Key Phytochemicals Contributing to the Bitter Off-Taste of Oat (Avena sativa L.)

Article abstract of DOI:10.1021/acs.jafc.6b04995

Sensory-directed fractionation of extracts prepared from oat flour (Avena sativa L.) followed by LC-TOF-MS, LC-MS/MS, and 1D/2D-NMR experiments revealed avenanthramides and saponins as the key phytochemicals contributing to the typical astringent and bitter off-taste of oat. Besides avenacosides A and B, two previously unreported bitter-tasting bidesmosidic saponins were identified, namely, 3-(O-α-l-rhamnopyranosyl(1→2)-[β-d-glucopyranosyl(1→3)-β-d-glucopyranosyl(1→4)]-β-d-glucopyranosid)-26-O-β-d-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol, and 3-(O-α-l-rhamnopyranosyl(1→2)-[β-d-glucopyranosyl(1→4)]-β-d-glucopyranosid)-26-O-β-d-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol. Depending on the chemical structure of the saponins and avenanthramides, sensory studies revealed human orosensory recognition thresholds of these phytochemicals to range between 3 and 170 μmol/L.

Full text of DOI:10.1021/acs.jafc.6b04995

Article
pubs.acs.org/JAFC
Key Phytochemicals Contributing to the Bitter Off-Taste of Oat
(Avena sativa L.)
Kirsten Gunther-Jordanland,§ Corinna Dawid,§ Maximilian Dietz,§ and Thomas Hofmann*,§,#
̈
§Chair for Food Chemistry and Molecular Sensory Science, Technische Universitat Munchen, Lise-Meitner-Straße 34, D-85354
̈
̈
Freising, Germany
#Bavarian Center for Biomolecular Mass Spectrometry, Gregor-Mendel-Straße 4, D-85354 Freising, Germany
ABSTRACT: Sensory-directed fractionation of extracts prepared from oat flour (Avena sativa L.) followed by LC-TOF-MS, LC-
MS/MS, and 1D/2D-NMR experiments revealed avenanthramides and saponins as the key phytochemicals contributing to the
typical astringent and bitter off-taste of oat. Besides avenacosides A and B, two previously unreported bitter-tasting bidesmosidic
saponins were identified, namely, 3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→3)-β-D-glucopyranosyl(1→4)]-β-D-
glucopyranosid)-26-O-β-D-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol, and 3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-
glucopyranosyl(1→4)]-β-D-glucopyranosid)-26-O-β-D-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol. Depending on the
chemical structure of the saponins and avenanthramides, sensory studies revealed human orosensory recognition thresholds
of these phytochemicals to range between 3 and 170 μmol/L.
KEYWORDS: oat, Avena sativa, saponin, avenanthramide, taste, taste dilution analysis, bitter, astringent, avenacoside
Therefore, the objective of the present study was to locate
the key molecules responsible for the typical astringent and
bitter taste of oats by application of the taste dilution technique,
to determine the chemical structure of the most taste-active
phytochemicals by means of LC-MS/MS, UPLC-TOF-MS and
1D/2D-NMR experiments, and to evaluate their sensory
activity on the basis of their human recognition thresholds.
INTRODUCTION
■
Within the past years, oats (Avena sativa L.) have become fairly
popular among cereal crops and are consumed as a key
ingredient of breakfast cereals such as porridge and muesli,
cookies, breads, and oat drinks, respectively. Besides nutrition-
ally favorable amounts of proteins, soluble fiber, unsaturated
fatty acids, vitamins, minerals, and polyphenols, oats have been
reported to show health-promoting attributes in blood
cholesterol and coronary heart diseases.1,2 Moreover, oats are
appreciated for their characteristic flavor, which is described
with a nutty, sweet, and cereal-like aroma.3 This aroma is paired
with a slight sweet, bitter, and astringent after-taste that is
considered an off-taste if perceived as being too intense.
Mainly lipid-derived compounds are assumed to contribute
to the typical bitterness of oat and oat fibers, among them 9-
hydroxy-trans,cis-10,12-octadecadienoic-1′-monoglyceride, 13-
hydroxy-cis,trans-9,11-octadecadienoic-1′-monoglyceride, 9-hy-
droxy-trans,cis-10,12-octadecadienoic acid, 13-hydroxy-cis,trans-
9,11-octadecadienoic acid, 9,12,13-trihydroxy-trans-10-octade-
cenoic acid, and 9,10,13-trihydroxy-trans-11-octadecenoic
acid.4−8 Moreover, glycosidic phytochemicals, such as avenaco-
sides A and B, have been isolated from oat and hypothesized to
contribute to the characteristic bitter taste.9,10 Although the
content of phenolic compounds seems to show some
correlation of avenanthramides with perceived bitterness,
freshness, and rancidity of oats,11 the knowledge of individual
key taste molecules imparting the bitter taste of oats is rather
fragmentary. To selectively sort out such key taste molecules
from the bulk of sensorially inactive components in foods, the
so-called taste dilution analysis (TDA)12 has been developed
and successfully applied in recent years to elucidate the most
important tastants such as bitter compounds in asparagus,13−15
avocado,16 and hops,17,18 pungent and tingling phytochemicals
in black pepper,19 and astringent molecules in roasted cocoa
nibs20 and red wine.21
MATERIALS AND METHODS
■
Chemicals and Materials. The following compounds were
obtained commercially: formic acid, hydrogen chloride, sodium
chloride, sodium bicarbonate, caffeine, L-lactic acid (Merck, Darmstadt,
Germany); L-glutamic acid monohydrate, anthranilic acid, 5-hydrox-
yanthranilic acid, Meldrum’s acid, p-hydroxybenzaldehyde, 3,4-
dihydroxybenzaldehyde, syringaldehyde, vanillin, p-anisaldehyde, pyr-
idine, toluene (Sigma-Aldrich, Steinheim, Germany); sucrose (Carl
Roth, Karlsruhe, Germany); sulfuric acid, hydrochloric acid, β-alanine,
and sodium bicarbonate (Merck). Deuterated solvents were supplied
by Euriso-Top (St. Aubin, France). Solvents for HPLC applications
were of HPLC grade (Merck), and solvents for LC-MS uses were of
LC-MS grade (J. T. Baker, Deventer, The Netherlands), whereas
solvents used for extraction (Merck) were distilled prior to use. Water
for HPLC separation was purified by means of a Milli-Q water
advantage A 10 water system (Millipore, Molsheim, France). Bottled
water (Evian, low mineralization = 405 mg/L) was adjusted to pH 6.4
with aqueous formic acid prior to gustatory analysis. Ethanol (absolute,
Merck) was utilized for sensory analysis. Ground oat flour from heat-
treated oat grains was used and stored at −20 °C until use.
Analytical Sensory Analyses. Training of the Sensory Panel.
According to the literature22 and to familiarize the panelists (six
females, six males; 22−35 years in age) with the taste language,
assessors were trained on different qualities of oral sensation. The
Received: November 7, 2016
Revised: December 2, 2016
Accepted: December 6, 2016
Published: December 6, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
panelists, who had no history of known taste disorders and had given
informed consent to participate in the following sensory studies, took
part in weekly sessions for at least two years. Aqueous solutions (2
mL; pH 6.4) of the following standard taste components were used for
training: sucrose (50 mM) for sweet taste, L-lactic acid (20 mM) for
sour taste, NaCl (20 mM) for salty taste, caffeine (1 mM) for bitter
taste, and sodium glutamate (3 mM) for umami taste. For the
astringent oral sensation, tannic acid (0.05%) was used to train the
panelists using the so-called half-tongue test.21,22 All sensory analyses
were performed in a sensory panel room at 22−25 °C in three
different sessions using nose clips to avoid cross-model interactions
with odorants.
Table 1. Yields and Sensory Evaluation of Fractions Isolated
from Oat Flour
a
fraction
b
taste intensity of
oat flour
I
II
III
IV
bitterness
astringency
sweetness
sourness
2.0
1.1
0.5
0.3
0.1
0.1
0.8
1.0
0.6
0.3
0.2
0.1
2.6
1.2
0.5
0.8
n.d.
0.4
1.2
0.7
0.5
0.4
0.2
0.3
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
saltiness
umami taste
Pretreatment of Food Fractions and Taste Compounds for
Sensory Analysis. Prior to sensory analysis, fractions or compounds
isolated from oat flour were suspended in water and freeze-dried twice,
after removal of the volatiles in high vacuum (<5 mPa). 1H NMR and
ion chromatography revealed that food fractions treated by that
procedure are essentially free of solvents and buffer compounds.
Taste Profile Analysis. Oat flour (15 g) was suspended in water (50
mL, pH 6.4) containing 3% ethanol and, after filtration (folded paper
a
Individual fractions obtained by solvent extraction of oat flour with n-
hexane (fraction I, yield 8.1 g/100 g), methanol/water (fraction II,
yield 2.5 g/100 g), and methanol (fraction III, yield 0.7 g/100 g) to
give the nonsoluble residue (fraction IV, yield 84.2 g/100 g). The
taste intensity of aqueous solutions of the “natural” concentrations of
individual fractions in bottled water (pH 6.4, 3% ethanol) was rated on
a scale from 0 (not detectable) to 5 (strongly detectable).
b
filter, 125 mm; Machery Nagel, Duren, Germany), the aqueous
̈
supernatant was presented to the trained sensory panel to evaluate the
taste sensations sweet, sour, umami, salty, bitter, and astringent on a
linear intensity scale from 0 (not detectable) to 5 (strongly
detectable). In the same way, aliquots of oat fractions I−IV, taken
up in “natural” concentrations in bottled water (30 mL; pH 6.4, 3%
ethanol), were offered to rate the taste intensities.
Taste Dilution Analysis (TDA). Aliquots of the HPLC fractions
were taken up in “natural” concentrations in bottled water (6.0 mL;
pH 6.4, 1% ethanol) and, then, sequentially diluted 1:1 (v/v) with
bottled water (pH 6.4, 1% ethanol). The series of dilutions were
presented randomly to the trained panel in order of increasing
concentration. Panelists were asked to evaluate the dilution step at
which a difference between the sample and a control (water, pH 6.4,
1% ethanol) could just be detected by means of the half-tongue
test.21−23 The so-called taste dilution (TD) factors were calculated by
the geometric mean of all individual threshold concentrations in
separate sessions and did not differ by more than plus/minus one
dilution step.24,25
Solvent Extraction of Oat Flour. A mixture of oat flour (300 g)
and freshly distilled n-hexane (1 L) was vigorously stirred for 3 h at
room temperature, followed by centrifugation for 25 min at 4 °C and
cold filtration. The filtrates were collected, and the procedure was
repeated three times to give the hexane extractables (fraction I). The
residual oat flour was extracted three times (1 L) for 8 h with a
methanol/water mixture (70:30, v/v) adjusted to pH 4.0 with aqueous
formic acid. After filtration, the combined filtrates were removed from
solvent by vacuum evaporation at 40 °C, followed by lyophilization to
give the methanol/water extractables (fraction II). The residue was
extracted with methanol (3 × 1 L) and, after filtration, the combined
filtrates were separated from solvent in vacuum, followed by
lyophilization to give the methanol extractables (fraction III). After
lyophilization of the remaining oat residue, insoluble fraction IV was
obtained. The individual fractions I−IV were freeze-dried twice to
remove trace amounts of solvents. Their yields were determined by
weight, and their taste profiles were evaluated in aqueous solutions as
given in Table 1.
