The Escherichia coli strain top 10 was used to propagate plasmids. B. circulans strain R1 was isolated from Okara and conserved in our laboratory. The B. subtilis strains RIK1285, pMD20T, and in-fusion cloning kit were purchased from Takara Biotechnology (Dalian, China). Q5® high-fidelity DNA polymerase and protein marker were purchased from New England Biolabs (NEB, Ipswich, MA, United States). Different inhibitors, such as pepstatin A, iodoacetamide, ethylenedia minetetraacetic acid (EDTA), and phenylmethanesulfonyl fluoride (PMSF), and substrates, such as casein, bovine serum albumin (BSA), keratin, soybean protein isolate, and chicken egg albumin, were purchased from Yuanye Biotechnology (Shanghai, China). Soybean dregs were by-products of manufacturing fermented soybean whey-based tofu, dried by microwave drying, ground, and sieved with 60 mesh. The soybean dregs comprised 22.03% protein and 10.23% water.

LBK (LB with 30 μg/ml kanamycin) was used to screen E. coli top 10 and B. subtilis RIK1285 contained plasmids with kanamycin-resistant gene (Kan^R). Fermentation medium 1 was used for the production of AprBcp in 24-well plates and shake flask, which was composed of 1.5% peptone, 2.5% yeast extract, 1% citrate sodium, 0.3% K₂HPO₄, and 4% maltose. Fermentation medium 2 was used for large-scale production of AprBcp in 7 and 50 l bioreactor, which contained 7% maltose, 2.5% soybean meal, 2% corn steep liquor, 0.3% citrate sodium, 1% K₂HPO₄, 2% wheat bran, and 0.3% CaCl₂.

The 24-well plate cultivation was used to screen the recombinant strains with different promoters and signal peptides. The shake flask cultivation was used to isolate the recombinant strains with higher alkaline protease activity. Large-scale production of AprBcp was carried out in 7- and 50-L bioreactor. The detailed protocol is provided in Supplementary Methods.

Based on the sequence of alkaline protease gene from B. circulans strain NCTC2610 (GenBank: UAPZ01000012.1: 120935-122077), two primers, bcp-fw and bcp-rev, were designed for PCR amplification. The gene encoding AprBcp was cloned by PCR amplification using genomic DNA of B. circulans R1 as template. The PCR product was ligated to pMD20T and sequenced. The primers used in this study are shown in Supplementary Table 1.

Sequence alignment was performed by blastn or blastx provided by the National Center for Biotechnology Information (NCBI). Amino acid sequences and signal peptide prediction of AprBcp were analyzed by Expasy.¹ The sequences were analyzed for identity using ClustalW² and ENDscript/ESPript.³ Homologous models were generated on the Swiss-model online server. PyMOL was used to analyze the obtained model. The alkaline serine protease from B. lentus was chosen as a template (Protein Data Bank code: 1st3.1.A).

The vector pHY was constructed as template to construct different plasmids (Figure 1). The vector pHY was composed of temperature-sensitive replicon PE194 and a kanamycin-resistant gene (Kan^R) which functions in B. subtilis and E. coli, as well as the pUC origin of replication from vector pBluescript IISK(+). Additionally, it included B. subtilis-derived promoter P₄₃, which located upstream from the multi-cloning site (MCS), the 5′-upstream repair, and 3′-downstream repair template of α-amylase from B. subtilis 168.

The gene bcp was ligated to pHY to form expression vector pHY-bcp. For promoter optimization, nine highly active promoters reported by previous studies were selected and synthesized for heterologous expression of AprBcp. The fragment pHY without promoter P₄₃ was obtained from pHY-bcp digested by Eco52I and MluI. And then, each different promoter flanked by Eco52I-MluI was separately ligated to fragment pHY to form nine expression plasmids. For dual promoter plasmids, the second promoter was inserted into the upstream of the single promoter expression plasmid (Figure 1). The sequences of nine highly active promoters are shown in Supplementary Table 1 and the primers used in this part are shown in Supplementary Table 1.

