Enhanced catalytic activity of a novel trypsin by semi-rational design with mechanistic insights from molecular simulations

Abstract

Trypsin is widely used in the food industry for meat processing, dairy production and seafood treatment. However, the industrial application of trypsin is constrained by the pathogenic risks associated with animal-derived trypsin and the low enzymatic activity of microbial-derived trypsin. This study aimed to enhance the catalytic activity of a novel trypsin heterologously expressed in Bacillus subtilis SCK6. Given the catalytic specificity of trypsin, numerous lysine and arginine residues within the trypsin are susceptible to autolytic cleavage, which may compromise the integrity and stability of its tertiary structure, thereby affecting its catalytic efficiency. To address this, a semi-rational design strategy was employed to introduce mutations at lysine and arginine residues. As a result, a trypsin variant with a 2.2-fold increase in enzymatic activity was obtained, reaching 93.9 U/ml. Further optimization of the fermentation process elevated the enzymatic activity to 132.8 U/ml. Additionally, this study pioneered molecular docking and molecular dynamics simulations in trypsin engineering, revealing that the introduced mutations stabilize the catalytic pocket and enhance enzyme activity. These findings demonstrate that structure-guided mutagenesis of autolysis-prone lysine and arginine residues can significantly improve the catalytic performance of microbial trypsin. This strategy provides a rational framework for the targeted engineering of trypsin variants and offers a practical approach for developing safer, high-activity preparations suitable for industrial food processing applications.

Highlights

• •A novel trypsin was expressed in Bacillus subtilis with high secretion efficiency.
• •Mutant trypsin activity rose 3.5-fold to 132.8 U/ml after optimization.
• •Semi-rational design yielded the highest trypsin activity in shake flasks.
• •Elucidated structural mechanisms of enhanced enzyme activity.

1. Introduction

Trypsin (EC 3.4.21.4), a serine protease, specifically hydrolyzes peptide bonds at the carboxyl termini of arginine and lysine residues. Due to its proteolytic activity, trypsin is widely utilized in the food industry for meat processing (Ming, Zeng & Li, 2009) and dairy processing (Ren, Wang, Guo & Huo, 2009). While traditionally extracted from animal tissues, microbial-derived trypsin offers significant advantages in terms of biosafety (He et al., 2024), large-scale production (Szkopek et al., 2024) and cost efficiency (Szkopek et al., 2024). However, the relatively low enzymatic activity and high production costs associated with microbial trypsin have limited its broader industrial adoption.

To address these challenges, protein engineering strategies, especially rational and semi-rational mutagenesis, have been extensively employed to improve enzymatic performance. Semi-rational design combines structural or sequence information with targeted mutagenesis to systematically and efficiently identify beneficial variants while minimizing screening efforts. This method proves especially effective when structural characteristics or conserved motifs inform the selection of mutation sites, thereby establishing itself as a robust strategy for improving enzymatic properties.

Due to trypsin's catalytic specificity, internal arginine and lysine residues are highly susceptible to autolytic cleavage, which disrupts the tertiary structure and compromises catalytic efficiency. Targeted modification of these residues has been shown to improve structural stability and better preserve enzymatic activity. For instance, Li et al. substituted Arg117 in rat trypsin with leucine (R117L) via site-directed mutagenesis, resulting in significantly enhanced enzymatic activity (Li, Nie & Tang, 1998). Zhang et al. performed site-directed mutagenesis at the R145 position of Streptomyces griseus trypsin (SGT) and combined it with propeptide engineering, achieving a 3.75-fold increase in activity (Zhang, Huang, Yao, Du & Kang, 2017). In a follow-up study, the same group used MALDI-TOF-MS to identify potential autolysis sites and substituted 15 additional lysine and arginine residues with alanine, ultimately obtaining a K101A mutant with substantially improved catalytic performance. Likewise, Niu et al. introduced R295A and R315A mutations and optimized the propeptide sequence in Pichia pastoris, significantly enhancing the catalytic activity of Streptomyces fradiae trypsin (Niu et al., 2024). These studies demonstrate that modification of arginine and lysine residues-frequently located at autolysis-prone sites-is an effective strategy for improving trypsin activity, highlighting the potential of mutation-guided enzyme engineering.

However, despite the promising results, previous research has primarily focused on alanine substitution at individual cleavage-prone sites or limited combinatorial mutagenesis strategies. The full potential of saturation mutagenesis at key residues has yet to be systematically explored. Furthermore, while improved activity has been reported, most studies lack in-depth structural analyses to elucidate the mechanistic basis underlying these enhancements. Specifically, the conformational alterations and changes in substrate-binding dynamics induced by mutations remain largely uncharacterized. This gap in knowledge hinders a comprehensive understanding of the structure–function relationships driving catalytic improvement and limits the effective application of semi-rational design strategies in trypsin engineering.

In this study, we explore the heterologous expression in Bacillus subtilis SCK6 of a novel trypsin homolog from Streptomyces sp. with the aim of enhancing catalytic performance. To achieve this, we implemented a semi-rational protein engineering strategy targeting lysine and arginine residues. This approach seeks to addresses limitations in prior studies by integrating systematic mutational analysis with computational structural insights to guide the engineering.

