Response Surface Methodology for Optimization of L-Arabinose/ Glycine Maillard Reaction through Microwave Heating

L-Arabinose is a low-calorie sweetener that inhibits sucrose absorption by inhibiting sucrase activity in the human intestinal tract. Response surface methodology (RSM) was applied to optimize the processing parameters of the L-arabinose/glycine Maillard reaction to improve the browning degree and antioxidant activity of Maillard reaction products (MRPs) through microwave heating. The eﬀect of heating time, volume ratio of propylene glycol to double distilled water (ddH2O), and pH on MRPs was evaluated. A change in the volume ratio of propylene glycol to ddH 2 O, heating time, and pH was associated with a largely changed browning degree and reducing power of the MRPs. RSM predicated optimum conditions that under substrates of L-arabinose/ glycine at a ratio of 2:1 (w/w) and concentration of 10% (w/v), a heating time of 7.44min, volume ratio of propylene glycol to ddH 2 O 0.93, and pH 10.44 were optimum conditions for the Maillard reaction. The predicted data from the optimum reaction conditions coincided well with the experiment results. The main ﬂavor of MRPs is roasted aroma, and the emulsifying ability of MRPs was 0.367 at 500nm by microwave heating under the optimal Maillard reaction conditions. MRPs derived from L-arabinose and D-glucose had similar activities. However, a slightly greater activity was found with MRP derived from L-arabinose-glycine with a more volume. This study provided a new direction for the development of sweeteners in the future.


Introduction
e Maillard reaction is a series of complex reactions that occur between the free amino groups of amino acids, peptides or proteins, and carbonyl groups of sugar, especially reducing sugars, to produce the Maillard reaction products (MRPs) [1]. Under different reaction conditions, the reaction pathway and the mechanism will be greatly different, and most of the formed products have a specific flavor and color. MRPs contain volatile substances including lowmolecular-weight hydrocarbons, alcohols, aldehydes, ketones, esters, ethers, and heterocyclic compounds, and some large-molecular-weight materials containing polyphenols, peptide polymers, melanoidins, and so on. Melanoidins (melanoidin) or melanoidins (melanoprotein) are brown part of the Maillard reaction in the ultimate stage polymer and copolymer containing nitrogen and are part of the brown macromolecular substances. Melanoidins (melanoidin) or melanoidins (melanoprotein) are the most common high-molecular-weight materials that are part of the brown macromolecular substances [2]. e time for the completion of a traditional Maillard reaction is generally some hours, but that could be reduced to 2-10 minutes if microwave heating is properly applied [3,4]. Zhao et al. [5] studied the Maillard reactions of different neutral amino acids with glucose and fructose under microwave conditions and found that both acidic and alkaline amino acids had strong characteristics and strong flavor sensation.
e Maillard reaction in aqueous systems has been used in the manufacture of food pigments for a long time, and flavors and MRPs are needed in certain food products with a certain emulsification, such as instant coffee products. is study suggests the feasibility of obtaining new compounds with potentially desirable characteristics through the Maillard reaction in ethanolic systems. Further investigation in this area is likely to be valuable. Z.-c. Tu et al. [6] described that the UV absorbance, browning intensity, and antioxidant activities as well as the emulsifying activity and emulsion stability of the Maillard reaction products (MRPs) were increased in accordance with the raise of microwave treatment power and time. e reaction time of microwave treatment is much shorter than those using traditional methods, suggesting that microwave irradiation is a novel and efficient approach to promote the Maillard reaction (MR). e high-molecular-weight melanoidins prepared from glucose and different amino acids (asparagine, glycine, and arginine) have been found to possess a higher browning degree, reducing power, and antioxidant activity [7]. Yen and Tsai [8] evaluated the antioxidant activity of partially fractionated MRPs prepared by refluxing glucose and tryptophan at pH 11.0 and 100°C for 10 h, and the results showed that, compared with low-molecular-weight fractions, the high-molecular-weight fractions achieved a higher reducing power and antioxidant activity. erefore, the preparation of high-molecular-weight MRPs might be an alternative method to develop valuable high antioxidant products.
Response surface methodology (RSM) is of considerable value for the improvement and optimization of complex processes that elucidate the causality between explanatory variables and response variables [9]. Furthermore, RSM is one of the best experiment design methods to reduce the number of experimental trials needed to evaluate multiple parameters and their interactions to provide sufficient information for statistically acceptable results [10,11]. e objective of the present study is to optimize the Maillard reactions between L-arabinose and glycine using microwave heating to develop high antioxidant MRPs. e relationship between browning degree and reducing power of the MRPs will also be evaluated.
Microwave radiation uses a heating mechanism that rotates and vibrates the electric dipole of target molecules. e reaction time of microwave treatment is much shorter than those using traditional methods. is study applies the microwave heating energy to the L-arabinose and glycine in the mixed solvent propylene glycol and ddH 2 O. Short reaction time and less on the substrate concentration have better reducing power and emulsifying ability.

