Novel Taste-Enhancing 4-Amino-2-methyl-5-heteroalkypyrimidines Formed from Thiamine by Maillard-Type Reactions

Article abstract of DOI:10.1021/acs.jafc.9b05896

Increasing the thiamine concentration in a respective process flavor yields a product with a significant higher kokumi activity. S-plot analysis of the mass spectrometric data revealed beside thiamine itself, 4-methyl-5-thiazoleethanol, (S)-((4-amino-2-methylpyrimidin-5-yl)methyl)-l-cysteine, N-((4-amino-2-methylpyrimidin-5-yl)methyl)formamide, 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one, and 2-methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine as marker molecules for a process flavor with higher thiamine concentration. Sensory-based targeted isolation revealed that (S)-((4-amino-2-methylpyrimidin-5-yl)methyl)-l-cysteine, 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one, and 2-methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine showed an influence on the kokumi taste activity with taste threshold concentrations between 35 and 120 μmol/L. An adapted mass spectrometric-based carbon module labeling experiment as well as quantitative studies clearly demonstrated thiamine as the only precursor and an intermolecular formation pathway for the compounds (S)-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one and 2-methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine. On the basis of the knowledge that several thiamine derivatives showed taste-modulating activity, selected thiamine-based binary model reactions and synthesis were carried out. This resulted in the isolation of further thiamine-derived taste modulators like (S)-((4-amino-2-methylpyrimidin-5-yl)methyl)-l-cysteinylglycine, (S)-3-((((4-amino-2-methylpyrimidin-5-yl)methyl)thio)methyl)piperazine-2,5-dione, 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)pentan-2-one, 5-(((furan-2-ylmethyl)thio)methyl)-2-methylpyrimidin-4-amine, and (4-amino-2-methylpyrimidin-5-yl)methanethiol, 2-methyl-5-((methylthio)methyl)pyrimidin-4-amine with taste thresholds ranging from 35 to 880 μmol/L.

Full text of DOI:10.1021/acs.jafc.9b05896

Subscriber access provided by - Access paid by the | UCSF Library
Food and Beverage Chemistry/Biochemistry
“Novel Taste Enhancing 4-Amino-2-methyl-5-heteroalkypyrimidines
formed from Thiamine by Maillard-type reactions”
Laura Brehm, Oliver Frank, Manon Juenger, Miriam Wimmer, Josef Ranner, and Thomas Hofmann
J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.9b05896 • Publication Date (Web): 11 Nov 2019
Downloaded from pubs.acs.org on November 14, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 44
Journal of Agricultural and Food Chemistry
1
2
3
4
Novel Taste Enhancing 4-Amino-2-methyl-5-
heteroalkypyrimidines formed from Thiamine by
Maillard-type reactions
1
1
1
1
1
5
6
7
8
9
Laura Brehm , Oliver Frank , Manon Jünger , Miriam Wimmer , Josef Ranner and
1,2
Thomas Hofmann *
1Chair of Food Chemistry and Molecular and Sensory Science, Technische Universität
2
München, Lise-Meitner-Str. 34, D-85354 Freising, Germany. Leibniz-Institute for Food
1
1
1
1
1
1
1
1
1
0
1
2
3
4
5
6
7
8
Systems Biology, Technical University of Munich, Lise-Meitner-Strasse 34, D-85354
Freising, Germany.
*
To whom correspondence should be addressed
PHONE
FAX
+49-8161/71-2902
+49-8161/71-2949
E-MAIL
thomas.hofmann@tum.de
1
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 2 of 44
2
2
0
1
ABSTRACT
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
Increasing the thiamine concentration in a respective process flavor yields in a product
with a significant higher kokumi activity. S-Plot analysis of the mass spectrometric data
revealed beside thiamine itself, 4-methyl-5-thiazoleethanol, S-((4-amino-2-
methylpyrimidin-5-yl)methyl)-L-cysteine, N-((4-amino-2-methylpyrimidin-5-yl)methyl)-
formamide, 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one
and 2-methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine as marker
molecules for a process flavor with higher thiamine concentration. Sensory based
targeted isolation revealed that S-((4-amino-2-methylpyrimidin-5-yl)methyl)-L-cysteine,
3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one
and
2-
methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine showed an influence on
the kokumi taste activity with taste threshold concentrations between 35 to 120 µmol/L.
An adapted mass spectrometric based carbon modul labeling (CAMOLA) experiment
as well as quantitative studies clearly demonstrated thiamine as the only precursor and
an intermolecular formation pathway for the compounds S-(((4-amino-2-
methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one
and
2-methyl-5-(((2-
methylfuran-3-yl)thio)methyl)pyrimidin-4-amine. Based on the knowledge that several
thiamine derivatives showed taste modulating activity, selected thiamine-based binary
model reactions and synthesis were carried out. This resulted in the isolation of further
thiamine derived taste modulators like S-((4-amino-2-methylpyrimidin-5-yl)methyl)-L-
cysteinylglycine,
(S)-3-((((4-amino-2-methylpyrimidin-5-yl)methyl)thio)methyl)-
piperazine-2,5-dione, 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)pentan-2-one,
5-(((furan-2-ylmethyl)thio)methyl)-2-methylpyrimidin-4-amine,
(4-amino-2-
2
ACS Paragon Plus Environment

Page 3 of 44
Journal of Agricultural and Food Chemistry
4
4
4
5
methylpyrimidin-5-yl)methanethiol, 2-methyl-5-((methylthio)methyl)pyrimidin-4-amine
with taste thresholds ranging from 35 to 880 µmol/L.
4
6
7
KEYWORDS: umami, kokumi, Process Flavors, thiamine, S-Plot, CAMOLA, model reactions, taste
4
enhancer, taste modulators, meat flavor
3
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 4 of 44
4
5
9
0
INTRODUCTION
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
The so-called process flavors (PFs) are food flavorings which are produced by heating
various precursors together under controlled reaction conditions, whereby the
reactions mimic kitchen-like conditions.1,2 The aroma and the taste properties of such
process flavors are similar to those found in thermally treated foodstuff like meat,
3
chocolate, coffee, caramel, popcorn or bread. Common precursors during the
manufacturing of the process flavors are amino acids, reducing sugars, thiamine,
ribonucleotides, phospholipids, yeast extracts or hydrolyzed vegetable proteins. The
most important reactions involved in the formation of process flavors include the
Maillard reaction, Strecker degradation, sugar degradation/fragmentation or thermal
conversions of thiamine.1–3
Along with the characteristic aroma, particularly the umami and the kokumi taste
perception are very important for a meaty flavor impression. The term kokumi is defined
as a perceived complexity, continuity and mouth fullness. It was first described by Ueda
et al. who isolated kokumi active compounds from a water extract of garlic.4,5 Since
that time many kokumi active compounds have been isolated from food stuff e.g. edible
beans, thermally processed avocados and gouda cheese.6–8 Recently, Mittermeier et
al. isolated kokumi modulating octadecadien-12-ynoic acids from chanterelles.9
Taste modulating compounds have not only been isolated from food stuff but also from
model reactions systems.10–12 A important representative of this compound class
formed during the Maillard reaction is the so called alapyridaine. It could be isolated
from an alanine/hexose model reaction mixture after thermal treatment.10 This
compound is tasteless itself but enhances the sweet, salty and umami taste perception.
2
Festring and Hofmann (2010) identified the umami modulating N -(1-
4
ACS Paragon Plus Environment

