Synthesis and sensory characterization of novel umami-tasting glutamate glycoconjugates

Article abstract of DOI:10.1021/jf0344441

Two glycoconjugates of glutamic acid, namely, the N-glycoside dipotassium N-(D-glucos-1-yl)-L-glutamate (1) and the corresponding Amadori compound N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid (2), have been synthesized in yields of 35 and 52%, respectively, using new Maillard-mimetic approaches, and their chemical structures have unequivocally been elucidated by 1D- and 2D-NMR and MS experiments. Systematic sensory studies revealed that both glycoconjugates exhibit pronounced umami-like taste with recognition taste thresholds of 1-2 mmol/L, close to that of monosodium glutamate (MSG). Contrary to an aqueous solution of MSG, 1 does not show the sweetish and slightly soapy by-note, but evokes an intense umami taste. Aqueous solutions of 2 were described by the descriptors umami, seasoning, and bouillon-like. Added to a bouillon base, which did not contain any taste enhancers, both glycoconjugates imparted a distinct umami character similar to the control sample containing the same amount of MSG on a molar basis. To the best of our knowledge, these types of glycoconjugates in general and, in particular, N-glucosyl glutamate and N-deoxyfructosyl glutamate have never been reported as taste active compounds having umami-like properties. Therefore, 1 and 2 represent a new class of umami-type taste compounds showing properties similar to the umami reference compound MSG. Systematic ¹³C NMR measurements revealed that 1 was fairly stable in aqueous solutions under alkaline conditions (pH 8-10) as well as in dry form. However, it rapidly hydrolyzes in neutral and acidic solutions, giving rise to glucose and glutamate. In contrast, glycoconjugate 2 was observed to be rather stable in aqueous solution as well as in the presence of human saliva.

Full text of DOI:10.1021/jf0344441

5428 J. Agric. Food Chem. 2003, 51, 5428−5436
Synthesis and Sensory Characterization of Novel
Umami-Tasting Glutamate Glycoconjugates
†
†,§
‡
‡,§
ERSAN BEKSAN, PETER SCHIEBERLE, FABIEN ROBERT, IMRE BLANK,
‡
‡
LAURENT BERNARD FAY, HEDWIG SCHLICHTHERLE-CERNY, AND
,
†,§,#
THOMAS HOFMANN*
Deutsche Forschungsanstalt f u¨ r Lebensmittelchemie, Lichtenbergstrasse 4, 85748 Garching, Germany,
and Nestec Ltd., Nestl e´ Research Center, Vers-chez-les-Blanc, 1000 Lausanne 26, Switzerland
Two glycoconjugates of glutamic acid, namely, the N-glycoside dipotassium N-(D-glucos-1-yl)-L-
glutamate (1) and the corresponding Amadori compound N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid
(2), have been synthesized in yields of 35 and 52%, respectively, using new Maillard-mimetic
approaches, and their chemical structures have unequivocally been elucidated by 1D- and 2D-NMR
and MS experiments. Systematic sensory studies revealed that both glycoconjugates exhibit
pronounced umami-like taste with recognition taste thresholds of 1-2 mmol/L, close to that of
monosodium glutamate (MSG). Contrary to an aqueous solution of MSG, 1 does not show the sweetish
and slightly soapy by-note, but evokes an intense umami taste. Aqueous solutions of 2 were described
by the descriptors umami, seasoning, and bouillon-like. Added to a bouillon base, which did not contain
any taste enhancers, both glycoconjugates imparted a distinct umami character similar to the control
sample containing the same amount of MSG on a molar basis. To the best of our knowledge, these
types of glycoconjugates in general and, in particular, N-glucosyl glutamate and N-deoxyfructosyl
glutamate have never been reported as taste active compounds having umami-like properties.
Therefore, 1 and 2 represent a new class of umami-type taste compounds showing properties similar
to the umami reference compound MSG. Systematic 13C NMR measurements revealed that 1 was
fairly stable in aqueous solutions under alkaline conditions (pH 8-10) as well as in dry form. However,
it rapidly hydrolyzes in neutral and acidic solutions, giving rise to glucose and glutamate. In contrast,
glycoconjugate 2 was observed to be rather stable in aqueous solution as well as in the presence of
human saliva.
KEYWORDS: Umami; taste; N-glucoside; Maillard reaction; Amadori product; glutamate; MSG; N-(D-
glucos-1-yl)-L-glutamate; N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid; saliva
INTRODUCTION
T1R1 and T1R3, are able to form a universal heterodimeric
sensor for L-glutamate (3, 4).
The Japanese word “umami” means “delicious” and is used
as a synonym for the characteristic sensory properties of
monosodium glutamate, abbreviated as MSG (1). Very recent
molecular-biological investigations led to the identification of
a truncated form of the metabotropic glutamate receptor
mGluR4, which is present in the brain, from rat taste buds and
suggested this protein as a taste receptor for L-glutamate (2).
The identification of this novel G-protein coupled receptor in
taste tissue provided evidence for the existence of glutamate-
like taste as a basic taste quality. This was further strengthened
by the recent finding that two taste receptor proteins, namely,
Besides glutamic acid, also adenosine-5′-monophosphate (5′-
AMP), inosine-5′-monophosphate (5′-IMP), and guanosine-5′-
monophosphate (5′-GMP) are well-known to show umami-like
sensory characteristics (5). These purine-5′-nucleotides occur
in many savory foods such as meat, fish, seafood, and
mushrooms and are able to enhance the flavor and mouthfeel
including the impression of creaminess and viscosity of savory
dishes. Therefore, these are widely used as ingredients and taste
enhancers in culinary and snack products.
In addition to their distinct sensory quality, another peculiar
effect of these compounds is their mutual taste synergism. The
synergistic effects between MSG and purine-5′-nucleotides and
analogues of both groups were observed more than 35 years
ago (6, 7), reporting an exponential increase in the glutamate-
like taste intensity of an aqueous solution of MSG when 5′-
IMP was added. According to these results the taste activity of
*
Corresponding author [telephone (49) 251-83-33-391; fax (49) 251-
8
3-33-396; e-mail thomas.hofmann@uni-muenster.de].
†
‡
§
#
Deutsche Forschungsanstalt f u¨ r Lebensmittelchemie
Nestec Ltd.
These authors contributed equally to this work.
Current address: Institut f u¨ r Lebensmittelchemie, Universit a¨ t M u¨ nster,
Corrensstr. 45, D-48149 M u¨ nster, Germany.
1
0.1021/jf0344441 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/19/2003

Umami-Tasting Glycoconjugates
J. Agric. Food Chem., Vol. 51, No. 18, 2003 5429
Figure 1. Formation of N-(D-glucos-1-yl)-L-glutamate (1) and N-(1-deoxyfructos-1-yl)-L-glutamate (2) from D-glucose and L-glutamic acid.
′-GMP is 2.3-fold the effect of 5′-IMP, while monosodium
5
the rat mGlu4a receptor and inhibited the binding of L-AP4, a
selective L-glutamate agonist for mGlu4 receptors (21).