Chemicals.
Glycine of food grade was purchased from Shanghai Weihong Bio-Sci and Tech Co., Ltd. (Shanghai, China) and L-arabinose from Jinan Shengquan Biotechnology Co., Ltd. (Jinan, China). All other chemicals were of analytical grade and purchased from Shanghai Chemical Reagent Co., Ltd. (Shanghai, China).

Preparation of Maillard Reaction Products (MRPs).
Based on our preliminary experiments, a L-arabinose: glycine ratio of 2 : 1 (w/w) and substrate concentration of 10% (w/v) were used in the Maillard reaction. is reaction model system has been reported by Peterson, Tong, Ho, and Welt [12] and modified in this study. Briefly, L-arabinose (2g) and glycine (1g) were dissolved in a certain amount of ddH 2 O and propanediol. e pH of the solution was adjusted with 5M NaOH and 1M HCl; 27 ml solutions were transferred to a 100 ml beaker and heated in a microwave oven with 500 W of power level (Galanz WD800 T model, 2450 MHz, 800 W, 305 mm × 508 mm×395 mm, Shunde, China) for a certain time. After heating, samples were collected and placed in an ice bath to cool down to stop the reaction before they were stored at a 4°C fridge. ese samples were referred to as MRPs.

Determination of Browning Degree (BD).
Samples of 1.0 ml MRPs were diluted to 100-fold with the addition of ddH 2 O. e browning degree was determined by measuring the absorbance at 420 nm using a UV-Vis spectrophotometer [13,14] (UV-2100 UNICO spectrophotometer, Jiangsu Scientific Instruments and Materials Co., LTD, Jiangsu, China).

Experimental Design.
According to our prior experimental findings, the most influential factors on the BD and RP of MRPs are heating time (factor A: 5 min, 7 min, 9 min), volume ratio of propylene glycol to ddH 2 O (factor B: 0.5, 1, 1.5 v/v), and pH (factor C: pH 8, pH 10, pH 12). e effects of interactions of these three factors were also considered in the RSM experimental design. e "Design-Expert" software (version 8.0.6, Stat-Ease, Inc., Minneapolis, USA) was used to generate the Box-Behnken experimental designs. e independent variables were heating time (A), volume ratio of propylene glycol to ddH 2 O (B), and pH (C). Each independent variable had coded levels of − 1, 0, and 1 and was constructed based on a 3 3 factorial design. Five replications of the central points were run, leading to 17 sets of experiments, allowing each experimental response to be optimized. e experimental designs of the coded factors and actual levels of variables are shown in Table 1. e two responses (Y) were browning degree (Y 1 , A 420nm ) and reducing power (Y 2 , A 700nm ). e response functions Y 1 and Y 2 were related to the coded variables (A, B, C) by a second-degree polynomial equation using the method of least squares: where Y is the response calculated by the model; A, B, and C are coded variables, corresponding to heating time, volume ratio of propylene glycol: ddH 2 O, and pH, respectively; a 1 , a 2 , and a 3 are the linear; a 4 , a 5, and a 6 are the quadratic, and a 7 , a 8 , and a 9 are the cross-product effects of the A, B, and C factors on the response.
Analysis of variance (ANOVA) was performed. ANOVA tables were generated, and the effect and regression coefficients of individual linear, quadratic, and interaction terms were determined. e statistical significance of the regression coefficients was determined by using the F-test, and the applicability of the model was checked with significant coefficients of determination (R 2 ) and the coefficient of variation (CV) values. e optimal processing conditions were obtained by using graphical and numerical analysis based on the criterion of desirability.

