Temperature Influences Chorda Tympani Nerve Responses to Sweet, Salty, Sour, Umami, and Bitter Stimuli in Mice

Abstract

Temperature profoundly affects the perceived intensity of taste, yet we know little of the extent of temperature’s effect on taste in the peripheral nervous system. Accordingly, we investigated the influence of temperature from 23 °C to 43 °C in 4 °C intervals on the integrated responses of the chorda tympani (CT) nerve to a large series of chemical stimuli representing sweet, salty, sour, bitter, and umami tastes in C57BL/J6 mice. We also measured neural responses to NaCl, Na-gluconate, Na-acetate, Na-sulfate, and MSG with and without 5 µM benzamil, an epithelial sodium channel (ENaC) antagonist, to assess the influence of temperature on ENaC-dependent and ENaC-independent response components. Our results showed that for most stimuli (0.5M sucrose, glucose, fructose, and maltose; 0.02M saccharin and sucralose; 0.5M NaCl, Na-gluconate, Na-acetate, Na-sulfate, KCl, K-gluconate, K-acetate, and K-sulfate; 0.05M citric acid, acetic acid, and HCl; 0.1M MSG and 0.05M quinine hydrochloride: QHCl), CT response magnitudes were maximal between 35 °C and 39 °C and progressively smaller at cooler or warmer temperatures. In contrast, the weakest responses to NH₄Cl, (NH₄)₂SO4, and K-sulfate were at the lowest temperature, with response magnitude increasing monotonically with increasing temperature, while the largest responses to acetic acid were at the lowest temperature, with response magnitude decreasing with increasing temperature. The response to sweet and umami stimuli across temperatures were similar reflecting the involvement of TRPM5 activity, in contrast to bitter stimuli, which were weakly affected by temperature. Temperature-modulated responses to salts and acids most likely operate through mechanisms independent of ENaC and TRPM5.

We know from experience that the serving temperature of foods and beverages has a powerful influence on taste especially its intensity, which in turn modulates food choice and consumption. Two generally accepted examples are the increased sweetness of ice cream as it melts and the increased bitterness of beer as it warms. Although it is generally accepted that changes in solution temperature activate trigeminal nerve fibers in the oral cavity (Lundy and Contreras 1994; Pittman and Contreras 1998), solution temperature modulates responses of peripheral-gustatory neurons to taste stimuli and can act as a stimulus alone in these gustatory neurons (Ogawa et al. 1968; Lundy and Contreras 1999; Breza et al. 2006). This latter point indicates that temperature’s influence on taste perception begins at a peripheral level.

Indeed, a mechanistic explanation of how temperature influences sweet–taste sensations starts at the taste-receptor cell, located within taste buds. Specifically, TRPM5, a temperature-sensitive member of the transient receptor family of ion channels, is expressed on taste cells that express the sucrose receptor (T1R2/T1R3) and is thought to be the molecule that modulates temperature-dependent responses as seen in neurophysiological recordings of isolated taste nerves. Talavera et al. (2005) showed that responses to sweeteners in the mouse chorda-tympani nerve (a taste nerve that innervates fungiform papillae on the anterior two-thirds of the tongue) were modulated by solution temperature (15–35 °C). In contrast, responses to sweeteners in mice lacking TRPM5, where unaffected by changes in solution temperature, indicating that TRPM5 is the locus for temperature modulation of sweet taste (Talavera et al. 2005).

Although TRPM5 is also present in cells expressing receptors for bitter or umami, direct evidence in support of TRPM5’s effect on modulation of bitter and umami taste is lacking. Furthermore, despite numerous lines of evidence that sucrose, salt, sour, and bitter taste neurons are modulated by temperature in the peripheral nervous system of rats (Ogawa et al. 1968; Lundy and Contreras 1999; Breza et al. 2006), the influence of temperature on taste responses representing the 5 basic taste qualities has not been reported in the peripheral nervous system of mice. Most electrophysiological investigations in the taste field are employed at room temperature. In rats, the majority taste stimuli are severely reduced if not abolished below room temperature (Breza et al. 2006). Additionally, we (Breza et al. 2006; Lu et al. 2012) and Lyall et al. (2004) have shown that responses to some taste stimuli increase with increasing temperature, whereas responses to other taste stimuli show an inverted U-shaped function. However, much is still unknown on how temperature influences taste intensity within (e.g., sucrose vs. sucralose or HCl vs. acetic acid) and across separate taste qualities in the peripheral nervous system.

