Chemical Senses Advance Access originally published online on March 30, 2005
Chemical Senses 2005 30(4):299-316; doi:10.1093/chemse/bji025
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Contribution of 
-Gustducin to Taste-guided Licking Responses of Mice
1 Department of Biological Sciences, Barnard College, Columbia University, 2 Department of Physiology and Biophysics, Mount Sinai College of Medicine, 3 Howard Hughes Medical Institute and 4 Department of Psychology, University of Florida, USA
Correspondence to be sent to: John I. Glendinning, Department of Biological Science, Barnard College, Columbia University, 3009 Broadway, New York, NY 10027, USA. e-mail: jglendinning{at}barnard.edu
 |
Abstract |
|---|
Top
Abstract
IntroductionWe examined the necessity of
-gustducin, a G protein -subunit expressed in taste cells, to taste-mediated licking responses of mice to sapid stimuli. To this end, we measured licking responses of -gustducin knock-out (Gus–/–) mice and heterozygotic littermate controls (Gus+/–) to a variety of ‘bitter’, ‘umami’, ‘sweet’, ‘salty’ and ‘sour’ taste stimuli. All previous studies of how Gus–/– mice ingest taste stimuli have used long-term (i.e. 48 h) preference tests, which may be confounded by post-ingestive and/or experiential effects of the taste stimuli. We minimized these confounds by using a brief-access taste test, which quantifies immediate lick responses to extremely small volumes of sapid solutions. We found that deleting -gustducin (i) dramatically reduced the aversiveness of a diverse range of ‘bitter’ taste stimuli; (ii) moderately decreased appetitive licking to low and intermediate concentrations of an ‘umami’ taste stimulus (monosodium glutamate in the presence of 100 µM amiloride), but virtually eliminated the normal aversion to high concentrations of the same taste stimulus; (iii) slightly decreased appetitive licking to ‘sweet’ taste stimuli; and (iv) modestly reduced the aversiveness of high, but not low or intermediate, concentrations of NaCl. There was no significant effect of deleting -gustducin on licking responses to NH4Cl or HCl.
Key words: taste, -gustducin, brief-access taste test, knock-out mice
|
Introduction |
|---|
In mammals, the ingestive response to sweeteners, amino acids and many ‘bitter’ compounds is initiated by the interaction of chemical compounds with G-protein-coupled receptors (GPCRs) on the apical membrane of taste cells (Gilbertson et al., 2000
; Glendinning et al., 2000). Two families of GPCRs are known to mediate this detection process—the T2Rs and T1Rs. A few T2Rs have been shown to respond selectively to compounds that elicit ‘bitter’ taste sensations in humans (Adler et al., 2000; Chandrashekar et al., 2000; Bufe et al., 2002). The T1Rs, which appear to function predominantly as heterodimers, are activated by amino acids and sweeteners (natural and artificial). The heterodimer of T1R2+T1R3 responds selectively to compounds that elicit ‘sweet’ taste sensations in humans, whereas that of T1R1+T1R3 responds selectively to compounds that elicit ‘umami’ taste sensations (Nelson et al., 2001; Zhao et al., 2003). Once these taste receptors are activated by their respective ligands, they activate one or more G-proteins (gustducin, rod-transducin or Gi2), which in turn stimulate one or more effector systems (adenylyl cyclase, phosphodiesterase or phospholipase C ß2; Huang et al., 1999; Clapp et al., 2001; Yan et al., 2001; Ogura et al., 2002). These early transduction steps initiate a cascade of downstream events, including the release of neurotransmitter from the taste cell, generation of action potentials in the postsynaptic afferent neuron, propagation of the action potentials up the gustatory neuraxis, and, ultimately, ingestion or rejection Glendinning et al., 2000; Spector, 2000).
There is compelling evidence that the -subunit of gustducin (-gustducin) contributes significantly to the transduction of ‘bitter’, ‘sweet’ and ‘umami’ taste stimuli (Wong et al., 1996; Ruiz-Avila et al., 2001; He et al., 2002, 2004; Ruiz et al., 2003). Gustducin is co-expressed with T1Rs or T2Rs in some taste cells (Adler et al., 2000; Max et al., 2001; Li et al., 2002; Kim et al., 2003) and its subunits have been shown to activate specific components of known gustatory effector systems (e.g. PLC-ß2, PDE and Ca2+ store-operated channels) in signaling pathways responsive to ‘sweet’ and ‘bitter’ compounds (Huang et al., 1999; Clapp et al., 2001; Yan et al., 2001; Ogura et al., 2002). -Gustducin knockout mice (i.e. Gus–/– mice) exhibit diminished behavioral and/or gustatory nerve responsiveness to natural and artificial sweeteners (Wong et al., 1996; He et al., 2002), ‘bitter’ compounds (Ruiz-Avila et al., 2001; Wong et al., 1996; He et al., 2002; Caicedo et al., 2003; Ruiz et al., 2003) and amino acids (Ruiz et al., 2003; He et al., 2004). Transgenic expression of wild-type -gustducin in Gus–/– mice restores normal behavioral responsiveness to these taste stimuli (Ruiz-Avila et al., 2001).