2. Materials and methods

2.1. Strains, plasmids and reagents

The trypsin amino acid sequence, along with the expression vector pUC980–2 (Fig. S1) and the host strain Bacillus subtilis SCK6 were obtained from our laboratory. High-fidelity DNA polymerase was purchased from Vazyme Biotech Co., Ltd. (Nanjing, China). The target gene and primer synthesis were performed by Sangon Biotech (Shanghai, China). N-α-Benzoyl-DL-arginine-p-nitroanilide (BAPNA) was purchased from Yuanye Biotech Co., Ltd. (Shanghai, China), and the benzamidine column (Benzamidine Agarose Beads 4FF) for trypsin purification were obtained from Qianzhusong Biotech Biotech Co., Ltd. (Jiangsu, China).

2.2. Media

Luria-Bertani (LB) medium, composed of 1 % (w/v) NaCl, 0.5 % (w/v) yeast extract, and 1 % (w/v) tryptone, was used for the routine cultivation of Bacillus subtilis SCK6. Competent cells were prepared using a modified LB medium supplemented with 1.5 % (w/v) xylose. The fermentation medium for recombinant trypsin expression consisted of 1 % (w/v) NaCl, 1 % (w/v) tryptone, 1.5 % (w/v) soluble starch, 0.502 % (w/v) KH₂PO₄, and 1.424 % (w/v) K₂HPO₄·3H₂O.

2.3. Plasmid construction and transformation

The target gene and vector backbone were individually amplified by PCR using primers containing homologous sequences, following the method (You, Zhang & Zhang, 2012). The resulting fragments were mixed at an equimolar ratio and subjected to in vitro homologous recombination to assemble plasmid multimers.

For the preparation of competent Bacillus subtilis SCK6 cells, an overnight culture was inoculated into fresh LB medium and grown to mid-logarithmic phase. The cells were then harvested and resuspended in modified LB medium (1 % NaCl, 0.5 % yeast extract, 1 % tryptone, and 1.5 % xylose), with the initial OD₆₀₀ adjusted to 1.0. After 1.5 h of incubation at 37 °C with shaking, the culture was used directly as competent cells.

Transformation was performed by mixing 1.5 μL of the assembled plasmid multimers with 100 μL of competent cells, followed by incubation at 37 °C with shaking for 1.5 h. The transformation mixture was plated on LB agar supplemented with kanamycin. After overnight incubation at 37 °C, transformants were selected and further verified by colony PCR and DNA sequencing.

2.4. Ala scanning and saturation mutation

Given the hydrolytic nature of trypsin, 18 arginine and lysine residues were individually mutated to alanine to evaluate their effects on enzymatic activity. Site-directed alanine mutagenesis was performed using overlap extension PCR, and the corresponding primers are listed in Table S1. After plasmid construction and transformation into Bacillus subtilis SCK6, the enzyme activity of each mutant was determined. The two mutants exhibiting the greatest enhancement in enzymatic activity were selected for subsequent saturation mutagenesis. Saturation mutagenesis at these sites was performed using the same overlap extension PCR approach as in alanine scanning, with primers provided in Table S2. Finally, selected beneficial single-site mutants were combined through stepwise mutagenesis to generate combinatorial variants for evaluating potential synergistic improvements in catalytic activity.

2.5. Expression and purification of trypsin

Each verified recombinant Bacillus subtilis SCK6 strain was inoculated into 50 ml of fermentation medium and incubated at 37 °C with shaking at 220 rpm until reaching the logarithmic growth phase. At this stage, the OD₆₀₀ was measured, and the inoculum volume was calculated to ensure all subsequent cultures started at a uniform initial OD₆₀₀. Fermentation was conducted under identical conditions (37 °C, 220 rpm) and culture samples were collected at defined intervals. The bacterial cultures were centrifuged at 12,000 ×g for 10 min at 4 °C, and the resulting supernatants were collected for enzyme activity assays. All fermentation experiments were carried out in triplicate to ensure reproducibility.

The fermentation supernatants of mutant strains were filtered through a 0.22 μm membrane and subjected to affinity chromatography for protein purification. Prior to sample loading, the Benzamidine Agarose Beads 4FF column was regenerated with 6 M guanidine hydrochloride and washed with ultrapure water, followed by equilibration with 20 column volumes of Buffer A (50 mM Tris-HCl, 0.5 M NaCl, pH 7.4). Crude enzyme samples were applied at a flow rate of 1 ml/min until the column was saturated. After washing, bound proteins were eluted using Buffer B (10 mM HCl, 0.5 M NaCl, pH 2.0). The eluates were immediately neutralized to pH 8.0 using 1 M Tris-HCl (pH 9.0) and stored at 4 °C for downstream analyses.

2.6. Enzyme activity determination

Preparing the BAPNA (N-α-Benzoyl-DL-arginine-p-nitroanilide) substrate solution (25 mM) and Tris-HCl buffer (50 mM, pH 8.0) separately. The enzyme activity was measured using the following procedure: A reaction mixture containing 2 ml of preheated Tris-HCl buffer, 50 μL of substrate solution, and 200 μL of enzyme solution was prepared in a cuvette with a 1 cm optical path length. The mixture was thoroughly mixed, and the reaction was conducted at 37 °C. The absorbance increase at 410 nm was recorded over 5 min. One unit of enzyme activity is defined as an increase in absorbance of 0.1 per minute under the assay conditions. The amidase activity of trypsin was calculated using the following formula:

$$\text{Amidase activity}\left(\mathrm{U}/\mathrm{ml}\right)=\frac{{\mathit{\Delta A}}_{410\mathit{nm}/\min \times K}}{0.1\times 0.2.}$$

Among these, ΔA410_nm/min represents the increase in absorbance per minute, while K denotes the dilution factor of the crude enzyme solution.

