Effect of Maillard reaction conditions on the solubility and molecular properties of wheat gluten–maltose conjugates

Abstract In this experiment, the conjugation reaction between gluten and maltose via Maillard reaction under dry‐heated condition was studied. The process conditions for the preparation of protein–maltose conjugates with optimum solubility were optimized by using Box‐Behnken model. The conjugation reaction and the structure changes of the protein–maltose conjugates were confirmed by infrared spectroscopy (FTIR) and scanning electron microscopy (SEM). The results showed that the process conditions for the preparation of protein–maltose conjugates with optimum solubility were as follows: temperature 50.72°C, time 1.92 days, and gluten/maltose (W/W) 267.36%. The infrared spectroscopy showed that the structure of the modified protein had a very obvious change, including the decrease in β‐fold and β‐turn and the increase in α‐helix at a certain degree. But the conjugation reaction has little effect on the irregular coiled structure. The scanning electron microscopy showed that the microstructure of gluten is small grainy, but gluten–maltose conjugate looks sheet with bigger volume.

or shear (Galazka, Dickinson, & Ledward, 2000;Haykawa, Linko, & Linko, 1996). Due to the appearance of the potential health hazards or dangerous products, most of the chemical modifications are not used in the food industry. The enzymatic treatments often lead to bitter taste, affecting the food's flavor. Therefore, more appropriate method needs to be applied to protein modification.
Protein and saccharide are two types of biological macromolecules in the food system, the main factors influencing food function and texture. From the beginning of the 90 s, several research groups in Europe and Japan began to covalent complexes of protein-saccharide.
The ε-lysyl amino groups of protein and the reductive terminal carbonyl group of carbohydrate were linked by Amadori-type linkage to form protein-saccharide complex (Kato, Minaki, & Kobayashi, 1993).
The functional properties of protein-saccharide conjugates are mainly based on the protein. The introduction of saccharide is modified or enhanced functions for proteins. At present, there are many studies using different sugars for protein modification, including monosaccharides and polysaccharides (Hiller & Lorenzen, 2010;Sheng et al., 2020).
The glycosylation of protein due to the mild reaction conditions does not add the foreign chemicals and is generally considered suitable for application in the food industry (Shepherd, Robertson, & Ofman, 2000). Because of the important role of Maillard reaction in food stability, flavor development, nutrition, and health, it is important to develop reasonable methods to reduce the adverse consequences of this reaction and to optimize the beneficial effects, while trying to establish conditions under which the highest yield of carbohydrate binding proteins can be produced (Scaman, Nakai, & Aminlari, 2006). Therefore, the purpose of this study was to evaluate the effects of temperature, time, and the addition ratio of reactants on the degree of glycosylation of gluten and maltose to obtain mixtures with different glycosylation levels and to evaluate their functional properties.

| Materials
Wheat gluten was obtained from Fengqiu County HuaFeng powder industry Co., Ltd, and contained 73.53% protein, 0.73% fat and 11.14% water content, and 1.04% ash content; maltose was purchased from Shanghai Jingchun biochemical technology Co. Ltd; corn oil was purchased from local markets and used without further refinement; and all other chemicals were of analytical grade.

| Preparation of protein-maltose conjugates
Gluten (1.0 g) was dissolved in 100 ml 0.05 mol/L sodium phosphate buffer with pH 12. Maltose powder (5, 3.6, and 3 mg) was added into gluten solution, respectively. The sample solutions were stirred at room temperature with a magnetic stirrer at 150 g for 2 hr to dissolve the mixture completely. The solutions were then freeze-dried.
Each experiment included a control sample without maltose. Freezedried powders were reacted at 45, 50, and 55°C at 79% relative humidity (over saturated potassium bromide) for up to 1, 2, and 3 days.

| Solubility
The solubility was determined by biuret method with bovine serum albumin as a standard. Native and glycated gluten were diluted in distilled water (1%, W/V). The samples were set to stand for 30 min.
And then samples were centrifuged for 10 min at 300 g . Protein content in the supernatant was determined. 1 ml of the supernatant (or water as the blank) put into the test tube, and 4 ml of biuret reagent was added. After that, the solution was mixed gently and thoroughly immediately. Incubation was carried out at 25°C for 30 min. Finally, the solution absorbance of each tube was measured at 540 nm by UV-VIS spectrophotometer.The measurement was carried out in triplicate for each sample.

| Fourier transform infrared spectroscopy (FTIR)
Fourier transform infrared spectroscopy was determined by using Madhav P. Yadav's method (Yadav, Strahan, Mukhopadhyay, Hotchkiss, & Hicks, 2012). The gluten and graft were thoroughly dried, and then, the samples were mixed with KBr at a mass ratio of 1:300 and then thoroughly ground and pressed into thin slices.
An infrared spectrometer was used to perform a full-band scan (400 ~ 4000 cm −1 ). The sample was scanned for 12 times.

| Experimental design
The Box-Behnken model in software Design Expert 7.1 was used to achieve the optimal condition for Maillard reaction. The independent variables were reaction temperature (X 1 , °C), gluten/ maltose ratio (X 2 , %), and reaction time (X 3 , d), while the dependent variable was solubility (Y, mg/ml). Replicates were performed for each experiment, and the average values were recorded as the response. The experimental Y value was fitted to the following quadratic equation:

| The data analysis
The data in the test were the mean value of three tests, and the software Design Expert 7.1 was used to analyze the regression model.

