|Home | About | Journals | Submit | Contact Us | Français|
Modulatory effects of pHi and [Ca2+]i on taste receptor cell (TRC) epithelial sodium channel (ENaC) were investigated by monitoring chorda tympani (CT) responses to NaCl and KCl at various lingual voltages, before and after lingual application of ionomycin and with 0–10mM CaCl2 in the stimulus and rinse solutions adjusted to pHo 2.0–9.7. 0.1 and 0.5M KCl responses varied continuously with voltage and were fitted to an apical ion channel kinetic model using the same parameters. ENaC-dependent NaCl CT response was fitted to the same channel model but with parameters characteristic of ENaC. A graded increase in TRC [Ca2+]i decreased the ENaC-dependent NaCl CT response, and inhibited and ultimately eliminated its pH sensitivity. CT responses to KCl were pHi- and [Ca2+]i-independent. Between ±60 mV applied lingual potential, the data were well described by a linear approximation to the nonlinear channel equation and yielded 2 parameters, the open-circuit response and the negative of the slope of the line in the CT response versus voltage plot, designated the response conductance. The ENaC-dependent NaCl CT response conductance was a linear function of the open-circuit response for all pHi-[Ca2+]i combinations examined. Analysis of these data shows that pHi and [Ca2+]i regulate TRC ENaC exclusively through modulation of the maximum CT response.
In the taste receptor cells (TRCs) of the lingual fungiform papillae of many species, the amiloride- and benzamil (Bz)-sensitive epithelial sodium channel (ENaC) is the Na+-specific salt taste receptor (DeSimone and Lyall 2008; Chandrashekar et al. 2010). TRC ENaC is constitutively active. ENaC conductance and hence the Bz-sensitive NaCl chorda tympani (CT) taste nerve responses are regulated by a variety of physiologically important modulators. These include intracellular pH (pHi) (Lyall et al. 2002), intracellular calcium ([Ca2+]i) (DeSimone et al. 2012a), osmolarity (Lyall et al. 1999), cAMP, insulin, arginine vasopressin (Gilbertson et al. 1993; Baquero and Gilbertson 2011; Mummalaneni et al. 2014), aldosterone (Herness 1992; Lin et al. 1999), angiotensin II (AngII) (Shigemura et al. 2013), and Na+ self-inhibition (Gilbertson and Zhang 1998; DeSimone and Lyall 2008). Synthetic low molecular weight ENaC agonists, N-(2-hydroxyethyl)-4-methyl-2-((4-methyl-1H-indol-3-yl)thio)pentanamide (S3969) (Lu et al. 2008), and N,N,N-trimethyl-2-((4-methyl-2-((4-methyl-1H-indol-3-yl)thio)pentanoyl)oxy)-ethanaminium iodide (S3559) also enhance the Bz-sensitive NaCl CT responses in rats (Mummalaneni et al. 2014). Here, we further examined the regulation of rat ENaC-dependent Bz-sensitive NaCl CT responses by H+ and Ca2+ in situ.
ENaC is an ion channel in the apical membrane of TRCs. The current density through it varies continuously with both stimulus Na+ concentration and electrical potential across the channel. Accordingly, we investigated the ENaC-dependent NaCl CT response as a function of Na+ concentration and the potential applied to the tongue using a flow chamber held in place by a vacuum (Ye et al. 1993). These responses were compared with KCl CT responses as a function of KCl concentration and applied voltage. While TRCs possess a variety of K+ selective ion channels, none of these appears to be an apical membrane taste transducer channel (Wang et al. 2009). Pharmacological studies suggested that the KCl taste receptor might involve nonspecific cation channel(s) that can be blocked by cetylpyridinium chloride, SB366791, and other modulators of the TRPV1 channel (Lyall et al. 2004a; DeSimone et al. 2012a). From the data we calculated 2 key variables, the CT responses at zero voltage clamp (open-circuit response) and the negative slope of the CT response with applied voltage, defined as the response conductance. We investigated the pH and Ca2+ dependence of these 2 variables and their unique relationship to one another to gain insight into how ENaC regulators in general influence one another.
Our results show that at the physiological TRC [Ca2+]i levels, the ENaC-dependent NaCl CT response displays its maximum sensitivity to changes in pHi. As TRC [Ca2+]i increases, it suppresses ENaC activity proportionately. However, [H+]i and [Ca2+]i do not function independently. We show that while increasing TRC [Ca2+]i suppresses the CT response to Na+, it also inhibits, and at high [Ca2+]i concentrations ultimately eliminates the ability of pHi to regulate the ENaC-dependent NaCl CT response. In contrast, KCl CT responses were also transduced by presumably an apical cation channel, that is Bz-insensitive and is not regulated by either pHi or [Ca2+]i. Our data further demonstrate similarities and differences between the ENaC transducer for Na+ and the apical ion channel mediating K+ taste transduction.