For each sampling time point, culture supernatants were collected from three independently cultured mutant strains (n = 3). Each supernatant sample was subjected to three independent enzyme activity assays. The mean activity value of each biological replicate, calculated from the three technical replicates, was used for statistical analysis.

2.7. Kenetic parameter determination

To accurately determine the kinetic parameters of each mutant, the purified enzyme solution was first diluted to an appropriate concentration to ensure substrate saturation under the assay conditions. The initial reaction rates were then measured at varying substrate concentrations using the enzyme activity assay described in Section 2.6. Kinetic constants were calculated by fitting the data to the Michaelis–Menten equation, with substrate concentration plotted on the x-axis and reaction rate on the y-axis using nonlinear regression analysis.

2.8. Molecular docking and MD simulation

The amino acid sequence of the wild-type trypsin and all mutants were uploaded to Alphafold3 (https://alphafoldserver.com/) to predict the three-dimensional structure, and the 3D structure of the trypsin ligand (N-α-Benzoyl-DL-arginine-p-nitroanilide, BAPNA) can downloaded from pubchem (https://pubchem.ncbi.nlm.nih.gov/). Then, the molecular docking of the protein receptor and ligand was performed using AutoDock Vina. Upon completion of the docking process, the suitable docking conformation was selected for subsequent molecular dynamic simulation analysis.

Molecular dynamics simulations were performed using GROMACS 2023.04 (Berendsen, Van Der Spoel & Van Drunen, 1995). The Amber14SB force field was applied to the protein, while the ligand topology was generated using the general Amber force field (GAFF) with Sobtop software (Lu, 2025). Atomic charges for the ligand were assigned through quantum mechanical (QM) calculations. The system was solvated in a cubic box containing TIP3P water molecules, ensuring a minimum distance of 0.8 nm between the solute and the box edges. To achieve system neutrality, Na⁺ and Cl⁻ counterions were added. Energy minimization was carried out using the steepest descent algorithm until the maximum force was reduced below 1000 kJ·mol⁻¹·nm⁻¹.

Subsequently, NPT equilibration was conducted for 100 ps with a 1-fs timestep, applying positional restraints to solute heavy atoms. The system temperature was maintained at 298.15 K using the V-rescale thermostat, while isotropic pressure coupling was regulated by the Berendsen barostat. Long-range electrostatic interactions were computed using the Particle Mesh Ewald (PME) method with the Verlet cutoff scheme. Additionally, all hydrogen bonds were constrained using the LINCS algorithm. Trajectory frames were recorded every 1 ps for further analysis.

Production MD was carried out for 100 ns under periodic boundary conditions with a 2-fs integration time step. The isotropic pressure coupling was regulated by the Berendsen barostat. Structural stability was assessed via root-mean-square deviation (RMSD) of the protein backbone and ligand-heavy atoms. Additional analyses included root-mean-square fluctuation (RMSF) and hydrogen bond number.

2.9. Optimization of fermentation medium

The effects of key fermentation parameter-including the types and concentrations of carbon and nitrogen sources, as well as the initial medium pH-on enzyme production were systematically investigated. Carbon sources evaluated included glycerol, maltose, glucose, sucrose, starch, lactose, malt extract, mannitol, yeast extract, oat, and corn starch. The concentration of the selected optimal carbon source was further optimized within the range of 0 % to 4.5 % (w/v). Similarly, various nitrogen sources such as peptone, tryptone, soybean peptone, beef extract, corn syrup, cottonseed meal, soybean meal and casein peptone were assessed, followed by concentration optimization of the most effective nitrogen source from 0 % to 4 % (w/v). Additionally, the impact of initial pH on enzyme production was examined across a pH range of 4 to 11.

3. Results and discussion

3.1. Alanine scanning

Alanine scanning is a broadly employed mutagenesis technique in protein engineering (Morrison & Weiss, 2001). Given the small size and moderate hydrophobicity of alanine's side chain, replacing other amino acid residues with alanine typically does not induce substantial structural alterations in proteins. This characteristic renders alanine scanning an efficient method for pinpointing critical residues that play a role in protein function (Hao et al., 2021). As shown in Fig. 1a, wild-type trypsin contains 18 arginine and lysine residues, which serve as potential hydrolysis sites and may impact enzyme activity. Preventing hydrolysis at these sites could help preserve the tertiary structure of trypsin, thereby maintaining its catalytic activity. Solvent accessibility describes the extent to which each amino acid residue in a protein is exposed to the surrounding solvent, providing critical insights into the protein's structure and function. Furthermore, analyzing the solvent accessibility of the key residues within the protein can help to identify suitable mutation sites to enhance enzyme activity, stability and specificity. With the help of DSSP, the solvent accessibility of all the lysine and arginine residues within wild-type trypsin was calculated and analyzed. As shown in Fig. 1b, amino acid residues R6, K64, R90, K98, R119 and R218 possess larger solvent-accessible surface areas, making them more exposed to the aqueous environment and thus more prone to hydrolysis.

