Simultaneous analysis of acrylamide and its key precursors, intermediates, and products in model systems by liquid chromatography-triple quadrupole mass spectrometry

Article abstract of DOI:10.1021/ac4019928

A new analytical setup allowing the simultaneous analysis of precursors, intermediates and products in Maillard reaction model systems was developed to investigate the formation and mitigation kinetics of acrylamide. It was based on high-performance liquid chromatography combined with electrospray ionization triple quadrupole tandem mass spectrometry (HPLC-MS/MS). Chromatography and mass spectrometry conditions were optimized to permit simultaneous monitoring of compounds relevant to acrylamide, such as asparagine, glucose, glycine, N-(1-deoxy-d-fructos-1-yl)-glycine (DFG), N-(1-deoxy-d-fructos-1-yl)-asparagine (DFA), and four reaction products between acrylamide and glycine. A fairly good separation was achieved on a Venusil MP-C₁₈ column using an isocratic eluent of 0.5% formic acid in water at a flow rate of 0.4 mL/min. Triple quadrupole mass spectrometry in multiple reaction monitoring mode (MRM) was applied to obtain good sensitivity. All of the ten key reaction compounds were separated and determined in a single run within 15 min. High correlation coefficients (r > 0.99) of the ten analytes were obtained in their respective linear ranges. The method was capable to accurately quantify the ten key compounds in model systems during heating. The results showed that acrylamide formed in Maillard reaction could be reduced by reacting directly with glycine, and acrylamide concentration constantly decreased by about 60% within the 180 min heating at 150 C when glycine was equal to asparagine and glucose in molecular concentration in model systems. This method helps achieve reduction in both time and labor of analysis of a large number of samples.

Full text of DOI:10.1021/ac4019928

Article
pubs.acs.org/ac
Simultaneous Analysis of Acrylamide and Its Key Precursors,
Intermediates, and Products in Model Systems by Liquid
Chromatography−Triple Quadrupole Mass Spectrometry
Jie Liu,†,‡ Yong Man,§ Yuchen Zhu,† Xiaosong Hu,† and Fang Chen*,†
†College of Food Science and Nutritional Engineering, National Engineering Research Center for Fruits and Vegetables Processing,
Key Laboratory of Fruits and Vegetables Processing, Ministry of Agriculture, Engineering Research Centre for Fruits and Vegetables
Processing, Ministry of Education, China Agricultural University, Beijing 100083, China
‡College of Grain, Oil and Food, National Engineering Laboratory for Wheat & Corn Further Processing, Henan University of
Technology, Zhengzhou 450001, China
§College of Chemistry and Chemical Engineering, Henan University of Technology, Zhengzhou 450001, China
S
* Supporting Information
ABSTRACT: A new analytical setup allowing the simultaneous
analysis of precursors, intermediates and products in Maillard
reaction model systems was developed to investigate the formation
and mitigation kinetics of acrylamide. It was based on high-
performance liquid chromatography combined with electrospray
ionization triple quadrupole tandem mass spectrometry (HPLC−
MS/MS). Chromatography and mass spectrometry conditions
were optimized to permit simultaneous monitoring of compounds
relevant to acrylamide, such as asparagine, glucose, glycine, N-(1-
deoxy-D-fructos-1-yl)-glycine (DFG), N-(1-deoxy-D-fructos-1-yl)-
asparagine (DFA), and four reaction products between acrylamide
and glycine. A fairly good separation was achieved on a Venusil
MP-C18 column using an isocratic eluent of 0.5% formic acid in water at a flow rate of 0.4 mL/min. Triple quadrupole mass
spectrometry in multiple reaction monitoring mode (MRM) was applied to obtain good sensitivity. All of the ten key reaction
compounds were separated and determined in a single run within 15 min. High correlation coefficients (r > 0.99) of the ten
analytes were obtained in their respective linear ranges. The method was capable to accurately quantify the ten key compounds in
model systems during heating. The results showed that acrylamide formed in Maillard reaction could be reduced by reacting
directly with glycine, and acrylamide concentration constantly decreased by about 60% within the 180 min heating at 150 °C
when glycine was equal to asparagine and glucose in molecular concentration in model systems. This method helps achieve
reduction in both time and labor of analysis of a large number of samples.
reduce acrylamide levels during food processing, such as
reducing the contents of precursors,6 mitigating formation of
acrylamide by adding other compounds,7−10 and adjusting
process parameters.11 In particular, the acrylamide content
could be reduced significantly when some amino acids other
than asparagine were added.12 It is more practical for the food
industry to add amino acids to inhibit the formation of
acrylamide among the strategies, since this operation would
raise little concern on the artificial chemicals in foods.