Besides the N-acylation of glutamic acid either by lactic acid
or by hydrophilic amino acids, the Maillard reaction between
glutamic acid and reducing sugars yields glutamic acid deriva-
tives that might exhibit umami-like and mouth-watering sensa-
tions as well as flavor-enhancing characteristics. It is well
accepted that the reaction between amino acids and aldoses first
proceedes through the corresponding N-glycoside, 1 (Figure
1), which is, however, not stable and can be rearranged to
produce the so-called Amadori product, 2 (Figure 1), as the
first relatively stable amino acid glycoconjugate (22). Such
Amadori products are important aroma precursors, formed as
intermediates during drying or heat treatment of foods such as
fruits, milk powder, cocoa, coffee, and meat (23-25). To the
best of our knowledge, Amadori compounds, however, have
not yet been reported as taste compounds exhibiting glutamate-
like organoleptic properties. The objectives of the present
investigation were, therefore, (i) to synthesize the glycoconju-
gates N-(glucos-1-yl)-L-glutamate and N-(1-deoxy-D-fructos-1-
yl)-L-glutamic acid, (ii) to evaluate their organoleptic properties,
and (iii) to study their stability in aqueous media.
aspartate possesses only about 7% of the potency of mono-
sodium glutamate.
Investigations on the structural requirements in compounds
exhibiting a taste-enhancing effect led to the suggestion that
two negative charges at a distance of 3-9, preferably 4-6,
carbon atoms are the essential structural motif required for taste
enhancement (8). Apart from L-glutamic acid, its lower homo-
logue L-aspartic acid also fulfills this postulate, as well as the
C4-dicarboxylic acids succinic acid and tartaric acid, exhibiting
some kind of umami taste and meeting the structural requirement
for two negative charges (8). Indeed, aqueous solutions of
sodium succinate show taste qualities similar to solutions of
sodium glutamate (9). Furthermore, lactic acid was found to
contribute to the glutamate-like taste of foods such as beef
bouillon and stewed beef juice (10, 11).
Besides some amino acids, organic acids, and purine-5′-
nucleotides, hydrophilic di-, tri-, and tetrapeptides containing
polar side chains, such as Glu-Asp, Glu-Glu, Asp-Glu, Thr-
Glu, Asp-Glu-Ser, Glu-Gly-Ser, and Asp-Asp-Asp-Asp, have
also been described as eliciting a lingering umami-like taste
and mouthfeel similar to that of MSG (12-14). The role of
these peptides as umami compounds is, however, controversial.
Some researchers described these di- and tripeptides as neutral
or even slightly bitter and generally questioned the existence
of umami peptides (15), whereas others reported that tripeptides
such as Glu-Leu-Glu or Glu-Asp-Phe, having a hydrophobic
amino acid residue and at least one acidic and another either
acidic or hydrophilic amino acid residue, impart mouthfeel and
a glutamate-like taste to foods (16). Particular emphasis was
put on the so-called “delicious” peptide Lys-Gly-Asp-Glu-Glu-
Ser-Leu-Ala, which was detected in papain-treated beef and
allegedly has taste-enhancing qualities comparable to MSG (17).
Recent synthetic and sensory experiments, however, could not
confirm the glutamate-like taste properties of the octapeptide,
but a sour and slightly astringent taste was detected (18-20).
Recently, N-lactoyl-L-glutamate, the condensation product of
lactic acid and glutamic acid, was reported to evoke a bouillon-
like umami taste and to have a long-lasting and mouth-watering
sensation in combination with flavor-enhancing characteristics,
although weaker compared to MSG (16). In agreement with
the sensory qualities of N-lactoyl-L-glutamate assessed by a
human taste panel, a comprehensive molecular-biological
investigation demonstrated that N-lactoyl-L-glutamate activated
MATERIALS AND METHODS
Chemicals. The following chemicals were obtained from commercial
sources: monopotassium L-glutamate (MPG) monohydrate, anhydrous
D-glucose, trifluoromethanesulfonic acid anhydride, trifluoro acetic acid
(
Aldrich, Buchs, Switzerland); anhydrous MPG (Sigma, Germany);
sulfuric acid (98%), acetonitrile, formic acid, acetone, methanol (H
2
O
<
0.02%), hexane, ethyl acetate, dichloromethane, paraffin, potassium
hydroxide (Merck, Darmstadt, Germany), D
2
O (99.8%); ammonium
acetate, sodium hydroxide, magnesium sulfate, hydrochloric acid, 2,6-
di-tert-butyl-4-methylpyridine, N,N-dimethylformamide, glutamic acid
dimethyl ester (Fluka, Buchs, Switzerland). HPLC grade water was
purified using a Milli-Q system (Millipore, Volketswil, Switzerland).
A model bouillon base without taste enhancers was used in sensory
experiments (PTC Kemptthal, Switzerland).
Synthesis of Dipotassium N-(D-Glucos-1-yl)-L-glutamate (1). A
solution of potassium hydroxide (167 mmol) in methanol (100 mL)
was added to MPG monohydrate (167 mmol), and the mixture was
stirred until the amino acid was completely dissolved. Methanol (300
mL) and anhydrous D-glucose (100 mmol) were then added, and the
suspension was refluxed for 1 h at 75 °C while stirring. After cooling
the reaction mixture to ambient temperature, the volume was reduced
to 100 mL under vacuum, the surplus of the amino acid was removed

5430 J. Agric. Food Chem., Vol. 51, No. 18, 2003
Beksan et al.
Figure 2. Preparation of dipotassium N-(D-glucos-1-yl)-L-glutamate (1) and N-(1-deoxyfructos-1-yl)-L-glutamic acid (2) using Maillard-mimetic chemistry.
Figure 3. Reaction sequence used in the synthesis of N-(1-deoxyfructos-1-yl)-L-glutamic acid (2).
1
by filtration, and the filtrate was cooled to -20 °C. Dropwise addition
of acetone precipitated the N-glucoside 1 as an amorphous powder.
Concentration, cooling, and addition of acetone were repeated and
yielded additional target compound. The two batches of crude 1 were
combined and, according to the procedure described above, recrystal-
lized twice from methanol/acetone, yielding pure dipotassium N-(D-
glucos-1-yl)-L-glutamate (1; yield: 35%) as a white amorphous powder.
(100), 292 (92), 274 (24), 226 (10), 148 (4). H NMR (360 MHz; D
2
O;
COSY; arbitrary numbering of carbon atoms refers to structure 2 in
Figure 2): δ 2.04-2.10 (m, 1H, H-C(3′)), 2.39-2.55 (m, 3H,
2
H-C(3′), H-C(4′)), 3.13 (d, 2H, J ) 14.5 Hz, H-C(1), 3.42-3.98
1
3
(m, 6H, H-C(2′), H-C(3), H-C(4), H-C(5), H-C(6)). C NMR
(360 MHz; D
2
O/H
2
O; HMQC, HMBC): δ 28.51 (CH
2
, C(3′), 37.0
, C(6));
(CH , C(4′)), 55.5 (CH
2
2
, C(1)), 66.2 (CH, C(2′)); 66.7 (CH
71.8 (CH, C(5)), 72.2 (CH, C(4)), 72.7 (CH, C(3)), 98.3 (C, C(2)),
176.3 (C, C(1′)), 183.9 (C, C(5′)).
