Kinetics of α‑dicarbonyl compounds formation in glucose‐glutamic acid model of Maillard reaction

Abstract As a potential health hazard, α‐dicarbonyl compounds have been detected in the thermally processed foods. In order to investigate the formation kinetics of α‐dicarbonyl compounds, liquid chromatography‐electrospray tandem mass spectrometry was employed to determine the content of α‐dicarbonyl compounds in glucose‐only and glucose‐glutamic acid (glucose‐Glu) thermal reaction models. The 3‐deoxyglucosone content was significantly higher than 6 α‐dicarbonyl compounds at 90–110℃, 0–6 hr in the two tested systems. The glutamic acid promoted the content accumulation of 1‐deoxyglucosone, diacetyl, methylglyoxal, and glyoxal, whereas inhibited the content of 3‐deoxyglucosone and 3,4‐dideoxyglucosone. Three‐fifths of the tested compounds content increased linearly with time increasing, but in glucose‐only system, the 1‐deoxyglucosone content increased logarithmically at 95–110℃ over reaction time. The formation of glucose (100–110℃, glucose‐only and glucose‐Glu), 5‐hydroxymethylfurfural (100–110℃, glucose‐only), 1‐deoxyglucose (105–110℃, glucose‐Glu), 3,4‐dideoxyglucosone (110℃, glucose‐Glu), glyoxal (95–110℃, glucose‐Glu) and diacetyl (90–95℃, glucose‐Glu) could be well fitted by exponential equation. Shortening the heating time and reducing heating temperature (except glyoxal in glucose‐only system) were the effective methods to decrease α‐dicarbonyl compounds content in the two tested systems. Additionally, high temperature could also reduce α‐dicarbonyl compounds content, such as 3‐deoxyglucosone (≥110℃, glucose‐only), 1‐deoxyglucosone (≥110℃, glucose‐only), glucosone (≥110℃, glucose‐only; ≥100℃, glucose‐Glu), methyloxyl (≥110℃, glucose‐only; ≥100℃, glucose‐Glu), diacetyl (≥110℃, glucose‐only), and glyoxal (≥100℃, glucose‐Glu).

Meat flavorings based on the animal and vegetable protein as precursors are highly welcomed by domestic and foreign consumers. In the first half of 2020, the total operating revenue of 17 major domestic savory flavoring and seasoning companies reached 42.015 billion yuan (the data are from http://news.foodm ate. net/2020/09/571161.html), which fully demonstrated the meat flavorings with various flavors dominate a huge market shares. Multiple carbon and nitrogen sources have been added during the Maillard reaction stage to obtain meat flavoring with good sensory quality. However, these preparation techniques involve many factors, including the reaction under high temperature, and adding acidic amino acid to the ingredient, which may promote the formation of α-dicarbonyl compounds and HMF (Marianou et al., 2018;Zhang et al., 2019). Although researchers tried to control the production of α-dicarbonyl compounds by adding polyphenoic extract without affecting the sensory properties of food (Pedreschi et al., 2018;Pu et al., 2020). In view of the negative effect of α-dicarbonyl compounds and HMF on human health, many Maillard models have been established to obtain methods that can reduce α-dicarbonyl compounds at the source. In the established model, glucose, fructose, maltose, lactose, and ribose are common sugars, and glycine, lysine, threonine, serine, and cysteine are common amino acids (Cha et al., 2019;Chen & Kitts, 2011;Jang et al., 2013). However, there are few models involving acidic amino acid especially glutamic acid.
Glutamic acid is the most commonly used umami amino acid in the preparation of meat flavoring. Considering the promotion of acidic conditions on for α-dicarbonyl compounds and HMF, it is important to monitor these compounds concentration under different thermal treatment conditions when making the meat flavoring.
To investigate the effect of acidic amino acid on the formation of α-dicarbonyl compounds in meat flavoring processing, the sugar (glucose) and acidic amino acid (glutamic acid) commonly used in meat flavoring preparation were selected in this work. Unitary (glucose-only) and binary (glucose-glutamic acid) models were constructed. Due to the formation of α-dicarbonyl compounds in a complex interconnected network, 10 main compounds (glucose, fructose, HMF, glucosone, 3-deoxyglucosone, 1-deoxyglucosone, 3,4-dideoxyglucosone, glyoxal, methylglyoxal, and diacetyl) were observed. Liquid chromatography-electrospray tandem mass spectrometry (LC-ESI MS/MS) was employed to determine the α-dicarbonyl compounds concentrations in all samples. This work could provide a promising strategy to reduce the amount of α-dicarbonyl compounds produced by glucose-glutamic acid system in meat flavoring.