Glycine is the smallest of the 20 amino acids commonly
found in proteins and has been shown to be one of the most
effective amino acid additives in mitigating acrylamide
formation in bread, potato crisps, wheat flake, biscuit, and
cracker at lab and pilot scale.13−15 It has been proved that
crylamide, a thermal process-induced contaminant in food,
Ahas attracted worldwide researchers to study the
mechanism of its formation in foods, the risks associated for
consumers and possible strategies to lower acrylamide levels in
foodstuffs,1 since acrylamide is a neurotoxic compound
classified as a probable human carcinogen and genotoxicant.2
It is well-known that Maillard reaction between asparagine and
reducing sugar is the main pathway for the formation of
acrylamide.3,4 The yield of acrylamide was sensitive to the
composition of free amino acid and sugar in food and to
conditions which are beneficial to promote Maillard reaction,
such as high temperature and low moisture level.5 Therefore,
the formation of acrylamide is one of a number of competing
processes as the desired consequences of the Maillard reaction
(color and flavor) share intermediates with acrylamide
formation.
Received: July 2, 2013
Taking the factors of effecting acrylamide formation into
Accepted: September 4, 2013
account, several effective measures have been suggested to
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
competition of glycine and asparagine for the reducing sugars
was responsible to decrease acrylamide in our previous study.16
Additionally, our research team first proposed the Michael
addition reactions between acrylamide and glycine with or
without the initial oxidation of glycine were the direct pathways
for the removal of acrylamide by glycine, and provided the
direct evidence for the pathways by the unequivocal
identification of the main reaction products between acrylamide
and glycine and intermediates formed in acrylamide/glycine
model systems.17
The knowledge of reaction kinetics is essential for under-
standing the acrylamide formation and elimination process.
Thus it badly needs an accurate and rapid analytical method.
The simultaneous quantification of multicomponents including
precursors, intermediates, as well as products in complex
reactions network is crucial to obtain concentrations of key
compounds rapidly and precisely. Although multiresidues
simultaneous analysis methods have been developed in food,
environmental monitoring, and metabolite analysis,18−21
methods for simultaneous determination of main reaction
components in successive critical reaction stages of the complex
chemical reactions are limited. Nikoline et al22 developed a
liquid chromatography-tandem mass spectrometry (LC-MS/
MS) method to simultaneously determine acrylamide and its
precursors (asparagine and reducing sugar). However, the
accuracies and precisions in determinations of saccharides were
expected to be improved. Zhang and co-workers23 reported an
efficient ultra high-performance liquid chromatography-tandem
mass spectrometry (UHPLC-MS/MS) method for the
simultaneous analysis of acrylamide and its precursors in
Maillard reactions using an isotope dilution technique. The
method did not yet deal with the determination of
intermediates and the products related to the elimination of
acrylamide after its formation.
the presence of DFG and DFA. For DFA, L-asparagine (10.59
g, 0.07 mol), D-glucose (21.02 g, 0.28 mol), and Na2S2O5-
sodium disulphite (7.6 g, 0.04 mol) were added to 80 mL of
MeOH−water solution (1:1). The mixture was then stirred for
20 min at room temperature and subsequently refluxed for 14
h. Purification methods was same to that of DFG.24 Analog
calculation for C10H18N2O8 (DFA): C, 40.82; H, 6.17; N, 9.52.
Found: C, 38.70; H, 6.59; N, 8.12.
Preparation of Standard Solutions. Standard substances
of acrylamide, asparagine, glycine, and glucose were commercial
products and that of reactions products were prepared on the
basis of our previous work.17 Stock solutions and calibration
standards of all ten analytes were prepared in water.
HPLC-MS/MS Analysis. The ten analytes were determined
by an Alliance 2695 Separations Module (Waters, Milford, MA,
U.S.A.) coupled to a Micromass Quattro Micro triple-
quadrupole mass spectrometer (Micromass Company Inc.,
Manchester, U.K.). HPLC separation was carried out on a
Venusil MP-C18 column (250 mm ×4.6 mm i.d., 5 μm particle
size, Agela Technologies, Newark, USA) maintained at 30 °C.
