Amino Acids

, Volume 43, Issue 6, pp 2349–2358

Gustatory sensation of l- and d-amino acids in humans

Authors

    • Section of Oral Neuroscience, Graduate School of Dental SciencesKyushu University
    • Institute for InnovationAjinomoto Co., Inc.
  • Yuki Sekine-Hayakawa
    • Institute for InnovationAjinomoto Co., Inc.
  • Atsushi Okiyama
    • Quality Assurance & External Scientific AffairsAjinomoto Co., Inc.
  • Yuzo Ninomiya
    • Section of Oral Neuroscience, Graduate School of Dental SciencesKyushu University
Original Article

DOI: 10.1007/s00726-012-1315-x

Cite this article as:
Kawai, M., Sekine-Hayakawa, Y., Okiyama, A. et al. Amino Acids (2012) 43: 2349. doi:10.1007/s00726-012-1315-x

Abstract

Amino acids are known to elicit complex taste, but most human psychophysical studies on the taste of amino acids have focused on a single basic taste, such as umami (savory) taste, sweetness, or bitterness. In this study, we addressed the potential relationship between the structure and the taste properties of amino acids by measuring the human gustatory intensity and quality in response to aqueous solutions of proteogenic amino acids in comparison to d-enantiomers. Trained subjects tasted aqueous solution of each amino acid and evaluated the intensities of total taste and each basic taste using a category-ratio scale. Each basic taste of amino acids showed the dependency on its hydrophobicity, size, charge, functional groups on the side chain, and chirality of the alpha carbon. In addition, the overall taste of amino acid was found to be the combination of basic tastes according to the partial structure. For example, hydrophilic non-charged middle-sized amino acids elicited sweetness, and l-enantiomeric hydrophilic middle-sized structure was necessary for umami taste. For example, l-serine had mainly sweet and minor umami taste, and d-serine was sweet. We further applied Stevens’ psychophysical function to relate the total-taste intensity and the concentration, and found that the slope values depended on the major quality of taste (e.g., bitter large, sour small).

Keywords

Amino acidTasteHuman psychophysics

Introduction

Taste stimuli are important for mastication, swallow, and digestion of foods through secretion of digestive fluids such as saliva, pancreatic juice (Ohara et al. 1988) etc. Recently, many of taste receptors have been detected in various tissues other than taste receptor cells in taste buds on tongue, and their contribution to nutrient sensing is becoming clear. For example, sweet receptor, T1r2/T1r3, in L cells of gut was found to regulate secretion of an incretin, GLP-1 (Jang et al. 2007). Amino acids elicit taste, however, very little has been known about the taste properties of amino acids because of their complexity.

Among the proteogenic amino acids, l-glutamate (Glu) has the best-known taste properties. Glu is one of prototypical substances responsible for umami (savory) taste, which has been proposed as a fifth basic taste, along with sweetness, saltiness, sourness, and bitterness (Ikeda 1909; Yamaguchi 1987; Ninomiya and Funakoshi 1989). A heterodimeric G-protein coupled receptor, T1r1/T1r3, has been proposed as a candidate receptor for Glu and its elicitation of umami taste (Li et al. 2002; Nelson et al. 2002). However, few human psychophysical studies have been conducted on the taste of amino acids other than Glu.

Solms showed that many l-amino acids elicit sweet or bitter tastes, while most d-amino acids have primarily a sweet taste (Solms et al. 1965). Schiffman and her colleagues (Schiffman and Dackis 1975; Schiffman et al. 1981) also examined the taste quality of l- and d-amino acids by asking subjects to rate amino acid powders for 45 attributes, including hedonic components such as good/bad, using a semantic differential scale. Taste intensity was evaluated using amino acid solutions and subjects were asked to evaluate the overall taste intensity by magnitude estimation without standard stimulus, i.e. modulus. Therefore, taste intensity could not be compared between amino acids. Ninomiya et al. (1966) measured the taste properties of l-amino acid solutions. They used the iso-intense NaCl concentration as a measure of the overall taste intensity and a subjective ratio of each basic taste to the overall taste intensity for taste quality. Yoshida and Saito (1969) used a multidimensional scaling method to analyze the relationship between taste qualities of l-amino acids. These studies provide information about the intensity of overall taste, but do not explain the intensity of each basic taste and its importance in defining the complex taste of an individual amino acid.