To establish a relatively complete picture of temperature’s influence on the CT nerve response to solutions representing the 5 basic taste stimuli in mice, we adopted a stimulus delivery system with which we can set and maintain (within 0.1 °C) the temperature of stimulus solutions, between a low at 23 °C (room temperature) and a high at 43 °C that begins to stifle taste. With this system, we investigated the influence of temperature from 23 °C to 43 °C on the integrated responses of the chorda tympani (CT) nerve to a large series of chemical stimuli in C57BL/J6 mice. We also examined temperature’s effect on ENaC-dependent and ENaC-independent response components to stimulation with 4 different sodium salts and monosodium glutamate (MSG).

Methods

Subjects

Adult male C57BL/J6 (Jackson Laboratory) mice weighing 26–32g at the start of the experiments were housed individually in transparent plastic cages in a temperature-controlled colony room (22–24 °C) and maintained on a 12:12-h light-dark cycle with lights on at 7:00 AM. All the animals were habituated to the animal facility for at least 1 week before CT recording. All mice had ad libitum access to Purina Rat Chow (no. 5001) and de-ionized water unless otherwise indicated. The Institutional Animal Care and Use Committee at Florida State University approved all procedures. We used a total of 39 mice across 5 experiments and used 7 mice to determine the temperature-response function for each chemical stimulus.

Electrophysiological recordings of taste responses of the mouse CT

The mice were anesthetized with intraperitoneal administration of ketamine (30mg/kg body weight) followed by urethane (1.2g/kg). Supplemental urethane injections were given to maintain a deep level of anesthesia without reflex response to foot pinch. The mice were tracheotomized with PE50 tubing, and secured in a nontraumatic head holder (Model 926B Mouse Nose/Tooth Bar Assembly, David Kopf Instruments). Using a mandibular approach, the right CT branch of the facial nerve was exposed and transected where it enters the tympanic bulla. The CT nerve was desheathed, and placed on a platinum wire electrode (positive polarity) and the entire cavity was then filled with high quality paraffin oil (VWR) to isolate the nerve signal from ground and maintain nerve integrity. An indifferent electrode (negative polarity) was attached to the skin overlying the cranium with a tinned-copper alligator clip. Neural activity was differentially amplified (×10000; A-M Systems, bandpass 300–5000 Hz), observed with an oscilloscope, digitized with waveform hardware and software (Spike 2; Cambridge Electronic Design), and stored on a computer for off-line analysis.

Chemical stimuli and solution delivery

We included 23 stimuli (6 sweet, 12 salt, 3 sour, 1 bitter, 1 umami) each at a single concentration to keep the project manageable. We used stimulus concentrations that could be analyzed over a large temperature range, because cool temperatures suppress the magnitude of peripheral taste responses in rats (Nakamura and Kurihara 1988; Lundy et al. 1997; Talavera et al. 2005; Breza et al. 2006). We used 0.5M sucrose, glucose, fructose, and maltose and 0.02M saccharin and sucralose, all highly preferred by B6 mice (Bachmanov et al. 2001; Inoue et al. 2001). Except for sucralose, these are the same stimuli and stimulus concentrations used by Talavera et al. (2005) in demonstrating the importance of TRPM5 on thermal enhanced sweet taste mice. Because the mouse CT nerve is poorly responsive to sugars and artificial sweeteners at cool temperatures (Talavera et al. 2005), we kept stimulus temperature between 23 °C and 43 °C (23 °C–27 °C–31 °C–34 °C–37 °C–40 °C–43 °C) for all stimuli even though the CT nerve retains some responsiveness to salty and sour stimuli delivered below room temperature conditions. Before embarking on the current study, we conducted some preliminary experiments and found CT nerve responses to all stimuli to be substantially reduced and more variable at stimulus temperatures above 43 °C. Because of this we chose 43 °C as the upper limit consistent with our prior study with mice using 44 °C as the highest temperature (Lu et al. 2012). We used 0.5M sodium and potassium salts bound to 1 of 4 anions (chloride, acetate, sulfate, and gluconate), as well as 0.5M CaCl₂, MgCl₂, NH₄Cl, and (NH₄)₂SO₄. We used high salt concentrations so that we could analyze salt responses with large anions, which are known to suppress the nonselective pathway in rats (Ye et al. 1991; Lundy and Contreras 1999; Breza and Contreras 2012b) and hamsters (Rehnberg et al. 1993). Sodium chloride is the archetypal stimulus for saltiness and more highly preferred than the other salts. However, mice find NaCl solution palatable only at low concentrations (Gannon and Contreras 1995), even though CT nerve responses increase with stimulus concentration up to 1.0M NaCl. We chose 0.05M acetic acid, citric acid, and HCl, the same 3 stimuli used by Horio et al. (2011) in demonstrating the importance polycystic kidney disease-like (PKD) ion channels to sour taste in mice. Finally, we used 0.1M MSG and 0.05M quinine hydrochloride as prototypical stimuli for umami and bitter taste, respectively.