2
-
LC/MS (ESI ): m/z 146 (100), 89 (62), 75 (61), 97 (58), 61 (53), 59
1
(
44) 121 (12), 179 (11), 326 (5), 308 (5). H NMR (360 MHz; D
2
O;
DQF-COSY; arbitrary numbering of carbon atoms refers to structure
Via D-Fructose. Following the general procedure described in the
literature (26), N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid (2) was
synthesized via the intermediates 2,3:4,5-di-O-isopropylidene-â-D-
fructopyranose (3), 2,3:4,5-di-O-isopropylidene-1-O-(trifluoromethansul-
fonyl)-â-D-fructopyranose (4; 76% yield), dimethyl N-(2,3:4,5-di-O-
isopropylidene-1-deoxy-D-fructos-1-yl)-L-glutamate (5), and disodium
N-(2,3:4,5-di-O-isopropylidene-1-deoxy-D-fructos-1-yl)-L-glutamate (6).
Deprotection of the acetal groups in 6 at mild acidic conditions yielded
the target product 2 in 80% yield.
1
3
1
in Figure 2): δ 1.82 (m, 2H, H-C(3′)), 2.22 (m, 2H, H-C(4′)),
3 3
.20 (dd, 1H, J ) 8.8 Hz, J ) 8.8 Hz, H-C(2)), 3.33-3.37 (m, 2 ×
H, H-C(4) and H-C(5)), 3.41 (m, 1H, H-C(2′)), 3.45 (m, 1H,
2
3
H-C(3)), 3.67 (dd, 1H, J ) 12.3 Hz, J ) 5.4 Hz, H -C(6)), 3.85
3
a
2
3
(
dd, 1H, J ) 12.3 Hz, J ) 0.8 Hz, H
b
-C(6)), 3.89 (d, 1H, J ) 8.8
O; HMQC, HMBC): δ 33.1
, C(4′)), 63.0 (CH, C(2′)), 63.9 (CH , C(6)),
2.7 (CH, C(5)), 75.8 (CH, C(2)), 79.5 (CH, C(3)), 79.7 (CH, C(4)),
1.7 (CH, C(1)), 185.0 (C, C(1′)), 185.6 (C, C(5′)).
Synthesis of N-(1-Deoxy-D-fructos-1-yl)-L-glutamic Acid (2). Via
1
3
Hz, H-C(1)). C NMR (360 MHz; D
CH , C(3′)), 37.1 (CH
2
(
7
9
2
2
2
2,3:4,5-Di-O-isopropylidene-â-D-fructopyranose, 3 (Figure 3). D-
Fructose (1, 66 mmol) was added to a cooled mixture of concentrated
sulfuric acid (11.6 mL) and acetone (233 mL). The suspension was
stirred at room temperature until the sugar was dissolved. The solution
was kept under these conditions for 90 min and then cooled in ice. An
ice-cooled aqueous solution of sodium hydroxide (36.6 g) was gradually
added while stirring. The acetone was removed by evaporation under
reduced pressure, and the resulting aqueous suspension was extracted
with dichloromethane (3 × 100 mL). The extracts were combined,
N-Glucoside 1. Dipotassium N-(D-glucos-1-yl)-L-glutamate (1; 7.9
mmol) was refluxed for 30 min in a mixture of methanol (30 mL) and
acetic acid (1 mL). After cooling to 0 °C, acetone (20 mL) was added,
and the precipitate formed was filtered under reduced pressure and dried
under high vacuum to yield a pale yellowish powder. This procedure
was repeated twice. After dissolving the powder in water, followed by
lyophilization, N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid (2; 1.23 g;
+
5
2% yield) was obtained as a white powder. LC/MS (ESI ): m/z 310
washed with water (2 × 50 mL), and dried over MgSO
4
. Evaporation

Umami-Tasting Glycoconjugates
of the solvent yielded 3 as a crystalline solid (14.9 g) in 86% yield. H
J. Agric. Food Chem., Vol. 51, No. 18, 2003 5431
1
Sensory Analyses. Taste Quality. The samples used for sensory
evaluation were passed through a sterile filter to obtain a micro-
biologically safe product. For preliminary taste testing, monosodium
glutamate (MSG), a binary mixture of MSG and sodium chloride
(NaCl), the dipotassium salt of N-glucosylglutamate (1), and N-(1-
deoxy-D-fructos-1-yl)-L-glutamic acid (2) were dissolved at a concentra-
tion of 10 mmol/L in a mixture (1:3, v/v) of bottled water (Vitel, low
mineralization: 405 mg/L) and deionized filtered water (Millipore) to
prevent a bitter taste produced by deionized water and to avoid a taste
influence of tap water with a higher mineral content. The solution was
adjusted to pH 6.0 by addition of aqueous solutions of sodium hydroxide
(1.0 mol/L) and hydrochloric acid (1.0 mol/L), respectively. The
samples were presented to the sensory panel in coded beakers in
alternating order at room temperature (22 °C), and five panelists were
asked to describe the taste of the samples using their own descriptors.
Determination of Recognition Taste Thresholds. Taste thresholds
were determined by eight panelists using the triangle test. Seven samples
of increasing concentrations were presented, i.e., from 0.16 to 12.5
mmol/L, without adjusting the pH of the solutions.
Sensory Experiments with a Model Bouillon Base. A preliminary
taste testing with the N-glucoside 1 (10 mmol/L) and the Amadori
product 2 (10 mmol/L), respectively, was performed using a model
bouillon base (19 g/L) that did not contain any taste enhancer, such as
MSG, IMP, or GMP. By the addition of an aqueous solution of
hydrochloric acid (1.0 mol/L) the pH of the mixture was adjusted to
5.8, corresponding to the pH of a solution of the bouillon base alone
(19 g/L). The bouillon containing one of the glycoconjugates was
compared to a solution of the bouillon base alone and to a solution of
the bouillon base with added MSG (10 mmol/L, pH 5.8). The bouillon
samples were tasted at 65 °C in coded beakers. The panelists were
asked to describe the taste of the samples using their own descriptors
and to compare the taste intensities of the solutions.
NMR (360 MHz, CDCl ): δ 1.32, 1.37, 1.45, 1.51 (4 s, 3H each,
C(CH
3
2
3
)
2
), 2.35 (s, 1H, OH), 3.64 (2×d, 2H, J ) 7.9 Hz, H-C(1)),
2
3
2
3
)
.74 (dd, 1H, J ) 13.2 Hz, J ) 1.0 Hz; H
13.2 Hz, J ) 2.0 Hz, H
a
-C(6)), 3.90 (dd, 1H, J
-C(6)), 4.21 (ddd, 1H, J ) 7.9, 2.0, 1.0
Hz, H-C(5)), 4.32 (d, 1H, J ) 2.6 Hz, H-C(3)), 4.59 (dd, 1H, J )
.9, 2.6 Hz, H-C(4)). C NMR (90 MHz, CDCl
5.3 (CH ), 25.7 (CH ), 26.4 (CH ), 61.2 (CH , C(6)), 65.4 (CH
0.0-70.9 (3 CH, C(3), C(4), C(5)), 103.1 (C, C(2)); 108.5 (C), 109.0
3
3
b
3
3
1
3
7
2
7
3
): δ 23.9 (CH
3
),
3
3
3
2
2
, C(1)),
(C).
