Chem. Senses 28: 301-313,
2003
© Oxford University Press 2003
A Psychophysical Investigation of Binary Bitter-compound Interactions
1 Monell Chemical Senses Center, 3500 Market St, Philadelphia, PA 19104, USA 2 Firmenich SA, Route des Jeunes 1, Geneva, Switzerland CH-1211
Correspondence to be sent to: Russell Keast, Monell Chemical Senses Center, 3500 Market St., Philadelphia, PA 19104, USA. e-mail: keast{at}monell.org
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The aim of this study was to determine if taste interactions occur when bitter stimuli are mixed. Eight bitter stimuli were employed: denatonium benzoate (DB), quinine-HCl (QHCl), sucrose octaacetate (SOA), urea, L-tryptophan (L-trp), L-phenylalanine (L-phe), ranitidine-HCl, and Tetralone. The first experiment constructed individual psychophysical curves for each subject (n = 19) for each compound to account for individual differences in sensitivities when presenting bitter compounds in experiment 2. Correlation analysis revealed two groupings of bitter compounds at low intensity (1, L-trp, L-phe, and ranitidine; 2, SOA and QHCl), but the correlations within each group decreased as the perceived intensity increased. In experiment 2, intensity ratings and two-alternative forced-choice discrimination tasks showed that bitter compounds generally combine additively in mixture and do not show interactions with a few specific exceptions. The methods employed detected synergy among sweeteners, but could not detect synergy among these eight bitter compounds. In general, the perceived bitterness of these binary bitter-compound mixtures was an additive function of the total bitter-inducing stimuli in the mouth.
Key words: additive, bitter taste, suppression, synergy, taste interactions, taste psychophysics
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Psychophysical investigations of same-quality taste-mixture interactions have revealed non-linear enhancements that implicate taste-integration mechanisms. For both the sweet and savory (umami) qualities, certain same-quality binary mixtures stimulate a perceived intensity in excess of predicted additivity (synergy). The binary mixture of the sweeteners aspartame and acesulfame-K results in a synergy of sweet taste (McBride, 1988
;
Ayya and Lawless, 1992;
Schiffman et al.,
1995; Schifferstein,
1996). Similarly, a binary mixture of monosodium glutamate and the
sodium salt of 5'-inosine or guanosine monophosphate results in the
synergy of savory taste (Yamaguchi,
1967; Rifkin and Bartoshuk,
1980). Very little is known, however, about same quality
interactions within bitterness (Keast and
Breslin, 2003), arguably the most physiologically complex taste.
The aim of this study was to determine if taste interactions occur when bitter
stimuli are mixed. One reason for the dearth of data may be this complexity. Any investigation of human bitterness perception must contend with three complicating factors:
- There are many chemically distinct compound classes that elicit bitter
taste: alkaloids, amino acids, iso-humulones, phenols, amines, thioureas,
carbamates, ionic salts, etc. (Belitz and
Wieser, 1985; Spielman et
al., 1992).
- Bitter taste transduction involves many proteins. A large family
(30–40) of putative `bitter-compound' receptors (T2Rs) have been
discovered (Adler et al.,
2000; Chandrashekar et
al., 2000). There is also more than one post-receptor
transduction sequence (Spielman et
al., 1992). With regard to coding, many different T2Rs were
identified within individual bitter-sensitive cells
(Adler et al., 2000),
indicating that each cell may respond to many bitter compounds (broad cellular
tuning) (Chandrashekar et al.,
2000). An alternate hypothesis was suggested by Caicedo and Roper
(Caicedo and Roper, 2001), who
reported that bitter-sensitive taste cells generally responded to only one of
five bitter stimuli, indicating that these stimuli activate different
subpopulations of cells (more selective cellular tuning).
- Individual observers vary in the quantity and presumably functionality of
taste cells and receptors (Kim et
al., 2003), which causes large individual variation in bitter
taste perception (Yokomukai et
al., 1993; Bartoshuk
et al., 1998;
Delwiche et al.,
2001; Keast and Breslin,
2002a,b).
To address factor 3 above and determine if taste interactions occur, concentration–intensity psychophysical curves were constructed for each individual and each bitter compound in experiment 1, thereby allowing compounds to be mixed at the same perceived intensity for subjects with different sensitivities. Experiment 2 investigated whether binary bitter-compound mixtures combined additively, or interacted synergistically or suppressively. This is a comprehensive investigation of binary interactions among eight compounds that stimulate bitter taste.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Subjects
Twenty-two non-smoking volunteers (13 females, 9 males) between 21 and 52 years old (mean 30.1 years) were paid to participate in the study. Subjects were mostly employees of the Monell Chemical Senses Center (primarily Caucasian and African–American). They provided informed consent on an Institutional Review Board approved form. The subjects were asked to refrain from eating, drinking or chewing gum for at least 1 h before testing.