Since 2000, in addition to the above umami receptor T1r1/T1r3, the gustatory receptors for the other four basic tastes have been identified by genetic analysis and molecular biology techniques (Chandrashekar et al. 2000; Nelson et al. 2001; Li et al. 2002; Huang et al. 2006; Chandrashekar et al. 2010). Mechanisms of receptor binding also have been explored by studies on the response of mutated receptors to taste molecules and by study of the docking of a taste molecule onto the predicted receptor structure especially about umami and sweet tastes (Zhao et al. 2003; Zhang et al. 2008). However, little is yet known about the taste of amino acids other than Glu.

In the present study, we addressed the potential relationship between the structure and the taste properties of proteogenic amino acids by measuring of human gustatory responses elicited by aqueous solutions. Taste of their d-enantiomers was also measured for comparison. A more in-depth evaluation of taste complexity was obtained by rating of the intensities of total taste (TT) and each basic taste, that is, sweetness (Sw), saltiness (Sa), sourness (So), bitterness (Bi) and umami taste (Um), using a category-ratio scale, labeled magnitude scale (LMS) (Green et al. 1993).

Materials and methods

Subjects

All psychophysical tests followed the regulations for sensory evaluation of the Ajinomoto Co., Inc. Eight subjects (4 men and 4 women, non-smokers, 20–40 years old) who were working for food research and development at Ajinomoto Co., Inc. and familiar to the sensory evaluation of foods participated as subjects after providing informed consent. The subjects had no chronic disease and required no medical treatment. They were given a suprathreshold solution for each basic taste in order to instruct them with regard to each basic taste; these solutions consisted of 150 mmol/l sucrose (Sw), 100 mmol/l NaCl (Sa), 0.67 mmol/l l-tartaric acid (So), 3.1 mmol/l caffeine (Bi), and 100 mmol/l monosodium l-glutamate (MSG) (Um). All subjects were able to discriminate the five basic taste qualities from the following dilute aqueous solutions: 11.7 mmol/l sucrose, 22.4 mmol/l NaCl, 0.33 mmol/l l-tartaric acid, 1.04 mmol/l caffeine, and 2.7 mmol/l MSG, and water.

Prior to evaluation, the subjects were trained to rate the intensity of each quality of taste with the LMS following a previously published procedure (Green et al. 1993, 1996) using suprathreshold solutions representing each basic taste or a mixture of tastes. They were instructed to choose the nearest descriptor to the perceived intensity, then to adjust the position on the LMS. They were strictly instructed to imagine that the descriptor “strongest imaginable” at the top of the LMS meant the strongest oral sensation that would be equivalent to severe pain. Before the sessions for amino acids, subjects had experienced the evaluation of each taste quality of single or mixture solutions (e.g., various concentrations of NaCl solutions, NaCl plus sucrose mixed solutions) using LMS repeatedly. As all of the subjects rated the 150 mmol/l NaCl solution near “moderate” on the LMS, 150 mmol/l NaCl solution was provided to the subjects as the reference for “moderate” in TT intensity at the start of every session. Subjects were asked to quit session when they perceived that the first sample was different from usual. They were asked to refrain from eating or drinking 1 h prior to the evaluation procedure.