Solid Phase Separation of Oat Fraction II. An aliquot (900 mg)
of freeze-dried oat fraction II was suspended in 0.1% aqueous formic
acid (10 mL) and applied onto a Strata C18-E cartridge (10 g/60 mL,
Phenomenex, Aschaffenburg, Germany), preconditioned with meth-
anol (60 mL), followed by methanol/water (50:50, v/v, 60 mL), and
0.1% aqueous formic acid (60 mL). Using a vacuum extraction box (J.
T. Baker, Phillipsburg, NJ, USA), the cartridge was flushed stepwise
with volumes (30 mL) of 0.1% aqueous formic acid to give fraction II-
A (yield 67.6%), methanol/aqueous formic acid (40:60, v/v) to give
fraction II-B (yield 7.8%), methanol/aqueous formic acid (60:40, v/v)
to give fraction II-C (yield 3.0%), methanol/aqueous formic acid
(80:20, v/v) to give fraction II-D (yield 3.0%), and methanol to give
fraction II-E (yield 3.9%). Fractions were collected separately, freed
from solvent under vacuum, taken up in water, and lyophilized twice
for sensory analysis. Yields were determined gravimetrically and
extracts kept at −20 °C until used for sensory and chemical analysis,
respectively.
Identification of Key Taste Compounds in Subfraction II-D.
Oat subfraction II-D was dissolved in acetonitrile/water (20:80, v/v;
24 mg/600 μL) and, after membrane filtration, was injected onto a 250
× 21.0 mm, 5 μm, Nucleodur C18 Pyramid column (Machery-Nagel,
Duren, Germany) equipped with a guard column of the same type and
̈
operated with a flow rate of 19 mL/min. Using 0.1% aqueous formic
acid as solvent A and acetonitrile as solvent B, chromatography was
performed with the following gradient: 0 min, 23% B; 6 min, 30% B;
16 min, 36% B; 20 min, 65% B; 22 min, 100% B; 24 min, 100% B; 26
min, 23% B; 30 min, 23% B. The effluent was separated into 16
subfractions (II-D-1−II-D-16), which were collected individually in
multiple HPLC runs. Fractions were combined, removed from solvent
under vacuum evaporation at 40 °C, and freeze-dried twice. The TDA
was performed by dissolving fractions in water containing 1% ethanol
in their “natural” concentration ratios. Bitter-tasting fractions (II-D-4,
II-D-5, II-D-8, II-D-9) were further separated by means of preparative
RP-HPLC using a 250 × 21.0 mm, 5 μm, Nucleodur C18 Pyramid
column (Machery-Nagel) equipped with a guard column of the same
type. Using 0.1% aqueous formic acid as solvent A and acetonitrile as
solvent B, chromatography was performed running an isocratic
gradient of 19% (II-D-4, II-D-5) and 33% solvent B (II-D-8, II-D-
9), respectively, for 12 min with a flow rate of 19 mL/min. Fractions
were combined and separated from solvent under vacuum. After
lyophilization, the structures of the key bitter compounds 1 (II-D-4), 2
(II-D-5), 3 (II-D-8), and 4 (II-D-9) were identified as 3-(O-α-L-
rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→3)-β-D-
glucopyranosyl(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyrano-
syl-(25R)-furost-5-ene-3β,22,26-triol (1), 3-(O-α-L-rhamnopyranosyl-
(1→2)-[β-D-glucopyranosyl(1→4)]-β-D-glucopyranoside)-26-O-β-D-
glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol (2), and avenaco-
sides B (3) and A (4) by means of LC-MS/MS, UPLC-TOF-MS,
and 1D/2D NMR experiments.
3-(O-α-L-Rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→3-β-D-
glucopyranosyl(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyra-
nosyl-(25R)-furost-5-ene-3β,22,26-triol, 1 (Figure 1): LC-MS (ESI−)
m/z 1225.5 [M − H ]−, 1063.5 [M − Glc − H]−, 901.4 [M − 2Glc −
H]−, 755.4 [M − 2Glc − Rha − H]−, 531.2, 441.1; LC-MS (ESI+) m/
z 1249.5 [M + Na]+, 1209.5 [M − 18 + H]+, 1047.5 [M − Glc − 18 +
H]+, 901.4 [M − Glc − Rha − 18 + H]+, 739.4 [M − 2Glc − Rha −
18 + H]+, 577.3 [M − 3Glc − Rha − 18 + H]+, 415.3 [M − 4Glc −
Rha − 18 + H]+; MS/MS (DP = −225 V) m/z (%) 1225.6 (100),
1063.6 (3), 901.4 (11), 755.4 (9), 575.6 (2), 289.0 (3); LC-TOF-MS
m/z 1225.5867 ([M − H]−, measured), m/z 1225.5853 calculated for
[C57H94O28 − H]−. 1H and 13C NMR data are given in Tables 2 and 3.
B
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 1. Chemical structures of 3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→3)-β-D-glucopyranosyl(1→4)]-β-D-glucopyranosid)-26-
O-β-D-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol (1), 3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→4)]-β-D-glucopyranosid)-
26-O-β-D-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol (2), nuatigenin-3-O-(α-L-rhamnopyranosyl-(1→2)-[β-D-glucopyranosyl-(1→3)-β-D-glu-
copyranosyl-(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyranoside (avenacoside B, 3), nuatigenin-3-O-(α-L-rhamnopyranosyl-(1→2)-[β-D-
glucopyranosyl-(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyranoside (avenacoside A, 4), avenanthramide 2c (5), avenanthramide 2p (6),
avenanthramide 2f (7), avenanthramide 2f (7), avenanthramide 1p (8), avenanthramide 1c (9), avenanthramide 1f (10), avenanthramide 1s (11),
and avenanthramide 2s (12).
3-(O-α-L-Rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→4)]-β-
D-glucopyranoside)-26-O-β-D-glucopyranosyl-(25R)-furost-5-ene-
3β,22,26-triol, 2 (Figure 1). LC-MS (ESI−) m/z 1063.5 [M − H]−,
901.4 [M − Glc − H]−, 755.1 [M − Glc − Rha − H]−, 593.3 [M −
2Glc − Rha − H]−, 298.0; LC-MS (ESI+) m/z 1087.5 [M + Na]+,
1047.5 [M − 18 + H]+, 885.4 [M − Glc − 18 + H]+, 739.4 [M − Glc
− Rha − 18 + H]+, 577.3 [M − 2Glc − Rha − 18 + H]+, 415.3 [M −
3Glc − Rha − 18+H]+; MS/MS (DP = −185 V) m/z (%) 1064.2
(100), 902.2 (48), 755.8 (2), 575.2 (5), 431.2 (10), 289.6 (13); LC-
TOF-MS m/z 1063.5334 ([M − H]−, measured), m/z 1063.5325
1061.5169 ([M − H]−, measured), m/z 1061.5168 calculated for
[C51H82O23 − H]−. 1H and 13C NMR data are given in Tables 4 and 5.
Identification of Key Taste Compounds in Fraction II-C.
Fraction II-C was dissolved in acetonitrile/water (20:80, v/v; 3 mg/80
μL) and, after membrane filtration, was injected onto a 250 × 10.0
mm, 4 μm, Synergi Hydro RP column (Phenomenex) equipped with a
guard column of the same type operated with a flow rate of 6.5 mL/
min. Using 0.1% aqueous formic acid as solvent A and acetonitrile as
solvent B, chromatography was performed with the following gradient:
0 min, 20% B; 1 min, 22% B; 16 min, 30% B; 19 min, 40% B; 21 min,
100% B; 23 min, 100% B; 25 min, 20% B; 27 min, 20% B. The effluent
was separated into 17 subfractions, namely, II-C-1−II-C-17, which
were collected individually in several runs. The solvent was removed
under vacuum evaporation at 40 °C. After freeze-drying twice, the
fractions were used for sensory experiments, and fractions 5 (II-C-11),
6 (II-C-15), and 7 (II-C-16) were identified to contain avenan-
thramides 2c (5), 2p (6), and 2f (7) by means of UV−vis, LC-MS/
MS, UPLC-TOF-MS, and 1D/2D NMR experiments, respectively.
N-(3′,4′-Dihydroxy-(E)-cinnamoyl)-5-hydroxyanthranilic acid
(avenanthramide 2c), 5 (Figure 1): UV−vis (ACN) λmax = 216 nm,
232 nm, 316 nm; LC-TOF-MS m/z 314.0669 ([M − H]−, measured),
m/z 314.0665 calculated for [C16H13NO6 − H]−; 1H NMR (500
MHz, MeOD-d4) δ 6.46 [d, 1H, J = 15.6 Hz, H−C(8′)], 6.79 [d, 1H, J
= 8.3 Hz, H−C(5′)], 6.97 [dd, 1H, J = 2.1, 8.3 Hz, H−C(6′)], 7.00
[dd, 1H, J = 2.9, 9.0 Hz, H−C(4)], 7.07 [d, 1H, J = 2.0 Hz, H−C(2′)],
7.50 [d, 1H, J = 15.7 Hz, H−C(7′)], 7.51 [d, 1H, J = 2.9 Hz, H−
C(6)], 8.44 [d, 1H, J = 9.0 Hz, H−C(3)]; 13C NMR [125 MHz,
1
calculated for [C51H84O23 − H]−. H and 13C NMR data are given in
Tables 2 and 3.
Nuatigenin-3-O-(α-L-rhamnopyranosyl-(1→2)-[β-D-glucopyrano-
syl-(1→3)-β-D-glucopyranosyl-(1→4)]-β-D-glucopyranoside)-26-O-
β-D-glucopyranoside (avenacoside B), 3 (Figure 1): LC-MS (ESI+)
m/z 1063.5 [M − Glc + H]+, 901.4 [M − 2Glc + H]+, 739.4 [M −
3Glc + H]+, 593.3 [M − 3Glc − Rha + H]+, 431.3 [M − 4Glc − Rha +
H]+, 413.3, 271.2; MS/MS (DP = −200 V) m/z (%) 1223.6 (100),
1061.4 (9), 899.6 (12), 753.4 (2), 591.0 (1); LC-TOF-MS m/z
1223.5695 ([M − H]−, measured), m/z 1223.5697 calculated for
[C57H92O28 − H]−. 1H and 13C NMR data are given in Table 4 and 5.
