Isolation, structure determination, and sensory activity of mouth-drying and astringent nitrogen-containing phytochemicals isolated from red currants (Ribes rubrum)

Article abstract of DOI:10.1021/jf0632076

Application of chromatographic separation and taste dilution analyses recently revealed, besides a series of flavon-3-ol glycosides and (E)/(Z)-aconitic acid, four nitrogen-containing phytochemicals as the key astringent and mouth-drying compounds in red currants (Ribes rubrum). The isolation and structure determination of the astringent indoles 3-carboxymethyl-indole-1-N-β-D-glucopyranoside (1) and 3- methylcarboxymethyl-indole-1-N-β-D-glucopyranoside (2), as well as the astringent, noncya-nogenic nitriles 2-(4-hydroxybenzoyloxymethyl)-4-β-D- glucopyranosyloxy-2(E)-butenenitrile (3) and 2-(4-hydroxy-3- methoxybenzoyloxymethyl)-4-β-D-glucopyranosyloxy-2(E)-butenenitrile (4) by means of 1D/2D NMR, LC-MS/MS, and UV-vis spectroscopy are reported. The structures of compounds 1 and 2 were confirmed by synthesis. Using the recently developed half-tongue test, human recognition thresholds for the astringent and mouth-drying nitrogen compounds were determined to be between 0.0003 and 5.9 μmol/L (water). In particular, the extraordinarily low threshold of 0.0003 μmol/L evaluated for the indole 1 represents the lowest recognition threshold of any astringent phytochemical reported to date.

Full text of DOI:10.1021/jf0632076

J. Agric. Food Chem. 2007, 55, 1405−1410 1405
Isolation, Structure Determination, and Sensory Activity of
Mouth-Drying and Astringent Nitrogen-Containing
Phytochemicals Isolated from Red Currants (Ribes rubrum)
BERND SCHWARZ AND THOMAS HOFMANN*
Institut f u¨ r Lebensmittelchemie, Universit a¨ t M u¨ nster, Corrensstrasse 45, D-48149 M u¨ nster, Germany
Application of chromatographic separation and taste dilution analyses recently revealed, besides a
series of flavon-3-ol glycosides and (E)/(Z)-aconitic acid, four nitrogen-containing phytochemicals as
the key astringent and mouth-drying compounds in red currants (Ribes rubrum). The isolation and
structure determination of the astringent indoles 3-carboxymethyl-indole-1-N-â-D-glucopyranoside (1)
and 3-methylcarboxymethyl-indole-1-N-â-D-glucopyranoside (2), as well as the astringent, noncya-
nogenic nitriles 2-(4-hydroxybenzoyloxymethyl)-4-â-D-glucopyranosyloxy-2(E)-butenenitrile (3) and
2-(4-hydroxy-3-methoxybenzoyloxymethyl)-4-â-D-glucopyranosyloxy-2(E)-butenenitrile (4) by means
of 1D/2D NMR, LC-MS/MS, and UV-vis spectroscopy are reported. The structures of compounds 1
and 2 were confirmed by synthesis. Using the recently developed half-tongue test, human recognition
thresholds for the astringent and mouth-drying nitrogen compounds were determined to be between
0.0003 and 5.9 µmol/L (water). In particular, the extraordinarily low threshold of 0.0003 µmol/L
evaluated for the indole 1 represents the lowest recognition threshold of any astringent phytochemical
reported to date.
KEYWORDS: Red currants; astringency; taste dilution analysis; half-tongue test; 3-carboxymethyl-indole-
1-N-â-D-glucopyranoside; 3-methylcarboxymethyl-indole-1-N-â-D-glucopyranoside; 2-(4-hydroxybenzoy-
loxymethyl)-4-â-D-glucopyranosyloxy-2(E)-butenenitrile; 2-(4-hydroxy-3-methoxybenzoyloxymethyl)-4-â-
D-glucopyranosyloxy-2(E)-butenenitrile
INTRODUCTION
hydroxy-4-(O-â-D-glucopyranosyl)phenyl]-3,5-hexadien-2-
one, (E)/(Z)-aconitic acid, and a series of flavon-3-ol mono-,
di-, tri-, and tetraglycosides as the key players imparting the
astringent oral taste sensation during consumption of red currant
juice (2). Because the N-containing compounds 1-4 have as
yet not been reported as key taste compounds in foods, the
details on their isolation, structure determination, and sensory
activity are presented.
Besides the attractive aroma, the sensory quality of red
currants (Ribes rubrum) as well as products made therefrom,
such as juices or jams, is driven by its typical astringency as
well as its sour taste. The astringency is perceived as a long-
lasting puckering and mouth-drying sensation in the oral cavity
and can enhance the complexity and palate length of fruit
products.
To answer the question as to which nonvolatile, key taste
compounds are responsible for the typical astringent taste of
red currants, we recently applied the so-called taste dilution
analysis (1) on chromatographic fractions isolated from freshly
prepared red currant juice (2). This bioassay-guided fractionation
led to the detection of the previously unknown nitrogen-
containing astringent compounds 3-carboxymethyl-indole-1-N-
â-D-glucopyranoside (1, Figure 1), 3-methylcarboxymethyl-
indole-1-N-â-D-glucopyranoside (2), 2-(4-hydroxybenzoyloxy-
methyl)-4-â-D-glucopyranosyloxy-2(E)-butenenitrile (3), and
MATERIALS AND METHODS
Chemicals. The following compounds were obtained commer-
cially: silver trifluoromethane sulfonate, 1,2-dichloroethane, 2,6-di-
tert-butyl-4-methylpyridine, hydrochloric acid, sodium hydroxide,
indole-3-acetic acid (Sigma, Steinheim, Germany); R-bromo-tetra-O-
acetyl-D-glucose (Fluka, Taufkirchen, Germany). Solvents were of
HPLC grade (Merck, Darmstadt, Germany). Deuterated solvents were
supplied by Euriso-Top (Gif-Sur-Yvette, France). Fresh puree, made
from red currant fruits harvested in 2003, was obtained from the food
industry and kept frozen at -26 °C until used. Bottled water (Evian;
low mineralization, 484 mg/L) adjusted to pH 4.5 with aqueous
hydrochloric acid (0.1 mol/L) was used for the sensory experiments.