Subject training
Subjects were initially trained in the use of the Labeled Magnitude Scale
(LMS) (Green et al.,
1993,
1996) except the top of the
scale was described as the `strongest imaginable' sensation of any kind
(referred to as the general LMS, or gLMS)
(Bartoshuk, 2000). The gLMS is
a computerized psychophysical tool that requires subjects to rate the
perceived intensity along a vertical axis lined with adjectives: barely
detectable = 1, weak = 5, moderate = 16, strong = 33, very strong = 51,
strongest imaginable = 96; the adjectives are spaced semi-logarithmically,
based upon experimentally determined intervals to yield ratio quality data
(Green et al., 1993,
1996). The gLMS only shows
adjectives, not numbers, to the subjects, but the experimenter receives
numerical data from the computer program.
Subjects were trained to identify each of the five taste qualities by presenting them with 10 ml of prototypical stimuli: 150 mM sodium chloride (NaCl) salty, 0.05 mM quinine-HCl (QHCl) bitter, 300 mM sucrose sweet, 3 mM citric acid sour, and 100 mM monosodium glutamate (MSG) savory. In all cases, subjects were instructed to identify the labeled quality as the dominant one, but others may also be perceived to a lesser degree. To help subjects understand how a stimulus could elicit multiple taste qualities, 300 mM urea (usually bitter and slightly sour) and 50 mM NH4Cl (usually salty, bitter, and slightly sour) were also employed as training stimuli.
A computerized data-collection program simultaneously presented subjects with 5 gLMSs corresponding to SWEET, SALTY, SOUR, SAVORY and BITTER. The order of the five scales on the monitor was randomized from session to session but remained constant within each test session.
Stimuli
Acesulfame-K, ammonium chloride, aspartame, citric acid, denatonium benzoate (DB), MSG, L-phenylalanine (L-phe), sucrose, sucrose octaacetate (SOA), NaCl, L-tryptophan (L-trp), and urea were all purchased from Sigma (St Louis, MO) and were Sigma-ultra grade. QHCl was purchased from Fluka (Buchs, Switzerland), ranitidine from Medisca (New York) and Tetralone from Kalsec (Kalamazoo, MI). All solutions were prepared with deionized MilliporeTM (Bedford, MA) filtered water and stored in amber glass bottles at 4–8°C and brought up to room temperature prior to testing with the aid of a water bath. Solutions were made fresh every 5 days. MilliporeTM filtered deionized water was used as the blank stimulus and the rinsing agent in all experiments.
Stimulus delivery
An aliquot of 10 ml of each solution was presented in 30 ml polyethylene medicine cups (Dynarex, Orangeburg, NY) on a numbered tray. All samples were presented in random order with an interstimulus interval of 90 s unless otherwise stated. The tasting protocol asked subjects to sip, rate and expectorate each solution. On each trial, subjects held 10 ml of solution in their mouth for 5 s and rated the intensity of the taste qualities of the solution (sweet, bitter, sour, salty, savory) before expectorating. Subjects wore nose-clips (GaleMed, Taipei, Taiwan) to eliminate olfactory input while rating.
Experiment 1: covariation of bitterness among compounds at three concentrations
Bitterness perception among individuals is highly variable, but the bitterness elicited by two compounds may correlate. For example, at a fixed concentration of QHCl and a fixed concentration of DB one individual may be sensitive to the bitterness of both (rate them as `strong' on the gLMS), while a second individual may be insensitive to the bitterness of both (rate them as `weak' on the gLMS). While there are large differences in the perceived bitterness of DB and QHCl between the two individuals, each individual responds similarly to the two.
Psychophysical curves were constructed for each bitter compound for each individual subject to enable us to deliver bitter additives that were in the same intensity range for all subjects (experiment 2). These functions provided the opportunity to investigate bitterness correlations as a function of individual sensitivities among bitter compounds at three different concentration levels. First, we adjusted intensity ratings for bias in scale use.
PROP (n-propylthiouracil) bitterness ratings and standardization of
gLMS ratings with tone and weight ratings
The PROP assessment and gLMS standardization followed previously published
methods used in our laboratory (Delwiche
et al., 2001). Briefly, subjects rated the bitterness and
total intensity of 10 ml samples of five concentrations of PROP (5.5 x
10–5, 1.7 x 10–4, 5.5 x
10–4, 1.7 x 10–3 and 5.5 x
10–3 M). Between each sample, subjects rinsed four times with
deionized water. Subjects also rated the loudness of six tones [generated by a
Maico Hearing Instruments tone generator (Minneapolis, MN), presented via
head-phones at 4000 Hz for 2 s at levels 0, 20, 35, 50, 65 and 80 dB] and the
heaviness of six visually identical weights (opaque, sand-filled jars at
levels 225, 380, 558, 713, 870 and 999 g). All three types of ratings were
made on a computerized gLMS. Subjects were asked to rate the intensity of
taste, or loudness, or heaviness, and all judgments were made within the
context of the full range of sensations experienced in life on the gLMS. All
stimuli were presented twice in blocks of ascending order. Subjects first
rated the intensity of weights, then tones, and finally PROP solutions.
There were significant correlations between PROP bitterness ratings, heaviness ratings and loudness ratings. Since these three sensory modalities were assumed to be unrelated, the significant correlations indicated that the gLMS ratings were subject to individual scale-use bias and required standardization across subjects.