While there is no doubt that -gustducin contributes to the transduction of ‘sweet’, ‘bitter’ and ‘umami’ taste stimuli, the following observations indicate that other G proteins are also involved in these signaling pathways. First, in anterior tongue, most taste cells doubly positive for T1R2+T1R3 (i.e. presumed ‘sweet’ responders) also express -gustducin (Kim et al., 2003). Yet, in posterior tongue, very few taste cells doubly positive for T1R2+T1R3 also express -gustducin—instead it is the T2R-positive taste cells that co-express -gustducin (Kim et al., 2003). Second, Caicedo et al. (2003) reported that only half of the mouse circumvallate taste cells that displayed increases in intracellular Ca2+ concentration (in response to stimulation with ‘bitter’ taste stimuli) also expressed gustducin; the other half of the taste cells expressed a different G protein, Gi2. Third, electrophysiological recordings from taste nerves and long-term preference tests reveal that knocking out gustducin diminishes (but does not eliminate) responses to a range of concentrations of sucrose, SC45647 (an artificial sweetener), quinine, denatonium benzoate and MSG (Wong et al., 1996; He et al., 2002, 2004; Ruiz et al., 2003). Fourth, expression of a dominant-negative form of -gustducin in Gus–/– mice reduced these responses further still (Ruiz-Avila et al., 2001), indicating that another G protein in the -gustducin-positive taste cells mediated the responses. Fifth, double knockout mice lacking both -gustducin and -transducin showed responses to ‘umami’ taste stimuli that were diminished relative to single knockout mice lacking only one of the G proteins (He et al., 2004).
The purpose of this study was to clarify the necessity of -gustducin to the taste-mediated ingestive response of mice to a diversity of taste stimuli. Although the molecular, Ca2+-imaging and nerve recording findings discussed above strongly suggest that the peripheral taste system of mice contains both -gustducin-dependent and -gustducin-independent pathways for transducing taste stimuli, the nature of the perceptions generated by the gustducin-independent pathways remains to be determined. All previous studies of the taste-related behavioral responses of Gus–/– mice to sapid stimuli have used long-term (i.e. 48 h) preference tests as the behavioral measure. Owing to the protracted nature of these preference tests, the specific contribution of taste to the observed intake may be confounded by post-ingestive feedback and/or experiential effects of the chemical stimuli (Mook, 1963; Rabe and Corbit, 1973; Weingarten and Watson, 1982; Flynn and Grill, 1988; Fregly and Rowland, 1992; Bachmanov et al., 2000; Spector 2000). Here we used a brief-access taste test, in which chemical stimuli are presented during a series of brief trials (i.e. 5 s), to measure unconditioned lick responsiveness to an array of taste stimuli in Gus –/– mice and their heterozygotic littermate controls. Although this test does not assess taste quality per se, it does provide an index of the taste-related affective potency of the stimuli, while minimizing the influence of post-ingestive feedback or experiential effects. This approach has been used successfully in the past to analyze taste-mediated ingestive responses of inbred strains of mice (Boughter et al., 2002; Glendinning et al., 2002; Nelson et al., 2003; Dotson and Spector, 2004).
|
Materials and methods |
|---|
Subjects
We used genetically engineered Gus–/– mice that were bred at the Mount Sinai School of Medicine. The genetic background of the mice was 99.2% C57BL/6J and 0.8% 129/SvEmsJ (for more husbandry details, see He et al., 2002). All mice were housed individually in standard shoebox cages (27.5 x 17 x 12.5 cm) in a room with automatically controlled temperature, humidity and lighting (12 h:12 h light:dark cycle). All subjects were adult males. The sample sizes for each experiment are indicated in the figure legends. Except where noted otherwise, the mice were maintained ad libitum on TestDiet laboratory chow (5012, Purina Mills Inc., Richmond, IN) and water. We tested mice during the light phase of their light–dark cycle, except during experiment 6. In the latter experiment, we tested mice during the dark phase (see below for details).
Lickometry
Each test was conducted in a commercial gustometer (Davis MS160-Mouse; DiLog Instruments, Tallahassee, FL). This device provided each mouse with access to a single sipper tube during successive 5 s trials. Immediately prior to a trial, the computer positioned a sipper tube directly behind a slot (1.5 cm wide, 4.0 cm high) in the back of the testing chamber, and then opened the shutter. The trial ‘began’ when the mouse took its first lick from the sipper tube, and ended 5 s later when the shutter closed. During the 7.5 s inter-trial interval, a different sipper tube was positioned behind the shutter in preparation for the next trial. Each mouse was offered a range of different concentrations of each taste stimulus during a test session (see below for details). The order of presentation was randomized without replacement in blocks so that every concentration of a taste stimulus and water was presented once before the initiation of a second block. Unconditioned licking responses were recorded for later analysis. Each test session lasted 30 min, during which the mouse could initiate as many 5 s trials as possible.
