Modeling plays a crucial role in our project. Molecular docking is an important molecular simulation method, mainly used to study the binding between small molecules (ligands) and large molecules (receptors). We performed molecular docking of PET hydrolase (dsPETase) with polyethylene terephthalate (PET)^[1-2]. The docking results can predict the key active sites of the wild-type PET hydrolase (dsPETase), thereby providing theoretical support for our next step in screening dsPETase mutants. Next, Autodock software was used to observe the changes in the amino acid interaction sites and binding effects between the eight mutants and PET. Finally, experimental data was utilized to verify the efficiency of dsPETase and the eight mutants in degrading PET plastic. Ultimately, the mutant with the best PET degradation performance was selected. The model construction plays a crucial role in optimizing and advancing the research.

Background:

Molecular docking represents a principal methodology in molecular simulations, fundamentally characterizing the reciprocal recognition process between molecules through geometric and energetic complementarity. Geometric matching establishes the prerequisite for molecular interactions, enabling computational prediction of optimal spatial binding conformations between ligands and receptors. Energetic matching further ensures binding stability by quantifying weak intermolecular forces—including hydrogen bonding, van der Waals interactions, and π–π stacking—within receptor-ligand complexes. This integrated approach identifies the lowest-energy conformation corresponding to the most stable complex. Through computational simulation, ligand placement within receptor binding sites allows robust prediction of both binding affinity and molecular recognition patterns[3-5].

Goal:

Here, we leverage a PET hydrolase (dsPETase) identified from a deep-sea hydrothermal vent at a depth of 1199 meters. This enzyme catalyzes the hydrolysis of polyethylene terephthalate (PET) into soluble monomeric units, offering a promising avenue for plastic recycling. To enhance the catalytic efficiency of dsPETase against PET, we predicted key active site residues responsible for binding the PET polymer and performed structure-guided site-directed mutagenesis to engineer optimized variants.

The process is as follows :

1. Preparation of the dsPETase protein structure:

The amino acid sequence of dsPETase was retrieved from the NCBI database (https://www.ncbi.nlm.nih.gov/) and subsequently submitted to AlphaFold for prediction of its three-dimensional protein structure(Figure 1). The amino acid sequence is as follows:

MTNPGGGGGGSNPDTGTGFPVVSSFSADGSFATTSGSAGLSCVFFSRAPWARMAKTSDYCVGQRRHRRPLAPYAGILEHWASQGFVVIAANTNNGHRQRYVKLRGYLTNQNNRSTGTYANKLDNRNRIGAAGHSQGGGGTIMAGQDYAISVTRPFQPYTIGLGHNSSSRSIQNGPLFLMTGSADTIASPTLNALPVYNRANVPVFWGELSGASHFEPVGSAVDFAVLAPRGFVIMLMDDASREDTFMASNCDLCTDNDWDVRRKGLT



Figure 1. Three-dimensional structure of dsPETase predicted by AlphaFold.

The protein structure was preprocessed using PyMOL 2.3.0 to remove crystalline water molecules and original ligands. Subsequently, the structure was imported into AutoDockTools (v1.5.6) for hydrogen addition, charge calculation, charge assignment, and atom type designation, followed by conversion to the “pdbqt” format.

2. Preparation of the PET Ligand Structure:

The monomeric unit of polyethylene terephthalate (CAS: 25038-59-9) was sketched using ChemBioDraw Ultra 14.0(Figure 2). The structure was subsequently imported into ChemBio3D Ultra 14.0 for energy minimization, with the convergence criterion set to a minimum RMS gradient of 0.001. The optimized structure was saved in MOL2 format and then processed in AutoDockTools-1.5.6 to add hydrogen atoms, compute partial charges, assign atomic types, and define rotatable bonds. The final ligand structure was exported in “pdbqt”format for docking simulations.



Figure 2. Chemical structure of the polyethylene terephthalate

3. Molecular Docking
   The binding site of the protein was predicted using POCASA 1.1. Molecular docking was performed with AutoDock Vina 1.1.2, with the following receptor parameters: center_x = 5.4, center_y = −13.0, center_z = 11.8; grid box dimensions: size_x = 60, size_y = 60, size_z = 60 (grid spacing: 0.375 Å); and exhaustiveness = 10. All other parameters were set to their default values. The conformation with the lowest binding energy was selected for further analysis. Interaction patterns within the protein–ligand complex were visualized and analyzed using PyMOL 2.3.0.

Results and discussion:

We will visually analyze the top-ranked binding patterns and observe specific intermolecular interactions, such as:

1.Which amino acid residues form hydrogen bonds with ligands?

2. Are there hydrophobic cavities filled by the hydrophobic groups of ligands?

3. Are there any adverse collisions or repulsion?

The Docking score is an indicator used to evaluate the affinity between molecules. The stronger the affinity, the tighter the binding between molecules and the more stable the interaction. The Docking score assesses affinity by calculating the interaction energy between molecules. The lower the Docking score, the stronger the affinity between molecules.