Nuatigenin-3-O-(α-L-rhamnopyranosyl-(1→2)-[β-D-glucopyrano-
syl-(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyranoside or
(avenacoside A), 4 (Figure 1): LC-MS (ESI+) m/z 1085.5 [M + Na
+ H]+, 1063.5 [M + H]+, 901.4 [M − Glc + H]+, 739.4 [M − 2Glc +
H]+, 593.3 [M − 2Glc − Rha + H]+, 431.3 [M − 3Glc − Rha + H]+,
413.3, 271.2; MS/MS (DP = +85 V) m/z (%) 1064.4 (48), 902.4 (21)
756.2 (8), 594.4 (28), 431.4 (77) 413.4 (30); LC-TOF-MS m/z
C
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Table 2. 1H NMR (500 MHz, Pyridine-d6) and 13C NMR Signals (125 MHz, Pyridine-d6) of the Aglycone Moiety of Saponins 1
and 2
ab
,
ab
,
1
2
position
1
δC
HSQC
[CH2]
δH
M [J, Hz]
δC
HSQC
[CH2]
δH
M [J, Hz]
37.8
0.94−1.04
1.70−1.79
1.85−1.90
2.06−2.12
3.83−3.93
2.68−2.81
2.68−2.81
−
5.27−5.32
1.39−1.52
1.82−1.93
1.52−1.63
0.87−0.95
−
1.40−1.52
1.40−1.52
1.07−1.19
1.72−1.80
−
1.04−1.13
1.42−1.52
1.99−2.10
4.92−5.00
1.91−1.99
0.88−0.93
1.02−1.11
2.22−2.29
1.28−1.42
−
m
m
m
m
m
37.8
0.94−1.04
1.70−1.79
1.85−1.91
2.08−2.14
3.83−3.93
2.68−2.81
2.68−2.81
−
5.27−5.32
1.39−1.52
1.82−1.93
1.52−1.63
0.87−0.95
−
1.38−1.52
1.38−1.52
1.08−1.17
1.72−1.79
−
1.00−1.11
1.41−1.52
1.96−2.07
4.92−5.00
1.91−1.99
0.87−0.94
1.02−1.11
2.20−2.27
1.29−1.40
−
m
m
m
m
m
2
30.4
[CH2]
30.5
[CH2]
3
4
78.4
39.2
[CH]
78.4
39.2
[CH]
[CH2]
[CH2]
5
6
7
141.1
122.1
32.7
[C]
[C]
[CH]
[CH2]
d
122.1
32.7
[CH]
[CH2]
d
m
m
8
32.0
50.7
37.4
21.3
[CH]
[CH]
[C]
m
m
32.0
50.7
37.4
21.3
[CH]
[CH]
[C]
m
m
9
10
11
[CH2]
m
[CH2]
m
12
40.2
[CH2]
m
o
40.2
[CH2]
m
o
13
14
15
41.0
56.9
32.8
[C]
41.0
56.9
32.8
[C]
[CH]
[CH2]
m
m
o
[CH]
[CH2]
m
m
o
16
17
18
19
20
21
22
23
81.4
64.1
16.7
19.0
41.1
16.7
111.0
37.4
[CH]
[CH]
[CH3]
[CH3]
[CH]
[CH3]
[C]
−
t
81.4
64.1
16.7
19.0
41.1
16.7
111.0
37.4
[CH]
[CH]
[CH3]
[CH3]
[CH]
[CH3]
[C]
−
t
s
s
s
s
m
d
m
d
[CH2]
1.70−1.78
2.00−2.09
1.68−1.76
2.04−2.11
1.89−1.99
3.58−3.66
3.92−3.98
0.98−1.02
o
[CH2]
1.70−1.78
2.00−2.09
1.67−1.73
2.03−2.08
1.89−1.97
3.58−3.66
3.92−3.98
0.98−1.02
o
o
o
24
28.6
[CH2]
m
o
28.6
[CH2]
m
o
25
26
34.6
75.5
[CH]
m
dd
dd
34.6
75.5
[CH]
m
[CH2]
[CH2]
dd
dd
27
17.7
[CH3]
d [J = 6.6]
b
17.7
[CH3]
d [J = 6.6]
a
Arbitrary numbering according to structures 1 and 2 in Figure 1. Assignments were based on HSQC, HMBC, J-res, and COSY experiments. d,
doublet; m, multiplet signals; o, overlapped with other signals; s, singlet; t, triplet.
MeOD-d4) δ 115.1 [C(2′)], 116.5 [C(5′)], 118.2 [C(6)], 119.5
[C(8′)], 120.4 [C(1)], 121.6 [C(4)], 122.5 [C(6′)], 123.3 [C(3)],
128.1 [C(1′)], 134.7 [C(2)], 143.3 [C(7′)], 146.8 [C(3′)], 149.1
[C(4′)], 154.1 [C(5)], 166.8 [C(9′)], 171.7 [C(7)].
MeOD-d4) δ 3.92 [s, 3H, H−C(10′)], 6.56 [d, 1H, J = 15.6 Hz, H−
C(8′)], 6.82 [d, 1H, J = 8.2 Hz, H−C(5′)], 6.99 [dd, 1H, J = 3.0, 9.0
Hz, H−C(4)], 7.09 [dd, 1H, J = 2.0, 8.1 Hz, H−C(6′)], 7.22 [d, 1H, J
= 1.9 Hz, H−C(2′)], 7.51 [d, 1H, J = 3.0 Hz, H−C(6)], 7.56 [d, 1H, J
= 15.6 Hz, H−C(7′)], 8.45 [d, 1H, J = 9.0 Hz, H−C(3)]; 13C NMR
(101 MHz, MeOD-d4) δ 56.5 [C(10′)], 111.6 [C(2′)], 116.5 [C(5′)],
118.2 [C(6)], 119.3 [C(8′)], 119.6 [C(1)], 122.0 [C(4)], 123.5
[C(3)], 123.8 [C(6′)], 128.1 [C(1′)], 134.8 [C(2′)], 143.4 [C(7′)],
149.4 [C(3′)], 150.2 [C(4′)], 154.2 [C(5)], 166.8 [C(9′)], 171.2
[C(7)].
N-(4′-Hydroxy-(E)-cinnamoyl)-5-hydroxyanthranilic acid (ave-
nanthramide 2p), 6 (Figure 1): UV−vis (ACN) λmax = 216 nm,
320 nm; LC-TOF-MS m/z 298.0716 ([M − H]−, measured), m/z
1
298.0715 calculated for [C16H13NO5 − H]−; H NMR (500 MHz,
MeOD-d4) δ 6.53 [d, 1H, J = 15.6 Hz, H−C(8′)], 6.82 [d, 2H, J = 8.6
Hz, H−C(3′/5′)], 6.98 [dd, 1H, J = 9.0, 3.0 Hz, H−C(4)], 7.48 [d,
2H, J = 8.6 Hz, H−C(2′/6′)], 7.51 [d, 1H, J = 3.0 Hz, H−C(6)], 7.56
[d, 1H, J = 15.6 Hz, H−C(7′)], 8.45 [d, 1H, J = 9.0 Hz, H−C(3)]; 13C
NMR (125 MHz, MeOD-d4) δ 116.8 [C(3′/5′)], 118.3 [C(6)], 119.6
[C(8′)], 120.5 [C(1)], 121.3 [C(4)], 123.2 [C(3)], 127.6 [C(1′)],
130.8 [C(2′/6′)], 134.8 [C(2)], 142.8 [C(7′)], 154.2 [C(5)], 160.9
[C(4′)], 166.5 [C(9′)], 171.1 [C(7)].
Thin Layer Chromatography (TLC). According to the
literature,26 TLC analysis was performed using a silica gel TLC plate
60 coated with fluorescent indicator F254 (Merck) and chloroform/
methanol/water (8:4:1; v/v/v) as the mobile phase. After chromato-
graphic separation, the TLC plate was sprayed with reagent solution
comprising a mixture (85:10:5:0.8; v/v/v/v) of methanol, acetic acid,
sulfuric acid, and p-anisaldehyde and heated in an oven for 2 min at
110 °C.
N-(4′-Hydroxy-3′-methoxy-(E)-cinnamoyl)-5-hydroxyanthranilic
acid (avenanthramide 2f), 7 (Figure 1): UV−vis (ACN) λmax = 216
nm, 332 nm; LC-TOF-MS m/z 328.0851 ([M − H]−, measured), m/z
Acidic Hydrolysis of Compounds 1−4 and Carbohydrate
1
328.0821 calculated for [C17H15NO6 − H]−; H NMR (400 MHz,
Analysis. Trifluoroacetic acid (4 mL) was added to a mixture of the
D
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Table 3. 1H NMR (500 MHz, Pyridine-d6) and 13C NMR Signals (125 MHz, Pyridine-d6) of the Sugar Moieties of Saponins 1
and 2
ab
,
ab
,
1
2
position
δC
HSQC
δH
M [J, Hz]
δC
HSQC
δH
M [J, Hz]
β-D-GlcI (at C-26)
1′
2′
3′
4′
5′
6′
105.2
75.3
78.9
71.9
78.6
62.8
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.80−4.85
4.02−4.10
3.93−3.98
4.19−4.28
4.19−4.28
4.29−4.35
4.53−4.59
d [J = 7.7]
105.2
75.3
78.9
72.0
78.6
62.8
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.80−4.85
4.02−4.10
3.93−4.00
4.18−4.28
4.18−4.28
4.30−4.37
4.50−4.56
d [J = 7.7]
m
o
o
o
m
m
o
m
m
o
o
o
β-D-GlcII (at C-3)
1″
2″
3″
4″
5″
6″
100.3
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.92−4.98
4.20−4.27
3.82−3.88
4.20−4.26
4.19−4.26
4.43−4.50
4.50−4.55
d [J = 7.2]
100.3
77.9
76.5
82.3
78.5
62.2
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.92−4.98
4.19−4.27
3.82−3.88
4.19−4.26
4.19−4.27
4.43−4.50
4.50−4.55
d [J = 7.0]
77.9
76.5
81.8
77.6
61.8
m
m
m
m
m
m
m
m
m
m
m
m
β-D-GlcIII (at C-4GlcII)
1‴
2‴
3‴
4‴
5‴
6‴
104.8
74.1
88.6
69.6
78.3
62.4
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
5.08−5.13
4.04−4.09
4.16−4.21
4.14−4.20
3.86−3.93
4.29−4.36
4.42−4.49
d [J = 7.8]
105.5
75.3
78.0α
71.5
78.8
62.4
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
5.10−5.18
4.01−4.10
4.19−4.26
4.22−4.30
3.91−4.01
4.29−4.36
4.42−4.49
d [J = 7.8]
m
m
m
m
m
m
m
m
m
m
m
m
α-L-Rha (at C-2GlcII)
1⁗
2⁗
3⁗
4⁗
5⁗
6⁗
102.1
72.7
73.1
74.4
69.8
18.9
[CH]
[CH]
[CH]
[CH]
[CH]
[CH3]
6.22−6.27
4.72−4.78
4.55−4.63
4.31−4.38
4.90−4.98
1.72−1.80
bs [J = 2.2]
102.1
72.7
73.1
74.4
69.8
18.9
[CH]
[CH]
[CH]
[CH]
[CH]
[CH3]
6.22−6.27
4.72−4.78
4.55−4.63
4.31−4.38
4.90−4.98
1.72−1.80
bs [J = 2.3]
dd
dd
t.o
o
dd
dd
t.o
o
d
d
β-D-GlcIV (at C-3GlcIII)
1⁗′
2⁗′
3⁗′
4⁗′
5⁗′
6⁗′
106.2
75.5
78.4
72.0
78.9
62.8
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
5.27−5.32
4.05−4.11
4.20−4.28
4.17−4.27
4.21−4.29
4.29−4.35
4.53−4.59
d [J = 8.0]
o
o
o
o
o
o
a
b
Arbitrary numbering according to structures 1 and 2 in Figure 1. Assignments were based on HSQC, HMBC, J-res, and COSY experiments. d,
doublet; m, multiplet signals; o, overlapped with other signals; bs, broad singlet; t, triplet.
target saponin (1 mg) dissolved in methanol (3 mL). After 5 h of
stirring at 90 °C, the mixture was cooled to room temperature and
diluted with water (20 mL), and the pH value was adjusted to 7.0 by
adding an aqueous sodium hydroxide solution. Aliquots (1 mL) were
analyzed by means of high-performance anion exchange chromatog-
raphy using a Dionex ICS-5000 IC (Thermo Fisher Scientific,
Waltham, MA, USA) consisting of a Dionex ICS-5000 DP dual
pump, a capillary and microbore eluent generator, an AS-AP
autosampler, and an ICS-5000 CD type conductivity detector as
well as an ICS-5000 EC type electrochemical detector operating in
pulsed amperometric detection mode. The detector was equipped with
a gold working electrode and a PdH reference electrode supplied by
the manufacturer. Data acquisition and instrumental control were
completed with Chromeleon software (version 6.80, Thermo Fisher
Scientific). Chromatographic separation was performed at 30 °C on a
150 × 0.4 mm CarboPac PA-20 column (Thermo Fisher Scientific)
connected to a CarboPac PA-20 guard column (10 × 0.4 mm, Thermo
Fisher Scientific). The following gradient was applied: a 10 mM
potassium hydroxide solution was run isocratically for 12 min,
increasing to 100 mM in 0.1 min and keeping conditions for 4.9 min;
thereafter, the gradient was adjusted to the initial 10 mM
concentration in 0.1 min and kept for another 19.9 min. After each
sample, the column was washed with a sodium hydroxide solution
(200 mM) and equilibrated with potassium hydroxide solution (10
mM) for 10 min prior to injection. Chromatography was performed
with an injection volume of 0.4 μL and a flow rate of 8 μL/min. Using
carbohydrates as reference chemicals, glucose and rhamnose were
unequivocally identified as the sugars in saponins 1−4.