Training procedure
Before taste testing was conducted, the mice were given 2 days of training with water as the taste stimulus. This served to familiarize the mice with the gustometer and trained them to lick from the sipper tube to obtain fluid. Because the mice were placed on a water-restriction schedule (see below) throughout training, they were highly motivated to lick from the sipper tube. Each training session began when the mouse took its first lick, and lasted 30 min. On training day 1, the mouse could drink freely from a single sipper tube throughout the session as the shutter permanently open. On training day 2, the mouse could only drink from a sipper tube during 5 s trials. All mice adapted readily to these procedures and took >250 licks per training session.
Testing
We used the brief-access taste test described in Glendinning et al. (2002). Testing began once training was completed. Multiple concentrations of a single taste stimulus were presented during the 30 min test session. The mouse was permitted to initiate as many trials (and hence, blocks) as possible throughout the 30 min test session. We included water as a solution within each block of taste stimulus concentrations.
Previous work with C57BL/6J mice (Glendinning et al., 2002) demonstrated that it is necessary to pool results from three test sessions to obtain reliable estimates of lick responsiveness to aversive taste stimuli. This is because the mice initiate relatively few trials (i.e. 10–25) with aversive taste stimuli. In contrast, one can obtain reliable estimates of lick responsiveness during a single test session with normally preferred taste stimuli because mice initiate relatively large numbers of trials (i.e. 30–75).
Taste stimuli
Taste stimuli were obtained from Sigma-Aldrich (St Louis, MO), unless stated otherwise. Immediately prior to testing, the stimuli were dissolved in deionized water and presented at room temperature. In all experiments, mice were exposed to a range of concentrations of each stimulus. The specific concentrations were selected because they elicited the full dynamic range of response in Gus+/– mice during preliminary trials.
Standardization of licking responses
For aversive taste stimuli, we controled for individual differences in lick rate and motivation to lick by calculating a tastant/water lick ratio. This involved dividing the average number of licks to each taste stimulus concentration (across all trials with that concentration) by the average number of licks to water alone (across all trials), separately for each mouse. A lick ratio of 1.0 signifies that licks to the taste stimulus equaled licks to water, whereas a lick ratio <1.0 signifies that there were fewer licks to the taste stimulus relative to that for water.
For normally preferred taste stimuli, we controlled for individual differences in lick rate by calculating the standardized lick ratio (SLR; Glendinning et al., 2002) separately for each mouse. To this end, the average number of licks per trial for each concentration was divided by that animal's maximum potential lick rate per trial based on the mean of the interlick interval (ILI) distribution measured during training day 1 with water (only ILIs > 50 ms and < 200 ms were used) (see Glendinning et al., 2002, for details). A SLR approaching 0 indicates that the taste stimulus elicited only sporadic licking, whereas an SLR near 1.0 indicates that the taste stimulus elicited nearly continuous licking across each 5 s trial. Although the SLR does not necessarily control for variation in the motivational state arising from differential responses to the food and water restriction schedule, it does control for individual differences in local lick rate.
Water-deprivation procedure
To encourage sampling from the sipper tube during training and during tests with aversive taste stimuli (in experiments 1–4), we water-deprived the mice for 22.5 h, subjected the animals to a 30 min test session, and then gave them water ad libitum for 1 h. We usually repeated this procedure across 4–5 successive days/week. Under this restricted water-access schedule, all mice maintained their body mass between 85–90% of their baseline values across successive days and showed no overt signs of distress.
Food and water-restriction procedure
To encourage sampling from the sipper tube during tests with normally preferred taste stimuli (i.e. experiments 5–7), we food- and water-restricted each mouse for 23.5 h (see below for details), subjected it to a 30 min test session, and then gave it a recovery day over which it had food and water ad libitum. We usually repeated this procedure three times per week. Under these conditions, the mice: (i) maintained their body mass at >80% of their baseline values; (ii) regained 100% of their baseline body mass over each recovery day; (iii) failed to show any overt signs of distress; and (iv) showed virtually no interest in water alone, but exhibited vigorous concentration-dependent increases in lick responsiveness for normally preferred taste stimuli.
The food- and water-restriction procedure involved limiting each mouse to 1 g of laboratory chow (dustless precision 1 g food-pellets; BioServ) and 2 ml of water. This equaled approximately 20 and 30% of the their normal daily food and water intake, respectively (J. Glendinning, unpublished data).
