Discovery of N2-(1-Carboxyethyl)guanosine 50-monophosphate as an umami-enhancing maillard-modified nucleotide in yeast extracts

Article abstract of DOI:10.1021/jf102899j

Sensory-guided fractionation of a commercial yeast extract involving medium-pressure RP-18 chromatography and ion-pair chromatography, followed by LC-MS/MS, LC-TOF-MS, 1D/2D-NMR, and CD spectroscopy, led to the discovery of the previously not reported umami-enhancing nucleotide diastereomers (R)- and (S)-N²-(1-carboxyethyl)guanosine 5'-monophosphate. Model experiments confirmed the formation of these diastereomers by a Maillard-type glycation of guanosine 50-monophosphate with dihydroxyacetone and glyceraldehyde, respectively. Sensory studies revealed umami recognition threshold concentrations of 0.19 and 0.85 mmol/L for the (S)- and (R)-configured diastereomers, respectively, and demonstrated the taste-enhancing activity of these nucleotides on monosodium L-glutamate solutions.

Full text of DOI:10.1021/jf102899j

10614 J. Agric. Food Chem. 2010, 58, 10614–10622
DOI:10.1021/jf102899j
Discovery of N 2-(1-Carboxyethyl)guanosine
50-Monophosphate as an Umami-Enhancing Maillard-Modified
Nucleotide in Yeast Extracts
D
ANIEL
F
ESTRING AND
T
HOMAS
H
OFMANN
*
€
€
Chair of Food Chemistry and Molecular Sensory Science, Technische Universitat Munchen,
Lise-Meitner-Strasse 34, D-85354 Freising-Weihenstephan, Germany
Sensory-guided fractionation of a commercial yeast extract involving medium-pressure RP-18
chromatography and ion-pair chromatography, followed by LC-MS/MS, LC-TOF-MS, 1D/2D-NMR,
and CD spectroscopy, led to the discovery of the previously not reported umami-enhancing nucleo-
tide diastereomers (R)- and (S)-N 2-(1-carboxyethyl)guanosine 50-monophosphate. Model experi-
ments confirmed the formation of these diastereomers by a Maillard-type glycation of guanosine
50-monophosphate with dihydroxyacetone and glyceraldehyde, respectively. Sensory studies revealed
umami recognition threshold concentrations of 0.19 and 0.85 mmol/L for the (S)- and (R)-configured
diastereomers, respectively, and demonstrated the taste-enhancing activity of these nucleotides on
monosodium L-glutamate solutions.
KEYWORDS: Yeast; umami; taste enhancer; taste dilution analysis; N 2-(1-carboxyethyl)guanosine
50-monophosphate
INTRODUCTION
yeast extracts, comparatively little is known about the presence of
additional, yet unknown, taste molecules and/or taste-modulating
compounds in yeast products.
In the food industry, yeast extracts, made from the soluble
fraction of yeasts, are used as a savory ingredient to impart a
broth-like, meaty taste impression, referred to as umami, to
soups, sauces, and snack products (1, 2). Depending on the
starting material used, namely, spent brewer’s yeast, baker’s
yeast, and torula yeast (2, 3), respectively, as well as the manu-
facturing process, such as autolysis, plasmolysis, or hydrolysis
(2-4), the yeast extracts available on the market strongly differ in
cost as well as flavor profile.
In recent years, activity-guided fractionation of foods combin-
ing liquid chromatographic separation techniques and analytical
sensory tools enabled the straightforward discovery of the key
taste molecules in red wine (18), Gouda cheese (19), dried morel
mushrooms (20), and beef broth (21), as well as previously not
reported taste modulators in edible beans (22), matured Gouda
cheese (23), chicken broth (24), and stewed beef juice (25),
respectively.
To identify previously unknown taste modulators in yeast
extract, the objective of the present study was to apply a
sensory-guided fractionation approach to an intensely umami-
like-tasting yeast extract.
Although a series of studies have been performed on the odor-
active volatile fraction of yeast extract (5-12), knowledge on the
taste-active and/or taste-modulating nonvolatiles is rather lim-
ited. The amino acids L-glutamic acid (1) and L-aspartic acid (2)
formed upon protein hydrolysis and the organic acid succinic acid
(3), as well as the purine ribonucleotides inosine 50-mono-
phosphate (4) and guanosine 50-monophosphate (5, 50-GMP),
generated upon breakdown of nucleic acids and showing a
MATERIALS AND METHODS
Chemicals and Materials. Unless stated otherwise, all chemicals were
obtained from Sigma-Aldrich (Steinheim, Germany) and were of puriss
grade; formic acid, hydrochloric acid, and sodium hydroxide were from
remarkable synergistic effect on L-glutamate (Figure 1), have been
well-known for a long time to contribute to the umami taste of
yeast extracts (2, 3, 13-16). Very recently, the molecular mecha-
nism of the synergistic effect of purine ribonucelotides on the
€
Grussing GmbH (Filsum, Germany), and disodium hydrogen phosphate
dihydrate and potassium dihydrogen phosphate were from Merck KGaA
(Darmstadt, Germany). Deuterium oxide and sodium deuteroxide (40%
w/w solution in D2O) were obtained from Euriso-Top (Gif-Sur-Yvette,
France), HPLC grade solvents were from Mallinckrodt Baker (Griesheim,
Germany), and membrane filter disks (0.45 μm) were purchased from
Satorius AG (Goettingen, Germany). Water used for chromatography
was purified by means of a Milli-Q Advantage A10 water purification
system (Millipore, Molsheim, France), and bottled water (Evian) was used
for sensory analysis. A commercially available yeast extract (Springer
2012/20-MG-L) was obtained in powdered form from the food industry in
France.
umami perception of L-glutamate was found to be mediated by a
positive allosteric modulation of the umami taste receptor due to the
fixation of the closed conformation of the receptor’s venus flytrap
domain by interaction of the negatively charged phosphate group
of the ribonucleotide with the positively charged pincer residues
of the receptor protein (17). Although many scientists have been
focusing on increasing the yields of the umami compounds 1-4 in
*Corresponding author [phone (49) 8161-71-2902; fax (49) 8161-71-
2949; e-mail thomas.hofmann@tum.de].
pubs.acs.org/JAFC
Published on Web 09/14/2010
© 2010 American Chemical Society

Article
Fractionation of Yeast Extract by Means of Medium-Pressure
Liquid Chromatography (MPLC). A sample of the yeast extract (2 g)
was dissolved in water (20 mL) and injected onto a 150 ꢀ 40 mm i.d.,
LiChroprep, 25-40 μm, RP-18 cartridge of a medium-pressure liquid
chromatographic system (RP-MPLC). Using 1% formic acid in water
(solvent A) and methanol (solvent B), chromatography (flow rate = 50
mL/min) was performed starting with 100% solvent A for 5 min and then
J. Agric. Food Chem., Vol. 58, No. 19, 2010 10615
increasing the content of solvent B to 20% within 15 min and to 100%
within additional 15 min, followed by a 5 min isocratic elution with 100%
solvent B. Ten subfractions (I-X; Figure 2) were collected, separated from
solvent in vacuum, and lyophilized twice to afford the individual fractions
in yields given in parentheses (in g/100 g): I (28.2), II (24.0), III (24.5), IV
(3.9), V (7.7), VI (2.8), VII (1.8), VIII (2.4), IX (3.3), and X (1.4). The
residues obtained were kept at -20 °C until used for sensory analysis
(Figure 3) and chemical analysis, respectively.
