Evolution of nonenzymatic browning during the simulated Msalais‐production process in models of grape juice

Abstract Msalais is a wine fermented from boiled grape juice. Nonenzymatic browning (NEB) greatly affects the quality of Msalais, but to date its mechanism has not been systematically analyzed. In the current study, the evolution of NEB during Msalais production was investigated using models of grape juice. 5‐Hydroxymethylfurfural (5‐HMF), browning index (BI), yellowness index (YI), sample absorbance at 420 (A420), and b* increased during heating, with a clear transition point at 110 min. The Maillard reaction (MR) was the major contributor to NEB. Vitamin C (VC) facilitated NEB in the late stage of heating. During heating that lasted over 130 min, glucose contributed to NEB more than fructose, while the reverse was true for heating lasting less than 130 min. Proline (Pro) was the most important amino acid in facilitating NEB. BI and A420 decreased during fermentation, while increasing slightly during wine storage. In conclusion, this study identified the evolution of NEB during the Msalais process, which will facilitate the control of traditional Msalais production for improved wine quality.

achieved if the grape juice is not sufficiently heated or, conversely, is overheated. Traditionally, the color of Msalais varies from light red to dark brown, and the aroma of Msalais is of a caramel odor (Zhu et al., 2013). However, how NEB evolves during the Msalais process has not, to date, been systematically analyzed.
Furthermore, polyphenols, fatty acids, and other carbonyl compounds can participate in the MR, even if formed via oxidative reactions (Bozkurt et al., 1999;Kader et al., 1997;Sajib & Undeland, 2020;Wang et al., 2021;Zamora & Hidalgo, 2005). In fact, molecules other than reducing sugars participate in the MR. This is particularly true for some lipid oxidation products, phenolic compounds, or any other molecules with carbonyl groups (Sajib & Undeland, 2020). MR and VC degradation, rather than caramelization, are reportedly the main NEB reactions during the concentration of grape juice (Bozkurt et al., 1999). In our previous work, the typical color and flavor of Msalais are formed as a result of NEB during heat concentration, and are due to MR, caramelization, and amino acid degradation (Mengmeng et al., 2018;Zhang, 2016). Also, during fermentation and aging, polyphenol oxidation may be an important pathway for additional NEB (Zhang, 2016). Nevertheless, the available information on various NEB reactions in different food matrices or models under different heating conditions emphasizes the complexity of NEB in fruit juice processing, which has hindered the development of an effective solution to mitigate browning (Paravisini & Peterson, 2018).
Thus, in this study, we focused on NEB evolution from the MR during the Msalais process in a model without polyphenols, as it is well understood that NEB caused by complex polyphenols is a complicated process.
Historically, the formation of brown color has been determined by using direct and indirect approaches. The former includes chemical methods for measuring the concentration of browning reaction products, such as 5-hydroxymethylfurfural (5-HMF) (Wang et al., 2006). 5-HMF is an intermediate of brown pigment formation during the MR, and the level of 5-HMF is strongly correlated with the degree of browning (Lee & Nagy, 1988;Wang et al., 2006).
Conversely, indirect approaches focus on registering the variation of color associated with NEB by measuring sample absorbance at 420 nm (A420), a wavelength widely used for the detection of browning pigments (Ibarz et al., 2000). Another indirect technique is based on the CIE L*a*b* coordinates. Furthermore, BI associated with fruit has been extensively developed to evaluate both NEB and enzymatic browning (Karabagias et al., 2018;Maskan, 2001;Pathare & Opara, 2013).
Color resulting from the concentration of grape juice is the most important characteristic of Msalais quality. However, no systematic studies into Msalais color development have been reported to date. Accordingly, with the increase in Msalais consumption, oenologists are becoming increasingly interested in understanding the ways in which browning reactions can be controlled to achieve a beverage with improved organoleptic and health properties. The current study was designed to analyze the evolution of NEB during simulated Msalais processing. We analyzed multiple indicators, including L*a*b*, 5-HMF, and A420, and used suitable model grape juice models to elucidate the processes that underpin color changes during Msalais production.

