Simultaneous generation of acrylamide, β-carboline heterocyclic amines and advanced glycation ends products in an aqueous Maillard reaction model system

Article abstract of DOI:10.1016/j.foodchem.2020.127387

The simultaneous formation of acrylamide; β-carboline heterocyclic amines (HAs): harmane and norharmane; and advanced glycation end products (AGEs) (N^ε-(carboxymethyl)lysine (CML) and N^ε-(carboxyethyl)lysine (CEL)) was analyzed based on an aqueous model system. The model systems included lysine–glucose (Lys/Glu), asparagine–glucose (Asn/Glu), tryptophan–glucose (Trp/Glu), and a mixture of these amino acids (Mix/Glu). Only AGEs were generated when heated at 100 °C, Asn and Trp competed with Lys for glucose and methylglyoxal (MGO), and glyoxal (GO) decreased AGE content. The k value of CML, CEL, and acrylamide decreased when heated at 130 °C, whereas that of harmane increased in the Mix/Glu, owing to the competition between Lys and Asn for glucose, GO, and MGO. Harmane preferably formed via the Pictet–Spengler condensation between Trp and acetaldehyde, which further reduced acrylamide formation via the acrolein pathway.

Full text of DOI:10.1016/j.foodchem.2020.127387

Journal Pre-proofs
Simultaneous generation of acrylamide, β-carboline heterocyclic amines and
advanced glycation ends products in an aqueous Maillard reaction model sys‐
tem
Wei Quan, Yong Li, Ye Jiao, Chaoyi Xue, Guoping Liu, Zhaojun Wang,
Zhiyong He, Fan Qin, Maomao Zeng, Jie Chen
PII:
S0308-8146(20)31249-8
DOI:
Reference:
https://doi.org/10.1016/j.foodchem.2020.127387
FOCH 127387
To appear in:
Food Chemistry
Received Date:
Revised Date:
Accepted Date:
28 March 2020
14 June 2020
17 June 2020
Please cite this article as: Quan, W., Li, Y., Jiao, Y., Xue, C., Liu, G., Wang, Z., He, Z., Qin, F., Zeng, M., Chen,
J., Simultaneous generation of acrylamide, β-carboline heterocyclic amines and advanced glycation ends products
in an aqueous Maillard reaction model system, Food Chemistry (2020), doi: https://doi.org/10.1016/j.foodchem.
2020.127387
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Ltd.

1
2
Simultaneous generation of acrylamide, β-carboline heterocyclic amines
and advanced glycation ends products in an aqueous Maillard reaction
model system
3
4
5
Wei Quana, Yong Lia, Ye Jiaoa, c, Chaoyi Xuea, Guoping Liud, Zhaojun Wang a,
Zhiyong Hea, Fan Qina, Maomao Zeng*, a, Jie Chen*, a, b
6
7
8
a State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi,
Jiangsu 214122, P R China
9
10
11
12
13
14
15
16
b Synergetic Innovation Center of Food Safety and Nutrition, Jiangnan University,
Wuxi, Jiangsu 214122, P R China
c School of Chemistry and Food Engineering, Changsha University of Science &
Technology, Changsha 410114, P R China
d Wuxi People’s Hospital, Nanjing Medical University, Wuxi, Jiangsu 214023, P R
China
*Corresponding author: Prof. Maomao Zeng and Prof. Jie Chen
17
18
E-mail: mmzeng@jiangnan.edu.cn Chenjie@jiangnan.edu.cn
Tel.: +86 510 85919065; Fax: +86 510 85919065
1

20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Abstract
The simultaneous formation of acrylamide; β-carboline heterocyclic amines (HAs):
harmane and norharmane; and advanced glycation end products (AGEs): ) (Nε-
(carboxymethyl)lysine (CML) and Nε-(carboxyethyl)lysine (CEL); )) was analyzed
based on an aqueous model system. The model systems was established to included
lysine–glucose (Lys/Glu), asparagine–glucose (Asn/Glu), tryptophan–glucose
(Trp/Glu), and a mixture of these amino acids (Mix/Glu). Only AGEs were generated
when heated at 100 °C, Asn and Trp competed with Lys for glucose and methylglyoxal
(MGO), and glyoxal (GO) decreased AGE content. The k value of CML, CEL, and
acrylamide decreased when heated at 130 °C, whereas that of harmane increased in the
Mix/Glu, owing to the competition between Lys and Asn for glucose, GO, and MGO.
Harmane preferably formed via the Pictet–Spengler condensation between Trp and
acetaldehyde, which further reduced acrylamide formation via the acrolein pathway.
Keywords: Multiresponse kinetic models; Nε-(carboxymethyl)lysine; Nε-
(carboxyethyl)lysine; acrylamide; harmane; norharmane
2

36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
1. Introduction
The Maillard reaction (MR) is a non-enzymatic browning reaction due to the
complex interaction between free amine groups of amino acids/proteins and carbonyl
group of sugars/carbohydrates under the influence of heat (Aljahdali, & Carbonero,
2017; Starowicz, & Zieliński, 2019). MR plays an important role in food processing,
because it can form desirable flavors and attractive color compounds in food products
(Starowicz, & Zieliński, 2019). However, despite these positive aspects, MR can also
form undesirable, potentially toxic, and harmful compounds (Aljahdali, & Carbonero,
2017). One of the Maillard reaction harmful products (MRHPs) that is widely
researched is acrylamide, a toxin with mutagenic and carcinogenic properties (Zamani
et al., 2017), which was first founded and mainly formed in carbohydrate-rich cereals
and potato-based foods, such as french fries, potato chips, and baked cookies (Stadler
et al., 2002; Zhang, Ren & Zhang, 2009), because asparagine (Asn) is an important
precursor for acrylamide (Friedman, 2003; Yaylayan, Wnorowski, & Perez Locas,
2003).
Moreover, advanced glycation end products (AGEs) are a group of structurally
complex and chemically stable MRHPs generated at the final MR stage (Poulsen et al.,
2013). The most common AGE markers are Nε-(carboxymethyl)lysine (CML) and Nε-
(carboxyethyl)lysine (CEL), which have been detected in various kinds of foods, such
as cookies, tea, meat products, and potato-based foods (Jiao et al., 2017; Yu et al., 2018).
The accumulation of AGEs from the human diet may increase the risk of several chronic
diseases in the human body (Poulsen et al., 2013; Singh et al., 2001). Additionally, β-
3