2,3:4,5-Di-O-isopropylidene-1-O-trifluoromethanesulfonyl)-â-D-fruc-
topyranose, 4 (Figure 3). Trifluoromethanesulfonic acid anhydride (1.35
mL, 11 mmol) was added dropwise at -10 °C under an atmosphere of
nitrogen to a solution of 2,6-di-tert-butyl-4-methylpyridine (2.26 g, 11
mmol) in dry dichloromethane (30 mL). The mixture was stirred and
resulted in the formation of a precipitate. To this mixture was added
dropwise a solution of 3 (1.25 g, 5.3 mmol) in dichloromethane (10
mL) while stirring. The reaction mixture was stirred for another 2 h
before adding ice-cooled water. The solution was extracted with
dichloromethane (8 × 30 mL). The combined organic extracts were
4
dried over MgSO , and the solvent was evaporated. Column chroma-
tography of the crude product on silica gel (hexane/ethyl acetate, 4:1,
v/v) and evaporation of the solvent from the appropriate fractions gave
1
4
1
(1.45 g, 76%) as a yellow syrup. H NMR (360 MHz, CDCl
3
): δ
2
.34, 1.39, 1.45, 1.55 (4 s, 3H each, C(CH
3
2
)
2
), 3.77 (dd, 1H, J ) 12.9
3
3
Hz, J ) 0.9 Hz, H
a
-C(6)), 3.92 (dd, 1H, J ) 12.9 Hz, J ) 1.7 Hz,
3
H
b
-C(6)), 4.22 (ddd, 1H, J ) 7.9, 1.7, 0.9 Hz, H-C(5)), 4.30 (d, 1H,
3
2
J ) 2.6 Hz, H-C(3)), 4.38 (d, 1H, J ) 10.5 Hz, H
a
-C(1)), 4.50 (d,
-C(1)), 4.62 (dd, 1H, J ) 7.9, 2.6 Hz, H-C(4)).
C NMR (90 MHz, CDCl ): δ 23.9 (CH ), 25.0 (CH ), 25.7 (CH ),
6.5 (CH ), 61.6 (CH , C(6)), 69.7-70.5 (3 CH, C(3), C(4), C(5)),
4.1 (CH , C(1)), 99.8 (C, C(2)), 109.2 (C), 109.8 (C), 118.0 (CF , q,
2
3
1
H, J ) 10.5 Hz, H
b
1
3
3
3
3
3
2
7
3
2
2
3
J ) 319.7 Hz).
Stability of Dipotassium N-(D-Glucos-1-yl)-L-glutamate (1) and
N-(1-Deoxy-D-fructos-1-yl)-L-glutamate (2). Solutions of 1 (50 mg)
Dimethyl N-(2,3:4,5-Di-O-isopropylidene-1-deoxy-D-fructos-1-yl)-
L-glutamate, 5 (Figure 3). The triflate derivative 4 (620 mg, 1.58 mmol)
was dissolved in anhydrous DMF (20 mL), and glutamic acid dimethyl
ester (554 mg, 2 equiv) was added to the solution. The mixture was
refluxed for 2 h, and the product formation was monitored by TLC
or 2 (50 mg) in D
2 2
O/H O (10:50, v/v; 0.7 mL) were prepared and
transferred to an NMR tube after pH adjustment with sulfuric acid.
13
C NMR experiments were performed on a regular basis over a defined
period of time.
Influence of Human Saliva on the Stability of Dipotassium N-(D-
Glucos-1-yl)-L-glutamate (1) and N-(1-Deoxy-D-fructos-1-yl)-L-
glutamate (2). Solutions of glycoconjugates 1 and 2 (10 mg/mL each),
respectively, in either phosphate buffer (pH 7.4) or human saliva, which
was obtained from 3 women and 3 men, were incubated at 37 °C in
closed vials. After 0, 5, and 20 min, samples were withdrawn and
subsequently analyzed by an LC 3000 type amino acid analyzer
(
pentane/diethyl ether, 3:2). The reaction was stopped by adding water
(
2
25 mL), and the solution was extracted with dichloromethane (8 ×
0 mL). The combined extracts were washed with tap water (6 × 20
mL), and the solvent was evaporated under reduced pressure. Column
chromatography of the crude product on silica gel (hexane/ethyl acetate,
1
yielded 5 (140 mg, 23%) as a yellow solid. H NMR (360 MHz,
CDCl ): δ 1.24, 1.31, 1.37, 1.43 (4 s, 3H each, C(CH ), 1.74 (s, 1H,
NH); 1.92-1.95 (m, 2H, H-C(3′)); 2.34 (t, 2H, H-C(4′), J ) 7.5
:1, v/v) and evaporation of the solvent from the appropriate fractions
1
(Biotronic, Maintal, Germany) using ninhydrin detection.
3
3 2
)
3
Direct Infusion Mass Spectrometry. Mass spectrometric experi-
2
ments were carried out on a Micromass Quattro-LC triple quadrupole
mass spectrometer (Micromass, Manchester, UK) equipped with a “Z-
Spray” electrospray ion source. The electrospray capillary voltage was
set to 4.0 kV, and the source block temperature to 80 °C. The cone
gas was operated at 70 L/h, and desolvation gas at 550 L/h. The
desolvation temperature was set to 150 °C. The samples were introduced
at 10 µL/min in a mixture of methanol/water (1:1, v/v). Data were
acquired in negative mode from 50 to 800 Da with a dwell time of 1
s. Data acquisition was performed using the software package MassLynx
Hz); 2.72 (2×d, 2H, J ) 12.7 Hz, H-C(1)); 3.30 (t, 1, NCH), 3.57
2
(
H
s, 3H, OCH
a
3
), 3.61 (s, 3H, OCH
3
), 3.65 (dd, 1H, J ) 13.1, 0.9 Hz,
2
-C(6)), 3.78 (dd, 1H, J ) 13.1, 2.0 Hz, H -C(6)); 4.12 (ddd, 1H,
3
b
3
J ) 7.9, 2.0, 0.9 Hz, H-C(5)), 4.21 (d, 1H, J ) 2.4 Hz, H-C(3)),
3
13
4
.47 (dd, 1H, J ) 7.9, 2.4 Hz, H-C(4)). C NMR (90 MHz, CDCl
δ 24.4 (CH ), 25.8 (CH ), 26.2 (CH ), 26.8 (CH ), 27.8 (CH , C(3′)),
0.5 (CH , C(4′)), 52.0 (OCH ), 52.1 (OCH ), 53.6 (CH , C(1)), 60.8
NCH), 61.5 (CH , C(6)), 70.4-71.9 (3 CH, C(3), C(4), C(5)), 103.7
3
):
3
3
3
3
2
3
(
(
2
3
3
2
2
C, C(2)), 108.4 (C), 109.2 (C), 173.8 (CO); 175.2 (CO).
Disodium N-(2,3:4,5-Di-O-isopropylidene-1-deoxy-D-fructos-1-yl)-
3
.4 (Micromass, Manchester, UK).
Nuclear Magnetic Resonance Spectroscopy (NMR). The samples
L-glutamate, 6 (Figure 3). 5 (140 mg, 0.36 mmol) was refluxed for 1
h in aqueous sodium hydroxide (1.0 mL, 2 N) and methanol (3.0 mL).