| Materials and chemicals
Glucose (food grade) was provided by Jizhou Huayang Chemical Co., Ltd. Glutamic acid (food grade) was provided by Jizhou City Huaheng Biological Technology Co., Ltd. Formic acid (98%), glucose, and fructose standard were provided by Sinopharm Chemical Reagent Co., Ltd. High-purity (99%) HMF standard was purchased from Sigma-

| Preparation of model systems
In 250 ml four-neck flasks, equipped with a reflux cooler, the systems of glucose-only (0.3 M) and glucose-Glu (0.3 M each) were prepared using water, respectively. Based on the study of the preparation of meat flavoring at atmospheric pressure, the temperature/ time combination of this work was determined. Under the condition of standard atmospheric pressure, glucose-only and glucose-Glu systems were heated at 90, 95, 100, 105, and 110℃ for 0, 1, 2, 3, 4, 5, and 6 hr, respectively. The heating was carried out in a silicone oil bath. After the end of the reaction, samples were immediately cooled and stored at 4℃ before analysis. Each temperature/time combination was carried out in triplicates. Additional experiments were conducted when needed to obtain reliable results.

| Analysis of glucose and fructose
The reaction mixture samples were diluted twofold with acetonitrile, filtered by 0.22 μm Nylon syringe filter, then determined the content of glucose and fructose by high performance liquid chromatography (HPLC).
The Agilent 1260 HPLC system (Agilent Corp.) equipped with a chromatography column (Innoval NH 2 , 250 mm × 4.6 mm i.d., 5 μm particle size) from Bonna-Agela Technologies Inc. (Tianjin, China) was employed. The mobile phase was acetonitrile-water (80:20, v/v); flow rate was 1.0 ml/min; injection volume was 20 μL; column temperature was 30℃; and the temperature of refractive index detector was 40℃. Glucose and fructose contents were calculated by an external calibration curve prepared with a concentration range from 28 to 7, 000 µg/ml.

| Analysis of HMF
The reaction mixture samples were filtered by a 0.22 μm Nylon syringe filter, then the content of HMF was determined by HPLC.

| Analysis of α-dicarbonyl compounds
Derivatization of α-dicarbonyl compounds was carried out with o-phenylenediamine according to the methods of Kocadağlı and Gökmen (2016a). The derivatization of a 0.5 ml reaction mixture was performed by adding 150 μl of 0.1 M phosphate buffer (pH 7) and 150 μl of 0.2% ο-phenylenediamine in 10 mM DETAPAC. After mixing, the mixture was kept in the dark at indoor temperature for 2 hr before analysis.
The chromatographic separation was performed on an XBP C18 column (150 mm × 4.6 mm i.d., 5 μm) using a gradient mixture of (A) 0.5% formic acid in water and (B) 0.5% formic acid in methanol as the mobile phase at a flow rate of 0.6 ml/min at 30°C. The gradient mixture started from 30% B and increased to 60% B at 10 min, and then it decreased to 30% B at 2 min, and the 30% B remained for 3 min. The chromatographic run was completed within 15 min. The injection volume was 10 μL. The electrospray source had the following settings: drying gas (N 2 ) flow of 10 L/min at 350°C, nebulizer pressure of 40 psig, and capillary voltage of 3,500 V. Fragmentor voltage was set to 100 V. MS data were acquired in the positive mode, and α-dicarbonyl compounds were identified by selected ion monitoring (SIM) mode. The SIM ions [M + H] + of the quinoxaline derivatives were as follows: glucosone, 251; 1-or 3-deoxyglucosone, 235; 3,4-dideoxyglucosone, 217; dimethylglyoxal, 159; methylglyoxal, 145; and glyoxal, 131. Dwell time was set at 50 ms for each compound.