58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
carboline heterocyclic amines (HAs), including harmane and norharmane, are
carcinogenic and neurotoxic MRHPs (Totsuka et al., 1999), which exist in both heat-
processed animal- and plant-based foods (Chen et al., 2017; Gibis et al., 2016).
The general reaction scheme of MR and its corresponding MRHPs’ formation
mechanisms have has been investigated using aqueous model systems of individual
amino acids and reducing sugars (Martins, Jongen, & Van Boekel, 2000; Amrein et al.,
2006). In detail, tThe formation of acrylamide has been reportedly follows a generic
amino acid route. A Schiff base is formed and then rearranged to yield Amadori
products (AP) when a reducing sugar reacts with any amino acid group, followed by
decarboxylation and heat-mediated elimination of ammonia to substituted imine,
forming acrylamide, with asparagine (Asn) as an important precursor (Yaylayan et al.,
2003; Zhang et al., 2009; Zyzak et al., 2003). Several studies revealed that acrylamide
can be formed via the acrolein pathway, and acrolein was reportedly generated through
ethanalacetaldehyde oxidation (Becalski et al., 2003; Zyzak et al., 2003). Similar to
acrylamide, CML and CEL can be formed through AP formation directly or through a
reaction between lysine (Lys) and reactive α-dicarbonyl compounds, including
methylglyoxal (MGO) and glyoxal (GO), which are generated in the middle MR stage
(Singh et al., 2001; Srey et al., 2010). Moreover, β-carboline HAs (harmane and
norharmane) formation is partly different from that of acrylamide and AGEs. Herraiz
(2000) indicated that tetrahydro-β-carboline (THβC) oxidation is another way to
produce β-carboline HAs, and that the Pictet–Spengler condensation between Trp and
acetaldehyde is the most effective way to form THβCs, except for AP dehydration and
4

80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
β-elimination (Gibis et al., 2016).
The precursors, including Lys, Asn, and Trp, always exist together in some real food
systems, especially for cereals and potato-based foods (Choi et al., 2016). Several
studies have indicated that adding Lys and Trp was effective in inhibiting the formation
of acrylamide in an aqueous MR model system, but AGE and β-carboline HA formation
was not considered (Kim, Hwang, & Lee., 2005; Koutsidis et al., 2009). Furthermore,
in several cereals and potato-based foods, the concentration of reducing sugars is
relatively lower than that of the free amino acids, and the content of each amino acid is
different (Choi et al., 2016). Thus, with such limited amounts of reducing sugars, Lys,
Asn, and Trp competed with each another to participate in the MR, which could show
different kinetic results MRHP formation, compared with that of model systems of
individual amino acids and reducing sugars. However, less attention has been given to
the simultaneous acrylamide, AGE, and β-carboline HA generation in under such
conditions.
Therefore, we aimed to investigate simultaneous acrylamide, β-carboline HA (harmane
and norharmane), and AGE (CML and CEL) formation based on an aqueous model
system with a limited content of reducing sugars and different concentrations of
precursor amino acids, using multi-response kinetic modeling. Compared with single-
response modeling, multi-response modeling can be more rigorously tested, and the
estimation of model parameters can be accurately performed (Martins et al., 2000;
Nguyen & Van Boekel, 2017). The present study would gives an insight into the effects
of precursor amino acids on simultaneous acrylamide, β-carboline HA (harmane and
5

102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
norharmane), and AGE (CML and CEL) formation, which were found to be generated
in cereals and potato-based foods.
2. Materials and methods
2.1 Chemicals
Chemical standards of glucose and 13C6-glucose, amino acids (Lys, d4-Lys, Trp, and
Asn), intermediates (acetaldehyde, acrolein, glyoxal, and methylglyoxal), HAs
(harmane and norharmane), and nonafluoropentanoic acid (NFPA, purity over 97%)
were purchased from J&K Scientific Co., Ltd. (Beijing, China). The standards of CML,
d4-CML, CEL, and d4-CEL were provided by Santa Cruz Biotechnology Co. (Paso
Robles, CA). acrylamide Acrylamide and 13C3-acrylamide were purchased from Sigma-
Aldrich. High-performance liquid chromatography (HPLC)–)-grade methanol and
acetonitrile were obtained from Tedia Company Incorporated (Fairfield, OH, USA).
The other chemicals were all analytical grade and were purchased from Sinopharm
Chemical Reagent Co., Ltd. (Shanghai, China).
2.2 Model systems and reaction conditions
The aqueous model system was established as previously described (Nguyen et al.,
2016; Yu et al., 2020), with slight modifications. Four amino acid–glucose reaction
solutions containing Lys–glucose (Lys/Glu; 30 mmol/100 mmol), Asn–glucose
(Asn/Glu; 200 mmol/100 mmol), Trp–glucose (Trp/Glu; 5 mmol/100 mmol), or a
combination of Lys–Asn–Trp–glucose (Mix/Glu; 100 mmol/200 mmol/5 mmol/100
mmol) were prepared in phosphate buffer (0.1 M, pH 7.0). The Lys, Asn, Trp, and
glucose concentrations selected in the model system were based on their actual
6

124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
proportions in real cereal and potato-based foods (Choi et al., 2016; Halford et al., 2012),
and their concentrations were appropriately magnified to better evaluate the
intermediates and final products. The 10-mL reaction samples were added to glass
reaction vials with Teflon caps (Synthware, Beijing, China) to avoid side phenomena,
such as water evaporation or oil absorption. The samples were pre-heated at 100 °C and
130 °C, respectively, in a thermostat-controlled oil bath with a magnetic stirrer (Jintan,
Zhejiang, China) for 5 min. The purpose of pre-heating is to raise the temperature of
the samples to the intended temperature level. The temperature of the oil bath was
checked every 1 min, and the temperature fluctuations were within 3 °C. The reaction
time was recorded after the pre-heated treatment. Samples were taken at each
predetermined heating time (3, 6, 9, 12, 15, 18, and 21 min), the vials were immediately
cooled in an ice bath to stop any further reaction, and then they were stored at −20 °C
prior to analysis. The model system experiments were performed in triplicate.
2.3 Analysis of melanoidins
Melanoidins were measured by using an UV-Vis spectrophotometer (UV-1601, ;
Shimadzu, Kyoto, Japan) with a path length of 1 cm at 470 nm, and then the samples
were diluted. The results were calculated using the Lambert––Beer equation with an
extinction coefficient of 282 L mol−1cm−1, as reported by Zhang et al. (2015), with the
following equation:
A = ε × bc
where, A is the absorbance value at 470 nm, ε is an extinction coefficient, b is the
thickness of the container’s absorbent layer, and c is the melanoid molar concentration.
7