Fig. 1.

(a) Lysine and arginine residues in wild-type trypsin. (b) Solvent accessible surface area(ASA)of lysine and arginine residues. (c) Relative enzyme activity of alanine mutants.

All lysine and arginine residues were mutated to alanine using overlapping extension PCR, resulting in a total of 18 mutant strains. The enzyme activity results, shown in Fig. 1c, indicate that alanine substitutions at these 18 sites led to varying changes in enzymatic activity among the different mutants. Mutants K66A, K79A, K81A, R90A, R119A, K130A, K178A, R199A, and R218A exhibited an increase in enzymatic activity, whereas the remaining mutants showed varying degrees of activity reduction. Notably, the most significant enhancement in enzymatic activity was observed in mutants K81A, R119A, R199A and K79A. Among them, the two mutants with the highest activity improvement were selected for subsequent saturation mutagenesis and combinatorial mutagenesis. Among the mutation sites with enhanced enzyme activity, K79, K81, R90, R119 and R199 are situated in the random coil region, K66, R178 and K130 in the β-sheet region and R218 in the α-helix region. Notably, the two residues with the most pronounced improvements in enzyme activity are both located within the random coil region.

3.2. Saturation mutation and combination mutation

The two mutation sites, K81 and R119, which showed the most significant improvements in enzyme activity, were selected for subsequent saturation mutagenesis, resulting in a total of 34 mutants. The results of the saturation mutagenesis for K81 and R119 are presented in Fig. 2a,b, compared to the wild-type trypsin (WT), only some of the saturation mutants exhibited enhanced enzyme activity, and even fewer mutants showed improved basal enzyme activity relative to the alanine mutants. At the K81 site, three saturation mutants demonstrated increased enzyme activity, with K81P showing the most significant improvement over the K81A mutant. For the R119 site, seven saturation mutants with enhanced enzyme activity were identified, with R119E and R119T showing further improvements compared to the R119A mutant. Specifically, the enzyme activities of K81P, R119E and R119T increased by 83.2 %, 43.7 % and 52.0 %, respectively. Based on these single-point mutations, combinatorial mutagenesis was performed to construct two combined mutants-K81P-R119E and K81P-R119T. As shown in Fig. 2c, the combined mutations exhibited an additive effect, with the enzyme activities of the combined mutants reaching 83.28 U/ml and 93.87 U/ml, respectively, representing 1.40-fold and 2.16-fold increases compared to wild-type trypsin (WT) (Fig. 2d).

Fig. 2.

Saturation mutations of lysine and arginine residues. (a) Relative activities of saturation mutants of K81. (b) Relative activities of saturation mutants of R119. (c) Curves of enzyme activities of some single-point mutants and combined mutants changing with time. (d) Comparison of relative activities of combined mutants and single point mutants.

Compared with previously reported trypsin heterologous expression systems, the trypsin mutant strain constructed in this study using Bacillus subtilis SCK6 as the host exhibited significantly enhanced enzyme activity under shake flask fermentation conditions, reaching 93.87 U/ml. This activity is notably higher than that of the GM2938 strain, which achieved 33.8 U/ml using the same host and the pWB980 plasmid (Wang et al., 2020). In addition, the enzyme activity obtained in this study also exceeds that of many trypsin variants expressed in yeast systems under similar conditions. For example, Niu et al. developed a Streptomyces fradiae mutant strain (R295A/R315A) through propeptide engineering and modification of specific hydrolysis sites, yet the final enzyme activity was only 0.59 U/ml after conversion, which is substantially lower than the value reported in this study (Niu et al., 2024). Similarly, Zhang et al. engineered the trypsin from Streptomyces griseus by introducing a hydrophobic leader peptide (FVEF) at the N-terminus and constructing mutants R145I and K101A. These strains exhibited enzyme activity levels of 6.87 ± 0.11 U/ml and 40.57 ± 2.97 U/ml, respectively (Zhang, Huang, Yao, Du & Kang, 2017), both of which remain significantly lower than the activity achieved by the strain developed in the present work. These comparisons underscore the significant advancement achieved in this study, positioning our engineered strain among one of the highest-yielding trypsin expression systems reported to date in shake flask fermentation. Unlike previous methods that relied solely on alanine scanning or single-site saturation mutagenesis, this study employed an integrated strategy combining alanine scanning, targeted saturation mutagenesis, and combinatorial mutagenesis to systematically enhance trypsin activity. This synergistic approach proved essential, as neither alanine scanning nor single-site saturation mutagenesis alone resulted in comparable improvements.

To further characterize the impact of these mutants, wild-type trypsin and selected mutants were purified using highly specific affinity chromatography, followed by a comprehensive analysis of their enzymatic kinetics. As summarized in Table 1, the K81P-R119T mutant exhibited superior enzymatic properties, with a specific activity of 1021.75 U/mg and a Kcat/Km value of 1.002 × 10⁶ min⁻¹·mM⁻¹, significantly surpassing those of wild-type trypsin (300.45 U/mg, 3.244 × 10⁵ min⁻¹·mM⁻¹). These findings confirm that site-directed mutagenesis effectively modulates enzymatic properties, leading to a substantial enhancement in catalytic efficiency.