Page 5 of 44
Journal of Agricultural and Food Chemistry
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
Carboxyethyl)guanosine-monophosphate- a nucleotide modified by the Maillard
reaction.11 By thermal treatment of creatinine and various sugars it was possible to
identify N-(1-methyl-4-oxoimidazolidin-2-ylidene)-α-amino acids as another taste
modulating substance class. Sensory evaluations of these compounds revealed that
12
they have a kokumi modulating effect. Very recently, an activity guided fractionation
of a meaty process flavor system lead to the isolation and identification of a new taste
modulating substance derived from thermally induced reaction of thiamine and
cysteine, the S-((4-amino-2-methylpyrimidin-5-yl)methyl)-L-cysteine.13
One precursor of this compound, thiamine, plays an important role during the formation
of meaty process flavors, as it is also a precursor for many (meaty) aroma compounds
such as aliphatic sulphides and thiols, sulphur-containing carbonyl compounds,
sulphur-substituted furans, thiophenes, thiazolees, bicyclic compounds and alicyclic
sulphur.14-16 The meaty aroma compound 2-methyl-3-furanthiol (MFT) is one of the
most important aroma compounds formed by thiamine degradation.14 However, the
focus of thiamine degradation in flavor research has mostly been directed towards the
volatile fraction only. Whereas for a meaty impression especially the umami and the
kokumi taste perception seemed to be important as well. The isolation of the taste
modulating compound S-((4-amino-2-methylpyrimidin-5-yl)methyl)-L-cysteine showed
that thiamine also plays an important role as a precursor for taste-modulating
13
compounds. Hence, the current study focusses on the formation of non-volatile taste-
modulating compounds derived from thiamine. Therefore, binary model reactions
consisting of thiamine and various further compounds were performed. The reaction
partner for thiamine was selected from substances, which play a key role in the flavor
perception of meaty process flavors, e.g. the kokumi active glutathione or the meat
aroma compounds 2-methyl-3-furanthiol (MFT) and furan-2-ylmethanethiol (FFT).
5
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 6 of 44
9
9
MATERIALS AND METHODS
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
Chemicals.
The following compounds were obtained commercially: acetonitrile, methanol (J.T.
Baker, Netherlands), ethyl acetate (VWR, Darmstadt, Germany), formic acid, (Merck,
Darmstadt, Germany). Solvents used for LC-MS/MS analysis were of LC-MS grade
(Honeywell, Seelze, Germany). Water for HPLC separation was purified using a Milli-
Q water advantage A10 water system (Millipore, Schwalbach, Germany). For sensory
analysis, bottled water (Evian, Danone, Wiesbaden, Germany) was used. Yeast
extract (Gistex XII LS) was obtained from Food Ingredients Distribution (Werne,
Germany). Deuterated solvents, thiamine hydrochloride (1, Figure 1), L-Cysteine, 4-
methyl-5-thiazoleethanol (2, Figure 1), 2-methyl-3-furanthiol (MFT, 20, Figure 4),
sodium thioacetate, thiomethoxide, methylamine (dissolved in THF), thiamine-(4-
13
13
methyl- C-thiazole-5-yl- C ) hydrochloride, water free Dimethylformamide (DMF) and
3
sodium hydroxide were obtained from Sigma Aldrich (Steinheim, Germany). 3-
mercapto-2-pentanone was obtained from ABCR-Chemicals (Karlsruhe, Germany). 2-
furfurylthiol (FFT) was purchased from Givaudan (Dübendorf, Switzerland). N-((4-
amino-2-methylpyrimidin-5-yl)methyl)formamide
(4,
Figure
1),
(4-amino-2-
methylpyrimidin-5-yl)methanol (11, Figure 1) and 5-(aminomethyl)-2-methylpyrimidin-
4-amine (12, Figure 1) were obtained from Carbosynth (Berkshire, United Kingdom).
Ethyl thiamine (16, Figure 1), 5-(bromomethyl)-2-methylpyrimidin-4-amine were
obtained from Santa Cruz Biotechnology (Heidelberg, Germany).
Two different process flavors (PF1 and PF2) were kindly provided by Lucta S.A.
(Barcelona, Spain).
Model Reaction Systems.
6
ACS Paragon Plus Environment

Page 7 of 44
Journal of Agricultural and Food Chemistry
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
Binary Reaction Systems: For the preparation of the binary model reaction systems
thiamine hydrochloride (1 mmol) or ethyl thiamine (1 mmol) and a second compound
(L-cysteine, glutathione, 3-mercapto-2-pentanone, MFT or FFT; 1 mmol each) were
dissolved in aqueous KH PO -buffer (0.1 M, 10 mL). The pH value was adjusted to 7
2
4
with a NaOH-solution (1 M). Afterwards the mixture was heated in an aluminum block
at 120 °C for 120 minutes while stirring in a closed glass vessel.
Adapted CAMOLA Experiment: For the adapted CAMOLA experiment 13C labeled
4
thiamine hydrochloride (0.1 mmol) and ethyl thiamine (0.1 mmol) were dissolved in
aqueous KH PO -buffer (0.1 M, 1 mL). After adjusting the pH-value to 7 (with a 1 M
2
4
NaOH solution), the mixture was heated at 120 °C for 120 minutes in a closed glass
vessel.
Fractionation of the Thiamine and Cysteine Reaction Mixture.
After cooling the thiamine cysteine reaction mixture to room temperature, the reaction
mixture was extracted twice with ethyl acetate (EtOAc). The combined organic layers
were freed from solvent in vacuum to obtain the ethyl acetate extractables. After
lyophilization the EtOAc-extract was stored at -20 °C until further use.
Aimed at locating the targeted compounds the ethyl acetate fraction was dissolved in
acetonitrile/water (50/50, v/v, membrane-filtered (0.45 µm) and separated by means of
preparative RP-HPLC on a MonoChrom MS (250 x 21.2mm i.d., 5u; Varian,
Darmstadt, Germany) column into 7 fractions (F1 – F7). Monitoring the effluent with an
UV-detector at 254 nm, chromatography was performed at a flow rate of 20 mL/min
using aqueous formic acid (0.1%) as solvent A and acetonitrile as solvent B. Starting
with 0% B for 2 minutes the gradient was increased to 40% B in 12 minutes. Within
another 2 minutes, the gradient was increased to 100% B and maintained for 1 minute.
7
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 8 of 44
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
Sub Fractionation and Structure Determination of Fraction F5. The lyophilized HPLC-
fraction F5 was dissolved in acetonitrile/water (50/50, v/v), membrane-filtered
(0.45 µm) and separated by means of preparative RP-HPLC on Luna Pentafluorphenyl
(250 x 10 mm i.d., 5µm, 100 Å; Phenomenex, Aschaffenburg, Germany) column into
three sub fractions F5-a - F5-c. For the separation aqueous formic acid (0.1%) was
used solvent A and methanol as solvent B. The effluent was set to 20 mL/min and was
monitored with an UV-detector (λ= 254 nm). The gradient started with 25% B for 2
minutes, afterwards solvent B was increased to 65% in 6 minutes. Within 1 minute
solvent B was increased to 100% B. After separation the organic solvent was
evaporated, and fractions were lyophilized. Based on UV-Vis, LC-MS/MS, TOF-MS
and 1D/2D-NMR experiments, the structure of the compound isolated from fraction F5-
c, could be identified as 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)-5-hydroxy-
pentan-2-one (5a, Figure 1).
UV-Vis
λmax
=
254 nm
(MeOH/H O,
50/50,
v/v),
UPLC-TOF-MS:
2
+
1
m/z=256.1232 [M+H] ; H NMR (500 MHz, D O, 300 K): δ (ppm) 8.07 [s, 1H, H-C(6)],
2
3.74 – 3.64 [m, 4H, H-C(5’,7)], 3.54 [t, J = 7.4 Hz, 1H, H-C(3‘)], 2.57 [s, 3H, H-C(8)],
2.30 [s, 3H, H-C(1‘)], 2.12 [ddt, J = 14.3, 7.4 Hz, 5.5 Hz 1H H -C(4‘)], 1.88 [ddt, J =
a
13
14.3, 7.4, 5.5 Hz, 1H, H -C(4‘)]. C NMR (125 MHz, D O, 300 K): δ (ppm) 210.66
b
2
[C(2‘)] , 163.09 [C(4)], 162.18 [C(2)], 143.40 [C(6)], 111.47 [C(5)], 58.53 [C(5‘)], 50.56
[C(3‘)], 31.72 [C(4‘)], 27.54 [C(7)], 26.01 [C(1‘)], 21.04[C(8)].
Structure Determination of Fraction F6. After separation, evaporation of the organic
solvent and lyophilization the structure of F6 was identified by means of UV-Vis, LC-
MS/MS, TOF-MS and 1D/2D-NMR experiments as 2-methyl-5-(((2-methylfuran-3-
yl)thio)methyl)pyrimidin-4-amine (6, Figure 1).
8
ACS Paragon Plus Environment