Isolation of (S)-N2-(1-Carboxyethyl)guanosine 50-Monophos-
phate, (S)-6. RP-MPLC fraction VIII was dissolved in a mixture
(93:7, v/v) of an aqueous triethylammonium acetate solution (250 mM
TEAA, pH 6.0) and methanol and then fractionated by means of
semipreparative ion pair high-performance liquid chromatography (IP-
HPLC) using a Microsorb-MV C18 column, 250 ꢀ 10 mm i.d., 5 μm
(Varian, Darmstadt, Germany). Monitoring the effluent at 260 nm,
chromatography was performed at a flow rate of 3.8 mL/min using an
aqueous TEAA buffer (250 mM, pH 6.0, solvent A) and methanol (solvent
B) as solvents. After isocratic elution with 7% solvent B for 10 min, the
content of solvent B was increased to 25% within 20 min and, finally, to
40% within an additional 2 min. Within 8 min the organic solvent content
was reduced to starting conditions. As shown in Figure 4A, a total of 16
fractions were collected individually, separated from solvent in vacuum,
and lyophilized to remove the buffer. Subfraction VIII-8 was subjected to
rechromatography using a preparative Microsorb-MV C18 column, 250 ꢀ
21.2 mm i.d., 5 μm (Varian) as stationary phase. Chromatography was
performed with a mixture of 1% formic acid in water (solvent A) and
acetonitrile (solvent B) at a flowrate of 18 mL/min, monitoring the effluent
at 260 nm. After isocratic elution with 0% B for 5 min, the content of B was
increased to 5% in 10 min and to 10% B within an additional 5 min,
followed by a linear gradient to 30% solvent B within 10 min and to 100%
within an additional 10 min. The UV-absorbing fractions (Figure 4B) were
collected, separated from organic solvent in vacuum, diluted 5-fold with
water, and thenlyophilized twice. UV-vis, LC-MS, and 1H NMR analysis
of subfraction VIII-8/5, obtained as an amorphous white powder after
freeze-drying, led to the identification of (S)-N2-(1-carboxyethyl)guanosine
50-monophosphate, (S)-6. Comparison of chromatographic (RP-18) and
spectroscopic data (UV-vis, LC-MS/MS, 1H NMR) with those obtained
for the reference compound generated by a Maillard-type reaction con-
firmed the identity of (S)-6.
Figure 1. Chemical structures of naturally occurring umami-active com-
pounds L-glutamic acid (1), L-aspartic acid (2), succinic acid (3), inosine
50-monophosphate (4), and guanosine 50-monophosphate (5).
(S)-N2-(1-Carboxyethyl)guanosine 50-monophosphate, (S)-6, Figure 6:
UV-vis (100 mM PO43- buffer, pH 7.5), λmax = 254, 276 nm (sh); LC-MS
(ESI-), m/z (%) 434 (100) [M - H]-; LC-TOF-MS, m/z 434.0705 (found),
m/z 434.0719 (calcd for [C13H17N5O10P]-); 1H NMR (500 MHz, D2O/
NaOD, COSY), δ 1.45 [d, 3H, J = 7.2 Hz, H-C(300)], 3.97-4.10 [m, 2H,
Figure 2. Chromatogram (λ = 260 nm) of the yeast extract separated by
means of gradient RP-MPLC.
Figure 3. (A) Taste dilution (TD) factors determined for the umami taste of the RP-MPLC fractions isolated from yeast extract in water (TDAwater) and in an
aqueous monosodium -glutamate solution (3 mmol/L in water; TDAMSG). (B) Quotient of the TD factors obtained from TDAMSG and TDAwater
L
.

10616 J. Agric. Food Chem., Vol. 58, No. 19, 2010
Festring and Hofmann
Figure 4. Semipreparative IP-HPLC chromatogram (λ = 260 nm) of fraction VIII (A) and excerpt of preparative RP-HPLC chromatogram (λ = 260 nm) of
fraction VIII-8 (B).
H-C(50)], 4.25 [q, 1H, J = 7.2 Hz, H-C(200)], 4.31 [m, 1H, H-C(40)], 4.44
[pt, 1H, J = 4.8 Hz, 1H, H-C(30)], 4.72 [pt, 1H, J = 5.2 Hz, H-C(20)],
6.02 [d, 1H, J = 5.2 Hz, H-C(10)], 8.17 [s, 1H, H-C(8)]; 13C NMR (125
MHz, D2O/NaOD, HMQC, HMBC), δ 20.2 [C(300)], 55.4 [C(200)], 66.7 [d,
2JC,P = 4.6 Hz, C(50)], 73.1 [C(30)], 77.1 [C(20)], 86.7 [d, 3JC,P = 8.7 Hz,
C(40)], 89.9 [C(10)], 118.8 [C(5)], 140.4 [C(8)], 154.8 [C(4)], 155.0 [C(2)],
162.0 [C(6)], 184.2 [C(100)].
Preparation of Compounds (R)-6 and (S)-6 by a Maillard Reac-
tion of Guanosine 50-Monophosphate. Guanosine 50-monophosphate
disodium salt hydrate (2 mmol, calculated on water-free nucleotide) and
DL-glyceraldehyde (6 mmol) were suspended in phosphate buffer (5 mL, 1
mol/L, pH 7.0) and incubated in a closed vessel at 40 °C. After 10 days, the
reaction mixture was diluted with water (10 mL) and, after membrane
filtration, fractionated by means of preparative RP-HPLC using the con-
ditions described above. UV-absorbing fractions were collected individu-
ally and, after separation of the organic solvent in vacuum, water and
formic acid residues were removed by repeated lyophilization, yielding the
title compounds (R)-6 and (S)-6 as amorphous white powders in a purity
of >95% (RP-HPLC/ELSD, 1H NMR).
some minor modifications, a mixture of 20-deoxyguanosine monohydrate
(2 mmol) and 1,3-dihydroxyacetone dimer (3 mmol) was suspended in
phosphate buffer (5 mL, 1 mol/L, pH 7.0) and thermally treated for 24 h at
70 °C in a closed vessel. After cooling, the reaction mixture was diluted
with water (10 mL), membrane filtered, and fractionated by means of
preparative RP-HPLC. Monitoring the effluent at 260 nm, chromatogra-
phy was performed using aqueous ammonium formate (5 mM, pH 7) as
eluent A and methanol as eluent B at a flow rate of 15 mL/min. After
starting chromatography with an isocratic step of 5% B for 5 min, the
content of B was increased linearly to 100% within 35 min and, then, was
kept isocratic for an additional 2 min. The UV-absorbing substances were
collected individually and, after separation of the solvent in vacuum, the
residue was repeatedly lyophilized to afford the title compounds as
amorphous yellowish powders in a purity of >95% (RP-HPLC/ELSD,
1H NMR).