Syntheses of Avenanthramides (5−12). With some modifica-
tions, avenanthramides 5−12 were synthesized according to the
literature.27 Anthranilic acid (21 mmol) and Meldrum’s acid (21
mmol) were mixed in dried toluene (15 mL) and refluxed for 4 h.
After cooling to room temperature, a saturated NaHCO3 (15 mL)
solution was added, followed by stepwise addition of concentrated
HCl (10 mL). The white precipitate was filtered, washed with water,
and dried at 100 °C to obtain 2-(2-carboxyacetyl)aminobenzoic acid
E
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Table 4. 1H NMR (500 MHz, Pyridine-d6) and 13C NMR Signals (125 MHz, Pyridine-d6) of the Aglycone Moiety of Saponins 3
and 4
ab
,
ab
,
3
4
position
1
δC
HSQC
[CH2]
δH
M [J, Hz]
δC
HSQC
[CH2]
δH
M [J, Hz]
38.5
1.03−1.12
1.84−1.92
1.56−1.65
1.88−1.95
3.54−3.63
2.25−2.33
2.41−2.48
−
5.35−5.41
1.64−1.70
2.01−2.09
1.61−1.69
0.93−1.01
−
1.47−1.61a
1.47−1.61a
1.17−1.24
1.75−1.82
−
1.10−1.18
1.20−1.27
1.94−2.00
4.43−4.50
1.72−1.80
0.78−0.85
0.99−1.10
2.14−2.20
0.96−1.01
−
d
38.5
1.03−1.12
1.84−1.92
1.57−1.63
1.89−1.95
3.56−3.62
2.25−2.33
2.41−2.48
−
5.35−5.41
1.65−1.69
2.03−2.09
1.61−1.69
0.93−1.01
−
1.46−1.63a
1.46−1.63a
1.17−1.24
1.75−1.82
−
1.10−1.18
1.19−1.27
1.95−2.00
4.43−4.50
1.72−1.80
0.77−0.85
0.99−1.10
2.14−2.20
0.96−1.02
−
d
dt
o
dt
o
2
30.7
[CH2]
30.8
[CH2]
o
o
3
4
79.4
39.6
[CH]
o
79.4
39.5
[CH]
o
[CH2]
t
[CH2]
t
dd
dd
5
6
7
141.9
122.6
33.6α
[C]
141.9
122.6
33.6α
[C]
[CH]
[CH2]
d
[CH]
[CH2]
d
o
o
o
o
8
32.8
51.7
38.0
22.0
[CH]
[CH]
[C]
m
o
32.8
51.7
38.0
22.0
[CH]
[CH]
[C]
m
o
9
10
11
[CH2]
m
m
o
[CH2]
m
m
o
12
40.9
[CH2]
40.9
[CH2]
o
o
13
14
15
41.6
57.7
33.2α
[C]
41.5
57.7
33.2α
[C]
[CH]
[CH2]
m
o
[CH]
[CH2]
m
o
o
o
16
17
18
19
20
21
22
23
82.1
63.3
[CH]
[CH]
[CH3]
[CH3]
[CH]
[CH3]
[C]
ddd
o
82.1
63.3
[CH]
[CH]
[CH3]
[CH3]
[CH]
[CH3]
[C]
ddd
o
16.6
s
16.6
s
19.8
s
19.8
s
39.4
m
d
39.4
m
d
15.1
15.1
121.8
33.2α
121.7
33.2α
[CH2]
1.19−1.27
1.94−2.00
1.64−1.70
2.01−2.09
−
3.45−3.50
3.82−3.88
1.18−1.27
o
o
o
o
[CH2]
1.20−1.28
1.94−2.01
1.65−1.69
2.02−2.09
−
3.45−3.50
3.82−3.88
1.18−1.27
o
o
o
o
24
33.7α
[CH2]
33.7α
[CH2]
25
26
85.2
77.6
[C]
85.2
77.6
[C]
[CH2]
d
d
[CH2]
d
d
s
27
24.2
[CH3]
s
24.2
[CH3]
a
b
Arbitrary numbering according to structures 3 and 4 in Figure 1. Assignments were based on HSQC, HMBC, J-res and COSY experiments. d:
duplet, m: multiplet signals; o: overlapped with other signals; s: singlet; t: triplet.
(12 mmol; 57% yield) as the intermediate for the synthesis of
avenanthramides 8−11. The same protocol was applied to 5-
hydroxyanthranilic (21 mmol) and Meldrum’s acid (21 mmol) to
afford 5-hydroxy-2-(2-carboxyacetyl)aminobenzoic acid (11 mmol;
52% yield) for further synthesis of avenanthramides 5−7 and 12,
respectively.
(effluent A) and acetonitrile (effluent B) at a flow rate of 19 mL/min
and isocratic conditions of 45, 50, and 65% of effluent B, respectively.
Collected fractions containing the target compounds were combined
and separated from solvent under vacuum. After lyophilization, the
avenanthramides 5−12 were obtained as white, amorphous powder in
1
purity of >98% (HPLC-DAD, H NMR) and their structures verified
Pyridine (3 mL) and catalytic amounts of β-alanine were added to
an equimolar mixture of 2-(2-carboxyacetyl)aminobenzoic acid (2.5
mmol) and either p-hydroxybenzaldehyde, 3,4-dihydroxybenzalde-
hyde, vanillin, or syringaldehyde, 2.5 mmol each) to give
avenanthramides 8−11. For synthesis of 5−7 and 12, pyridine (3
mL) and β-alanine were added to 5-hydroxy-2-(2-carboxyacetyl)-
aminobenzoic acid (2.5 mmol) and syringaldehyde (2.5 mmol),
respectively. After heating for 110 min at 115 °C under reflux, the
mixture was cooled with ice and acidified with concentrated HCl (3
mL). The precipitated avenanthramides (5−12) were filtered, washed
with water, dried in a laboratory oven at 100 °C, and purified by means
of preparative HPLC on a 250 × 21.0 mm, 5 μm, Phenyl-Hexyl Luna
column (Phenomenex). Monitoring the effluent at 280 and 340 nm,
chromatography was performed using 0.1% formic acid in water
by means of LC-MS and NMR spectroscopy. UV−vis, LC-MS, and
1D/2D-NMR data of N-(3′,4′-dihydroxy-(E)-cinnamoyl)-5-hydrox-
yanthranilic acid (avenanthramide 2c, 5), N-(4′-hydroxy-(E)-cinna-
moyl)-5-hydroxyanthranilic acid (avenanthramide 2p, 6), and N-(4′-
hydroxy-3′-methoxy-(E)-cinnamoyl)-5-hydroxyanthranilic acid (ave-
nanthramide 2f, 7) were identical to those recorded for the
compounds isolated above from oats.
N-(4′-Hydroxy-(E)-cinnamoyl)anthranilic acid (avenanthramide
1p), 8 (Figure 1): UV−vis (ACN/0.1% HCOOH, 70/30 v/v) λmax
=
236 nm, 332 nm; LC-MS (ESI−) m/z 282.0 ([C16H13NO4 − H]−),
238.0 ([C15H13NO2 − H]−), 161.8 ([C9H9NO2 − H]−); MS/MS (DP
= −85 V) m/z (%) 282.0 (100), 238.0 (43), 210.0 (15), 161.8 (8),
144.0 (22), 92.0 (27); LC-TOF-MS m/z 282.0804 ([M − H]−,
measured), calculated for [C16H13NO4 − H]− m/z 282.0766; 1H
F
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Table 5. 1H NMR (500 MHz, Pyridine-d6) and 13C NMR Signals (125 MHz, Pyridine-d6) of the Sugar Moieties of Saponins 3
and 4
ab
,
ab
,
3
4
position
δC
HSQC
δH
M [J, Hz]
δC
HSQC
δH
M [J, Hz]
β-D-GlcI (at C-26)
1′
2′
3′
4′
5′
6′
105.0
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.26−4.32
3.15−3.29
3.30−3.39
3.23−3.32
3.34−3.41
3.63−3.70
3.82−3.91
d [J = 7.7]
105.0
75.3
77.8
71.7
76.2
62.8
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.27−4.33
3.17−3.27
3.33−3.39
3.25−3.32
3.36−3.42
3.63−3.68
3.83−3.90
d [J = 7.7]
75.3
77.8
71.6
76.2
62.8
dd
dd
dd
o
dd
dd
dd
o
o
o
o
o
β-D-GlcII (at C-3)
1″
2″
3″
4″
5″
6″
100.4
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.48−4.53
3.38−3.44
3.62−3.68
3.52−3.57
3.36−3.41
3.63−3.70
3.82−3.91
d [J = 7.7]
100.5
78.6
77.9
81.0
76.2
61.9
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.48−4.54
3.39−3.45
3.62−3.68
3.52−3.57
3.36−3.41
3.63−3.68
3.80−3.90
d [J = 7.6]
78.6
77.9
81.1
76.2
61.9
t
t
o
t
o
t
o
o
o
o
o
o
β-D-GlcIII (at C-4GlcII)
1‴
2‴
3‴
4‴
5‴
6‴
104.2
74.4
87.8
69.9
77.8
62.4
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.43−4.48
3.33−3.46
3.52−3.59
3.33−3.43
3.38−3.42
3.78−3.90
3.78−3.90
d [J = 7.9]
104.7
75.1
78.0
71.4
78.0
62.4
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.36−4.42
3.15−3.26
3.33−3.40
3.25−3.33
3.32−3.39
3.63−3.68
3.80−3.90
d [J = 7.8]
t
dd
t
t
o
o
o
o
o
o
o
o
α-L-Rha (at C-2GlcII)
1⁗
2⁗
3⁗
4⁗
5⁗
6⁗
102.1
72.2
72.4
73.9
69.7
17.9
[CH]
[CH]
[CH]
[CH]
[CH]
[CH3]
5.20−5.26
3.85−3.90
3.63−3.68
3.34−3.43
4.09−4.15
1.19−1.26
bs [J = 2.5]
102.0
72.2
72.4
73.9
69.7
17.9
[CH]
[CH]
[CH]
[CH]
[CH]
[CH3]
5.22−5.26
3.87−3.91
3.62−3.68
3.39−3.45
4.10−4.15
1.18−1.27
bs [J = 2.5]
dd
dd
o
dd
dd
o
m
d
m
d
β-D-GlcIV (at C-3GlcIII)
1⁗′
2⁗′
3⁗′
4⁗′
5⁗′
6⁗′
105.3
75.5
76.2
71.7
77.8
62.6
[CH]
[CH]
[CH]
[CH]
[CH]
[CH2]
4.51−4.58
3.23−3.29
3.37−3.42
3.23−3.32
3.33−3.39
3.59−3.70
3.78−3.91
d [J = 7.7]
t
o
o
o
o
o
a
b
Arbitrary numbering according to structures 3 and 4 in Figure 1. Assignments were based on HSQC, HMBC, J-res, and COSY experiments. d,
doublet; m, multiplet signals; o, overlapped with other signals; bs, broad singlet; t, triplet.
