺, Ca²⁺, and NH₄⁺. At low concentrations (≤1 × 10⁻⁶ M), RTX is a TRPV1t agonist, and at high concentrations (>1 × 10⁻⁶ M), it behaves as an antagonist of the channel. TRPV1t activation by RTX does not depend on the anion associated with Na⁺. TRPV1t activity is most likely not associated with TRCs that express GPCRs for sweet, bitter, or umami taste. We conclude that, similar to the case with TRPV1, PIP₂ is a specific blocker of TRPV1t activity. An increase in PIP₂ specifically inhibited the TRPV1t-dependent NaCl CT response and desensitized it to further stimulation by RTX. These data strengthen the hypothesis that TRPV1t is a candidate mammalian amiloride- and Bz-insensitive salt taste receptor and is regulated by changes in cell [Ca²⁺]_i, PKCε, calcineurin (Lyall et al. 2009), and PIP₂.

Table 5.

Comparative effects of PIP₂ modulators on the CT response to salty, sweet, bitter, and umami stimuli

Drug	Bz-insensitive NaCl CT (−RTX)	Bz-insensitive NaCl CT (+RTX)	Bz-sensitive NaCl CT	Sucrose CT	SC45647 CT	Glycine CT	Quinine CT	MSG CT
Bz	↔	↔	↓	↔	↔	↔	↔	↓
SB-366791	↓	↓	↔	↔	↔	↔	↔	↓
RTX	↑	biphasic	↔	↔	↔	↔	↔	↔
PAO	↑	↑	↔
U73122	↓	↓	↔	↓	↓	↓	↓
U73343	↔	↔	↔	↔	↔	↔	↔	↔
diC8-PIP2	↓	↓	↔	↑		↑	↑
MSG	↔	↔						↑
SC45647	↔	↔			↑
Quinine	↔	↔					↑
IMP	↔	↔						↑

Bz and SB-366791 decrease MSG response indirectly by inhibiting Na⁺ flux via ENaC and TRPV1t, respectively. SB-366791 also decreases part of the CT response to K⁺, Ca²⁺, NH₄⁺. See Table 2 for abbreviations.

GRANTS

This work was supported by National Institute on Deafness and Other Communication Disorders Grants DC-000122 and DC-005981 to V. Lyall and a grant from the Campbell Soup Company to J. A. DeSimone.