(R)-N2-(1-Carboxyethyl)-20-deoxyguanosine ammonium salt, (R)-7,
3-
Figure 7: UV-vis (100 mM PO4 buffer, pH 7.5), λmax = 251, 278 nm
(sh); LC-MS (ESI-), m/z (%) 338 (100) [M - H]-; 1H NMR (500 MHz,
D2O, COSY), δ 1.50 [d, 1H, J = 7.3 Hz, H-C(300)], 2.48 [ddd, 1H, J = 3.9
Hz, 6.8 Hz, 14.1 Hz, H-C(20a)], 3.13 [pdt, 1H, J = 6.8 Hz, 13.8 Hz,
H-C(20b)], 3.82-3.93 [m, 2H, H-C(50)], 4.13 [m, 1H, H-C(40)], 4.27 [q,
1H, J = 7.3 Hz, H-C(200)], 4.69 [m, 1H, H-C(30)], 6.35 [t, 1H, J = 6.9 Hz,
H-C(10)], 7.95 [s, 1H, H-C(8)]; 13C NMR (125 MHz, D2O, HMQC,
HMBC), δ 20.5 [C(300)], 40.4 [C(20)], 55.6 [C(200)], 64.7 [C(50)], 74.1[C(30)],
87.7 [C(10)], 89.8 [C(40)], 119.7 [C(5)], 142.1 [C(8)], 154.4 [C(4)], 154.8
[C(2)], 162.0 [C(6)], 184.0 [C(100)].
(R)-N2-(1-Carboxyethyl)guanosine 50-monophosphate, (R)-6, Figure 6:
UV-vis (100 mM PO43- buffer, pH 7.5), λmax = 254, 276 nm (sh); LC-MS
(ESI-), m/z (%) 434 (100, [M - H]-); LC-TOF-MS, m/z 434.0722 (found),
m/z434.0719 (calcd for [C13H17N5O10P]-);1H NMR (500 MHz, D2O/NaOD,
COSY), δ 1.45 [d, 3H, J = 7.2 Hz, H-C(300)], 3.95-4.07 [m, 2H, H-C(50)],
4.26 [q, 1H, J = 7.2 Hz, H-C(200)], 4.30 [m, 1H, H-C(40)], 4.47 [pt, 1H, J =
4.8 Hz, 1H, H-C(30)], 4.83 [pt, 1H, J=5.3Hz,H-C(20)],5.97[d,1H,J=5.3
Hz, H-C(10)], 8.14 [s, 1H, H-C(8)]; 13C NMR (125 MHz, D2O/NaOD,
HMQC, HMBC), δ 20.3 [C(300)], 55.5 [C(200)], 66.7 [d, 2JC,P = 4.5 Hz, C(50)],
73.3 [C(30)], 76.4 [C(20)], 87.0 [d, 3JC,P = 8.2 Hz, C(40)], 90.5 [C(10)], 119.0
[C(5)], 141.2 [C(8)], 154.6 [C(4)], 154.9 [C(2)], 162.1 [C(6)], 184.1 [C(100)].
(S)-N2-(1-Carboxyethyl)guanosine 50-monophosphate, (S)-6, Figure 6:
spectroscopic data were identical to those obtained for the sample of (S)-6
isolated from the yeast extracts.
(S)-N2-(1-Carboxyethyl)-20-deoxyguanosine ammonium salt (S)-7,
3-
Figure 7: UV-vis (100 mM PO4 buffer, pH 7.5), λmax = 251, 278 nm
(sh); LC-MS (ESI-), m/z (%) 338 (100) [M - H]-; 1H NMR (500 MHz,
D2O, COSY), δ 1.46 [d, 1H, J = 7.3 Hz, H-C(300)], 2.51 [ddd, 1H, J = 4.6
Hz, 6.8 Hz, 14.0 Hz, H-C(20a)], 2.94 [pdt, 1H, J = 7.0 Hz, 13.8 Hz
H-C(20b)], 3.71 [dd, 1H, J = 6.1 Hz, 12.3 Hz, H-C(50a)], 3.80 [dd, 1H,
J = 4.0 Hz, 12.2 Hz, H-C(50b)], 4.07 [m, 1H, H-C(40)], 4.23 [q, 1H, J =
7.2 Hz, H-C(200)], 4.63 [m, 1H, H-C(30)], 6.30 [t, 1H, J = 6.6 Hz,
H-C(10)], 7.92 [s, 1H, H-C(8)]; 13C NMR (125 MHz, D2O, HMQC,
Preparation of (R)- and (S)-N2-(1-Carboxyethyl)-20-deoxygua-
nosine ((R)- 7 and (S)-7). Following a literature procedure (26) with

Article
J. Agric. Food Chem., Vol. 58, No. 19, 2010 10617
HMBC), δ 20.4 [C(300)], 40.9 [C(20)], 55.3 [C(200)], 64.6 [C(50)], 73.9 [C(30)],
87.0 [C(10)], 89.6 [C(40)], 119.2 [C(5)], 141.1 [C(8)], 154.2 [C(4)], 154.7
[C(2)], 161.8 [C(6)], 184.0 [C(100)].
stepwise diluted 1 þ 1 with water (pH 6.0), and the dilutions of each
fraction were presented to the sensory panel in order of increasing
concentrations by means of a duo test using water (pH 6.0; for TDAwater
)
Acidic Hydrolysis of (R)-6, (S)-6, (R)-7, and (S)-7. A portion (10 mg)
of compound of (R)-6, (S)-6, (R)-7, and (S)-7, respectively, was mixed with
hydrochloric acid (1 mol/L; 1 mL) and heated in closed glass vials (1.5 mL)
for 2.5 h at 100 °C. The individual hydrolysates were allowed to cool to
room temperature, diluted with water (2 mL), and adjusted to pH 3 by the
addition of sodium hydroxide solution (1 mol/L), and, then, the effluent of
the peak showing UV adsorption at 260 nm was isolated by means of
preparative RP-HPLC, using aqueous formic acid and acetonitrile as
described. After solvent removal in vacuum and lyophilization, (R)-8 and
(S)-8 were obtained as amorphous white powders.
or a serial dilution of monosodium -glutamate (3 mmol/L, pH 6.0; for
L
TDAMSG) as blank solutions. The participants were instructed to deter-
mine the dilution step at which a taste difference between sample and blank
solution could be detected. This so-called taste dilution (TD) factor
determined by the sensory subjects in three separate sessions was averaged.
Taste Recognition Threshold Concentrations. Determination of the
taste threshold concentrations of the nucleotides was performed in bottled
water (pH 6.0) using a triangle test with ascending concentrations of the
stimulus, as reportedin previous papers (27,29). The threshold value of the
sensory group was approximated by averaging the threshold values of the
individuals in three independent sessions. Values between individuals and
separate sessions differed by not more than plus or minus one dilutionstep;
as a result, a threshold value of 0.15 mmol/L for 5 represents a range of
0.075-0.30 mmol/L.
(R)-N2-(1-Carboxyethyl)guanine, (R)-8, Figure 7: UV-vis (100 mM
PO43- buffer, pH 7.5), λmax = 247, 279 nm (sh); LC-MS (ESI-), m/z (%)
222 (100) [M - H]-; 1H NMR (400 MHz, D2O, COSY), δ 1.44 [d, 3H, J =
7.2 Hz, H-C(300)], 4.24 [q, 1H, J = 7.2 Hz, H-C(200)], 7.98 [s, 1H,
13
H-C(8)]; C NMR (100 MHz, D2O, HMQC, HMBC), δ 20.4 [C(300)],
Determination of Taste-Enhancing Activity. To determine potential
taste-enhancing properties of the nucleotide derivatives, the purified target
compounds were subjected to an iso-umami test. To achieve this, an
aqueous binary solution (pH 6.0) containing the test nucleotide (0.1 mmol/L)
55.0 [C(200)], 116.0 [C(5)], 141.6 [C(8)], 155.1 [C(4)], 155.8 [C(2)], 160.9
[C(6)], 183.4 [C(100)].
