To identify potential isozymes of papain, the NCBI (National Center for Biotechnology Information) Blast tool was employed for sequence homology search in this study. The amino acid sequence of papain from Carica papaya was used as the query sequence to search the non-redundant protein sequence database (nr database), with the sequence similarity threshold set at 50%. This ensured that the screened sequences had a relatively close evolutionary relationship with papain and might possess similar structural and functional characteristics. After rigorous screening, 30 homologous sequences from different species were identified. Subsequently, the detailed amino acid sequence information of these homologous sequences was obtained through the GenBank platform of the NCBI database, providing the basic data for the subsequent structural prediction and functional analysis.

To obtain the three-dimensional structural models of these isozymes, the AlphaFold tool was used in this study. The downloaded amino acid sequences were input into the AlphaFold platform in FASTA format. Prediction parameters were set, including model selection (the AlphaFold2 model was adopted, which has significant advantages in prediction accuracy) and sequence coverage (to ensure that the predicted structure could cover the complete amino acid sequence). The prediction task was initiated, and the AlphaFold platform, with its powerful computing capabilities and deep learning algorithms, performed structural predictions on the input sequences and output the predicted three-dimensional structural files (in PDB format). The predicted structural files were preliminarily checked to ensure the integrity and rationality of the structures. Using the visualization tools of the AlphaFold platform, key regions of the predicted structures (such as active sites and binding pockets) were observed and analyzed in detail, providing an accurate structural basis for the subsequent molecular docking experiments.

After obtaining the three-dimensional structural models of the isozymes, molecular docking experiments were conducted using AutoDock Vina software in this study to evaluate the binding affinity between these isozymes and the substrate Pglu-phe-leu-pna.

Substrate preparation: First, the molecular structure of the substrate Pglu-phe-leu-pna was imported into AutoDock Vina software in PDB format. The substrate molecule was optimized using chemical drawing software such as ChemDraw to ensure its accuracy and stability. The optimized substrate molecular structure was used as the ligand in the docking experiment.

Receptor preparation: The three-dimensional structural file of the isozyme predicted by AlphaFold was imported into AutoDock Vina software as the receptor in the docking experiment. During the import process, necessary preprocessing was performed on the receptor structure, including adding hydrogen atoms and assigning charges, to ensure the accuracy and rationality of the receptor structure during the docking process.

Docking process and result analysis: The molecular docking task was run, and AutoDock Vina software searched and optimized the possible binding modes of the ligand in the receptor's binding pocket through genetic algorithms and local search algorithms. After docking, the software output the docking results, including the docking score (in kcal/mol, with lower values indicating stronger binding affinity) and the optimal binding pose of the ligand in the receptor's binding pocket. By statistically analyzing the molecular docking results of 30 isozymes with the substrate, the isozymes with stronger binding affinity were screened out, providing a basis for further binding force analysis.

For the isozymes that exhibited strong binding affinity in the molecular docking experiments, detailed binding force analysis was conducted using Discovery Studio software in this study to gain a deeper understanding of the molecular interaction mechanisms between them and the substrate Pglu-phe-leu-pna. Discovery Studio is a powerful molecular simulation and drug design software that offers a variety of analysis tools for studying protein-ligand interactions. The binding force analysis included hydrogen bond, hydrophobic interaction, π-π stacking interaction, and electrostatic interaction analysis. The Discovery Studio software's relevant modules identified, calculated, and statistically analyzed the number, strength, distribution, and synergistic relationships of each interaction, to assess the binding situation between the isozymes and the substrate.

First, we did the sequence homology search and structure prediction. As shown in the figure 1 about sequence comparison analysis, we searched for 30 homologous sequences with a similarity of more than 50% to the papain sequence through NCBI Blast, and used AlphaFold to predict their three-dimensional structures. The predicted structure provides a solid foundation for the subsequent molecular docking experiments.

Figure 1 Sequence alignment analysis of papain isoenzymes

The molecular docking results as shown in table 1. According to this table which use molecular docking predicts the affinity of papain isoenzymes to substrates. We can reach the first conclusion: the binding force stability values of each isoenzyme are different. Obviously, the negative value of papain is the largest. It indicates that its binding affinity is the strongest, and the following four are the isoenzymes we found.

The docking score indicates that these isoenzymes have formed stable complexes with the substrate, suggesting that they may have similar functions to papain.

Table 1 Molecular docking predicts the affinity between papain isoenzyme and substrate Pglu-phe-leu-pna

This figure 2 shows that these several isoenzymes have all formed dense hydrogen bonds with the substrate. It is notable that papain not only forms hydrogen bonds but also generates additional π bonds stacking interactions with the substrate. These interactions of π bonds may help enhance binding stability, thereby explaining why papain has a stronger affinity compared to other isoenzymes.

It can be concluded from the table that the affinities of other isoenzymes are weaker than that of papain because other isoenzymes only form dense hydrogen bonds and do not form π bonds.

Moreover, from the results of molecular docking prediction, it can be found that the affinities of other isoenzymes are different because their binding sites are distinct.