2-(4-hydroxy-3-methoxybenzoyloxymethyl)-4-â-D-glucopyrano-
syloxy-2(E)-butenenitrile (4), besides (E)-6-[3-hydroxy-4-
(O-â-D-glucopyranosyl)phenyl]-5-hexen-2-one, (3E,5E)-6-[3-
Sensory Analyses. Training of the sensory panel as well as the half-
tongue test used for taste dilution analysis and taste threshold
determination was done closely following the procedure reported
recently (2, 3).
*
Author to whom correspondence should be addressed (telephone +49-
2
51-83-33-391; fax +49-251-83-33-396; e-mail thomas.hofmann@
uni-muenster.de).
1
0.1021/jf0632076 CCC: $37.00
© 2007 American Chemical Society
Published on Web 01/30/2007

1406 J. Agric. Food Chem., Vol. 55, No. 4, 2007
Schwarz and Hofmann
Figure 2. (A) GPC chromatogram and (B) taste dilution (TD) chromato-
gram of the ethyl acetate extractables isolated from red currant puree.
Figure 1. Chemical structures of the astringent indoles 1 and 2 and the
astringent non-cyanogenic nitriles 3 and 4 isolated from red currant.
Solvent Extraction of Red Currant Puree. A mixture of red currant
puree (10 kg) and methanol (40 L) was vigorously stirred for 1 h at
room temperature under an atmosphere of argon. After filtration, the
residual fruit material was extracted for 1 h with a mixture (7:3, v/v;
3
× 10 L) of methanol and water adjusted to pH 4.0 with aqueous
formic acid (1% in water). After filtration, the liquid layers were
combined and freed from methanol under vacuum to obtain the
methanol extractables. After the addition of water (2 L), the methanol
solubles were extracted with ethyl acetate (5 × 1.5 L), and the combined
organic layers were freed from solvent under high vacuum to give the
intensely astringent tasting solvent extractables after freeze-drying twice.
Gel Permeation Chromatography (GPC). An aliquot (4 g) of the
ethyl acetate extractable fraction isolated from red currant puree was
dissolved in a mixture (40:60, v/v; 20 mL) of methanol and water
adjusted to pH 4.0 with aqueous formic acid (1% in water) and then
applied onto the top of a water-cooled 400 × 50 mm glass column
filled with a slurry of Sephadex LH-20 (Amersham Pharmacia Biotech,
Uppsala, Sweden) conditioned with the same solvent mixture. Chro-
matography was performed with methanol/water (40:60, v/v; pH 4.0;
Figure 3. RP-HPLC chromatogram (272 nm) of GPC fraction II and peaks
of the astringent-tasting compounds 1−4.
(5:95, v/v; 5 mL) of acetonitrile and water and, after membrane
filtration, was fractionated by preparative HPLC on ODS-Hypersil RP-
18, 250 × 21.2 mm i.d., 5 µm (ThermoHypersil, Kleinostheim,
Germany) using an acetonitrile/aqueous formic acid gradient at a flow
rate of 20 mL/min. Using an aqueous formic acid (1.5% in water) as
solvent A and acetonitrile as solvent B, chromatography was started
with a linear gradient from 5% B to 17% B within 35 min and then
maintained isocratically for 15 min, thereafter increasing the amount
of solvent B to 50% B within 20 min and then to 100% within 10 min.
The compounds eluting as peaks 1-4 (Figure 3) and imparting intense
astringency were collected, freed from solvent under vacuum, and
freeze-dried two times to give compounds 1-4 as white amorphous
powders. After rechromatography using the same chromatographic
procedure, the taste compounds were obtained with a purity of >99%
(based on HPLC analysis).
9
00 mL), followed by methanol/water (50:50, v/v; pH 4.0; 900 mL),
methanol/water (60:40, v/v; pH 4.0; 2700 mL), methanol/water (80:
0, v/v; pH 4.0; 900 mL), and, finally, methanol (2700 mL) with a
2
flow rate of 3 mL/min. Monitoring the effluent by means of an L-7420
type UV-vis detector (Merck Hitachi, Darmstadt, Germany) operating
at 272 nm, the seven fractions I-VII were collected as shown in Figure
2
, and the individual fractions were freed from solvent under vacuum
and were then freeze-dried twice. The residue of each GPC fraction
was used for the taste dilution analysis as well as for chemical analysis.
Taste Dilution Analysis (TDA). Aliquots of the seven GPC fractions
were dissolved in 5.0 mL of bottled water (pH 4.5) in their “natural”
ratios, that is, the concentration ratios present in the authentic fruit puree.
These solutions were then used for the TDA following precisely the
procedure reported recently (1, 2). The TD factors evaluated by three
different assessors in three different sessions were averaged. The TD
factors between individuals and separate sessions did not differ by more
than plus/minus one dilution step.