(S)-N2-(1-Carboxyethyl)guanine, (S)-8, Figure 7: spectroscopic data of
(S)-8 were identical to those of the enantiomer (R)-8.
and monosodium
solutions containing ascending concentrations (in 5 mmol/L steps) of
monosodium -glutamate alone (10-100 mmol/L; pH 6.0) in a duo test.
The sensory panel was asked to determine the concentration of mono-
sodium -glutamate required to match the umami taste intensity of the
binary nucleotide/monosodium -glutamate mixture. This so-called iso-
L-glutamate (10 mmol/L) was compared to aqueous
Sensory Analyses. Precautions Taken for Sensory Analysis of Food
Fractions and Taste Compounds. Prior to sensory analysis, the fractions or
compounds isolated were, after removal of the volatiles in high vacuum
(<5 mPa), diluted with water and freeze-dried twice. GC-MS and ion
chromatographic analysis revealed that food fractions treated by that
procedure are essentially free of the solvents and buffer compounds used.
For pH adjustment of all samplesapplied in human sensory experiments to
6.0, trace amounts of formic acid (1 g/100 g) or sodium hydroxide (1 mmol/
L or 0.1 mmol/L) were used. Formic acid, which is GRAS listed as a
flavoring agent for food and feed applications, was used, because trace
amounts of this acid do not influence the sensory profiles of the test
solution. To minimize the uptake of any toxic compound to the best of our
knowledge, all of the sensory analyses were performed by using the sip-
and-spit method, which means the test materials were not swallowed but
expectorated.
L
L
L
umami concentration was approximated by averaging the values obtained
for each individual in three independent sessions.
RP-MPLC. Medium-pressure chromatography was performed on a
€
preparative Sepacore chromatography system (Buchi, Flawil, Switzerland)
consisting of two C-605-type pumps, a C-615 pump manager, a C-660
fraction collector, a manual injection port equipped with a 20 mL loop,
and a C-635-type UV detector monitoring the effluent at 260 nm. Chro-
matography (flow rate = 50 mL/min) was performed on LiChroprep,
25-40 μm, RP-18 bulk material (Merck KGaA), filled in a 150 ꢀ 40 mm i.d.
€
€
polypropylene cartridge (Buchi) using a C-670 cartridge (Buchi).
HPLC. The analytical HPLC system (Jasco, Gross-Umstadt,
Germany) consisted of a PU-2080 Plus pump, a DG-2080-53 degasser,
an LG-2080-02 gradient unit, an AS-2055 Plus autosampler with a 100 μL
loop, and an MD-2010 Plus detector. The HPLC apparatus (Jasco) for
semipreparative and preparative liquid chromatography consisted of two
PU-2087 pumps, a Degasys DG-1310 online degasser (Uniflows Co., Tokyo,
Japan), a 1000 μL gradient mixer, a 7725 i injection valve (Rheodyne,
Bensheim, Germany), and an MD-2010 Plus detector. Semipreparative
separations (flow rate = 3.8 mL/min for IP-HPLC) were performed using
a Microsorb-MV C18column, 250ꢀ 10 mm i.d., 5 μm (Varian, Darmstadt,
Germany) column and preparative chromatography (flow rate = 15 or
18 mL for RP-HPLC) using a Microsorb-MV C18 column, 250 ꢀ 21.2 mm
i.d., 5 μm (Varian). For chromatographic analysis of the isolated fractions
an aliquot was dissolved in water and analyzed on an Outstanding B C18
column, 250 ꢀ 4.6 mm i.d., 5 μm (Trentec, Rutesheim, Germany).
For reversed phase separations, chromatography was performed using
1% formic acid in water (solvent A) and acetonitrile (solvent B) as
effluents at a flow rate of 1 mL/min, starting with 100% solvent A for
5 min and increasing solvent B to 5% within 10 min, to 10% within an
additional 5 min, to 30% within the following 10 min, and, finally, to
100% B within 10 min.
Training of the Sensory Panel. Thirteen subjects (11 women and 2 men,
ages 22-30 years), who gave consent to participate in the sensory tests of
the present investigation and have no history of known taste disorders,
were trained to evaluate the taste of aqueous solutions of the following
standard taste compounds in bottled water (pH 6.0): sucrose (50 mmol/L)
for sweet taste, lactic acid (20 mmol/L) for sour taste, NaCl (20 mmol/L)
for salty taste, caffeine (1 mmol/L) for bitter taste, monosodium
L
-glutamate (3 mmol/L) for umami taste. For the puckering astringency
and the velvety astringent mouth-drying oral sensation, the panel was
trained by using tannic acid (0.05%) and quercetin-3-O-β- -glucopyrano-
D
side (0.01 mmol/L), respectively, using the half-tongue test. Additionally
the panel was familiarized to differentiate several umami taste qualities by
tasting solutions of monosodium
L-glutamate (3 mmol/L), a binary
mixture of monosodium
L
-glutamate (3 mmol/L) and guanosine 50-mono-
phosphate (0.1 mmol/L), and a disodium succinate solution (5 mmol/L).
The assessors had participated earlier at regular intervals for at least
18 months in sensory experiments (triangle tests, iso-intensity testings,
scale training) and were, therefore, familiar with the techniques applied.
To prevent cross-modal interactions with olfactory inputs, the panelists
wore nose clips.
Taste Profile Analysis. To evaluate the taste profile of an aqueous
solution of the yeast extract, the yeast powder was dissolved in bottled
water to give a final concentration of 1%. Afteradjustment of the pH value
to 6.0, the trained sensory panel was asked to rate the taste qualities
umami, salty, sweet, bitter, sour, and astringent on an intensity scale
ranging from 0 (not detectable) to 5 (strongly detectable).
For ion pair chromatography, an aqueous triethylammonium acetate
buffer (40 mM TEAA, pH 6) was used as solvent A and methanol as
solvent B (flow rate = 0.8 mL/min). After starting with 3% B for 5 min,
the organic solvent content was increased linearly to 40% B within 25 min
and finally to 100% B in 7 min.
LC-Time-of-Flight Mass Spectrometry (LC-TOF-MS). Mass
spectra of the target compounds were measured on a Bruker Micro-TOF-
Q (Bruker Daltonics, Bremen, Germany) mass spectrometer with flow
injection referenced on sodium formate. Data processing was performed
by using Daltonics DataAnalysis software (version 3.4, Bruker Daltonics).