The solvent was evaporated, and the crude salt (6, 23% yield) was
used without further purification for the next step. C NMR (90 MHz,
D
C(3′)), 34.9 (CH , C(4′)), 53.5 (CH
C(6)), 74.7-76.6 (3 CH, C(3), C(4), C(5)), 108.6 (C, C(2)), 115.1 (C),
1
for NMR spectroscopy were prepared in Wilmad 528-PP 5 mm Pyrex
NMR tubes, using D O/H O (3:7, v/v) as the solvent (0.7 mL). The
2
2
1
3
NMR spectra were acquired on a Bruker AM-360 spectrometer,
equipped with a quadrinuclear 5 mm probe head, at 360.13 MHz for
2
O): δ 28.2 (CH
3
), 28.6 (CH
3
), 29.4 (CH
3
), 30.3 (CH
3
), 30.6 (CH
2
,
,
1
13
H and at 75.56 MHz for C. Chemical shifts given in ppm are relative
2
2
, C(1)), 65.9 (NCH), 69.6 (CH
2
1
13
to the solvent signal. One-dimensional H NMR, C NMR, distortion-
less enhancement by polarization transfer (DEPT 135), and two-
dimensional COSY and HETCOR spectra were acquired as described
previously using standard conditions (27).
15.2 (C), 184.6 (CO), 184.8 (CO).
N-(1-Deoxy-D-fructos-1-yl)-L-glutamic Acid (2). The disodium salt
6) was dissolved in trifluoroacetic acid/water (9:1, v/v, 1 mL) and
(
stirred for 2 h at room temperature. The reagent and the solvent were
evaporated under reduced pressure, and the Amadori compound was
crystallized from diethyl ether (92 mg, 82% yield) to give a pale yellow
solid.
RESULTS AND DISCUSSION
Syntheses of Glycoconjugates. It is well-known in the
literature that the reaction of amino compounds with reducing

5432 J. Agric. Food Chem., Vol. 51, No. 18, 2003
Beksan et al.
sugars leads to N-glycosides, 1 (Figure 1) as the first reaction
products (22), which, however, rapidly rearrange into the so-
called Amadori products, 2 (Figure 1) (28). Our preliminary
studies on the formation of these glycoconjugates indicated that
it is possible to stop the reaction of glucose and glutamate at
the level of N-glucosyl glutamate (1), when the reaction was
performed under nonaqueous, alkaline conditions (29). Using
a Maillard-mimetic approach, dipotassium N-(D-glucos-1-yl)-
L-glutamate (1) was obtained by a new and simple synthetic
route in two steps starting from monopotassium L-glutamate,
which was first solubilized in anhydrous methanol in the
presence of potassium hydroxide and then coupled with D-
glucose under reflux conditions (Figure 2). After precipitation,
the N-glucoside (1) was obtained in an overall yield of about
4
0%. Absolute anhydrous conditions were mandatory to avoid
Figure 4. COSY spectrum of dipotassium N-(D-glucos-1-yl)-L-glutamate
(1).
hydrolysis of the N-glucoside into the educts or its rearrangement
into the Amadori compound. On the other hand, the latter
reaction offered an elegant synthetic approach toward the N-(1-
deoxy-D-fructos-1-yl)-L-glutamic acid, 2 (Figure 2). Acidic, but
anhydrous conditions allowed dehydration and enolization,
affording the Amadori rearrangement product 2 in about 50%
yield (Figure 2). This is a new synthetic pathway to obtain
Amadori compounds in relatively high yields, which also allows
the preparation of 2 in a multigram pilot scale as well (data not
shown).
As an alternative to the Maillard-mimetic approach, a
synthetic route to 2 was developed starting from â-D-fructose,
which first was successively protected in the positions C(2)/
C(3) and C(4)/C(5) with acetone (26). The intermediate 2,3:
4,5-di-O-isopropylidene-â-D-fructopyranose, 3 (Figure 3), was
then activated at the C-1 position by a triflate group, giving
rise to 2,3:4,5-di-O-isopropylidene-1-O-(trifluoromethansul-
fonyl)-â-D-fructopyranose (4) in 76% yield. The leaving group
was then substituted by glutamic acid dimethyl ester in a 23%
yield. The substitution reaction resulted in dimethyl N-(2,3:4,5-
di-O-isopropylidene-1-deoxy-D-fructos-1-yl)-L-glutamate (5).
However, an intramolecular cyclization of the glutamate moiety
was also observed leading to the corresponding pyrrolidone
derivative. Alkaline treatment of 5 gave rise to the acetal-
protected glutamate derivative disodium N-(2,3:4,5-di-O-iso-
propylidene-1-deoxy-D-fructos-1-yl)-L-glutamate (6) by saponi-
fication and, in the case of the pyrrolidone derivative, opening
of the lactam ring. Deprotection of the acetal groups in 6 under
mild acidic conditions yielded the target product 2 in 80% yield
Figure 5. MS-ESI(−) spectrum of the dipotassium salt of N-(D-glucos-1-
yl)-L-glutamic acid (1; 10 ng/µL). D, daughter ions; P, parent ions.
the carbons of the five CHOH groups of the sugar moiety were
shifted between 79.7 and 72.7 ppm, whereas the CH OH group
2
was detected at 63.9 ppm. The signals at 63.0, 37.1, and 33.1
ppm were assigned as the methine and the two methylene groups
of the glutamate backbone, respectively. The presence of only
three methylene protons ruled out the Amadori structure in 1
and confirmed the structure of an N-glucoside. Moreover, the
anomeric proton at 3.89 ppm was observed as a doublet with a
3
coupling constant of J ) 8.8 Hz, which is characteristic for a
trans axial-axial coupling. These data suggest the â-glucopy-
ranose as the main form of N-glucoside 1.
(
Figure 3).
Direct infusion mass spectrometry operating in the negative
electrospray ionization mode confirmed the molecular weight
and the proposed structure of 1. In dilute solutions of 10 ng/
µL, the ion at m/z 326, corresponding to the molecular ion of
the open-chain form, was found to be associated with an ion at
m/z 308, which was suggested to be the cyclic form of the
molecule (Figure 5). MS/MS experiments of the ions m/z 308
and 326 confirmed the presence of both compounds occurring
individually in the synthesized sample. Fragmentation of the
molecular ion m/z 326 resulted only in m/z 146 as the daughter
ion, thus confirming that m/z 308 was not formed by dehydration
of m/z 326. In agreement with these observations, parent ion
measurement of m/z 308 did not yield the ion m/z 326, thereby
confirming the presence of an equilibrium between the cyclic
and the open-chain form of 1 (Figure 6).
Structural Analysis of Glycoconjugates 1 and 2. The
purified dipotassium salt of N-(D-glucos-1-yl)-L-glutamic acid
1) was characterized by various 1D- and 2D-NMR as well as
(
1
3
MS techniques. C NMR spectroscopy in combination with a
DEPT-135 experiment showed 11 resonance signals, six of
which corresponded to methine and methyl carbon atoms, three
as methylene carbon atoms, and two as quaternary carbon atoms.