3-Carboxymethyl-indole-1-N-â-D-glucopyranoside, 1 (Figure 1):
UV-vis (acetonitrile), λmax ) 215, 267 nm; LC-TOF/MS, C16
H
19NO
7
;
-
-
-
LC-MS (ESI ), m/z 336 (100, [M - H] ), 174 (52, [M - H - 162] );
1
H NMR (400 MHz, D O; COSY), δ 3.59 [dd, 1H, J ) 9.2, 9.2 Hz,
2
H-C(4′)], 3.67 [m, 1H, H-C(5′)], 3.69 [dd, 1H, J ) 9.1, 9.1 Hz,
H-C(3′)], 3.75 [m, 1H, H-C(6a′)], 3.81 [s, 2H, H-C(8)], 3.86 [m,
1H, H-C(6b′)], 4.01 [dd, 1H, J ) 9.1, 9.1 Hz, H-C(2′)], 5.56 [d, 1H,
J ) 9.1 Hz, H-C(1′)], 7.19 [dd, 1H, J ) 8.1, 8.1 Hz, H-C(6)], 7.29
[dd, 1H, J ) 8.1, 8.1 Hz, H-C(5)], 7.39 [s, 1H, H-C(2)], 7.57 [dd,
2H, H-C(4, 7)]; 13C NMR (100 MHz, MeOD; HMQC, HMBC), δ
30.9 [C(8)], 60.6 [C(6′)], 69.3 [C(4′)], 71.6 [C(2′)], 77.1 [C(3′)], 77.1
Isolation of Nitrogen-Containing Astringent Compounds 1-4
from GPC Fraction II. GPC fraction II was dissolved in a mixture

N-Containing Astringent Tastants in Red Currants
J. Agric. Food Chem., Vol. 55, No. 4, 2007 1407
Figure 4. Synthetic sequence used for the preparation of 3-carboxymethyl-indole-1-N-â-D-glucopyranoside (1).
[
[
[
C(5′)], 84.4 [C(1′)], 110.0 [C(3)], 110.2 [C(7)], 119.1 [C(4)], 120.6
C(6)], 122.5 [C(5)], 124.5 [C(2)], 128.0 [C(3a)], 136.4 [C(7a)], 177.6
C(9)].
[C(6′)], 63.2 [C(1)], 66.7 [C(4)], 70.0 [C(5′)], 73.4 [C(2′)], 76.3 [C(4′)],
76.5 [C(3′)], 102.0 [C(1′)], 111.7 [C(2)], 112.0 [C(2′′)], 114.6 [C(5′′)],
114.8 [C(5)], 120.2 [C(1′′)], 123.8 [C(6′′)], 147.5 [C(3′′)], 147.6 [C(3)],
152.0 [C(4′′)], 165.5 [C(7′′)].
3-Methylcarboxymethyl-indole-1-N-â-D-glucopyranoside, 2 (Figure
1
): UV-vis (acetonitrile), λmax ) 215, 267; LC-TOF/MS, C17
H
21NO
7
;
Analysis of Glycosidically Bound Carbohydrates. An aliquot (2
mg) of the target compound, dissolved in aqueous hydrochloric acid
(2 mol/L; 1 mL), was heated at 110 °C for 120 min. After cooling, an
-
-
-
LC-MS (ESI ), m/z 350 (100, [M - H] ), 188 (54, [M - H - 162] );
1
H NMR (400 MHz, D O; COSY), δ 3.52 [dd, 1H, J ) 9.3, 9.3 Hz,
2
H-C(4′)], 3.59 [m, 1H, H-C(5′)], 3.62 [dd, 1H, J ) 9.0, 9.0 Hz,
H-C(3′)], 3.71 [s, 3H, H-C(10)], 3.73 [m, 1H, H-C(6a′)], 3.80 [s,
2 3
aqueous solution of Na CO (4 mol/L; 300 µL), followed by pyridine
(50 µL), was added, and the solution was freeze-dried. The water-free
residue was dissolved in a solution (100 µL) of 1% hydroxylamine
hydrochloride in water-free pyridine and was then heated at 70 °C for
30 min. After cooling, 1-(trimethylsilyl) imidazole (100 µL) was added,
and the solution was heated for an additional 30 min at 70 °C and,
after cooling, was directly injected into the HRGC-MS system.
Synthesis of 3-Carboxymethyl-indole-1-N-â-D-glucopyranoside.
By adoption of the modified Koenigs-Knorr methodology reported
for O-glycoside synthesis (4) (Figure 4), an activated molecular sieve
(4 Å; 1.5 g) was added to a suspension of silver trifluoromethane
sulfonate (2.0 mmol) in anhydrous 1,2-dichloroethane (10 mL) under
an atmosphere of argon in the dark with stirring at room temperature.
The mixture was then cooled to -20 °C, R-bromotetra-O-acetyl-D-
glucose (2.0 mmol) and indole-3-acetic acid (1.0 mmol) were added,
and, after 15 min of stirring at -20 °C, 2,6-di-tert(butyl)-4-methylpy-
ridine (2.0 mmol) was added. After removal of the cooling bath, the
mixture was stirred at room temperature for 48 h and then filtered,
water (10 mL) was added to the filtrate, the pH value was adjusted to
5.0 with sulfuric acid (1 mol/L), and the organic solvent was removed
under vacuum. The aqueous layer was freeze-dried, and the residue
was dissolved in chloroform (30 mL) and cooled to 0 °C in an ice
bath. After the addition of a methanolic solution of sodium methylate
(1 mol/L, 3 mL), the mixture was stirred for 2 h at 0 °C, and then an
equivalent amount of sulfuric acid (1 mol/L) was added and the pH
value was adjusted to 5.0 with aqueous sulfuric acid (1 mol/L). The
organic solvent was removed under vacuum and, after membrane
filtration (0.45 µm), the title compound was isolated from the aqueous
layer by preparative HPLC on ODS-Hypersil RP-18, 250 × 21.2 mm
i.d., 5 µm (ThermoHypersil, Kleinostheim, Germany), using an
acetonitrile/aqueous formic acid gradient at a flow rate of 20 mL/min
and monitoring the effluent at 272 nm. Using an aqueous formic acid
(1.5% in water) as solvent A and acetonitrile as solvent B, chroma-
tography started isocratically with 5% B for 10 min, and then solvent
B was increased to 20% within 10 min and then to 70% within 5 min.
