Soybeans were purchased from the China Resources Vanguard market (Hangzhou, China) in September 2020. A standard 16 amino acid mixture was purchased from Hitachi Company (Tokyo, Japan). Standards for the quantitative determination of six isoflavones [i.e., daidzin, glycitin, genistein, glycitein, daidzein, and genistin (≥98.0%)] and molecular weight determination (L-reduced glutathione (GSH), cytochrome C, aprotinin, bacitracin, Gly–Gly–Tyr–Arg, and Gly–Gly–Gly) were purchased from Shanghai Yuanye Bio-Chem Technology Co., Ltd. Acetonitrile and DPPH (>97.0%) were purchased from Sigma Company (St. Louis, USA) and were of high-performance liquid chromatography (HPLC) grade. Other chemicals used in this study were of analytical grade. Total Antioxidant Capacity Kit (FRAP) was purchased from Nanjing Jiancheng Biological Engineering Research Institute Co. Ltd. A commercial enzyme preparation containing papain (80%) and amylase (20%) was purchased from Nanning Donghenghuadao Biotechnology Co., Ltd. Normal soybean protein hydrolysates (NSPH) were purchased from Inner Mongolia Keran Biological High-tech Co., Ltd.

Yellow soybean seeds were prepared by soaking in distilled water for 12 h at room temperature. The germination process was performed in an automatized germination chamber (CB-A360B, Foshan, China) in the dark at 30°C with 10 s of irrigation every 1 h (27). The germinated seeds were harvested after 12, 24, 36, 48, 60, and 72 h (Figure 1).

The length of soybean buds was determined by the following rule: 100 soybeans were randomly selected at a certain time, and if more than 80% of the bud length was between 0.0 and 0.5, 0.5–1.0, 1.0–2.0, 2.0–3.0, and 3.0–4.0 cm, the bud length of soybean at this germination time was considered in the corresponding length ranges. Following germination, samples were homogenized and kept frozen at −20 °C until further use. Soybeans germinated at 0 h (i.e., the soybeans were soaked for 12 h) were used as control.

A total of 5.000 g of soybeans were placed in an air-blast drying oven at 105°C for 4 h. After cooling, the weight of soybeans was recorded. Then the soybeans were placed back into the air-blast drying oven and weighed every 30 min till they reached a constant weight. The moisture content was calculated based on the difference in quality before and after drying.

The protein content was determined using the Kjeldahl method (28). The soybeans (0.50 g) were digested in a digestion flask with sulfuric acid (10 ml), followed by the addition of 7.0 g of potassium sulfate and copper sulfate. The soybeans were heated at 420°C using a heating block. Then, the soybean digestive solution was boiled for a further 60 min. Next, the soybean digestive solution was cooled by cautiously adding 90 ml of water. Of note, 20 ml of the digestive solution was taken in Kjeltec Analyzer (FOSS 2300, Beijing, China) for titration. The following formula was used to calculate the protein contents:

where X is the protein content of the sample (%), V₁ is the volume of hydrochloric acid standard solution that expend by sample solution (ml), V₂ is the volume of hydrochloric acid standard solution that expend by reagent blank solution (ml), c is the concentration of the hydrochloric acid standard solution (mol/L), 0.0140 is the weight of nitrogen equivalent to 1.0 ml hydrochloric acid standard solution (1.000 mol/L) (g), m is the weight of the sample (g), V₃ is the volume of the digestive solution that is taken (ml), V₄ is the total volume of the digestive solution (ml), and F is the conversion factor (6.25).

The amino acid content was determined using an amino acid analyzer (Hitachi LA8080, Tokyo, Japan) after hydrochloric acid hydrolysis. The acid hydrolysis was carried out by assessing 0.200 g of soybeans in an ampoule with 6 mol/L hydrochloric acid solution, followed by vacuuming and sealing the ampoule on the alcohol lamp. The soybeans were heated at 110°C using the air-blast drying oven for 22 h. After cooling, it was transferred to a volumetric flask and diluted into deionized water until the volume reached 100 ml. After filtration, 5 ml of the solution was taken to a round bottom flask and evaporated using a rotary evaporator under vacuum at 40°C, followed by the addition of 5 ml of pH 2.2 sodium citrate buffer. Finally, the solution was filtered through a 0.45 μm filtration membrane before being subjected to the amino acid analyzer.

The method described by XU (29) was used with appropriate modifications to detect isoflavones in soybeans. Homogenates of the samples (1.00 g each) were put into conical flasks with 30 ml of 70% methanol. After ultrasonic treatment for 30 min, the samples were shaken for 2 h in the thermostatic oscillator (60°C). When the conical flasks were cooled to room temperature, the volume was adjusted with 70% methanol to 50.0 ml and centrifuged at 2,795 × g for 10 min. The supernatant was filtered through a 0.45-μm filtration membrane before being subjected to HPLC.