The tongue was slightly extended and held in place by withdrawing the thread, sutured to the tongue’s ventral surface, and securing it to the surgical table. Solutions were presented to the anterior tongue at a constant flow rate (50 μL/s) and controlled temperature by an air-pressurized 32-channel commercial fluid-delivery system and heated perfusion cube (OctaFlow Multi-function Multi-valve Perfusion System, ALA Scientific Instruments). All solutions were made from reagent-grade chemicals and dissolved in a dilute salt mixture (0.015M NaCl, 0.022M KCl, 0.003M CaCl₂, and 0.0006M MgCl₂) of artificial saliva (AS). We recorded CT nerve responses to 10s applications of each taste stimulus. Each stimulus was followed by a rinse of AS for 150s to ensure that nerve activity returned to stable baseline levels. Amplified nerve activity was monitored on-line, digitized using Spike II, and integrated with a root mean square (RMS) calculation with a time constant of 200ms. The average baseline neural activity immediately preceding each chemical stimulus presentation was subtracted from the integrated response resulting from the 10s stimulus to calculate the area under the curve (AUC).

Data analysis

All data are presented as group means ± SEM. Differences between the responses at different temperatures were performed using 1-way repeated measure ANOVA followed by post hoc pairwise comparisons using contrast coefficients. All analyses were performed using Statistical Program for Social Sciences software (SPSS 13.0) with statistical significance accepted when P < 0.05.

Experiments: Subjects and taste protocol

For each mouse, we recorded CT nerve responses to solutions representing a single taste quality, at 6 different temperatures (in 4 °C intervals) beginning with the lowest and increasing to the highest temperature (ranging from 23 °C to 43 °C). The tongue was pre-adapted to each stimulus temperature with 60s of AS preceding stimulation. The AUC responses to each stimulus were normalized to the average response to 0.1M NH₄Cl at 35 °C before and after each temperature series for each animal. The response to 0.1M NH₄Cl at 35 °C was also used to evaluate the stability of the preparation. A recording was considered stable when the responses to 0.1M NH₄Cl at 35 °C at the beginning and the end of each stimulation series deviated by no more than 15%. Only responses from stable recordings were used for data analysis. For sodium salts and MSG, we tested CT nerve responses at each temperature 3 times, the first to salt alone, the second to the salt mixed with 5 µM benzamil, and the third to the salt alone.

Results

Figure 1 shows integrated CT-nerve recordings to NH₄Cl (A) and sucrose (B) at 23 °C, 27 °C, 31 °C, 35 °C, 39 °C, and 43 °C. As shown in the Figure 1, NH₄Cl responses increase with increasing temperature, whereas sucrose responses have a quasi-inverted U shape.

Figure 1.

Integrated CT–nerve responses to NH₄Cl (top) and sucrose (bottom) at 23 °C, 27 °C, 31 °C, 35 °C, 39 °C, and 43 °C. NH₄Cl responses increase with increasing temperature, whereas sucrose responses have a quasi-inverted U shape.

As shown in Figure 2, the temperature–response functions were similar for the 6 sweeteners with some notable exceptions. For all 6 stimuli, response magnitudes increased with temperature, peaked between 31 °C and 39 °C and declined to a variable degree at higher stimulus temperatures depending on the stimulus. The function for the 4 sugars were similar to each other but differed somewhat from that of the 2 artificial sweeteners, which were similar to each other. The temperature–response function for the artificial sweeteners was shifted to the left (perhaps sucralose more than saccharin), peaking at a slightly lower temperature and dropping more precipitously at the 2 higher temperatures than that for the 4 sugars.

Figure 2.