Liquid Chromatography-Mass Spectrometry (LC-MS). Electro-
spray ionization (ESI) spectra were acquired on an API 3200 type LC-MS/
MS system (AB Sciex Instruments, Darmstadt, Germany) coupled to an
Taste Dilution Analysis (TDA). The lyophilized fractions obtained by
RP-MPLC were dissolved in water (10 mL; TDAwater) to perform the
TDA (27) or in an aqueous monosodium
L-glutamate solution (3 mmol/L
in water; 10 mL) to perform a modified taste dilution analysis (TDAMSG
)
as an alternative to the previously reported cTDA (28). After adjustment
of the pH value to 6.0 by adding trace amounts of either sodium hydroxide
(1 mol/L) or aqueous formic acid (1% in water), the parent solutions were

10618 J. Agric. Food Chem., Vol. 58, No. 19, 2010
Festring and Hofmann
Agilent 1100 pump, an Agilent 1100 degasser, and an Agilent 1200
autosampler (Agilent, Waldbronn, Germany). The spray voltage was set
at -4500 V in ESI- mode. Zero grade air served as nebulizer gas (35 psi)
and as turbo gas (350 °C) for solvent drying (45 psi). Nitrogen served as
curtain (20 psi) and collision gas (4.5 ꢀ 10-5 Torr). Both quadrupoles were
set at unit resolution. The declustering potential was set at -10 to -85 V in
ESI- mode. The mass spectrometer was operated in the full-scan mode
monitoring negative ions. ESI- mass and product ion spectra were
acquired with direct flow infusion. Fragmentation of [M - H]- molecular
ions into specific product ions was induced by collision with nitrogen and a
collision energy of -15 to -70 V. For instrumentation control and data
aquisition, the Analyst software version 1.5.1 was used.
HPLC-MS/MS analysis separation was performed on an analytical
Luna PFP column, 150 ꢀ 2 mm i.d., 3 μm (Phenomenex, Aschaffenburg,
Germany), prior to introduction of the eluent into the mass spectrometer
operating in the multiple reaction monitoring mode (MRM) detecting
negative ions. Three mass transitions, namely, m/z 434f339, m/z 434f
204, and m/z 434f79, were selected for the identification of (R)- and (S)-6,
respectively. After sample injection (5 μL), chromatography was per-
formed with a flow rate of 200 μL/min starting with a mixture (97:3, v/v) of
aqueous formic acid (0.1% in water) and acetonitrile containing 0.1%
formic acid. After isocratic elution for 3 min, the acetonitrile content was
increased to 100% within 22 min and, finally, kept isocratic for 2 min.
Circular Dichroism Spectroscopy (CD). Measurements of the
optical rotation were performed using a J-810 spectropolarimeter (Jasco)
equipped with a 0.1 cm path length quartz cuvette 106-QS (Hellma,
localize the fractions, contributing to the intense umami taste
impression induced by the yeast extract, aliquots of the individual
fractions I-X were used for taste dilution analysis using water
(TDAwater) and an aqueous monosodium L-glutamate solution
(3 mmol/L; TDAMSG) as the matrix, respectively. Application of
the TDAwater revealed an intrinsic umami taste of fractions
I-VIII (Figure 3A). The highest TD factors of 256 and 128 were
observed for fractions V and III, followed by fractions I, II, and
IV, all of which were evaluated with a TD factor of 64. The less
polar fractions VI-VIII were judged with comparatively low TD
factors of 8 and 16, respectively, and the nonpolar fractions IX
and X did not show any umami taste at all. In comparison, the
results of the TDAMSG demonstrated each of the 10 MPLC
fractions to have a taste-enhancing activity on the umami taste of
the L-glutamate matrix solution (Figure 3A). By far the highest
TD factors of 2048 and 1024 were found for fractions V, III, and
IV. Calculation of the ratio of the TD factors obtained from
TDAMSG and TDAwater revealed values of 8 and 16 for fractions
III-VIII, whereas for fractions I and IIthe TDfactor of TDAMSG
was only 2-fold above that of the TDAwater (Figure 3B). Most
interesting, the less polar fractions VI-VIII showed an increase of
the TD factors of 8-16 times when evaluated in the L-glutamate
matrix (TDAMSG; Figure 3B), although the TDAwater revealed
these fractions to exhibit only a marginal intrinsic umami taste
judged with TD factors of <32 (Figure 3A). As LC-MS/MS
screening and cochromatography with the corresponding refer-
€
Mullheim, Germany). Spectra were recorded from 220 to 360 nm at
20 °C by accumulating 16 runs (bandwidth = 1 nm, scan speed = 100 nm/
min, DIT = 4 s). For measurement, the samples were dissolved in phos-
phate buffer (100 mmol/L, pH 7.5), which was ultrasonicated for at least
30 min and filtered through a 0.45 μm membrane filter disk prior to use.
Final concentration of the sample solutions, which were freshly prepared
prior to analysis, was ∼500 μmol/L. After subtraction of the background
noise induced by the buffer, the spectra were analyzed using the instrument
software. The molar circular dichroic absorption Δε (cm2/mmol) was
calculated according to the quotient θobs/32980cl, where θobs (mdeg)
denotes the measured ellipticity, c (mol/L) the sample concentration,
and l (cm) the path length of the quartz cuvette.
ence substances revealed the well-known umami stimuli
L-gluta-
mic acid (1), -aspartic acid (2), and succinic acid (3) in fractions
L
I and II, the umami-enhancing purine ribonucleotide inosine
50-monophosphate (4) in fractions III-V, and guanosine 50-mono-
phosphate (5) in fractions V-VII, MPLC fraction VIII, lacking all
of these compounds, was expected to contain unknown umami
taste modulators.
Discovery of N2-(1-Carboxyethyl)guanosine 50-Monophosphate
(6) as an Unkown Umami Enhancer in Fraction VIII. Aimed at the
discovery of the umami taste-enhancing principle in fraction VIII,
this fraction was further separated on RP-18 material by means of
IPC using triethylammonium acetate. To answer the question as
to whether the unknown umami taste modulator might be based
on the purine backbone as found for 4 and 5, the effluent of the
RP18-IPC separation was monitored at 260 nm and heart-cut
into 16 subfractions (Figure 4A). Each fraction was lyophilized
twice and analyzed for the presence of purine nucleotide deriva-
tives by means of RP-HPLC/DAD analysis. Rechromatography
of fraction VIII-8 revealed the subfraction 5, which eluted after
22 min (Figure 4B) and exhibited an intense umami taste as well as
a guanine-like UV-vis absorption spectrum with a maximum at
255 nm and a shoulder at 280 nm. Preparative isolation of this
compound by means of RP-HPLC, followed by lyophilization,
afforded an amorphous white powder, which was subjected to
LC-TOF-MS, LC-MS, and 1D/2D-NMR spectroscopic experiments.
LC-TOF-MS analysis revealed m/z 434.0705 as the pseudo-
molecular ion [M - H]-, fitting well with a molecular formula of
[C13H17N5O10P]-. In addition, LC-MS/MS studies, performed in
the ESI- mode, revealed a fragmentation of the pseudomolecular
ion m/z 434 to give m/z 336 as the predominant daughter ion, thus
implying the cleavage of phosphoric acid. This was further
strengthened by the detection of the fragment ion m/z 97 found
for [H2PO4]-. In addition, the 13C NMR spectrum revealed split
signals for the carbon atom C(40) and C(50) with coupling
Nuclear Magnetic Resonance Spectroscopy (NMR). The 1H, 13C,
COSY, DEPT, HMQC, and HMBC spectroscopic experiments were
performed on either a 400 MHz DRX 400 or a 500 MHz Avance III
NMR spectrometer from Bruker (Rheinstetten, Germany). Samples were
dissolved in D2O using 3-(trimethylsilyl)propionic-2,2,3,3-d4 acid sodium
salt (TMSP) as reference. In the case of (R)- and (S)-N2-(1-carbox-
yethyl)guanosine 50-monophosphate, the solvent was prepared by mixing
10 μL of a 40% w/w solution of sodium deuteroxide in D2O with 750 μL of
D2O. Data processing was performed using Topspin version 1.3 (Bruker),
and the individual data interpretation was done with MestReNova 5.1.0-
2940 (Mestrelab Research S.L., Santiago de Compostela, Spain).