Because each mouse was subjected to multiple test sessions (one for each of the six taste stimuli), it was necessary to interject a recovery day between successive test sessions, during which food and water were available ad libitum. After the recovery day, each mouse was once again food- and water-restricted.
|
Experiment 1: Do Gus–/– mice show attenuated lick responsiveness to aversive taste stimuli? |
|---|
Based on findings from long-term preference tests, investigators have reported that knocking out
-gustducin diminishes the aversion to a broad range of concentrations of quinine hydrochloride and denatonium benzoate (Wong et al., 1996; Ruiz-Avila et al., 2001; He et al., 2002; Ruiz et al., 2003). We asked whether this finding would generalize to a more diverse range of aversive taste stimuli in a brief-access taste test. Methods
Taste stimuli and test concentrations
The ‘bitter’ taste stimuli were quinine hydrochloride, QHCl (0.006, 0.01, 0.03, 0.1, 0.3, 1 mM), denatonium benzoate (0.1. 0.3, 1, 3, 10 mM), sparteine (0.1, 0.3, 1, 3, 10 mM), salicin (1, 10, 100 mM), cycloheximide (0.01, 0.03, 0.1, 0.3, 1, 3 µM), caffeine (10, 30, 100 mM), quinacrine (10, 30, 100, 400 µM) and urea (0.3, 1, 3 M). All of these compounds taste bitter to humans (the authors, unpublished data), and all but caffeine and urea have been shown to activate -gustducin in either an in vitro assay involving isolated taste membranes (Ruiz-Avila et al., 2000) or G protein-coupled T2R taste receptors in a heterologous cell reporter system (Chandrashekar et al., 2000; Bufe et al., 2002).
Testing procedure
After training, each mouse was run through three test sessions with each of the following taste stimuli in the following order: QHCl, denatonium benzoate, sparteine, salicin, cycloheximide, caffeine, quinacrine and then urea. The mice were all water-deprived prior to testing (see above for details).
Data analysis
We analyzed the tastant/water lick ratios from each taste stimulus (e.g. the range of QHCl concentrations) with a two-way ANOVA. We treated stimulus concentration as a within factor and genotype as a between factor. To decompose significant interactions, we ran each main effect through a one-way ANOVA. The level was set at 0.05 for this and all subsequent statistical tests.
We also performed one-sample paired t-tests to determine whether the mean tastant/water lick ratios for each QHCl concentration were significantly above or below the indifference point (i.e. 1.0) when the concentration main effect from the ANOVA was significant. To control for the use of multiple paired comparisons on the same data set, we performed a Bonferroni correction (i.e. divided the level by the number of comparisons that were made).
Results
Gus–/– mice were significantly less responsive than Gus+/– mice to all eight aversive taste stimuli (Figure 1). A two-way ANOVA revealed significant main effects of genotype and stimulus concentration for all compounds; there was also a significant interaction of the main effects (Table 1). To explore the nature of the interactions, we further analyzed the main effect of stimulus concentration, separately for each genotype and aversive taste stimulus. Gus+/– mice showed similar responses to all of the aversive taste stimuli—i.e. they all exhibited significant concentration-dependent reductions in lick responsiveness (in all one-way ANOVAs, P < 0.05). Gus–/– mice, on the other hand, showed more variable responses to the eight aversive stimuli. For sparteine and salicin, they exhibited indifference followed by increases in lick responsiveness at the highest concentration (in both one-way ANOVAs, P < 0.05); for QHCl, cycloheximide and caffeine, they failed to show any concentration-dependent changes in lick responsiveness (in all three one-way ANOVAs, P > 0.05); and for denatonium benzoate, quinacrine and urea, the mice exhibited concentration-dependent decreases in lick responsiveness, although their avoidance curves were right-shifted in comparison to their +/– littermates (in all three one-way ANOVAs, P < 0.05). While the concentration–response curve for QHCl was virtually flat, all of the mean values were significantly greater than 1.0 (one-sample paired t-test: in all cases, t-value 
3.32, df = 9, P < 0.01).
|
|
Taken together, these results demonstrate that for all eight aversive taste stimuli, Gus–/– mice show significantly less avoidance than do Gus+/– mice. Further, Gus–/– mice actually lick more vigorously from some concentrations of QHCl, sparteine and salicin than from water alone.
|
Experiment 2: Are high concentrations of QHCl, cycloheximide and sparteine aversive to Gus–/– mice? |
|---|
In the previous experiment, Gus –/– mice did not exhibit a concentration-dependent decrease in licking in responsiveness to QHCl, sparteine, cycloheximide, caffeine and salicin. Here, we tested the hypothesis that knocking out
-gustducin completely eliminated the aversive taste of all concentrations of these compounds. To this end, we asked whether Gus–/– mice would display lick avoidance to QHCl, cycloheximide and sparteine when offered concentrations higher than those used in experiment 1. We could not test caffeine because concentrations higher than those used in experiment 1 (i.e. 300 mM) produced toxic effects during 30 min test sessions (J. Glendinning, unpublished data). Salicin also had to be excluded because concentrations higher than those used in experiment 1 were insoluble in water. Methods
We used the same mice and multi-bottle no-choice testing procedure as in experiment 1. The only difference was that we used a higher range of concentrations of QHCl (1, 3 and 10 mM), cycloheximide (3, 10, 20 and 30 µM) and sparteine (10, 30 and 100 mM). The order of testing the different taste stimuli was as follows: QHCl, cycloheximide and then sparteine.