Animals were housed in the Virginia Commonwealth University (VCU) animal facility in accordance with institutional guidelines. All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of VCU. Twenty female Sprague–Dawley rats (150–200g) were used in this study. For CT recordings, rats were anesthetized by intraperitoneal injection of sodium pentobarbital (60mg/kg), and supplemental sodium pentobarbital (20mg/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). The left CT nerve was exposed laterally as it exited the tympanic bulla and was placed onto a 32-gauge platinum-iridium wire electrode. An indifferent electrode was placed in nearby tissue. Neural responses were differentially amplified with an optically coupled isolated bioamplifier (ISO-80; World Precision Instruments). Stimulus solutions were injected into a Lucite chamber (3mL; 1mL/s) affixed by vacuum to a 30-mm2 patch of anterior dorsal lingual surface. The chamber was fitted with separate Ag–AgCl electrodes for passing current and measuring potential. These electrodes served as inputs to a voltage- and current-clamp amplifier that permitted the recording of neural responses with the chemically stimulated receptive field under zero current clamp (0 cc) or ±60 mV voltage clamp. Voltages were referenced to the mucosal side of the tongue (Ye et al. 1993, 1994). While recording a CT response to a stimulus, clamp voltages were set relative to the zero current clamp potential (same as the open-circuit voltage). These were typically less than 5 mV relative to the mucosal side. Accordingly, zero voltage (V = 0) was the open-circuit voltage referenced to itself and all nonzero clamp voltages were similarly referenced to the open-circuit voltage.
The compositions of the rinse solutions (R), control stimuli (CS), and NaCl stimulus solutions (S) used in CT nerve recordings are shown in Table 1. Each experiment began and ended with the application of control stimuli, consisting of 0.3M NH4Cl and 0.3M NaCl, used to assess the recording stability. The preparation was considered stable if the difference between the magnitude of the responses to the control stimuli at the beginning and end of the experiment was less than 10%. CT responses to lingual stimulation with 0.1M NaCl (S) were recorded at V = 0 and with V = ±60 mV each expressed relative to the open-circuit potential as previously described. Potentials were applied across the NaCl-stimulated anterior tongue in the flow chamber, and responses at a given voltage were obtained relative to the baseline response in the rinse solution containing 0.01M KCl (R). The pH (pHo) of the rinse (R) and NaCl stimulus (S) were adjusted to 2 (R2 and S2), 7 (R7 and S7), or 9.7 (R9.7 and S9.7). In addition, the rinse and NaCl stimulus solutions contained either: 0, 0.001, 0.005, or 0.01M CaCl2. KCl stimuli were studied at pH 7 and also contained CaCl2 as described for NaCl. The CT responses were recorded under control conditions (pre-ionomycin) or after topical lingual application of 150×10–6 M ionomycin (Sigma-Aldrich) dissolved in dimethyl sulfoxide (DMSO; Sigma-Aldrich) for 30min (post-ionomycin) (DeSimone et al. 2012a, 2012b; Mummalaneni et al. 2014). Before ionomycin treatment, the presence of 0.01M CaCl2 or less in the rinse and NaCl stimulus solution did not affect the CT response relative to the NaCl response without added CaCl2 (DeSimone et al. 2012a, 2012b). This suggests that TRC apical membranes have extremely low permeability to Ca2+.
