The process for preparation of RSS was same as our previous study (13). For spontaneous fermentation (named RSS-SF), firstly, the tomatoes (3.5 kg) and chili peppers (0.5 kg) were washed, mixed and grinded into jam like liquid. Then, ingredients such as salt, wine, ginger, garlic, glucose and lactose were added into the obtained jam like liquid with ratio of 1, 1, 0.75, 0.75, 0.25, and 0.75% (w/v), respectively. Finally, the mixture was incubated at 85°C for 20 min and put into fermentation containers. The process of preparation of RSS for inoculant fermentation (named RSS-IF) was same as RSS-SF except addition with L. rhamnosus. The RSS-SF and RSS-IF both were fermented at 37°C for 28 days. Different samples were randomly withdrawn at 0, 3, 7, 14, 21, and 28 days for subsequent analysis.

The method for determination of browning degree was same as previous study (9, 10). To begin with, 10 g RSS was mixed with 20 mL 95% alcohol and homogenized for 20 min. After that, the mixture was centrifuged at 10,000 × g for 10 min at 4°C. Lastly, the obtained supernatant was monitored at 420 nm using a spectrophotometer (UV-2600, Shimadzu Company, Japan). The value of browning degree was according to equation 1.

where A_420nm is the absorbance at 420 nm.

The activity of PPO (14), peroxidase (POD) (15), phenylalaninammo-nialyase (PAL) (16) and lipoxygenase (LOX) (16) were determined according to the already reported research and the detail processes for determination of those enzymes were provided in Supplementary material.

The method for determination of amino nitrogen, carbonyl value and Vc were in accordance with previous study (10). The detail protocols for those methods were provided in Supplementary material.

The methods for determination of reducing sugar (17) and total polyphenol (18) were based on the previous studies and the detail protocols for those methods were provided in Supplementary material.

All results were presented as average values from three replications with ± standard deviation (SD). Data was analyzed by one way analysis of variance (ANOVA) of the SPSS 25.0 (International Business Machines Corporation, Amunk, New York, United States) and mean separations were performed by Duncan’s multiple-range test. Different letters were used to indicate the significant differences at p ≤ 0.05 of each group.

Path analysis was performed to investigate the main factors related to browning of RSS during fermentation. The content of amino nitrogen (X₁), reducing sugar (X₂), total polyphenol (X₃), Vc (X₄), carbonyl value (X₅) were taken as the independent variables, and browning degree (Y) was taken as the dependent variable. The direct path coefficients (P₁–P₅) and indirect path coefficients (P_ij) between all these independent variables and dependent variable were analyzed by SPSS 25.0 (International Business Machines Corporation., Amunk, New York, United States). The determination coefficients (d_ij) reflected the degree of independent variables to dependent variable and were calculated by correlation coefficients and path coefficients. The determination coefficient indicated the influence degree of the factor on the result (expressed in d), and it can be determined by formula through correlation coefficient and path coefficient:

Annotation:

r_ij: the correlation coefficients between all factors.

P_i: the influence degree of single factor acting independently on Y.

P_j: the influence degree of other factors acting independently on Y (i ≠ j).

P_ij: the influence of one factor X_i on Y through another factor X_j (i ≠ j).

d_i: the influence of single factors on Y.

d_ij: the influence of two factors on Y.

P_e: the coefficients of residual factors that not consider during the process of analysis.