The whole nerve responses of the CT to lingual application of 0.5M sucrose (A), 0.5M glucose (B), 0.5M fructose (C), 0.5M maltose (D), 0.02M saccharin (E), and 0.02M sucralose (F) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05, **P < 0.01, ***P < 0.001.

A 1-way, repeated measures ANOVA showed that the CT nerve responses to all 6 sweeteners varied significantly as a function of stimulus temperature [sucrose: F(5, 30) = 7.112, P < 0.001; glucose: F(5, 30) = 35.657, P < 0.001; fructose: F(5, 30) = 14.099, P < 0.001; maltose: F(5, 30) = 21.364, P < 0.001; saccharin: F(5, 30) = 16.991, P < 0.001; sucralose: F(5, 30) = 26.165, P < 0.001]. As depicted in Figure 2A–D, the average sugar responses were maximal approximately at 35 °C, with post hoc comparison tests showing that the average sugar response at 35 °C was significantly greater than that at 23 °C, 27 °C, and 31 °C. As depicted in Figure 2E, F, the average response to saccharin at 35 °C was significantly greater than that at 23 °C (P < 0.001), 27 °C (P < 0.01), and 43 °C (P < 0.001), while the average response to sucralose at 35 °C was significantly greater than that at 23 °C (P < 0.05), 39 °C (P < 0.001), and 43 °C (P < 0.001).

Figure 3 shows the results from the CT nerve recordings to stimulation with the 4 sodium salts at 6 different temperatures. Because we also examined the effect of benzamil antagonism of ENaC, the results from each salt are described by 3 separate temperature–response functions—1 for the entire salt response, and 1 each for ENaC-dependent (ENaC+) and ENaC-independent (ENaC–) component responses. The ENaC-independent function was obtained experimentally from the average responses to the salt mixed with benzamil. The ENaC-dependent function was obtained by subtracting the responses to the salt-benzamil mixture from the responses to the salt alone.

Figure 3.

The whole nerve responses of the CT to lingual application of 0.5M NaCl (A), 0.5M Na-acetate (B), Na-sulfate (C), and 0.5 M Na-gluconate (D) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C for salt alone and ENaC-independent response functions, *P < 0.05, **P < 0.01, ***P < 0.001.

As shown in Figure 3, the temperature–response functions were similar for the 4 salts with some subtle differences. For all 4 stimuli, response magnitudes increased with stimulus temperature [NaCl: F(5, 30) = 37.206, P < 0.001; Na-gluconate: F(5, 30) = 34.155, P < 0.001; Na-acetate: F(5, 30) = 23.039, P < 0.001; Na-sulfate: F(5, 30) = 38.860, P < 0.001], peaked at 39 °C and declined modestly at 43 °C. In contrast, the ENaC-dependent response component remained relatively stable across temperature and was always a small portion (20–30%) of the overall salt response. The ENaC-independent response component paralleled the overall salt response function; it was largely responsible for the differences in CT nerve responses across stimulus temperature. This was reflected with post hoc comparison tests showing that the average salt responses at 35 °C were significantly greater than those at lower temperatures for both the salt alone and ENaC-independent response functions to all 4 salts.

Figure 4 shows the results from the CT nerve recordings to stimulation with the 4 potassium salts at 6 different temperatures. For all 4 stimuli, response magnitudes varied significantly [KCl: F(5, 30) = 32.725, P < 0.001; K-glu: F(5, 30) = 27.063, P < 0.001; K-ace: F(5, 30) = 17.502, P < 0.001; K-sulf: F(5, 30) = 49.976, P < 0.001] with stimulus temperature. In general, response magnitudes increased with stimulus temperature, peaked between 35 °C and 39 °C and declined modestly at 43 °C. The exception to this pattern was the temperature–response function to K-sulfate where response magnitudes continued rising to the highest temperature at 43 °C.

Figure 4.

The whole nerve responses of the CT to lingual application of 0.5M KCl (A), 0.5M K-acetate (B), 0.5M K-sulfate (C), and 0.5M K-gluconate (D) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05, **P < 0.01, ***P < 0.001.