The chemical shifts of the two carboxylate groups resonating
at 185.6 and 185.0 ppm clearly showed that these are present
in their deprotonated form. By means of homonuclear H,H-
correlation spectroscopy (COSY) (Figure 4) and heteronuclear
C,H-correlation spectroscopy (HETCOR), the complete set of
1
13
H and C signals could be unequivocally assigned. HETCOR
(data not shown) defined the anomeric proton at 3.89 ppm,
which coupled with the carbon at 91.7 ppm. This carbon
resonance signal was assigned as the anomeric carbon, thus
proving the N-glucosyl-type structure of compound 1. From this
signal, all other proton shifts of the sugar skeleton could be
attributed in the COSY experiment (Figure 4). As expected,
The Amadori-type structure of N-(1-deoxy-D-fructos-1-yl)-
13
L-glutamic acid (2) was mainly determined by C NMR/DEPT
135 data indicating four methylene carbons out of 11 carbon
signals, instead of three as detected for the N-glucoside 1. In
agreement with these suggestions, C(1) in compound 2 was

Umami-Tasting Glycoconjugates
J. Agric. Food Chem., Vol. 51, No. 18, 2003 5433
Figure 6. Mutarotation of dipotassium N-(D-glucos-1-yl)-L-glutamate (1).
Table 1. Recognition Threshold Concentrations for the Umami Taste
of N-Glucoside 1, the Amadori Product 2, and Monosodium Glutamate
Table 2. Sensory Profiling of Aqueous Solutions of Monosodium
Glutamate (MSG), Dipotassium N-(D-Glucos-1-yl)-L-glutamate (1), and
N-(1-Deoxy-D-fructos-1-yl)-L-glutamate (2)
a
threshold
concentration
sample
MSG
quality attributes
intensity
compound
[mmol/kg]b
umami, sweet, slightly soapy,
mouthfeel, lingering sensation
salty, umami, bouillon-like, mouthfeel,
long-lasting taste sensation
1−2
a
dipotassium N-(D-glucos-1-yl)-L-glutamate (1)
N-(1-deoxy-D-fructos-1-yl)-L-glutamate (2)
1.6
1.8
1.5
a
MSG+NaCl
3
monosodium glutamate (MSG)
1
bouillon-like, sweet, mouthfeel
salty, umami, bouillon-like, glutamate-
like, long-lasting taste sensation
umami, bouillon-like, sweet, mouthfeel,
viscous and lingering taste sensation
umami, bouillon-like, seasoning-like,
mouthfeel
2
3
a
Structure of compound is given in Figure 1. b The umami taste threshold was
1+NaCl
determined in water using the triangle test.
1+MSG
3
clearly defined as a methylene carbon, instead of the methine
carbon in 1. Furthermore, carbon C(2) was represented by a
quaternary carbon in 2, whereas C(2) was assigned as a methine
carbon in 1.
2
1−2
a
The samples (10 mmol/L each) were dissolved in water (pH 6.0) and
Sensory Properties of the Glycoconjugates. Prior to sensory
analysis, the purity of the glycoconjugates 1 and 2 was checked
instantaneously evaluated at 22 °C by six sensory panelists on a scale from 0
(not perceived) to 3 (intensely perceived).
1
by LC/MS (30) and H NMR spectroscopy. Preliminary sensory
experiments revealed a pronounced umami-like taste character
of aqueous solutions of 1 and 2, respectively. To study the
sensory activity of these compounds, the recognition thresholds
of 1 and 2 for the umami taste were determined by means of
triangle tests. Both glycoconjugates showed taste thresholds
between 1.5 and 2.0 mmol/L, which is close to the threshold
concentration of MSG (Table 1).
In addition, the sensory panel was asked to describe the taste
profile and to rate the perceived taste intensity of aqueous,
equimolar solutions of the glycoconjugates as well as MSG,
either alone or in mixture with NaCl, on a scale from 0 (not
perceived) to 3 (perceived as intense). Dipotassium N-(D-glucos-
Table 3. Influence of Monosodium Glutamate (MSG),
N-(D-Glucos-1-yl)-L-glutamate (1), and
N-(1-Deoxy-D-fructos-1-yl)-L-glutamate (2) on the Taste Profile of a
Model Bouillon Base
additivea
quality attributes
intensity
(
no additive)
salty, bland, fatty, weak umami taste
umami, bouillon-like, salty, slightly sweet
umami, bouillon-like, salty
1
3
2−3
2−3
MSG
1
2
umami, bouillon-like, salty, fresh
a A bouillon base (19 g/L; pH 5.8) was evaluated in either the absence or
presence of additives (10 mmol/L) at 65 °C by five sensory panelists on a scale
1
-yl)-L-glutamate (1) on its own revealed a bouillon-like, long-
lasting taste sensation and a more pronounced umami character
intensity 2.0) as compared to MSG, which was evaluated with
from 0 (not perceived) to 3 (intensely perceived).
(
which was essentially free of any taste enhancer, such as MSG
or purine-5′-ribonucleotides, was spiked with either the N-
glucoside (1) or the Amadori product 2, and the taste qualities
were compared either to a bouillon base containing MSG or to
an aqueous solution of the bouillon base alone. The pure
bouillon base was evaluated with an overall intensity of 1 and
was perceived as bland and very salty with only a weak
glutamate-like taste (Table 3). The bouillon spiked with
glycoconjugate 1, however, elicited a bouillon-like umami taste
with an intensity of 2-3, which was close to that with added
MSG (intensity 3). Similar to the effect of 1, the addition of
N-(1-deoxy-D-fructos-1-yl)-L-glutamate (2) imparted a pro-
nounced umami and bouillon-like taste (Table 3).
To the best of our knowledge, these types of glycoconjugates
in general and N-glucosyl glutamate and N-deoxyfructosyl
glutamate in particular have never been reported as taste active
compounds exhibiting umami-like properties. Therefore, gly-
cosylglutamates represent a new class of umami-type tasting
compounds showing properties similar to the umami reference
compound MSG. While the occurrence of the N-glucoside (1)
a score of 1-2 (Table 2). Contrary to an aqueous solution of
MSG, 1 did not show the sweetish and slightly soapy note, but
evoked an intense umami taste in aqueous solutions. In addition,
1
elicited a distinct mouthfeel effect comparable to that observed
for MSG (Table 2). The bouillon-like taste of 1 was further
enhanced when sodium chloride was present in equimolar
concentrations; that is, the mixture was evaluated with an
intensity of 3 and was similar to the binary mixture of MSG
and NaCl (Table 2). Also the binary mixture of 1 and MSG
was rated with an intensity of 3 and was described to impart an
umami, bouillon-like, viscous, and lingering taste sensation. The
solution containing MSG alone was judged only with an
intensity of 1-2. Similar to glycoconjugate 1, the N-(1-deoxy-
D-fructos-1-yl)-L-glutamate (2) showed the same umami and
bouillon-like overall sensation with an intensity of 1-2 as
observed for MSG (Table 2).
To evaluate the potential of these glycoconjugates in food-
type applications, the impact of 1 and 2 on the taste profile of
a model bouillon base was evaluated. The model bouillon base,

5434 J. Agric. Food Chem., Vol. 51, No. 18, 2003
Beksan et al.
indicated L-glutamate and glucose as the major hydrolysis
products (data not shown), thus suggesting that the N-glucoside
1
was relatively unstable in water. To gain a more detailed
insight into the stability of the N-glucoside, aqueous solutions
of 1 were incubated in NMR tubes as function of temperature,
reaction time, and pH value, and the hydrolysis of 1 was
13
followed by C NMR spectroscopy by monitoring the chemical
shift range between 90 and 100 ppm (Figure 7). The release of
free glucose was indicated by the increase of the resonance
signals at 95.0 and 98.8 ppm corresponding to the R- and
â-anomeric carbon atom of the hexose, respectively. As
displayed in Figure 7, the N-glucoside 1 was found to be stable
after an incubation time of 15 min at pH 9.3 (sample A) as the
signals representing the anomeric carbon atoms of free glucose
remained unchanged. In contrast, the lower pH of 6.5 (sample
B) induced a partial hydrolysis of 1 as indicated by the two
anomeric carbon atoms of glucose. Under more acidic conditions
(sample C), hydrolysis occurred nearly instantaneously; for
example, no resonance signal of 1 was detectable when the
N-glucoside was maintained at pH 2.3.