Qualitative Analysis of Volatiles Compounds MRPs by GC/ MS.
A manual solid-phase microextraction (SPME) device and divinylbenzene/carbo-xen/polydimethylsiloxane (DVB/ CAR/PDMS) fibres (100 μm film thickness) were obtained from Supelco Co. (Bellefonte, PA, USA). e fibre was conditioned for 1 h at 270°C as recommended by the manufacturer. Five milliliter of MRPs was placed in a 10-ml vial closed by a PTFE/silicone septum (Supelco). Before the extraction process, a time of 30 min at 40°C was requested for headspace equilibration. After 1 h of fibre exposure in the sample headspace, the fibre was thermally desorbed in a gas chromatography (GC) injection port for 20 min. e injector was set at 250°C and operated in a splitless mode for 3 min. e GC/MS analyses were carried out using an Agilent 7890A gas chromatograph(NYSE : A, USA), equipped with an FID and coupled to a quadrupole Agilent 5975C Network mass selective detector (NYSE : A, USA). e gas chromatography was equipped with a fused silica capillary column HP-INNOWAX (PEG, 60 m × 0.32 mm i.d. film thickness � 0.25 μm, NYSE : A, USA). e carrier gas was helium (head pressure for both columns � 25 psi); oven temperature was programmed from 60°C (2 min) to 200°C at 2°C/min and then at 5°C/min to 230°C and held isothermal for 5 min. e FID temperature was set at 250°C, and the temperatures of the ion source and the transfer line were 170 and 280°C, respectively. Energy was set at 70 eV and mass range of 35～350 amu. Qualitative method is done by comparing the spectrum of the detected substance with the standard spectrum in the NIST 05al and Wiley 7n databases (Agilent, USA). By comparing retention time with standard products, the results were compared with the retention index (RI) of standard substances, and those without standard substances were compared with RI in the reported literature.

Determination of Reducing Power (RP).
e reducing power of the MRPs was determined according to the method previously reported [14,15] with slight modification. Samples of 1 ml MRPs (100-fold dilution) were mixed with 1.0 ml of 0.2 M sodium phosphate buffer (pH 6.6) and 1.0 ml of 1% potassium ferricyanide (K 3 Fe(CN) 6 ) in a test tube and sealed. e reaction mixtures were incubated in a water bath at 50°C for 20 min, followed by rapid cooling to 25°C. e solution of 1.0 ml was further mixed with 1.0 ml ddH 2 O and 200 μl 0.1% FeCl 3 (w/v), and the absorbance was read at 700 nm with a spectrophotometer.
e reducing power of MRPs was expressed as absorbance (A 700 ) using the mean values of three determinations.

DPPH Radical Scavenging Activity of MRPs.
DPPH radical-scavenging activity of MRPs was determined according to the method of Yen and Hsieh [16] with a slight modification. An aliquot of 80 μl MRP sample was diluted with 320 μl of ddH 2 O and 2 ml of 0.12 mM DPPH in methanol was added. e solution was then mixed vigorously and allowed to stand at room temperature in the dark for 30 min. e absorbance of mixtures was measured at 517 nm on the UNICO UV-2100 spectrophotometer. e control was prepared in the same way, except that ddH 2 O was used instead of MRP samples. For the blank sample, the assay was conducted in the same way, but methanol was added instead of DPPH solution. e percentage of DPPH radical-scavenging activity is calculated as follows: radical scavenging activity(%)