The results from the CT nerve recordings to stimulation with CaCl₂ and MgCl₂ at 6 different temperatures are shown in Figure 5 and those to stimulation with NH₄Cl and (NH₄)₂SO₄ are shown in Figure 6. For all 4 stimuli, response magnitudes varied significantly [CaCl2: F(5, 30) = 20.329, P < 0.001; MgCl2: F(5, 30) = 16.848, P < 0.01] with stimulus temperature. The temperature–response functions for CaCl₂ and MgCl₂ were similar to each other, with response magnitude increasing with stimulus temperature, peaking at 39 °C and declining modestly at 43 °C. The temperature–response functions for NH₄Cl and (NH₄)₂SO₄ were also similar to each other, but with response magnitude increasing progressively as temperature increased from the lowest to highest temperature [NH₄Cl: F(5, 30) = 42.144, P < 0.001; (NH4)2SO4: F(5, 30) = 47.705, P < 0.001].

Figure 5.

The whole nerve responses of the CT to lingual application of 0.5M CaCl₂ (A) and 0.5M MgCl₂ (B) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6.

The whole nerve responses of the CT to lingual application of 0.5M NH₄Cl (A) and 0.5M (NH4)2SO4 (B) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05, **P < 0.01, ***P < 0.001.

The results from the CT nerve recordings to stimulation with the 3 acids at 6 different temperatures are shown in Figure 7. Unsurprisingly, response magnitudes varied significantly (citric acid: F(5, 30) = 16.740, P < 0.001; HCl: F(5, 30) = 18.450, P < 0.001; acetic acid: F(5, 30) = 38.409, P < 0.001) with stimulus temperature for all 3 acids, but the temperature–response function for acetic acid was unexpected. Instead of increasing with stimulus temperature, response magnitude decreased progressively with increasing stimulus temperature with the largest response at 23 °C and the smallest response at 43 °C. In contrast, response magnitude increased progressively with increasing stimulus temperature peaking at 43 °C for citric acid and at 39 °C for HCl. It should be noted that the largest responses to HCl were twice as big as the largest responses to citric and acetic acid.

Figure 7.

The whole nerve responses of the CT to lingual application of 0.05M citric acid (A), 0.05M acetic acid (B), and 0.05M HCl (C) at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05.

The results from the CT nerve recordings to stimulation with 0.1M MSG at 6 different temperatures are shown in Figure 8. Because we also examined the effect of benzamil antagonism of ENaC, the results are described by 3 separate temperature–response functions—1 for the entire MSG response, and 1 each for ENaC-dependent (ENaC+) and ENaC-independent (ENaC−) component responses. The ENaC-independent function was obtained experimentally from the average responses to MSG mixed with benzamil. The ENaC-dependent function was obtained by subtracting the responses to the salt-benzamil mixture from the responses to MSG alone.

Figure 8.

The whole nerve responses of the CT to lingual application of 0.1 M MSG at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C for MSG alone and ENaC-independent response functions, *P < 0.05, **P < 0.01, ***P < 0.001.

As shown in Figure 8, response magnitude increased progressively with stimulus temperature [F(5, 30) = 41.532, P < 0.001], peaked at 35 °C and declined modestly between 39 °C and 43 °C. In contrast, the ENaC-dependent response component remained relatively stable across temperature and was always a small portion (20–30%) of the overall MSG response. The ENaC-independent response component paralleled the MSG response function [F(5, 30) = 25.124, P < 0.001]; it was largely responsible for the differences in CT nerve responses across stimulus temperature. This was reflected with post hoc comparison tests showing that the average MSG response at 35°C was significantly greater than those between 23 °C –31 °C and at 43 °C for both the MSG alone and ENaC-independent response functions.

As shown in Figure 9, response magnitude increased gradually with stimulus temperature [F(5, 30) = 17.114, P < 0.01], peaked at 39 °C and declined modestly at 43 °C. In general, the CT nerve responded less strongly to QHCl than any another chemical stimulus used in this investigation.

Figure 9.

The whole nerve responses of the CT to lingual application of 0.05M QHCl at different stimulating temperatures in C57BL/J6 mice. Significantly different from 35 °C, *P < 0.05.