The following rinse and NaCl lingual stimulation series were used in CT experiments:
R (0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV) → R (0 mV) → R7(0 mV) → S7(0 mV) → S7(−60 mV) → S7(+60 mV) → S7(0 mV) → R7(0 mV) → R(0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) →0.3M NaCl(0 mV)
R (0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV) → R (0 mV) → R7(0 mV) → R2(0 mV) → S2(0 mV) → S2(−60 mV) → S2(+60 mV) → S2 (0 mV) → R2(0 mV) → R7(0 mV) → R(0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV)
R (0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV) → R (0 mV) →R7(0 mV) → R9.7(0 mV) → S9.7(0 mV) → S9.7(−60 mV) → S9.7(+60 mV) → S9.7(0 mV) → R9.7(0 mV) → R7(0 mV) → R(0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV)
R (0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV) → R(0 mV) → R7(0 mV) → S7 KCl(0 mV) →S7 KCl(−60 mV) → S7 KCl(+60 mV) → S7 KCl(0 mV) → R7(0 mV) → R(0 mV) → 0.3M NH4Cl(0 mV) → R(0 mV) → 0.3M NaCl(0 mV)
Typically, a stimulus solution at the same pH as the rinse was then applied to the tongue at zero voltage clamp and allowed to remain from 1 to 2min in that state. Without rinsing, the system was then switched to −60 mV voltage-clamp for 1–2min and then to +60 mV voltage clamp for 1–2min. (For the 0.5M KCl stimulus, the additional clamp voltages of ±20, ±40, and ±80 mV were included, see Figure 1A). The starting rinse was then reapplied for at least 1min. If the protocol included a pH change, then another rinse solution at the new pH was applied and the stimulation series was continued at that pH. The mean tonic CT response of a stimulus (final 30 s period) was then normalized by dividing it by the mean tonic CT response to 0.3M NH4Cl over a similar 30 s period. The normalized data were reported as the mean ± standard error of the mean (SE) of the number of animals. Integrated neural responses and lingual current and voltage changes were captured on disk using LabView software (National Instruments) and analyzed off-line as described previously (Lyall et al. 2004a; Mummalaneni et al. 2014). Student’s t-test was employed to analyze the difference between mean values. The values of the open-circuit and response conductance for the ENaC-dependent part of the NaCl response versus voltage data (see Table 2) were obtained by linear regression. In some cases, the response versus voltage curves were fitted to the nonlinear ion channel equation using the nonlinear regression curve fitting routine of SigmaPlot 8 (see Figure 4; Table 3).
We investigated the effect of systematic increases in TRC [Ca2+]i over a range of pH values on the rat CT response to NaCl and KCl. Ionomycin (150×10–6 M) was topically applied to the dorsal surface of the rat tongue to increase the TRC apical membrane Ca2+ permeability. Then adding CaCl2 to both the rinse and stimulus solutions at a given pH produced a graded linear decrease in the CT response to NaCl with increasing outside Ca2+ concentration, [Ca2+]o (in M). Without ionomycin, no effect of [Ca2+]o is observed on the NaCl CT response (DeSimone et al. 2012a, 2012b). The NaCl CT response, r NaCl, is given by: r NaCl = 0.534–0.0378[Ca2+]o (see Figure 2B). Ionomycin establishes the following linear relationship between [Ca2+]i and [Ca2+]o, viz.: [Ca2+]i = [Ca2+]iu + a([Ca2+]o − [Ca2+]iu). Here [Ca2+]iu is the unperturbed intracellular Ca2+ concentration, and a is the increase in the intracellular Ca2+ concentration per unit increase in outside Ca2+ concentration. The parameter a is determined by both the amount of ionomycin in the TRC membranes and by the intracellular Ca2+ buffer capacity.
Figure 1A shows that replacing 0.5M KCl with 0.5M KCl + 5×10–6 M Bz had no effect on the CT response, indicating that ENaC plays no role in K+ salt taste response. Figure 1A also shows the effect of a series of voltage perturbations (−80 mV to +80 mV) on the CT response to 0.5M KCl at pH 7 and [Ca2+]o = 0. Increasing the electronegativity of the tongue interior increased the CT response continuously in proportion to the size of the voltage perturbation, while increasing electropositivity proportionately decreased the CT response. Perturbations of either polarity relaxed rapidly to the unperturbed state when the applied voltage circuit was opened. Figure 1B shows the open-circuit response (V = 0) and 8 additional responses at both positive and negative V values. Assuming that the CT response to a salt is proportional to the apical channel cation influx, a kinetic model can be derived that yields an expression for the response, r, as a function of the stimulus salt concentration, n, and the applied translingual voltage, V, as shown previously (Hendricks et al. 2000; Mummalaneni et al. 2014):
Here ϕ = FV/RT is a dimensionless potential with the voltage scaled as usual by the physical constants RT/F (=26 mV) and δ is the fraction of V dropped across the apical membranes of cation-detecting taste cells. With V = 0, the open-circuit response, r o is:
Where, r m is the maximum response and K r is operationally defined as the stimulus concentration at which r o is half-maximal. More fundamentally, it can also be interpreted in kinetic and thermodynamic terms (see eq. 7). The solid line in Figure 1B was drawn from the least squares fit of equation (1) to the data (R 2 = 0.99) where the parameter values are: r m = 1.49, K r = 0.4M, and δ = 0.203. For a useful range of voltages about V = 0 for which δ ϕ < 1, the response will be a quasilinear function of voltage. Under these conditions, equation (1) becomes:
Because the CT response is assumed to be proportional to the ion flux through the apical membrane channels mediating transduction, the slope in equation (3) is analogous to a channel conductance. For that reason, we call the slope the response conductance and designate it by κ, that is
The dashed line in Figure 1B was drawn using the linear approximation (eq. 3) and the same fit parameters obtained in fitting the data to the nonlinear function (eq. 1). For 0.5M KCl, the linear approximation is good over most of the voltage range and only begins to diverge from the nonlinear expression for r at about −60 mV, but the difference is still less than 1%.