146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
2.4 Analysis of CML and CEL
CML and CEL sample analysis were performed according to the methods previously
described by Jiao et al (2017), and Yu et al (2018). Each 150-μL reaction sample was
mixed with 150 μL d4-CML and d4-CEL (internal standards), followed by
centrifugation at 10,000 rpm for 10 min and filtration through a 0.22-μm membrane.
The obtained filtrate was separated using Waters 2695 liquid chromatography (LC)
module with a Waters X-Bridge C18 column at 35 °C (2.1 mm × 100 mm; 3.5 μm;
Milford, MA, USA). The mobile phase consisted of acetonitrile as solvent A and 5 mM
NFPA as solvent B, and gradient elution with a flow rate of 0.3 mL/min was performed
as follows: 0–5 min, 5%–60% A; 5–7 min, 60%–100% A; 7–9 min, 100% A; 9–10 min,
100%–5% A; and 10–16 min, 5% A. Injection volume was set to 5 μL, and detection
was performed using a Micromass Quattro micro triple quadrupole mass spectrometer
(Manchester, UK) using the electrospray ionization positive mode with the multiple
reaction monitoring (MRM) model. The capillary voltage was 3.5 kV, and the
ionization source and desolvation gas temperatures were 110 °C and 400 °C,
respectively. The nitrogen flow rates were 50 and 600 L/h for the cone gas and
desolvation gas, respectively, whereas the cone and collision energies were 20 V and
18 V, respectively. For CML, d4-CML, CEL, and d4-CEL, the fragments m/z 205→84,
m/z 209→88, m/z 219→84, and m/z 223→88 were used for quantification, respectively.
CML and CEL quantification was done using isotope internal standard method.
2.5 Analysis of GO and MGO
The α-dicarbonyl compound derivatization was performed with 100 μL of 5 μM o-
8

168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
phenylenediamine, as previously described (Scheijen & Schalkwijk, 2014). Reaction
samples were centrifuged at 10,000 ×g for 10 min, and the 200-μL supernatant was
derivatized, then 100 μL 2,3-hexanedione (4 μM) was added as the internal standard
and immediately mixed and kept at 4 °C in the dark for 12 h prior to HPLC–MS/MS
measurement.
According to Jiao et al (2019) the quinoxaline derivatives of the α-dicarbonyl
compounds were determined using Waters 2695 LC system interfaced with a
Micromass Quattro micro triple quadrupole mass spectrometer. Separation was
performed on a Waters X-Bridge C18 column (2.1 mm × 100 mm; 3.5 μm) using a
gradient mixture of (A) 1% formic acid in water and (B) methanol at 0.3- mL/min flow
rate. The elution gradient was as follows: 0–5 min, 30%–90% A; 5–8 min, 90%–100%
A; 8–10 min, 100% A; 10–12 min, 100%–30% A; and 12–15 min, 30% A. The α-
dicarbonyl compound identification was performed using the MRM mode with the
same MS parameters as those used in CML and CEL detection. The fragments m/z of
the quinoxaline derivatives of α-dicarbonyl compounds were m/z 131→77 (GO),
145→77 (MGO), and 187→77 (internal standard).
2.6 Simultaneous determination of acrylamide, harmane, norharmane, amino
acids, and glucose using UHPLC-MS/MS
The reaction sample (150 μL) was added to the 150- μL internal standard solution
containing 200 μg/mL 13C3-acrylamide, 500 μg/mL d4-Lys, and 500 μg/mL 13C6-
glucose, centrifugated at 10,000 rpm for 10 min, and filtered through a 0.22-μm
membrane. The obtained filtrate was collected for UHPLC-–MS analysis as previously
9

190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
described by Zhang et al (2011), with slight modifications.
The identification and quantification were conducted using an ACQUITY UHPLC
system equipped with a triple quadrupole mass spectrometer (Waters, Milford, MA,
USA). Analyte separation was performed on a Waters Atlantics dC18 column (250 mm
× 4.6 mm i.d.; 3.0 μm particle size) at 30 °C. The gradient elution was obtained with a
binary mobile phase of 0.1% formic acid (pH 6.8; solvent A) and acetonitrile (solvent
B). The solvent composition was 0–0.1 min, 5% A; 0.1–10 min, 5%–100% A; 10–10.5
min, 100%−5% A; and 10.5–15 min, 5% A. The flow rate was set to 0.5 mL/min, with
a 5-μL injection volume. Mass spectrometric detection was performed in the MRM
mode. The capillary voltage was 3.5 kV, and the ion source and desolvation gas
temperatures were 130 °C and 350 °C, respectively. The fragments used for
quantification were m/z 133→74 (Asn), 147→84 (Lys), 151→88 (d4-Lys), 205→188
(Trp), 72→55 (acrylamide), 75→58 (13C3-acrylamide), 151→88 (d4-Lys), 169→115
(norharmane), 183→115 (harmane), 383→203 (glucose), and 395→209 (13C6-glucose).
2.7 HPLC analysis of acrolein
By using a previously reported model with slight modifications (Barman, 2014;
Uchiyama, Inaba & Kunugita, 2011), a 2.5-mL reaction sample was placed in a 10-mL
centrifugation tube. Afterward, 1 mL 2% NaCl–0.7 mol/L HCl solution, 0.5 mL 18%
Na2HPO4, and 0.5 mL 0.2% 2,4-dinitrophenylhydrazine solution were sequentially
added. Water was added to a volume of 10 mL, and the contents were mixed through
shaking before heating in a water bath at 60 °C for 20 min. Upon cooling to room
temperature, 3 mL of hexane was were twice added for acrolein extraction, and 1 mL
10