Stimuli

Aqueous solutions were prepared using ultrapure water purified using a Milli-Q water purification system (Nihon Millipore K.K, Japan) and guaranteed reagent grade l-form protein amino acids and their d-form enantiomers (Table 1). l- and d-tyrosine were omitted because they were barely soluble in water and their saturated solutions had no taste. Amino acids in free form (i.e., not in salt form) were purchased from Nacalai Tesque, Inc. (Japan). The purity of each amino acid was higher than 98.0 %, except for l-lysine (97.0 %). Three geometrically progressive concentrations, low (L), medium (M), and high (H), for each amino acid were determined by two or three trained subjects before the sessions. The range of concentration was set so that TT intensity was stronger than “weak” and weaker than “strong” on the LMS to avoid the range subthreshold or over sensationally saturated concentration.
Table 1

Test solutions

Typea

Amino acid

Physicochemical properties

l-form

d-form

  

Size

Hydrophobicity

pI

Functional group

prb

Lc

Mc

Hc

prb

Lc

Mc

Hc

I

Gly

Small

Low

Neutral

 

4.0

63

250

1,000

    

Ala

Aliphatic

3.0

67

200

600

3.0

67

200

600

Ser

Hydroxyl

4.0

63

250

1,000

2.0

250

500

1,000

Thr

Hydroxyl

2.5

100

250

625

2.5

112

280

700

II

Cys

Small–medium

Medium

Neutral

Sulfhydryl

2.0

100

200

400

Met

Sulfide

4.0

12.5

50

200

2.0

50

100

200

Val

Aliphatic (branched)

2.0

25

50

100

2.0

50

100

200

III

Leu

Large

High

Neutral

Aliphatic (branched)

2.5

16

40

100

2.0

30

60

120

Ile

Aliphatic (branched)

2.5

16

40

100

90

Phe

Aromatic

3.0

8.3

25

75

2.5

8.0

20

50

Tyr

Aromatic, hydroxyl

Trp

Heterocyclic (indol)

2.5

4.0

10

25

4.0

1.0

4.0

16

IV

Asp

Small–medium

Low

Acidic

Carboxyl

2.0

1.5

3.0

6.0

3.0

0.67

2.0

6.0

Glu

Carboxyl

4.0

0.63

2.5

10

3.0

0.67

2.0

6.0

V

Asnd

Medium

Low

Neutral

Amide

3.0

16.7

50

150

3.0

16.7

50

150

Gln

Amide

2.0

62.5

125

250

2.0

50

100

200

VI

His

Medium–large

Medium–high

Basic

Heterocyclic (imidazole)

2.0

25

50

100

3.0

10

30

90

Lys

Amine

2.0

50

100

200

Arg

Guanidino

2.0

5.0

10

20

2.5

8.0

20

50

VII

Pro

Medium

Medium

Neutral

Imino acid

4.0

63

250

1,000

2.0

95

190

380

L low, M medium, H high concentration (mmol/l)

aClassified by psychophysical properties

bProgressive ratio for concentration of the test solutions

cConcentration of the test solutions

dBoth l-and d-forms of Asn monohydrate were used

Psychophysical methods

Solutions and ultrapure water for a mouth rinse were kept at room temperature, 20–25 °C, when they were presented to the subjects. Approximately 40 ml of each solution was served in a plastic cup. After three mouth rinses with water, subjects sipped approximately 10 ml of each amino acid solution, tasted the solution, and then spat it out. To reduce the effects of odor on taste ratings, subjects wore nose clips during the tasting. Subjects were asked to rate the intensity of each taste quality, TT, Sw, Sa, So, Bi and Um, using an LMS 100 mm in length. Oral sensation which could not be ascribed to basic tastes was also evaluated as other taste (Ot). The relative position of each descriptor, “barely detectable (BD)”, “weak (W)”, “moderate (Mod)”, “strong (S)”, “very strong (VS)” and “strongest imaginably (SI)”, was the same as that of the original LMS. Subjects were allowed to taste repeatedly after the mouth rinse until all taste qualities were rated. Solutions of 2 or 3 amino acids (6–9 solutions) were presented in one session. The order of presentation of l- or d-amino acids was randomized. The subjects participated the sessions up to once a day and up to three times a week.