We also obtained data for 0.1M KCl at V = 0 and ± 60 mV. The mean response ± SE was plotted for each of the 3 voltages in Figure 1C (see also Figure 3D below). The solid curve in Figure 1C was drawn from equation (1) for n = 0.1M KCl, and it also represents the data at 0.1M KCl quite well. The dashed curve shows the linear approximation drawn from equation (3) using the calculated values: r o = 0.30, and κ = 0.00186. The approximation falls within the SE of the data points at ±60 mV. The linear approximation deviates from the nonlinear curve by 3% at V = −60 mV and by 7% at V = 60 mV. This demonstrates the robustness of the model and validity of the linear approximation for 0.1M KCl.
A similar analysis was performed for the ENaC-dependent part of the CT response to 0.1M NaCl for pHo = 7 and [Ca2+]o = 0. For these conditions, the mean response ± SE for the ENaC-dependent part of the response was plotted for the 3 voltages: V = −60, 0, and 60 mV in Figure 1D. From an independent study of the open-circuit response (r o, see eq. 2) as a function of the NaCl concentration (n) (Mummalaneni et al. 2014), we already know the values of r m (1.20), K r (0.268M), and δ (~0.2). Therefore, using these values of r m and K r in equation (1), we obtained the solid curve in Figure 1D as the best fit to the data points using δ as the single adjustable parameter. The least squares fit value of δ was 0.246, confirming that its best fit value should be close to 0.2. The dashed curve shows the linear approximation drawn from equation (3) which yields r o = 0.325 and κ = 0.0023. The deviation of the linear approximation from the nonlinear curve is just 3% at V = −60 mV and 8% at V = 60 mV. The regression line connecting the response points at −60, 0, and 60 mV (not shown here) is a very good representation of the linear approximation to the exact nonlinear response curve within the same voltage limits (see Table 3 below). Linear regression methods simplify the analysis by allowing us to obtain the open-circuit response and the response conductance without first obtaining parameters available only from a nonlinear fit to the data.
Figure 2A shows that at V = 0 and [Ca2+]o = 0, the CT response to 0.1M NaCl increased linearly with pHo over the range where measurements were made, that is, between pHo 2.0 and 10.3 (R 2 = 0.93; slope = 0.0417±0.0056, and intercept = 0.310±0.038). Because no data were obtained outside the pHo 2–10.3 range, the regression line does not apply outside that range. The pHo dependence resides exclusively in the ENaC dependent Bz-sensitive NaCl CT response (Lyall et al. 2002). Similarly, Figure 2B shows that at V = 0 and [pH]o = 7, the CT response linearly decreased in ionomycin-treated rat tongues (R 2 = 0.96; slope = −0.0378±0.0054, and intercept = 0.534±0.030). Like the pHo dependence, the [Ca2+]o dependence resides exclusively in the ENaC-dependent Bz-sensitive part of the NaCl CT response (DeSimone et al. 2012a). Because we operationally altered NaCl CT responses by changing pHo and [Ca2+]o it is convenient to refer the changes that were noted back to these parameter values, but in each case a change in pHo or [Ca2+]o produced the corresponding change in pHi and [Ca2+]i that is the actual modulator of the NaCl CT response (Lyall et al. 2002).
Figure 3A and andBB show the variations in a representative NaCl CT response when the tongue was stimulated using lingual stimulation Series 1, 2, and 3. As shown in Figure 2A, the open-circuit (0 mV) response increased as pHo increased from 2 to 9.7. Also the excursion in the response due to either a +60 or −60 mV applied voltage increased as the open-circuit response increased. Replacing R7 with R2 elicited a small CT response. This is due to the presence of HCl in R2. Accordingly, the response to 0.1M NaCl at pHo 2 (S2) has as its baseline the acid response to R2. Thus, the tonic response to S2 was calculated with respect to the R2 baseline. In our previous studies (Lyall et al. 2002), the CT response to NaCl + Bz between pHo 2 and 10.3 did not differ statistically from the NaCl response at pHo 2. This indicates that the Bz-insensitive part of the NaCl CT response is pHo independent. This further indicates that pH-sensitivity resides solely in the ENaC-dependent component of the NaCl CT response.