The mobile phase was 0.5% formic acid in water at a flow rate
of 0.4 mL/min. Detection was performed by MS/MS after
selection and optimization of multiple reaction-monitoring
mode (MRM) traces from MS and daughter scans. The
optimized MS instrument parameters obtained after tuning
were as follows: capillary voltage, 3 kV; source temperature, 110
°C; desolvation temperature, 400 °C; desolvation gas flow, 600
L/h nitrogen; and cone gas flow, 50 L/h; argon collision gas
pressure to 2 × 10−3 mbar for MS/MS. In the MRM
transitions, the dwell and inter scan time were 0.4 and 0.1 s,
respectively. Data acquisition was performed with MassLynx
software (version 4.1, Micromass, Manchester, U.K.). The cone
voltage and collision energy for each monitored transition of
the ten analytes were shown in Table 1. The collision energy of
the analytes was optimized according to the maximal
abundance of their quantitative ions.
The aim of this study was to develop a high-performance
liquid chromatography-tandem mass spectrometry (HPLC-
MS/MS) method to simultaneously quantify ten key
compounds in one run: precursors (asparagine, glycine,
glucose), two intermediates N-(1-deoxy-D-fructos-1-yl)-glycine
(DFG) and N-(1-deoxy-D-fructos-1-yl)-asparagine (DFA),
acrylamide, and four reaction products (RP-1, RP-2, RP-3,
and RP-4) in model systems for mitigating acrylamide by
glycine. The developed method would have the potential to
generate data rapidly, to provide a tool performing thorough
research on the mitigating mechanism by glycine, and to
monitor reaction process.
Table 1. Parameters of MRM Mode to Acquire Data
traces
(m/z units)
cone voltage
(V)
collision energy
(eV)
AA [M + H]+
Gly [M + H]+
Asn [M + H]+
Glc [M + Na]+
DFG [M + H]+
DFA [M + H]+
RP-1 [M + H]+
RP-2 [M + H]+
RP-3 [M + H]+
RP-4 [M + H]+
72 > 55
76 > 30
20
20
15
32
12
13
18
22
30
30
13
35
15
20
10
15
15
15
10
10
133 > 74
203 > 143
238 > 220
295 > 259
147 > 88
218 > 159
205 > 88
276 > 217
EXPERIMENTAL SECTION
■
Chemicals and Materials. LC solvents were of HPLC
grade and purchased from Honeywell (Seoul, South Korea).
Formic acid (≥98%) was obtained from Sigma-Aldrich
(Steinheim, Germany). Acrylamide (>99.9%) was purchased
from Bio Basic Inc. (Markham, Canada), L-asparagine and
glycine (≥99%) from Amresco (Solon, U.S.A.), and D-glucose
from J&K Scientific Ltd. (Beijing, China). Ion-exchange resin
(Dowex 50W × 8, H+) was purchased from Huaerbo Ltd.
(Beijing, China). Na2S2O5-sodium disulphite, anhydrous
alcohol, and aqueous ammonia (≥25−28%) were of analytical
grade and obtained from Beijing Chemicals Co. (Beijing,
China).
Analysis of Samples. An aliquot of reaction solution (1.0
mL) from acrylamide/glycine (AA/Gly), asparagine/glucose
(Asn/Glc), glycine/glucose (Gly/Glc), and asparagine/glu-
cose/glycine (Asn/Glc/Gly) model systems heating at 150 °C
from 15 to 180 min were passed through 0.22 μm filtration
membranes, and diluted with the mobile phase. The filtrate was
transferred to an autosampler vial for HPLC-MS/MS analysis.
RESULTS AND DISCUSSION
Optimization of LC Conditions. The simultaneous
analysis performed on the present chromatographic system
Preparation of Intermediates. DFG and DFA were
prepared according to the method described by Martins et al24
with the modification that mass spectrometer was used to check
■
B
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 1. continued
C
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 1. MRM chromatograms of simultaneous determination of the ten analytes. AA, acrylamide; Glu, glucose; Asn, asparagine; Gly, glycine;
DFG, N-(1-deoxy-D-fructos-1-yl)-glycine; DFA, N-(1-deoxy-D-fructos-1-yl)-asparagine; RP-1−RP-4, four reaction products between acrylamide and
glycine. The concentrations of analytes were at levels when Asn/Glu/Gly model systems heated for 15 min: CAA = 0.05 mM, CGlu = 61.55 mM, CAsn
= 100.01 mM, CGly = 100.21 mM, CDFG = 0.24 mM, CDFA = 0.043 mM, CRP‑1 = 0.023 mM, CRP‑2 = 0.0049 mM, CRP‑3 = 0.091 mM, CRP‑4 = 0.14 mM.
seems difficult because of the significant difference in polarity
among the ten compounds. Besides, the separation among the
components with the adjacent retention time is a great
challenge. Therefore, the selection of the chromatographic
column and mobile phase was a prerequisite to obtain a good
separation. Acrylamide is a highly polar molecule with poor
retention in conventional LC reversed phase sorbents.