RESULTS AND DISCUSSION
To evaluate the taste modalities of a yeast extract, a 1% aque-
ous solution ofthe extract was subjected to a taste profile analysis.
To achieve this, a trained sensory panel was asked to rate the taste
descriptors umami, salty, sweet, sour, bitter, and astringent on a
linear scale from 0 (no taste impression) to 5 (strong taste
impression). Whereas the umami taste impression was rated with
the highest intensity of 5.0, sweet and salty tastes were evaluated
with intensities of 1.6 and 1.5, respectively. Bitter and sour taste
impressions were judged with comparably low intensities of 0.6
and 0.5, whereas an astringent orosensation was not detectable at
all.
Sensory-Guided Fractionation of a Yeast Extract. To gain first
insight into the polarity of the taste-active and/or taste-modulat-
ing compounds, the yeast extract was separated by means of
MPLC on RP-18 material to give 10 fractions, I-X (Figure 2),
with yields ranging from 1.4 to 28.2%. Polar fractions I-V
accounted for a total yield of 88.3%, whereas the total yield of
the less polar fractions VI-X accounted for 11.7% only. To
3
2
constants of JC,P = 8.7 Hz and JC,P = 4.6 Hz, respectively,
thus verifying the presence of the phosphate moiety at C(50) of the
50-GMP moiety.
1
The H NMR spectrum of the unknown compound showed
eight resonance signals with integrals between one and three

Article
J. Agric. Food Chem., Vol. 58, No. 19, 2010 10619
Figure 5. Excerpt from the HMBC spectrum (500 MHz, D2O/NaOD) of fraction VIII-8/5. Arbitrary numbering of the carbon atoms refers to Figure 6.
protons. The signals of the 50-GMP moiety were easily assigned
by comparing the spectral data of the isolated compound with the
1H NMR spectrum recorded for the disodium salt of 50-GMP (5).
As expected for the purine nucleotide, the signal of the aromatic
proton H-C(8) resonated at 8.17 ppm and the protons of the
ribose moiety were observed between 6.02 and 3.97 ppm. The
anomeric proton H-C(10) of the ribose moiety was assigned to
the doublet found at 6.02 ppm and showing a coupling constant
Figure 6. Structures of (R)-N2-(1-carboxyethyl)guanosine 50-monopho-
sphate (6a) and (S)-N2-(1-carboxyethyl)guanosine 50-monophosphate
(6b).
of 5.2 Hz. In addition to the 50-GMP related signals, two
additional proton signals were observed at 1.45 and 4.25 ppm,
respectively. The resonance signal at 4.25 ppm, integrating for
one proton and showing the multiplicity of a quartet, was assig-
ned toH-C(200) and showed homonuclear 3J coupling (7.2 Hz) to
the methyl group H-C(300) resonating at 1.45 ppm. Furthermore,
an HMBC experiment revealed that the proton H-C(200) showed
heteronuclear coupling to the guanine carbon atom C(2), ob-
served at 155.0 ppm, as well as an additional C,H coupling to the
carboxy carbon C(100) resonating at 184.2 ppm (Figure 5). In
addition, the 13C NMR spectrum revealed C(200) with a chemical
shift of55.4 ppm as expected for the R-carbon atomofan R-amino
acid. These data verified the 50-GMP derivative to be decorated
with an aminocarboxylic acid structure in contrast to the recently
reported 50-GMP lactoylamide showing a resonance signal of
70.7 ppm for the R-carbon atom of the lactic acid amide moiety
(30). With all of these spectroscopic data taken into considera-
tion, the target compound was identified as the previously not
reported (R)- or (S)-configured N2-(1-carboxyethyl)guanosine
50-monophosphate, (R)- or (S)-6 (Figure 6), but its stereochemistry
at C(200) could not be clarified.
structure of compound 6 was not yet reported in literature, the
structurally related compounds N2-(1-carboxyethyl)-20-deoxy-
guanosine (7) and N2-(1-carboxyethyl)guanine (8; Figure 7) are
known as advanced glycation end products, derived from the
reaction of 20-deoxyguanosine and guanine with highly reactive
C3 sugar breakdown products such as glyceraldehyde and dihy-
droxyacetone, respectively (26). To investigate whether (R)- or
(S)-6 is generated by a Maillard-type model reaction of 50-GMP
with C3 sugar breakdown products, a binary aqueous solution of
50-GMP (5) and DL-glyceraldehyde or 1,3-dihydroxyacetone,
respectively, was incubated at pH 7.0 for 10 days at 40 °C. RP-
HPLC analysis revealed that 50-GMP was completely converted
into two reaction products (Figure 7), which were isolated and
purified by means of preparative RP-HPLC and, then, used for
MS and NMR spectroscopic experiments. Comparison of chro-
1
matographic (RP-HPLC) and spectroscopic data (LC-MS, H
and 13C NMR) as well as cochromatography demonstrated the
later eluting compound to be identical with the stereoisomer of 6
isolated above from the yeast extract.
Maillard-Type Generation of N2-(1-Carboxyethyl)guanosine
50-Monophosphate (6) and Stereochemical Investigations. To unequi-
vocally confirm the proposed structure of 6 and to clarify its
stereochemistry at C(200), authentic reference compounds needed
to be synthesized for (R)- or (S)-6 (Figure 6). Although the
Next, the absolute stereochemistry of the chiral center C(200) of
the isomers of 6 isolated from the yeast extract and the Maillard
reaction system, respectively, should be unequivocally identified.

10620 J. Agric. Food Chem., Vol. 58, No. 19, 2010
Festring and Hofmann
Figure 7. RP-HPLC chromatogram (λ = 260 nm) of an aqueous solution of 50-GMP (5) and DL-glyceraldehyde before (A) and after incubation for 10 days at
40 °C (B).
To confirm that the CD spectra of the earlier as well as the later
eluting stereoisomers of 6, showing the same orientations of the
CD curve as (R)- and (S)-7, respectively, are not influenced by the
presence of the phosphate moiety as well as the additional chiral
center at C(20) ofthe ribose backbone in 6, both stereoisomers of 6
and 7 structures were truncated to release the common feature of
the N2-(1-carboxyethyl)guanine, 8 (Figure 8). To achieve this
Figure 8. Acidic depurination of the stereoisomers of N2-(1-carbox-
depurination, both the (R)- and (S)-7 and the earlier as well as the
yethyl)guanosine 50-monophosphate (6) and N2-(1-carboxyethyl)-20-
deoxyguanosine (7) to give N2-(1-carboxyethyl)guanine (8) as their
common hydrolysis product.
later eluting stereoisomers of 6 were individually hydrolyzed by
treatment with hydrochloric acid, and the enantiomeric hydro-
lysis products (R)- and (S)-8 were isolated individually and puri-
fied by means of preparative RP-HPLC. LC-MS and NMR spec-
troscopic analyses revealed the identical spectroscopic data for
(R)- and (S)-8, being well in line with previous papers (26). The
CD spectra of the hydrolysis products (R)- and (S)-8 (Figure 9C)
formed upon hydrolysis of (R)- and (S)-7 were found to be
identical to those of the hydrolysis products released from the
earlier and later eluting stereoisomers of 6 (Figure 9D). These data
clearly confirmed (R)-N2-(1-carboxyethyl)guanosine 50-mono-
phosphate, (R)-6, as the earlier eluting isomer and (S)-N2-(1-
carboxyethyl)guanosine 50-monophosphate, (S)-8, as the later
eluting isomer of the umami enhancer isolated from yeast.