212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
of hexane was filtered through a 0.22-μm organic microporous membrane and subjected
to HPLC analysis.
A LiChrospher® C18 (4.6 mm × 250 mm; , 5.0 μm) column was used to separate
2,4-dinitrophenylhydrazones. The injection volume was set to 10 μL, and the column
temperature was maintained at 30 °C. The solvent flow rate was 1.0 mL/min under an
isocratic water–acetonitrile elution (38:62; v/v) for 20 min. The acrolein derivatives
(2,4-dinitrophenylhydrazones) were detected using a UV detector at 365 nm.
2.8 Headspace gas chromatography analysis of acetaldehyde
According to Qin et al (2020), acetaldehyde analysis was performed with an HS-10
automated headspace sampler (Shimadzu, Kyoto, Japan) and GC-2010 gas
chromatography (Shimadzu, Kyoto, Japan), with a flame ionization detector and a DB-
23 capillary column (i.d. 60 m × 0.32 mm i.d., × 0.25 μm film thickness; 60 m) (;
Agilent Technologies, California, USAFolsom, CA). The 3 mL reaction solution
aliquot was added to a 20-mL vial, sealed, and then heated at 75 °C for 30 min on the
headspace sampler. The transfer line temperature was 120 °C. The gas flow was set to
2.0 mL/min, and the detector and injection temperatures were 260 °C and 205 °C,
respectively. The column was maintained at 35 °C for 6 min and was heated to 185 °C
at an incremental rate of 10 °C/min. Furthermore, the analysis was set to 1:1 split mode.
2.9 Method validation
The proposed method’s linearity, recovery, limit of detection, and limit of
quantification were evaluated. Calibration curve was established by plotting the peak
area ratios or only the peak area of the standards to the internal standard against the
11

234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
eight analyte concentrations. Sample preparation recoveries were determined by
analyzing blank samples spiked with two concentrations of mixed low and high
standards, and they were calculated as [(levels in spiked samples − levels in blank
samples)/amount added] × 100. The overall method’s precision was determined by
implementing intra-day (repeatability) and inter-day (reproducibility) precision
experiments. Intra-day precision was evaluated at a low spiked level through a recovery
study with three replicates, whereas inter-day precision was evaluated at a high spiked
level through a recovery study that was consecutively carried out for 3 days.
2.10 Multi-response kinetic modeling
A comprehensive reaction mechanism was built, comprising the major formation
pathways of the MR and caramelization (see Fig. 5). The average duplicate analysis
data results were normalized and expressed in millimoles per liter (mmol/L). All
experiment data were made to fit the proposed kinetic model, and the kinetic parameters
(k) were estimated using the non-linear curve fitting function of the Origin 9.0 software
(OriginLab, Northampton, MAUSA) and SAS 8.0 software (SAS Institute Inc., China),
and the R-squared (R2) values were calculated to determine the fitting performance of
the non-linear fitted models (Yu et al., 2020; Zhang et al., 2015). Furthermore, the
experiments were performed at only one or two temperatures; hence, specific activation
energies cannot be estimated.
2.11 Statistical analysis
The component analysis was conducted in triplicate, with values reported as the mean
± standard deviation of three independent experiments. Statistical analysis was
12

256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
performed using the general linear model protocol of Statistix 9.0 software (Analytical
Software, Tallahassee, FL, USA). Analysis of variance of the results and subsequent
least significant difference post hoc test were performed to determine significant
differences (p < 0.05) among different MR model systems.
2.11 Statistical analysis
The analysis of all components was conducted in triplicate, with values reported as
the mean ± standard deviation (SD) of three independent experiments. Statistical
analysis was performed using the general linear model procedure of Statistix software
9.0 (Analytical Software, Tallahassee, FL, USA). Analysis of variance of the results
followed by the least significant difference post hoc test was performed to determine
whether if there were significant differences (p < 0.05) among different MR model
system.
3. Results and discussion
3.1 Reactant concentration during heating
Figure 1 shows the amino acid and glucose kinetic profiles during heating. As
expected, amino acid and glucose concentrations decreased with increased heating time,
which was consistent with a previous report (Zyzak et al., 2003). As shown in Figure
1A, the final Trp, Lys, and Asn concentrations in the Trp/Glu, Lys/Glu, and Asn/Glu
models decreased by 20.7%, 29.1%, and 33.3%, respectively, when heated at 100 °C,
compared with their initial concentrations; however, that concentrations in the Mix/Glu
model were reduced by 6.0%, 19.3%, and 30.9%, respectively, compared with their
13

278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
initial concentrations. Trp, Lys, and Asn concentrations in the Trp/Glu, Lys/Glu, and
Asn/Glu decreased by 79.2%, 56.1%, and 52.0%, respectively, when the heating
temperature changed to 130 °C (Fig. 1B). However, Trp, Lys, and Asn concentrations
in the Mix/Glu model were only reduced by 10.9%, 38.7%, and 50.1%, respectively.
The results suggest that Trp and Lys degradation in the Mix/Glu model was
significantly less than that for the Trp/Glu and Lys/Glu models. However, Asn
consumption in the Asn/Glu and Mix/Glu models showed insignificant differences.
Glucose concentration was monitored during heating (Figs. 1C and 1D). Glucose
consumption levels in the Asn/Glu, Mix/Glu, Lys/Glu, and Trp/Glu models were only
24.1%, 28.1%, 15.7%, and 15.8% of their initial concentrations, respectively, when
heated at 100 °C. However, the glucose consumption in each of the Trp/Glu, Asn/Glu,
Mix/Glu, and Lys/Glu models increased by 66.1%, 73.2%, 95.5%, and 63.1%,
respectively, when heated at 130 °C. Glucose reduction levels in the Trp/Glu and
Lys/Glu models were significantly less than that those for the Asn/Glu and Mix/Glu
models.
3.2 Formation of acrylamide and HAs during heating
The formation of acrylamide was measured and shown in Figure 2A. There was no
acrylamide produced at 100 °C due to the low reaction temperature. This finding is
consistent with previous literature (Zhang et al., 2009). When heated at 130 °C,
acrylamide was generated after 6 min at 130 °C. The content of acrylamide in the
Asn/Glu and Mix/Glu models reached a maximum of 22.9 ± 2.55 μg/mL and 12.6 ±
2.80 μg/mL, respectively. In comparison, the production of acrylamide was
14