To determine a standardization factor, each subject's average intensity for heaviness was divided by the grand mean for heaviness across weight levels and subjects. This procedure for determining a correction factor was repeated with loudness ratings (averaging across decibel levels). The two correction factors (one for weights and one for tones) were averaged, and each individual's bitter intensity ratings for all eight bitter compounds, in subsequent tests, and all five levels of PROP were multiplied by his or her personal standardization factor for scale-use bias.
Psychophysical curves for bitter compounds
The concentration ranges for constructing a psychophysical curve for the
bitter stimuli were: DB (7.5 x 10–8–1 x
10–4 M), L-phe (0.016–0.16 M),
L-trp (0.01–0.06 M), SOA (1 x
10–5–1 x 10–3 M), urea
(0.15–2.5 M), QHCl (1 x 10–5–1 x
10–2 M), ranitidine (1 x 10–4–2
x 10–2 M), Tetralone (1.37 x
10–5–1 x 10–2 M). Subjects were
presented with numbered trays that contained 10 randomized solutions (10 ml)
of one bitter stimulus (nine concentrations from the psychophysical curve and
one deionized water control). The nine concentrations for each bitter stimulus
ranged from below `weak' on the gLMS to maximum solubility (L-trp,
L-phe, SOA) or maximum practical tasting limit (near `very
strong'). Each point on an individual psychophysical curve was tested at least
four times. Subjects were excluded from the study (3 of 22 subjects screened),
if bitterness concentration–intensity curves were not ordinal (defined
here as a change of direction of slope >30% of the y-axis values)
over the range of concentrations tested.
Statistical analysis
Data used for correlation and cluster analysis were the individual
bitterness intensity ratings of concentration levels (associated with average
ratings of gLMS 4, 8 and 12). Note that individual ratings of the compounds
were free to vary at each level; the concentrations were selected so that the
average ratings would be perceived at particular intensities. Correlation
analysis (Pearson's product moment coefficients) and cluster analysis (single
linkage joining, Euclidean distances) were performed using Statistica version
6.0. To reduce Type I errors, a Bonferroni correction for multiple comparisons
was made by dividing the P value (P < 0.05) by 36, the
total number of correlations. Statistical significance of correlation
therefore was P < 0.0014.
Experiment 2: bitter–bitter interactions
Subjects
All subjects had participated in experiment 1. Due to the large number of
sessions to complete experiment 2 (eight sub-experiments each comprising at
least 16 sessions) and some subject's insensitivity to the bitterness of
certain compounds, only five subjects completed all of the sub-experiments
(128 sessions). Other subjects completed partial sets of separate
sub-experiments. For each bitter stimulus used as a target compound to which
other compounds were added, the number of subjects who completed each test
matrix was: DB n = 14 (eight females), L-phe n =
15 (seven females), L-trp n = 14 (seven females), SOA
n = 15 (nine females), urea n = 10 (seven females), QHCl
n = 15 (nine females), ranitidine n = 15 (nine females),
Tetralone n = 14 (eight females).
Design and rationale
All bitter compounds were both a `target' (four concentrations from the
dynamic portion of the psychophysical curve) and an `additive' (a weak
intensity added to the four concentrations of the target compound). During
each session, subjects were presented with the target concentrations of a
bitter compound, and binary combinations of the target concentrations with the
weakly bitter additives [including self-addition of a weak intensity (the
additive control)]. There were some binary combinations that were not included
due to physical limitations: QHCl–Tetralone mixtures at all
concentrations precipitated when mixed, and the amino acids (L-phe
and L-trp) when combined with the additive urea at their highest
concentration also precipitated.
The group psychophysical curves for all eight bitter compounds were examined and four concentrations corresponding to varying bitter intensities were chosen for the bitter–bitter interaction experiment. The four concentrations were from the dynamic phase of the group psychophysical curve, determined in experiment 1, and corresponded to increasing bitter intensity (Figure 1C1–C4). These are referred to as the `target' compound concentrations.
|
A weak intensity `additive' control concentration of each compound was also required for the study so that a compound could be added to itself. Due to the large individual differences in bitterness perception at a single concentration of stimulus (as detailed in experiment 1), it was necessary to divide the subject population into three sub-groups, a sensitive group, an insensitive group, and the middle group (Figure 2). Psychophysical curves were plotted for the sub-groups for each compound and the three concentrations that corresponded to a `weak' intensity were determined, one for each sub-group for each compound. Thus, the insensitive group had a concentration for their additive that was higher than the average, and the sensitive group an additive concentration that was lower. Across these sub-groups the average bitterness experienced for each additive was the same intensity, `weak'. This approach was necessary since the intensity of the additive could influence the type of perceived interaction that would occur between bitter compounds. Although it would be theoretically ideal, the preparation of individual concentrations of additives for every subject would have greatly increased the stimulus preparation time. The `additive' control concentration was mixed with the four `target' concentrations and subjects rated the taste intensities of sweet, sour, salty, bitter, and savory.
|
The additive control bitter mixture was made by adding a compound to itself at the four target concentrations.
A set concentration of sucrose corresponding to `weak' sweetness (gLMS =
5.76) was included as a taste quality control and a confirmation of the
methods. It was expected that the cognitive influence of sweetness would
inhibit bitterness in general (Kroeze and
Bartoshuk, 1985; Calvino et al.,
1990,
1993;
Frijters and Schifferstein,
1994; Breslin and Beauchamp,
1997).