Sensory Activity of (R)- and (S)-6. To gain a first insight into the
taste properties of (R)- and (S)-6, the recognition threshold
concentrations for umami taste were determined in water (pH
6.0). Whereas the (R)-configured isomer showed an umami thres-
hold of 0.85 mmol/L, the corresponding (S)-configured molecule
exhibited an umami taste at a comparatively lower level of 0.19
mmol/L, which is identical to the threshold concentration of the
parent nucleotide 50-GMP (Table 1).
To achieve this, the (R)- and (S)-isomers of N2-(1-carboxyethyl)-
20-deoxyguanosine (7; Figure 8), both of which have been synthe-
sized by independent chemical synthesis and assigned in their
absolute configuration (31), were generated by the Maillard-type
reaction of 20-deoxyguanosine and 1,3-dihydroxyacetone and,
after purification, the CD spectra of both stereoisomers were
recorded. Comparison of the CD spectra obtained for the two
diastereomers of the Maillard-type generated N2-(1-carboxy-
ethyl)-20-deoxyguanosine (Figure 9A) with the CD data published
in literature for the synthetic analogues (31) clearly revealed (R)-7
as the earlier and (S)-7 as the later eluting stereoisomer. Although
(R)- and (S)-7 were lacking the phosphate moiety as well as the
chiral center at C(20) of the sugar backbone when compared to the
50-GMP derivative 6, the earlier eluting stereoisomer of 6 showed
the same orientations of the CD curve as (R)-7 with a positive
absorption maximum at 249 nm and a negative absorption
maximum at 282 nm, and the CD spectrum of the later eluting
diastereomer of 6 was the corresponding mirror image with a
positive absorption maximum at 282 nm and a negative absorp-
tion maximum at 249 nm (Figure 9B).
Although several purine 50-ribonucleotides exhibit an intrinsic
taste, the hallmark of umami taste is the pronounced synergism

Article
J. Agric. Food Chem., Vol. 58, No. 19, 2010 10621
Figure 9. CD spectra of (A) N 2-(1-carboxyethyl)-20-deoxyguanosine, (R)-7 (solid line) and (S)-7 (dotted line), (B) N 2-(1-carboxyethyl)guanosine
50-monophosphate, (R)-6 (solid line), and (S)-6 (dotted line), as well as N 2-(1-carboxyethyl)guanine, (R)-8 (solid line), and (S)-8 (dotted line), obtained by
acidic hydroysis of (C) (R)-7 (solid line) and (S)-7 (dotted line) and (D) (R)-6 (solid line) and (S)-6 (dotted line).
Table 1. Umami Recognition Threshold Concentration and Iso-umami Con-
centrations of 50-GMP (5) and the Maillard Reaction Products (R)-6 and (S)-6
5a
(R)-6
(S)-6
taste thresholdb (mmol/L)
iso-umami concnc (mmol/L)
0.15
45
0.85
25
0.19
45
a Applied as the disodium salt. b Taste recognition threshold concentration
determined in bottled water by means of a triangle test. c Iso-umami concentration
determined by comparison of the taste intensity of an aqueous binary solution of the
nucleotide (0.1 mmol/L) and monosodium L-glutamate (10 mmol/L) with that of
solutions of ascending monosodium L-glutamate concentrations (10-100 mmol/L,
in 5 mmol/L steps; pH 6.0) using a duo test.
Figure 10. HPLC-MS/MS analysis of (R)-6 and (S)-6 in a reference
solution (A) and a sample of the authentic yeast extract (B). Signal
intensity of each mass transition is normalized. Peak numbering refers to
the chemical structure given in Figure 6.
between those compounds and L-glutamate (17). To identify
possible taste-enhancing properties of the 50-GMP derivatives,
an iso-umami test was performed. To achieve this, an aqueous
binary solution of the test nucleotide (0.1 mmol/L) and mono-
sodium
solutions containing ascending concentrations of monosodium
-glutamate alone in a duo test, and the sensory panelists were
asked to determine the concentration of monosodium -gluta-
mate required to match the umami taste intensity of the binary
nucleotide/monosodium -glutamate mixture. As shown in Ta-
ble 1, the sensory panel rated the mixture containingmonosodium
-glutamate and 50-GMP (5) with an iso-umami value of 45 mmol/
L. In the case of the glycated compounds, once again a difference
between both diastereomers was observed. Whereas (R)-6 ex-
hibited a significantly reduced taste-enhancing activity with an
iso-umami value of 25 mmol/L, the enhancement activity of (S)-6
was equivalent to that of 5, but the umami taste sensation of the
L
-glutamate (10 mmol/L) was compared to aqueous
and separated by means of HPLC using a pentafluorophenyl-
propyl (PFP) column, and the effluent was screened for three
selected mass transitions, namely, m/z 434f339, m/z 434f204,
and m/z 434f79, for the identification of (R)- and (S)-6 by
means of tandem mass spectrometry. Comparison of the mass
transitions and retention times of the synthesized reference
compounds (Figure 10A) with those of the authentic yeast
sample (Figure 10B) revealed two baseline-separated peaks for
(R)- and (S)-6, thus demonstrating the presence of both N2-(1-
carboxyethyl)guanosine 50-monophosphate diastereomers in
the yeast extract. To the best of our knowledge, these umami-
active sensometabolites were detected for the first time in a food
product.
L
L
L
L
monosodium
L
-glutamate solution was described to be slightly
In conclusion, the present study shows for the first time
the isolation and identification of umami-active N2-(1-carbox-
yethyl)guanosine 50-monophosphate diastereomers in a commer-
cial yeast extract and its formation by a Maillard-type glycation
of 50-GMP with a C3 sugar breakdown product. Studies on
structure-activity relationships of a series of umami-tasting
glycated 50-GMP derivatives are ongoing and will be published
elsewhere.
sweetish and more full-bodied in the presence of (S)-6 when
compared to the parent nucleotide 5. This needs to be more
precisely investigated in future studies.
Verification of the Occurrence of (R)- and (S)-6 in Yeast Extract
by Means of LC-MS/MS. To unequivocally confirm the pre-
sence of both diastereomers of 6 in the yeast extract, an aqueous
solution of a sample of the yeast extract was membrane filtered

10622 J. Agric. Food Chem., Vol. 58, No. 19, 2010
Festring and Hofmann
ACKNOWLEDGMENT
(18) Hufnagel, J. C.; Hofmann, T. Orosensory-directed identification of
astringent mouthfeel and bitter-tasting compounds in red wine.
J. Agric. Food Chem. 2008, 56, 1376-1386.
We thank Bio Springer for providing yeast extract samples as
well as Sabine Rauth for technical support in the CD spectro-
scopic analyses.