Generally, browning of food was mainly composed of enzymatic and non-enzymatic browning. For enzymatic browning, those enzymes, such as PPO, POD, and PAL, were considered to be the most important factors associated with browning (19). PPO shows either mono-or di-phenol oxidase activity that could catalyze oxidation of phenolic compounds and convert into pigments (19). Besides, it is reported that POD exhibits synergistic action with PPO during the process of browning. Different from PPO and POD, PAL is a committed enzyme in phenyl propanoid metabolism and its activity is related to the concentration of phenolic compounds which are the substrates for PPO and POD (19). Previous studies indicated that the activities of PPO, POD, and PAL are closely related with the browning of different vegetables and fruits (20). In order to investigate the relationship of PPO, POD, and PAL with browning of RSS, the activities of those enzymes were determined during the process of fermentation. As shown in Figure 1A, no enzyme activities of PPO, POD, and PAL were detected during the process of whole fermentation. Research has shown that PPO, POD, and PAL are easily deactivation at high temperature (above 60°C) and acid conditions (below pH4.0) (19, 21). In this study, the final step of preparation of RSS was incubated at 85°C for 20 min that may denature completely of PPO, POD, and PAL. Meanwhile, the values of pH were totally below 4.0 during the process of fermentation (Figure 1B). The activities of PPO, POD, and PAL indicated that browning of group RSS-SF and RSS-IF may not related to enzymatic browning.

The results of changes in total polyphenol during fermentation were depicted in Figure 1C. Interestingly, at the first stage of fermentation (from 0 to 3 days), the contents of total polyphenol from group RSS-SF and RSS-IF were gradually increased, which is similar to previous research (22). As the fermentation time above 3 days, the concentration of total polyphenol was gradually decreased and kept in the ranges from 21.38 to 28.98 mg/mL. It is well known that polyphenol is a critical factor related to browning of different food (20). The polyphenol are easily oxidized and further polymerized to form macromolecular brown compounds, which cause color changes. Generally, the oxidation of polyphenols can be divided into enzymatic and non-enzymatic oxidation of polyphenols (23). In this study, oxidation of polyphenol was mainly caused by non-enzymatic. On the one hand, degradation of polyphenol could react at a slow speed in acidic conditions due to its multiple weakly acidic phenolic hydroxyl groups (24). On the other hand, the multiple phenolic hydroxyl groups had strong reducibility which could cause the polyphenol automatic oxidation and decrease the content (25).

As natural antioxidant, Vc is one of the main nutrition components of different food and also closely related to non-enzymatic browning during the process of food production. According to the previous research, the degradation of Vc could convert to derivative and intermediate to react with other components to lead to browning (26). In summary, there are two pathways of Vc oxidation, which include the aerobic and anaerobic degradation pathways. Therefore, determination of changes in Vc is helpful to elucidate the browning of RSS. As depicted in Figure 1D, the changes of Vc were similar to total polyphenol during the whole process of fermentation. The concentrations of Vc of group RSS-SF and RSS-IF were gradually increased at the initial stage of fermentation (from 0 to 3 days). Then, the concentration of Vc of group RSS-SF and RSS-IF gradually decreased and kept in the ranges from 0.85 to 2.97 mg/100 g when the fermentation time above 3 days.

Previous studies demonstrated that Maillard reaction is closely associated with non-enzymatic browning during the storage of many foods (9, 27). As the most important precursors of Maillard reaction, amino acids and reducing sugars could undergo browning reaction to influence the color of different foods. The changes in in reducing sugars and amino nitrogen during fermentation were illustrated in Figures 1E, F. The content of reducing sugar in both groups were keeping decreasing during whole fermentation time and shown obviously difference at the day of 0 and 14 days (p < 0.05). Additionally, the amino nitrogen content in group RSS-SF increased after 7 days, and then decreased until 21 days, while the group RSS-IF continually decreased until 21 days, and with a slight increase at 28 days (Figure 1F).