NMR (400 MHz, MeOD-d4) δ 6.54 [d, 1H, J = 15,6 Hz, H−C(8′)],
6.83 [d, 2H, J = 8.6 Hz, H−C(3′/5′)], 7.15 [m, 1H, H−C(5)], 7.50
[d, 2H, J = 8.6 Hz, H−C(2′/6′)], 7.57 [m, 1H, H−C(4)], 7.61 [d, 1H,
J = 15.6 Hz, H−C(7′)], 8.11 [dd, 1H, J = 8.0, 1.6 Hz, H−C(6)], 8.69
[dd, 1H, J = 8.5, 0.7 Hz, H−C(3)]; 13C NMR (100 MHz, MeOD-d4)
δ 116.8 [C(3′/5′)], 117.4 [C(1)], 119.2 [C(8′)], 121.5 [C(3)], 123.9
[C(5)], 127.4 [C(1′)], 131.0 [C(2′/6′)], 132.6 [C(6)], 135.2 [C(4)],
142.4 [C(2)], 143.7 [C(7′)], 161.0 [C(4′)], 167.2 [C(9′)], 171.5
[C(7)].
N-(3′,4′-Dihydroxy-(E)-cinnamoyl)anthranilic acid (avenanthra-
mide 1c), 9 (Figure 1): UV−vis (ACN/0.1% HCOOH, 70:30 v/v)
λmax = 244 nm, 340 nm; LC-MS (ESI−) m/z 298.0 ([C16H13NO5 −
H]−), 161.6 ([C8H5NO3 − H]−), 134.8 ([C8H8O2 − H]−); MS/MS
(DP = −65 V) m/z (%) 298.0 (34), 161.8 (54), 134.8 (100), 117.8
(22), 91.8 (36); LC-TOF-MS m/z 298.0733 ([M − H]−, measured),
calculated for [C16H13NO5 − H]− m/z 298.0715; 1H NMR (400
MHz, MeOD-d4) δ 6.47 [d, 1H, J = 15.6 Hz, H−C(8′)], 6.80 [d, 1H, J
= 8.2 Hz, H−C(5′)], 6.99 [dd, 1H, J = 2.0, 8.3 Hz, H−C(6′)], 7.08 [d,
1H, J = 2.0 Hz, H−C(2′)], 7.14 [m, 1H, H−C(5)], 7.52 [d, 1H, J =
15.6 Hz, H−C(7′)], 7.56 [m, 1H, H−C(4)], 8.10 [dd, 1H, J = 8.0, 1.5
Hz, H−C(6)], 8.68 [dd, 1H, J = 8.5, 0.7 Hz, H−C(3)]; 13C NMR
[100 MHz, MeOD-d4) δ 115.2 [C(2′)], 116.6 [C(5′)], 117.5 [C(1)],
119.2 [C(8′)], 121.6 [C(3)], 122.7 [C(6′)], 123.9 [C(5)], 127.9
[C(1′)], 132.6 [C(6)], 135.2 [C(4)], 142.8 [C(2)], 144.0 [C(7′)],
146.8 [C(3′)], 149.3 [C(4′)], 167.2 [C(9′)], 171.5 [C(7)].
N-(4′-Hydroxy-3′-methoxy-(E)-cinnamoyl)anthranilic acid (ave-
nanthramide 1f), 10 (Figure 1): UV−vis (ACN/0.1% HCOOH, 70/
30 v/v) λmax = 240 nm, 340 nm; LC-MS (ESI−) m/z 312.0
([C17H15NO5 − H]−), 267.8 ([C16H15NO3 − H]−), 251.8
([C16H15NO2 − H]−); MS/MS (DP = −90 V) m/z (%) 312.0
(100), 267.8 (41), 251.8 (68), 175.0 (10), 133.0 (43), 91.8 (20); LC-
TOF-MS m/z 312.0868 ([M − H]−, measured), calculated for
[C17H15NO5 − H]− m/z 312.0872; 1H NMR (400 MHz, MeOD-d4) δ
3.93 [s, 3H, H−C(10′)], 6.59 [d, 1H, J = 15.6 Hz, H−C(8′)], 6.83 [d,
1H, J = 8.1 Hz, H−C(5′)], 7.11 [dd, 1H, J = 8.1, 1.9 Hz, H−C(6′)],
7.15 [m, 1H, H−C(5)], 7.25 [d, 1H, J = 1.9 Hz, H−C(2′)], 7.57 [m,
G
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
1H, H−C(4)], 7.61 [d, 1H, J = 15.6 Hz, H−C(7′)], 8.12 [dd, 1H, J =
8.0, 1.5 Hz, H−C(6)], 8.69 [dd, 1H, J = 8.5, 0.7 Hz, H−C(3)]; 13C
NMR [100 MHz, MeOD-d4) δ 56.5 [C(10′)], 111.6 [C(2′)], 116.5
[C(5′)], 117.7 [C(1)], 119.5 [C(8′)], 121.6 [C(3)], 123.9/124.0
[C(5/6′), may be interchangeable], 127.9 [C(1′)], 132.6 [C(6)],
135.2 [C(4)], 142.7 [C(2)], 144.0 [C(7′)], 149.4 [C(3′)], 150.4
[C(4′)], 167.2 [C(9′)], 171.6 [C(7)].
linked to the mass spectrometer operated in the multiple reaction
monitoring mode (MRM) in negative electrospray ionization. Using
acetonitrile containing 0.1% formic acid as solvent A and 0.1% formic
acid in water as solvent B, chromatography was performed using the
following gradient at a flow rate of 0.3 mL/min: 0 min, 40% A (1 min
isocratically); in 1 min to 70% A (2 min isocratically); in 2 min to
100% A (3 min isocratically); within 2 min back to 40% A (3 min
isocratically). Nitrogen served as curtain gas (25 psi), nebulizer gas (55
psi), and turbo gas (45 °C). Fragmentation of the pseudo molecular
ions [M − H]− into specific product ions was induced by collision
with nitrogen (4.5 × 10−5 Torr). The following mass transitions were
recorded for the taste compounds using the declustering potential
(DP), collision energy (CE), and collision cell exit potential (CXP)
given in parentheses: 5, m/z 314.2 → 178.0 (DP/CE/CXP −65/−16/
−11); 6, m/z 298.1 → 253.9 (DP/CE/CXP − 50/−22/−15); 7, m/z
328.0 → 283.8 (DP/CE/CXP −95/−24/−19); 8, m/z 282.0 → 238.0
(DP/CE/CXP − 85/−24/−7); 9, m/z 298.0 → 161.7 (DP/CE/CXP
−65/−16/−11); 10, m/z 312.0 → 251.7 (DP/CE/CXP −90/−34/−
17); 11, m/z 342.0 → 298.1 (DP/CE/CXP − 85/−24/−9); 12, m/z
358.1 → 314.1 (DP/CE/CXP −80/−24/−9).
N-(4′-Hydroxy-3′,5′-dimethoxy-(E)-cinnamoyl)-anthranilic acid
(avenanthramide 1s), 11 (Figure 1): UV−vis (ACN/0.1%
HCOOH, 70:30 v/v) λmax = 244 nm, 344 nm; LC-MS (ESI−) m/z
341.8 ([C18H17NO6 − H]−), 297.8 ([C17H17NO4 − H]−), 281.8
([C16H13NO4 − H]−), 266.8 ([C16H13NO3 − H]−); MS/MS (DP =
−85 V) m/z (%) 341.8 (100), 297.8 (35), 281.8 (29), 266.8 (32),
143.8 (27), 120.8 (45); LC-TOF-MS m/z 342.1025 ([M − H]−,
measured), calculated for [C18H17NO6 − H]− m/z 342.0978; 1H
NMR (400 MHz, MeOD-d4) δ 3.91 [s, 6H, H−C(10′/11′)], 6.61 [d,
1H, J = 15.6 Hz, H−C(8′)], 6.95 [s, 2H, H−C(2′/6′)], 7.15 [m, 1H,
H−C(5)], 7.57 [m, 1H, H−C(4)], 7.60 [d, 1H, J = 15.6 Hz, H−
C(7′)], 8.11 [dd, 1H, J = 8.0, 1.5 Hz, H−C(6)], 8.69 [d, 1H, J = 8.2
Hz, H−C(3)]; 13C NMR [100 MHz, MeOD-d4) δ 56.9 [C(10′/11′)],
106.8 [C(2′/6′)], 117.6 [C(1)], 119.9 [C(8′)], 121.6 [C(3)], 123.9
[C(5)], 126.9 [C(1′)], 132.6 [C(6)], 135.2 [C(4)], 139.3 [C(4′)],
142.7 [C(2)], 144.2 [C(7′)], 149.5 [C(3′/5′)], 167.1 [C(9′)], 171.6
[C(7)].
Nuclear Magnetic Resonance Spectroscopy (NMR). One- and
1
two-dimensional H and 13C NMR spectra were acquired on a 500
MHz Bruker Avance III, equipped with a triple-resolution cryo probe
(TCI) (Bruker, Rheinstetten, Germany). Samples were dissolved in
methanol-d4 containing 0.03% trimethylsilane (TMS) or pyridine-d5.
The chemical shifts are referenced to the TMS or the solvent signal
N-(4′-Hydroxy-3′,5′-dimethoxy-(E)-cinnamoyl)-5-hydroxyanthra-
nilic acid (avenanthramide 2s), 12 (Figure 1): UV−vis (ACN) λmax
=
236 nm, 348 nm; LC-MS (ESI−) m/z 358.0 ([C18H17NO7 − H]−),
314.0 ([C17H17NO5 − H]−), 298.0 ([C17H17NO4 − H]−), 133.8
([C7H5NO2 − H]−); MS/MS (DP = −80 V) m/z (%) 358.0 (100),
314.0 (42), 298.0 (15), 190.8 (20), 159.8 (36), 133.8 (24); LC-TOF-
MS m/z 358.0926 ([M − H]−, measured), calculated for [C18H17NO7
− H]− m/z 358.0927; 1H NMR (400 MHz, MeOD-d4) δ 3.90 [s, 6H,
H−C(10′/11′)], 6.59 [d, 1H, J = 15.6 Hz, H−C(8′)], 6.93 [s, 2H, H−
C(2′/6′)], 7.00 [dd, 1H, J = 2.9, 9.0 Hz, H−C(4)], 7.52 [d, 1H, J = 2.9
Hz, H−C(6)], 7.54 [d, 1H, J = 15.5 Hz, H−C(7′)], 8.46 [d, 1H, J =
8.9 Hz, H−C(3)]; 13C NMR (101 MHz, MeOD-d4) δ 56.8 [C(10′/
11′)], 106.7 [C(2′/6′)], 118.3 [C(6)], 119.4 [C(8′)], 121.7 [C(4)],
123.4 [C(3)], 127.9 [C(1′)], 134.7 [C(2)], 139.2 [C(4′)], 143.4
[C(7′)], 149.5 [C(3′/5′)], 154.2 [C(5)], 166.6 [C(9′)], 171.8 [C(7)].
High-Performance Liquid Chromatography (HPLC). Prepara-
tive analyses of fractions II-C and II-D was done on a HPLC apparatus
(Jasco, Groß-Umstadt, Germany) consisting of two PU-2087 Plus
pumps and an MD 2010 Plus photodiode array detector as well as a
Sedex LT-ELSD detector model 85 (Sedere, Alfortville, France) and
an Rh 7725i type Rheodyne injection valve (Rheodyne, Bensheim,
Germany). The split ratio was set to 1 mL/min for the ELSD detector.
Data acquisition was executed by means of Chrompass 1.9. (Jasco).
UPLC/Time-of-Flight Mass Spectrometry (UPLC/TOF-MS).