For signal peptide optimization, the expression vector with the highly active dual promoters was used to screen the efficient signal peptide. Nine signal peptides were selected and synthesized by over-lapping PCR, and ligated to dual promoter expression plasmid by in-fusion cloning. The process for construction of plasmids with signal peptides is shown in Figure 1. Plasmids with different promoters and signal peptides used in this study are shown in Table 1. The sequences of nine signal peptides are shown in Supplementary Table 3 and the primers used in this part are also shown in Supplementary Table 1.

The bcp expression cassette, with highly active dual promoter, efficient signal peptide and bcp, was integrated to α-amylase locus of B. subtilis 1285. The method for chromosomal integration of bcp expression cassette was similar to previous studies and the detailed process is provided in Supplementary Methods and Figure 2 (Zakataeva et al., 2010; Zhang et al., 2016a; Zhou et al., 2021). The recombinant strains with bcp expression cassette integrated to α-amylase locus were cultivated in a 500 ml flask for 48 h. The recombinant strain with the highest activity at 500 ml shake flask was further cultivated in 7 and 50 l bioreactor, respectively.

The fermentation supernatant after centrifugation was filtered through with a 0.22 μm membrane, then concentrated and buffer-exchanged to Tris–HCl buffer (20 mM, pH 7.5) by ultrafiltration with a membrane of 10 kDa cutoff. The AprBcp was purified by HiTrap SP Fast Flow chromatography. The purified AprBcp was collected and analyzed by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

The protease activity was detected according to the Folin’s phenol method (Wang et al., 2019). A quantity of 1 g casein was dissolved in 100 ml boric acid buffer (pH 10.0, 50 mM) and used as substrate. Firstly, 200 μl of diluted enzyme was added to 200 μl of 1% (w/v) casein solution and incubated at 60°C for 10 min. Secondly, 400 μl of 400 mM trichloroacetic acid (TCA) was added to end the reaction, and then centrifuged at 11,000 × g for 5 min. Finally, 200 μl of the supernatant was added into 1 ml of 400 mM Na₂CO₃ and 200 μl of Folin’s phenol reagent and incubated at 40°C for 20 min. The reactions with enzymes added after TCA were used as controls. One unit of alkaline protease activity was defined as the amount of enzyme that releases 1 μg of tyrosine from casein at 60°C and pH 10.0.

The optimal pH of AprBcp was assayed in different 50 mM buffer from pH 6.0 to 11.0 (Na₂HPO₄–NaH₂PO₄ for pH 6.0–7.0, Tris–HCl for pH 8.0–9.0, and boric acid-NaOH for pH 10.0–11.0). The relative activity at different pH was calculated by setting pH 10.0 as 100%. For the pH stability, AprBcp was incubated at 30°C in 50 mM buffer with different pH from 6.0 to 11.0 for 6 h, and then the residual activity was determined. The highest enzyme activity was considered as 100%, and other values were recorded as a percentage of the highest value.

The optimal temperature of AprBcp was measured at different temperatures from 40°C to 75°C. The relative activity at different temperatures was calculated by setting 60°C as 100%. The thermal stability of AprBcp was investigated by incubating at temperatures from 40°C to 70°C for 10 min, and the residual activity was determined at pH 10.0 and 60°C. Furthermore, thermal stabilities of AprBcp at 60°C in presence of 2, 5, and 10 mM Ca²⁺ were determined by incubation at different time intervals. The residual activity was calculated by taking the non-heated AprBcp activity as 100%.

Several inhibitors, such as pepstatin A, PMSF, EDTA, and iodoacetamide, were used to investigate the inhibition effects on the activity of AprBcp. The residual activity of AprBcp was detected in presence of these inhibitors and calculated by taking the non-treated AprBcp activity as 100%.

The effect of different metal ions on the activity of AprBcp was analyzed by incubating enzyme samples for 6 h at room temperature in 50 mM Tris–HCl buffer (pH 8.0), containing 1 and 5 mM of Ca²⁺, Mg²⁺, Na⁺, K⁺, Li⁺, Zn²⁺, Mn²⁺, Co²⁺, Hg²⁺, Ag⁺, and Cu²⁺. The residual activity was determined as described previously.