Discussion

Temperature has a profound influence on taste perception in humans (Cruz and Green 2000) and modulates consummatory behaviors in rats (Torregrossa et al. 2012). Despite a longstanding interest by a variety of research specialists on the topic, much is still unknown about the molecular and physiological mechanisms underlying temperature’s influence on different taste modalities. Much of the background literature in the peripheral nervous system is dominated by electrophysiological studies using rats as the model system. More recently, however, the effort has shifted to mice because of their usefulness to address underlying genetic mechanisms. Few studies in mice have investigated the influence of temperature on taste nerve responses, and a thorough description of responses to the 5 basic taste stimuli across a dynamic range of temperatures has not yet been employed. In the present study, we performed a comprehensive examination of temperature’s influence on CT nerve responses to a wide range of chemical stimuli representing the 5 basic taste stimuli in mice. It has been reported that temperature’s effect varied across the basic taste stimuli, but the extent of that variation has not yet been elucidated. In the present study, we found that temperature not only differentially affected taste modalities, but also differentially affected exemplars within each modality, especially for sweet, salt, and sour stimuli where we focused most of our attention. These temperature-mediated differences in CT nerve responses are most likely due to underlying molecular mechanisms on the plasma membrane of taste-receptor cells. Some of the molecular mechanisms underlying temperature modulation of taste have been identified (i.e., TRPM5), whereas others are controversial (TRPV1 in salt taste) or completely unknown, such as the mechanism(s) of temperature sensitivity of sour taste.

It is well established that sweetness depends on stimulus concentration and based upon a growing amount of electrophysiological evidence on temperature as well. Sweeteners bind to T1R2/T1R3 receptors and activate downstream TRMP5 cation channels. Using a heterologous expression system, Talavera et al. (2005) showed that TRPM5-mediated inward currents increase with increasing temperatures. This results in more robust cell depolarization and a stronger afferent signal, as shown in CT nerve recordings to sweeteners. In a recent study (Lu et al. 2012), we confirmed the findings from Talavera et al. (2005) showing that the CT nerve responded with greater magnitude to sucrose and saccharin as stimulus temperature increased from 23 °C to 35 °C. In that study we also used higher stimulus temperatures and discovered that response magnitude declined as temperature increased from 38 °C to 44 °C resulting in an inverted U-shaped function across temperature for 0.1, 0.3, 0.6, and 1.0M sucrose (Lu et al. 2012). Additionally, it should also be noted that responses to 0.5M sucrose were also quite similar to those reported by Ohkuri et al. (2009), where responses to sucrose peaked at 35 °C and decreased at higher temperatures, despite huge differences quantification of CT nerve responses (10s that included initial phasic responses of CT-nerve activity in the current study compared with the Ohkuri study, which analyzed 25s of data ignoring phasic responses).

In the present study, we replicated our observations on sucrose and saccharin and expanded it to include observations on 3 other sugars (glucose, fructose, maltose) and 1 additional artificial sweetener (sucralose). We discovered that the temperature-dependent responses for all 6 stimuli were similar exhibiting a quasi-inverted U-shaped function. Except for sucralose, response magnitude was at its nadir at the lowest temperature. Response magnitude increased steadily as temperature rose from 23 °C to about 35°C, and then at higher temperatures, response magnitude took a downward slide. In fact, this downward slide was pronounced for saccharin and sucralose and modest for the 4 sugars. The 4 sugars had similar temperature-response functions across the full temperature range even though response magnitude differed with sucrose > fructose > maltose > glucose as shown previously (Talavera et al. 2005). The saccharin response at 43 °C was similar to that at 23 °C, while the sucralose response at 43 °C was significantly lower than that at 23 °C. Clearly, high stimulus temperatures compromise CT nerve responses to the 2 artificial sweeteners more than to the 4 sugars, which may reflect the differential effects on the sucrose receptor (T1R2/T1R3), and/or a T1R independent mechanism, such as glucose transporters (Damak et al. 2003; Toyono et al. 2011; Yee et al. 2011).

With respect to saltiness, we found that CT nerve responses to all 12 salts were dependent on stimulus temperature. Except for the 2 ammonium salts, the temperature-dependent response functions for Na+, K+, Ca+, and Mg+ salts responses were similar within and across anion species. Responses were maximal at 39 °C and progressively smaller as temperature dropped or rose from this adapted temperature. This pattern was especially evident for the 4 sodium salts with a distinct peak response at 39 °C and sharp declines on both sides of the temperature scale. In contrast, response magnitude increased monotonically from lowest to highest temperature for both NH₄Cl and (NH₄)₂SO₄. This replicates our previous finding with NH₄Cl (Lu et al. 2012) and extends to (NH₄)₂SO₄ in the present study. While the temperature-dependent response functions are similar, CT nerve responses to NH₄Cl were greater than those to (NH₄)₂SO₄, presumably because anion size influences response magnitude. In fact, this anion effect was so pronounced that despite having twice the number of cations in (NH₄)₂SO₄, responses were still less than those to NH₄Cl. This pronounced anion effect was also apparent in K+ responses and ENaC-independent Na+ responses, consistent with our studies as well as others in rats (Elliott and Simon 1990; Ye et al. 1991; Lundy and Contreras 1997; Breza and Contreras 2012b), and hamsters (Rehnberg et al. 1993) demonstrating that the function of the ENaC-independent pathway for salts is a highly conserved mechanism across several species of rodents.