Page 9 of 44
Journal of Agricultural and Food Chemistry
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
UV-Vis λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=236.0974
2
+
1
[M+H] ; H NMR (400 MHz, MeOD-d 298 K): δ (ppm) 7.35 [d, J = 1.9 Hz, 1H, H-C(5’)],
4,
7.29 [s, 1H, H-C(6)], 6.29 [d, J = 1.9 Hz, 1H, H-C(4’)], 3.65 [s, 2H, H-C(7)], 2.38 [s, 3H,
13
H-C(8)], 1.97 [s, 3H, H-C(6’)]. C NMR (100 MHz, MeOD-d , 298 K): δ (ppm) 166.73
4
[C(2)] , 163.11 [C(4)], 158.23 [C(2’)], 153.99 [C(6)], 142.24 [C(5’)], 116.42 [C(4’)],
111.94 [C(5)], 109.92 [C(3’)], 34.09 [C(7)], 24.57 [C(8)], 11.24 [C(6’)].
Fractionation of Binary Model Reaction Systems. The cooled reaction
mixtures were separated by means of HPLC. The exact parameters are disclosed in
the supporting information. Following substances could be isolated and confirmed by
means of UV-Vis, LC-MS/MS, TOF-MS and 1D/2D-NMR experiments.
S-((4-amino-2-methylpyrimidin-5-yl)methyl)cysteinylglycine (7, Figure 1). UV-Vis
+
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=300.1129 [M+H] ;
2
1
H NMR (500 MHz, D O, 300 K): δ (ppm) 8.07 [s, 1H, H-C(6)], 4.23 [t, J = 6.4 Hz, 1H,
2
H-C(5‘)], 3.91 – 3.71 [m, 4H, H-C(7), H-C(2‘)], 3.03 [d, J = 6.4 Hz, 2H, H-C(6‘)], 2.58
13
[s, 3H. H-C(8)]. C NMR (125 MHz, D O, 300 K): δ (ppm) 175.68 [C(1‘)], 167.94 [C(4‘)],
2
163.24 [C(4)], 161.75 [C(2)], 141.74 [C(6)], 111.20 [C(5)], 52.23 [C(5‘)], 43.26 [C(2‘)],
31.11 [C(6‘)], 28.38 [C(7)], 20.73 [C(8)].
3-((((4-amino-2-methylpyrimidin-5-yl)methyl)thio)methyl)piperazine-2,5-dione
(8,
Figure 1). UV-Vis λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS:
2
+
1
m/z=282.1030 [M+H] ; H NMR (400 MHz, D O, 298 K): δ (ppm) 8.09 [s, 1H, H-C(6)],
2
4.52 [dd, J = 8.5, 5.2 Hz, 1H, H-C(6‘)], 3.82 [s, 2H, H-C(3‘)], 3.73 [m, 2H, H-C(7)], 2.58
13
[s, 3H, H-C(8)]. C NMR (100 MHz, D O, 298 K): δ (ppm) 175.04 [C(5‘)], 171.82 [C(2‘)],
2
163.93 [C(4)], 161.64 [C(2)], 141.33 [C(6)], 111.75 [C(5)], 52.66 [C(6‘)], 42.90 [C(3‘)],
31.11 [C(7‘)], 28.48 [C(7)], 20.48 [C(8)].
9
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 10 of 44
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)pentan-2-one (9, Figure 1). UV-Vis
+
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=240.1170 [M+H] ;
2
1
H NMR (400 MHz, D O, 298 K): δ (ppm) 7.92 [s, 1H, H-C(6)], 3.68 [d, J = 14.7 Hz, 1H
2
H -C(7)], 3.62 [d, J = 14.7 Hz, 1H, H -C(7)], 3.32 [dd, J= 7.35, 7.35, 1H, H-C(3‘)], 2.41
a
b
[s, 3H, H-C(8)], 2.20 [s, 3H, H-C(1‘)], 1.81 [dq, J = 14.5, 7.3 Hz, 1H, H -C(4‘)], 1.67 [dq,
a
13
J = 14.5, 7.3 Hz, 1H, H -C(4‘)], 0.88 [t, J = 7.3 Hz, 3H, H-C(5‘)]. C NMR (100 MHz,
b
D O, 298 K): δ (ppm) 213.25 [C(2‘)] , 167.35 [C(4)], 162.157 [C(2)], 155.11 [C(6)],
2
111.58 [C(5)], 56.56 [C(3‘)],29.18 [C(7)], 26.85 [C(1‘)], 24.16 [C(1‘)], 17.45[C(8)], 11.54
[C(5‘)].
5-(((furan-2-ylmethyl)thio)methyl)-2-methylpyrimidin-4-amine (10, Figure 1). UV-Vis
+
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=236.0911 [M+H] ;
2
1
H NMR (400 MHz, MeOD-d 298 K): δ (ppm) 7.87 [s, 1H, H-C(6)], 7.41 [dd, J = 1.9,
4,
0.9 Hz, 1H, H-C(5’)], 6.33 [dd, J = 3.2, 1.9 Hz, 1H, H-C(4’)], 6.20 [dd, 1H, J=3.2, 0.9
13
Hz, H-C(3’)], 3.67 [s, 2H, H-C(6’)], 3.61 [s, 2H, H-C(7)], 2.41 [s, 3H, H-C(8)]. C NMR
(100 MHz, MeOD-d , 298 K): δ (ppm) 166.61 [C(2)], 163.55 [C(4)], 153.56 [C(6)],
4
152.99 [C(2’)], 143.41 [C(5’)], 111.80 [C(C4’)], 111.51 [C(5)], 108.62 [C(3’)], 29.79
[C(7)], 28.32 [C(6’)], 24.43 [C(8)].
2-ethyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine (17, Figure 1), UV-Vis
+
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z= 250.1024 [M+H] ;
2
1
H (400 MHz, MeOD-d , 298 K): δ (ppm) 7.33 [d, 1 H, J = 1.9 Hz, H-C(5‘)] , 7.27 [s,
4
1 H, H-C(6)], 6.28 (d, 1 H, J = 1.9 Hz, H-C(4‘)], 3.63 [s, 2 H, H-C(7)], 2.61 [q, 2 H,
13
J = 7.7 Hz, H-C-(8)], 1.91 [s, 3 H, H-C(6‘)], 1.22 [t, 3 H, J = 7.7 Hz, H-C(9)]. C
(125 MHz, MeOD-d , 298 K): δ (ppm) 171.3 [C(2)], 163.2 [C(4)], 158.3 [C(2)],
4
154.3 [C(6)], 142.2 [C(5‘)], 116.4 [C(4‘)], 112.0 [C(5)], 109.9 [C(3‘)], 34.2 [C(7)],
32.5 [C(8)], 13.3 [C(9)], 11.2 [C(6‘)].
10
ACS Paragon Plus Environment