High-resolution mass spectra were measured on a Waters Synapt
G2-S HDMS time-of-flight mass spectrometer (Waters, Manchester,
UK) coupled to an Acquity UPLC Core system (Waters, Milford, MA,
USA). Data acquisition and interpretation were performed by using
MassLynx software v4.1 SCN 851 (Waters, Milford, USA) and the
tool “Elemental Composition”.
1
(pyridine-d5: H 7.22 ppm; 13C 123.87 ppm). TOPSPIN version 2.1
1
(Bruker) was used for data processing. H, 13C, 135DEPT, COSY, J-
RESOLVED, ROESY, HSQC, and HMBC spectroscopies were
recorded using standard pulse sequences of the Bruker library.
Interpretation of the obtained spectra was performed with
MestReNova 8.1.0-11315. (Mestrelab Research, Santiago de Compos-
tela, Spain).
RESULTS AND DISCUSSION
■
A freshly prepared suspension of oat flour, imparting a typical
astringent and bitter taste, was evaluated by means of a taste
profile analysis. Therefore, a trained sensory panel was asked to
rate the intensities of the taste modalities bitter, sweet, sour,
salty, umami, and astringent on a linear 5-point intensity scale.
Bitter and astringent notes were rated with the highest taste
intensities of 2.0 and 1.1, respectively (Table 1). A
comparatively low intensity was reported for sweetness (0.9),
whereas sour (0.3), umami (0.2), and salty taste (0.1) were
almost not detectable. To gain a first insight into the
hydrophobicity of the compounds imparting the typical bitter
and astringent orosensation, oat flour was extracted with
solvents of different polarities.
Solvent Fractionation of Oat Flour. Oat flour was
extracted sequentially with n-hexane, methanol/water (70:30,
v/v), and methanol to give the hexane solubles (fraction I), the
methanol/water extractables (fraction II), the methanol
extractables (fraction III), and the nonsoluble residue (fraction
IV) after solvent separation under vacuum and lyophilization.
Taste profile analysis of the individual fractions, each dissolved
in water in its “natural” concentration, revealed no taste activity
for fraction IV, thus indicating an exhaustive solvent extraction
of taste compounds from oat flour (Table 1). Highest sensory
scores were reported for bitterness (2.6) and astringency (1.2)
in fraction II, whereas fractions I and III exhibited lower bitter
intensities; for example, fractions I and III showed astringency
scores of 1.0 and 0.7, respectively (Table 1). Aimed at
identifying the key molecules inducing the bitter and astringent
taste in oat flour, the most taste-active fraction II was subjected
to a sensory-guided fractionation.
High-Performance Liquid Chromatography/Tandem Mass
Spectrometry (LC-MS/MS). For structure elucidation, mass and
product ion spectra were acquired on an API 4000 Q Trap triple-
quadrupole/linear ion trap mass spectrometer (AB Sciex, Darmstadt,
Germany). The isolated fractions were dissolved in a mixture of
acetonitrile/water (50:50, v/v) and directly introduced into the mass
spectrometer by flow infusion using a syringe pump. For electrospray
ionization, the ion spray voltage was set at −4500 V in the negative
mode and at 5500 V in the positive mode. Both quadrupoles operated
at unit mass resolution, and nitrogen served as a curtain gas (25 psi)
and as a turbo gas (425 °C). Fragmentation of the pseudo molecular
ions [M + H]+ or [M − H]− into specific product ions was induced by
collision with nitrogen (4.5 × 10−5 Torr). Data acquisition and
instrumental control were performed with Analyst 1.5.1 (AB Sciex).
HPLC-MS/MS analysis of taste compounds was performed on a
150 × 2.0 mm, 5 μm, Luna C18 PhenylHexyl column (Phenomenex)
H
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 2. Taste profile analysis of SPE fractions II-A−II-E isolated from oat flour. Sensory data are given as the mean of triplicates; error bars indicate
the 95% confidence interval of the arithmetical mean.
Figure 3. RP-HPLC chromatograms of SPE-fractions II-D (A) and II-C (B) and taste dilution (TD) factors of subfractions II-D-1−II-D-16 and II-C-
1−II-C-17.
Sensory-Guided Separation of Oat Fraction II. To sort
out the bitter and astringent nonvolatiles from the bulk of
tasteless or less active compounds present in fraction II, the
methanol/water solubles isolated from oat flour were separated
by means of solid phase extraction on an RP-18 material using
water, methanol/water mixtures, and methanol as eluents to
afford the five subfractions II-A, II-B, II-C, II-D, and II-E. After
solvent evaporation under vacuum, the individual subfractions
were taken up in water in their “natural” concentration ratios
and, then, evaluated in their taste profile (Figure 2). The bitter
and astringent sensation of fraction II-D was rated with the
highest scores of 2.7 and 1.8, respectively, whereas bitterness
and astringency of all other fractions were evaluated with lower
intensities ranging between 0.7 and 1.3.
means of preparative RP-HPLC/ELSD to give 16 subfractions,
namely, II-D-1−II-D-16 (Figure 3A), which were collected
separately, freed from solvent under vacuum, and used for a
TDA. To achieve this, the subfractions were taken up in water
in their “natural” concentrations, stepwise 1+1 diluted with
water and, then, presented in order of ascending concentrations
to a trained sensory panel who was asked to determine the TD
factors by means of the half-tongue test.22−24 Fractions II-D-4−
II-D-6, II-D-8−II-D-10, and II-D-14 were evaluated with the
highest TD factors for bitterness as well as astringency (Figure
3A). As subfractions II-D-6 and II-D-14 were too unstable to
enable an unequivocal identification and subfraction II-D-10
was found to contain the same key compounds as subfraction
II-D-9 by carry-over, the following experiments focused on the
structure determination of the key taste compounds in
subfractions II-D-4, II-D-5, II-D-8, and II-D-9, respectively.
Identification of Taste Compounds in Fraction II-D. To
identify the phytochemicals imparting the bitter taste of the
most taste-active fraction II-D, this fraction was separated by
I
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 4. Excerpt of HMBC spectra of the purified saponins 4 (A) and 2 (B).
Preparative RP-HPLC of subfractions II-D-4 and II-D-5,
followed by rechromatography, revealed two bitter compounds,
1 and 2 (Figure 1), as amorphous powders. Preliminary
separation by means of TLC and spraying with p-anisaldehyde/
sulfuric acid gave intense green spots suggesting the presence of
saponine structures in compounds 1 and 2.26 LC-TOF-MS
analysis revealed m/z 1225.5867 as pseudomolecular ion ([M
− H]−) for 1, indicating a molecular formula of C57H94O28.
Furthermore, fragment ion peaks with m/z 1063.5 ([M − Glc
− H]−), 901.4 ([M − 2Glc − H]−), and 755.4 ([M − 2Glc −
Rha − H]−) were observed in the negative ESI mode, thus
demonstrating the presence of at least two hexose and one
deoxyhexose moieties in the saponin. For unequivocal
identification of the glycosidically bound carbohydrates,
aliquots of the isolates were hydrolyzed with aqueous
trifluoroacetic acid, followed by high-performance ion chroma-
tography to D-glucose and L-rhamnose in a ratio of 4:1 for
saponin 1 and 3:1 for saponin 2. The identity of the
monosaccharides was confirmed by comparison of their
retention times with those found for the corresponding
reference compounds as well as by cochromatography.
interglycosidic linkage between the monomers were accom-
plished by means of COSY, HMBC, and ROESY experiments.
For example, interglycosidic homonuclear coupling of vicinal
protons could be allocated by means of a COSY experiment,
homonuclear correlation between the anomeric proton and the
protons at C(3Glc) and C(5Glc) was observed in the ROESY
spectrum, and heteronuclear 2J,3J couplings were observed
between C(2GlcII) and H−C(1Rha), between C(4GlcII) and H−
C(1GlcIII), and between C(3GlcIII) and H−C(1GlcIV), respec-
tively.
The germinal protons H−C(26a) and H−C(26b) observed at
3.58−3.66 and 3.92−3.98 ppm matched the data reported for
(25R)-configured saponins (3.63 and 3.96 ppm).14,28 As the
germinal protons H−C(26a) and H−C(26b) of (25S)-
configured furostanol saponins were reported to resonate at
3.46 and 4.07 ppm,29,30 the carbon atom C(25) of the
furostanol saponin 1 was assigned to be (R)-configured. In
addition, the methyl group H−C(27), resonating at 0.98−1.02
ppm, supports the (25R)-configuration of 1. By combined
analysis of COSY, HSQC, HMBC, DEPT, and J-RES data, the
aglycone for saponin 1 could be assigned to proto-diosgenin.31
Taking all MS and 1D/2D-NMR spectroscopic data into
consideration, the structure of the bitter compound isolated
from fraction II-D-4 was identified as 3-(O-α-L-
rhamnopyranosyl(1→2)-[β-D-glucopyranosyl(1→3)-β-D-
glucopyranosyl(1→4)]-β-D-glucopyranosid)-26-O-β-D-gluco-
pyranosyl-(25R)-furost-5-ene-3β,22,26-triol (1, Figure 1).
Although the (25S)-epimer of 1 has been reported as
trigoneoside XIIIa in fenugreek seeds of Trigonella foenum-
graecum L.,32 the (25R)-congener has to the best of our
knowledge not yet been reported in literature.
The 13C NMR spectrum for 1 displayed a total of 57 signals
resonating between 16.7 and 141.1 ppm (Tables 2 and 3); 27
of those signals were assigned to the aglycone moiety, and the
four signals at 37.4 (C(10)), 41.0 (C(13)), 111.0 (C(22)), and
141.1 ppm (C(5)) were assigned as quarternary carbons by
means of an HSQC experiment. The 1H NMR spectrum
revealed the presence of two tertiary methyl groups at 0.90
(H−C(18)) and 1.06 ppm (H−C(19)), three secondary
methyl groups at 1.00 (H−C(27)), 1.35 (H−C(21)), and
1.77 ppm (H−C(6Rha)), and one olefinic proton at 5.30 ppm
(H−C(6)). Moreover, five anomeric carbohydrate protons
were observed and assigned as one α-configured rhamnopyr-
anoysl proton (6.24 ppm) showing a coupling constant of 2.2
Hz and four glucopyranosyl moieties (4.83, 4.95, 5.10, and 5.29
ppm) with coupling constants of 7−8 Hz, thus indicating a β-
configuration. The bidesmosidic saponin structure was
disclosed by means of an HMBC experiment showing long-
range correlation between carbon C(26) resonating at 75.5
ppm and the anomeric proton H−C(1GlcI) observed at 4.83
ppm, as well as between carbon atom C(3) at 78.4 ppm and the
proton H−C(1GlcII). Full assignment of the remaining
carbohydrate atoms and unequivocal identification of the
Mass spectrometric analysis of the bitter compound 2,
purified from fraction II-D-5, revealed m/z 1063.5334 as the
pseudomolecular ion ([M − H]−), thus indicating a molecular
formula of C51H84O23. The MS daughter ions observed with m/
z 901.4 ([M − Glc − H]−), 755.1 ([M − Glc − Rha − H]−)
1
and 593.3 ([M − 2Glc − Rha − H]−) as well as the H NMR
resonance signals at 4.82, 4.96, 5.14, and 6.25 ppm
demonstrated only one rhamnose and three glucose moieties
in saponin 2, but the same aglycone as found for the furostanol
saponin 1 (Tables 2 and 3). The HMBC spectrum revealed the
aglycone glycosylation by the observed correlation between the
proton H−C(1GlcII)) at 4.96 ppm and the carbon atom C(3) at
J
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 5. UV−vis spectra (A) of isolated fractions II-C-11 (a), II-C-15 (b) and II-C-16 (c) and MS/MS spectra of compounds 5 (B), 6 (C), and 7
(D).