For substrate specificity, casein, BSA, keratin, soy protein isolate, and chicken egg albumin were used as substrate, respectively. The relative activity on each substrate was expressed as the percentage of that on casein.

The soybean dregs powders were mixed with deionized water at a ratio of 3% (w/w), and homogenized using a homogenizer at a speed of 11,000 × g for 3 min before it was adjusted pH to 9.0 with 1 M NaOH. The mixtures were gradually raised to 50°C prior to the addition of proteinase AprBcp, and then continuously shook at that temperature for 3 h. The resulting suspensions were incubated in boiling water for 10 min to inactivate the proteinase, and subsequently centrifuged at 11,000 × g for 30 min to remove the insoluble residue. The collected supernatant was acidified with 3 M HCl to pH 7.0.

Different ratios of enzyme to substrate (500, 1,000, 1,500, 2,000, and 2,500 U/g), ratio of substrate to water (1, 2, 3, 4, and 5%, w/w), hydrolysis time (2, 3, 4, 5, and 6 h), hydrolysis temperature (40, 45, 50, 55, and 60°C), and pH (7, 8, 9, 10, and 11) were selected successively for optimization.

The protein recovery (RP, %) was used as a measurement of the soybean dregs’ protein extraction efficiency and defined as the ratio of the protein mass in the supernatant to that in the soybean dregs. The protein content was determined using the GB 5009.5-2016 of the Chinese standard.

All measurements were done in triplicate. The data were analyzed by the Statistical Package for the Social Sciences (SPSS) software (version 18.0, SPSS Inc., United States) and were expressed as mean ± standard deviation (SD).

The results of NCBI-blastn and open reading frame finder exhibited that the open reading frame of bcp was 1,143 bp, which encoded a polypeptide with 381 amino acid residues. The sequence of bcp was deposited in the GenBank of NCBI (accession no. ON733249). Based on the results of NCBI-blastp, we deduced that AprBcp is an alkaline serine protease, which belonged to S8 family peptidase. In addition, AprBcp displayed 92.3, 89.4, and 78.7% identity against alkaline serine protease from B. circulans NCTC2610 (accession no. SPU21234.1), Alkalihalobacillus clausii (accession no. WP_094423791), and Alkalihalobacillus patagoniensis (accession no. WP_09419032 9.1), respectively. Multiple alignment analyses of different alkaline serine protease with AprBcp displayed that alkaline proteases from bacillus are composed of three different domains, which include signal peptide, pro-peptide, and mature peptide domain (Figure 2A). For AprBcp, the first 27 amino acid residues, the sequence from Ala28 to Met111, and the sequence from Ala112 to Arg380 were signal peptide, pro-peptide, and mature peptide domain, respectively.

The tertiary structure of AprBcp with mature peptide domain was obtained by homology modeling using the alkaline proteases from Bacillus lentus as a template (PDB deposition: 1st3.1.A). As shown in Figure 2B, the catalytic triad consisted of D32, H62, and S215. In addition, two surface loops, located from position G98 to S104 and F122 to S128, were responsible for substrate binding. Similar to already reported alkaline proteases (Zhang et al., 2019), AprBcp also has two calcium-binding sites. The first calcium-binding site with high affinity is involved of a loop from position A72 to L80 and side chains of Gln2 and Asp40. The second calcium-binding site, a low-affinity binding site, was formed by two bound water and residues Tyr165, Ala168 Asp191, and Arg241.

In this study, 10 promoters were chosen to drive the expression of AprBcp in B. subtilis 1285 (Table 1). The process for construction of expression vectors with different promoters was depicted in Figure 1. The obtained 10 expression vectors were transformed to B. subtilis 1285 and the strengths of those 10 promoters were evaluated by extracellular activity of AprBcp. As exhibited in Supplementary Figure 2 and Figure 3A, the recombinant strain Bs7 containing plasmid pHY7 with promoter P_cry3A exhibited the highest activity at both 24-well plate and shake flask cultivation, followed by Bs8 (P_Bsapr), Bs3 (P_Bsamy), Bs4 (P_Baamy), and Bs6 (P_HpaII). For cultivated 500 ml shake flask, the highest activity of Bs7 was 5,580 U/ml, which is 1.13- to 2.73-fold higher than that of another recombinant strain containing different promoters (Figure 3A). Until now, there are few reports showed that promoter P_cry3A from Bacillus thuringiensis could drive high-level expression of alkaline protease in B. subtilis (Widner et al., 2000). The results of this study demonstrated that promoter P_cry3A was more suitable for heterologous expression of recombinant AprBcp in B. subtilis 1285 than another nine promoters. Our results were similar to previous studies, which demonstrated that optimization of promoter could effectively improve the production of recombinant protein in B. subtilis (Song et al., 2015; Song et al., 2017; Zhang et al., 2017; Liu et al., 2019; Kang et al., 2020).