In our hands, responses to NaCl peaked at 39 °C, which is in contrast to studies by Ohkuri et al. (2009) in mice and Nakamura and Kurihara (1988) in rats where responses to NaCl peaked at 30 °C. There are, however, several methodological differences between these studies exist that must be considered. Foremost, we analyzed 10s of CT-nerve data, which included initial transients, whereas Ohkuri et al. (2009) and Nakamura and Kurihara (1988) measured nerve responses that ignored initial transients. Specifically, Ohkuri et al. (2009) applied solutions for 30s and measured 25s of CT-nerve data 5s after the start of the neural response and Nakamura and Kurihara (1988) applied solutions for 40s and measured 20s of tonic CT-nerve responses 20s after the start of the neural response. Secondly, in the present experiment, AS served as the rinse and solvent for all taste solutions, whereas Ohkuri et al. (2009) and Nakamura and Kurihara (1988) used deionized water.

It is well known that sodium-mediated salt taste depends upon ENaC-dependent and ENaC-independent response components. We used the selective ENaC antagonist, benzamil, to separate the 2 as a function of stimulus temperature. We found that the ENaC-dependent component was a small portion of the overall response (probably due to the stimulus concentration) for all 4 sodium salts and was temperature-independent. In contrast, the ENaC-independent component was large component of the integrated response and temperature-dependent, consistent with rat CT-nerve recordings to high salt concentrations (Lundy and Contreras 1997), particularly at elevated temperatures. The effects of temperature on salt responses are also consistent with single-neuron recordings in the rat geniculate ganglion (Sprague Dawley), where responses to NaCl in ENaC-mediated NaCl-specialist neurons were virtually unaffected by changes in adapted temperature, whereas NaCl responses in nonselective generalist neurons were profoundly affected by adapted temperature (Lundy and Contreras 1999; Breza et al. 2006). These data contrast with those from Nakamura and Kurihara (1988), a CT-nerve study in Wistar rats, where responses to NaCl affected the amiloride-sensitive component of the response. Importantly, as mentioned above, Nakamura and Kurihara (1988) analyzed only tonic CT-nerve responses 20s after stimulus onset, whereas we analyzed the entire 10s response, which included the phasic portion of the response.

Since studies with rodents suggest that in addition to a unique taste, MSG may also evoke NaCl-like taste (Yamamoto et al. 1988), we used benzamil to block the ENaC channels to eliminate the interference effects of sodium ions on the temperature dependence of MSG. Results showed that 0.1M MSG + benzamil-evoked CT activity was temperature dependent. Interestingly, the results of MSG with and without ENaC blockade closely resemble those to natural sugars, despite the remaining influence of ENaC-independent responses to sodium ions.

Among the 5 basic tastes, sour, which is associated with acid stimuli, is the least understood. We chose to examine the influence of temperature on an equimolar concentration (0.05M) of HCl, citric acid, and acetic acid, as these acids are effective sour stimuli at these concentrations at our temperature range (Horio et al. 2011), but are quite different from another in terms of solution pH (HCl = 1.3, citric acid = 2.24, and acetic acid = 3.03). Importantly, acid responses from the rat CT neurons are not directly correlated to solution pH, because weak acids are more effective stimuli than their pH would suggest (Lyall et al. 2001; Breza and Contreras 2012a). A decrease in the intracellular pH appears to be the proximate stimulus in sour taste transduction (Lyall et al. 2001), but weak acids, such as acetic acid and carbonic acid, are more effective (at the same solution pH) at decreasing intracellular pH than HCl, a strong mineral acid (Lyall et al. 2001).