After freeze-drying, the title compound (0.5 mmol; 50% yield) was
obtained as a white, amorphous powder in high purity of >99%.
2
1
7
H, H-C(8)], 3.89 [dd, 1H, J ) 2.0, 12.1 Hz, H-C(6b′)], 3.92 [dd,
H, J ) 9.0, 9.0 Hz, H-C(2′)], 5.45 [d, 1H, J ) 9.1 Hz, H-C(1′)],
.10 [dd, 1H, J ) 8.0, 8.0 Hz, H-C(6)], 7.20 [dd, 1H, J ) 8.0, 8.0
Hz, H-C(5)], 7.38 [s, 1H, H-C(2)], 7.57 [dd, 2H, J ) 8.6, 8.6 Hz,
1
3
H-C(4, 7)]; C NMR (100 MHz, d
3
-MeOD; HMQC, HMBC), δ 30.0
[C(8)], 50.9 [C(10)], 61.2 [C(6′)], 70.0 [C(4′)], 72.2 [C(2′)], 77.7 [C(3′)],
79.2 [C(5′)], 85.2 [C(1′)], 108.3 [C(3)], 110.1 [C(7)], 118.3 [C(4)], 119.6
[
[
C(6)], 121.7 [C(5)], 124.0 [C(2)], 128.2 [C(3a)], 136.9 [C(7a)], 173.3
C(9)].
2
-(4-Hydroxybenzoyloxymethyl)-4-â-D-glucopyranosyloxy-2(E)-
butenenitrile, 3 (Figure 1): UV-vis (acetonitrile), λmax ) 211, 255
+
nm; LC-TOF/MS, C18
H21NO
9
; LC-MS (ESI ), m/z 396 (100, [M +
+
+
1
H] ), 234 (59, [M + H - 162] ); H NMR (400 MHz, d
3
-MeOD;
COSY), δ 3.23 [dd, 1H, J ) 8.4, 8.4 Hz, H-C(2′)], 3.30 [m, H-C(5′)],
.31 [m, 1H, H-C(4′)], 3.39 [dd, 1H, J ) 8.4, 8.4 Hz, H-C(3′)], 3.70
dd, 1H, J ) 4.9, 12.1 Hz, H-C(6a′)], 3.89 [dd, 1H, J ) 1.8, 12.1 Hz,
H-C(6b′)], 4.36 [d, 1H, J ) 8.0 Hz, H-C(1′)], 4.56 [dd, 1H, J ) 6.4,
3
[
1
4.6 Hz, H-C(4a)], 4.68 [dd, 1H, J ) 6.4, 14.6 Hz, H-C(4b)], 4.93
[
7
s, 2H, H-C(1)], 6.87 [m, 2H, H-C(3′′, 5′′)], 6.90 [m, 1H, H-C(3)],
1
3
3
.93 [m, 2H, H-C(2′′, 6′′)]; C NMR (100 MHz, d -MeOD; HMQC,
HMBC), δ 61.1 [C(6′)], 62.9 [C(1)], 66.6 [C(4)], 70.0 [C(5′)], 73.5
[C(2′)], 76.6 [C(4′)], 76.6 [C(3′)], 102.8 [C(1′)], 111.5 [C(2)], 114.8
[C(5)], 114.8 [C(3′′)], 114.8 [C(5′′)], 119.8 [C(1′′)], 131.4 [C(2′′)], 131.4
[C(6′′)], 147.7 [C(3)], 162.4 [C(4′′)], 165.5 [C(7′′)].
2-(4-Hydroxy-3-methoxybenzoyloxymethyl)-4-â-D-glucopyranosyloxy-
2
2
+
(E)-butenenitrile, 4 (Figure 1): UV-vis (acetonitrile), λmax ) 211,
+
63 nm; LC-TOF/MS, C19
H
23NO10; LC-MS (ESI ), m/z 426 (100, [M
+
+
1
H] ), 264 (62, [M + H - 162] ); H NMR (400 MHz, d
3
-MeOD;
COSY), δ 3.23 [dd, 1H, J ) 8.4, 8.4 Hz, H-C(2′)], 3.30 [m, 1H,
H-C(5′)], 3.31 [m, 1H, H-C(4′)], 3.39 [m, 1H, H-C(3′)], 3.70 [dd,
1
H, J ) 4.9, 12.1 Hz, H-C(6a′)], 3.89 [dd, 1H, J ) 1.8, 12.1 Hz,
H-C(6b′)], 3.92 [s, 3H, H-C(8′′)], 4.36 [d, 1H, J ) 8.0 Hz, H-C(1′)],
.57 [dd, 1H, J ) 6.4, 14.6 Hz, H-C(4a)], 4.68 [dd, 1H, J ) 6.4, 14.6
Hz, H-C(4b)], 4.95 [s, 2H, H-C(1)], 6.89 [m, 1H, H-C(5′′)], 6.90
m, 1H, H-C(3)], 7.59 [m, 1H, H-C(2′′)], 7.61 [m, 1H, H-C(6′′)];
4
[
1
3
C NMR (100 MHz, d
3
-MeOD; HMQC, HMBC), δ 54.9 [C(8′′)], 61.1
R
Spectroscopic (NMR, LC-MS, UV-vis) and chromatographic data (t )

1408 J. Agric. Food Chem., Vol. 55, No. 4, 2007
Schwarz and Hofmann
as well as the taste threshold concentration of the synthetic title
compound were identical to the data obtained for the 3-carboxymethyl-
indole-1-N-â-D-glucopyranoside (1) isolated from red currants.