| Kinetics and statistical analysis
Kinetics analysis was conducted according to the method described by Zhang et al. (2019) and Chi et al. (2011). In a non-enzymatic browning reaction, there is an initial induction period corresponding to the stage at which α-dicarbonyl compounds forms. After this relatively quick induction period, the content of α-dicarbonyl compounds increases linearly, exponentially, or logarithmically with time, as shown in Equation 1, 2 and 3, respectively: where C 0 refers to the initial content of compounds (µg/ml); t is the time of heat treatment (h); C t is the content of compound (µg/ml) at time t; k 0 , k 1 , and k 2 are rate constants (h -1 ).
The concentrations of reactants and products were expressed as µg/ml of reaction mixture. The experimental data were compared with the mathematical model, and the dynamic model was evaluated by correlation coefficients (R 2 ), accuracy factor (Af), bias factor (Bf), root mean square error (RMSE), and sum of squares of the differences of the natural logarithm of observed and predicted values (SS) (Kavousi et al., 2015;McClure et al., 1993). In general, a value of Af, Bf and R 2 closer to 1, together with a lower value for SS and RMSE, indicated a better-fit model. (1)

| Glucose consumption and effect of glutamic acid
The markers in Figure 1 respectively showed experimental observation data of reactants and products formed in heated glucose-only and glucose-Glu systems. Goodness-of-fit evaluation of glucose-only and glucose-Glu systems was shown in Table 1 and Table 2.
The initial contents of glucose in glucose-only and glucose-Glu systems were 53, 898.08 µg/ml and 53, 877.05 µg/ml, respectively (Figure 1 (1A and 2A)). In both the glucose-only system and the glucose-Glu system, the glucose content decreased with increasing heating time and increasing temperature. This is consistent with the trends that have been reported for glucose in glucose-only and glucose-NaCl systems (Kocadağlı, & Gökmen, 2016a). Levels of glucose decreased by 4.21% (90℃, 6 hr), 4.67% (95℃, 6 hr), 5.89% (100℃, 6 hr), 8.52% (105℃, 6 hr), and 8.59% (110℃, 6 hr) in glucose-only system at five different heating temperatures, respectively. The glucose level in the glucose-Glu system severally reduced by 4.66%, 8.66%, 12.21%, 17.35%, and 19.43%. The degradation rate of glucose content in the glucose-Glu system was faster than that of glucose-only system. Kocadağlı and Gökmen (2016a) also reported that the degradation of glucose was apparently faster in the presence of NaCl.
The kinetics analysis results of glucose consumption in the system of glucose-only and glucose-Glu were shown in Table 1 and 2, respectively. At the given temperature, the glucose consumption exhibited zero-order kinetics at 90-95℃ (0-6 hr) in both glucose-only and glucose-Glu systems. And the consumption rate of glucose increased with the increase of temperature. In contrast, when the reaction temperature was higher than 100℃, the kinetics of glucose consumption was fitted by equation (2) in two systems. At the same temperature, the consumption rate of glucose in glucose-Glu system was significantly higher than that of glucose-only system. Therefore, it could be concluded that the presence of glutamic acid would accelerate the consumption of glucose.