Following the ionomycin-induced increase in TRC [Ca2+]i (Figure 3B), the larger tonic CT responses at pH o > 2 were markedly less pH-dependent compared with the preionomycin case (Figure 3A). In fact, the tonic CT responses at pHo 7 and 9.7 were of similar magnitude as the response at pHo 2. These results suggest that an increase in TRC [Ca2+]i blocks the ENaC-dependent NaCl CT response even at higher pHo, conditions normally favoring increased ENaC activity (Figures 2A and and3A).3A). The minimum CT response to NaCl + 0.01M CaCl2 was obtained at pHo 2 postionomycin treatment. In our previous studies, we observed that the CT response to 0.1M NaCl at pH 2 was essentially equal to the Bz-insensitive part of the CT response (Lyall et al. 2002). In this study, the joint condition: pHo = 2, [Ca2+]o = 0.01M resulted in the minimum response to NaCl and this did not differ significantly from the response to 0.1M NaCl + Bz. The mean response to 0.1M NaCl + 0.01M CaCl2 at pHo 2 post-ionomycin treatment was 0.176±0.012 (N = 3). This is not significantly different from 0.160±0.004, (N = 32) the response to 0.1M NaCl + Bz found in earlier studies (P = 0.2477) (Mummalaneni et al. 2014). Therefore, the mean response to 0.1M NaCl + 0.01M CaCl2 at pHo 2 post-ionomycin treatment was subtracted from responses at all other conditions of pHo and [Ca2+]o. This enabled us to ascribe the effects of pHo and [Ca2+]o solely to the ENaC-dependent part of the NaCl CT response.
The ionomycin-induced increase in TRC [Ca2+]i also eliminated the acid-induced response when R2 was replaced by R7 (Figure 3B). This is due to the [Ca2+]i-induced increase in the activity of the basolateral Na+-H+ exchanger-isoform 1 (NHE-1) in sour transducing TRCs, which increases pHi and consequently reduces the CT response to acidic stimuli (Lyall et al. 2004b; Vinnikova et al. 2004). The experiment shown in Figure 3B was also performed in other rats with rinse and NaCl stimulus solutions containing 0.001 and 0.005M CaCl2 before and after topical lingual application of 150×10–6 M ionomycin (data not shown).
Figure 3C shows the KCl CT response when the tongue was stimulated according to Series 4. The CT response to 0.1M KCl + 0.01M CaCl2 at pHo 7 (S7 KCl) was unaffected by exposure of the rat tongue to ionomycin. Data were also obtained in other rats with CaCl2 at 0.001M and 0.005M and these gave the same negative results (data not shown). Figure 3D shows the CT response to 0.1M KCl at clamp voltages of −60, 0, and 60 mV before and after ionomycin treatment when the stimulus contained either 0.001, 0.005, or 0.01M CaCl2. Reponses before ionomycin treatment did not differ significantly from those after ionomycin treatment at any of the 3 clamp voltages. Comparing the r o and κ values found by linear regression in Figure 3D from responses before ionomycin treatment ([Ca2+]o = 0 and pHo = 7) with the linear approximation (eq. 3) shown in Figure 1C (see Table 3) confirms that the regression line from response data at 3 voltages can adequately yield good approximations to both the open-circuit response and response conductance.
Both pHo and [Ca2+]o are potent modulators of the ENaC-dependent part of the response to NaCl. Figure 4A shows that at [Ca2+]o = 0, increasing pHo from 2 to 7 to 9.7 caused both r o and κ to increase maximally. However, post-ionomycin treatment increasing [Ca2+]o (Figure 4B,,C)C) caused the same pHo progression to have significantly less effect on r o and κ, until at 0.01M [Ca2+]o (Figure 4D) both r o and κ approached zero and changing pHo had no effect at all. Table 2 gives all 12 r o and κ values for each pHo (p), [Ca2+]o (c) pair obtained from the individual regression lines. We can assess the effect of increasing [Ca2+]o by comparing κ(p f, c) for a fixed value of pHo (p f) relative to κ(p f, 0). None of the values of κ(p f, 1) were statistically different from the corresponding κ(p f, 0) values, that is, c = 0.001M [Ca2+]o was not sufficient to raise [Ca2+]i to a concentration great enough to diminish the CT response at any pHo. At c = 0.005M κ(p f, 5) was significantly less than κ(p f, 0) for p f = 7 and 9.7, but not for p f = 2. At c = 0.01M κ(p f, 10) was significantly less than κ(p f, 0) at all 3 p f values, and as Figure 4D shows pHo had lost its capacity to modulate the CT response. Increasing [Ca2+]i, therefore, diminished the CT response to NaCl and, in addition, blocked the effectiveness of pHi as a modulator of ENaC.