High-Performance Liquid Chromatography (HPLC). The HPLC
apparatus (Jasco, Gross-Umstadt, Germany) consisted of an MD-2010
plus photodiode array detector and two PU 2087 pumps. Chromato-
graphic separations were performed on stainless steel columns packed
with ODS-Hypersil, 5 µm, RP-18 material (ThermoHypersil, Kleinos-
theim, Germany) either in analytical (250 × 4.6 mm i.d., flow rate )
solution was stepwise diluted 1:1 with water, and the dilutions
were then presented in order of increasing concentrations to
trained sensory panelists, who were asked to judge the astringent
oral sensation by application of the recently developed half-
tongue test (3) until the threshold was reached. As this so-called
TD factor, obtained for each fraction, is related to its taste
activity in water, the GPC fractions I-VII were rated according
to their relative astringency impact (Figure 2). Due to its high
TD factor of 32, GPC fraction II was evaluated with the highest
taste impact for astringency. In comparison, the other fractions
were evaluated with a significantly lower astringent taste impact.
1
.0 mL/min) or preparative scale (250 × 21.2 mm i.d., flow rate ) 20
mL/min).
LC-Time-of-Flight/Mass Spectrometry (LC-TOF/MS). High-
resolution mass spectra of the compounds were measured on a Bruker
Micro-TOF (Bruker Daltronics, Bremen, Germany) mass spectrometer
and referenced to sodium formate. The deviation of the measured from
the calculated molecular mass was <1.5 ppm.
LC-MS/MS. Electrospray ionization (ESI) mass and product ion
spectra were acquired on an API 4000 QTRAP mass spectrometer
To isolate the taste compounds exhibiting the astringent oral
sensation, GPC fraction II was further fractionated by HPLC
on RP-18 material (Figure 3). By monitoring the effluent at
2
72 nm, the effluents of the individual peaks were collected
and freed from solvent under vacuum, and aliquots of the
isolates were taken up in water in their “natural” concentrations
and were then sensorially evaluated for astringency. Finally,
the compounds isolated from peaks 1-4 (Figure 3) were
purified by RP-18 rechromatography, thus affording the pure
astringent stimulants, which were analyzed by means of UV-
vis, LC-MS/MS, and NMR spectroscopy.
(Applied Biosystems, Darmstadt, Germany) with direct flow infusion.
For ESI, the ion spray voltage was set at -4500 V in the negative
mode and at 5500 V in the positive mode. The mass spectrometer was
operated in the full-scan mode detecting positive or negative ions. The
-
MS/MS parameters were set to induce fragmentation of the [M - H]
+
or [M + H] molecular ions into specific product ions after collision
-
5
with nitrogen as collision gas (4 × 10 Torr).
Identification of the Astringent Indoles 1 and 2. LC-MS
analysis of the astringent compound 1 (Figure 1), exhibiting a
UV-vis absorption maximum at 267 nm, showed the pseudo-
High-Resolution Gas Chromatography-Mass Spectrometry
HRGC-MS). Electron impact (EI) GC-MS data were acquired on an
(
HP 6890 series gas chromatograph and an HP 5973 mass spectrometer
Hewlett-Packard/Agilent, B o¨ blingen, Germany). Chromatographic
+
molecular ion [M + H] with m/z 338, indicating the presence
(
of one nitrogen atom in the target molecule. Further LC-MS/
separation was performed on a 60 m × 0.25 mm fused-silica capillary
with a 0.25 µm methyl silicone coating (J&W Scientific DB-1 column,
Agilent) using 0.6 mL/min helium as carrier gas. The injector
temperature was set at 250 °C, and the injection volume was 1 µL
with split injection (1:50). The initial oven temperature was set to
+
MS studies performed in the ESI mode revealed a loss of 162
amu to give the daughter ion m/z 176 and an additional loss of
46 amu to give the daughter ion m/z 130, thus indicating the
cleavage of a hexose moiety and one molecule of CO2. Acidic
hydrolysis of an aliquot of the isolated compound, derivatization
of the liberated sugar to give the persilylated aldoxime, followed
by HRGC-MS analysis, unequivocally identified glucose as the
carbohydrate moiety present in the tastant.
1
40 °C and then raised at a rate of 4 °C/min to 210 °C, thereafter at a
rate of 8 °C/min to 300 °C, and, finally, held isothermally for 10 min
at 300 °C. Heating the transfer line at 300 °C, the mass spectrometer
was operated in the electron impact mode (EI; 70 eV electron energy)
with a source temperature of 230 °C and the quadrupole heated at 150
1
D and 2D NMR spectroscopic experiments revealed the
°
C. Mass spectra were acquired in the full-scan mode ranging from
anomeric proton signal H-C(1′) resonating at 5.56 ppm, the
large coupling constant (9.1 Hz) of which indicated a â-con-
figuration of the hexoside. Furthermore, heteronuclear multiple
bond correlation experiments (HMBC) demonstrated a long-
range correlation between the anomeric proton H-C(1′) of the
â-D-glucose moiety and the carbon atoms C(2) and C(7a) of
the aglycone. All of the protons and carbons in the molecule
m/z 40 to 800 with a scan rate of 2.0 scans/s.