| Fructose formation and effect of glutamic acid
In     The kinetics analysis results of fructose formation in the system of glucose-only and glucose-Glu were shown in Table 1 and 2, respectively. At the given temperature, the fructose formation exhibited zero-order kinetics at 90-105℃ (1-6 hr) in glucose-only system. The formation rate of fructose increased with the increase of temperature at lower heating temperature (90-95℃). When the heating temperature was 100 and 105℃, fructose consumption was greater than its formation. The content of fructose showed a trend of accumulation at 110℃. The formation of fructose could not be well fitted by zero-order kinetics at 110℃ (1-6 hr) (R 2 = 0.7489).
In the glucose-Glu system, except for 100℃, the curve-fitting of fructose formation was not ideal (R 2 < 0.9). The presence of acidic amino acids (glutamic acid) might contribute to the Maillard reaction, making the reaction of the entire system more complicated (Kavousi et al., 2015).

3-Deoxyglucosone, 3,4-dideoxyglucosone, and glucosone
were rapidly accumulated with the increasing of heating time at the beginning. The 3-deoxyglucosone was predominant whether in glucose-only system or in glucose-Glu system at five different heating temperatures. The same observation was also reported in glucose-lysine system (Gobert & Glomb, 2009). Gobert and Glomb (2009) also reported that 3-deoxyglucosone-quinoxaline increased independent from oxygen during the 7-day observation Note: -: Multiple data were below the lowest detection limit of the analytical instruments, which makes it impossible for these models to be analyzed by the curve-fitting method applied to the other model systems.
/: The data cannot be calculated by formula.  period. The content of 3-deoxyglucosone was the highest with heating at 121℃, 90 min in glucose-glycine system, while at 121℃, 30 min in glucose-lysine system (Chen & Kitts, 2011). The highest contents of 3-deoxyglucosone in glucose-only and glucose-Glu systems were observed at the maximum heating time tested (Figure 1 (1C and 2C)), which was consistent with the study of 3-deoxyglucosone in glucose-glycine system (Chen & Kitts, 2011). At five different heating temperatures, the highest content of 3,4-dideoxyglucosone was 0.0136 µg/ml (110℃, 6 hr) in glucose-only system (Figure 1 (1D)). The highest content of glucosone was 0.0045 µg/ml (105℃, 6 hr) in glucose-only system ( Figure 1 (1F)). However, the glucosone content was significantly higher than 3,4-dideoxyglucosone at 90-95℃ in glucose-Glu system. The 3,4-dideoxyglucosone content was higher than the glucosone content with heating at > 100℃ in glucose-Glu system. The highest content of 3-deoxyglucosone (105℃, 6 hr) was about 108.36 times as high as that of 1-deoxyglucosone (105℃, 6 hr) in glucose-only system. The highest content of 3-deoxyglucosone (110℃, 6 hr) was about 31.85 times as high as that of 1-deoxyglucosone (110℃, 6 hr) in glucose-Glu system. The highest content of glucosone was observed with heating at 105℃, 6 hr in glucose-only system, while the highest content of glucosone was observed with heating at 95℃, 6 hr in glucose-Glu system. The ratio of 3-deoxyglucosone (105℃, 6 hr) to glucosone (105℃, 6 hr) was 10.80 in glucose-only system. The ratio of 3-deoxyglucosone (95℃, 6 hr) to glucosone (95℃, 6 hr) was 3.17 in glucose-Glu system. It has been reported that the ratio of 3-deoxyglucosone to 1-deoxyglucosone was 3, the ratio of 3-deoxyglucosone to glucosone was 20 during heating aqueous solution of glucose and alanine in pH 7 phosphate buffer under reflux. The ratio of 3-deoxyglucosone to 1-deoxyglucosone and the ratio of 3-deoxyglucosone to glucosone had nothing to do with heating time (Hofmann, 1999). In this work, the ratio of 3-deoxyglucosone to 1-deoxyglucosone and the ratio of 3-deoxyglucosone to glucosone were different from those reported by Hofmann (Hofmann, 1999). The difference might be due to the formation and degradation of α-dicarbonyl compounds during heating had complex kinetics and thermodynamic behaviors under the experimental conditions in this work.
At higher temperature, equation (2) had better fitting effect on the formation kinetics of 3,4-dideoxyglucosone (110℃) and 1-deoxyglucosone (105-110℃) ( Table 2). The formation rate of 3-deoxygluconone and 3,4-dideoxygluconone increased with the increase of temperature in the glucose-only system, especially when the heating temperature was lower than 105℃ (Table 1). At the 5 given temperatures, the promotion effect of high temperature on the formation of 1-deoxyglucosone, 3-deoxygluconone, and 3,4-dideoxygluconone was significantly in glucose-Glu system. The formation rate of glucosone was irregular with the change of heating temperature in both glucose-only and glucose-Glu systems, which might be related to the complexity of the Maillard reaction under the test condition. In the glucose-only system, the promotion effect of high temperature on the formation of 1-deoxyglucosone was significantly at 95-105℃, while the higher temperature reduced the formation rate of 1-deoxyglucosone (Table 1). The data of 5-hydroxymethylfurfural are from our previous work (Zhang et al., 2019).