It is important to note that at c = 0.001M [Ca2+]o post-ionomycin treatment, no significant effect was observed on the CT response at any pHo relative to control ([Ca2+]o, pre-ionomycin treatment) (Figures 2B, ,4A,4A, and B). This indicates that ionomycin by itself does not affect pHo-induced changes in the ENaC-dependent NaCl CT response. The increased influx of Ca2+ through the Ca2+-ionophore, ionomycin and the subsequent increase in TRC [Ca2+]i inhibits the pHo-induced changes in the ENaC-dependent NaCl CT response.
Figure 5 shows that with pHo and [Ca2+]o modulation of ENaC activity κ varies linearly with r o with the slope of the regression line as 0.0059±0.0004. The slope of the corresponding regression line for the ENaC modulators previously studied was also 0.0059±0.0005 (Mummalaneni et al. 2014). This indicates that for a variety of ENaC modulators that either enhance or suppress its activity, any change observed in the open-circuit response to NaCl produces the same proportional change in the response conductance. Moreover, the constancy of the proportionality between κ and r o means that neither enhancers nor suppressors affect the values of K r and δ, that is, their actions on ENaC are exerted through changes in r m only.
To obtain the maximum CT response for a given parameter pair, pHo and [Ca2+]o, that is, r m(p, c), we make use of the fact that the slope of the line in Figure 5 requires that from equation (4):
Using equation (2) with K r = 0.268M and n = 0.1M, gives r o = 0.272 r m. With this, equation (3) becomes:
We used equation (6) to fit r as a function of V for each of the 12 p–c pairs in Figure 4. Table 4 shows the r m value for each p–c pair derived from the least squares fit of the data in Figure 4. Because of the constraint imposed by the constancy of the slope of the κ versus r o plot, equation (6) indicates that as in our previous study (Mummalaneni et al. 2014), pHo and [Ca2+]o also modulated ENaC activity by modulating r m.
From the channel model expression for K r (eq. 7), for K r to remain unchanged, changes in intracellular Na+ concentration and apical membrane potential cannot occur independently. This limits the extent to which an agonist can enhance ENaC function. According to the model (Hendricks et al. 2000):
Here k = k c/k mo where k c is the Na+ dissociation rate constant between the channel cytosolic side and the cytoplasm, k mo is Na+ dissociation constant between the channel mucosal side and the stimulus solution at zero apical membrane potential, c c is intracellular Na+ concentration, K e is the thermodynamic dissociation constant between Na+ bound to ENaC and Na+ in solution, and θ is the potential normalized to RT/F across the apical membrane. K e may be expressed as either k c/f c or k mo/f mo where f c and f mo are, respectively, the Na+ association rate constant for the binding of Na+ from the cytoplasm to the channel cytosolic side, and the Na+ association rate constant for the binding of Na+ from the stimulus solution to the channel mucosal side at zero apical membrane potential. To prevent a violation of the second law of thermodynamics k c/f c must equal k mo/f mo. Thus if an agonist increases k c by some factor, it must increase k mo, f c, and f mo by the same factor. Therefore, k and K e are invariant on thermodynamic grounds. The additional empirical invariance of K r that emerges from the experiments, therefore, suggests that c c and θ cannot vary independently but are rather constrained so as to satisfy the constancy of K r. If the net result of some agonist acting on ENaC is an increase in k c (and therefore r m) which we might assume leads to further depolarization of θ then that must be accompanied by a further decrease in c c for K r to remain unchanged as experiments show it does. Constraints such as these must limit the extent to which any agonist can increase k c and, therefore, the ENaC-dependent CT response to NaCl in situ regardless of how the agonist might perform with ENaC expressed in cell culture.
Figure 6A shows the maximum response as a function of pHo (p) at fixed [Ca2+]o, (c f), that is, r m(p, c f). For each c f r m(p, c f) varies linearly with p. The lines through the points in Figure 6A are the regression lines corresponding to each c f. The slope of each regression line is a measure of the pH sensitivity (ρm) of r m for a given value of [Ca2+]o, that is, ρm(c f) = (r m/p)cf. Clearly the greater the slope is, the greater the pH sensitivity. We note that c f = 0 and c f = 0.001M yield responses with the highest pH sensitivity, at c f = 0.005M the pH-sensitivity of the response was significantly diminished and at c f = 0.01M it was essentially zero.
From the slopes of the regression lines, we can conclude that the pH sensitivity of the maximum response at c f = 0 is about 2.5 times greater than at c f = 0.005M, and 40 times greater than at c f = 0.01M. Figure 6B shows that ρm(c f) declines linearly with increasing c f. This suggests that as the TRC Ca2+ increases there is a proportionate decrease in the sensitivity of ENaC to an increase in pHi. Thus, at the unperturbed TRC Ca2+ (Figure 6A for c f = 0) and for small increases in [Ca2+]i (Figure 6A for c f = 0.001M), changes in pHi effectively regulate ENaC conductance, however, as [Ca2+]i increases above a threshold value (Figure 6A for c f = 0.005 and 0.01M), as might occur in situ due to hormone action or some other Ca2+-dependent regulatory process, increasing [Ca2+]i progressively blocks pH regulation, and then ultimately supersedes it as a key ENaC conductance regulator.