Because of concerns about the toxicity of the high concentrations of each taste stimulus, we calculated two separate tastant/water lick ratios for each taste stimulus. Ratio 1 was calculated based on licking responses during the first block of trials in test session 1. Ratio 2 was based on licking response during all subsequent blocks of trials across test sessions 1–3. If a taste stimulus elicited an aversive response through a post-ingestive toxicity mechanism, then we expected ratio 2 to be substantially lower than ratio 1 across all concentrations of a taste stimulus. On the other hand, if a taste stimulus elicited an aversive response through a gustatory mechanism, then we expected that ratios 1 and 2 would be similar across all concentrations of a taste stimulus. We tested these predictions with a two-way ANOVA, treating both Ratio and Stimulus Concentration as within factors. We analyzed each taste stimulus separately.
Results
Gus–/– mice displayed robust concentration-dependent inhibition of licking in response to all three bitter taste stimuli during both the first block of trials (in test session 1) and all subsequent blocks of trials (in test sessions 1–3) (Figure 2). For each ‘bitter’ taste stimulus, the two-way ANOVA revealed a significant effect of stimulus concentration (P < 0.05), but no significant effect of trial (P > 0.05) or interaction of stimulus concentrationxtrial (P > 0.05).
|
These findings, together with those from experiment 1, demonstrate that QHCl, cycloheximide and sparteine all inhibit licking in Gus–/– mice, but only at relatively high concentrations. Further, they provide support for the hypothesis that the inhibition of licking at these high concentrations is mediated by taste, and not post-ingestive toxicity. This is because the concentration–response curves derived from the first block of trials were statistically indiscriminable from those derived from all subsequent blocks of trials.
|
Experiment 3: Do Gus–/– mice show attenuated lick responsiveness to NaCl, HCl and MSG? |
|---|
In long-term preference tests, Gus–/– mice exhibit normal preference–response curves across a broad range of NaCl and HCl concentrations (Wong et al., 1996
; Ruiz-Avila et al., 2001; He et al., 2002; Ruiz et al., 2003), but attenuated preferences for midrange concentrations of monosodium glutamate (MSG; 60 – 100 mM; Ruiz et al., 2003; He et al., 2004). The present experiment sought to address two interpretive limitations of these studies. First, because of their long duration, one cannot determine the relative contribution of taste versus postingestive and/or experiential factors to the ingestive response (see e.g. Bachmanov et al., 2000; Ruiz et al., 2003). Second, because MSG was not always presented with an epithelial sodium channel blocker (e.g. amiloride hydrochloride) for the long-term preference studies, it is impossible to determine the extent to which the taste of the Na+ ion in MSG contributed to the ingestive response. To address these limitations, we examined the short-term lick response of Gus–/– mice to MSG in the presence of varying concentrations of amiloride. Methods
Taste stimuli and test concentrations
We tested a range of concentrations of NaCl and NH4Cl (0.01, 0.1, 0.2, 0.3, 0.6, 1.0 M), HCl (0.1, 1.0, 10, 100 mM) and MSG (0.01, 0.03, 0.1, 0.3, 1.0 M). We ran three separate tests with MSG. In these tests, the five concentrations were mixed with water alone, 10 µM amiloride or 100 µM amiloride (test 3). We used 10 µM amiloride because it has been found to reduce the response of the chorda tympani nerve to the Na+ component of MSG in B6 mice (He et al., 2004). We also used 100 µM amiloride because it has been found to (i) eliminate the response of the chorda tympani nerve to the Na+ component of MSG in T1R1 and T1R3 knock-out mice (Zhao et al., 2003), (ii) increase the behavioral detection threshold for sodium in C57BL/6 mice (Eylam and Spector, 2003), but (iii) lack a perceptible taste in mice (Eylam et al., 2003). In all test sessions, we included a water control. When amiloride was mixed with the taste stimulus, it was added to all test solutions, including the water control.
Testing procedure
We used the same mice and multi-bottle no-choice testing procedure as in experiment 1. We ran mice through a single test session with NaCl, NH4Cl and MSG (with and without amiloride), but had to use three test sessions with HCl to obtain reliable estimates of lick responsiveness. Each mouse was tested with the taste stimuli in the following order: NaCl, NH4Cl, HCl, MSG without amiloride, MSG in 10 µM amiloride, and MSG in 100 µM amiloride.
Data analysis
The tastant/water lick ratios from each taste stimulus were analyzed separately using a two-way ANOVA. Stimulus concentration was a within factor and genotype a between factor. To explore the factors contributing to a significant interaction, we ran unpaired t-tests (two-tailed) across genotype at specific MSG concentrations. In addition, to determine whether the mice licked more vigorously from specific concentrations of MSG than water alone, we used a one-sample paired t-test (t-tailed) to determine whether the tastant/water lick ratios at each MSG concentration differed significantly from 1.0 (i.e. indifference). To control for the use of multiple t-tests on a related data set, we used the Bonferroni correction to adjust the level (i.e. 0.05/number of comparisons).