1
Nuclear Magnetic Resonance Spectroscopy (NMR). H, 13C, and
2
D NMR data were acquired on a Bruker DPX-400 (Bruker BioSpin,
2 3
Rheinstetten, Germany). D O and d -MeOD were used as solvents, and
chemical shifts were referenced to the solvent signal. For structural
elucidation and NMR signal assignment, COSY, HMQC, HMBC, and
ROESY experiments were carried out using the pulse sequences taken
from the Bruker software library. Data processing was performed by
using XWin-NMR software (version 3.5; Bruker, Rheinstetten, Ger-
many) as well as Mestre-C (Mestrelab Research, A Coru n˜ a, Spain).
1
1
were undoubtedly assigned by means of H/ H (COSY) and
1
13
H/ C correlation spectroscopy (HMQC and HMBC).
To further confirm the structure of compound 1, 3-carboxy-
methyl-indole-1-N-â-D-glucopyranoside was synthesized using
a modified Koenigs-Knorr methodology as outlined in Figure
4. Two equivalents of R-bromotetra-O-acetyl-D-glucose were
reacted with 1 equiv of indole-3-acetic acid in the presence of
silver trifluoromethane sulfonate and 2,6-di-tert-butyl-4-meth-
ylpyridine in anhydrous 1,2-dichloroethane. Treatment of the
intermediary diglycoside with a methanolic solution of sodium
methylate afforded 3-carboxymethyl-indole-1-N-â-D-glucopy-
ranoside after HPLC purification in yields of about 50%. The
LC-MS, NMR, and chromatographic data, as well as the
threshold concentration of that compound, were identical to
those determined for the authentic compound 1 isolated from
red currants.
RESULTS AND DISCUSSION
With the aim of isolating the astringent nitrogen-containing
phytochemicals, red currant puree was extracted with methanol/
water to obtain the methanol solubles, which were further
separated into a highly polar fraction and a semipolar fraction
by means of ethyl acetate fractionation as reported recently (2).
After removal of trace amounts of solvents under vacuum, the
ethyl acetate extractables, exhibiting an intense astringent and
mouth-drying oral sensation, were further separated by means
of GPC.
GPC. To locate the key astringent compounds in the ethyl
acetate extractables, this fraction was further separated by GPC
using Sephadex LH-20 as the stationary phase and a methanol/
water gradient as the mobile phase. Monitoring the effluent at
Taking all of these data into consideration, LC-MS/MS and
1D and 2D NMR experiments, as well as synthesis, led to the
unequivocal identification of the astringent compound isolated
from HPLC fraction 1 of the GPC fraction II as 3-carboxy-
methyl-indole-1-N-â-D-glucopyranoside (1, Figure 1).
272 nm, seven fractions I-VII (Figure 2) were collected,
individually freeze-dried, then dissolved in the same amount
of water, and, finally, used for the TDA. To achieve this, each

N-Containing Astringent Tastants in Red Currants
J. Agric. Food Chem., Vol. 55, No. 4, 2007 1409
The astringent compound isolated from HPLC fraction 2
Table 1. Human Recognition Taste Thresholds of Nitrogen-Containing
Astringent Compounds
(
Figure 3) showed a UV-vis absorption maximum at 267 nm
being well in line with the data obtained for compound 1. LC-
+
threshold concn
MS and LC-MS/MS (ESI ) analysis of compound 2 (Figure
a
+
for astringency
1
) showed a pseudomolecular ion [M + H] with m/z 352 and
taste compound (no.)
-(4-hydroxybenzoyloxymethyl)-4-
pyranosyloxy-2(E)-butenenitrile (3)
(mmol/L)
a daughter ion with m/z 190. 1D and 2D NMR spectroscopic
experiments revealed an anomeric proton signal at 5.45 ppm
showing a large coupling constant of 9.1 Hz and indicating the
existence of a â-glucoside as already found for compound 1.
In addition, the quaternary carbon resonating at 173.3 ppm as
well as the proton signal observed at 3.71 ppm and integrating
for three protons indicated the presence of a methyl ester.
Careful consideration of all the data obtained from COSY,
HMQC, and HMBC experiments clearly demonstrated the
structure of taste compound 2 as the 3-methylcarboxymethyl-
indole-1-N-â-D-glucopyranoside (Figure 1). To check that
compound 2 is not formed as an artifact upon esterification of
compound 1 with methanol during workup, the extraction of
red currant puree was repeated by using ethanol instead of
methanol as solvent. As HPLC-DAD and HPLC-MS analysis
demonstrated the presence of compound 2 also when using
ethanol as solvent (data not shown), the indole ester 2 has to
be considered as a naturally occurring phytochemical in red
currants.
2
2
â-D-gluco-
5.9
-(4-hydroxy-3-methoxybenzoyloxymethyl)-4-
â
-
D-
1.2
glucopyranosyloxy-2(E)-butenenitrile (4)
3-methylcarboxymethyl-indole-1-N-
pyranoside (2)
â
-
D
-gluco-
0.001
0.0003
0.5b
3
-carboxymethyl-indole-1-N-
-tryptophan
indole-3-acetic acid
â-D-glucopyranoside (1)
L
0.01
3
5
3
-
â
-
D
-glucopyranosyloxyindole
-hydroxyindole-3-acetic acid
-(2-aminoethyl)-5-hydroxyindole (serotonin)
0.001
0.005
0.005
0.001
5-hydroxyindole
a
Taste threshold concentrations were determined by means of the half-tongue
-tryptophan showed bitter taste above 6000
test (3). b In addition to astringency,
mol/L.