Table 1.

Enzyme kinetic parameters and specific enzyme activities of wild-type trypsin and mutants.

Trypsin	Specific activity (U/mg)	Km(mM)	Kcat (min−1)	Kcat /Km (min−1·mM−1)
WT	300.45	0.510900	1.658E+05	3.244E+05
K81P	686.20	0.346400	3.296E+05	9.516E+05
R119E	759.33	0.447500	2.320E+05	5.184E+05
R119T	638.25	0.455600	2.648E+05	5.812E+05
K81P-R119E	797.84	0.466400	4.035E+05	8.652E+05
K81P-R119T	1021.75	0.218000	2.183E+05	1.002E+06

3.3. Molecular docking

The amino acid sequence of wild-type trypsin and the above mutants was uploaded to the official website of Alphafold3 to predict the three-dimensional structure. After the prediction was completed, the conformation with the best score was selected for subsequent molecular docking. The substrate for the activity determination of trypsin in this experiment was BAPNA (N-benzoyl-DL-arginyl-4-nitroaniline hydrochloride), so the BAPNA was used as a ligand for molecular docking with the protein receptor. Trypsin specifically hydrolyzes peptide bonds on the carboxyl side of lysine or arginine residues through a catalytic triad composed of Ser172, His37 and Asp82 (Kraut, 1977). During catalysis, the hydroxyl group of Ser172 acts as a nucleophile, attacking the carbonyl carbon of the substrate peptide bond to form a tetrahedral intermediate. Concurrently, His37 abstracts a proton from Ser172, while Asp82 stabilizes the conformation of His37, enhancing its proton transfer ability. The intermediate subsequently collapses, leading to the release of the amine product and the formation of an acyl-enzyme intermediate. Finally, a water molecule, activated by His37, hydrolyzes the acyl-enzyme intermediate, regenerating the enzyme and completing the catalytic cycle (Craik, Roczniak, Largman & Rutter, 1987).

To investigate the structural basis for enhanced catalysis, the distance between the hydroxyl oxygen atom of the Ser172 side chain and the carbonyl carbon atom of the amide bond in the arginine side chain of the substrate BAPNA was analyzed. As shown in Fig. 3 (a, c, e), compared to the wild-type trypsin, the distance between the hydroxyl oxygen atom of Ser172 and the carbonyl carbon atom of the substrate in the K81P-R119E and K81P-R119T mutants decreased from 4.6 Å to 2.8 Å and 3.1 Å, respectively. This reduction is expected to facilitate proton transfer, lower the catalytic barrier, and thereby enhance the trypsin's catalytic efficiency. Lu et al. employed molecular docking analysis to investigate the UGT30033 mutant M1–1 and observed a reduction in the distance between the O6 atom at the C24 position of glucose and the N22 atom of His21 from 3.6 Å in the wild-type enzyme to 2.9 Å (Lu et al., 2024). The shortened distance between key catalytic residues facilitated proton transfer, thereby significantly enhancing catalytic efficiency. Similarly, Guo et al. utilized structure-guided engineering to construct the UGT76G1-T284S/M88L/L200A mutant, which exhibited improved enzymatic performance (Guo et al., 2022). Specifically, the reduced spatial distance between the nucleophilic attack site of the substrate Reb D and the sugar donor UDPG enabled the formation of a more stable and efficient catalytic complex, ultimately resulting in a marked increase in catalytic activity. The reduced distance between key nucleophilic attack sites in the substrate Reb D and the sugar donor UDPG contributed to the formation of a more stable and efficient catalytic complex, leading to a substantial increase in catalytic activity. Through molecular docking and molecular dynamics simulations, analysis of the atomic distances between ligand and receptor molecules, particularly those critical for nucleophilic attack, revealed that enhancements in enzymatic catalytic efficiency are closely associated with a reduction in key atomic distances. This mechanism has been experimentally validated in the engineering and optimization of various enzymes, including glycosyltransferases (Guo et al., 2023) and dehydrogenases (Wu et al., 2022), providing a robust strategy for the rational design of enzymes with improved catalytic performance.

Fig. 3.

Schematic diagram of molecular docking between receptor protein and ligand. (a) Molecular docking diagram of wild-type trypsin and ligand. (b) Overlay of ligand trajectories of wild-type trypsin at different times. (c) Molecular docking diagram of mutant K81P-R119E and ligand. (d) Overlay of ligand trajectories of mutant K81P-R119E at different times. (e) Molecular docking diagram of mutant K81P-R119T and ligand. (f) Overlay of ligand trajectories of mutant K81P-R119T at different times.