Page 11 of 44
Journal of Agricultural and Food Chemistry
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
3-(((4-amino-2-ethylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one (18, Figure 1)
UV-Vis λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z= 270.1014
2
+
1
[M+H] ; H (400 MHz, D O, 298 K): δ (ppm) 7.98 [s, 1H, H-C(6)], 3.71 – 3.62 [m, 4H.
2
H-C(7), H-C(5’)], 3.51 [t, 1H, H-C(3’)], 2.73 [q, J = 7.6 Hz, 2H, H-C(8)], 2.23 [s,
2H,H-C(1’)], 2.08 [ddt, 1H, H -C(4’), J=7.6, 7.6, 13.74], 1.84 [ddt, 1H, H -C(4’), J = 7.6,
a
b
7.6, 13.74 Hz], 1.26 [t, J = 7.6 Hz, 3H, H-C(9)].
Synthesis of further Thiamine Derivates. Further thiamine derivatives were
synthesized according to Zheng et al.17 The synthesis pathway and the isolation
parameters can be found in the supporting information. Following substances could be
1
isolated and confirmed by means of UV-Vis, LC-MS/MS, TOF-MS and H-NMR
experiments.
(4-amino-2-methylpyrimidin-5-yl)methanethiol (13, Figure 1) UV-Vis λmax = 254 nm
+
1
(MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=156.056 [M+H] ; H NMR (400 MHz,
2
MeOD-d , 298 K): δ (ppm) 7.83 [s, 1H, H-C(6)], 3.54 [s, 2H, H-C(7)], 2.43 [s, 3H, H-
4
C(8)].
2-methyl-5-((methylamino)methyl)pyrimidin-4-amine (14, Figure 1), UV-Vis
+
1
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=153.1141 [M+H] ; H
2
NMR (400 MHz, D O, 298 K): δ (ppm) 8.11 [s, 1H, H-C(6)], 4.11 [s, 2H, H-C(7)], 2.70
2
[s, 3 H, H-C(1’)], 2.37 [s, 3 H, H-C(8)].
2-methyl-5-((methylthio)methyl)pyrimidin-4-amine (15, Figure 1),
UV-Vis
+
1
λmax = 254 nm (MeOH/H O, 50/50, v/v), UPLC-TOF-MS: m/z=170.0781 [M+H] ; H
2
NMR (400 MHz, MeOD-d , 298 K): δ (ppm) 7.96 [s, 1H, H-C(6)], 3.61 [s, 2H, H-C(7)],
4
2.46 [s, 3H, H-C(8)], 2.02 [s,H- C(1’)].
Quantitation of Thiamine and Thiamine Derivates. For screening and
quantitation of the thiamine derivatives 1-15 an LC-MS/MS method was developed on
11
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 12 of 44
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
a 5500 QTrap LC-MS/MS system (AB Sciex, Darmstadt, Germany) equipped with
Luna PFP column a (150 × 2.0 mm i.d., 3 μm, Phenomenex, Aschaffenburg,
Germany). Chromatography was performed with acetonitrile containing 1% formic acid
and as solvent B and water containing 1% formic acid as solvent A. The flow rate was
set to 0.3 mL/min. Starting with 5% B and holding it for 3 min, the ratio of solvent B
was linearly increased to 40% within 8 min and held for 1 min. Thereafter starting
conditions were readjusted within 0.5 min and equilibrated for 3.5 min. The compounds
1-15 were quantitated by means of multiple reaction monitoring (MRM) with defined
+
mass transitions in an ESI mode. The parameters like declustering potential (DP, in
V), collision energy (CE in V) and cell exit potential (CXP in V) are disclosed in the
supporting information.
Aliquots (1 mg) of the process flavors or diluted reaction mixtures were
dissolved in a water/methanol-mixture (1 mL, 50/50, v/v). After adding the internal
standard (17), the samples were membrane filtered (0.45 µm) and injected to the LC-
MS/MS-system.
Sensory Analyses. Sensory Panel Training. Eight female and nine male
panellists (23-32 years in age), who had no history of known taste disorders and who
had given the informed consent to participate in the present sensory tests, were trained
in weekly training sessions for at least two years in order to became familiar with the
taste language and methodologies used, to evaluate the taste of aqueous reference
solutions (2.0 mL; pH 5.9): sucrose (50 mmol/L) for sweet taste, L-lactic acid
(20 mmol/L) for sour taste, NaCl (20 mmol/L) for salty taste, caffeine (1 mmol/L) for
bitter taste, monosodium L-glutamate (3 mmol/L) for umami taste. To train the activity
of mouth fullness enhancement and complexity increase, coined kokumi activity, the
12
ACS Paragon Plus Environment

Page 13 of 44
Journal of Agricultural and Food Chemistry
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
panelists were requested to compare the gustatory impact of a blank model broth
(control) with a solution of reduced glutathione (5 mmol/L) in model broth.4,5
Preparation of Model Broth. For the detection of umami modulating substances, a
model broth was prepared. This model broth served as reference for all sensory
analyses. Therefore sodium chloride (2.9), monosodium glutamate (MSG 1.9 g),
maltodextrin (6.4 g) and yeast extract (2.1 g) were solved in 1 L of ultrapure water.
Duo-Trio-Test. To evaluate the sensory differences within between PF1 and PF2 duo-
trio tests were used. Therefore, the respective process flavors were dissolved in model
broth (5 g/L).
Determination of Taste Threshold. For the determination of the taste threshold, a
solution with known concentration was prepared in water or in model broth and
successively diluted 1:1 (v/v). The solutions were presented to the panelists in order
of increasing concentrations in a duo-trio test. Water or model broth was used as blank
and reference, respectively. The panelists were asked to describe the difference
between the blank and the sample. The first dilution step where a difference between
the blank and the sample was detectable was defined as the taste threshold. The taste
threshold was calculated by the geometric mean.
High-Performance Liquid Chromatography (HPLC). The HPLC apparatus
(Jasco, Gross-Umstadt, Germany) consisted of a binary high-pressure HPLC pump
system PU-2080 Plus, an AS-2055 Plus autosampler, a DG-2080-53 degasser, a MD-
2010 Plus type diode array detector (Jasco, Gross-Umstadt, Germany) and a Rh 7725i
type Rheodyne injection valve (Rheodyne, Bensheim, Germany). Data acquisition was
performed by means of Chrompass Chromatography Data System, Version 1.9
(Jasco, Gross-Umstadt, Germany).
13
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 14 of 44
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
Liquid Chromatography-Mass spectrometry (LC-MS). LC-MS/MS analysis
was performed on a 5500 QTrap MS/MS system (AB Sciex, Darmstadt, Germany)
connected to a Shimadzu Nexera X2 system (Shimadzu, Duisburg, Germany) system
+
running in the positive electrospray ionization (ESI ) mode. For measurements on the
QTrap system the following conditions were used. Zero grade air served as nebulizer
gas (55 psi), and as turbo gas for solvent drying (65 psi, 450 °C). Nitrogen served both
−
7
as curtain gas (35 psi) and collision gas (8.7× 10 psi) dissociation potential (-2 V) and
+
entrance potential (-10 V). Both quadrupoles were set at unit resolution. ESI mass
+
and product ion spectra were acquired with direct flow infusion. For ESI , the ion spray
voltage was set at +5500 V. Energies for declustering potential (DP), entrance potential
(EP), collision energy (CE), and cell exit potential (CXP) as well as MS/MS parameters
for measuring in the MRM mode were optimized for each compound individually, to
detect the fragmentation of molecular ions into specific product ions after collision with
nitrogen. For instrumental control and data acquisition, Sciex Analyst software v1.6
was used.
UPLC/Time-of-Flight Mass Spectrometry (UPLC/TOF-MS). An aliquot (0.1 mg
- 1mg) of the sample, dissolved in methanol/water (30/70, v/v; 1mL), was injected into
an Acquity UPLC core system (Waters UK Ltd., Manchester, UK) connected to a
SYNAPT G2 HDMS spectrometer (Waters UK Ltd., Manchester, UK) operating in a
+
positive electrospray (ESI ) modus with the following parameters: capillary voltage
(+2.0 kV), sampling cone (20 V), source temperature (120 °C), desolvation
temperature (450 °C), cone gas (5 L/h), and de-solvation gas (850 L/h).
Chromatographic separations were performed on a BEH C18 column (2.1 x 150 mm,
1.7 µm, Waters UK Ltd., Manchester, UK) operated at 45 °C with a solvent gradient
(flow rate 0.4 mL/min) of 0.1% aqueous formic acid (solvent A) and 0.1% formic acid
14
ACS Paragon Plus Environment