300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
significantly reduced by 45.2% (p < 0.05) when Trp and Lys were involved in the
reaction. This is consistent with previous studies by Koutsidis (2009) and Mestdagh et
al (2008), who reported that Trp and Lys may reduce the concentration of acrylamide.
Two HAs, harmane and norharmane, were undetectable when the model mixtures
were heated at 100 °C (Fig. 2A). Several published studies have reported that HAs were
generated at temperatures generally higher than 125 °C (Gibis, 2016). When the models
were heated at 130 °C, norharmane could be detected after 6 min, and its highest
concentrations reached in the Trp/Glu and Mix/Glu models were 27.9 ± 2.03 ng/mL
and 22.6 ± 1.74 ng/mL, respectively. Norharmane content in the Trp/Glu model was
significantly higher (p < 0.05) than that in the Mix/Glu model; however, harmane was
detected only after 15 min of heating time in the Trp/Glu and Mix/Glu models (Fig.
2B). Harmane content in the Trp/Glu model ranged from 13.5 ± 0.90 to 13.9 ± 0.68
ng/mL and was significantly increased (p < 0.05) by 6.6%–13.1% in the Mix/Glu model,
which was exactly in contrast to the norharmane results.
3.3 Formation of AGEs and melanoidins in the model systems
As shown in Figure 2D and 2E, CML and CEL were generated at both 100 °C and
130 °C, respectively. CML contents in the Mix/Glu and Lys/Glu models were 5.11 ±
1.03 μg/mL and 11.6 ± 2.21 μg/mL, respectively, when heated at 100 °C, whereas
thatose of CML in the same models were 9.06 ± 1.89 μg/mL and 27.2 ± 2.31 μg/mL,
respectively, when heated at 130 °C. Furthermore, CEL content was 8.30 ± 3.15 μg/mL
for the Mix/Glu model and 10.6 ± 0.49 μg/mL for the Lys/Glu model when heated at
100 °C. Moreover, it significantly increased to 31.9 ± 8.09 μg/mL and 51.7 ± 8.74
15

322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
μg/mL for the Mix/Glu and Lys/Glu models, respectively, when heated at 130 °C. These
results show that CML and CEL contents generated at 130 °C were higher than their
contents at 100 °C. CML formation in the Mix/Glu model was significantly lower (p <
0.05) than that for the Lys/Glu model.
Melanoidins, generated by aldol condensation and carbonyl compound
polymerization, were are considered the main products in the final MR stage (Martins
et al., 2000; Wang, Qian & Yao, 2011). Generally, melanoidin generation in all groups
exhibited a minimal increase during heating (Fig. 2F and 2G), and melanoidin content
reached a maximum at the end of the reaction. The highest melanoid concentrations
was were observed in the Mix/Glu model (at 3.37 ± 0.94 mmol/L and 30.8 ± 0.75
mmol/L) when melanoidins were heated at 100 °C and 130 °C, whereas the lowest
concentrations was were found in the Trp/Glu model.
3.4 Formation of α-dicarbonyl compounds
The α-dicarbonyl compounds, including GO and MGO, are highly reactive
intermediates that play an important role in MR (Martins et al., 2000). They are mainly
produced by thermal glucose degradation and Schiff base decarboxylation, which reacts
with amino acids to form a series of MRHPs (Chen & Kitts, 2011). Figure 3 shows the
GO and MGO amounts that were detected. A higher GO concentration can be found at
130 °C than at 100 °C. The GO concentration in different model systems increases with
reaction time, except for the Trp/Glu and Lys/Glu models, which showed a maximum
GO concentration at the initial stage, but significantly decreased when heated at 100 °C
and 130 °C, respectively. A comparison of the different model systems showed that the
16

344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
Lys/Glu model has the highest GO contents when heated at both 100 °C and 130 °C,
with a range of 3.01 ± 0.50 to 7.38 ± 1.54 μg/mL and 8.33 ± 1.81 to 35.1 ± 0.43 μg/mL,
respectively, followed by the Mix/Glu model.
A significant increase in MGO concentration was observed with increased reaction
time, reaching a relatively high value at the end of the reaction (Figs. 3B and 3D).
Similar to GO, higher temperatures induced easier MGO formation of MGO. MGO
generation in the different model systems varied relatively. . The MGO levels produced
in the Mix/Glu and Lys/Glu models were higher than that those in other systems when
heated at 100 °C. However, the Mix/Glu and Asn/Glu models showed the highest MGO
content when heated at 130 °C.
3.5 Formation of aldehyde compounds
Aldehydes are also important MR intermediates and are known as glucose
degradation products at high temperatures (Martins et al., 2000; Van Boekel, 2006).
Lys/Glu had higher reactivity levels and generated more ethanalacetaldehyde than the
other models at the two heating temperatures (Figs. 4C and 4D). The
ethanalacetaldehyde contents in the Lys/Glu model system was were 3.02 ± 0.12 μg/mL
at 100 °C and 13.1 ± 1.13 μg/mL at 130 °C. Moreover, Asn/Glu also produced a higher
ethanalacetaldehyde content of 12.2 ± 1.53 μg/mL when heated at a higher temperature.
Acrolein is an important intermediate compound that can form acrylamide via the
acrolein pathway, which is mainly produced by ethanalacetaldehyde oxidation
(Becalski et al., 2003; Zyzak et al., 2003). The acrolein content showed an ia non-
significant difference among the model systems when heated at 100 °C (Fig. 4A).
17

366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
However, the highest acrolein concentration was observed in the Lys/Glu and Trp/Glu
models, which ranged from 0.06 ± 0.01 to 1.29 ± 0.12 μg/mL and from 0.20 ± 0.09 to
1.25 ± 0.09 μg/mL, respectively, when heated at 130 °C (Fig. 4B).
3.6 Validation of the proposed analytical methods
The characteristics of the proposed method were demonstrated by a validation
procedure of the linearity, LOD, LOQ, recovery, and precision of the method. The
linearity was evaluated by injecting different concentrations of standards and plotting
the peak areas versus the concentrations (Yan et al., 2014). As shown in Supplemental
Tables 1 and 2, the detector response for individual compounds was linear over a broad
concentration range, with coefficients higher than 0.991. The LODs and LOQs, which
are respectively defined as the amount of an analyte producing a signal-to-noise ratio
of ≥3 and ≥10, were determined using standards. The recoveries of compounds in the
model system ranged from 87.1% to 116% at a low spiked level and from 83.1% to 123%
at a high spiked level. The precisions, which are expressed as relative standard
deviations (RSDs) were lower than 15.3% (14.6%) for intraday (interday) studies. By
summarizing all of these method validation results, the developed method proposed
here can be regarded a potent quantification method that could be successfully applied
to the analysis of reactants, products, and intermediate compounds in a Maillard
reaction model system corresponding to the kinetic study, with appropriate recoveries
and excellent method repeatability.
3.7 Reaction kinetic modeling and related parameters
Referring to the general mechanisms of MRHPs reported in previous literature (Gibis
18