Methodology
Subjects were given numbered trays of randomized bitter tasting solutions.
For each session, the solutions included deionized water as a control for
spurious ratings (n = 1), self-addition concentrations of the target
bitter stimuli (n = 4), and one `target' concentration with the
`additive' concentrations of the other seven bitter compounds (n =
7). The testing protocol was as follows. Randomized solutions (12 solutions
containing 10 ml) were presented in 30 ml plastic medicine cups on numerically
labeled trays. Subjects rinsed with deionized water at least four times over a
2 min period prior to testing. Each subject tasted and then rated each
solution for sweetness, sourness, saltiness, bitterness and savoriness, on the
gLMS before expectorating, while wearing nose-clips (GaleMed) to minimize any
olfactory input. All subjects rinsed with deionized water four times during
the interstimulus interval of 85 s. All binary bitter combinations were tasted
on at least four separate occasions.
Method verification
To ensure the method could detect non-additive interactions in taste
intensity, we conducted a parallel experiment with aspartame and acesulfame-K
(both sweeteners), which, when mixed, exhibit synergy of sweet taste
(McBride, 1988;
Ayya and Lawless, 1992;
Schiffman et al.,
1995; Schifferstein,
1996). Sucrose was used as a control sweetener, since it does not
synergize with either sweetener
(Schifferstein, 1995). All
subjects (n = 16) matched the intensity of sweeteners to gLMS 5 and
10 prior to the experiment. The group mean concentration required for each of
the sweeteners to elicit gLMS 5 or 10 intensity was determined. The method for
intensity matching followed previously published methods used in our
laboratory (Keast and Breslin,
2002a). During each session, subjects were presented with a single
concentration of a sweetener, a double concentration of the same sweetener
(self-addition control) and binary combinations of sweeteners. The tasting
procedure was the same as above. Each sample was tasted only once per session
and every binary sweet combination was tasted on at least three separate
occasions. There were a total of six sessions, three for gLMS 5 and three for
gLMS 10 solutions.
Alternative forced-choice methodology
Subjects (n = 10) were asked to determine whether a bitter-tasting
additive was more bitter than a self-addition control with a two-alternative
forced-choice (2-AFC) method. The 2-AFC method is more sensitive than the
rating method and could identify deviations from bitter-taste additivity that
were not statistically significant using the rating data. The 2-AFC procedure
was used to determine if either urea (as a bitterness inhibitor) or DB (as a
bitterness enhancer) could be distinguished from the self-addition target. The
choice of urea and DB provided the best chance to confirm a suppression or
enhancement of taste because urea tended to suppress and DB tended to enhance
bitterness. Each session consisted of six discrimination tasks with an
interstimulus interval of 85 s. Each sample pair was repeated three times for
the 10 subjects yielding 30 trials per pair. For a result to be statistically
significant (P < 0.05) using a chi-square test, one of the two
samples must be chosen as more bitter on 20 or more of the 30 trials. All
sample pairs were presented in random order.
Normalization of gLMS ratings
The standardized bitterness rating for bitter compounds tended to follow a
log-normal distribution. A normal distribution was approximated by taking the
log value of the ratings. Therefore, the log was taken of all standardized
gLMS ratings before any statistical analyses were conducted. Before taking the
log, all zero values were converted to 0.24, the lowest possible value above
zero that can be measured on the computerized gLMS.
Statistical analysis
Numerical results are expressed as geometric means + geometric standard
error [see Breslin and Tharp (Breslin and
Tharp, 2001) for calculation of geometric standard error].
Statistical variation was determined by one- or two- or three-way analysis of
variance (ANOVA) using Statistica 6.0 software package. P values <
0.05 were considered statistically significant. Individual's mean bitterness
intensity data from the binary bitter-compound experiment were analyzed by an
8 x 8 x 4 (target x additive x concentration)
repeated-measures ANOVA.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Experiment 1
Table 1,
Figure 2 (top) and
Figure 3 illustrate the wide
range in the perceived bitterness intensity of compounds used in this study.
Table 1 shows concentrations of
the bitter compounds that correspond to three intensities, gLMS 4, 8 and 12,
as well as the range of individual ratings of bitterness at those
concentrations. Figure 3 shows
psychophysical curves plotted for the group, and representative curves from
typical insensitive and sensitive subjects (sensitivities for an individual
varied from compound to compound). These results complement other studies that
illustrate the high variability of bitterness perception within a population
(Yokomukai et al.,
1993; Delwiche et
al., 2001; Keast and
Breslin, 2002b). PROP's psychophysical curve was included in this
phase of the research, although PROP was not one of the compounds used in the
binary bitter interactions phase due to the high proportion of the population
that is insensitive. Urea, L-phe, and L-trp were
perceived as being the least bitter. The limitations of solubility for
L-trp, L-phe, and SOA in aqueous solutions determined
the maximum bitterness of those compounds. Thus, for these three compounds,
the highest concentration tested was the maximum practical solubility.