Mathematic Model of Maillard
Reaction. RSM experiments of L-arabinose/glycine were carried out in a random order. Values obtained from the Maillard reaction system are given in Table 1, while characteristics of the model for BD and RP are shown inTable 2 and 3, respectively. e ANOVA confirmed the adequacy of the statistical models since their Prob > F values were less than 0.05 and statistically significant at the 95% confidence level. e models presented high determination coefficients (R 2 ) and low coefficients of variation (CV). ese values are listed as follows: R2 � 0.9967 and CV% � 3.4 for BD; R2 � 0.9957 and CV% � 5.01 for RP. ese results indicated a good precision and reliability for the experiment. e fitted model equations are as follows:    volume ratio had a negative but not significant effect equation (3) on BD. However, the linear terms and quadratic terms of volume ratio have a significantly negative effect on BD. Figure 1(a) shows the dependence of BD on the reaction factors of heating time (time) and the volume ratio of propylene glycol to ddH 2 O (Ratio) at a fixed pH. It is clear that at a constant pH value and heating time, BD increased slightly with the increase in the volume ratio. It was also observed that the BD increased quickly at the beginning of the experiment and then decreased slightly with the extending of heating time at a fixed volume ratio and pH (Figure 1(a)). e variation is curvilinear in nature. e variation in the BD with heating time and pH at a constant volume ratio of propylene glycol to ddH 2 O is presented in Figure 1(b). It is evident that at a fixed volume ratio and pH, the BD increased rapidly with heating time at the first stage and then decreased slowly. At a fixed volume ratio and heating time, the BD increased with slow increment of pH but decreased slightly in later stages. Figure 1(c) shows the effects of the pH value and the volume ratio of propylene glycol to ddH 2 O on BD at a fixed heating time. e BD increased slowly at the beginning and decreased slightly afterwards with an increased volume ratio at a constant pH and heating time. e same trend can be seen for the variable of pH at a fixed heating time and volume ratio. e results indicated that the linear effects of pH value, volume ratio of propylene glycol to ddH 2 O, and heating time were dominant over the interaction terms. e interaction effects between heating time and volume ratio were not significant (p ﹥0.05), but they slightly influenced the BD. e quadratic effects were significantly negative to the browning degree (p ﹤0.05).

Reducing Power.
e reducing power (RP) of the MRPs has a positive linear effect on the variation in heating time, volume ratio of propylene glycol to ddH 2 O, and pH, as shown in Table 3. e quadratic effects also have a significant effect on the RP of MRPs. e interaction terms of the variables of volume ratio and pH were found to have significant effects on RP. Figure 2(a) presents the value of RP with the variation of heating time and the volume ratio of propylene glycol to ddH 2 O at a given pH. e RP value increased rapidly with the heating time at a given volume ratio and pH, while at a fixed volume ratio and heating time, the RP value slightly increased with increased pH values (Figure 2(b)). However, the RP value increased slightly at the beginning and decreased slowly with the variation in the volume ratio at a fixed heating time and pH Figure 2(c).
It has been widely recognized that melanoidins from Maillard reactions possess high reducing power [18,19]. e present experimental results showed that the browning degree and reducing power of MRPs from the L-arabinose/ glycine system have a good positive correlation with each other (higher browning degree and higher reducing power) and are consistent with the reported results of Yamaguchi, Koyama, and Fujimaki [7].