We utilized the voltage dependence of ENaC to investigate the combined effects of pH and Ca2+ on both the CT response to NaCl and the response conductance. To determine if these effects of pH and Ca2+ are specific for Na+ and ENaC, parallel studies were conducted with KCl as the salt stimulus.
The CT response to KCl was sensitive to applied lingual voltage in the same way as NaCl (Figures 1A and and3A)3A) but unlike the NaCl response, the KCl response was Bz-insensitive. This indicates that the taste transducer for K+ is also a constitutively active ion channel that is different from ENaC. Accordingly, the ion channel model (eq. 1; Figures 1B and and1C)1C) accounts as well for the variation of the KCl response with voltage as it does for the voltage dependence of the ENaC-dependent NaCl response. The KCl CT response and the ENaC-dependent NaCl CT response were continuous functions of voltage as well as of concentration for both channel types. The fit of the 0.5M KCl data yielded K r = 0.4M (eq. 1 and Figure 1B), a larger value than K r for ENaC, further indication that these are different channels. However, in both cases the fraction of the voltage dropped across the apical membrane was about 0.2. The KCl CT response also differed from the ENaC-dependent NaCl CT response in its insensitivity to changes in [Ca2+]i (Figure 3D). This indicates that unlike ENaC, the ion channel mediating KCl CT responses is not regulated by changes in TRC Ca2+.
While pHi (Figure 2A) and [Ca2+]i (Figure 2B) are individually effective regulators of ENaC activity, [Ca2+]i in turn is also a potent regulator of the pH-sensitivity of ENaC. With [Ca2+]o = 0.01M, the CT response to 0.1M NaCl post-ionomycin treatment was suppressed between pH 2 and 10.3 (Figure 4D). Thus, at high TRC [Ca2+]i ENaC conductance cannot be regulated by changes in pHi. However, changes in pHi normally regulate the channel activity under the conditions that do not perturb TRC [Ca2+]i (Figure 4A). Our results demonstrate that of the 3 parameters that are characteristic of the ENaC channel in situ: the maximum ENaC-dependent response (r m), K r, the NaCl concentration at which the open-circuit response (r o) is one-half of r m, and the fraction of the applied voltage dropped across the channels in the TRC apical membranes (δ), the ENaC sensitivity to pH and Ca2+ is due to changes in r m alone (eqs. 1 and 3). At present the precise sites in the αβγ ENaC subunits where H+ and Ca2+ exert their effect on ENaC activity are not known. Histidine residues in the cytoplasmic NH2 terminus of the α-rENaC have been suggested to be potential candidates for its pH-sensitivity (Chalfant et al. 1999). Consistent with this, Zn2+ and diethylpyrocarbonate, modification reagents for histidine residues in proteins, attenuated the CO2-induced inhibition of NaCl CT responses and the pHi-induced inhibition of apical Na+ influx in polarized fungiform TRCs (Lyall et al. 2002).
Figure 5 demonstrates that the response conductance varies linearly with the open-circuit response as changes in pH and Ca2+ have no effect on K r and δ (eq. 3). The relation between κ and r o was linear with slope equal to K rδ F/(RT(K r + n) = 0.0059±0.0004. This is exactly the slope found earlier in our investigation of various ENaC modulators (Mummalaneni et al. 2014). This suggests that most, if not all modulators (agonists and antagonists) of ENaC activity alike, fall on a common linear plot of response conductance versus open-circuit response indicating that they all exert their influence on ENaC activity by modulating r m. Analysis shows that r m is proportional to Nk c where N is the number of ENaC channels per unit area of apical membrane and k c is the rate constant for the dissociation of Na+ from ENaC into the cell interior (Hendricks et al. 2000). A hormone, capable of changing the channel density, can modulate r m in that manner. Other modulators may operate through changes in k c alone and some could affect both channel density and k c. In the principal cells of the rat renal collecting duct, ENaC is found in both the cell apical membranes and in intracellular vesicles. ENaC membrane density can be regulated by vesicle to apical membrane trafficking of ENaC (Hager et al. 2001).
The rapidly decreasing pH-sensitivity of r m as [Ca2+]o increases (Figure 6A,,B)B) was designated ρm and was quantified as the slope of the r m versus p curve for a fixed value of [Ca2+]o (c f). The ρm decreases linearly with [Ca2+]o and is essentially zero as [Ca2+]o approaches 0.01M (Figure 6B). Thus elevating [Ca2+]i essentially displaces pHi as an ENaC regulator.