Isoflavones were separated using the Capcell PAK MG C18 column (250 × 4.6 mm i.d., 5 μm, Shiseido, Japan) with a mobile phase of 0.1% aqueous phosphoric acid (A) and acetonitrile (B). The gradient elution conditions were as follows: 18% phase B from 0 to 5 min, 18–32% phase B from 5 to 6 min, and then a holding time of 6–13 min; 32%−30% B from 13 to 14 min, and then a holding time of 2 min; 30–80% phase B from 16 to 17 min, and then a holding time of 3 min; 80–18% phase B from 20 to 21 min followed by a holding time of 21–25 min. The flow rate was set at 1.0 ml/min. The injection volume was 10 μl. The temperature of the column was set at 34°C. The UV detection wavelength was 260 nm.

A single-factor test of enzymatic hydrolysis was employed in this study using a commercial enzyme preparation containing papain (80%) and amylase (20%). The effects of three factors, including enzymolysis time (X₁), dosages of proteinase (X₂), and the ratio of material to liquid (X₃), were investigated. The RSM coupled with the Box-Behnken design was utilized to obtain optimal process parameters for enzymatic hydrolysis. With the hydrolysis degree (DH) as the response value, each factor corresponded to lower (−1) and higher (+1) values (Table 1). Three-factor, three-level experiments were performed, and the DH was taken as the response value. The determination of the DH was based on the method described by Lin (30).

The soybeans were germinated for 24 h before they were homogenized. The homogenate was mixed with distilled water according to the material to liquid ratio of 1:10. After mixing well, the mixture (30 g) was put into a conical flask with the compound protease (3% of the dry matter weight) and was stirred for 5 h at 50°C in a magnetic stirrer. Then the conical flask was put into boiling water for 10 min to end the enzymatic hydrolysis, followed by centrifugation at 2,795 × g for 5 min. The supernatant was freeze-dried and then stored at −20°C before use.

The GSPH obtained in 2.4 was redissolved and then filtered using a membrane (Thermo Fisher Scientific, Waltham, USA) with a molecular weight cutoff of 3 kDa. The ultrafiltrate (<3 kDa) was collected, freeze-dried, and then stored at −20°C until further analysis.

The Sephadex G-15 gel was used to separate the hydrolysates collected in the previous step (GSPH after ultrafiltration, UGSPH). A certain amount of Sephadex G-15 gel dry powder was weighed and made fully swollen. Then it was loaded into the column (1.6 × 100 cm). In case of the appearance of bubbles in the gel chromatography column, the gel was reloaded, and the gel was balanced with eluent 3 times before loading samples. The separation conditions were as follows: the loading concentration was 100 mg/ml and the volume of the loading sample was 1 ml. The mobile phase was deionized water, and the flow rate was 0.5 ml/min. The detection wavelength was 280 nm. Each peak in the chromatogram (Figure 2A) was considered one fraction of UGSPH. Every fraction was collected and freeze-dried to determine the molecular weight distribution and In vitro antioxidant activity.

The sample (10 mg) was accurately weighed and diluted into deionized water until the volume reached 10 ml. The molecular weight distribution of protein hydrolysates was carried out using gel filtration chromatography with HPLC. Separation work was performed using a TSK gel G2000SWXL column (7.8 mm I. D. × 30 cm, 5 μm, Japan), with a mobile phase of 0.1% aqueous trifluoroacetic acid (A) and acetonitrile (B). The gradient elution was A: B = 55:45 (v/v). The flow rate was set at 0.5 ml/min. The injection volume was 10 μl. The temperature of the column was set at 30°C. The UV detection wavelength was 220 nm. The calibration curves were drawn using cytochrome C (Mw = 12, 400 Da), aprotinin (Mw = 6, 512 Da), bacitracin (Mw = 1, 423 Da), Gly–Gly–Tyr–Arg (Mw = 451 Da), and Gly–Gly–Gly (Mw = 189 Da). The X-axis of the calibration curve was the retention time, and the Y-axis was the lg value of relative molecular mass (lg Mw) of the above standards.

Each experimental procedure was performed three times, and data were expressed as the mean ± SD. Statistical analyses were conducted by one-way analysis of variance using IBM SPSS 20 (SPSS Inc., Chicago, USA). Duncan's test was employed to analyze the significance for multiple groups. The P-value < 0.05 was considered statistically significant.

Figure 3 shows variations in the nutritional status of soybeans during germination. Soybeans, after 12 h soaking in water, attained favorable internal environmental conditions for later germination. The protein content of non-germinated soybean was 35.43%, which was reduced during germination and early growth of the seedlings as a result of the high activity of proteases (33). The content of soybean protein showed an upward trend during germination; however, the soybean protein content remained statistically stable until 24 h following germination (Figure 3B). At 24 h following germination, there was an increase (6.32%) in soybean protein content compared to 0 h, and then a decrease (2.89%) was observed at 60 h compared to 24 h. It has been reported that soybean germination for 18 h activates endogenous proteases involved in the cleavage of storage soybean proteins and the production of bioactive peptides with antioxidant and anti-inflammatory activities (34, 35). In the beginning of seed germination, endogenous proteases stored in protein may be activated. During the enzymatic hydrolysis, these endogenous proteases can cooperate with exogenous protease to improve the efficiency of enzymatic hydrolysis. As a result, germination had positive enhancement effects on bioactive compounds in beans (36).