As a key regulatory element, promoter plays an important role on the expression level of heterologous gene in B. subtilis. Many studies demonstrated that the strength of dual promoter was normally higher than the single in B. subtilis (Guan et al., 2016; Zhang et al., 2017; Liu et al., 2019; Kang et al., 2020). For further improvement of the production of AprBcp, expression vectors with dual promoter were constructed (Table 1; Figure 1). According to previous work, the transcriptional strength of a dual promoter depends on the promoter adjacent to the gene, so promoter P_cry3A was located adjacent to bcp. Promoter P_Bsapr, P_Bsamy, P_Baamy, and P_HpaII were inserted to pHY7 upstream P_cry3A, generating plasmids with dual promoter such as P_Bsapr-cry3A, P_Baamy-cry3A, P_Bsamy-cry3A, and P_HpaII-cry3A, which were in accordance with plasmids pHY11, pHY12, pHY13, and pHY14, respectively. As exhibited in Supplementary Figure2 and Figure 3A, the extracellular activities of recombinant strains with dual promoter were higher than Bs7 (P_cry3A), which demonstrated that dual promoter could further improve the production of recombinant AprBcp. Among those recombinant strains, Bs11 (P_Bsapr-cry3A) with highest extracellular activity (7,068 U/ml), which was 1.26-fold higher than that of Bs7 (P_cry3A). According to the previous literatures, two tandem promoters can improve the production of recombinant protein in B. subtilis. As reported by Kang et al. (2020), the production of recombinant amidase drove by dual promoter P_amyE-cdd was 1.37-fold higher than that of the control. Similarly, the activity of extracellular BcaPRO from recombinant strain containing dual promoter P_Bsamy-Baamy was 1.5-fold of the activity of strain containing single promoter P_Baamy (Liu et al., 2019). In recent years, different dual tandem promoters have been reported to drive expression of recombinant protein in B. subtilis such as P_Bsamy-Baamy, P_amyE-cdd, P_{spoVG-spoVG142}, P_HpaII-amyQ, and P_gsiB-HpaII (Guan et al., 2016; Zhang et al., 2017; Liu et al., 2019; Kang et al., 2020). To our best knowledge, there are no reports about dual promoter P_Bsapr-cry3A. The results of this study revealed that dual tandem promoter P_Bsapr-cry3A with great potential for heterologous expression of recombinant protein in B. subtilis.

As a key regulatory element, signal peptide plays an important role on efficient extracellular expression of target protein in B. subtilis (Song et al., 2017; Fu et al., 2018). To further improve the production of AprBcp in B. subtilis, eight efficient sec-type signal peptides from B. subtilis were selected to replace the native signal peptide of AprBcp (Table 1; Figure 1). As shown in Supplementary Figure 3 and Figure 3B, the recombinant strain BS17 (containing signal peptide SP_Bschi) exhibited the highest activity among those eight recombinant strains containing different signal peptides. The maximum extracellular activity of recombinant strain BS17 was 8,253 U/ml, which was almost about 1.09- to 1.33-fold higher than that of the other strain with different signal peptides. Until now, there are few studies reported the signal peptide SP_Bschi of chitinase from B. subtilis. Generally, signal peptides of amylase and protease from B. subtilis, B. licheniformis, or Bacillus amyloliquefaciens are usually used to the secretion of recombinant protein in B. subtilis. The result of this study revealed that signal peptide SP_Bschi of chitinase is more efficient than that from α-amylase, alkaline, and neutral protease.