| Glyoxal, methylglyoxal, and diacetyl formation and effect of glutamic acid
The formation rates of shorter chain (C 2 , C 3 , and C 4 ) α-dicarbonyl compounds glyoxal, methylglyoxal, and diacetyl were different at five different heating temperatures in glucose-only system (Figure 1 (1G-1I and 2G-2I)). The highest contents of methylglyoxal and diacetyl were observed with heating at 105℃, 6 hr, whereas the highest content of glyoxal was observed with heating at 95℃, 6 hr in glucoseonly system. When heating temperature was higher than 105℃, the formation rates of methylglyoxal and diacetyl were decreased in glucose-only system. Martins et al. reported that the content of methylglyoxal increased by heating time (within 4 hr) at 100℃ in heated glucose/glycine equimolar solution (Martins & Van Boekel, 2005).
The changing trend of methylglyoxal content (within 4 hr of heating) in the study of Martins et al. was consistent with that of methylglyoxal content in this study. Glyoxal and methylglyoxal, as intermediates of glycosylation and lipid peroxidation, were more likely to take part in secondary reactions and converted into advanced glycosylation products, advanced lipid peroxidation products, and secondary products at high temperatures (Andrewes, 2012). This might be the reason that the highest contents of glyoxal and methylglyoxal were not observed at the highest temperature in this study. From Figure 1 (1G), the generation of glyoxal was negatively correlated with heating temperature in glucose-only system. The highest contents of glyoxal were observed with heating at 95℃, 5 hr in the glucose-Glu system (Figure 1 (2G)). Glyoxal formation rate decreased with heating at 95℃, 5 hr in glucose-Glu system. It was worth mentioning that the formation rate of diacetyl also decreased with heating at 95℃, 5 hr. The reasons for this result remained to be further studied. The contents of methylglyoxal were the highest observed with heating at 95℃, 6 hr in the glucose-Glu system (Figure 1 (2H)).
The kinetics analysis results of 3 shorter chain α-dicarbonyl compounds formation in the system of glucose-only and glucose-Glu were shown in Table 1 and 2, respectively. At the 5 given temperatures, the glyoxal, methylglyoxal, and diacetyl formation both exhibited zero-order kinetics when heated for 1-6 hr in glucose-only system. The promotion effect of high temperature on the formation of methylglyoxal and diacetyl was significantly at 95-105℃, while the higher temperature reduced the formation rate of methylglyoxal and diacetyl (Table 1). However, high temperature can inhibit the formation of glyoxal (95-110℃) in the glucose-only system ( Table 1). The methylglyoxal formation also exhibited zero-order kinetics, but the promotion effect of high temperature on the formation of methylglyoxal was not significantly in the glucose-Glu system ( Table 2). The glyoxal formation was fitted to equation (1) at 90℃ (1-6 hr), while equation (2) at 95-105℃ (1-4 hr) in the glucose-Glu system ( Table 2). The fitting effect of the first-order kinetics on glyoxal formation at 110℃ (1-6 hr) was not ideal (R 2 = 0.7474) ( Table 2). The diacetyl formation was fitted to equation (2) at 90℃ (1-6 hr) and 95℃ (1-5 hr), while equation (2) at 100-110℃ (1-6 hr) in the glucose-Glu system (Table 2). And the promotion effect of high temperature on the formation of diacetyl was significantly. From Tables 1 and 2, the formation rate of glyoxal, methylglyoxal, and diacetyl had a tendency to decline with the increase of temperature, except for diacetyl in glucose-Glu system, possibly due to the high volatility of these compounds.