388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
et al., 2016; Srey et al., 2010; Yaylayan et al., 2003; Zyzak et al., 2003), the reaction
scheme was simplified into an apparent scheme (Fig. 5) by keeping only the reactants
(Lys, Trp, Asn, and glucose), aldehydes, and ketone intermediates (acrolein,
acetaldehyde, GO, and MGO), and final products (CML, CEL, harmane, norharmane,
acrylamide, and melanoidins). Due to the different reaction rate constants, several
intermediates showed higher uncertainty and, thus, could not be estimated in this
interval, such as AP which could not be quantified analytically. Kinetic data were fitted,
and the related kinetic parameters (k1–k11) were calculated by the non-linear least
squares regression, according to previous studies (Yu et al., 2020; Zhang et al., 2015).
R2 values were calculated to evaluate the model-fitting performance. As shown in Table
1, most models showed high R2 values, ranging from 0.80–0.99, which reflected good
fitting to the proposed reaction scheme.
Table 1 shows the k values for the formation and elimination of the reactants,
intermediates, and MRHPs. The k values generally increased with temperature in all
the model systems, which may be due to the enhanced molecular mobility at high
temperatures, and this was consistent with the general rule of kinetics (Martins et al.,
2000). Mel formation in all models showed the highest k value. Furthermore, owing to
the lower heating temperature (100 °C), acrylamide, harmane, and norharmane were
not detected; hence, the values of k5, k8, k9, and k10 were considered zero.
Simultaneously, comparing the rate constants of the different model systems at 100 °C
revealed that another k value in the Lys/Glu model was significantly higher than that
for the other model systems, although the k11 value of the Lys/Glu model is lower than
19

410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
that of the Mix/Glu model. The results indicated that the MR in the aqueous model
system mainly involved a reaction with Lys and glucose to form CML and CEL when
heated at 100 °C, which may be due to the high activity of Lys in the MR, dependingent
on the number of carbon (C) atoms and its epsilon amino group in the amino acid
molecule (Claeys et al., 2005; De Sa et al., 2010). Moreover, compared with the
Lys/Glu model, the lower k values for GO, MGO, CML, and CEL (k1, k2, k3, and k4) in
the Mix/Glu model indicated that the decreased CML and CEL concentrations in the
Mix/Glu may be attributed to Asn and Trp competing with Lys for the GO and MGO
in the MR when heated at 100 °C.
The comparison of kinetic parameters between the Lys/Glu, Asn/Glu, and Mix/Glu
models when heated at 130 °C revealed that the k values for CML, CEL, and acrylamide
(k3, k4, and k5) and those for α-dicarbonyl compounds (k1 and k2) in the Mix/Glu model
system were significantly lower (p < 0.05) than that those in the Lys/Glu and Asn/Glu
model systems. Previous studies found that adding Lys to the Asn/Glu model can
significantly reduce the acrylamide production due to the high Lys activity in the MR
from competing with Asn for the reducing sugar (Koutsidis et al., 2009; Mestdagh et
al., 2008). However, those studies have not considered CML and CEL formation,
except for the competitive inhibition between Lys and Asn on acrylamide formation.
Several studies have reported that Lys could decrease acrylamide formation via adduct
formation (Claeys et al., 2005; Friedman et al., 2003). These adducts are due to the
electrophilic double bond of acrylamide, which can participate in nucleophilic reactions
with active hydrogen-bearing functional groups, such as the –NH2 group of Lys.
20

432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
Furthermore, the adduct reaction between Lys and acrylamide not only decreased
acrylamide formation but may also reduce CML and CEL generation by Lys
consumption. Furthermore, the significantly decreased CML, CEL, and acrylamide
formation may be due to their competition for α-dicarbonyl compounds between Asn
and Lys (particularly GO) (Amrein et al., 2006).
The k values for norharmane (k10) and acrylamide (k8) in the Mix/Glu model were
significantly reduced, compared with the Trp/Glu and Asn/Glu models. Previous
studies reported that Trp may compete with Asn for reducing sugars, and a non-covalent
interaction occurred between acrylamide and Trp (Koutsidis et al., 2009). Therefore,
adding Trp to the Asn/Glu model can significantly reduce acrylamide production
(Koutsidis et al., 2009). However, a significant increase in harmane content that was
generated in the Mix/Glu model was observed. Table 1 shows that acetaldehyde
formation (k6) in the Mix/Glu model was significantly increased, compared with the
Trp/Glu model, which contributed to harmane formation, because acetaldehyde is an
important intermediate for β-carboline HA formation. The details of this reaction show
that acetaldehyde can be used to form THβC through Pictet–Spengler condensation,
and can further generate harmane through THβC oxidation (Herraiz, 2000). The k
values of acetaldehyde and acrolein in the Mix/Glu model were significantly decreased,
compared with the Asn/Glu model, and the resulting k value of acrylamide (k8)
produced by the acrolein oxidation pathway was significantly reduced. Therefore, we
speculated that the competition for acetaldehyde between Asn and Trp and further
inhibition of acrolein formation may be the reason for the increased harmane content
21

454
455
and decreased acrylamide content in the Mix/Glu model when heated at 130 °C.
4. Conclusions
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
In this study, aqueous MR model systems, including Asn/Glu, Lys/Glu, Trp/Glu, and
Mix/Glu, were established to investigate simultaneous acrylamide, CML, CEL,
harmane, and norharmane formation. The results indicated that only CML and CEL
were generated when heated at 100 °C, and they would compete with Lys for GO and
MGO in the presence of Asn and Trp, thus leading to significantly decreased CML and
CEL concentrations. acrylamideAcrylamide, CML, and CEL were similarly reduced
when heated at 130 °C because of the competition between Asn and Lys for glucose
and α-dicarbonyl compounds. However, harmane levels were increased through Pictet–
Spengler condensation when Trp reacted with acetaldehyde in the Mix/Glu system.
Moreover, it would further reduce acrolein formation by consuming acetaldehyde, thus
inhibiting acrylamide formation via the acrolein oxidation pathway. Therefore, this
study revealed the effects of Asn, Lys, and Trp on simultaneous AGEs, β-carboline
HAs, and acrylamide generation. Further studies are warranted to examine the obtained
insights regarding the complex interactions and key factors that affect MRHP
generation in real food systems, which in turn will influence the identification of
mitigation strategies during food processing.
Acknowledgements
The National Key R & D Program of China (Grant No.2017YFC1600405, Grant No.
22