|
|
Table 2 shows the results of the correlation analyses at gLMS 4, 8 and 12. In general, the correlations between bitter compounds were more frequent at gLMS 4 and diminished as the intensity increased. For example, at gLMS 4 the bitterness of L-phe was correlated with five other compounds. At gLMS 8 (Table 2), the bitterness of L-phe was only correlated with one other compound, and at gLMS 12 (Table 2) L-phe did not correlate with any compounds. This illustrates that the concentrations of bitter compounds is an important variable to account for when assessing bitter taste interactions. The bitterness of PROP did not correlate with the other bitter compounds at any intensity.
|
Figure 4 shows the results of the descriptive cluster analyses (single linkage, Euclidean Distance) at the three concentration levels. The placement of compounds at the three intensities is similar to results from the correlation matrices. As the perceived intensity increased, the linkage distance among compounds also increased. There were two tight groupings at gLMS 4, the first being ranitidine, L-trp and L-phe, while the second was SOA and QHCl. As the intensity of bitterness increased, the separation of these tight groupings was evident. PROP was always the outlier in these analyses.
|
The analysis indicates that at higher concentrations the compounds become more distantly connected and linkages appear more uniform. Data from Figure 2 (bottom graph) support these observations where three groupings of subjects are evident according to perceived intensity at low concentrations of ranitidine, while at higher concentrations of ranitidine (upper graph), the perceived bitterness intensity for the majority of subjects is more evenly distributed over a wide range of intensities. Thus, at low concentrations, some low sensitivity subjects become moderately sensitive at high concentrations, and some high sensitivity subjects become moderately sensitive at high concentrations. This results in both weaker correlations and weaker linkages among compounds at higher concentrations.
Experiment 2
Figure 5 shows the pooled (across four target concentrations and across all the target compounds) effects of the bitter compounds as additives. This figure illustrates the overall influence of these additives on bitterness in mixture. There were no significant differences between bitter compounds as additives. Figure 6A–H shows the effects of additives on specific target compounds pooled across all four concentrations of the targets, which indicates how each target compound was generally influenced by each additive. The bitter additives did not significantly alter the bitterness of the target compound.
|
|
Verification of the method with sweetness
The results reveal that there are significant differences in sweetness of
binary mixtures of sweeteners: gLMS5 [F(5,55) = 9.75, P <
0.05]; gLMS10 [F(5,55) = 12.4, P < 0.05]
(Figure 7). The mixture of
aspartame and acesulfame-K significantly (P < 0.05) increase
sweetness (synergy) relative to the self-addition controls, which verifies
that the methodology is sensitive enough to confirm non-linear taste
interactions that are known to exist.
|
Binary bitter interactions
Results from an 8 x 8 x 4 (target x additive x
concentration) repeated-measures ANOVA follow. There was a significant main
effect of the `target' compounds [F(7,35) = 3.2, P <
0.05]. This indicates that the bitterness of the `target' compounds differed
overall.
There was a significant main effect of concentration [F(3,15) = 19.4, P < 0.001], indicating that the bitterness significantly increased as the concentration of the target compound increased.
There was no main effect of the `additive' [F(7,35) = 1.9, P = 0.09] and no interaction between the `target' compound and the `additive' [F(49,245) = 1.4, P = 0.051], indicating that additives affected the bitterness of all compounds equally (Figures 5 and 6A–H).
There was a significant interaction between the `target' compound and the concentration [F(21,105) = 5.9, P < 0.001], indicating the bitterness intensity of target compounds increased differentially as the concentration increased. There was a significant interaction between the `additive' compound and concentration [F(21,105) = 1.93, P < 0.05], indicating the some additives interact with target concentrations differently than other additives.
There was a significant three-way interaction between the `target' compound, the `additive' compound, and the concentration [F(147,735) = 1.3, P < 0.05], indicating that specific `target', `additive' and `concentration' combinations were different in bitterness from each other. Overall there were very few significant differences among the bitter compounds (see below for specific interactions). Note that these effects do not appear in Figure 6, since responses have been averaged across concentration levels in the figure.
Bitter compounds as `additives'
Figure 5 shows the average
bitterness intensity ratings when the bitter stimuli and sucrose were added to
the target bitter compounds. There were no significant differences between
bitter compounds (8 x 8 x 4 ANOVA). Results from an 8 x 9
x 4 (target x additive x concentration) repeated-measures
ANOVA revealed that sucrose (sweet), as an additive, was significantly
(P < 0.05) more effective at suppressing bitterness than most
bitter compounds, except urea and L-trp.
There were concentration specific non-additive binary interactions (results not shown). Tukey HSD analysis of target–additive concentration interactions revealed that urea inhibited the bitterness of L-phe, QHCl and ranitidine at low intensities (P < 0.05) (see below for urea's forced choice results). SOA suppressed the bitterness of urea and QHCl at low intensities (P < 0.05). In addition, the amino acids L-trp and L-phe suppressed QHCl bitterness at low intensity (P < 0.05).
In general, the vast majority of the 218 unique binary interactions between bitter compounds were not statistically significant, meaning that the bitterness among these compound mixtures at a variety of concentrations and intensities combined additively.