To investigate the kinetics of the degradation of the di-
potassium salt of N-(D-glucos-1-yl)-L-glutamate by hydrolysis,
1
was maintained at room temperature under different pH
13
conditions, and the hydrolysis was monitored by C NMR
spectroscopy over time. As shown in Figure 8A, 1 was
hydrolyzed relatively slowly under alkaline conditions. After
2
4 h at pH 10, the N-glucoside was still the major constituent
with some glucose liberated. With time, however, the degree
of hydrolysis increased. After about 1 week, glucose prevailed.
The stability of 1 at pH 10 is well reflected in the long half-
life, τ1/2, of about 4 days (Figure 8A). In contrast, decomposition
of 1 was found to be faster under neutral and slightly acidic
conditions (Figure 8B). The low pH of 2.5 induced a spontane-
ous decomposition of 1, which did not allow the measurement
of the half-life (data not shown). Between these extreme pH
values, the rate of decomposition increased with acidity; for
example, the τ1/2 of 1 was about 10 h at pH 7.5 and 2.5 h at pH
5.0 (Figure 8B). The hydrolytic instability of the N-glycosidic
bond in N-(D-glucos-1-yl)-L-glutamate (1) under acidic and
neutral pH conditions is well in line with earlier observations
on the stability of N-glycosides of amino acids in aqueous
solutions (32, 33).
Figure 7. 13C NMR spectra of dipotassium N-(D-glucos-1-yl)-L-glutamate
in D O at (A) pH 9.3, (B) pH 6.5, and (C) pH 2.3 after 15 min of sample
2
preparation. Full arrow indicates the C(1) signal at 91.7 ppm for 1; the
dotted arrows represent the anomeric C atoms of free glucose (95.0,
98.8 ppm).
in foods has never been reported, the Amadori compound 2 is
a known constituent of processed foods; for example, high
amounts of up to 3.6 g of 2 per 100 g of dry matter have been
reported in dried tomatoes (31), but it has also been found in
dried celery, asparagus, cauliflower, carrots, and red pepper as
well as in dark malts (25, 31).
Stability of Glycoconjugates 1 and 2. Preliminary analytical
measurements on the stability of 1 by means of RP-HPLC and
high-performance anion exchange chromatography, respectively,
As the N-glucoside 1 was shown to be rapidly degraded by
acid-catalyzed hydrolysis, the question arose as to whether the
umami taste of aqueous solutions of 1 is imparted by the
N-glucoside itself or by the free L-glutamate liberated upon
Figure 8. Influence of the pH on the time course of dipotassium N-(D-glucos-1-yl)-L-glutamate (1) decomposition.

Umami-Tasting Glycoconjugates
J. Agric. Food Chem., Vol. 51, No. 18, 2003 5435
L-glutamic acid. These quantification experiments revealed that
after 5 min no significant glutamate release could be measured
from the N-glucoside in the presence of either saliva or water
(control). Increasing the incubation time to 20 min led, however,
to the liberation of 12 and 9% glutamate, respectively, from
glycoconjugate 1 when maintained in the presence of saliva or
water (data not shown). These data show a stability of the
N-glucoside in human saliva similar to that found in pure
aqueous solution. In contrary, neither in the presence of human
saliva nor in the aqueous control system were significant
amounts of free L-glutamic acid formed from compound 2 (data
not shown), thereby clearly demonstrating that the Amadori
product is also stable in the presence of human saliva. In
summary, these data clearly show that the umami taste of 1
and 2, which is perceived immediately after taking these
compounds into the oral cavity, is caused by the glycoconjugates
thereselves, and not by free glutamic acid liberated therefrom.
ACKNOWLEDGMENT
We are grateful to F. Arce Vera and D. Welti (both Nestle´
Research) for the NMR service.
LITERATURE CITED
(
1) Ikeda, K. On a new seasoning (in Japanese). J. Tokyo Chem.
Soc. 1909, 30, 820-826.
(
2) Chaudari, N.; Landin, A. M.; Roper, S. D. A metabotropic
glutamate receptor variant functions as a taste receptor. Nat.
Neurosci. 2000, 3, 113-119.
Figure 9. 13C NMR spectra of N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid
(
3) Nelson, G.; Chandrashekar, J.; Hoon, M. A.; Feng, L.; Zhao,
G.; Ryba, N. J. P.; Zuker, C. S. An amino-acid taste receptor.
Nature 2002, 416, 199-202.
(2) in aqueous solution (pH 5.0) after 3 h (A), 24 h (B), and 96 h (C). The
full arrow indicates the C(2) signal at 98.3 ppm; the dotted arrows represent
the anomeric C atoms of free glucose (95.0, 98.8 ppm). Trace amounts
of glucose were present in the purified target compound.
(4) Li, X.; Staszewski, L.; Xu, H.; Durick, K.; Zoller, M.; Adler, E.
Human receptors for sweet and umami taste. Proc. Natl. Acad.
Sci. 2002, 99, 4692-4696.
hydrolysis. To answer that question, the pH of an aqueous
solution of 1 was measured depending on its concentration.
Because the dipotassium N-(D-glucos-1-yl)-L-glutamate (1) is
a base, as expected, the pH of an aqueous solution was
significantly changed with its concentration. The pH was found
to be stable at concentrations above 1.3 g/L (3.3 mmol/L),
slowly approximating a value of 9.1. Under these conditions,
the half-life of 1 was τ1/2 g 3 days (data not shown). The same
concentration of 1 (>3.3 mmol/L) in a buffered solution at pH
(5) Kuninaka, A. Flavor Potentiators. In Symposium on Foods: The
Chemistry and Physiology of FlaVors; Schultz, H. W., Day, E.
A., Libbey, L. M., Eds.; AVI Publishing Company: Westport,
CT, 1967; pp 515-535.
(6) Yamaguchi, S. The synergistic taste effect of monosodium
glutamate and disodium 5′-inosinate. J. Food Sci. 1967, 32, 473-
4
78.
(
7) Yamaguchi, S.; Yoshikawa, T.; Ikeda, S.; Ninomiya, T. Mea-
surement of the relative taste intensity of some L-R-amino acids
and 5′-nucleotides. J. Food Sci. 1971, 36, 1761-1765.
7
.0 resulted in a half-life of τ1/2 ≈ 10 h. Consequently, the
(8) Ney, K. H. Flavor enhancing effect of L-glutamate and similar
compounds. Z. Lebensm. Unters. Forsch. 1971, 146, 141-143
(in German; Chem. Abstr. 1971, 75, 497341).
umami taste threshold determined to be 1.6 mmol/L is without
any doubt evoked by compound 1 and not by free L-glutamate
(
9) Vel ´ı sˇ ek, J.; Dav ´ı dek, J.; Kubelka, V.; Tran Thi Bich Thu;
Haj sˇ lov a´ , J. Succinic acid in yeast autolysates and its sensory
properties. Nahrung 1978, 22, 735-743.
possibly released upon hydrolysis.