Therefore, based on our experience on chromatographic
separation of acrylamide25 and the relationships among
structures of the ten compounds, several columns specially
designed to have enhanced retention and selectivity for polar
compounds were tested for their separation efficiencies: (i)
Thermo Hypersil ODS-C18 column (250 mm ×4.6 mm i.d., 5
μm), (ii) Waters Atlantis dC18 column (150 mm ×2.1 mm i.d.,
3 μm), (iii) Agela Venusil MP-C18 column (250 mm ×4.6 mm
i.d., 5 μm), and (iv) Agela HILIC column (250 mm ×4.6 mm
i.d., 5 μm). The peak elution time of acrylamide on Atlantis
dC18 column was about 3 min, which was faster than that on
other three columns. While the amino acids and reactions
products were difficult to be retained and separated efficiently
on Atlantis dC18 column as its column length was limited. RP-1
and RP-2 were not separated completely on HILIC column and
ODS-C18 column. So Venusil MP-C18 was selected as the
separation column with a flow rate of 0.4 mL/min which
matched with the MS/MS electroscopy ionization.
Besides the optimization of separation efficiencies of the
chromatographic column, the choice of mobile phase should be
concerned based on the ionization efficiency of analytes in MS/
MS system in order to obtain nice resolution and high
sensitivity. The mobile phase was adjusted from the mixture of
10% acetonitrile and 90% water (used to analyze acrylamide in
our early study)25 to 100% aqueous mobile phase. Results
showed that the retention of acrylamide could be improved and
the MS sensitivity did not change significantly by avoiding
organic modifiers like acetonitrile and methanol in the aqueous
mobile phase. Venusil MP-C18 can be used with pure water as
the mobile phase without suffering from the “phase collapse”
phenomena associated with traditional reversed phase columns.
Therefore, the combination of Venusil MP-C18 column and
water as a mobile phase gave good retention of polar
compounds and maintain good peak shapes and intensities.
Mobile phase containing organic acid can often improve
ionization when using electrospray, and consequently increase
the sensitivity of a method. Results from the MS full scan of ten
compounds showed that the responses of corresponding
molecular ions under the ESI+ mode were greatly improved
and stabilized, and high sensitivity was subsequently obtained
when formic acid was added. Finally, water containing 0.5%
formic acid as the mobile phase appeared optimal with regard
to retention, separation efficiency, and MS response of
compounds.
The chromatograms of ten analytes obtained from Asn/Glc/
Gly model systems heated 15 min were shown in Figure 1.
Glycine, asparagine, DFG, RP-1, DFA, glucose, RP-2, RP-3, RP-
4, and acrylamide successively eluted from the column. The
whole separation of the all ten analytes was completed when
acrylamide eluted at about 15 min in a chromatographic cycle.
Though the retention time of glucose, asparagine, glycine,
DFG, DFA, RP-1, and RP-2 distributed around 6−7 min
closely, the characteristic ions of each compound were different
and there was no interference among each other for the
quantification of the seven compounds in MRM mode.
Optimization of MS/MS Conditions. In this study, triple
quadrupole mass spectrometry was applied to increase
sensitivity in the determination. As the presence of nitrogen
atom(s), having an important proton affinity (basic character),
and carboxyl groups (acidic character) in the ten compounds
except glucose, the MS spectra could be acquired in both
positive and negative ion modes. While most of them could be
monitored under both ESI+ and ESI− modes, the stronger
response was observed under the ESI+ mode than ESI− mode,
such as asparagine described by others.22,23 The fragmentations
of the corresponding characteristic ions used to quantification
in MRM were presented in Figure 2.
The ion fragments of acrylamide were in agreement with the
previous work.26 The fragment ion m/z 55 showed a relatively
high intensity and was selected to quantify acrylamide.