Generally, the signal peptide is composed of positively charged N-region, hydrophobic H-region, and C-region. The positively charged N-region is responsible for the interaction with the translocation machinery. Additionally, the hydrophobic H-region, following the charged N-region, functions on the process of insertion of signal peptide into the membrane. Previous studies indicated that the efficiency of signal peptide is related to its positively charged N-region and hydrophobic H-region (Guo et al., 2021). In this study, the N-domain charge and the ratio of hydrophobic amino acids of eight efficient Sec-type signal peptides used in this study were from 2 to 4, and 54.5 to 72.7%, respectively (Supplementary Table 4). The most efficient signal peptide SP_Bschi with the highest N-domain charge, but lowest ratio of hydrophobic amino acids, which indicated that the efficiency of secretion may not direct relative to the properties described above. Based on the results of previous (Liu et al., 2019; Kang et al., 2020) and this study, we concluded that each target protein has its preference for signal peptide and it is necessary to screen a suitable signal peptide for efficient extracellular production of target protein.

During the processes of promoter and signal peptide optimization, the expression of AprBcp was mediated by free plasmids, and the kanamycin was required to maintain expression vectors. However, cultivation added with kanamycin and the plasmid with kanamycin resistance gene marker are both prohibited in many application industries (Yang et al., 2016). Therefore, it is necessary to develop a mark-free expression way to produce AprBcp. The expression vector pHY17 contains temperature-sensitive replicon PE194, which will be inactivated when cultivation temperature is up to 45°C. Then, the fragment with dual promoter P_Bsapr-cry3A, signal peptide SP_Bschi, bcp, 5′-upstream and 3′-downstream homologous template of α-amylase will be integrated to α-amylase locus of B. subtilis1285. A total of 10 mark-free strains named Bs23 to Bs32 were selected and cultivated in shake flask. The extracellular activities of these 10 mark-free strains were from 8,123 to 8,365 U/ml and the strain Bs27 with highest activity (8,365 U/ml) was chosen for large-scale fermentation.

The large-scale fermentation of Bs27 was carried in 7- and 50-L bioreactor. The maximum activity produced by recombinant strain Bs27 in 7- and 50-L bioreactor were 161,661 and 165,870 U/ml, respectively (Figure 3C). In recent years, several engineered Bacillus strains, such as B. subtilis, B. licheniformis, and B. amyloliquefaciens, were cultivated in a bioreactor to prepare alkaline protease (Liu et al., 2019; Wang et al., 2019; Zhou et al., 2020, 2021). Previous studies reported that the extracellular alkaline protease activity of engineered B. licheniformis 2709, B. amyloliquefaciens K11, and B. subtilis WB600 reached 57,763, 30,200, and 27,860 U/ml, respectively, in 5 or 15 l fermenter (Wang et al., 2016, 2019; Liu et al., 2019; Zhou et al., 2021). However, for most of the studies, the method for alkaline protease activity assay was based on the method of the People’s Republic of China GB/T 23527-2009, in which one unit of alkaline protease activity was defined as the amount of enzyme that releases 1 μg of tyrosine from casein at 40°C and pH 10.5. Thus, our results were not fully comparable with these values since activity assays with different conditions have been used. SDS-PAGE analysis of fermented supernatant from different cultivation times in 50-L bioreactor revealed that AprBcp was approximately free from contaminating proteins, which facilitates downstream processing (Figure 3D).

The process for purification of AprBcp is shown in Supplementary Table 5. After ultrafiltration and affinity chromatography, AprBcp was purified 2.09-fold to homogeneity and specific activity reached 58,621 U/mg, which was higher than many previous reported alkaline proteases but lower than those from Idiomarina sp. C9-1 (99,511.9 U/mg), Termitomyces albuminosus (180,000 U/mg), Trametes cingulata CTM10101 (94,000 U/mg), and Vibrio metschnikovii J1 (120,000 U/mg; Jellouli et al., 2009; Zheng et al., 2011; Omrane Benmrad et al., 2016; Zhou et al., 2018). SDS-PAGE analysis showed that the purified AprBcp was about 30 kDa corresponding to molecular mass of mature peptide of AprBcp, which is similar to other alkaline proteases from Bacillus (Contesini et al., 2018). It has been reported that alkaline proteases from Bacillus species almost consist of a signal peptide, pro-peptide, and subtilisin domain. Theoretically, alkaline proteases from Bacillus are activated by three steps: (1) folding of alkaline protease containing different domains, (2) auto-processing between the pro-peptide and subtilisin domain, and (3) degradation of the pro-peptide by activated subtilisin (Uehara et al., 2013).