Two-alternate forced-choice method assessing urea and denatonium
benzoate as `additives'
Figure 5 shows that bitter
mixtures with DB as an additive were rated on average LMS15 and bitter
mixtures with urea as a component were on average LMS10. While an ANOVA failed
to find a significant difference in bitterness between these additives, the
difference was large enough to warrant further investigation. A
two-alternative forced-choice procedure was used to directly assess whether
the bitter compounds DB or urea, as additives, significantly affected
bitterness in relation to self-addition controls. Results from this highly
sensitive method showed that subjects were unable to distinguish between the
intensities of DB as an additive or the self-addition control, thereby
illustrating that the bitterness of DB was perceptually additive. Urea
suppressed the bitterness of QHCl and L-phe at all four
concentrations, SOA and ranitidine at all concentrations except the lowest,
and DB and L-trp all concentrations except the highest (P
< 0.05). Addition of urea to Tetralone had no effect on bitterness. This
demonstrated that urea inhibits bitterness, although the effect is both
compound and concentration dependent.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Experiment 1
Increasing the concentration of bitter compounds decreases the
differences among individuals in bitterness sensitivities
The correlation and cluster analysis from the lowest intensity level (LMS
4) supports the hypothesis that bitterness in humans appears to be transduced
via several heterogeneous mechanisms. The individual differences in bitter
intensity ratings of the nine compounds indicate three clusterings: one for
PROP; one for L-trp, L-phe and ranitidine; and one for
SOA and QHCl.
When comparing experiment 1 to the parallel study of Delwiche et
al. (Delwiche et al.,
2001) there were 29 binary combinations of bitter compounds in
common, and on only five occasions were there differences in binary-pair
bitterness correlations between the two experiments. Cluster analysis also
revealed strong similarities between the two studies. Delwiche et al.
reported tight clusters among L-trp, L-phe and urea and
among QHCl, SOA and DB. In the present experiment,
Figure 4A shows that
L-phe and L-trp cluster tightly with urea less related,
and SOA and QHCl cluster tightly with DB somewhat less related.
Interestingly, as the concentration of the bitter compounds was increased, the correlations between bitter compounds decreased (Table 2). For example, no inter-compound correlations persisted at all three intensity levels; and only three pairs of compounds correlated at two intensities (ranitidine and L-phe, QHCl and SOA, and Tetralone and SOA). Cluster analyses in Figure 4A–C, show a similar pattern; the tight clusters loosen as the bitterness intensity increases. At the highest intensity, the clusters of bitter compounds are more evenly distributed (except for PROP), essentially forming one large cluster. These data indicate that individual differences to bitter tasting compounds that were evident at low intensity levels become less prominent the more intense the bitter compounds are. That is, the population becomes more evenly distributed about the y-axis at higher concentrations (see Figure 2 for example).
PROP
Many studies report that sensitivity to the compound PROP correlates with
sensitivities to several other bitter compounds
(Hall et al., 1975;
Bartoshuk, 1979;
Lawless, 1979;
Gent and Bartoshuk, 1983;
Leach and Noble, 1986;
Bartoshuk et al.,
1988) and an equal number of studies show no correlations with
PROP (Mela, 1989;
Schifferstein and Frijters,
1991; Yokomukai et
al., 1993; Schiffman
et al., 1994;
Delwiche et al.,
2001). In the present study, the perceived bitterness of PROP did
not correlate or cluster with the bitterness of any other compounds at any
intensity. We conclude that one's sensitivity to PROP does not predict
sensitivity to the bitterness of these other compounds
(Delwiche et al.,
2001).
Experiment 2: bitter–bitter interactions
While there were exceptions, most binary bitter mixtures combined additively with respect to taste and did not show interactions. The few interactions that occurred were suppressive and only occurred at weak intensities, with the added compound decreasing the bitterness in comparison to the target compound's self-addition control.
Urea as a component in a binary mixture of bitter compounds
Urea was effective at suppressing the bitterness of most compounds with the
exception of Tetralone using 2-AFC. Therefore, we suggest that the bitter
tasting compound urea is a bitter taste suppressor
(Keast and Breslin, 2002a).
Urea's influence over bitterness may be due to an oral peripheral effect,
rather than a cognitive effect. The primary reason for suggesting an oral
peripheral effect is that urea did not suppress the bitterness of Tetralone.
Such compound-specific differences indicate that the site of urea's bitterness
suppression is likely in the oral periphery and is independent of mechanisms
involved with Tetralone, rather than a cognitive influence affecting perceived
bitterness generally. This latter type of cognitive interaction was found with
the additive sucrose, which was effective at inhibiting the bitterness of all
compounds tested, including Tetralone. At present, the mode of bitterness
inhibition by urea is unknown.
Rejection of false negatives
The primary finding of this study is that bitter-tasting compounds do not
interact when in binary mixtures. There were a couple notable exceptions to
this rule, mentioned above, but they were suppressions rather than synergies.
Therefore, the question arises as to whether the methods employed in the
present study could detect taste synergy. The sweet taste control study
demonstrated that compounds that are expected to show synergy (aspartame and
acesulfame-K) in fact do, and those that are not expected to show synergy
(sucrose and aspartame or sucrose and acesulfame-K) do not
(Figure 7). Thus, it appears
that if bitter mixtures were synergizing perceptually, the present methods
would have detected this.