Saccharides could be ionized as sodium ion adducts without
the addition of sodium ions into the mobile phase in advance
and the sodium ion adducts are usually more abundant than the
hydrogen ion adducts. Hence the glucose was detected as
sodium adducts. Ions [M + Na]+ at m/z 203 indicated the
D
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 2. continued
E
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 2. Fragmentation of the ten analytes. DFG, N-(1-deoxy-D-fructos-1-yl)-glycine; DFA, N-(1-deoxy-D-fructos-1-yl)-asparagine; RP-1−RP-4,
four reaction products between acrylamide and glycine.
formation of sodium ion adducts of glucose, which was also
detected by Nielsen22 in a LC-MS/MS method for
simultaneous analysis of acrylamide and its precursors. The
inferior signal-to-noise ratio in the MRM chromatography of
glucose at significant levels showed that the monosodiated
adduct ions [M + Na]+ was unstable, although their response
was high in full scan. The fragment ion m/z 74 of asparagine
corresponded to the loss of acid amide (−CH2CONH2). The
similar fragmentation of glycine was observed that ion m/z 30
corresponded to the loss of carboxyl group (−COOH) in the
present work. The quantitative ions of both DFG and DFA
corresponded to the loss of water from the molecular ions,
respectively. The characteristic ions of the reactions products
have been discussed in our previous work17 and their
quantitative ions were listed in Table 1.
and the regression lines were calculated by no weight. The
concentrations of the standards were selected in a way that they
could cover a wide range of concentrations detected.
Calibration standards solutions were 1−15 μmol/L for
acrylamide, 10−40 μmol/L for asparagine, 100−500 μmol/L
for glycine, 50−550 μmol/L for glucose, 1−15 μmol/L for
DFG, 1−15 μmol/L for DFA, 1−15 μmol/L for RP-1, 0.05−
0.25 μmol/L for RP-2, 2−25 μmol/L for RP-3, and 2−20
μmol/L for RP-4. Strong linear relationships and good
coefficients of determination were achieved for all ten
compounds over the concentration range (Table 2).
Samples were spiked with the standards of ten analytes at
different low, intermediate and high levels to determine their
recoveries, respectively. The low, intermediate, and high spiking
levels of acrylamide, asparagine, DFG, DFA, RP-1, RP-3, and
RP-4 standards were 2, 10, and 20 μmol/L, respectively.
Meanwhile, the low, intermediate, and high spiking levels of
glycine and glucose were 50, 200, and 500 μmol/L, and those
Calibration and Validation. The quantification of all
compounds was achieved by means of seven-point calibration
curves. All calibration curves were not forced through the origin
F
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
a
systems. At the same time, the changes in concentrations of
compounds related to acrylamide should be compared in AA/
Gly, Asn/Glc, and Gly/Glc model systems.
Table 2. Calibration Results for Ten Analytes
linear range
y-
intercept coefficient(r)
analyte
(μmol/L)
slope
Results appearing from Figure 3−5 showed the variance of
precursors, intermediates, and products in different model
systems with heating time. The kinetics of the ten compounds
was clearly profiled. The changes of acrylamide, glycine, and
reactions products in AA/Gly model systems have been
discussed in previous work.17 In the absence of heating
(heating time, 0 min), the concentrations of three precursors
(asparagine, glucose, and glycine) were quantified as approx-
imately 0.1 mol/L, which is the addition level used in this study,
while acrylamide could not be detected, indicating there was no
contamination in the experiments.
acrylamide
asparagine
glycine
glucose
DFG
1−15
10−40
100−500
50−550
1−15
1−15
1−15
0.05−0.25
2−25
1774.13
1098.41
1.27
218.18
3109.24
51.71
0.99907
0.99861
0.99099
0.99758
0.99958
0.99870
0.99804
0.98857
0.99415
0.99326
0.29
24.93
5688.61
3649.93
4754.44
16409.51
655.67
84.56
4876.55
792.32
2672.74
−21.02
−745.89
−72.55
DFA
RP-1
RP-2
RP-3
RP-4
2−20
a
DFG, N-(1-deoxy-D-fructos-1-yl)-glycine; DFA, N-(1-deoxy-D-fruc-
The concentrations of acrylamide in both Asn/Glc/Gly and
Asn/Glc model systems were increased with heating time, and
that in Asn/Glc/Gly model systems were always about half of
that in Asn/Glc model systems (Figure 3A). If the
concentrations of acrylamide in Asn/Glc/Gly and Asn/Glc
model systems were set as CAA3 and CAA2 respectively, the
tos-1-yl)-asparagine; RP-1−RP-4, four reaction products between
acrylamide and glycine.
levels of RP-2 were 0.05, 0.1, and 0.2 μmol/L. The recoveries of
all analytes ranged from 96.9% to 108.2% at three spiking levels
(Supporting Information Table S1).