Notably, although docking results indicate that the key atomic distance in the K81P-R119T mutant (3.1 Å) is slightly larger than that in K81P-R119E (2.8 Å), the trypsin activity of K81P-R119T (93.87 U/ml) surpasses that of K81P-R119E (83.28 U/ml). This discrepancy may not only stem from the intrinsic differences between threonine and glutamate but also be associated with the proximity of R119 to the critical residue D166 at the base of the substrate-binding pocket. D166 plays a pivotal role in recognizing substrates with positively charged side chains, such as arginine and lysine (Hartley, 1979). Consequently, the presence of R119, which carries the same charge as the substrate, may interfere with substrate binding by disrupting the interaction between D166 and the substrate. Substituting R119 with threonine or glutamate likely alleviates this unfavorable effect, thereby enhancing D166's ability to recognize and position the substrate more effectively, ultimately improving catalytic efficiency. Given the longer, negatively charged side chain of glutamate, its substitution may alter the electrostatic environment within the substrate-binding pocket, potentially affecting substrate affinity and catalytic efficiency. In contrast, threonine, due to its shorter side chain and milder electrostatic characteristics, may alleviate steric hindrance and reduce unfavorable electrostatic interactions, thereby promoting a more favorable conformation of the substrate-binding pocket for efficient substrate recognition. Although R119 is not directly adjacent to the catalytic triad, its mutation may alter the local structural or electrostatic environment, indirectly modulating enzymatic activity.

3.4. MD simulation

3.4.1. RMSD analysis

The binding of the protein and ligand was simulated using the GROMACS software package. The protein backbone was employed as the least-squares fitting object, and the relative position of the ligand BAPNA with respect to the protein was analyzed to preliminarily assess the binding interaction between the ligand and the protein receptor. At the same time, the trajectories of the ligand at different times were extracted at regular intervals. Subsequently, the ligand trajectories at different times were superimposed and finally the superimposed conformation of the ligand was visualized in the active pocket of the protein. As we can see in Fig. 3 (b, d, f) the ligand fluctuation amplitude in the docking conformation of the wild-type trypsin and mutants was small, and there was no phenomenon of ligand falling off from the active pocket due to the large amplitude of ligand movement, indicating that the binding of the ligand to the receptor protein was relatively stable.

Furthermore, during the molecular dynamics (MD) simulations, the relative motion of the ligands with respect to the protein backbone was analyzed. As shown in Fig. 4a, the RMSD trajectories of all ligands exhibited moderate fluctuations. Although the RMSD values are somewhat higher than those typically reported in similar studies (Kato et al., 2021), they remain within an acceptable range. Studies have shown that the presence of phenyl groups in ligands increases their conformational flexibility. Therefore, an RMSD fluctuation of 5.6 Å can still be considered indicative of stable binding (Wieder et al., 2016). Notably, the ligand in the mutant trypsin complex displayed a lower RMSD value compared to that in the wild-type trypsin, suggesting a more stable interaction between the ligand and the mutant protein. This reduced conformational fluctuation may facilitate more efficient and stable mass transfer during catalysis.

Fig. 4.

Analysis of dynamic simulation results. (a) The relative distance (RMSD) between the ligand and the receptor protein backbone. (b) Frequency distribution of the actual number of hydrogen bonds between the receptor protein and the ligand during the dynamics simulation. (c) Frequency distribution of the potential number of hydrogen bonds between the receptor protein and the ligand during the dynamics simulation. (d) The distance of hydrogen bonds.

3.4.2. Analysis of hydrogen bond network

The changes in the hydrogen bond network between the ligand and the receptor protein during the simulation were analyzed. As shown in Fig. 5, the number of hydrogen bonds formed between the ligands and both wild-type trypsin and mutant trypsin fluctuated dynamically. The number of stable hydrogen bonds between the small-molecule ligand and the receptor protein remained in the single digits, which is consistent with previous literature reports (Jiang et al., 2024; Salmaso et al., 2017). Among them, the mutant K81P-R119T had the largest number of confirmed hydrogen bonds, with the largest number of hydrogen bonds being 5, and showing a certain normal distribution trend (Fig. 4b). In contrast, the wild-type trypsin and the mutant K81P-R119E had fewer hydrogen bonds, with the largest number of hydrogen bonds being only 2, and more hydrogen bonds would provide stronger interaction forces. The formation of hydrogen bonds requires a certain distance and angle to be satisfied at the same time, with an angle of 180 degrees being the best (Steiner, 2002). Fig. 4c shows the number of hydrogen bonds that meet the distance condition but temporarily do not meet the angle condition. It can be seen that the frequency distribution of the number of potential hydrogen bonds in the mutant K81P-R119E has the largestsss range of hydrogen bonds. Since the docking of receptor protein and ligand is a dynamic process, at certain moments, some hydrogen bond donors and receptors may temporarily meet the angle conditions to form hydrogen bonds.

Fig. 5.

Curve of the number of hydrogen bonds between the receptor and the ligand changing with time during the dynamic simulation. (a) Curve of actual hydrogen bond number in wild-type trypsin. (b) Curve of actual hydrogen bond number in mutant K81P-R119E. (c) Curve of actual hydrogen bond number in mutant K81P-R119T. (d) Curve of the number of potential hydrogen bonds in wild-type trypsin. (e) Curve of the number of potential hydrogen bonds in mutant K81P-R119E. (f) Curve of the number of potential hydrogen bonds in mutant K81P-R119T.