78.4 ppm, as well as by the heteronuclear connectivity between
H−C(1GlcII), resonating at 4.82 ppm, and C−26, observed at
75.5 ppm. Comparison of 1D/2D-NMR data of saponin 2 with
those recorded for 1 confirmed the lack of the hexose moiety
bound to C(3GlcIII) in 1. In addition, the resonance signal of
C(3GlcIII) was upfield shifted to 81.4 ppm (2) when compared
to saponin 1 (88.6 ppm). Comparison of spectroscopic data
with those observed for 1 and with those reported for
structurally related derivatives in the literature14,28−30,33
revealed a (25R)-configuration of saponin 2 and allowed its
structure determination as 3-(O-α-L-rhamnopyranosyl(1→2)-
[β-D-glucopyranosyl(1→4)]-β-D-glucopyranosid)-26-O-β-D-
glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol (2, Figure 1).
To the best of our knowledge, this bitter saponin isolated from
A. sativa L. has not yet been described elsewhere.
Purification of the key bitter and astringent phytochemicals
from fractions II-D-8 and II-D-9 revealed the target compounds
3 and 4, respectively (Figure 1). LC-TOF-MS analysis
exhibited m/z 1223.5695 and 1061.5169 as the pseudomo-
lecular ions ([M − H]−) and an elemental composition of
C57H92O28 and C51H82O23, respectively. Additional LC-MS/MS
experiments, performed in the ESI+ mode, revealed m/z 1063.5
([M − Glc + H]+), 901.4 ([M − 2Glc + H]+), 739.4 ([M −
3Glc + H]+), 593.3 ([M − 3Glc − Rha + H]+), and 431.3 ([M
− 4Glc − Rha + H]+) as main daughter ions of compound 3.
Whereas compound 4 exhibited exactly the same daughter ions
at m/z 901.4, 739.4, 593.3, and 431.3, the lower mass by 162
Da indicated one hexose moiety less than found for compound
3. These findings were well in line with the nuatigenin-type
steroidal saponins avenacosides B and A, which had been
previously isolated from oat.10,34−36
after acidic hydrolysis revealed glucose and rhamnose moieties
as part of the sugar units. 1H and 13C NMR data (Tables 4 and
5) were well in line with those published previously.34 Whereas
Pecio et al.34 referenced the signals of C(12) and C(13) to be
41.6 and 40.3 ppm, we assigned carbon C(12) and C(13) as
shifted to 40.9 and 41.6 ppm, respectively. This alignment is
endorsed by previous studies showing that carbon C(12) in
steroidal saponins is high-field shifted.14,32,37,38 With all
spectroscopic data taken into consideration, the chemical
structure of the bitter and astringent molecules in fractions II-
D-8 and II-D-9 were determined to be nuatigenin-3-O-(α-L-
rhamnopyranosyl-(1→2)-[β-D-glucopyranosyl-(1→3)-β-D-glu-
copyranosyl-(1→4)]-β-D-glucopyranoside)-26-O-β-D-glucopyr-
anoside (3), and nuatigenin-3-O-(α-L-rhamnopyranosyl-(1→
2)-[β-D-glucopyranosyl-(1→4)]-β-D-glucopyranoside)-26-O-β-
D-glucopyranoside (4, Figure 1). Although both saponins have
been earlier reported in the seeds, grains, and leaves of A. sativa
L.,10,34,36,39,40 avenacosides B (3) and A (4) have not yet been
demonstrated to contribute to the bitter taste of oats.
Identification of Taste Compounds in Fraction II-C. As
HPLC analysis of the bitter and astringent fractions II-B and II-
C showed more quantitative rather than qualitative differences
in composition, the most intense taste compounds were located
in fraction II-C by means of RP-HPLC-TDA (Figure 3B).
Among the 17 subfractions collected, by far the highest TD
factor of 32 was found for bitter fraction II-C-11, followed by
fractions II-C-5, II-C-16, and II-C-5, exhibiting a bitter
impression with TD factors of 16 and 8, respectively. Next to
their bitter taste, these fractions exhibited a puckering
astringent orosensation evaluated with TD factors between 8
and 16. The following experiments were focused on the
identification of the key taste compounds in fractions II-C-5, II-
C-11, II-C-15, and II-C-16.
The bitter and astringent compounds detected in sub-
fractions II-C-11, II-C-15 and II-C-16 gave similar UV−vis
absorbance spectra exhibiting maxima between 310 and 360 nm
(Figure 5). LC-TOF-MS analysis showed pseudomolecular ions
([M − H]−) with m/z 314.0669, 298.0716, and 328.0851 for
the target taste compounds 5−7 in fractions II-C-11, II-C-15,
and II-C-16, thus indicating the incorporation of one nitrogen
atom in each molecule. Furthermore, MS/MS analysis
demonstrated losses of 136, 120, and 150 amu, respectively,
to result in a characteristic fragment ion with m/z 178 for all
three molecules 5−7 (Figure 5), matching well the phenyl-
isocyanate fragment ion reported to be formed upon MS
analysis from the 5-hydroxyanthranilic acid moiety of
avenanthramides.41 To further confirm the structure of the
1H NMR data indicated the identical aglycone for both
saponins 3 and 4. In contrast to the furostanol-type saponin
skeletons of 1 and 2, analysis of the HMBC signals recorded for
3 and 4 demonstrated a spirostanol-type aglycone as reported
earlier.34 Due to the spiro ring in compounds 3 and 4, carbon
C(22) observed at 121.7 ppm and C(25) resonating at 85.2
ppm were shifted to lower fields. Long-range correlation signals
for C(22) in compounds 3 and 4 (Figure 4A) and, in
comparison, in compounds 1 and 2 (Figure 4B) are displayed
in the excerpts of the HMBC experiments. In addition, the
bidesmosidic structure of saponins 3 and 4 was confirmed by
the observed heteronuclear connectivity between glycosidic
proton H−C(1GlcI) at 4.30 ppm and carbon atom C(26) at 24.2
ppm, as well as the correlation between H−C(1GlcII) at 4.51
ppm and carbon atom C(3) at 79.4 ppm. High-performance
ion chromatographic analysis of monosaccharides obtained
K
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Figure 6. Reaction sequence used for synthesis of avenanthramides 5−12.
suggested 5-hydroxyanthranilic acid structures, 1D- and 2D-
NMR experiments were performed.
1
The H NMR spectra of each compound revealed three
proton signals showing ortho and meta couplings detected in
the aromatic region at 6.9 and 8.4 ppm, which were attributed
to H−C(4) and H−C(3) in a COSY experiment. Furthermore,
the olefinic protons H−C(8′) and H−C(7′), resonating at 6.5
and 7.5 ppm, demonstrated a coupling constant of 15.7 Hz,
indicating the characteristic (E)-configuration for cinnamic acid
derivatives. The 1H NMR of compound 7 revealed an
additional singlet at 3.9 ppm with an intensity of three protons,
thus indicating a methoxy group (C−10′) in the chemical
structure. With these data taken into consideration, the
structures of these bitter and astringent compounds were
assumed to be N-(3′,4′-dihydroxy-(E)-cinnamoyl)-5-hydrox-
yanthranilic acid (5), N-(4′-hydroxy-(E)-cinnamoyl)-5-hydrox-
yanthranilic acid (6), and N-(4′-hydroxy-3′-methoxy-(E)-
cinnamoyl)-5-hydroxyanthranilic acid (7) as shown in Figure
1. Coined avenanthramides 2c, 2p, and 2f, these compounds
have been previously reported as phytochemicals in oats.42,43
To confirm the proposed structures and to study the
presence of other avenanthramides reported in oats,42,44
reference compounds of compounds 5−7 as well as additional
avenathramides (8−12) were chemically synthesized (Figure
6). Following a literature protocol with some modifications,27
malonylation of anthranilic acid or 5-hydroxyanthranilic acid
with Meldrum’s acid, followed by a condensation reaction with
benzaldehyde derivatives, gave the crude targeted amides,
which were purified by RP-18 chromatography to afford the
individual avenanthramides 2c (5), 2p (6), 2f (7), 1p (8), 1c
(9), 1f (10), 1s (11), and 2s (12) in a high purity of >98%.
Consequently, the identities of 5−12 were verified by means of
UV−vis, LC-MS/MS, and NMR experiments.
Figure 7. HPLC-MS/MS (ESI−) analysis of phytochemicals 5−12.
Oral Threshold Concentrations of Oat Saponins and
Avenanthramides. Prior to sensory evaluation, the purity of
the taste compounds was confirmed by means of HPLC-ELSD
and LC-MS. To determine the recognition threshold
concentrations of the bitter taste and astringent orosensation
of the saponins (1−4) and avenanthramides (5−12), aqueous
solutions (pH 6.4) of the target compounds were evaluated by
a trained sensory panel using the half-tongue test.22,23
The orosensory threshold concentrations of the avenan-
thramides 5−12 were strongly dependent on their chemical
structure and ranged from 38 to 135 μmol/kg for astringency
and from 60 to 170 μmol/kg for bitterness (Table 6). The
lowest bitter taste threshold of 60 μmol/kg was found for
avenanthramide 2p (6), showing a para-hydroxylated cinnamic
acid moiety. Additional hydroxylation (5) or methoxylation (7)
in meta position of the cinnamic acid resulted in somewhat
higher orosensory threshold concentrations; for example,
To validate the presence of the avenanthramides in oat flour,
an oat extract was analyzed by LC-MS/MS-MRM using
characteristic mass transitions of the individual target
compounds 5−12. With the exception of 11 and 12 (data
not shown), all other avenanthramides (5−10) have been
detected in the oat extract and unequivocally confirmed by
comparing retention times with the reference compounds as
well as by cochromatography (Figure 7).
L
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
Table 6. Human Recognition Threshold Concentrations of
Taste Compounds 1−12
ACKNOWLEDGMENTS
■
We are thankful to Daniela Gunzkofer for her excellent
technical assistance. We are grateful to General Mills Inc. for
supporting this research project.
̈
a
taste threshold
(μmol/kg)
b
taste compound
astringency
bitterness
3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-
glucopyranosyl(1→3)-β-D-glucopyranosyl(1→4)]-β-
D-glucopyranosid)-26-O-β-D-glucopyranosyl-(25R)-
furost-5-ene-3β,22,26-triol (1)
3
2
1
4
9
2
REFERENCES
■
(1) Butt, M.; Tahir-Nadeem, M.; Khan, M.; Shabir, R.; Butt, M. Oat:
unique among the cereals. Eur. J. Nutr. 2008, 47, 68−79.
(2) Zhang, M.; Liang, Y.; Pei, Y.; Gao, W.; Zhang, Z. Effect of process
on physicochemical properties of oat bran soluble dietary fiber. J. Food
Sci. 2009, 74, C628.
3-(O-α-L-rhamnopyranosyl(1→2)-[β-D-
glucopyranosyl(1→4)]-β-D-glucopyranosid)-26-O-β-
D-glucopyranosyl-(25R)-furost-5-ene-3β,22,26-triol
(2)
3
21
(3) Schuh, C.; Schieberle, P. Characterization of (E,E,Z)-2,4,6-
nonatrienal as a character impact aroma compound of oat flakes. J.
Agric. Food Chem. 2005, 53, 8699−8705.
(4) Biermann, U.; Grosch, W. Bitter-tasting monoglycerides from
stored oat flour. Z. Lebensm.-Unters. Forsch. 1979, 169, 22−26.
(5) Biermann, U.; Wittmann, A.; Grosch, W. Occurrence of bitter
hydroxy fatty acids in oat and wheat (in German). Fette, Seifen,
Anstrichm. 1980, 82, 236−240.