The values of K_m and V_max of purified AprBcp were 2.35 mg/ml and 63,038 μM min⁻¹ ml⁻¹, respectively. Kinetic parameters of purified AprBcp indicated that it has high substrate affinity and catalytic efficiency toward casein. Based on previous studies, we found that K_m values of Bacillus alkaline proteases vary differently. The K_m values of alkaline proteases from B. cereus-S6-3, Bacillus stearothermophilus, Bacillus safensis RH12, and B. circulans M34 are 3.07, 0.62, 2.2, and 0.96 mg casein/mL, respectively (Sari et al., 2015; Abdel-Naby et al., 2017, 2020; Rekik et al., 2019).

The optimum pH for maximum activity of AprBcp was 10.0, and the relative activities were above 65% in the range from pH 7.0 to 11.0 (Figure 4A). Generally, the most reported alkaline proteases exhibit higher activity in the range of 8.0–11.0. The maximum activities of alkaline proteases from B. pumilus AR57 and Bacillus megaterium RRM2 are detected at 9.0 and 10.0, respectively (Rajkumar et al., 2011; Jagadeesan et al., 2020). While, the optimal pH of an alkaline protease from Bacillus halodurans C-125 is 12.0 (Tekin et al., 2021). For pH stability, AprBcp exhibited high residual activity (above 80%) in the range from pH 8.0 to 11.0 after 5 h of incubating at 25°C (Figure 4B), which is similar to previous reported alkaline protease (Rajkumar et al., 2011; Jagadeesan et al., 2020; Tekin et al., 2021).

For optimum temperature, AprBcp exhibited maximum relative activity at 60°C and remained highly active from 50 to 65°C (Figure 4C). The optimum temperature of already studied alkaline proteases verifies differently according to their origin. Normally, alkaline proteases from Bacillus showed high activity in the range from 50 to 80°C (Contesini et al., 2018). As far as thermal stability analysis is concerned, the purified AprBcp was incubated in buffers with and without Ca²⁺. The residual activities of AprBcp in the absence of Ca²⁺ at 50, 55, 60, and 65°C were 98.2, 75.3, 5.2, and 0.5%, respectively, after heat treatment for 10 min (Figure 4C). As shown in Figure 4D, it was obvious that addition with Ca²⁺ could effectively improve the thermal stability of AprBcp. The residual activities of AprBcp in the presence of 2, 5, and 10 mM Ca²⁺ were 45.8, 82.3, and 100.3%, respectively, after heat treatment at 60°C for 40 min. It has been reported that Ca²⁺ plays an important role on the stability of alkaline protease. For instance, the stability of alkaline protease from Trametes cingulata CTM10101, Nocardiopsis alba OM-5, and B. safensis S406 was enhanced by 1.24-to 4.52-fold, respectively (Omrane Benmrad et al., 2016; Mhamdi et al., 2017; Chauhan et al., 2021). According to the results from Zhang et al. (2019), Ca²⁺ could as an activator to improve the stability and activity of alkaline protease from Bacillus species is attributed to its electrostatic interaction with alkaline protease. In this study, there were two calcium-binding sites in AprBcp to interact with Ca²⁺, which improved the stability of AprBcp in extreme conditions.