The results of molecular docking are consistent with those of the ELISA mentioned earlier, strongly supporting the experimental results. It proves that the papain synthesised by microorganisms has the activity of papain extracted from natural plants, providing a research basis for subsequent studies.

1. Papain

It has got a binding affinity of -7.246 kcal/mol. From the perspective of hydrogen bonds, Cys25 in the catalytic center forms a stable hydrogen bond with carboxyl group of the phenylalanine in the substrate. From the perspective of the π-π stacking, Trp177 in papain forms a parallel π-π stacking interaction with the aro matic ring of the substrate’s phenylalanine residue. Val133 and Leu141 create a hydrophobic pocket that stabilizes the substrate’s central segment. The carboxyl group of the substrate interacts with Lys157 to form a salt bridge, enhancing substrate positioning.

The combination of strong hydrogen bonds, π-π stacking, and salt bridges leads to highly stable substrate binding, explaining why papain has the strongest affinity among all isozymes.

2. Actinidia eriantha

It has got a binding affinity of -7.192 kcal/mol. From the perspective of hydrogen bonds, there are five hydrogen bonds, mostly at the N-terminal of the substrate. Active site architecture is conserved but Trp is replaced by Ile—no π-π stacking observed. Ile178 and Leu130 stabilize the substrate via van der Waals forces.

To conclude, there is strong affinity, but the lack of π-π stacking reduces overall stability compared to papain.

3. Abrus precatorius

It has got a binding affinity of -6.777 kcal/mol. From the perspective of hydrogen bonds, there are only three hydrogen bonds, and mainly at the mid-section of the substrate. The active site is slightly open, resulting in tilted substrate orientation.

To conclude, there is Moderate binding affinity and weaker interaction due to partial misfit of substrate in the pocket.

4. Cicer arietinum

It has got a binding affinity of -6.763 kcal/mol. From the perspective of electrostatic interactions, N-terminal of substrate is relatively stable and C-terminal is loosely bound. There is no π-π stacking present in the structure.

To conclude, binding force of cic is weaker due to mismatch between substrate shape and pocket geometry.



Figure 2 Analysis of the interaction between different sources of papain isoenzymes and the substrate Pglu-phe-leu-pna. (A) cicer arietinum source, (B) Abrus precatourius source, (C) Actinidia eriantha source, (D) Cajanus cajan source, (E) Carica papaya source.

To quantify enzymatic activity, a standard curve was established using a range of known substrate concentrations. The equation derived from the curve was:

Y=0.2787X+0.1783

where Y represents the absorbance at 450 nm (A450), and X represents the product concentration in ng/mL. The high degree of linearity in the curve confirmed the accuracy and reliability of subsequent concentration calculations based on absorbance data.

Figure 3 The standard curve of papain concentration

Individual enzyme assay revealed that papain exhibited the highest catalytic activity toward the substrate Pglu-phe-leu-pna, with an A450 value of 1.036. This result aligns perfectly with the molecular modeling prediction, where papain showed the strongest binding affinity and additional π-π stacking interactions. Other isozymes, such as those from Actinidia and Cicer ,also demonstrated moderate to high activity levels. Overall, a strong relationship was observed between predicted binding affinity and experimental enzymatic efficiency.

Figure 4 Graphs of single-enzyme activity comparison

Among all dual enzyme combinations tested, the pairing of Cajanus cajan and Cicer arietinum (Caj + Cic) yielded the highest substrate conversion rate with a A450 value of 1.126, even higher than papain-inclusive combinations. This suggests a potential synergistic effect, likely due to complementary structural features and minimized competition for the active site. In contrast, combinations like papain + Abrus(1.063) or papain + Actinidia (0.999) showed slightly reduced activity, possibly due to overlapping substrate binding preferences or partial active-site interference. These results suggest that pairing enzymes with structurally complementary features can significantly enhance catalytic performance.

Figure 5 Graphs of dual-enzyme combination comparison

Introducing a third enzyme into the system did not lead to further improvements in enzyme activity. In some cases, such as papain + Abr + Cic(0.907), the total activity was lower than that of their dual-enzyme counterparts(1.063). This reduction is likely the result of substrate competition or competition between active sites, which impairs the overall coordination and efficiency of the reaction.

Figure 6 Graphs of Three-enzyme combinations comparison

Combinations involving four or five enzymes showed a clear decline in enzyme activity. Although extensive substrate identification is theoretically beneficial, the increase in complexity leads to a decrease in activity. This may be attributed to excessive substrate competition, structural mismatches, and the lack of spatial coordination among enzymes. These observations highlight the importance of precise selection and designing when constructing multi-enzyme systems.

Figure 7 Graphs of four-enzyme combinations comparison

To sum up, the ELISA experiment does validate our previous predictions. It is matched that papain demonstrate the highest activity when there is single enzyme. The combination of Caj + Cic seems to be the most effective one according to the result. However adding more enzyme in did not enhance the enzyme activity and even decrease the effect. These findings suggest that the compacity between enzymes and the coordinate ability of different enzymes should be take into consideration when designing a multi-enzyme system.

Figure 8 Graphs of optimal combinations comparison