Data analysis

Labeled magnitude scale ratings were transformed into logarithmic values after zero (“no sensation”) was replaced with a very small number, “0.01”. Average and standard error of were calculated using logarithmic values for the intensity of TT, Sw, Sa, So, Bi, Um, and Ot. Stevens’ law was applied to relate concentration and TT intensity. The goodness of fit of the linier regression was evaluated by one-way ANOVA. The concentration at Mod-TT intensity was calculated according to the psychophysical functions for the comparison of absolute intensity. The correlation of the slope values or the concentrations at Mod-TT intensity between l- and d-forms was evaluated by Spearman’s correlation coefficient (r).

Results and discussion

In this study, we asked subjects to taste each amino acid solution and to rate the intensities of TT and each basic taste, Sw, Sa, So, Bi and Um, and Ot using a category-ratio scale, LMS. This method allowed measurement and comparison of the qualitative and quantitative properties of taste for each amino acid.

The results are summarized by the type of amino acids, classified by physicochemical properties, that is, hydrophobicity, size, isoelectric point (pI) and the functional group of the side chain (Table 1).

Taste quality

Mean intensities of TT, Sw, Sa, So, Bi, Um and Ot of each amino acid solution are shown in Fig. 1a for l-amino acids, and Fig. 1b for d-amino acids. The taste quality of each solution was summarized in Table 2.
https://static-content.springer.com/image/art%3A10.1007%2Fs00726-012-1315-x/MediaObjects/726_2012_1315_Fig1a_HTML.gif
https://static-content.springer.com/image/art%3A10.1007%2Fs00726-012-1315-x/MediaObjects/726_2012_1315_Fig1b_HTML.gif
Fig. 1

Taste of each amino acid solution. Taste of l-amino acids (a), and d-amino acids (b). Each bar represents geometric mean ± geometric standard error for each quality of taste, TT (white), Sw (pink), Sa (light blue), So (red), Bi (yellow green), Um (orange) and Ot (gray). Horizontal axis represents three concentration levels L, M and H (see Table 1). The labels on the right handy axes represent the descriptors on the LMS. Graph for Gly is presented in both a and b

Table 2

Summary of taste quality

Type

Amino acid

l-form

d-form

L

M

H

L

M

H

I

Gly

Sw

Sw

Swum

 

Ala

Sw

Sw

Swum

Sw

Sw

Sw

Ser

Sw

Sw

Swumso

Sw

SwSo

SwSo

Thr

Sw

Swso

Swso

Sw

Sw

Sw

II

Cys

Bi

Bi

BiSw

Met

Bi

Bisw

Bisw

Swbi

SwBi

SwBi

Val

Bi

Bi

Bisw

SwBi

Swbi

Swbi

III

Leu

Bi

Bi

Bi

Swbi

Swbi

Swbi

Ile

Bi

Bi

Bi

Swbi

Phe

Bi

Bi

Bi

Swbi

Swbi

Swbi

Trp

Bi

Bi

Bi

Sw

Sw

Swbi

IV

Asp

So

So

So

So

So

So

Glu

So

So

Soum

So

So

So

V

Asn

So

So

Soum

Sw

Sw

SwSo

Gln

Sw

Sw

Swum

Sw

Swso

SwSo

VI

His

Bi

Bi

Bi

Sw

Sw

Swbi

Lys

SwBi

BiSw

Bi

Arg

Bi

Sw

Bi

BiSw

SwBi

BiSw

VII

Pro

BiSw

BiSw

SwBi

Bi

Bi

Bi

Taste qualities are written according to intensity. Tastes in large letters represent the strongest taste, and the taste that was stronger both than half of the strongest taste and BD. Tastes in small letters represent the taste that was weaker than half of the strongest taste, and stronger both than 1/10 of the strongest taste and BD

For l-amino acids, the most predominant taste qualities could be categorized based on their physicochemical properties; type I (Gly, Ala, Ser, Thr): Sw, type II (Cys, Met, Val): Bi, type III (Leu, Ile, Phe, Trp): Bi, type IV (Asp, Glu): So, type V (Asn, Gln): So or Sw, type VI (His, Lys, Arg): Bi, type VII (Pro): Sw. On the other hand, most of d-amino acids had predominantly Sw except for types IV and VII amino acids. l-Amino acids of types I, IV and V elicited Um at high concentration. On the contrary, none of d-amino acids elicited Um.