| HMF formation and effect of glutamic acid
As one of the major degradation products of carbohydrates, the HMF was produced unexpectedly during thermal treatment of carbohydrate-containing food (Zhang et al., 2019). Compared with glucose, the conversion of fructose to HMF was efficient, and fructose could synthesize HMF without catalyst. This might be caused by the energy barrier of HMF formed by isomerization of glucose to fructose and dehydration of the intermediate was lower than that of other pathways (Mayes et al., 2015). Fructose tended to form HMF via the fructofuranosyl cation pathway in the presence of glutamic acid. Based on our previous studies on the content of HMF in the glucose-Glu system (Zhang et al., 2019), the formation of HMF in glucose-only and glucose-Glu systems was positively correlated with temperature and time. The highest content of HMF in glucoseonly system was 5.73 µg/ml (110℃, 6 hr) (Figure 1 (1J)). The content of HMF in glucose-Glu system was much higher than that of glucose-only system. The contents of HMF in glucose-Glu system were 920.83 (90℃, 6 hr), 291.47 (95℃, 6 hr), 90.46 (100℃, 6 hr), 80.13 (105℃, 6 hr), and 63.56 times (110℃, 6 hr) as high as that of glucoseonly system, respectively. Based on situ NMR analysis, Li et al. have confirmed that the transformation of fructose to HMF was a highly selective reaction that proceeded through the cyclic fructofuranosyl intermediate pathway (Li et al., 2010).
The kinetics analysis results of HMF formation in the system of glucose-only were shown in Table 1. In our previous studies, it can be known that the HMF formation exhibited first-order kinetics at 90-105℃ (1-6 hr), while zero-order kinetics at 110℃ (1-6 hr) in glucose-Glu system (Zhang et al., 2019). In contrast with the glucose-Glu system, the HMF formation exhibited zero-order kinetics at 95℃ (2-6 hr), while first-order kinetics at 100-105℃ (2-6 hr) and 110℃ (1-6 hr) in glucose-only system. The promotion effect of high temperature on the formation of HMF was significantly at the five given temperatures. Although there were differences in the models of HMF fitted under different heating conditions, the formation of HMF increased with the increase of heating time and temperature.
In summary, based on the kinetics analysis, the content of threefifths of the tested compounds increased linearly with time. The content of 1-deoxyglucosone increased logarithmically at 95-110℃ over reaction time in the glucose-only system. The exponential equation

| CON CLUS ION
The effect of glutamic acid on the formation of α-dicarbonyl compounds in thermal reaction was investigated in this work. The content of 3-deoxyglucosone was significantly higher than 6 α-dicarbonyl compounds at 90-110℃, 0-6 hr in both glucose-only system and glucose-Glu system. The glutamic acid promoted the content accumulation of 1-deoxyglucosone, diacetyl, methylglyoxal, and glyoxal, whereas inhibited the content of 3-deoxyglucosone and 3,4-dideoxyglucosone. The accumulation of 7 α-dicarbonyl compounds was positively correlated with time in the two different tested systems.