According to the hydrogen bond formation theory, a strong hydrogen bond can be formed when the distance between the donor and the receptor is between 0.22 and 0.25 nm (Steiner, 2002); a weaker hydrogen bond is formed when the distance is greater than 0.32 nm; and a moderate hydrogen bond is formed when the distance is between the 0.25–0.32 nm. As shown in Fig. 4d, the hydrogen bonds formed between each receptor protein and the ligand are of medium strength. In summary, compared with wild-type trypsin (WT), mutants K81P-R119E and K81P-R119T can form a stronger hydrogen bond network with the ligand, which helps to enhance the interaction between the receptor and ligands. This enhanced interaction may provide more effective support for proton transfer and the maintenance of the catalytic conformation, thereby indirectly promoting the catalytic activity of the trypsin. Li et al. systematically analyzed the interactions between eight glucose dehydrogenase (GDH) mutants and the substrate NADP⁺ (Li et al., 2022). The study revealed that, compared with the wild-type enzyme, certain mutants (such as P45A) formed additional hydrogen bonds with NADP⁺ involving several key amino acid residues. Notably, the catalytic activity of the P45A mutant was significantly enhanced, with specific activity and kcat/km increased by approximately 28-fold and 7-fold, respectively. This enhancement is likely attributable to the increased number of hydrogen bonds formed between the enzyme and substrate, which stabilizes the enzyme–substrate complex and improves both substrate affinity and catalytic efficiency. These findings are consistent with the conclusions of the present study. Furthermore, other studies also demonstrate that increasing the number of hydrogen bonds or reorganizing the hydrogen bond network is an effective strategy for enhancing enzymatic activity (Xu, Liu, Jiang & Liang, 2024).

3.4.3. Analysis of RMSF

By employing molecular dynamics simulations, the fluctuations of each amino acid residue in both the wild-type trypsin and its mutants were analyzed throughout the simulation process. As shown in Fig. 6a, due to the different structural regions and amino acid properties, the fluctuation of each amino acid residue is quite different. In general, the fluctuation of the N-terminal and C-terminal regions is small, and the fluctuation of the middle region is large. Compared to the wild-type trypsin, the fluctuations of amino acid residues in the R1, R2, and R3 regions of the mutants are significantly reduced, particularly in the R1 region.

Fig. 6.

Three-dimensional structural characteristics of wild-type trypsin and its mutants. (a) The RMSF of wild-type trypsin and its mutants. (b) Reigon 1 (R1). (c) Reigon 2 (R2). (d) Reigon 3 (R3).

As shown in Fig. 6b, the amino acid residues with reduced fluctuations in region R1 are mainly amino acid residues at positions 73–81 and the mutation site K81 is located in this region. It can be seen that the R1 region adopts a random coil secondary structure with relatively low stability, while the lysine-to-proline substitution at position 81 markedly enhances its stability. It is well known that the random coil region in the protein structure is less stable than the β-fold and α-fold. Structurally, the proline side chain forms a five-membered ring (pyrrolidine ring) with the main chain nitrogen atom. Due to the presence of this five-membered ring, the dihedral angle of proline is restricted to a smaller range, which makes the conformational freedom of the amino acid residue significantly smaller, thereby significantly reducing its spatial fluctuation (Yun, Anderson & Hermans, 1991). From the perspective of amino acid properties, proline is a hydrophobic residue known to enhance protein stability through hydrophobic interactions, particularly in maintaining the integrity of core structural regions (Deber, Brodsky & Rath, 2010). This is consistent with the observation that the mutation site K81 is located near the catalytic active center. Typically, reduced structural flexibility of amino acid residues surrounding the catalytic active site facilitates more stable substrate binding and proton transfer, as well as enhanced stabilization of the transition state, thereby improving the catalytic efficiency of the enzyme to some extent. Kamal et al. constructed a lipase mutant (6B) and demonstrated, through enzymatic activity assays combined with molecular dynamics simulations, that a reduction in active site flexibility was accompanied by an increase in catalytic efficiency (Kamal et al., 2012). This finding is consistent with the conclusions of the present study.

As shown in Fig. 6c, the mutation site R199 lies within the R2 region, where the observed reduction in fluctuation may be attributed to the physicochemical characteristics of the substituted residues. The bulky side chain of arginine likely introduced significant steric hindrance at its original position. Upon substitution with threonine or glutamate, which possess smaller side chains, steric constraints are alleviated, leading to a more compact local structure and reduced fluctuation. As for the R3 region, which is more distant from the mutation site (Fig. 6d), the decrease in residue fluctuation may involve long-range interactions and rearrangement of the hydrogen bond network. Studies have shown that even when a mutation site is spatially distant from a target region, alterations in charge distribution or conformational adjustments can propagate through hydrogen bonding and hydrophobic interaction networks. For instance, in T4 lysozyme, mutations at K85 and D89 were found to influence the stability of distal residues by modulating the electrostatic environment (Mooers, Baase, Wray & Matthews, 2009). Similarly, the mutations at K81 and R119 may induce structural stabilization in the R3 region via long-range interaction networks.

In conclusion, although an increase in the overall structural stability of a protein may reduce the flexibility of certain regions such as the active site, it does not necessarily compromise enzymatic activity. On the contrary, a moderate enhancement of structural rigidity while maintaining adequate local flexibility may lead to improved catalytic efficiency. This observation provides useful guidance for the rational design and engineering of enzymes.