Results
The Gus–/– and Gus+/– mice both exhibited strong concentration-dependent inhibition of licking for NaCl, NH4Cl and HCl (Figure 3). Indeed, the analysis of the NH4Cl and HCl data revealed a significant main effect of concentration, but a non-significant main effect of genotype; the interaction of the main effects was non-significant as well (Table 2). For NaCl, the main effect of genotype was not significant, but the genotype x concentration interaction was significant; this latter finding reflects the fact that 0.6 and 1.0 M NaCl inhibited licking less effectively in the Gus–/– mice (for both concentrations, unpaired t-value > 5.4, df = 34, P < 0.01). Taken together, these data show that Gus–/– mice exhibit normal lick avoidance responses to NH4Cl and HCl, but attenuated lick avoidance responses to high concentrations of NaCl (300 mM).
|
|
When MSG was mixed with 0 or 10 µM amiloride, mice from both genotypes showed indifference at low concentrations and avoided high concentrations (Figure 4, left and middle panels). Indeed, a two-way ANOVA of the data revealed a non-significant main effect of genotype, but a significant main effect of concentration (Table 3). The significant interactions of genotype x MSG concentration are explained by the fact that 1.0 M MSG was less aversive to Gus–/– mice than to Gus+/– mice, when mixed with 0 or 10 µM amiloride (in both cases, P < 0.05/5).
|
|
When MSG was mixed with 100 µM amiloride, the mice from each genotype showed qualitatively distinct responses. The Gus+/– mice exhibited an appetitive licking response (i.e. a tastant/water lick ratio > 1.2) to MSG concentrations
0.3 M, but an aversion to 1.0 M MSG (Figure 4). On the other hand, the Gus–/– mice exhibited a weak appetitive response to all MSG concentrations. These inferences are supported by two statistical analyses. First, a one-way ANOVA revealed a significant main effect of concentration on the tastant/water lick ratios for Gus+/– mice [F(4,40) = 31.1, P < 0.05], but not for Gus–/– mice [F(4,28) = 0.8, P > 0.05]. These latter findings indicate that the affective value of MSG, when mixed with 100 µM amiloride, decreases with increasing concentrations (>100 mM) in Gus+/– mice, but remains stable in Gus–/– mice. Second, a series of one-sample paired t-tests revealed that Gus+/– mice licked significantly more vigorously from the MSG concentrations 0.3 M than they did from water alone, but significantly less vigorously from 1 M MSG (all t-values > 3.4; df = 10; P < 0.01). Gus–/– mice, on the other hand, did not exhibit a concentration-dependent change in affective response to MSG; for all concentrations of MSG, except 0.3 M, they licked at slightly higher rates than they did from water alone (all t-values > 3.6; df = 7; P < 0.01). Taken together, these findings indicate that it is necessary to present MSG in amiloride concentrations of 100 µM to observe robust genotypic differences in lick responsiveness. Under these conditions, Gus–/– mice show weak appetitive responses to a broad range of MSG concentrations (0.01–1.0 M). In contrast, Gus+/– mice show strong appetitive responses to MSG concentrations ranging from 0.01 to 0.1 M, and strong aversive responses to 1.0 M MSG.
|
Experiment 4: Does knocking out -gustducin selectively eliminate the aversive taste of MSG?
|
|---|
The results of experiment 3 indicate that knocking out gustducin eliminates the aversive response to 1.0 M MSG, when mixed with 100 µM amiloride. This finding contrasts with earlier reports involving long-term preference tests, which indicated that Gus–/– show normal aversions to 1.0 M MSG, when mixed with 10 µM amiloride (He et al., 2004
). In the present experiment, we tested the hypothesis that the results from experiment 3 were an artifact of the water-deprivation procedure, which was used to motivate stimulus-sampling in the gustometer. To this end, we examined the lick response of Gus+/– and Gus–/– mice to MSG when they were in a different physiological state—i.e. one that strongly motivates the mice to ingest nutrients, but not water. This physiological state is produced by 23 h of food- and water-restriction (see below for details; Glendinning et al., 2002). Methods
We used mice that were naïve to the taste stimuli. Following training, the mice were subjected to the same multi-bottle no-choice testing procedure as in experiment 3, except that they were food- and water-restricted over the 23 h period immediately prior to testing (see above for details). Each mouse was exposed to five different concentrations of MSG (0.01, 0.03, 0.1, 0.3, 1 M), all mixed with 100 µM amiloride.
Data analysis
We treated the SLR of each mouse to the control solution (i.e. water plus 100 µM amiloride) as a baseline response, and then sought to determine the lowest concentration of MSG that elicited a significantly elevated (or depressed) SLR. To this end, we used the paired t-test, separately for each concentration. We controlled for the use of multiple t-tests with a Bonferroni correction.