In this study, pHi and [Ca2+]i were varied by changing their extracellular concentrations and, in the case of Ca2+, by also increasing its apical membrane permeability using ionomycin. However, changes in [H+]i and [Ca2+]i occur in situ. In mice, AngII type 1 receptors (AT1) are coexpressed in TRCs containing α-ENaC (Shigemura et al. 2013). AT1 is a G-protein-coupled receptor that, depending on the effecter cell type, utilizes a variety of second messenger pathways. One pathway involves inositol trisphosphate synthesis and the subsequent increase in [Ca2+]i (Higuchi et al. 2007), which presumably inhibits TRC ENaC activity (Figures 3 and and4).4). Changes in pHi in ENaC expressing TRCs occur in the presence of acidic stimuli (Lyall et al. 2002) or during osmotically induced changes in cell volume (Lyall et al. 1999). Changes in TRC volume can induce secondary changes in pHi due to the activation of pH-regulatory mechanisms, such as the basolateral Na+-H+ exchanger-1 (NHE-1) (Vinnikova et al. 2004). In addition, a subset of TRCs demonstrate a temporal relationship between [H+]i and [Ca2+]i. An increase in [H+]i (intracellular acidification) increases [Ca2+]i and a decrease in [H+]i (intracellular alkalinization) decreases [Ca2+]i (DeSimone et al. 2012a). However, changes in pHi in ENaC containing TRCs are not associated with a neural response, as these cells lack the sour taste transduction machinery (Oka et al. 2013).
An increase in ionomycin-induced [Ca2+]i can activate basolateral NHE-1 which should result in a decrease in [H+]i (intracellular alkalinization) that can increase ENaC activity and hence the magnitude of the CT response to NaCl (Lyall et al. 2002). At [Ca2+]o = 0.01M and pHo = 7, an increase in TRC [Ca2+]i almost completely inhibited the ENaC-dependent NaCl CT response (Figures 2A and and4D)4D) (DeSimone et al. 2012a). Because ENaC becomes pH-insensitive in the presence of high TRC [Ca2+]i, changes in TRC pHi induced by the activation of the basolateral membrane NHE-1 do not contribute to ENaC activity.
In our studies, a decrease in pHo or pHi inhibited and an increase in pHo and pHi enhanced TRC ENaC activity and the Bz-sensitive NaCl CT response (Lyall et al. 2002; Mummalaneni et al. 2014). In Xenopus oocytes, only the rat αβ, but not the αγ or αβγ ENaC expressed channels were inhibited by low pHo (Fyfe et al. 1999). In another study (Collier and Snyder 2009), in Xenopus oocytes expressing the rat αβγ ENaC channel, the amiloride-sensitive Na+ currents were not altered by changes in pHo over the range 8.5–6.0. However, Xenopus oocytes expressing the human αβγ ENaC channel, demonstrated an opposite effect of pHo on the amiloride-sensitive Na+ currents. Increasing pHo from 7.4 to 8.5 attenuated and decreasing pHo to 6.0 enhanced the amiloride-sensitive Na+ currents. In addition, pHo failed to alter amiloride-sensitive Na+ currents when human α and β ENaC subunits were coexpressed with the rat γ ENaC subunit (Collier and Snyder 2009). This suggests that pHo regulates ENaC in a species-specific manner. Sequence differences in the γ-ENaC subunit and also in the α-ENaC subunit may account for the species variations between the sensitivity of rat and human ENaC to pHo. Furthermore, differences in the proteolytic cleavage states of ENaC in different tissues and species may also contribute to differences in the sensitivity of ENaC to pHo.
Human TRCs express an additional ENaC subunit, δENaC (Huque et al. 2009), which may be activated by acidic pH (Ji and Benos 2004; Ji et al. 2012). ENaC in rat TRCs (Lyall et al. 2002), mouse kidney cortical collecting cells (Gu 2008) and principal cells of frog skin (Lyall et al. 1994) is inhibited by a decrease in pHi. This is consistent with the observation that rat αβγ and αγ ENaC channels are known to be inhibited by intracellular acidification (Palmer and Frindt 1987). Further studies are needed to resolve the differences between human and rat ENaC regulation by pHo, pHi, and [Ca2+]i.
This work was supported by the National Institute on Deafness and other Communication Disorders (NIDCD) [DC-005981 and DC-011569 V.L.), the Jeffress Memorial Trust [J-1031 to V.L.] and the Korea Food Research Institute (KFRI) [E094101 and E0111501 to M.R.].