| Models with different main compositions of grape juice
The 'complete' model was designed as the our previous formulation to simulate the grape juice of Vitis vinifera Hetianhong (Zhu et al., 2017). The other models were prepared by subtracting specific

| Determination of the CIE color coordinates and BI
The samples were analyzed using a Hunter Lab model D25 L optical sensor (Hunter Associates Laboratory, Inc., Reston, VA, USA).
For the analysis, 60 ml of a sample was placed in a cylindrical optical cell. Reflectance values were obtained using a 45-mm viewing aperture. The colorimeter had been operating for 1 h prior to use and was calibrated using black-and-white reference plates obtained from the manufacturer. The reported data are the average values of five measurements. Tristimulus color parameters L*, a*, and b* (Karabagias et al., 2018) were determined. The yellowness index (YI, indicating the degree of yellowness) (Rhim et al., 1999) and the browning index (BI) were calculated by using the following equations (Maskan, 2001):
When necessary, appropriate sample dilutions were prepared with an optical density less than 1.5. Each measurement was performed in triplicate and the mean values were used to construct kinetic plots. The relative standard deviations of the triplicates were below 8.33%. Sample absorbance at 294 (A294) indicates the formation of The concentrations of the components in the grape juice models

Items Compounds Concentration
Sugars (g/L) Glucose 110 Fructose 110 Acids (  Vitamin ( Apart from sugars, amino acids, and VC, the 'complete' model contained phenolic acids and salts. NEB in the 'complete' model was not as pronounced as in the G + F + A and G + F + A + VC models, suggesting that phenolic acids and/or salts inhibited the reaction (Figure 1 b-f). In other models containing only one sugar (G + F model), amino acids (A model), VC model, or phenolic acids (P model), NEB was much less pronounced than in the three models of G + F + A, G + F + A + V, and 'complete' (Figure 1b-f). NEB in G + F and VC models was slightly more pronounced than that in A and P models (Figure 1b-f). These findings suggest that in model grape juice, the MR of amino acids and sugars was the dominant contributor to NEB and that caramelization of sugars and VC degradation only slightly contributed to NEB.

| Contribution of reducing sugars and amino acids to NEB
As shown in Figure  130 min, and a much greater consumption of glucose than fructose after heating for more than 130 min, especially in the late stages of heating.
The degree of NEB in models containing individual amino acids and reducing sugars is shown in Figure 2b. In these models, the initial concentration of amino acids differed according to the composition of the natural grape juice (Figure 2b). The proline (Pro) model exhib-

| NEB during fermentation in a concentrated 'complete' model
BI and A420 decreased rapidly during the first 7 days of fermentation, from 21 to 10.8, and 0.65 to 0.24, respectively. Then, BI slowly increased to 13.5, while A420 slowly decreased to 0.21 (Figure 3a).
Even though BI and A420 values decreased during fermentation,