475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
2018YFC1600405). The Key Research and Development Plan of Guangdong Province,
China (Grant No.2019B020212011). The National First-class Discipline Program of
Food Science and Technology (No. JUFSTR20180201).
References
Aljahdali, N., & Carbonero, F. (2017). Impact of Maillard reaction products on nutrition and health:
Current knowledge and need to understand their fate in the human digestive system. Critical
Review in Food Science and Nutrition, 1-14.
Amrein, T. M., Andres, L., Manzardo, G. G., & Amadò, R. (2006). Investigations on the promoting
effect of ammonium hydrogencarbonate on the formation of acrylamide in model
systems. Journal of Agricultural and Food Chemistry, 54(26), 10253-10261.
Barman, B. N. (2014). Accurate determination of aldehydes in amine catalysts or amines by 2, 4-
dinitrophenylhydrazine derivatization. Journal of Chromatography A, 1327, 19-26.
Becalski, A., Lau, B. P. Y., Lewis, D., & Seaman, S. W. (2003). Acrylamide in foods: Occurrence,
sources, and modeling. Journal of Agricultural and Food Chemistry, 51(3), 802-808.
Chen, X. M., & Kitts, D. D. (2011). Identification and quantification of α-dicarbonyl compounds
produced in different sugar-amino acid Maillard reaction model systems. Food Research
International, 44(9), 2775-2782.
Choi, S. H., Kozukue, N., Kim, H. J., & Friedman, M. (2016). Analysis of protein amino acids, non-
protein amino acids and metabolites, dietary protein, glucose, fructose, sucrose, phenolic, and
flavonoid content and antioxidative properties of potato tubers, peels, and cortexes
(pulps). Journal of Food Composition and Analysis, 50, 77-87.
23

497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
Claeys, W. L., De Vleeschouwer, K., & Hendrickx, M. E. (2005). Effect of amino acids on
acrylamide formation and elimination kinetics. Biotechnology Progress, 21(5), 1525-1530.
De Sa, P. F., Treubig Jr, J. M., Brown, P. R., & Dain, J. A. (2001). The use of capillary
electrophoresis to monitor Maillard reaction products (MRP) by glyceraldehyde and the epsilon
amino group of lysine. Food Chemistry, 72(3), 379-384.
Friedman, M. (2003). Chemistry, biochemistry and safety of acrylamide. A review. Journal of
Agricultural and Food Chemistry, 51, 4504−4526.
Gibis, M. (2016). Heterocyclic aromatic amines in cooked meat products: causes, formation,
occurrence, and risk assessment. Comprehensive Reviews in. Food Science and Food
Safety, 15(2), 269-302.
Halford, N. G., Muttucumaru, N., Powers, S. J., Gillatt, P. N., Hartley, L., Elmore, J. S., & Mottram,
D. S. (2012). Concentrations of free amino acids and sugars in nine potato varieties: effects of
storage and relationship with acrylamide formation. Journal of Agricultural and Food
Chemistry, 60(48), 12044-12055.
Herraiz, T. (2000). Analysis of the bioactive alkaloids tetrahydro-β-carboline and β-carboline in
food. Journal of Chromatography A, 881(1-2), 483-499.
Jiao, Y., He, J., Li, F., Tao, G., Zhang, S., Zhang, S., Fan, Q., Zeng, M., & Chen, J. (2017). Nε-
(carboxymethyl) lysine and Nε-(carboxyethyl) lysine in tea and the factors affecting their
formation. Food Chemistry, 232, 683-688.
Jiao, Y., Quan, W., He, Z., Gao, D., Qin, F., Zeng, M., & Chen, J. (2019). Effects of Catechins on
N ε-(Carboxymethyl) lysine and N ε-(Carboxyethyl) lysine Formation in Green Tea and Model
Systems. Journal of Agricultural and Food Chemistry, 67(4), 1254-1260.
24

519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
Kim, C. T., Hwang, E., & Lee, H. J. (2005). Reducing Acrylamide in Fried Snack Products by
Adding Amino Acids. Journal of Food Science, 70(5), 354-358.
Koutsidis, G., Simons, S. P., Thong, Y. H., Haldoupis, Y., Mojica-Lazaro, J., Wedzicha, B. L., &
Mottram, D. S. (2009). Investigations on the effect of amino acids on acrylamide, pyrazines, and
Michael addition products in model systems. Journal of Agricultural and Food Chemistry, 57(19),
9011-9015.
Martins, S. I., Jongen, W. M., & Van Boekel, M. A. (2000). A review of Maillard reaction in food
and implications to kinetic modelling. Trends in Food Science and Technology, 11(9-10), 364-
373.
Mestdagh, F., Maertens, J., Cucu, T., Delporte, K., Van Peteghem, C., & De Meulenaer, B. (2008).
Impact of additives to lower the formation of acrylamide in a potato model system through pH
reduction and other mechanisms. Food Chemistry, 107(1), 26-31.
Nguyen, H. T., Van Der Fels-Klerx, H. J., & van Boekel, M. A. (2016). Kinetics of Nε-
(carboxymethyl) lysine formation in aqueous model systems of sugars and casein. Food
Chemistry, 192, 125-133.
Nguyen, H. T., & Van Boekel, M. A. J. S. (2017). Acrylamide and 5-hydroxymethylfurfural
formation during biscuit baking. Part II: Effect of the ratio of reducing sugars and
asparagine. Food Chemistry, 230, 14-23.
Poulsen, M. W., Hedegaard, R. V., Andersen, J. M., de Courten, B., Bügel, S., Nielsen, J., &
Dragsted, L. O. (2013). Advanced glycation endproducts in food and their effects on health. Food
and Chemical Toxicology, 60, 10-37.
Qin, Y., Shin, J. A., & Lee, K. T. (2019). Determination of acetaldehyde, methanol and fusel oils in
25