L
µ
the methylene protons H-C(1) and the olefinic proton H-C(3),
but not between the methylene protons H-C(4) and proton
H-C(3). In addition, the NMR data of the 2-butanenitrile
substructure were consistent with literature data reported for
Following precisely the method reported in the literature (5),
-methylcarboxymethyl-indole-1-N-â-D-glucopyranoside was
3
2
2
-[(E)-p-coumaroyloxymethyl]-4-O-â-D-glucopyranosyloxy-
(E)-butenenitrile, coined nigrumin-5-p-coumarate, and the
synthesized (data not shown). The LC-MS/MS and NMR as
well as sensory data of the synthetic compound were identical
with those obtained for the taste compound 2 (Figure 3) isolated
from red currants. Taking all of these data into consideration,
the astringent compound isolated from HPLC fraction 2 of the
GPC fraction II was unequivocally identified as 3-methylcar-
boxymethyl-indole-1-N-â-D-glucopyranoside (2, Figure 1).
Although indole-3-acetic acid has long been known as a plant
hormone of the auxin family, neither its N-glucoside 1 nor its
N-glucoside methyl ester 2 was reported earlier as an astringent
taste compound in foods.
corresponding nigrumin-5-ferulate isolated from black currant
seed (6). Therefore, the structure of astringent compound 3 was
unequivocally identified as the previously not reported noncya-
nogenic nitrile 2-(4-hydroxybenzoyloxymethyl)-4-O-â-D-glu-
copyranosyloxy-2(E)-butenenitrile, named nigrumin-4-hydroxy-
benzoate (Figure 1).
LC-MS and LC-MS/MS analysis of taste compound 4 in the
+
ESI mode showed a pseudomolecular ion with m/z 426 and a
loss of 162 amu to give the daughter ion m/z 264, thus indicating
the presence of a hexose moiety in the molecule. In addition,
Identification of the Astringent Nitriles 3 and 4. After
HPLC purification of the astringent compound 3 isolated from
GPC fraction II, UV-vis and LC-MS analysis showed an
absorption maximum at 255 nm and a pseudomolecular ion with
m/z 396 indicating the presence of one nitrogen atom in the
molecule. 1D and 2D NMR spectroscopic experiments led to
the identification of an anomeric proton resonating at 4.36 ppm.
This proton signal showed a large coupling constant of 8.0 Hz,
thus indicating a â-configuration of a glycoside. Acidic hy-
drolysis of an aliquot of the isolated compound and derivati-
zation of the liberated sugar to give the persilylated aldoxime,
followed by HRGC-MS analysis, unequivocally identified
glucose as the carbohydrate moiety present in the tastant. The
assignment of all the protons and carbons in compound 3 was
successfully achieved by means of COSY, HMQC, and HMBC
experiments. For example, the HMBC experiment showed a
heteronuclear coupling between the proton signals observed at
+
the ESI-TOF/MS spectrum of compound 4 gave a [M + Na]
ion with m/z 448.1228 consistent with the molecular formula
of C19H23NO10. 1D/2D NMR spectroscopic experiments were
performed and showed wide similarity to the data found for
compound 3. As the major difference, the protons of a methoxy
group resonating at 3.92 ppm were detectable and the aromatic
proton H-C(3′′) was lacking. Considering all of the spectro-
scopic data obtained, the structure of taste compound 4 was
identified as the previously unreported noncyanogenic nitrile
2-(4-hydroxy-3-methoxybenzoyloxymethyl)-4-O-â-D-glucopy-
ranosyloxy-2(E)-butene nitrile, named nigrumin-4-hydroxy-3-
methoxybenzoate (Figure 1).
Sensory Activity of Astringent Nitrogen Compounds. Prior
to sensory analysis, the purity of all compounds was checked
1
by HPLC-MS as well as H NMR spectroscopy. To study the
sensory activity of these nitrogen-containing compounds, the
human recognition thresholds for astringency were determined
in bottled water (pH 4.5) using the half-tongue test (Table 1).
The oral sensation imparted by the nitrogen-containing
compounds was described as mouth-drying and astringent with
threshold concentrations ranging from 0.0003 to 5.9 µmol/L.
The nitriles 3 and 4 showed threshold concentrations of 5.9 and
1.2 µmol/L, which is rather similar to the values found for most
of the naturally occurring flavon-3-ol glycosides (3). Surpris-
ingly, two indole derivatives, 1 and 2, were found to induce a
mouth-drying astringency at extraordinarily low threshold
concentrations of 0.0003 and 0.001 µmol/L (Table 1). On the
6
.87 and 7.93 ppm and the quaternary carbon atoms resonating
at 165.5 and 162.4 ppm, thus indicating 4-hydroxybenzoic acid
as part of the molecular structure. In addition, the HMBC
spectrum displayed a long-range correlation between H-C(1)
and C(7′′) and between the anomeric proton resonating at 4.36
and carbon atom C(4) of the aglycone, being well in line with
the structure of 2-(4-hydroxybenzoyloxymethyl)-4-O-â-D-glu-
copyranosyloxy-2(E)-butenenitrile given in Figure 1. The
E-configuration of the 2-butenenitrile moiety was confirmed by
a ROESY experiment that clearly showed a NOE effect between

1410 J. Agric. Food Chem., Vol. 55, No. 4, 2007
Schwarz and Hofmann
basis of these data, the N-glucoside 1 was found with the lowest
astringency threshold concentration reported for any astringent
phytochemical so far.
the perception of astringent-type oral sensations will be a
challenging, but important, task for future investigations.