Results
The Gus–/– and Gus+/– mice both showed concentration-dependent increases in SLR for the lower range of MSG concentrations (Figure 5). The lowest concentration that elicited a SLR significantly above the baseline response was 0.03 M for Gus+/– mice and 0.1 M for Gus–/– mice. The most notable genotypic difference occurred at the high MSG concentrations (Figure 5). While the SLR of Gus–/– mice increased linearly with MSG concentration, that of Gus+/– mice showed a much more shallow increase reaching a maximum at 0.1 M and then plummeted to a value significantly below baseline at 1 M.
|
) SLR relative to water. (These findings corroborate those from the previous experiment, and thus demonstrate that knocking out
-gustducin selectively eliminates the aversive taste of 1.0 M MSG when mixed with 100 µM amiloride. |
Experiment 5: Do Gus–/– mice show attenuated lick responsiveness to natural and artificial sweeteners? |
|---|
Previous studies have shown that knocking out
-gustducin markedly attenuates the preference of mice for a broad range of concentrations of sucrose and SC45647 (an artificial sweetener) in 48 h preference tests (Wong et al., 1996; Ruiz-Avila et al., 2001; He et al., 2002; Ruiz et al., 2003). Here, we examined the lick response of Gus+/– and Gus–/– mice to a broad range of concentrations of sucrose, maltose, fructose, SC45647 and two mixtures of varying-length glucose polymers (Polycose and maltooligosaccharide).
Polycose and maltooligosaccharide are highly palatable to rodents (Davis and Breslin, 2000; Bachmanov et al., 2001; Ramirez, 1994), but not to humans (Feigin et al., 1987). Polycose is a mixture of glucose polymers, ranging from glucose (G1) to G30 or higher; 
9% of Polycose is glucose and maltose. Maltooligosaccharide is a more uniform mixture of glucose polymers, ranging from G2 to G8; only 1.5% is G2. Behavioral and electrophysiological studies indicate that the taste quality of Polycose differs from that of sucrose in rats and hamsters (Nissenbaum and Sclafani, 1987; Sako et al., 1994; Rehnberg et al., 1996), but its taste quality has not been investigated in mice.
Methods
The test stimuli were as follows: sucrose (0.03, 0.1, 0.2, 0.3, 0.6 and 1 M), SC45647 (0.01, 0.03, 0.1, 0.3, 1 and 3 mM), maltose (0.03, 0.1, 0.2, 0.3, 0.6 and 1 M), fructose (0.03, 0.1, 0.2, 0.3, 0.6 and 1 M), Polycose (0.01, 0.02, 0.04, 0.08, 0.16 and 0.32 M) and maltooligosaccharide (14, 28, 55, 111, 221, 442 mM). We determined the molar concentrations of Polycose based on an average molecular weight of 1000 daltons (provided by the manufacturer), and that of maltooligosaccharide based on an average molecular weight of 724 daltons (using the distribution of polymer lengths provided by the manufacturer). The SC45647 was obtained as a gift from Goran Hellekant (University of Wisconsin), the Polycose was purchased from Abbott Laboratories (Columbus, OH), and the maltooligosaccharide was purchased from Pfanstiehl Laboratories (Waukegan, IL).
Testing procedure
We tested the same mice, and used the same multi-bottle no-choice testing procedure, as in experiment 4. Each mouse was tested with all six taste stimuli, in the following order: sucrose, SC45647, maltose, fructose, Polycose and then maltooligosaccharide.
Data analysis
The SLR from each taste stimulus was analyzed separately using a two-way ANOVA; stimulus concentration was a within factor and genotype a between factor.
Results
The mice showed robust concentration-dependent increases in lick responsiveness to the taste stimuli (Figure 6). For all stimuli, except sucrose, there was a significant main effect of concentration, but a non-significant main effect of genotype and a non-significant interaction of concentration x genotype (Table 4). The non-significant main effect of genotype indicates that both genotypes had equal lick responsiveness to a broad range of natural and artificial sweeteners. For sucrose, on the other hand, there was a significant main effect of concentration and genotype (Table 4). Further, the interaction of concentration x genotype was significant.
|
|
Taken together, the results of this assay indicate that knocking out
-gustducin has little impact on lick responsiveness of mice to a structurally diverse array of natural and artificial sweeteners. The only exception was sucrose—Gus–/– mice exhibited a small, but significant, attenuation in lick responsiveness to this natural sweetener. |
Experiment 6: Is lick responsiveness of Gus–/– mice to sucrose and SC45647 influenced by water- and food-restriction? |
|---|
Taste nerve recordings and long-tem preference tests indicate that Gus–/– mice are substantially less responsive than Gus+/– mice to sucrose and SC45647 (Wong et al., 1996
; Ruiz-Avila et al., 2001; He et al., 2002, 2004; Ruiz et al., 2003). In contrast, the results of experiment 5 indicate that Gus–/– mice show only moderately attenuated lick responsiveness to sucrose and normal lick responsiveness to SC45647. To explain these contradictory findings, we hypothesized that the food- and water-restriction procedure (used in experiment 5) disproportionately elevated lick responsiveness of the Gus–/– mice to sweeteners. If so, then we predicted that non-restricted Gus–/– mice should generate results more consistent with those reported in taste-nerve recordings and long-tem preference tests—i.e. they should lick substantially less vigorously than Gus+/– mice across a range of sucrose and SC45647 concentrations. Methods
Training procedure
We used the same mice as in experiment 5, but trained them to perform in the gustometer in a non-deprived state. At the beginning of the dark cycle, we removed the mouse from its home cage and placed it in a gustometer. The mouse was given 60 min to acclimate to the gustometer prior to each training session. During training days 1 and 2, we presented a 0.3 M Polycose solution in a stationary spout over a 60 min test session. We used the Polycose solution because it stimulates licking, but was different than the test stimuli used in the experimental phase of testing. During training day 3, the 0.3 M polycose solution was presented during successive 5 s trials, separated by a 7.5 s inter-trial interval (as described in experiment 1). The mouse could initiate as many trials as it wanted over the 60 min session. By the end of training day 3, all mice had learned to drink vigorously from the sipper tubes. Parenthetically, we had to illuminate the room containing the gustometer with a 40 W red light bulb as testing was conducted in the dark phase of the light:dark cycle.