Amino acid composition of soybean at different germination states (all in dry matter) and their significant markers are presented in Table 2. Cys and Trp cannot be determined by acid hydrolysis and therefore were not included in the analysis. Asn and Gln were converted to Asp and Glu after the hydrolysis, respectively. Therefore, the Asx and Glx presented in Table 2 include Asp+Asn and Gln+Glu, respectively. The total amino acid content in germinated soybean increased with an increase in germination time. A total of 34 proteins related to amino acid metabolism have been reported to be involved in the process of soybean germination (27). The highest essential amino acids (EAA)/total amino acids (TAA) ratio was 0.36 at 36 h of germination, and EAA/non-essential amino acid (NEAA) ratio was 0.56. The glutamic acid content was highest followed by proline and aspartic acid. Germination in darkness has been reported as a useful method to increase the number of amino acids and thus the nutritional status of soybean products, which is congruent with our findings (37). These results show that soybean germination is associated with an increase in threonine (40.13%), aspartic acid (36.36%), and glutamic acid (33.45%), and a decrease in proline (64.66%) and histidine (18.32%), as well as the least change in the content of methionine, compared to non-germinated soybeans. Mostafa et al. (38) reported an increasing trend in different amino acid components with germination compared with the contents of the same amino acids in dry seeds. A decrease in proline content is in line with the observation of Yang et al. (39).

Isoflavones are a large and very distinctive subgroup of flavonoids that are commonly found in soybeans (40), and they have a significant correlation with the germination of soybeans, which has not been thoroughly studied yet. Vermont et al. observed an increase of 16.2% in soybean isoflavones following 18 h of germination (16). Genistin accounted for the largest proportion (40-60%) of soybean isoflavones during 0–36 h, followed by genistein and daidzein (Figure 3C). During soybean germination, the contents of genistein and daidzein were significantly increased. Sharma (41) reported an increasing trend of isoflavones in chickpea (Cicer arietinum) during germination, which is consistent with our findings. The observed decrease in the content of isoflavones could be due to the conversion of one type of isoflavones into other types of isoflavones. Terrence (42) reported that soybean primary leaf tissues underwent a programmed shift from isoflavonoid to flavonoid metabolism 3 days after germination. After 36 h, genistein became the highest soybean isoflavone component. An increase in total isoflavones in small-seeded soybeans after 7 days of germination under light conditions has been reported (19). Lee et al. (43) found the contents of phenols and flavonoids dramatically increase in black soybeans germinated for 5 days and may contribute to enhancing antioxidant activity. For the elimination of protease inhibitors, 18 h of germination is recommended (16).

In conclusion, germinated beans are better materials to prepare health products than non-germinated beans. In addition, the improvement of bioactive compounds of soybeans depends on the timing of germination. Soybeans were germinated for 24 h, the bud length was about 0.5–1 cm, and the seeds' protein and isoflavone contents were at their peak. At this stage, germinated soybeans are suitable for the hydrolysis of protein.

The effects of three variables, namely, the enzymolysis time (X₁), dosages of proteinase (X₂), and the ratio of material to liquid (X₃), were studied to optimize the hydrolysis conditions using Box-Behnken design and DH as an indicator. The results of the analysis for hydrolysis are presented in Table 1. By comparing the F-value of each factor, it can be concluded that the dosage of proteinase is the most influential factor for DH, followed by the material-liquid ratio and enzymolysis time. A three-dimensional response surface diagram and a contour plot (Figure 4) were plotted based on changes in any two independent variables, which shows the influence of two-factor interaction on the DH of GSPH. As shown in Figure 4, the DH of GSPH gradually increased with an increase in the quantity of proteinase. Figures 4A,E show that the slope of the response surface along the axis of dosages of proteinase was steep, indicating its more influence on the DH of GSPH compared to enzymolysis time and the material-liquid ratio. Figure 4C showed that the slope of the response surface along the axis of the material-liquid ratio was steep, indicating its significant influence on the DH of GSPH.

Based on data analysis using design-export 8.0.6 software, the maximum predictive value of the DH of GSPH was 57.46% when the enzymolysis time was 4.98 h, the dosage of proteinase was 2.95%, and the ratio of material to liquid was 1:11.87. The optimal values of the selected parameters were further subjected to a verification test. The DH of GSPH in this study was (56.49 ± 0.25)%, which was not significantly different from the value of DH 57.46% predicted by the model. There was an increase in the DH following optimization, and the value reached 56.49%. The model established in this study could accurately and reliably predict the DH of GSPH.