Generally, the carbonyl value is related to the oxidation of lipid. Previous study demonstrated that oxidation of lipid is involved with browning of food (11). In summary, the whole networks of lipid oxidation to final colored products include different reactions which are mainly divided into three steps: (i) oxidation of lipid to form lipid hydroperoxides; (ii) rearrangement, fragmentation or degradation of lipid hydroperoxides to produce variety of volatiles and non-volatile monomers; (iii) aldol condensation and/or carbonyl-amine polymerization of products from second step to form colored polymers (11). Briefly, lipids can be oxidized by two main ways which include non-enzymatic oxidation and enzymatic-catalyzed oxidation (11). In this study, the oxidation of lipid is mainly non-enzymatic autoxidation due to no enzyme activities of LOX were detected during the whole fermentation (Figure 1A). As depicted in Figure 1G, the carbonyl value in group RSS-SF increased continually during the whole process of fermentation. Different from group RSS-SF, the carbonyl value of group RSS-IF reached maximum (14.4 meq/kg) at 14 days and then decreased to 9.4 meq/kg as the fermentation time reached 28 days. It is reported that oxidation of lipid is involved with non-enzymatic browning (11). The lipid oxidation occurred during fermentation may contribute to the non-enzymatic browning of RSS-SF and RSS-IF.

The changes in browning degree of different sample were shown in Figure 1H. For group RSS-SF, the value of browning degree increased gradually from 0 to 7 days. The values of browning degree at 0 and 7 days were 0.206 and 0.346, respectively. As the fermentation time above 7 days, the value of browning degree changed little and kept in the range from 0.328 to 0.346 (Figure 1H). Similar to group RSS-SF, the value of browning degree of group RSS-IF also increased gradually from 0 to 14 days. For group RSS-IF, the highest value of browning degree was 0.397 when the fermentation time was 14 days (Figure 1H). However, as the fermentation time above 14 days, the value of browning degree decreased slowly and reached 0.312 at the end of fermentation (Figure 1H).

Based the results of activities of PPO, POD, and PAL (Figure 1A), we conclude that browning of group RSS-SF and RSS-IF during fermentation is non-enzymatic browning. Generally, non-enzymatic browning is composed of oxidation of polyphenol, degradation of ascorbic acid, Maillard reaction and lipid oxidation (10). According to the obtained results, we deduced that the main factors related to non-enzymatic browning of group RSS-SF and RSS-IF maybe oxidation of polyphenol, degradation of ascorbic acid and lipid oxidation. However, it is essential to figure out exactly the main reasons for non-enzymatic browning of group RSS-SF and RSS-IF by a statistical method.

Previous research indicated that path analysis is a suitable method for analysis and isolation of the crucial factors closely associated with browning (10). Usually, the process of path analysis includes analysis and calculation of correlation coefficients, direct coefficients and determination coefficients. The value of determination coefficients exhibits the contribution of different factors related to the final results. In this study, the correlation coefficients of different factors related to browning degree were firstly analyzed and the results shown in Table 1. Based on the results of correlation coefficients (Table 1), the direct path coefficients were secondly calculated. As shown in Table 2, the order of direct path coefficients of group RSS-SF was P₄ > P₅ > P₃ > P₂ > P₁, which exhibited the main direct browning factor was degradation of Vc, followed by oxidation of lipid and polyphenol. For group RSS-IF, the order of direct path coefficients was P₄ > P₃ > P₂ > P₅ > P₁, which indicated that degradation of Vc and oxidation of polyphenol played an important role in browning.

The determination coefficients of different factors related to browning were calculated based on the results of direct path coefficients (Table 2) through the formulas from (2) to (4) and the results were shown in Table 3. The top three of determination coefficients in group RSS-SF were d₃₄ > d₄ > d₄₅, which suggested that the main browning factors were the interaction between oxidation of polyphenol and degradation of Vc, then came after the degradation of Vc. Meanwhile, the top three of determination coefficients in group RSS-IF were d₃₄ > d₄ > d₃, which indicated that the main browning factors also included the interaction between oxidation of polyphenol and degradation of Vc, degradation of Vc, and oxidation of polyphenol. In summary, non-enzymatic browning is complicated and always involved with multiple reactions such as Maillard reaction, degradation of Vc, oxidation of polyphenol and lipids (10, 23). In this study, the main factors related to non-enzymatic browning of group RSS-SF and RSS-IF during fermentation were oxidation of polyphenol and degradation of Vc.