| Discussion
NEB contributes to the desired aroma, taste, and color of food to meet the expectations of consumers (Ismarti et al., 2021). In addition to this, the biological activities of some NEB products generated during the heating of food positively impact human health (Arihara et al., 2019;Dhungel et al., 2020;Nowak et al., 2021). Yet, on the other hand, the toxicity and undesired health effects of NEB products have been reported (Li et al., 2021;Nowak et al., 2021;Parisi & Luo, 2018). Currently, controlling NEB is critical for maximizing profit and avoiding formation of harmful compounds during food processing (Nowak et al., 2021;Parisi & Luo, 2018). During Msalais processing, regulating NEB is essential to take advantage of the MR to improve the sensory characteristics and to avoid harmful effects, such as eliminating or reducing the level of 5-HMF. The data presented in this study describe the specific evolution of NEB in a model of heated grape juice, and form a firm basis for controlling, and thereby improving the traditional technique of Msalais production.
In the current study, it is well established that NEB increases with heating time of the food matrix. The level of 5-HMF, BI, YI, A420, and b* all increased during heating of the 'complete' model, with one obvious transition point at 110 min. On the other hand, the a* value displayed a downward trend during the concentration stage. These observations agree well with the reported phenomenon of NEB, that is, a change of color, shifting from red to yellow and brown (Pathare & Opara, 2013). Besides the heating time, NEB was affected by the specific matrix components. The MR was the major contributor to NEB. In fact, the MR effectively takes place in boiling grape juice at temperatures >50°C and at a pH range of approximately 4-7 (Ajandouz et al., 2001), and NEB is still increasing in a concentrated grape juice at a very low temperature such as 5°C (Bakeshlou et al., 2020). VC is easily oxidized and decomposed through aerobic decomposition and anoxic decomposition to form dehydroascorbic acid, and then dehydrated to produce reductive ketones. These reductive ketones participate in the intermediate and final stages of the MR (Hrynets et al., 2019). In the current study, VC facilitated the MR during the late stage of heating instead of during early stage of heating, and this may be the high temperature to inhibit pathways of oxidative decomposition of VC in a degree . There are few reports on caramelization that occurred during the heating of juice or jam. That is because the boiling point of grape juice is far lower than the sugar melting point (>120°C) under acidic (pH 3) or alkaline (pH 9) conditions (Hrynets et al., 2019). In the current study, phenolic acids and/or salts inhibited the NEB reaction in boiling models (Figure 1b-f). In boiled grape juice, fructose is more reactive than glucose with respect to both 5-HMF accumulation and NEB, during heating for 10 days at 55, 65, and 75°Cat pH 3.5 (Göğüş et al., 1998). Similarly, in sugar-catalyst systems, fructose is the major reactant, leading to a more rapid formation of 5-HMF than sucrose and glucose at pH 3.5 (Lee & Nagy, 1990). We showed here that the heating time is another important factor of NEB evolution.
Fructose participated more actively in NEB evolution than glucose when model grape juice was heated for less than 130 min. However, glucose contributed to NEB more than fructose when heating lasted for more than 130 min.
The effect of amino acids in the model on NEB has often been investigated under the conditions of the same food matrix or food models (Hrynets et al., 2019). For example, under the same conditions (including concentration) glutamine (Gln) was more reactive than arginine with fructose as a substrate, while arginine was more reactive than glutamine with glucose as a substrate (Göğüş et al., 1998). According to the reaction rate of NEB, β-alanine and tryptophan have been reported to be the most important amino acids for facilitating NEB due to their ability to enhance browning via the MR (Ajandouz & Puigserver, 1999). In contrast to these findings, we observed that the higher the amino acid content, the higher the degree of NEB, with proline being the most important amino acid contributor to NEB, possibly due to its high content in the models.
The other grape components in grape juice, such as polyphenol, are important for the color of regular wine, mainly due to anthocyanidins and flavonols (Gutiérrez-Escobar et al., 2021). These factors should not be ignored for the NEB of Msalais, more possible with different pathways. According to the local oenologists of Msalais, the color of concentrated grape juice after the addition of grape residue extract is redder than that of the concentrated grape juice alone. This could be explained by the fact that polyphenol substances are extracted from grape skin and seeds.
In the wine process, the substances including polyphenol in skins are often extracted during grape must fermentation, heating grape must before fermentation is sometimes used to increase the extraction and stabilization of wine color (Lisov et al., 2021).
However, for Msalais, we have previously shown that the strong NEB occurred during the concentration of grape juice, although the total polyphenol content increases considerably, from less than 400 mg/L in grape juice to approximately 1300 mg/L in the final grape juice concentrate (Zhang, 2016). The NEB directly results in the deep brown color of Msalais wine (Zhang, 2016), instead of the red color or yellow color such as the red or yellow color in wine, attributed more to the anthocyanidins and flavonols (Gutiérrez-Escobar et al., 2021). The increasingly concentrated polyphenol may be directly incorporated into MR, as nonenzymatic transglycosylation reactions, the main mechanism of phenolics' incorporation into melanoidins (Moreira et al., 2017). Besides, F I G U R E 3 Nonenzymatic browning (NEB) evolution during fermentation of the concentrated 'complete' model. (a) Browning index (BI) and sample absorbance at 420 (A420) evolution; (b) changes in 5-hydroxymethylfurfural (5-HMF) evolution, fermentation after inoculation with vigorously fermenting cultures from a modern production plant (DaolangMslais Limited (Md)) and a traditional craft workshop (Abudu•GayitMsalais (Ma)); and (c) NEB polymerization degree (sample absorbance at 290 (A290)/A420) the more concentrated the juice, the higher the anthocyanin and phenolic compound degradation (Mirzaee et al., 2016), these degraded compounds could also be incorporated into MR. In addition to these prior studies, our enhanced understanding of the evolution of NEB by the MR during the Msalais process provides a clear premise to further investigate the extent to which polyphenols contribute to NEB in Msalais. Additionally, the interaction between polyphenols and sugars during Msalais production, and its effect on color constitute an additional important research question.

| CON CLUS IONS
The MR was the main contributor to NEB during Msalais production.

ACK N OWLED G EM ENTS
We sincerely thank the DaoLangMsalais Limited and Abudu•GayitMsalais traditional craft workshop for the kind provision of Msalais cultures for this study. The authors thank AiMi Academic Services (www.aimie ditor.com) for English language editing and review services.

CO N FLI C T O F I NTE R E S T
The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript or in the decision to publish the results.