Page 15 of 44
Journal of Agricultural and Food Chemistry
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
in acetonitrile (solvent B). Starting with 5% B the ratio was increased in 4 min to
100% B. The instrument was calibrated over a mass range from m/z 100 to 1200 using
a solution of sodium formate (0.5 mmol/L) in 2-propanol/water (9/1, v/v). All data were
lock mass corrected using leucine enkephaline as the reference (m/z 556.2771 for
+
-
[M+H] ; m/z 554.2615 for [M-H] ). Data acquisition and analysis were done by using
the MassLynx software (version 4.1; Waters).
For the calculation of the group differences the PF1 and PF2 were dissolved in
methanol/water (30/70, v/v; 0.5 mg/mL) and injected 5 times. The raw data the process
flavors and their replicates obtained from UPLC-ESI-TOF-MS analysis were processed
with Progenesis QI. The processed data were exported to EZinfo. The group
differences between the two process flavors were calculated using orthogonal partial
least-squares discriminant analysis (OPLC-DA) highlighted as S-Plot.
Nuclear Magnetic Resonance Spectroscopy (NMR). 1D- and 2D-NMR
experiments were performed on a Bruker 400 MHz AV III or 500 MHz AV III
spectrometer (Bruker, Rheinstetten, Germany) equipped with a Z-gradient 5 mm
multinuclear observe probe (BBFOplus) or a triple resonance cryo-Probe (TCI, H/C/N) .
MeOD-d (600 µL) or D O (600 µL) were used as solvent and chemical shifts are
4
2
reported in parts per million referenced to the MeOD-d or D O solvent signals,
4
2
respectively. Data processing was performed by using Topspin NMR software (version
3.2; Bruker, Rheinstetten, Germany) and MestReNova 10.0 (Mestrelab Research,
Santiago de Compostela, Spain). For quantitative NMR spectroscopy (qHNMR), the
spectrometer (400 MHz, BBFOplus) was calibrated by using the ERETIC 2 tool based
18
on the PULCON methodology as reported earlier. The isolated signal at 3.96 ppm (t,
J = 5.3 Hz, 1H) was used for absolute quantitation of 1-15 by using a sample of L-
15
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 16 of 44
3
3
3
42
43
44
tyrosine at a defined concentration (5.21 mmol/L) as the external standard and its
specific resonance signal at 7.10 ppm (m, 2H) for analyses.
3
3
45
46
RESULTS AND DISCUSSION
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
Recently a reaction product of thiamine (1, Figure 1) and L-cysteine, S-((4-amino-2-
methylpyrimidin-5-yl)methyl)-L-cysteine (3, Figure 1) was identified as a new taste
modulating compound in a complex model reaction system, a so-called process flavor
13
(PF). This compound is formed during heating of thiamine (1, Figure 1) and cysteine,
wherein cysteine substitutes the thiazole moiety (2, Figure 1) of thiamine. It is of
tremendous interest for the food industry to increase the formation of taste modulating
substances in savory systems like process flavors. One possible way to enhance the
formation of 3 in process flavors is to increase the amount of thiamine. In order to
evaluate whether an increased thiamine concentration leads to other savory products,
two different process flavors were produced. One with the original thiamine
concentration (PF1) and one with a ten times higher thiamine concentration (PF2). All
other ingredients and the heating parameters were kept constant.
To evaluate the sensory impact of the increased thiamine concentration on the taste
activity of the process flavors a sensory duo-trio test was performed. With a
significance value of α=0.1 the panelists could detect a difference between PF1 and
PF2. The panelists characterized PF1 as less savory and described a higher kokumi
activity in PF2.
However, the composition of the process flavors is very complex and other compounds
might be formed during thermal thiamine conversion as well. To identify other potential
savory impact compounds which are higher abundant in PF2 an UPLC-TOF-MS
16
ACS Paragon Plus Environment