Bitter taste as a linear, additive combinatorial system
The majority of `bitter' compound binary mixtures did not interact
significantly (bitterness was additive). Therefore, taste receptor cells and
higher taste relays generally act as simple, additive, bitter-taste
integrators and convey a signal to higher cognitive centers that reflects the
total amount of bitterness-inducing compounds present in the mouth. Since it
may be important to accurately relay information regarding amounts of toxins
being ingested in foods (including foods with multiple classes of toxins),
this strategy may be the most informative and maximize survival. Although we
recognize that not all bitter-tasting compounds are toxic and not all toxins
taste bitter, we believe that the bitter taste system evolved to detect toxins
in foods. Virtually all foods contain relatively low levels of bitter-tasting
toxins (Leiner, 1969); yet we
must eat them. The strategy of the taste system appears to be to keep an
additive tally of what bitter toxins are in the mouth and track total levels
of different potential toxins ingested.
|
Acknowledgments |
|---|
The authors wish to thank Gary Beauchamp and Beverly Cowart for their comments on a draft of this manuscript. In addition, many thanks are given to Melissa Tepper for her technical assistance. This research was supported by a grant from NIH DC02995 to P.A.S.B. and a research grant from Firmenich SA to R.S.J.K. and P.A.S.B.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Adler, E., Hoon, M., Mueller, K., Chandrashekar, J., Ryba, N. and Zuker, C. (2000) A novel family of mammalian taste receptors. Cell,100 ,693 –702.[CrossRef][Web of Science][Medline]
Ayya, N. and Lawless, H. (1992)
Quantitative and qualitative evaluation of high-intensity sweeteners and
sweetener mixtures. Chem. Senses,17
, 245–259.
Bartoshuk, L. (1979) Bitter taste of
saccharin related to the genetic ability to taste the bitter substance
6-n-propylthiouracil. Science,205
,934
–935.
Bartoshuk, L. (2000) Comparing sensory
experience across individuals: recent psychophysical advances illuminate
genetic variation in taste perception. Chem. Senses,25
, 447–460.
Bartoshuk, L., Rifkin, B., Marks, L. and Hooper, J.
(1988) Bitterness of KCl and benzoate: related to genetic
status for sensitivity to PTC/PROP. Chem. Senses,13
, 517–528.
Bartoshuk, L., Duffy, V., Lucchina, L., Prutkin, J. and Fast, K. (1998) PROP (6-n-propylthiouracil) supertasters and the saltiness of NaCl. Ann. N.Y. Acad. Sci., 855,793 –796.[CrossRef][Web of Science][Medline]
Belitz, H.D. and Wieser, H. (1985) Bitter compounds: occurrence and structure–activity relationship. Food Rev. Int., 1,271 –354.
Breslin, P. and Beauchamp, G. (1997) Salt enhances flavour by suppressing bitterness.Nature , 387,563 .[Medline]
Breslin, P. and Tharp, C. (2001)
Reduction of saltiness and bitterness after a chlorhexidine rinse.Chem. Senses
, 26,105
–116.
Caicedo, A. and Roper, S. (2001)
Taste receptor cells that discriminate between bitter stimuli.Science
, 291,1557
–1560.
Calvino, A., Garcia-Medina, M., Cometto-Muniz, J. and Rodriguez, M. (1993) Perception of sweetness and bitterness in different vehicles. Percept. Psychophys.,54 , 751–758.[Web of Science][Medline]
Calvino, A.M., Garcia-Medina, M.R. and Cometto-Muniz,
J. (1990) Interactions in caffeine-sucrose and
coffee–sucrose mixtures: evidence of taste and flavor suppression.Chem. Senses
, 15,505
–519.
Chandrashekar, J., Mueller, K., Hoon, M., Adler, E., Feng, L., Guo, W., Zuker, C. and Ryba, N. (2000) T2rs function as bitter taste receptors. Cell,100 ,703 –711.[CrossRef][Web of Science][Medline]
Delwiche, J.F., Buletic, Z. and Breslin, P.A.S. (2001) Covariation in individuals' sensitivities to bitter compounds: evidence supporting multiple mechanisms. Percept. Psychophys., 63,761 –776.[Web of Science][Medline]
Frijters, J. and Schifferstein, H. (1994) Perceptual interactions in mixtures containing bitter tasting substances. Physiol. Behav.,56 ,1243 –1249.[CrossRef][Medline]
Gent, J.F. and Bartoshuk, L.M. (1983)
Sweetness of sucrose, neohesperidin dihydrochalcone, and saccharin is
related to the bitter substance 6-n-propylthiouracil.Chem. Senses
, 7,265
–272.
Green, B., Dalton, P., Cowart, B., Shaffer, G., Rankin, K.
and Higgins, J. (1996) Evaluating the `labeled
magnitude scale' for measuring sensations of taste and smell.Chem. Senses
, 21,323
–334.
Green, B.G., Shaffer, G.S. and Gilmore, M.M.
(1993) Derivation and evaluation of a semantic scale of oral
sensation magnitude with apparent ratio properties. Chem.
Senses, 18,683
–702.
Hall, M., Bartoshuk, L., Cain, W. and Stevens, J. (1975) PTC taste blindness and the taste of caffeine.Nature , 253,442 –443.[CrossRef][Medline]
Keast, R.S.J. and Breslin, P.A.S.