LOD and LOQ were determined as the amounts of analytes
that produced a signal-to-noise ratio of 3:1 and 10:1,
respectively. The values of LOD of acrylamide, asparagine,
DFG, DFA, RP-1, RP-2, and RP-3 were less than 10 ng/mL,
and that of glycine and glucose were at levels of mg/mL
(Supporting Information Table S1).
inhibitory rate (γI) of acrylamide by glycine was γI = [(CAA2
−
CAA3)/CAA2] × 100%. The inhibitory rate varied between 55%
and 64% within the 180 min reaction time (Figure 3B). It
suggested that the obvious mitigating effect of glycine on
acrylamide in complex reactions was constant even with the
dynamic glycine concentrations.
The concentrations of asparagine decreased more than 90%
up to 135 min and then changed little in both model systems
(Figure 4A). Especially in the first 60 min, the first order
observed rate coefficient (k) for asparagine loss in Asn/Glc/Gly
model systems (k = 0.75) was significantly less than that in
Asn/Glc model systems (k = 1.91), indicating the competition
of glycine with asparagine for the reducing sugars occurred in
Asn/Glc/Gly model systems.
Glycine decreased by about 10% up to 180 min in both Asn/
Glc/Gly and Gly/Glc model systems (Figure 4B). In Asn/Glc/
Gly model systems, the glycine was involved in competing with
asparagine for glucose and combining the formed acrylamide.
So the k of glycine in Asn/Glc/Gly model systems (k = 0.07)
Precision was calculated in terms of intraday repeatability and
interday reproducibility as RSD% at three concentration levels.
The reaction solutions were sampled in 15, 30, and 45 min
heating for the detection of DFG and DFA, and in 60, 120, and
180 min heating for the other eight compounds, respectively.
The values of RSD ranged from 0.0% to 10.3% for the intraday
precision test (n = 6) and 0.1% to 10.4% for the interday
precision tests (n = 6) (Supporting Information Table S2).
Analysis of Ten Analytes and Kinetic Evaluation. As
the main pathway for the formation of acrylamide in foods is
related to the Maillard reation between asparagine and reducing
sugar, the effect of glycine on the formation and reduction of
acrylamide should be investigated in Asn/Glc/Gly model
Figure 3. Time courses of AA in Asn/Glc and Asn/Glc/Gly model systems (A) heated at 150 °C with different heating time. Each experiment was
performed in triplicate repeats (n = 3). Inhibition rates of AA (B) by adding Gly in Asn/Glc/Gly model systems were obtained in comparison with
that in Asn/Glc model systems. AA, acrylamide; Asn, asparagine; Glc, glucose; Gly, glycine.
G
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
Figure 4. Time courses of Asn (A), Gly (B), Glc (C), DFG (D), and DFA (E) in Asn/Glc, Gly/Glc, and Asn/Glc/Gly model systems heated at 150
°C with different heating time. Each experiment was performed in triplicate repeats (n = 3). Asn, asparagine; Gly, glycine; Glc, glucose; DFG, N-(1-
deoxy-D-fructos-1-yl)-glycine; DFA, N-(1-deoxy-D-fructos-1-yl)-asparagine.
was slightly greater than that (k = 0.06) in Gly/Glc model
systems.
The concentrations of glucose decreased more than 90% in
Asn/Glc/Gly model systems, 80% in Asn/Glc model systems,
and 50% in Glc/Gly model systems up to 180 min, respectively
(Figure 4C). After the reaction lasting for 70 min, the
concentrations of glucose in Asn/Glc/Gly model systems
were less than that in Asn/Glc and Glc/Gly model systems. It
was the result of glycine and asparagine consuming glucose
together in Asn/Glc/Gly model systems.