Many of the l- and d-amino acids were evaluated as having more than just a single basic taste. In addition, changes in taste quality, the ratio of intensity of each basic taste, were noted in response to the concentration.

Sweetness and bitterness

Sweetness and Bi of neutral amino acids of types I, II and III were linked to hydrophobicity, size and chirality, similar to previous reports (Shallenberger et al. 1969; Birch and Kemp 1989). In l-amino acid cases, Sw and Bi strongly depended on hydrophobicity (>Val: Bi, Cys>: Sw) and size (>Cys: Bi, Met>: Sw). Thus, hydrophilic small type I amino acids elicited Sw without Bi, and hydrophobic large type III amino acids elicited Bi without Sw. Type II amino acids whose properties are between types I and III had Sw and Bi mixed taste. Meanwhile, Sw of d-amino acids had reverse tendency, hydrophobic and large amino acid showed strong Sw, and d-Trp was the sweetest as reported by Deutsch and Hansch (1966). His, a type VI amino acid with a imidazol group, also elicited Bi as the l-form, while the d-form elicited Sw comparable to d-Phe. The particularly strong Sw of d-Trp, d-Phe and d-His might be due to pi–pi interaction between the side chain of these amino acids and the aromatic residues on the Sw receptor proteins. The Bi side taste of d-amino acids of types II and III, and d-His was supposed to be derived from the hydrophobic large side chain like Bi of l-amino acids.

In the case of imino acid Pro, d-Pro was expected to be Sw because its hydrophobicity and size were intermediate between types I and III. However, Pro was predominantly Sw in l-form, and Bi in d-form. This relationship between l- and d-forms was reverse of other amino acids.

In the cases of Lys and l-/d-Arg of basic type VI amino acids, dominant taste was evaluated as Bi and/or Sw, although these amino acids showed complicated taste.

Sourness and umami taste

Amino acids with lower pI than Ser and lower hydrophobicity than Thr, except for Gln and d-Thr had So. Acidic amino acids of type IV (Asp, Glu) elicited primarily a So response.

Though Asp elicited BD Um within the range of the experimental concentration, neutralized salt of Glu and Asp e.g., monosodium l-glutamate and monosodium l-aspartate, are well known to have Um without So (Yamaguchi et al. 1971). In addition to type IV l-amino acids, types I and V amino acids of higher pI than Ala and of smaller than Gln were found to elicit Um at high concentrations. We have shown previously by a human psychophysical experiment that synergistic Um enhancement (Kuninaka 1960; Yamaguchi 1967) occurs when l-amino acids of types not only IV but also types I and V are mixed with 5′-inosine monophosphate (Kawai et al. 2002). This phenomenon is similar to responses seen in in vitro experiments with murine T1R1/T1R3-expressing cells, in which the cellular response to many l-amino acids, such as Gly and Ser, can be enhanced by 5′-guanosine monophosphate (Nelson et al. 2002). The weak Um response of l-amino acids alone is therefore possibly enhanced by adding nucleotides, suggesting that the T1R1/T1R3 heterodimeric receptor for acidic l-amino acids might be involved in the reception of these l-amino acids. However, unlike a previous report on human T1R1/T1R3 by Li et al. (2002), we found that a wider variety of amino acids could elicit the umami taste in our experiment.

On the other hand, no d-amino acid elicited Um, even if a response was elicited for their l-forms. The phenomenon whereby many l-amino acids, apart from the well known type IV Asp and Glu, elicited umami taste might be consistent with the concept that umami is a signal for protein nutrition.