To gain some insights into the structural requirements for
the low taste threshold of 1, indole-3-acetic acid lacking the
glucose moiety was analyzed by the sensory panel. As shown
in Table 1, the threshold of 0.01 µmol/L found for indole-3-
acetic acid was also rather low, but was 33 times higher
compared to that of the N-glucoside 1. These data clearly
confirm earlier observations made for flavon-3-ol O-glycosides
LITERATURE CITED
(1) 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.
(
2) Schwarz, B.; Hofmann, T. Sensory-guided decomposition of red
currant juice (Ribes rubrum) and structure determination of key
astringent compounds. J. Agric. Food Chem. 2007, 55, 1394-
(
(
3), ellagitannin C-glycosides (7), and flavan-3-ol C-glycosides
8), showing that the astringency threshold of the corresponding
phenolic aglycone is significantly lowered when glycosidically
linked to a carbohydrate moiety. To confirm that assumption
with an indole-O-glycoside, 3-â-D-glucopyranosyloxyindole was
evaluated, too, and, indeed, this O-glucoside showed a rather
low threshold of 0.001 µmol/L (Table 1).
1
404.
(
3) Scharbert, S.; Holzmann, N.; Hofmann, T. Identification of the
astringent taste compounds in black tea infusions by combining
instrumental analysis and human bioresponse. J. Agric. Food
Chem. 2004, 52, 3498-3508.
Differing from the indole-3-acetic acid by just the presence
of the additional amino function, the amino acid L-tryptophan
was also found to induce a mouth-drying and astringent
sensation above a threshold concentration of 0.5 µmol/L,
whereas bitter taste was perceived at concentration levels above
(4) Desmares, G.; Lefebvre, D.; Renevret, G.; Drian, C. Selective
formation of â-D-glucosides of hindered alcohols. HelV. Chim.
Acta 2001, 84, 880-889.
5) Franklin, R. L.; Sell, H. M. Methyl-1-(â-D-glucopyranosyl)-3-
indole acetate J. Chem. Eng. Data 1969, 2, 14.
(
(
6) Lu, Y.; Foo, Y.; Wong, H. Nigrumin-5-p-coumarate and nigru-
min-5-ferulate, two unusual nitrile-containing metabolites from
black currant (Ribes nigrum) seed. Phytochemistry 2002, 59,
6
000 µmol/L (Table 1). However the astringency threshold of
L-tryptophan was 50 times above the value found for the indole-
3-acetic acid.
4
65-458.
Because the astringent activity of indoles were reported in
(
(
(
7) Glabasnia, A.; Hofmann, T. Sensory-directed identification of
taste-active ellagitannins in American (Quercus alba L.) and
European oak wood (Quercus robur L.) and quantitative analysis
in bourbon whiskey and oak-matured red wines. J. Agric. Food
Chem. 2006, 54, 3380-3390.
8) Stark, T.; Hofmann, T. Application of a molecular sensory
science approach on alkalized cocoa (Theobroma cacao)s
structure determination and sensory activity of non-enzymatically
C-glycosylated flavan-3-ols. J. Agric. Food Chem. 2006, 54,
the present investigation for the first time and to gain some
insight into the influence of hydroxylation on the taste activity
of indoles, dilute aqueous solutions of 5-hydroxyindole-3-acetic
acid, 3-(2-aminoethyl)-5-hydroxyindole (serotonin), and 5-hy-
droxyindole were evaluated sensorially. These indoles were also
found to exhibit astringency at rather low threshold concentra-
tions between 0.001 and 0.005 µmol/L, thus indicating that a
hydroxyl group at the 5-position might be beneficial for the
astringency activity of indoles.
In summary, nitrogen-containing phytochemicals and, in
particular, those with indole structure were reported for the first
time as powerful astringent and mouth-drying compounds
exhibiting much lower threshold concentrations as compared
to polyphenols such as flavan-3-ols and flavon-3-ols. As
serotonin also showed mouth-drying and astringent activity, we
might hypothesize that certain serotonin receptor candidates,
which are known to be present in the oral cavity (9), are involved
in the human sensation of mouth-drying and astringency. The
involvement of receptor proteins in the human astringency
sensation is further corroborated by the recent finding that the
neurotransmitter γ-aminobutyric acid also induces a mouth-
drying and velvety astringent sensation (10) and also for this
amino acid multiple receptors are known to be present in the
oral cavity (11). Understanding the role of such receptors in
9
510-9521.
9) Ryotaroa, H.; Tsutomua, K.; Maia, M.; Yokoa, H.; Yoshitakaa,
O.; Takashib, K.; Kiyonori, Y. Cell-networks in mouse taste
buds. Int. Congress Series 2004, 1269, 53-56.
(10) Rotzoll, N.; Dunkel, A.; Hofmann, T. Activity-guided identifica-
tion of (S)-malic acid 1-O-D-glucopyranoside (morelid) and
γ-aminobutyric 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.
(
11) Nagai, T.; Delay, R.; Welton, J.; Roper, S. D. Uptake and release
3
3
of neurotransmitter candidates, [ H]serotonin, [ H]glutamate, and
[
3
H]-γ-aminobutyric acid, in taste buds of the mudpuppy,
Necturus maculosus. J. Comp. Neurol. 1998, 392, 199-208.
Received for review November 7, 2006. Revised manuscript received
December 8, 2006. Accepted December 20, 2006.
JF0632076