| Statistical analysis and the model fitting
The test is carried out according to the experimental design scheme, and the results are shown in Table 2. Using the software Design Expert 7.1 to perform regression analysis of solubility measured under different conditions in Table 2, the regression equation of solubility was obtained by software regression fitting: Then, ANOVA was used to analyze the significance and applicability of the experimental model, and statistical summary was provided (Table 3). As seen in Table 3, the p value of the model was much <.01, which indicates the significance of the model for Maillard reaction.
This shows that the model coincides with the experimental data very well. The linear variables X 1, X 2 , and X 3 showed statistically significant influences (p < .01); and the quadratic variables X 1 2 , X 2 2 , and X 3 2 also significantly influenced the solubility (p < .01). The interaction coefficients between the independent variables also showed statistically significant influences (p < .05). However, the p value for the model mismatch is not significant (>.05), indicating other factors have little effect on the model. Therefore, it was concluded that the model can be used to analyze and predict the dependent variables correctly.

| Response surface analysis
With solubility as the dependent variable, the 3 days response surface was drawn (Figure 1). Figure 1 shows
And Amide I (1600-1700 cm −1 ) can reflect the changes of secondary structure of glycosylated and unglycosylated proteins. This secondary structure is mainly composed of bed sheets (1600-1640 cm −1 ), bed helices (1650-1670 cm −1 ), and bed turns (1680-1685 cm −1 ). The bands at 1640-1650 cm −1 belong to the side chain structure (Yaping et al., 2014;Liu et al., 2012). Figure 2 shows the FTIR spectra of conjugate and gluten. The spectral was processed by the software Peak fit to carry out the convolution, the peak, and the fitting processing and to get the infrared fitting curve of the amide I band (Figure 3).
Then, the peak area is calculated to get the proportion of different secondary structure (Table 5). Table 5 shows the proportion of α-helices in the conjugate increased significantly, and β-sheets had a larger decline. But the proportion of β-turns and random coil had little change. In theory during the heating process, the protein peptide chain with thermal shock, the secondary bonds keeping the spatial structure of the protein were destroyed, and the molecular order  infrared spectrum (Pearce & Kinsella, 1978). When proteins and  absorption peak in the 1200-1000 cm −1 (Yaping et al., 2014). As shown in Figure 2, the conjugate had stronger absorption peak in 3700-3200 cm −1 and 1200-1000 cm −1 compared to gluten, which confirmed sugar molecules accessed protein molecules by covalent bonding.

| SEM analysis of gluten-maltose conjugates
The scanning electron micrographs (SEM) of gluten and gluten-maltose conjugate, taken at 500× magnification, respectively, are shown in the Figure 4. It reveals what happens when maltose and proteins are dry-heated under controlled conditions. The electron micrograph of the untreated gluten showed that the surface of the original gluten particles was massive, and the volume of the massive tissue was small. But the massive tissue of conjugates prepared by dryheating reaction under controlled conditions increased significantly.
This proved that the gluten is strongly linked to the maltose making their conjugates. It looks very clear that gluten protein molecules are intimately associated with maltose molecules giving their closely associated compact and nonhomogeneous microstructure, which further confirms a covalent linkage between them.

| CON CLUS IONS
The effects of reaction temperature, time, and gluten/maltose ratio (W/W) on the solubility of gluten-maltose conjugate were studied.
And the reaction conditions were optimized by using Box-Behnken model. The analysis of variance showed that the model was fit well with the experimental data and the experimental error was small.
By the verification of the model, the best reaction conditions were as follows: The temperature was 50.72°C, the gluten/maltose ratio W/W was 267.36%, and the reaction time was 1.92 days. And the solubility was 1.60 mg/ml under the optimized reaction condition.
The grafting reaction was proven by FTIR and SEM. FTIR also demonstrated that the incoming sugar chain made the protein spatial structure change, resulting in the change of the solubility of gluten.
Scanning electron microscopy showed the microstructure of gluten-maltose conjugate and gluten is obviously different. The volume of the massive tissue of gluten was small, but the massive tissue of gluten-maltose conjugate increased significantly, proving that sugar binds to protein. In this work gluten-maltose conjugates showed better solubility than native gluten. Therefore, grafting reaction between gluten and polysaccharide is an efficient way to develop new use of gluten in food. Further investigations on physicochemical properties and functional characteristics of gluten-polysaccharide conjugates will be conducted, such as TGA analysis, DSC analysis, emulsifying activity and emulsifying stability analysis, and foaming property and foaming stability analysis, to achieve more improvements in protein glycosylation and elucidate the grafting mechanism.

TA B L E 5
The secondary structure of gluten and gluten-maltose conjugate

ACK N OWLED G M ENTS
This work was supported by Henan Key Laboratory of Cereal Resource Transformation and Utilization (PL2017010).

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

E TH I C A L A PPROVA L
This study does not involve any human or animal testing.

I N FO R M ED CO NTENT
Written informed consent was obtained from all study participants.