Responses to citric acid increased with increasing temperature, while acetic acid decreased with increasing temperature. As for HCl, CT responses maximized at 39 °C, and then decreased at higher temperature, which is generally consistent with our CT responses to salts through the benzamil-insensitive pathway. The mechanism by which the acetic acid evoked CT activity decreased with increasing temperature is not clear at present, but it may reflect the magnitude and/or rate of decrease in intracellular pH, which may be near saturation at 0.05M HCl, and citric acid may have additional physiological effects extracellularly, which may explain why our results to acetic acid showed opposing response curves than those to HCl and citric acid. This may imply that the effects of temperature may have intracellular and extracellular sites of action or multiple receptor mechanisms, and warrants future investigation. It could also imply that the mechanisms of sour–taste transduction in rats and mice differ, since responses to acids in whole nerve and single cell recordings (Lyall et al. 2001; Breza and Contreras 2012a) of rats are similar when matched for concentrations, despite differences in solution pH.

In humans, the bitterness of caffeine is altered by temperature, but is less dependent compared with the other 4 tastes (Green and Frankmann 1987). According to our results, CT responses to QHCl were also dependent on temperature, but less than the salts and sweeteners. Although we’ve reported greater effects of temperature on QHCl in single neurons from the geniculate ganglion (Lundy and Contreras 1999; Breza et al. 2006) than can be ascertained by the current whole-nerve study, the concentrations and temperature ranges between the 2 studies differed by more than a factor of ~2.5. Specifically, in the current study, we delivered 0.05M QHCl to the tongue at 23–43 °C, whereas in single unit studies, we delivered 0.02M QHCl to the tongue at 10–40 °C. It should be noted that the greatest effect of temperature on QHCl responses in the rat geniculate ganglion was when the temperature fell below 25 °C (Breza et al. 2006)—the effects of temperature on QHCl between both studies are similar from 23 °C to 43 °C. Because the mouse CT is not as sensitive to bitters like it is for salts and sweeteners, it is difficult to draw conclusions about what these effects mean for bitter taste quality. We chose to use an ionic bitter stimulus (QHCl) as the representative bitter stimulus, because it has been used in numerous CT nerve recordings. Clearly, ionic bitters and nonionic bitters are not treated equally in awake behaving hamsters (Frank et al. 2004). Considering the multitude of bitter receptors and that many bitter stimuli are capable of activating central gustatory neurons (Lemon and Smith 2005), it will be interesting to discover whether responses to ionic bitter stimuli (such as QHCl) and nonionic stimuli (such as cycloheximide) are differentially modulated by temperature in the glossopharyngeal nerve of the mouse.

In summary, the present study indicated that all the 5 basic tastes peripheral perception were dependent on temperature. The results presented herein provide a much richer appreciation for the effects that temperature has on taste intensity within and between taste qualities. The CT responses to the majority of taste stimuli are higher at moderate temperatures (between 35 °C and 39 °C) than at colder or warmer temperatures in C57BL/J6 mice. Sodium responses that were dependent on ENaC were virtually unaffected by temperature. Although most salts through the ENaC-independent pathway exhibited similar quasi inverted u-shaped functions across the temperature range, responses to ammonium salts increased linearly with increasing temperature, suggesting that perhaps ammonium salts may act on another unidentified salt receptor/ion channel. Perhaps the most perplexing response characteristics rest with sour responses. Responses to acetic acid, the weakest of the 3 acid stimuli, decreased with increasing stimulus temperature, whereas more acidic stimuli (in terms of solution pH) increased with increasing stimulus temperature. Although the mechanisms of temperature’s effects on sour taste are unknown, the differential responses to weak and strong acids may reflect differences in intracellular and extracellular pH or animal species.

The possible mechanisms underlying the temperature dependence features of the taste perceptions for sweet stimuli most likely involve TRPM5. Although it has been proposed that TRPM5 may underlie temperature effects to bitter and umami stimuli, bitter stimuli have shown to inhibit TRPM5 function (Talavera et al. 2008), which does not propose a strong argument for the role of TRPM5 in enhancing bitter responses by warming. In contrast, the effects of temperature on MSG closely resemble those to natural sugars and this effect may, in part, reflect TRPM5 activity. Clearly, much of what we know regarding the effects of temperature on taste intensity within and between taste qualities is unresolved, as the mechanisms for temperatures effects on sweet (natural vs. artificial), salty (independent of ENaC), sour (weak and strong acids), bitter (ionic vs. nonionic), and umami have not been fully elucidated. Increasing our knowledge of how temperature interacts with taste quality and intensity has the potential to impact our knowledge of mechanisms involved in diet selection and the taste experience.

Funding

This work was supported by The National Institute on Deafness and Communication Disorders (DC-004785).