Page 17 of 44
Journal of Agricultural and Food Chemistry
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
screening of the two process flavors was performed. To visualize the differences of the
two process flavors a mass spectrometric based S-Plot, representing pairs of exact
molecular mass to charge ratio and retention time of each compound from the two
process flavors was calculated. The y-axis of the S-plot is a visualization of the
confidence of a compounds’ contribution to the group difference and the x-axis denotes
the contribution of a compound to the group difference.19,20 The S-Plot indicated that
the ions m/z 122.0845, 144.0612, 167.1061, 236.0974, 243.1032 and 256.1232 are
m/z ratios of compounds showing the highest difference in between PF1 and PF2 and
were higher abundant in PF2 (Figure 2).
By comparing the compounds with reference substances and co-chromatography, the
fragment ion with m/z 144.0610 (R =0.9 min) and the fragment ion m/z 122.0845
t
(R =0.9 min) could be explained by in-source fragmentation of thiamine).
t
Furthermore, it was possible to identify known degradation products of thiamine such
as 4-methyl-5-thiazoleethanol (2, Figure 1) with m/z 144.0612 (R =1.6 min) and N-((4-
t
amino-2-methylpyrimidin-5-yl)methyl)formamide (4, Figure 1) with m/z 167.1061
21–23
(R =1.12 min).
In addition, the already discovered taste modulating compound 3
t
13
could be identified as marker for PF2 with m/z 243.1032 (R =1.0 min).
t
Apart from that, the compounds with m/z 256.1232 (R =1.6 min) and m/z 236.0974
t
(R =2.1 min) could not be identified by comparing with reference substances. To
t
evaluate whether those highly abundant substances have an impact on the overall
taste of the PF2, the two unknown substances have to be isolated.
Isolation of unknown compounds. Since the composition of the process
flavors is very complex an isolation directly from the process flavors would be very
laborious and time consuming. Hence, a binary model reaction system of thiamine and
cysteine was prepared. Cysteine was chosen, as a second ingredient, since the
17
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 18 of 44
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
elemental composition of both compounds indicated a sulfur in the target molecule and
due to the fact that it play an important role during the formation of 3.
Through screening with UPLC-TOF-MS it was possible to confirm that the unknown
targeted substances with the m/z ratios of 256.1232 and 236.0974 are formed during
this reaction.
Aimed at locating the these compounds a sequential solvent fractionation was
performed. The model reaction was extracted twice with ethyl acetate (EtOAc) to
obtain a water and an organic solvent soluble fraction. UPLC-TOF-MS screening
revealed that the target compounds are extractable with ethyl acetate. This organic
extract was further fractionated by means of preparative HPLC into seven sub fractions
(F1-F7; supporting information). UPLC-TOF-MS screening revealed that fraction F5
contained the compound with m/z 256.1232 whereas, the compound with
m/z 236.0974 eluted in fraction F6.
Sub Fractionation and Structure Determination of Fraction F5. In order to
isolate the compound from fraction F5 the thiamine-cysteine model reaction system
was further fractionated by means of semipreparative HPLC into three sub fractions
(F5-a - F5-c). The target compound eluted in fraction F5-c (supporting information).
+
The UPLC-TOF-MS analysis in the ESI mode indicated a pseudo molecule ion of
+
+
m/z 256.1232 [M+H] with a calculated formula of [C H N O S+H] . The fragment ion
11
17
3
2
+
of m/z 154.044 [C H N S] indicated that a compound is linked via a sulfur to the
6
8
3
aminopyrimidine moiety of the thiamine.
For accurate structure determination of the unknown compound 1D- and 2D-NMR-
1
experiments were performed. In the H-NMR spectrum, a strong de-shielded single
aromatic proton at 8.03 ppm (H-C(6)) was observed, well in the line with the aromatic
proton of the aminopyrimidine moiety of thiamine. This assumption was further
18
ACS Paragon Plus Environment

Page 19 of 44
Journal of Agricultural and Food Chemistry
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
strengthened by the HMBC correlations of H-C(6) to C(5), C(2) and C(4) at 114.4 ppm,
162.18 ppm and 163.09 ppm, respectively (Figure 3A). With exception of the carbon
C(7) resonating at 27.54 ppm the chemical shifts of the protons and the carbon atoms
24
of the aminopyrimidine moiety were comparable to those of an intact thiamine. The
carbon atom C(7) at 27.54 ppm showed a correlation signal in the HMBC spectrum
with H-C(3’) at 3.54 ppm, which is most likely linked to a hetero-nucleus. In addition,
with the results of the high-resolution mass spectrometry, this correlation confirmed
the connection of the aminopyrimidine moiety via a sulfur atom to another constituent
of the molecule. The correlation of H-C(3’) in the HMBC spectrum with a quaternary
carbon C(2’) resonating at 210.7 ppm, indicated that a keto group was incorporated in
the structure. This keto group showed a 3JC,H coupling in the HMBC with a methyl group
H-C(1’) at 2.22 ppm as well as with a diastereotopic methylene group H-C(4’). This
methylene group showed a 3JC,H coupling to another another methylene group at 3.64
ppm (H-C(5’)). The shift of H-C(5’) to higher frequencies indicated that it is linked to a
hydroxy group (Figure 3A). Taking all spectrometric and spectroscopic data into
consideration the compound of fraction F5-c was confirmed as 3-(((4-amino-2-
methylpyrimidin-5-yl)methyl)thio)-5-hydroxypentan-2-one (5a, Figure 1), in which the
thiazole moiety (2) of the thiamine (1) is substituted by 5-hydroxy-3-mercapto-2-
1
pentanone (19a, Figure 4). H-NMR experiments revealed that this thiamine derivate
does not only occur in an open chained (5a) but also as closed cyclic ketal form (5b,
Figure 1). There is a clear chemical shift of the methyl group of the open-chained form
(2.30 ppm) and the cyclic ketal form (1.46 ppm). By means of qHNMR it was possible
to show that the ratio of the two tautomers (85 % open, 15% cyclic ketal) was favored
to the open form. To the best of our knowledge, these compounds have not been
described in literature before.
19
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 20 of 44
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
Cerny et al. (2008) isolated 5-hydroxy-3-mercapto-2-pentanone (19a, Figure 4) as an
instable intermediate during thiamine degradation.25 They also could show this
intermediate occurs in two different tautomeric forms, the open chained (19a) and
closed cyclic form (19b, Figure 4).25 The ratio of the two tautomeric forms in the
underivatized compound (79% open; 21 % cyclic ketal) is similar to the ratio found for
compound 5a/5b.
Since 19a is a degradation product of thiamine, UPLC-TOF-MS analysis confirmed
that this compound is also formed in a sole thiamine reaction system and the presence
of cysteine is not necessary.
Identification of Fraction F6. Due to the high purity of fraction F6 isolated from
the thiamine-cysteine model reaction system no further sub fraction was required.
+
UPLC-TOF-MS analysis in the ESI -mode of F6 revealed m/z 236.0974 as the pseudo
+
+
molecular ion ([M+H] ), and an elemental composition of [C H N OS] . To elucidate
11
14
3
the structure 1D- and 2D-NMR-experiments were performed. Compared to compound
5 the HMBC spectrum of the isolated substance showed also the typical signals of the
aminopyrimidine moiety of thiamine.24 Due to the correlation signals in the HMBC-
spectrum it was possible to confirm that a MFT-motif, which is linked to the
aminopyrimidine moiety at position C(7). The methylene group H-C(7) at 3.64 ppm
showed a correlation signal with the aromatic carbon atom C(3’) of the MFT-moiety at
109.9 ppm. This carbon atom showed further couplings to the two aromatic proton H-
C(4’) and H-C(5’) as well as to the methyl group H-C(6’) at 1.96 ppm. The two aromatic
protons H-C(4’) and H-C(5’) showed couplings in the proton NMR as well as in the H,H
COSY with typical coupling constants of 1.9 Hz for the furanyl moiety. Compared to H-
C(4’) the proton H-C(5’) is shifted to higher frequencies, due to the adjacent oxygen
atom. The long-range coupling of the methyl group H-C(6’) to C(7) confirmed the
20
ACS Paragon Plus Environment