It has been reported that proteases from different sources are sensitive to various inhibitors (Barzkar, 2020). Usually, serine proteases are inhibited by PMSF, diisopropyl fluorophosphate, and 3,4-dichloroisocoumarin. The relative activities of AprBcp treated with different inhibitors are shown in Table 2. AprBcp was strongly inhibited by PMSF and the relative activity was only 27.2 and 0.5%, respectively, in the presence of 1 and 5 mM PMSF, suggesting that serine is relative to the catalytic activity. Based on previous studies, we found that the activities of most serine proteases were completely inhibited by PMSF at a concentration of 5 mM (Barzkar, 2020; Jagadeesan et al., 2020; Chang et al., 2021). However, for an alkaline serine protease from Pseudoalteromonas sp. 129–1, the maximum activity inhibition is detected in the presence of 15 mM PMSF (Wu et al., 2015). In addition, the relative activities of AprBcp were 86.3 and 73.2%, respectively, in presence of 1 and 5 mM EDTA, revealing that some metallic ions are important for catalytic activity of AprBcp. The inhibitory effect of EDTA on AprBcp is similar to previous studies (Omrane Benmrad et al., 2018; Sharma et al., 2020; Chang et al., 2021). The alkaline proteases from Nocardiopsis dassonvillei OK-18, B. stearothermophilus CAU209, and Penicillium chrysogenum X5 are inhibited by 30.8, 40.5, and 22.2%, respectively, in presence of 1 to 10 mM EDTA (Omrane Benmrad et al., 2018; Sharma et al., 2020; Chang et al., 2021). Compared with PMSF and EDTA, pepstatin A and iodoacetamide exhibited little inhibitory effects on the activity of AprBcp, which indicated that-SH group and aspartate were not involved in the catalytic activity of AprBcp.

Several metal ions were assessed for their effects on the stability of AprBcp (Table 3). As activator, metal ions Ca²⁺, Mn²⁺, and Mg²⁺ could improve the residual activity of AprBcp by 115.2, 112.3, and 108.5%, respectively, at a concentration of 5 mM. Besides, AprBcp was inhibited by Cu²⁺ and the residual activities were 40.7 and 25.9%, respectively, in the presence of 1 and 5 mM Cu²⁺. Other metal ions, such as K⁺, Na⁺, Co²⁺, Zn²⁺, and Al³⁺ exhibited little effects on the stability of AprBcp. According to the previous studies, metal ions act as activator or inhibitor mainly depend on the specific target alkaline proteases (Contesini et al., 2018; Barzkar, 2020). The result of this study and some previous research found that Cu²⁺ is an inhibitor. Otherwise, the alkaline proteases from B. safensis RH12 and Penicillium chrysogenum X5 are activated by Cu²⁺ and the residual activities are upregulated by 34.2 and 5.3%, respectively, when treated with 1 or 5 mM Cu²⁺ (Omrane Benmrad et al., 2018; Rekik et al., 2019).

The substrate preferences of AprBcp were characterized with various natural proteinaceous substrates (Table 4). AprBcp exhibited the highest activity toward casein, followed by soybean protein isolate (61.2%) and bovine serum albumin (46.2%). Furthermore, AprBcp also displayed little activity against keratin and chicken egg albumin.

To search for the effect of the proteinase AprBcp enzymatic hydrolysis of soybean dregs on the protein recovery, single-factor experiments were carried out (Figure 5). The protein recovery (RP) gradually increased with the increase in the ratio of enzyme to substrate and leveled off after 2,000 U/g (Figure 5A). Therefore, the selected ratio of enzyme to substrate was 2,000 U/g. In addition, as ratio of substrate to water increased, the RP first increased and then decreased, reaching the maximum value at 2%, however, with no significant difference between 2 and 3% (Figure 5B). Considering the improvement of industrial production efficiency, ratio of substrate to water = 3% was selected. As shown in Figure 5C, when the hydrolysis time reached 4 h, the RP was the highest (69.64%). But, in industrial production, shorter enzymatic digestion time can prevent the growth of stray bacteria and reduce reaction costs. Thus, 3 h was selected as the hydrolysis time. Temperature is an important index affecting protease activity. The optimal temperature for enzymolysis of soybean dregs by proteinase AprBcp was 45°C, which was close to our above study (Figure 5D). When pH was 9.0, proteinase AprBcp exhibited the best hydrolysis efficiency in soybean dregs with a protein recovery of 71.81% (Figure 5E), which was higher than the results reported by Figueiredo et al. (2018).