541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
distilled liquors and sakès by headspace gas chromatography. Food Science and Biotechnology,
1-7.
Scheijen, J. L., & Schalkwijk, C. G. (2014). Quantification of glyoxal, methylglyoxal and 3-
deoxyglucosone in blood and plasma by ultra performance liquid chromatography tandem mass
spectrometry: evaluation of blood specimen. Clinical Chemistry and Laboratory Medicine, 52(1),
85-91.
Singh, R., Barden, A., Mori, T., & Beilin, L. (2001). Advanced glycation end-products: A review.
Diabetologia, 44, 129−146.
Srey, C., Hull, G. L., Connolly, L., Elliott, C. T., Del Castillo, M. D., & Ames, J. M. (2010). Effect
of inhibitor compounds on N ε-(carboxymethyl) lysine (CML) and N ε-(carboxyethyl) lysine
(CEL) formation in model foods. Journal of Agricultural and Food Chemistry, 58(22), 12036-
12041.
Stadler, R. H., Blank, I., Varga, N., Robert, F., Hau, J., Guy, P. A., Robert, M.C., & Riediker, S.
(2002). Food chemistry: acrylamide from Maillard reaction products. Nature, 419, 449−450.
Starowicz, M., & Zieliński, H. (2019). How Maillard reaction influences sensorial properties (color,
flavor and texture) of food products?. Food Reviews International, 35(8), 707-725.
Totsuka, Y., Ushiyama, H., Ishihara, J., Sinha, R., Goto, S., Sugimura, T., & Wakabayashi, K.
(1999). Quantification of the co-mutagenic β-carbolines, norharmane and harmane, in cigarette
smoke condensates and cooked foods. Cancer letters, 143(2), 139-143.
Uchiyama, S., Inaba, Y., & Kunugita, N. (2011). Derivatization of carbonyl compounds with 2, 4-
dinitrophenylhydrazine and their subsequent determination by high-performance liquid
chromatography. Journal of Chromatography B, 879(17-18), 1282-1289.
26

563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
Van Boekel, M. A. J. S. (2006). Formation of flavour compounds in the Maillard
reaction. Biotechnology Advance, 24(2), 230-233.
Wang, H. Y., Qian, H., & Yao, W. R. (2011). Melanoidins produced by the Maillard reaction:
Structure and biological activity. Food Chemistry, 128(3), 573-584.
Yan, Y., Zeng, M. M., Zheng, Z. P., He, Z. Y., Tao, G. J., Zhang, S., Gao, Y., & Chen, J. (2014).
Simultaneous Analysis of PhIP, 4′-OH-PhIP, and Their Precursors Using UHPLC–
MS/MS. Journal of Agricultural and Food Chemistry, 62(48), 11628-11636.
Yaylayan, V. A., Wnorowski, A., & Perez Locas, C. (2003). Why asparagine needs carbohydrates
to generate acrylamide. Journal of Agricultural and Food Chemistry, 51(6), 1753−1757.
Yu, H., Zhong, Q., Xie, Y., Guo, Y., Cheng, Y., & Yao, W. (2020). Kinetic study on the generation
of furosine and pyrraline in a Maillard reaction model system of D-glucose and L-lysine. Food
Chemistry, 126458.
Yu, L., Chai, M., Zeng, M., He, Z., & Chen, J. (2018). Effect of lipid oxidation on the formation of
Nε-carboxymethyl-lysine and Nε-carboxyethyl-lysine in Chinese-style sausage during
storage. Food Chemistry, 269, 466-472.
Zamani, E., Shokrzadeh, M., Fallah, M., & Shaki, F. (2017). A review of acrylamide toxicity and
its mechanism. Pharmaceutical and Biomedical Research, 3(1), 1-7.
Zhang Y., Ren Y., & Zhang Y. (2009). New research developments on acrylamide: analytical
chemistry, formation mechanism, and mitigation recipes. Chemical Review, 109(9): 4375-4397.
Zhang, Y., Ren, Y., Jiao, J., Li, D., & Zhang, Y. (2011). Ultra high-performance liquid
chromatography-tandem mass spectrometry for the simultaneous analysis of asparagine, sugars,
and acrylamide in Maillard reactions. Analytical Chemistry, 83(9), 3297-3304.
27

585
586
587
588
589
Zhang, Y., Wang, Q., Huang, M., & Chen, X. (2015). Unravelling the effect of flavonoids on the
kinetic profiles of acrylamide in the Maillard reaction. RSC Advance, 5(102), 84084-84092.
Zyzak, D. V., Sanders, R. A., Stojanovic, M., Tallmadge, D. H., Eberhart, B. L., Ewald, D. K.,
Gruber, D. C., Morsch, T. R., Strothers, M. A., Rizzi, G. P., & Villagran, M. D. (2003).
Acrylamide formation mechanism in heated foods. Journal of Agricultural and Food
Chemistry, 51(16), 4782-4787.
590
591
28

592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
Figure captions
Figure 1. Kinetic profiles of glucose and amino acid (Lys, Trp, and Asn) in the different
amino acid/glucose model systems when heated at 100 °C (A, C) and 130 °C (B, D).
Asn/Glu: asparagine and glucose model system, Trp/Glu: tryptophan and glucose
model system, Lys/Glu: lysine and glucose model system; Mix/Glu: asparagine,
tryptophan, lysine and glucose model system.
Figure 2. Kinetic profiles of Maillard reaction harmful products includes including
acrylamide (A), harmane (B), norharmane (C), CML (D), CEL (E), and melanoidins (F
and G) in the different amino acid/glucose model systems when heated at 100 °C and
130 °C. Different alphabetical symbols shows significant differences (p < 0.05) in the
content of acrylamide, CML, CEL, harmane and norharmane. Asn/Glu: asparagine and
glucose model system, Trp/Glu: tryptophan and glucose model system, Lys/Glu: lysine
and glucose model system; Mix/Glu: asparagine, tryptophan, lysine and glucose model
system.
Figure 3. Kinetic profiles of α-dicarbonyl intermediate compounds, including glyoxal
and methylglyoxal, in the different amino acid/glucose model systems when heated at
100 °C (A, C) and 130 °C (B, D). Asn/Glu: asparagine and glucose model system,
Trp/Glu: tryptophan and glucose model system, Lys/Glu: lysine and glucose model
system; Mix/Glu: asparagine, tryptophan, lysine and glucose model system.
Figure 4. Kinetic profiles of major aldehyde intermediate compounds, such as acrolein
and acetaldehyde, in the different amino acid/glucose model systems when heated at
100 °C (A, C) and 130 °C (B, D). Asn/Glu: asparagine and glucose model system,
29