Testing procedure
We used the same mice, multi-bottle no-choice testing procedure, and concentrations of sucrose and SC45647 as in experiment 5. The only difference was that we tested the mice on two successive days and pooled together the results for the analysis; this was necessary because the non-deprived mice initiated fewer numbers of trials. To control for any effects of stimulus testing order, we used a counter-balanced design (i.e. half of the mice were tested initially with sucrose, and the other half with SC45647). Because the mice were not food-and water-restricted, we could test them on successive days.
Data analysis
We ran a two-way ANOVA on the SLRs, with genotype as a between factor and concentration as a within factor. The ANOVAs were conducted separately for each taste stimulus.
Results
The Gus +/– and Gus –/– mice both exhibited strong concentration-dependent increases in SLR for sucrose and SC45647 (Figure 7). For sucrose, there was a significant main effect of genotype and concentration, and a significant interaction of genotype x concentration (Table 5). The significant interaction reflects the fact that lick responsiveness of both genotypes was identical at low concentrations but diverged at the midrange and high concentrations, with the Gus+/– mice licking slightly more vigorously. For SC45647, there was a significant main effect of concentration, but no significant main effect of genotype or interaction of genotype x concentration (Table 5).
|
|
Taken together, these results contradict the hypothesis that that the food- and water-restriction procedure disproportionately increased the lick responsiveness of Gus–/– mice. This is because the results from non-deprived mice in this experiment are qualitatively and quantitatively similar those from food- and water-restricted mice in the previous experiment.
|
Experiment 7: Is lick responsiveness of Gus–/– mice to natural and aritificial sweeteners influenced by the stimulus presentation method? |
|---|
When Gus+/– and Gus–/– mice were examined using long-term preference tests with sucrose and SC45647 (Wong et al., 1996
; Ruiz-Avila et al., 2001; He et al., 2002; Ruiz et al., 2003), they were (i) offered a choice between water and a taste stimulus over a 48 h test session, and (ii) tested with different concentrations of each taste stimulus during separate 48 h tests, in an ascending concentration series. In contrast, when Gus+/– and Gus–/– mice were tested with the brief-access procedure in experiments 5 and 6, they were offered a broad range of concentrations of each specific taste stimulus during a single 30 min test session (in a randomized block design). Here, we tested the hypothesis that the different methods of stimulus presentation contributed to the contradictory results generated by long-term preference tests and brief-access taste tests. To this end, we redesigned the brief-access taste test so that it used a method of stimulus presentation that more closely approximated that of the two-bottle preference test. We predicted that Gus–/– mice would be substantially less responsive than Gus+/– mice to sucrose, SC45647 and maltose.
Methods
We used the same mice as in experiments 4–6, but subjected them to a two-bottle no-choice testing procedure. That is, we presented each mouse with two solutions per 30 min test session—water and a specific sweetener concentration. To control for potential order effects across trials within the 30 min test session, we treated the two test solutions as a block, and randomized (without replacement) the presentation sequence of each solution within each block. The mouse was permitted to initiate as many trials (and hence, blocks) as possible throughout the test session.
We tested three sweeteners in the following order: sucrose, SC45647 and then maltose. To mimic the stimulus presentation sequence used in the long-term preference tests, we presented six concentrations of each taste stimulus in an ascending series (across successive test sessions). We used the same concentrations of each sweetener as in experiment 5.
Data analysis
To make our data comparable to those of the long-term preference tests, we calculated a preference ratio, separately for each sweetener concentration and mouse. We first calculated the mean number of licks across all trials with each sweetener solution and the mean number of licks across all trials with each corresponding water solution. Then, we divided the mean number of licks from the sweetener solution by the mean total number of licks from both the sweetener solution and water. A preference ratio of 0.5 indicates indifference to the sweetener solution, whereas a preference ratio of 1.0 indicates a strong preference for the sweetener solution. We subjected the preference ratio data to a two-way ANOVA, treating genotype as a between factor and concentration as a within factor. To help clarify significant interaction, we also made comparisons across genotypes with an unpaired t-test, separately for each sweetener concentration. To control for the use of multiple t-tests, we corrected the level with a Bonferroni correction (i.e. level =