In Asn/Glc/Gly model systems, the concentrations of
asparagine and glucose were quickly reduced in the first 75
H
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
kinetic profiles of RP-1 and RP-2 in Asn/Glc/Gly model
systems showed that the reaction between RP-1 and acrylamide
might be disturbed and the acrylamide concentration was so
little that the production of RP-2 was limited. According to the
structure of RP-2 and nucleophilic adduction, the ratio to
produce RP-2 for glycine and acrylamide was 1: 2, indicating
that excess acrylamide favors the formation of RP-2. While in
the Asn/Glc/Gly model systems, glycine outweighed acryl-
amide greatly. So the concentration ratio of reactants made
against the formation of RP-2 and the concentrations of RP-2
increased steeply within the initial 15 min followed by decrease
with the development of reaction. It was same to the formation
of RP-4 on the base of RP-3. Sufficient glycine favored the
formation of RP-3 so the concentrations of RP-1 and RP-3
were about the same amount all the time. The difference in the
changes of the four reaction products mean that acrylamide was
not only formed from asparagine and glucose through Maillard
reaction, but also might be released from RP-2 and RP-4
through Hoffmann elimination. Then acrylamide was used to
form RP-1 and RP-3.
Figure 5. Time courses of RP-1, RP-2, RP-3, and RP-4 in Asn/Glc/
Gly model systems heated at 150 °C with different heating time. Each
experiment was performed in triplicate repeats (n = 3). RP-1−RP-4,
four reaction products between acrylamide and glycine.
Compared to individual analyzing some of these analytes in
previous work,27 the current method greatly reduced the time
and cost. Besides, the present method also eliminated the error
of results from different quantitative methods. Although the
matrix effect is one of the limitations of LC-MS/MS with
electrospray interface (ESI) that accuracy, precision and
robustness cannot be guaranteed, high multiples of dilution
for samples eliminated matrix effect easily in this study.
min heating, and then decreased gradually slowly with the
development of reaction. The k of glucose (k = 0.87) was faster
than that of asparagine (k = 0.75), because glucose was
simultaneously involved in the isomerization and Maillard
reactions with glycine and asparagine. The loss of glycine was
far less than that of asparagine in the whole process, and k of
glycine in Gly/Glc model systems (k = 0.06) was significantly
less than that of asparagine in Asn/Glc model systems (k =
1.91), indicating that asparagine had the advantage in
competition with glycine for glucose.
The concentrations of the both intermediates (DFG and
DFA) detected decreased with heating time after their
formation indicating that they generated fast and were unstable.
DFG in Asn/Glc/Gly model systems was completely
consumed, and the reduction of that in Glc/Gly model systems
became moderate up to 135 min (Figure 4D). The
concentrations of DFG in Asn/Glc/Gly model systems were
less than that in Glc/Gly model systems from beginning to end.
In the meantime (0−135 min), the concentrations of glycine in
both model systems were about the same amount (Figure 4B),
but the glucose concentrations in Glc/Gly model systems was
40% higher than that in Asn/Glc/Gly model systems (Figure
4C). So the difference in the concentrations of DFG in both
model systems was the result of the difference of the
concentrations of glucose, indicating that asparagine competing
with glycine for glucose and inhibiting the formation of DFG.
DFA was completely consumed up to about 76 min in both
Asn/Glc/Gly and Asn/Glc model systems (Figure 4E). The
concentration of DFA at 15 min in Asn/Glc/Gly model
systems was almost five times as much as that in Asn/Glc
model systems, also indicating that glycine competing with
asparagine for glucose and inhibiting the formation of DFA.
The ending time suggested that DFG was more stable than
DFA.
CONCLUSIONS
■
In conclusion, a sensitive, specific and accurate analytical
method based on HPLC-MS/MS was developed and
demonstrated in this work to be an excellent analytical tool
to simultaneously monitor the key precursors, intermediates
and products related to acrylamide in a single analytical run.
Meanwhile, the quantitative results in model systems
demonstrated that simple and rapid sample preparation in
combination with high sensitivity of detection were the
principal advantages of this method as compared to HPLC or
GC. It could be applied to the trace analysis of the multiple key
compounds in complex reactions for a large number of samples,
generate kinetic data, and estimate reaction efficiency.
ASSOCIATED CONTENT
* Supporting Information
Additional material as described in the text. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Phone/Fax:+86 10 6273 7645, ext 18. E-mail: chenfangch@
sina.com.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The concentrations of RP-1 and RP-3 increased with heating
time, whereas the concentrations of RP-2 and RP-4 decreased
with heating after their formation in Asn/Glc/Gly model
systems (Figure 5). The concentrations of RP-2 were the least
compared with that of the other three reactions products. The
concentrations of the four reactions products increased linearly
with heating time in AA/Gly model systems.17 The different
■
This study was financially supported by the National Natural
Science Foundation of China (No.31222044 and
No.31071554), the National Basic Research Program of
China (973 Program, No. 2012CB720805), and the National
High Technology Research and Development Program of
China (863 Program, No. 2011AA100806).