Page 21 of 44
Journal of Agricultural and Food Chemistry
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
linkage of the MFT moiety to the pyrimidine moiety (Figure 3B). Taking all
spectrometric and spectroscopic data into consideration, the compound isolated from
F6 was confirmed as 2-methyl-5-(((2-methylfuran-3-yl)thio)methyl)pyrimidin-4-amine
(6, Figure 1), this compound can be classified as a derivate of thiamine where MFT
(20, Figure 4) substitutes the thiazole moiety (2) of thiamine (1). This compound was
isolated from Jhoo et al. for the first time from a heated thiamine solution, however no
sensory data were reported so far.26
As MFT is a known degradation product of thiamine, UPLC-TOF-MS analysis
confirmed, analogous to 5, that this compound is also formed in a sole thiamine
reaction system and the presence of cysteine is not necessary et all.
Sensory Evaluation of Thiamine Derived Compounds in process flavors.
To evaluate the impact of these compounds in the thiamine enriched PF2. The taste
thresholds of the isolated or commercially available compounds 1-6 were determined
in water as well as in model broth (Table 1).
Thiamine (1) itself showed no modulating activity in model broth, the taste activity was
only described as bitter and the taste threshold in model broth was determined at 1200
µmol/L. The astringent taste threshold of the thiamine degradation product 4-methyl-
5-thiazoleethanol (2) was 10 times lower in water (120 µmol/L) than in model broth
(1200 µmol/L). In addition, the panelists could not detect any modulating activity in
model broth. Literature already describes the kokumi taste modulating effects of
13
compound 3. With a taste threshold of 120 µmol/L in model broth and a 4-fold higher
taste threshold in water this compound could be confirmed as taste modulator. Up to
a concentration of 1000 µmol/L compound 4 did not show any taste activities neither
in model broth nor in water. The isolated compounds 5 and 6 showed strong taste
modulating effects in model broth with taste thresholds of 40 µmol/L and 50 µmol/L,
21
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 22 of 44
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
respectively. The panelists characterized the compounds with a strong kokumi effect,
which was described as long-lasting and mouth-filling. The two compounds did not
show any intrinsic taste activities in water up to a concentration of 710 µmol/L and 280
µmol/L, respectively (Table 1).
The untargeted screening of the PF1 and PF2 confirmed different thiamine conversion
products are higher abundant in the thiamine enriched PF2 compared to PF1.
However, not all conversions products contribute to the higher kokumi activity of PF2.
Only, the thiamine derivates 3, 5 and 6 showed a kokumi activity.
Formation Pathway of 5 and 6. In order to obtain highly kokumi active
products, it is essential to get a better understanding of the formation pathways of
these taste enhancers. The recently published reaction product of cysteine and
13
thiamine (3) is formed via an S 1 reaction of the two precursors. Literature describes
N
the degradation of thiamine to 5-hydroxy-3-mercapto-2-pentanone (19a, Figure 4).
Wherein, the degradation of thiamine to 5-hydroxy-3-mercapto-2-pentanone (19a)
occurs via water addition to the thiazole moiety, which leads to the intermediate 21.
This intermediate then undergoes a ring opening and tautomerization to 22. The
following hydrolysis of 22 leads to the targeted 5-hydroxy-3-mercapto-2-pentanone
(19a) and to the formamide derivate 4 or the amino derivate 12 (Figure 4).23,22 The
degradation product 5-hydroxy-3-mercapto-2-pentanone (19a) was identified from a
thiamine, cysteine and xylose reaction mixture by means of NMR from Cerny et al. in
25
2008. This compound is described as very reactive and an important intermediate. It
serves as precursor for numerous S-containing flavor compounds.15,25 For example it
reacts via its cyclic ketal form (19b) after dehydration and oxidation to MFT (20), a very
important meaty aroma compound (Figure 4).25
22
ACS Paragon Plus Environment

Page 23 of 44
Journal of Agricultural and Food Chemistry
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
As 19a and 20 are constituents of 5a and 6, respectively. The latter substitute the
thiazole part (2) of thiamine (1) and are linked to the pyrimidine motif via a sulfur atom.
As 19a and 20 are degradation products of thiamine themselves, two formation
pathways of 5a and 6 are possible. One is the intramolecular formation by a
rearrangement during the thermal process (Figure 5A). The other possible pathway is
an intermolecular pathway whereby first thiamine (1) degrades to the volatile
compounds 19a and 20 which than react via an S 1-reaction mechanism with thiamine.
N
The cationic intermediate 23 can be formed either from thiamine (1), the formamide
derivate (4) or the amino derivate (12) with 4-methyl-5-thiazoleethanol (2), ammonia
or formamide as leaving group, respectively (Figure 5B).
To elucidate the formation pathway of e.g. Maillard reaction products the so-called
CAMOLA (carbon module labeling) approach is widely used. Where fully labeled and
unlabeled analogues are heated in an equimolar ratio.27–30 Since in the reaction of
thiamine to the target compounds both, the thiazole and aminopyrimidine moiety are
13
involved it is necessary to obtain fully C-labeled thiamine. However, in purchasable
13
products only the thiazole part of thiamine is C-labeled. Therefore, it was necessary
to adapt the CAMOLA experiment. To get a different type of labeling into the pyrimidine
moiety, ethyl thiamine (16, Figure 1) was used. Here the methyl group at position C(2)
is substituted by an ethyl group. The degradation of the thiazole moiety to the sulfur
compounds 5-hydroxy-3-mercapto-2-pentanone (19a) and MFT (20) should not
affected by the alkyl chain in the pyrimidine moiety. Hence, they should be formed by
ethyl thiamine as well as by thiamine. By heating ethyl thiamine and 13C labeled
4
thiamine in an equimolar ratio it should be possible to clarify the reaction mechanism
by mass spectrometry. If the compounds 5 and 6 are formed via an intramolecular
mechanism only four m/z-signals in the mass spectra should be observed (unlabeled
23
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 24 of 44
*
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
ethyl thiamine derivatives (17, 18) and labeled thiamine derivatives (5*, 6 ), Figure5A).
In contrast, if the compounds 5 and 6 are generated by an intermolecular mechanism
labeled and unlabeled 5-Hydroxy-3-mercapto-2-pentanone and MFT would react with
ethyl thiamine and thiamine respectively and as a consequence to eight different m/z-
signals (ethyl thiamine-derivate unlabeled (17,18) and labeled (17*,18*), thiamine
*,
*
derivate unlabeled (5,6) and labeled (5 6 ), Figure5B) should be observed.
13
For this labeling experiment C -labeled thiamine and ethyl thiamine were heated in
4
an equimolar mixture at optimized conditions. A comparison of the pseudomolecular
+
13
ion [M+H] of 5a and 18 showed an enrichment of three C-carbon atoms in the two
derivates (Figure 6 A1&A2), which confirms that the reaction product of thiamine and
5-hydroxy-3-mercapto-2-pentanone, 5 is formed via an intermolecular mechanism
+
(Figure 5B). The comparison of the pseudomolecular ion [M+H] of 6 and 17 also
1
3
showed an incorporation of three C-carbon atoms (Figure 6 B1&B2). Concluding,
that derivate of thiamine and MFT, 6 is also formed via an intermolecular mechanism
(Figure 5B).
That new type of CAMOLA-experiments can be widely adapted to different reaction
mechanism where the reactant can be divided into two different reactive parts and no
fully labeled compounds are available but are required for this type of investigations.
The application is not only limited to model reaction system and can also be applied
for biological systems.
Formation of further Thiamine Derivates. By means of an activity guided
fractionation of a process flavor and untargeted screenings followed by targeted
isolation it was possible to identify several taste modulating compounds, which are
1
3
derived from thiamine (3, 5 and 6). Aimed at isolate further thiamine derived taste
modulating compounds several binary model reactions were performed. The binary
24
ACS Paragon Plus Environment