Taste intensity

The relationships between log-transformed concentration and TT intensity are shown in Fig. 2a for l-amino acids, and Fig. 2b for d-amino acids, with the regression lines of Stevens’ psychophysical function. The main effect of the regression was highly significant, except for Cys, Lys, Arg and d-Arg (p < 0.05). Slope value and estimated concentration at Mod-TT intensity are summarized in Table 3.
https://static-content.springer.com/image/art%3A10.1007%2Fs00726-012-1315-x/MediaObjects/726_2012_1315_Fig2_HTML.gif
Fig. 2

Total-taste intensity. TT intensity with regression lines for psychophysical function of l-amino acids (a), and d-amino acids (b)

Table 3

Summary of taste intensity

Type

Amino acid

l-form

d-form

Ratio (l/d)

Slopea

pb

Mod-TTc

Slopea

pb

Mod-TTc

Slope

Mod

I

Gly

0.78

1.64E−09

6.4E+02

  

Ala

0.57

6.21E−08

2.9E+02

0.62

3.91E−07

3.4E+02

0.92

0.85

Ser

0.54

7.34E−09

4.5E+02

0.69

3.45E−06

8.2E+02

0.78

0.54

Thr

0.84

4.81E− 07

4.6E+02

0.71

4.06E−06

7.3E+02

1.19

0.63

II

Cys

0.42

1.73E−02

2.6E+02

Met

0.72

2.07E−07

1.5E+02

0.52

4.39E−05

1.1E+02

1.37

1.41

Val

1.18

7.20E−05

8.1E+01

0.83

2.32E−06

2.4E+02

1.42

0.34

III

Leu

1.51

5.84E−05

8.2E+01

0.56

1.10E−03

8.8E+01

2.68

0.93

Ile

1.38

1.88E−04

6.8E+01

Phe

1.14

3.26E−03

5.0E+01

0.66

4.17E−09

2.6E+01

1.74

1.91

Trp

1.14

3.05E−07

1.4E+01

0.54

3.11E−07

7.3E+00

2.09

1.93

IV

Asp

0.57

3.08E−03

3.0E+00

0.42

2.72E−05

2.3E+00

1.35

1.29

Glu

0.43

2.08E−06

4.6E+00

0.54

2.50E−08

4.2E+00

0.81

1.10

V

Asn

0.49

1.50E−04

3.1E+02

0.66

5.09E−09

1.2E+02

0.75

2.67

Gln

0.79

9.44E−04

3.7E+02

0.69

9.61E−04

2.0E+02

1.14

1.82

VI

His

0.77

9.39E−03

1.1E+02

0.53

4.16E−05

4.2E+01

1.46

2.59

Lys

0.40

2.55E−02

1.2E+02

Arg

0.53

2.80E−02

4.0E+01

0.37

2.66E−02

4.9E+01

1.44

0.82

VII

Pro

0.62

3.87E−06

4.6E+02

0.95

1.11E−03

3.2E+02

0.66

1.43

aSlope value of the Stevens’ psychophysical function

bSignificance of fitting

cConcentration at Mod-TT intensity estimated by the psychophysical function (mmol/l)

Slope values for psychophysical functions

The order of mean ranks of slope value of psychophysical functions for l-amino acids was types III, I, II, VII, V, VI, IV. For d-amino acids, it was types VII, I, V, II, III, IV, VI. No correlation was noted for slope values between l- and d-amino acids [r = 0.314 (p = 0.240)] (Fig. 3). In particular, the difference in slope between l- and d-amino acids was substantial when their dominant quality of taste was different, as observed especially in types III and VII; e.g., one enantiomer was Sw, while the other was Bi. In summary, a relationship was noted between the predominant taste and the slope value common to l- and d-amino acids, with “Bi > Sw > So”. The phenomenon that amino acids have the same slope values for Stevens’ psychophysical functions can be interpreted as the amino acids having apparent multiple concentrations. That is, when amino acids A and B have similar taste quality, the following relationship holds:
$$ f_{\text{A}} \left( {x_{\text{A}} } \right) = f_{\text{A}} \left( {l \times x_{\text{B}} } \right) $$
where fA is the psychophysical function for amino acid A, xA is the concentration of amino acid A, and l is a constant.
https://static-content.springer.com/image/art%3A10.1007%2Fs00726-012-1315-x/MediaObjects/726_2012_1315_Fig3_HTML.gif
Fig. 3

Correlation of slope values between l- and d-amino acids. Symbols on the horizontal axis indicate that data are not available for d-amino acids

However, the observed tendency of the slope differed from that reported in previous studies by Schiffman and Clark (1980), Schiffman et al. (1982). For example, they reported that the slope values for most of the d-amino acids were larger than those for l-amino acids, and no conclusive relationship could be determined between slope values and either structure or taste quality. Contrary to our results, they reported a substantial difference between the slopes of Leu (0.475) and Ile (0.675), even though Leu and Ile are very similar in structure.