In contrast to N-glucosyl-L-glutamate (1), 13C NMR measure-
ments of aqueous solutions containing the Amadori compound
N-(1-deoxy-D-fructos-1-yl)-L-glutamic acid (2) revealed this
glycoconjugate to be rather stable (Figure 9). No increase of
the intensity of the resonance signal of the R- (95.0 ppm) and
â-anomeric carbon atom (98.8 ppm) of the free hexose could
be observed at pH 5.0, even after 3 days of storage at room
temperature.
(
10) Warendorf, T.; Belitz, H.-D.; Gasser, U.; Grosch, W. The flavor
of bouillon. Part 2. Sensory analysis of non-volatiles and
imitation of a bouillon. Z. Lebensm. Unters. Forsch. 1992, 195,
2
15-223 (in German; Chem. Abstr. 1993, 118, 5923).
(
11) Schlichtherle-Cerny, H.; Grosch, W. Evaluation of taste com-
pounds of stewed beef juice. Z. Lebensm. Unters. Forsch. 1998,
207, 369-376.
Although compounds 1 and 2 were evaluated as rather stable
at moderate pH values for at least some hours, it was further
investigated whether the umami taste is caused by these
glycoconjugates theirselves or by hydrolysis products formed
in the oral cavity, e.g., by an enzymatic cleavage of Amadoriase-
type enzymes. To achieve this, glycoconjugates 1 and 2,
respectively, were dissolved in human saliva and were incubated
in vitro at 37 °C. As the control sample, these compounds were
incubated separately in aqueous solution. After 0, 5, and 20
min of incubation at 37 °C, samples were analyzed for free
(12) Arai, S.; Yamashita, M.; Fujimaki, M. Glutamyl oligopeptides
as factors responsible for tastes of a proteinase-modified soybean
protein. Agric. Biol. Chem. 1972, 36, 1253-1256.
(
13) Nogushi, M.; Arai, S.; Yamashita, K.; Kato, H.; Fujimaki, M.
Isolation and identification of acidic oligopeptides occurring in
a flavor potentiating fraction from a fish protein hydrolysate. J.
Agric. Food Chem. 1975, 23, 49-53.
(
14) Matsushita, I.; Ozaki, S. Purification and sequence determination
of tasty tetrapeptide (Asp-Asp-Asp-Asp) from beer yeast season-
ing and its enzymatic synthesis. Pept. Chem. 1965, 323, 249-
251.

5436 J. Agric. Food Chem., Vol. 51, No. 18, 2003
Beksan et al.
(
15) van den Oord, A. H. A.; van Wassenaar, P. D. Umami peptides:
assessment of their alleged taste properties. Z. Lebensm. Unters.
Forsch. 1997, 205, 125-130.
16) Fr e´ rot, E.; Escher, S. D. Flavoured products and method for
preparing the same. WO 97/04667, 1997.
compounds. Z. Lebensm. Unters. Forsch. 1991, 192, 24-29 (in
German; Chem. Abstr. 1991, 115, 27981a).
(26) Lopez, M. G.; Gruenwedel, W. Synthesis of aromatic Amadori
compounds. Carbohydr. Res. 1991, 212, 37-45.
(27) Lin, J.; Welti, H. D.; Arce Vera, F.; Fay, L. B.; Blank, I.
Synthesis of deuterated volatile lipid degradation products to be
used as internal standards in isotope dilution assays. 1. Alde-
hydes. J. Agric. Food Chem. 1999, 47, 2813-2821.
(28) Hodge, J. E. The Amadori rearrangement. AdV. Carbohyd. Res.
1955, 10, 169-205.
(
(
(
17) Yamasaki, Y.; Maekawa, K. A peptide with delicious taste. Agric.
Biol. Chem. 1978, 42, 1761-1765.
18) van Wassenaar, P. D.; van den Oord, A. H. A.; Schaaper, W.
M. M. Taste of “delicious” beefy meaty peptide. Revised. J.
Agric. Food. Chem. 1995, 43, 2828-2832.
(
(
(
19) Hau, J.; Cazes, D.; Fay, L. B. Comprehensive study of the “Beefy
Meaty Peptide”. J. Agric. Food Chem. 1997, 45, 1351-1355.
20) Schlichtherle-Cerny, H. Taste compounds in beef juice (in
German). Thesis Technical University of Munich, 1996.
21) Monastyrskaia, K.; Lundstrom, K.; Plahl, D.; Acuna, G.;
Schweitzer, C.; Malherbe, P.; Mutel, V. Effect of the umami
peptides on the ligand binding and function of rat mGluR4a
receptor might implicate this receptor in the monosodium
glutamate taste transduction. Br. J. Pharmacol. 1999, 128, 1027-
(29) Weitzel, G.; Geyer, H.-U.; Fretzdorff, A.-M. Preparation and
stability of salts of amino acid N-glycosides. Chem. Ber. 1957,
90, 1153-1161 (in German; Chem. Abstr. 1958, 52, 65627).
(30) Schlichtherle-Cerny, H.; Affolter, M.; Cerny, C. Hydrophilic
interaction liquid chromatography coupled to electrospray mass
spectrometry of small polar compounds in food analysis. Anal.
Chem. 2003, 75, 2349-2354.
(31) Eichner, K.; Reutter, M.; Wittmann, R. In The Maillard Reaction
in Food Processing, Human Nutrition and Physiology; Finot,
P. A., Aeschbacher, H. U., Hurrel, R. F., Liardon, R., Eds.;
Birkhauser Verlag: Basel, 1990; pp 63-77.
(32) von Euler, H.; Brunius, E. Amino derivatives of sugars. Chem.
Ber. 1926, 59B, 1581-1585 (in German; Chem. Abstr. 1926,
20, 26457).
(33) Paulsen, H.; Pflughaupt, H. Glycosylamines. In The Carbohy-
drates-Chemistry and Biochemistry; Pigman, W., Horton, D.,
Eds.; Academic Press: New York, 1980; Vol. 1B, pp 881-927.
1
034.
(22) Ledl, F.; Schleicher, E. New aspects of the Maillard reaction in
foods and in the human body. Angew. Chem., Int. Ed. Engl. 1990,
2
9, 565-594.
(
23) van den Ouweland, G. M. A.; Peer, H. G.; Tjan, S. B. Occurrence
of Amadori and Heyns Rearrangement Products in Processed
Foods and their Role in Flavor Formation. In FlaVor of Foods
and BeVerages-Chemistry and Technology; Charalambous, G.,
Inglett, G. E., Eds.; Academic Press: New York, NY, 1978; pp
1
31-143.
Received for review April 30, 2003. Revised manuscript received July
(
24) Oberparleitner, S.; Ziegleder, G. Amadori compounds as aroma
precursors in cacao. Nahrung 1997, 41, 142-145.
3, 2003. Accepted July 3, 2003.
(25) Heinzler, M.; Eichner, K. The role of Amadori compounds during
cocoa processing. 1. Formation and decomposition of Amadori
JF0344441