I
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry
Article
REFERENCES
■
(1) Vinci, R. M.; Mestdagh, F.; Meulenaer, B. De. Food Chem. 2012,
133, 1138−1154.
(2) Monographs on the Evaluation of Carcinogenic Risks to Humans;
International Agency for Research on Cancer (IARC): Lyon, France,
1994; Vol. 60, pp 389−433.
(3) Mottram, D. S.; Wedzicha, B. I.; Dodson, A. T. Nature 2002, 419,
448−449.
(4) Stadler, R. H.; Blank, I.; Varga, N.; Robert, F.; Hau, J.; Guy, P. A.;
Robert, M.; Riediker, S. Nature 2002, 419, 449−450.
(5) Zhang, Y.; Ren, Y.; Zhang, Y. Chem. Rev. 2009, 109, 4375−4397.
(6) Muttucumaru, N.; Elmore, J. S.; Curtis, T.; Mottram, D. S.; Parry,
M. A.; Halford, N. G. J. Agric. Food Chem. 2008, 56, 6167−6172.
(7) Pedreschi, F.; Granby, K.; Risum, J. Food Bioprocess Technol.
2010, 3, 917−921.
(8) Levine, R. A.; Ryan, S. M. J. Agric. Food Chem. 2009, 57, 6823−
6829.
(9) Claeys, W.; De Vleeschouwer, K.; Hendrickx, M. E. Biotechnol.
Prog. 2005, 21, 1525−1530.
(10) Zhang, Y.; Ying, T.; Zhang, Y. J. Food Sci. 2008, 73, C60−C66.
(11) Williams, J. S. E. Food Chem. 2005, 90, 875−881.
(12) Kim, C. T.; Hwang, E. S.; Lee, H. J. J. Food Sci. 2005, 70, C354−
C358.
(13) Brathen, E.; Kita, A.; Knutsen, S. H.; Wichlund, T. J. Agric. Food
̊
Chem. 2005, 53, 3259−3264.
(14) Sadd, P. A.; Hamlet, C. G.; Liang, L. J. Agric. Food Chem. 2008,
56, 6154−6161.
(15) Brathen, E.; Kita, A.; Knutsen, S. H.; Wichlund, T. J. Agric. Food
̊
Chem. 2005, 53, 3259−3264.
(16) Zhang, H. X. Inhibitive Effect of Glycine on Acrylamide
Formation, Master’s Thesis in China Agricultural University, Beijing,
2007.
(17) Liu, J.; Chen, F.; Man, Y.; Dong, J.; Hu, X. Food Chem. 2011,
128, 442−449.
(18) Cai, Z.; Zhang, Y.; Pan, H.; Tie, X.; Ren, Y. J. Chromatogr., A
2008, 1200, 144−155.
(19) Ren, Y.; Zhang, Y.; Shao, S.; Cai, Z.; Liang, F.; Pan, H.; Wang, Z.
J. Chromatogr., A 2007, 1143, 48−64.
(20) Ogawa, T.; Yonemura, M.; Kawasumi, K.; Aoyagi, Y.; Ueno, S.;
Ichida, Y.; Ohe, Y.; Endo, K.; Izumi, K. Ann. Oncol. 2012, 23 (S11),
164−165.
(21) Xiong, X.; Yang, L.; Duan, J. Clin. Chim. Acta 2013, 423, 69−74.
(22) Nielsen, N. J.; Granby, K.; Hedegaard, R. V.; Skibsted, L. H.
Anal. Chim. Acta 2006, 557, 211−220.
(23) Zhang, Y.; Ren, Y.; Jiao, J.; Li, D.; Zhang, Y. Anal. Chem. 2011,
83, 3297−3304.
(24) Martins, S. I. F. S.; Marcelis, A. T. M.; van Boekel, M. A. J. S.
Carbohydr. Res. 2003, 338, 1651−1663.
(25) Liu, J.; Zhao, G.; Yuan, Y.; Chen, F.; Hu, X. Food Chem. 2008,
108, 760−767.
(26) Riediker, S.; Stadler, R. H. J. Chromatogr., A 2003, 1020, 121−
130.
(27) Elmore, J. S.; Koutsidis, G.; Dodson, A. T.; Mottram, D. S.;
Wedzicha, B. L. J. Agric. Food Chem. 2005, 53, 1286−1293.
J
dx.doi.org/10.1021/ac4019928 | Anal. Chem. XXXX, XXX, XXX−XXX