Among the Sw–Bi amino acids, type VI basic amino acids, Lys, Arg, and d-Arg, had Bi and/or Sw intensities that were weakly dependent on concentration as shown in Fig. 1. However, Bi intensity of neutralized salts of these basic amino acids showed a strong concentration-dependency (data not shown), suggesting that these bulky side chains might be linked to Bi taste. Alkaline pH of solution of these basic amino acids could perturb Sw and Bi taste sensation.

Relative taste intensity

The concentration at Mod-TT intensity, calculated according to psychophysical function, showed a strong correlation between l- and d-amino acids [r = 0.882 (p = 0.001), for 15 amino acids] (Fig. 4). This relationship was coincident with the correlation in threshold of l- and d-forms observed by Schiffman et al. (1981). In addition, a high correlation was noted between the concentrations at Mod-TT intensity and the threshold concentrations reported by Schiffman et al. (1981), r = 0.676 (p = 0.004) for 19 l-amino acids (except Tyr), r = 0.910 (p = 0.001) for 15 d-amino acids (except for Cys, Ile, Lys, Arg and Tyr). However, the mechanism that gives rise to the correlation in taste intensity between l- and d-amino acids that completely differ from each other in taste quality is not clear because it is likely that their receptors may differ from each other. Fine structure analysis of receptors will be necessary to determine the underlying mechanism.
https://static-content.springer.com/image/art%3A10.1007%2Fs00726-012-1315-x/MediaObjects/726_2012_1315_Fig4_HTML.gif
Fig. 4

Correlation of log-transformed concentrations at moderate total-taste intensity between l- and d-amino acids. Symbols on the horizontal axis indicate that data are not available for d-amino acids

Possible mechanism for the mixed taste of amino acids

The taste of amino acids was apparently perceived as a mixed sensation of basic tastes. The present results suggest that the taste quality and intensity are linked to the combination of the taste properties based on the partial structure of the amino acid. That is, amino acids may interact with more than one taste receptor, as occurs for the odorant receptor system, in which one odorant molecule is received by multiple odor receptors according to its structure (Malnic et al. 1999; Araneda et al. 2000). Hence, the change in taste quality by concentration could have been observed through the taste which was hidden at low concentration and perceived at high concentration.

However, several exceptions to this hypothesis were observed in the cases where Bi was a minor taste of type III d-amino acids, and for So of l- and d-enantiomers of Glu and Ser. In the former case, although Bi of d-Trp was expected to be the stronger than that of d-Phe due to the hydrophobicity and size of the side chain, Bi of d-Trp was in fact weaker than that of d-Phe. In the case of Glu and Ser, each enantiomer was expected to elicit the same So intensity as they have the same pH at a given concentration. However, the So of the l-form was weaker than that of the d-form. One possible mechanism might be a suppression of Bi by Sw (Kroeze and Bartoshuk 1985) or a suppression of So by Um (Yamaguchi and Kimizuka 1979) which are known to occur due to inter-molecular interactions. Such intra-molecular taste–taste interaction might be also involved in complex taste of amino acids.

Conclusion

We measured the qualitative and quantitative properties of taste of amino acids using a category-ratio scale, LMS. The complex taste of amino acids could be described by a combination of the intensities of five basic tastes according to taste properties that depended on the partial structure of each individual amino acid.

Conflict of interest

Misako Kawai, Yuki Sekine-Hayakawa, Atsushi Okiyama are employees of Ajinomoto Co., Inc.

Copyright information

© Springer-Verlag 2012