Page 25 of 44
Journal of Agricultural and Food Chemistry
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
model reactions are preferred over more complex reaction systems (e.g. PFs) because
it facilitates the isolation and structure elucidation step.
Formation of 7 and 8. Since the thiamine derivate of cysteine, 3, appears to be active
a further model reaction of thiamine and the tripeptide glutathione (GSH) was
performed. GSH degrades during heating into cysteinyl glycine and its respective
diketopiperazine (DKP).31,32 Due to this, it was not possible to isolate a direct derivate
of thiamine and GSH but derivatives of thiamine and cysteinyl glycine S-((4-amino-2-
methylpyrimidin-5-yl)methyl)-L-cysteinylglycine (7, Figure 1) and its respective DKP
(S)-3-((((4-amino-2-methylpyrimidin-5-yl)methyl)thio)methyl)piperazine-2,5-dione (8,
Figure 1). Both compounds have not been described in literature so far. Their
structures where confirmed by their characteristic TOF-MS spectra and 1D- and 2D-
NMR experiments.
+
Compound 7 showed a pseudo molecular ion ([M+H] ) at m/z 300.113 Da. By means
of HMBC-experiments it was possible to confirm the structure e.g. via coupling of the
aminopyrimidine moiety with cysteinylglycine. As an example, the carbon atom C(7) at
28.4 ppm showed a coupling to the methylene group H-C(6’) at 3.03 ppm. Compound
+
8 showed a pseudo molecular ion ([M+H] ) at m/z 282.103 Da. The HMBC experiments
confirmed the linkage of the aminopyrimidine moiety with the DKP by the connection
of C(7) with H-C(7’).
The isolated compounds showed a kokumi modulating effect in model broth with a
taste threshold of 880 µmol/L (7) and 255 µmol/L (8) respectively (Table 1). Compared
to the other taste modulating compounds 3, 5, and 8 the taste threshold of 7 was
higher. Both substances were detectable in the two process flavors by LC-MS/MS-
screening. Comparing PF1 and PF2 the derivates increased in the thiamine enriched
25
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 26 of 44
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
process flavor (PF2) from 44 µmol/kg to 81 µmol/kg for 7 and from 43 µmol/kg to 137
µmol/kg for 8, respectively.
Formation of 9. A further model reaction was performed with thiamine and the flavor
active compound 3-mercapto-2-pentanone, which is formed during Maillard reaction
from ribose with cysteine and plays an important role in the aroma profile of meat.14,33
By means of HPLC it was possible to isolate a derivate of thiamine and 3-mercapto-2-
pentanone 3-(((4-amino-2-methylpyrimidin-5-yl)methyl)thio)pentan-2-one (9, Figure
1) from a heated mixture of thiamine and 3-mercapto-2-pentanone. The structure was
confirmed by its TOF-MS spectra and 1D- and 2D-NMR experiments. The UPLC-TOF-
+
MS analysis revealed a pseudo molecular ion ([M+H] ) with 240.117 Da. The NMR
spectra of 9 was similar to the NMR spectra of 5a as they only differ in the hydroxy
1
group at position C(5’). In comparison to the H-NMR spectra of 9 H-C(5’) is shifted to
lower frequencies than in 5a.
In order to evaluate the taste enhancing effects of the compound the taste threshold
was determined in water for the intrinsic taste and in model broth to evaluate the taste
enhancing effects. The compound showed a kokumi effect with a taste threshold of
35 µmol/L in model broth. The taste threshold in water was nearly 7-fold higher than in
model broth (220 µmol/L), confirming the compound as a taste modulating compound
(Table 1). The new discovered taste modulator has not been described in literature so
far. LC-MS/MS screening revealed that the compound was detectable in a sole
thiamine reaction system as well as in PF2 (1 µmol/kg) but not in PF1.
Formation of 10. A further important Maillard reaction product with a roasty aroma is
Furfurylthiol (FFT). This compound is a key aroma compound in meaty dishes and in
process flavors.14,33,34 By means of HPLC it was possible to isolate the derivate of
thiamine and FFT 5-(((furan-2-ylmethyl)thio)methyl)-2-methylpyrimidin-4-amine (10,
26
ACS Paragon Plus Environment

Page 27 of 44
Journal of Agricultural and Food Chemistry
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
Figure 1) from a reaction mixture of thiamine and FFT. The structure of the compound
was confirmed by its TOF-MS spectra and 1D- and 2D-NMR experiments. As 10 is an
+
isomer of 6 it shows the same pseudo molecular ion ([M+H] ) at m/z 236.091 Da. 10
and 6 can be differentiated by their characteristic NMR spectra. The HMBC spectrum
of 10 showed a characteristic coupling of C(7) to the methylene group H-C(6)’. This
proton is further connected with the carbon atoms of the furan ring.
The Sensory evaluations of the compound revealed that it has a kokumi effect in model
broth (taste threshold 120 µmol/L) and is tasteless in water up to a concentration of
330 µmol/L. This new taste modulating compound has not been described in literature
so far. LC-MS/MS screening revealed that the compound was detectable in PF2
(1 µmol/kg) but not in PF1.
Structure-Activity-Relationship. All thiamine derived taste modulating
compounds that have been isolated so far, have in common that a thiol is linked to the
aminopyrimidine moiety of the thiamine. If an amino compound like formamide instead
of the thiol is linked to the aminopyrimidine moiety (4) no taste modulating effects were
detectable. To evaluate the sensory impact of the sulfur, several structural related
compounds, which only differ by the heteroatom were compared. Those derivates were
17
either commercially available or synthesized according to Zheng et al. The hydroxy
derivate (4-amino-2-methylpyrimidin-5-yl)methanol (11, Figure 1) and the amino
derivate 5-(aminomethyl)-2-methylpyrimidin-4-amine (12, Figure 1) were compared to
the respective thiol derivate (4-amino-2-methylpyrimidin-5-yl)methanethiol (13, Figure
1). In order to o check whether the polarity might also play a role the derivate of methyl
amine 2-methyl-5-((methylamino)methyl)pyrimidin-4-amine (14, Figure 1) was
compared to the derivate of thiomethoxide 2-methyl-5-((methylthio)methyl)pyrimidin-4-
amine (15, Figure 1). The sensory evaluation showed that all compounds which
27
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 28 of 44
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
contain a sulfur had in common that they showed a kokumi activity in model broth with
a low taste threshold of 70 - 80 µmmol/L. The taste thresholds of these compounds
were comparable to the taste threshold of the already identified thiamine derived
kokumi substances (35 - 255 µmol/L). The analogous compounds with other
heteroatoms (oxygen or nitrogen) did not show any kokumi effects (Table 1).
Therefore, it can be concluded that the sulfur definitively plays an important role in the
taste perception of the thiamine derived kokumi substances.
As many thiamine derivates show modulating effects, it seemed to be interesting to
know, whether the respective ethyl thiamine derivates also show kokumi effects in
model broth. Therefore, a model reaction of ethyl thiamine (16, Figure 1) and MFT
was performed to isolate the respective ethyl thiamine derivate (17, Figure 1).
Surprisingly, sensory evaluations of this compound revealed that is does not have any
taste modulating properties (Table 1). This leads to the assumption, that the alkyl chain
in the aminopyrimidine moiety plays an important role, during the kokumi taste
perception.
Untargeted screening followed by targeted isolation of a model reaction screening lead
to the identification of new thiamine derived taste modulating compounds 5 and 6 which
have not been described in literature before. By a novel adaption of the classic
CAMOLA-experiment it was possible to confirm the intermolecular formation pathway
of the compounds. Furthermore, binary model reactions with thiamine lead to the
isolation of additional taste modulating compounds. Those reactants, like glutathione,
3-mercapto-2-pentanone or FFT, chosen for the model reactions play a key role in the
development of meaty process flavors. Based on those structure-activity relationships
it was possible to confirm that the sulfur plays an important role in the kokumi taste
perception. The discovery of the new taste modulating compounds gives several
28
ACS Paragon Plus Environment

Page 29 of 44
Journal of Agricultural and Food Chemistry
6
6
6
6
65
66
67
68
possibilities to optimize the savory taste perception of thermally treated model
reactions systems like meaty process flavors.
29
ACS Paragon Plus Environment