The solvent accessibility of the mutation sites before and after modification was analyzed, and the results are presented in Fig. 7, When K81 was mutated to proline, the Solvent accessible surface area (ASA) at this position decreased from 85 Å² in wild-type trypsin to 25 Å² in the K81P-R119E mutant and 28 Å² in the K81P-R119T mutant. Similarly, when R119 was mutated to glutamic acid and threonine, the ASA at this position decreased from 186 Å² to 134 Å² (K81P-R119E) and 91 Å² (K81P-R119T), respectively. Heinemeyer et al. investigated the effects of active-site mutations in the eukaryotic 20S proteasome on precursor processing and found that structural changes in the active site influence enzyme maturation and stability (Heinemeyer et al., 1997). A reduction in solvent accessibility around the active site due to mutations may enhance enzyme stability, thereby improving enzymatic activity. Furthermore, studies have shown that the introduction of hydrophobic amino acid mutations can enhance hydrophobic interactions, promoting the formation of new hydrophobic cores that stabilize the active-site conformation (Mittal, 2012). Since the active site of a protease requires a specific physicochemical environment, including optimal solvent accessibility, changes in ASA may influence the conformational stability of the protease, ultimately affecting its catalytic activity.

Fig. 7.

Solvent accessible surface area (ASA) of wild-type trypsin and mutants.

3.5. Optimization of fermentation

Fermentation-based enzyme production involves complex biochemical processes within microorganisms, making the optimization of growth conditions essential for maximizing enzyme yield. To identify the optimal fermentation parameters, a systematic investigation was carried out to assess the effects of different carbon sources and their concentrations, nitrogen sources and their concentrations, as well as pH levels on microbial enzyme production.

As shown in Fig. 8a, starch was identified as the optimal carbon source for enzyme production. Under the same initial inoculation OD value, after 48 h of fermentation, the trypsin activity in the starch-containing medium was 61.29 % higher than that in the medium supplemented with the basal carbon source (yeast extract powder). These results suggest that starch serves as a more favorable carbon source for fermentation-based enzyme production in Bacillus subtilis SCK6. Consistent with previous studies, the use of starch has been shown to enhance enzyme yield (Salman et al., 2016). As a complex polysaccharide, starch may require enzymatic hydrolysis by amylase secreted by the bacteria, converting it into monosaccharides that can be directly utilized by microorganisms (Salman et al., 2016). Research has demonstrated that, compared to rapidly metabolizable carbon sources such as glucose, structurally more complex carbon sources are less prone to induce catabolite repression, thereby alleviating its inhibitory effects on microbial fermentation and enzyme production (Chen et al., 2012). These properties are particularly important for the biosynthesis of secondary metabolites, such as trypsin. The optimization results for different nitrogen sources (Fig. 8c) indicated that corn syrup powder was the most suitable nitrogen source for enzyme production. After 48 h of fermentation, the trypsin activity in the medium supplemented with corn steep liquor was 89.02 % higher than that in the medium containing the basal nitrogen source. Corn syrup powder is rich in amino acids, vitamins, and minerals, which promote microbial growth and metabolic activity (Bai, 2022). In contrast, peptone provides relatively simple nutrients. Additionally, corn syrup powder contributes to optimizing the carbon-to‑nitrogen (C/N) ratio in the culture medium. An appropriate C/N ratio is critical for the synthesis of metabolites, including enzymes (Song & Chang, 2003). The optimization experiments for carbon and nitrogen source concentrations (Fig. 8b) revealed that as the concentrations of starch and nitrogen source increased, the enzyme activity in the culture medium also gradually increased. However, when the starch and corn syrup powder concentrations reached approximately 3 %, the enzyme activity plateaued. Considering cost-effectiveness, an optimal supplementation of 3 % starch and 3 % corn syrup powder was determined to be the most economical choice for fermentation.

Fig. 8.

Optimization of Trypsin Fermentation. (a) Effect of different carbon source. (b) Effect of different carbon and nitrogen source concentration. (c) Effect of different nitrogen source. (d) Effect of different pH.

The pH optimization experiments (Fig. 8d) demonstrated that a pH range of 6–7 is the most favorable for trypsin fermentation. Deviations beyond this range, either towards more acidic or alkaline conditions, resulted in a significant decline in trypsin activity. This finding is consistent with previous reports, which have identified neutral pH as the optimal condition for Bacillus subtilis fermentation (Song & Chang, 2020; Zhang, Guo & Liu, 2018). Following fermentation optimization, the enzyme activity of the mutant K81P-R119T reached 132.8 U/ml, representing a 7.38-fold increase compared to pre-optimization levels.

4. Conclusion

In this study, a high-performance trypsin mutant, K81P-R119T, was successfully engineered through saturation mutagenesis and combinatorial mutagenesis targeting key amino acid residues of a novel trypsin variant. Compared to the wild-type trypsin (29.7 U/ml), the K81P-R119T mutant exhibited a 2.2-fold increase in enzymatic activity, reaching 93.9 U/ml. Molecular docking and molecular dynamics simulations revealed that the introduced mutations shortened the catalytic distance, reinforced the hydrogen bond network between the enzyme and its substrate, and reduced residual fluctuations in the catalytic pocket, thereby enhancing enzymatic stability and efficiency. Furthermore, fermentation process optimization further elevated the enzyme activity to 132.8 U/ml, underscoring the potential of this engineered trypsin for broader food industry.

CRediT authorship contribution statement

Jia Chen: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Guangyang Jiang: Writing – review & editing, Supervision, Conceptualization. Yongqiang Tian: Writing – review & editing, Resources, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary material

Data availability

Data will be made available on request.