(2002a) Cross adaptation and bitterness inhibition of
l-tryptophan, l-phenylalanine and urea: further support for shared peripheral
physiology. Chem. Senses, 27,123
–131.
Keast, R.S.J. and Breslin, P.A.S. (2002b) Modifying the bitterness of selected oral pharmaceuticals with cation and anion series of salts. Pharm. Res., 19,1020 –1027.
Keast, R.S.J. and Breslin, P.A.S. (2003) An overview of binary taste–taste interactions.14 , 111–124.
Kim, U., Jorgenson, E., Coon, H., Leppert, M., Risch, N. and
Drayna, D. (2003) Positional cloning of the human
quantitative trait locus underlying taste sensitivity to
phenylthiocarbamide. Science, 299,1221
–1225.
Kroeze, J. and Bartoshuk, L. (1985) Bitterness suppression as revealed by split-tongue taste stimulation in humans. Physiol. Behav., 35,779 –783.[CrossRef][Medline]
Lawless, H. (1979) The taste of creatine
and creatinine. Chem. Senses, 4,249
–258.
Leach, E.J. and Noble, A.C. (1986)
Comparison of bitterness of caffeine and quinine by a time–intensity
procedure. Chem. Senses, 11,339
–345.
Leiner, I. (1969) Toxic Constituents of Plant Foodstuffs. New York: Academic Press.
McBride, R.L. (1988) Taste reception of binary sugar mixtures: pyschophysical comparison of two models.Percept. Psychophys. , 44,167 –171.[Web of Science][Medline]
Mela, D. (1989) Bitter taste intensity: the effect of tastant and thiourea taster status. Chem. Senses, 14,339 –345.
Rifkin, B. and Bartoshuk, L. (1980) Taste synergism between monosodium glutamate and disodium 5'-guanylate. Physiol. Behav.,24 ,1169 –1172.[CrossRef][Medline]
Schifferstein, H. (1995) Prediction of
sweetness intensity for equiratio aspartame/sucrose mixtures.Chem. Senses
, 20,211
–219.
Schifferstein, H. (1996) An equiratio model
for non-additive components: a case study for aspartame/acesulfame-k
mixtures. Chem. Senses, 21,1
–11.
Schifferstein, H. and Frijters, J.E.R.
(1991) The perception of taste of KCl, NaCl, and quinine-HCl
is not related to prop sensitivity. Chem. Senses,16
, 303–317.
Schiffman, S., Gatlin, L., Frey, A., Heiman, S., Stagner, W. and Cooper, D. (1994) Taste perception of bitter compounds in young and elderly persons: relation to lipophilicity of bitter compounds. Neurobiol. Aging, 15,743 –750.[CrossRef][Web of Science][Medline]
Schiffman, S., Booth, B., Carr, B., Losee, M., Sattely-Miller, E. and Graham, B. (1995) Investigation of synergism in binary mixtures of sweeteners. Brain Res. Bull., 38,105 –120.[CrossRef][Web of Science][Medline]
Spielman, A., Huque, T., Whitney, G. and Brand, J.G. (1992) The diversity of bitter taste transduction mechanisms, In Corey, D.P. and Roper, S.D. (eds), Sensory Transduction. Rockefeller Press, New York, pp.307 –324.
Yamaguchi, S. (1967) The synergistic taste effect of MSG and disodium 5'-inosinate. J. Food Sci., 32,473 –478.[CrossRef][Web of Science]
Yokomukai, Y., Cowart, B.J. and Beauchamp, G.
(1993) Individual differences in sensitivity to bitter
tasting substances. Chem. Senses,18
, 669–681.
Accepted March 26, 2003
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  W. Meyerhof, C. Batram, C. Kuhn, A. Brockhoff, E. Chudoba, B. Bufe, G. Appendino, and M. Behrens The Molecular Receptive Ranges of Human TAS2R Bitter Taste Receptors Chem Senses, February 1, 2010; 35(2): 157 - 170. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Riera, H. Vogel, S. A. Simon, S. Damak, and J. le Coutre Sensory Attributes of Complex Tasting Divalent Salts Are Mediated by TRPM5 and TRPV1 Channels J. Neurosci., February 25, 2009; 29(8): 2654 - 2662. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Behrens, S. Foerster, F. Staehler, J.-D. Raguse, and W. Meyerhof Gustatory Expression Pattern of the Human TAS2R Bitter Receptor Gene Family Reveals a Heterogenous Population of Bitter Responsive Taste Receptor Cells J. Neurosci., November 14, 2007; 27(46): 12630 - 12640. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S.J. Keast and J. Roper A Complex Relationship among Chemical Concentration, Detection Threshold, and Suprathreshold Intensity of Bitter Compounds Chem Senses, March 1, 2007; 32(3): 245 - 253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S.J. Keast, T. M. Canty, and P. A.S. Breslin Oral Zinc Sulfate Solutions Inhibit Sweet Taste Perception Chem Senses, July 1, 2004; 29(6): 513 - 521. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S.J. Keast, T. M. Canty, and P. A.S. Breslin The Influence of Sodium Salts on Binary Mixtures of Bitter-tasting Compounds Chem Senses, June 1, 2004; 29(5): 431 - 439. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||