(6) Molteberg, E. L.; Magnus, E. M.; Bjorge, J. M.; Nilsson, A.
Sensory and chemical studies of lipid oxidation in raw and heat-treated
oat flours. Cereal Chem. 1996, 73, 579−587.
(7) Baur, C.; Grosch, W.; Wieser, H.; Jugel, H. Enzymatic oxydation
of linoleic acid: formation of bitter tasting fatty acids. Z. Lebensm.-
Unters. Forsch. 1977, 164, 171−176.
avenacoside B (3)
4
3
12
2
6
7
19
20
avenacoside A (4)
avenanthramide 2c (5)
avenanthramide 2p (6)
avenanthramide 2f (7)
avenanthramide 1p (8)
avenanthramide 1c (9)
avenanthramide 1f (10)
avenanthramide 1s (11)
avenanthramide 2s (12)
55 31
45 17
135 37
70 48
78 38
60 15
170 55
96 48
56
2
130
2
98 40
82 21
38 23
113 25
101 21
81 65
a
Taste threshold concentrations were determined in aqueous solution
(pH 6.4) containing 1 (saponins 1−4) or 2% ethanol (avenan-
thramides 5−12). Structures of compounds are given in Figure 1.
b
(8) Lehtinen, P.; Laakso, S. Agric. Food Sci. 2004, 13, 88−99.
(9) Rohrlich, M.; Train, G. Isolation of a bitter tasting glycoside from
oat (in German). Z. Lebensm.-Unters. Forsch. 1954, 99, 346−351.
(10) Tschesche, R.; Tauscher, M.; Fehlhaber, H.-W.; Wulff, G.
Steroid saponins with more than one sugar chain, IV. Avenacosid A, a
bisdesmosidic steroid saponin from Avena sativa (in German). Chem.
Ber. 1969, 102, 2072−2082.
avenanthramide 2f (7), exhibiting a ferulic acid moiety, showed
bitter and astringency thresholds of 170 and 135 μmol/kg,
respectively. With the exception of avenanthramide 2f, the 5-
hydroxyanthranilic acid amides, such as 2c (5) and 2p (6), with
thresholds of 60 and 78 μmol/kg were found with higher taste
activities when compared to anthranilic acid amides such as
avenanthramides 1p (8), 1c (9), and 1f (10), showing bitter
taste threshold concentrations of 96, 130, and 113 μmol/kg,
respectively.
Sensory evaluation of the saponins (1−4) revealed
astringency and bitter taste at much lower concentrations
when compared to the avenanthramides; for example, bitter
taste thresholds were found to range between 4 and 9 μmol/kg
(Table 6). Among the saponins, the furostanol saponin 1
imparted the lowest threshold concentrationa of 3 and 4 μmol/
kg for astringency and bitter taste, respectively.
In conclusion, sensory-guided fractionation of oat flour
extracts and chemical synthesis led to the identification of the
saponins 1−4 as well as the avenanthramides 5−10 as the most
intense bitter and astringent phytochemicals. Aimed at
demonstrating their relative contribution to the overall off-
taste of oat flour, quantitative studies are currently in progress
and will be published elsewhere.
(11) Molteberg, E. L.; Solheim, R.; Dimberg, L. H.; Frølich, W.
Variation in oat groats due to variety, storage and heat treatment. II:
Sensory quality. J. Cereal Sci. 1996, 24, 273−282.
(12) Schieberle, P.; Hofmann, T. Mapping the combinatorial code of
food flavors by means of the molecular sensory science concept. In
Food Flavors: Chemical, Sensory and Technological Properties; Jelen, H.,
Ed.; CRC Press, Taylor and Francis Group: 2011; pp 411−437.
(13) Dawid, C.; Hofmann, T. Identification of sensory-active
phytochemicals in asparagus (Asparagus officinalis L.). J. Agric. Food
Chem. 2012, 60, 11877−11888.
(14) Dawid, C.; Hofmann, T. Structural and sensory characterization
of bitter tasting steroidal saponins from asparagus spears (Asparagus
officinalis L.). J. Agric. Food Chem. 2012, 60, 11889−11900.
(15) Dawid, C.; Hofmann, T. Quantitation and bitter taste
contribution of saponins in fresh and cooked white asparagus
(Asparagus officinalis L.). Food Chem. 2014, 145, 427−436.
(16) Degenhardt, A. G.; Hofmann, T. Bitter-tasting and kokumi-
enhancing molecules in thermally processed avocado (Persea
americana Mill.). J. Agric. Food Chem. 2010, 58, 12906−12915.
(17) Dresel, M.; Dunkel, A.; Hofmann, T. Sensomics analysis of key
bitter compounds in the hard resin of hops (Humulus lupulus L.) and
their contribution to the bitter profile of Pilsner-type beer. J. Agric.
Food Chem. 2015, 63, 3402−3418.
(18) Dresel, M.; Vogt, C.; Dunkel, A.; Hofmann, T. The bitter
chemodiversity of hops (Humulus lupulus L.). J. Agric. Food Chem.
2016, 64, 7789−7799.
(19) Dawid, C.; Henze, A.; Frank, O.; Glabasnia, A.; Rupp, M.;
Buening, K.; Orlikowski, D.; Bader, M.; Hofmann, T. Structural and
sensory characterization of key pungent and tingling compounds from
black pepper (Piper nigrum L.). J. Agric. Food Chem. 2012, 60, 2884−
2895.
AUTHOR INFORMATION
■
Corresponding Author
*(T.H.) Phone: +49-8161/71-2902. Fax: +49-8161/71-2949.
E-mail: thomas.hofmann@tum.de.
ORCID
(20) Stark, T.; Bareuther, S.; Hofmann, T. Sensory-guided
decomposition of roasted cocoa nibs (Theobroma cacao) and structure
determination of taste-active polyphenols. J. Agric. Food Chem. 2005,
53, 5407−5418.
Thomas Hofmann: 0000-0003-4057-7165
Notes
The authors declare no competing financial interest.
M
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX

Journal of Agricultural and Food Chemistry
Article
(21) 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.
(22) 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.
(23) Stark, T.; Hofmann, T. Isolation, structure determination,
synthesis, and sensory activity of N-phenylpropenoyl-L-amino acids
from cocoa (Theobroma cacao). J. Agric. Food Chem. 2005, 53, 5419−
5428.
(40) Yang, J.; Wang, P.; Wu, W.; Zhao, Y.; Idehen, E.; Sang, S.
Steroidal saponins in oat bran. J. Agric. Food Chem. 2016, 64, 1549−
1556.
(41) Wise, M. L. Effect of chemical systemic acquired resistance
elicitors on avenanthramide biosynthesis in oat (Avena sativa). J. Agric.
Food Chem. 2011, 59, 7028−7038.
(42) Collins, F. W. Oat phenolics: avenanthramides, novel
substituted N-cinnamoylanthranilate alkaloids from oat groats and
hulls. J. Agric. Food Chem. 1989, 37, 60−66.
(43) Xu, J. G.; Tian, C. R.; Hu, Q. P.; Luo, J. Y.; Wang, X. D.; Tian,
X. D. Dynamic changes in phenolic compounds and antioxidant
activity in oats (Avena nuda L.) during steeping and germination. J.
Agric. Food Chem. 2009, 57, 10392−10398.
(24) Meyer, S.; Dunkel, A.; Hofmann, T. Sensomics-assisted
elucidation of the tastant code of cooked crustaceans and taste
reconstruction experiments. J. Agric. Food Chem. 2016, 64, 1164−1175.
(25) 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.
(44) Bratt, K.; Sunnerheim, K.; Bryngelsson, S.; Fagerlund, A.;
Engman, L.; Andersson, R. E.; Dimberg, L. H. Avenanthramides in
oats (Avena sativa L.) and structure−antioxidant activity relationships.
J. Agric. Food Chem. 2003, 51, 594−600.
(26) Schwarzbach, A. Comparative studies on saponins in Asparagus
officinalis (in German). Ph.D. thesis, Berlin, 2004.
(27) Kamat, S. P.; Parab, S. J. A simple two-step synthesis of
avenanthramides, constituents of oats (Avena sativa L). Indian J. Chem.
2007, 46B, 2074−2078.
(28) Agrawal, P. K. Assigning stereo-diversity of the 27-Me group of
furostane-type steroidal saponins via NMR chemical shifts. Steroids
2005, 70, 715−724.
(29) Shi, J.-g.; Li, G.-Q.; Huang, S.-Y.; Mo, S.-Y.; Wang, Y.; Yang, Y.-
C.; Hu, W.-Y. Furostanol oligoglycosides from Asparagus cochinchi-
nensis. J. Asian Nat. Prod. Res. 2004, 6, 99−105.
(30) Hayes, P. Y.; Jahidin, A. H.; Lehmann, R.; Penman, K.; Kitching,
W.; De Voss, J. J. Structural revision of shatavarins I and IV, the major
components from the roots of Asparagus racemosus. Tetrahedron Lett.
2006, 47, 6965−6969.
(31) Kang, L.-P.; Zhao, Y.; Pang, X.; Yu, H.-S.; Xiong, C.-Q.; Zhang,
J.; Gao, Y.; Yu, K.; Liu, C.; Ma, B.-P. Characterization and
identification of steroidal saponins from the seeds of Trigonella
foenum-graecum by ultra high-performance liquid chromatography and
hybrid time-of-flight mass spectrometry. J. Pharm. Biomed. Anal. 2013,
74, 257−267.
(32) Murakami, T.; Kishi, A.; Matsuda, H.; Yoshikawa, M. Medicinal
foodstuffs. XVII. Fenugreek seed. (3): structures of new furostanol-
type steroid saponins, trigoneosides Xa, Xb, XIb, XIIa, XIIb, and XIIIa,
from the seeds of Egyptian Trigonellafoenum-graecum L. Chem. Pharm.
Bull. 2000, 48, 994−1000.
(33) Simmons-Boyce, J. L.; Tinto, W. F.; McLean, S.; Reynolds, W.
F. Saponins from Furcraea selloa var. marginata. Fitoterapia 2004, 75,
634−638.
(34) Pecio, L.; Jedrejek, D.; Masullo, M.; Piacente, S.; Oleszek, W.;
Stochmal, A. Revised structures of avenacosides A and B and a new
sulfated saponin from Avena sativa L. Magn. Reson. Chem. 2012, 50,
755−758.
(35) Tschesche, R.; Lauven, P. Steroid saponins with more than one
sugar chain, V. Avenacosid B, another bisdesmosidic steroid saponin
from Avena sativa (in German). Chem. Ber. 1971, 104, 3549−3555.
(36) Oenning, G.; Asp, N. G.; Sivik, B. Saponin content in different
oat varieties and in different fractions of oat grain. Food Chem. 1993,
48, 251−254.
(37) Agrawal, P. K.; Jain, D. C.; Gupta, R. K.; Thakur, R. S. Carbon-
13 NMR spectroscopy of steroidal sapogenins and steroidal saponins.
Phytochemistry 1985, 24, 2479−2496.
(38) Saijo, R.; Fuke, C.; Murakami, K.; Nohara, T.; Tomimatsu, T.
Studies on the constituents of Solanum plants. Part 3. Two steroidal
glycosides, aculeatiside A and B from Solanum aculeatissimum.
Phytochemistry 1983, 22, 733−736.
(39) Tschesche, R.; Schmidt, W. Saponins of the spirostanol series.
XIII. Two new saponins from the aerial plant parts of oats (Avena
sativa), with nuatigenin as the aglycon. Z. Naturforsch. 1966, 21, 896−
897.
N
DOI: 10.1021/acs.jafc.6b04995
J. Agric. Food Chem. XXXX, XXX, XXX−XXX