(19) Toelstede, S.; Hofmann, T. Sensomics mapping and identification of
the key bitter metabolites in gouda cheese. J. Agric. Food Chem.
2008, 56, 2795-2804.
(20) Rotzoll, N.; Dunkel, A.; Hofmann, T. Activity-guided identification
LITERATURE CITED
of (S)-malic acid 1-O- -glucopyranoside (morelid) and γ-aminobu-
D
(1) Ikeda, K. New seasonings. Chem. Senses 2002, 27, 847-849.
(2) Nagodawithana, T. Yeast-derived flavors and flavor enhancers and
their probable mode of action. Food Technol. 1992, 46 (138), 140-
142, 144.
tyric acid as contributors to umami taste and mouth-drying oral
sensation of morel mushrooms (Morchella deliciosa Fr.). J. Agric.
Food Chem. 2005, 53, 4149-4156.
(21) Ottinger, H.; Hofmann, T. Identification of the taste enhancer alapyr-
idaine in beef broth and evaluation of its sensory impact by taste
reconstitution experiments. J. Agric. Food Chem. 2003, 51, 6791-6796.
(22) Dunkel, A.; Koester, J.; Hofmann, T. Molecular and sensory
characterization of γ-glutamyl peptides as key contributors to the
kokumi taste of edible beans (Phaseolus vulgaris L.). J. Agric. Food
Chem. 2007, 55, 6712-6719.
(23) Toelstede, S.; Dunkel, A.; Hofmann, T. A series of kokumi peptides
impart the long-lasting mouthfulness of matured gouda cheese.
J. Agric. Food Chem. 2009, 57, 1440-1448.
(24) Dunkel, A.; Hofmann, T. Sensory-directed identification of β-alanyl
dipeptides as contributors to the thick-sour and white-meaty oro-
sensation induced by chicken broth. J. Agric. Food Chem. 2009, 57,
9867-9877.
(25) Sonntag, T.; Kunert, C.; Dunkel, A.; Hofmann, T. Sensory-guided
identification of N-(1-methyl-4-oxoimidazolidin-2-ylidene)-R-amino
acids as contributors to the thick-sour and mouth-drying orosensa-
tion of stewed beef juice. J. Agric. Food Chem. 2010, 58, 6341-6350.
(26) Ochs, S.; Severin, T. Reaction of 20-deoxyguanosine with glycer-
aldehyde. Liebigs Ann. Chem. 1994, 851-853.
(27) Frank, O.; Ottinger, H.; Hofmann, T. Characterization of an intense
bitter-tasting 1H,4H-quinolizinium-7-olate by application of the
taste dilution analysis, a novel bioassay for the screening and
identification of taste-active compounds in foods. J. Agric. Food
Chem. 2001, 49, 231-238.
(28) Ottinger, H.; Soldo, T.; Hofmann, T. Discovery and structure
determination of a novel maillard-derived sweetness enhancer by
application of the comparative taste dilution analysis (cTDA). J. Agric.
Food Chem. 2003, 51, 1035-1041.
(29) Haseleu, G.; Intelmann, D.; Hofmann, T. Structure determination
and sensory evaluation of novel bitter compounds formed from
β-acids of hop (Humulus lupulus L.) upon wort boiling. Food Chem.
2009, 116, 71-81.
(30) De Rijke, E.; Ruisch, B.; Bakker, J.; Visser, J.; Leenen, J.; Haiber, S.; De
Klerk, A.; Winkel, C.; Koenig, T. LC-MS study to reduce ion suppres-
sion and to identify N-lactoylguanosine 50-monophosphate in bonito: a
new umami molecule? J. Agric. Food Chem. 2007, 55, 6417-6423.
(31) Cao, H.; Jiang, Y.; Wang, Y. Stereospecific synthesis and character-
ization of oligodeoxyribonucleotides containing an N2-(1-carboxyethyl)-
20-deoxyguanosine. J. Am. Chem. Soc. 2007, 129, 12123-12130.
(3) Nagodawithana, T. W. Savory flavors. In Bioprocess Production of
Flavor, Fragrance, and Color Ingredients; Gabelman, A., Ed.; Wiley:
New York, 1994; pp 135-168.
(4) Chae, H. J.; Joo, H.; In, M.-J. Utilization of brewer’s yeast cells for
the production of food-grade yeast extract. Part 1: Effects of
different enzymatic treatments on solid and protein recovery and
flavor characteristics. Bioresour. Technol. 2001, 76, 253-258.
(5) Davidek, J.; Hajslova, J.; Kubelka, V.; Velisek, J. Flavor significant
compounds in yeast autolyzate gistex X-II powder. Part I. Acidic
fraction. Nahrung 1979, 23, 673-680.
(6) Hajslova, J.; Velisek, J.; Davidek, J.; Kubelka, V. Flavor significant
compounds in yeast autolyzate gistex X-II powder. Part 2. Neutral
and basic fractions. Nahrung 1980, 24, 875-881.
(7) Izzo, H. V.; Ho, C. T. Ammonia affects Maillard chemistry of an
extruded autolyzed yeast extract: pyrazine aroma generation and
brown color formation. J. Food Sci. 1992, 57, 657-659, 674.
(8) Izzo, H. V.; Ho, C. T. Isolation and identification of the volatile
components of an extruded autolyzed yeast extract. J. Agric. Food
Chem. 1991, 39, 2245-2248.
(9) Mahadevan, K.; Farmer, L. Key odor impact compounds in three
yeast extract pastes. J. Agric. Food Chem. 2006, 54, 7242-7250.
€
(10) Munch, P.; Hofmann, T.; Schieberle, P. Comparison of key odorants
generated by thermal treatment of commercial and self-prepared
yeast extracts: influence of the amino acid composition on odorant
formation. J. Agric. Food Chem. 1997, 45, 1338-1344.
€
(11) Munch, P.; Schieberle, P. Quantitative studies on the formation of
key odorants in thermally treated yeast extracts using stable isotope
dilution assays. J. Agric. Food Chem. 1998, 46, 4695-4701.
(12) Ames, J. M.; MacLeod, G. Volatile components of a yeast extract
composition. J. Food Sci. 1985, 50, 125-131.
(13) Nagodawithana, T. Flavor enhancers: their probable mode of
action. Food Technol. 1994, 48 (79-80), 82-85.
(14) Velisek, J.; Davidek, J.; Kubelka, V.; Tran, T. B. T.; Hajslova, J.
Succinic acid in yeast autolysates and its sensory properties. Nahrung
1978, 22, 735.
(15) Yamaguchi, S. Synergistic taste effect of monosodium glutamate and
disodium 50-inosinate. J. Food Sci. 1967, 32, 473-487.
(16) Yamaguchi, S.; Yoshikawa, T.; Ikeda, S.; Ninomiya, T. Measure-
ment of the relative taste intensity of some
L
-R-amino acids and
50-nucleotides. J. Food Sci. 1971, 36, 846-849.
(17) Zhang, F.; Klebansky, B.; Fine, R. M.; Xu, H.; Pronin, A.; Liu, H.;
Tachdjian, C.; Li, X. Molecular mechanism for the umami taste
synergism. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 20930-
20934.
Received for review July 26, 2010. Revised manuscript received
August 31, 2010. Accepted September 01, 2010. We thank Bio Springer
for providing financial support.