Optimization and Experimental Validation.
e optimal processing parameters were obtained from SRM to the preparation of L-arabinose/glycine MRPs with high browning degree and reducing power. Browning degree can be optimized from the contour plot figures of Figure 1(a)-1(c). e pH region is in the range of 8~12, the volume ratio of propylene glycol to ddH 2 O is 0.5-1.5, and the heating time is 5-9min. e model describing the optimum conditions for BD was as follows: a heating time of 7.44 min; a volume ratio of propylene glycol to ddH 2 O of 0.93; and a pH of 10.44. Reducing power can be optimized from contour plots of Figure 2(a)-2(c), and the model describing the optimal conditions for RP were the same to those of BD: a heating time of 7.44 min; a volume ratio of propylene glycol to ddH 2 O of 0.93; and a pH of 10.44. e highly coincident data also suggested that the BD and RP of the MRPs are positively correlated with each other. erefore, the responses (Y 1 and Y 2 ) of the optimal conditions for both BD and RP could be expressed using the same model. e responses (Y 1 and Y 2 ) calculated from the final polynomial functions were a BD of 0.405 at 420 nm and a RP of 0.268 at 700 nm. e Maillard reaction conditions were experimentally validated, and the results were a BD of 0.461 ± 0.02 at 420 nm and a RP of 0.319 ± 0.01 at 700 nm. Based on the relative deviation values of BD (SD% � 3.96) and RP (SD% � 3.61), it could be tentatively concluded that the methodology employed for the optimization of the Maillard process conditions was satisfactory and that the surface responses obtained by the full experimental design were suitably validated. Table 4 shows the volatile chemicals of MRPs produced by microwave heating under the optimum Maillard reaction conditions obtained by the SRM design. e major compounds identified in the MRP sample were pyrazines (peak area-� 21.27%), furans (peak area � 4.49%), alcohols (7.8%), pyrroles (peak area � 10.34%), acids (peak area � 12.88%), ketones (3.51%) and phenols (peak area � 16.97%). The types of pyrazines of the result are more than the ascorbic acid/ glycine Maillard reaction [20]. And alkylated pyrazines are an important group of flavor compounds, which contribute substantially to the unique roasted aroma of various food products [21]. Pyrazines came from the Maillard reaction involving glycine [22]. Pyrroles were the key antioxidant activity compounds [23]. So the MRPs have the pyrazines characteristics of pyrazines flavor and fine antioxidant activity.

Emulsifying Ability Analysis of MRPs.
e emulsifying ability of MRPs was 0.367 absorbance at 500 nm by microwave heating under the optimal Maillard reaction conditions, which was very close to the emulsifying ability of the casein-glucose Maillard reaction product (the emulsifying ability was 0.397 absorbance at 500 nm) reported by Gu, Abbas, and Zhang [24]. So the MRPs have a good emulsify ability like the casein-glucose Maillard reaction product. We can also conclude that some hydrophilic and lipophilic substances must be formed by the reaction products. If it can

Changes in DPPH Radical-Scavenging
Activity. e DPPH radical was scavenged by MRPs by the donation of hydrogen to form a stable DPPH-H molecule. And the color changed from purple to yellow by the acceptance of a hydrogen atom from MRPs, and it became a stable diamagnetic molecule [25]. As shown in Figure 3, it can be observed that DPPH radical-scavenging activity of the MRPs of L-arabinose-glycine and D-glucose-glycine is positively related to the linear effect of heating time (Time), volume ratio of propylene glycol to ddH 2 O (ratio), and initial pH (pH) (p < 0.05). MRPs derived from L-arabinose and D-glucose had similar activities. However, a slightly greater activity was found with MRP derived from L-arabinose-glycine with more volume.

Conclusion
e response surface methodology has been demonstrated as a useful tool to optimize the reaction conditions of heating time, volume ratio of propylene glycol to ddH 2 O, and pH to improve the browning degree and reducing power in the L-arabinose/glycine Maillard reaction product. e coefficients of determinations and R 2 values showed a good fit of the models with the experimental data at the 95% confidence level. e different conditions for Maillard reaction revealed that heating time had the significant effect on browning degree and reducing power using microwave heating, while the other two variables (volume ratio of propylene glycol to ddH 2 O and pH) had an optimum zone for emulsifying ability and DPPH scavenging ability. ese results were well fitted with the experimental data, and the obtained models have the potential to be used to maximize the antioxidant activity of the Maillard reaction products. e peak area of pyrazines is 21.27%, and alkylated pyrazines are the unique roasted aroma of various food products, so the main flavor of MRPs is roasted aroma. MRPs have better reducing power and emulsifying ability. e reaction time of microwave treatment is much shorter than those using traditional methods, so microwave irradiation is a highly efficient approach to promote MR and has huge potential in MR. is study provides a new direction for the development of sweet flavor.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.