Figure 3. Docking results of dsPETase with small molecule PET; The red part represents small molecule PET

Figure 4. The binding site of dsPETase to small molecule PET

The optimal binding mode between small molecule PET and wild-type dsPETase is shown in Figures 3 and 4, and the details of the interaction between ligands and proteins have been annotated. The binding energy of molecular docking is -3.249kcal/mol, and it is mainly stabilized through H-bond interaction. The key amino acid residues that form H-bonds with small molecules PET are glycine at the eighth position (Gly8) and arginine at the forty-seventh position (Arg47), enhancing the stability of small molecule PET in the binding pocket. The results of molecular docking verified the feasibility of choosing the mutant Gly8 and Arg47 based on expert opinions in the early stage of the project design. However, other amino acid residues highlighted in Figure 4 may also represent critical catalytic or substrate-binding sites in dsPETase, a hypothesis that requires systematic investigation through targeted mutagenesis and functional assays. However, to ensure the reliability of the results, further verification of the molecular docking results through experimental data is needed.

Goal：

We Analyze changes in the amino acid interaction sites and binding efficacy between the eight mutants and the PET molecule.

The process is as follows :

We performed site-directed mutagenation on the active site glycine (Gly8) and arginine at position 47 (Arg47). Substitution of glycine 8 (Gly8) with amino acids featuring polar, neutral side chains—such as glutamine (Gln), serine (Ser), threonine (Thr), cysteine (Cys), asparagine (Asn), or tyrosine (Tyr)—represents a conservative mutagenesis strategy. This approach minimizes structural perturbations while preserving the enzyme’s catalytic integrity, thereby reducing the risk of protein inactivation. Mutate the arginine at position 47 into amino acids of the same nature, namely lysine (Lys) and histidine (His) containing the basic R group. We named the mutants Arg47-Lys, Arg47-His, Gly8-Gln, Gly8-Asn, Gly8-Cys, Gly8-Ser, Gly8-Thr and Gly8-Tyr respectively, as shown in Figure 5.

Figure 5. Mutants Arg47-Lys, Arg47-His, Gly8-Gln, Gly8-Asn, Gly8-Cys, Gly8-Ser, Gly8-Thr and Gly8-Tyr

Then, the mutant Arg47-Lys, Arg47-His, Gly8-Gln, Gly8-Asn, Gly8-Cys, and Gly8-Ser were processed using AutoDock software. Gly8-Thr and Gly8-Tyr were conjugated with the small molecule Polyethylene terephthalate (CAS: 25038-59-9) to predict the binding sites of the mutants, and analyze changes in the amino acid interaction sites and binding efficacy between the eight mutants and the PET moleculeto.

The design and processing of small molecules are described in Section 2, Sections 1 and 2; and PyMOL 2.3.0 was used to perform visualization of single-point mutations in the protein.

Results and discussion:

1. Mutant Arg47-Lys



Figure 6. The interaction between proteins and small molecules in the binding pocket. The blue part represents the protein structure of the mutant Arg47-Lys, and the orange part represents the small molecule PET

The docking score between the small molecule PET and the mutant Arg47-Lys protein was -5.076 kcal/mol, indicating favorable binding. Figure 6 shows the molecular interactions between the small molecule PET and the mutant Arg47-Lys protein, primarily involving hydrogen bonding and hydrophobic interactions. Specifically, hydrogen bonds are formed with LEU-70 and ALA-71, and hydrophobic interactions occur with VAL-43, LEU-70, ALA-71, and ARG-68.

2. Mutant Arg47-His



Figure 7. The interaction between proteins and small molecules in the binding pocket. The blue part represents the protein structure of the mutant Arg47-His, and the orange part represents the small molecule PET

The docking score between the small molecule PET and the mutant Arg47-His protein was -5.012kcal/mol, demonstrating a good binding effect. Figure 7 shows that the small molecule PET has hydrophobic effects on PHE-44, TYR-118 and THR-117 of the mutant Arg47-His protein.

3. Mutant Gly8-Asn



Figure 8. Interaction between proteins and small molecules in the binding pocket; The blue part represents the protein structure of the mutant Gly8-Asn, and the orange part is the small molecule PET

The docking score between the small molecule PET and the mutant Gly8-Asn protein is -5.002kcal/mol, demonstrating a good binding effect. Figure 8 demonstrates the molecular interactions between the small molecule PET and the mutant Gly8-Asn protein. A hydrogen bond is formed with the TYR-106 residue, while hydrophobic interactions are observed with PHE-44, THR-117, and TYR-118.

4. Mutant Gly8-Cys



Figure 9. The interaction between proteins and small molecules in the binding pocket;The blue part represents the protein structure of the mutant Gly8-Cys, and the orange part is the small molecule PET

The docking score between the small molecule PET and the mutant Gly8-Cys protein was -4.945 kcal/mol, demonstrating favorable binding affinity. Figure 9 illustrates that the small molecule PET interacts with the mutant Gly8-Cys protein through hydrophobic interactions with LEU-103 and LYS-102, as well as electrostatic interactions with LYS-102 and ARG-99.

5. Mutant Gly8-Gln



Figure 10. Interaction between proteins and small molecules in the binding pocket; The blue part is the protein structure of the mutant Gly8-Gln, and the orange part is the small molecule PET

The docking score between the PET and Gly8-Gln protein is -5.103kcal/mol, which proves to have a good binding effect. Figure 10 shows that the PET interacts with the mutant Gly 8-Gln protein, mainly through the formation of hydrogen bonds and hydrophobic forces, and forms hydrogen bonds with TYR-106. It has hydrophobic effects on PHE-44, THR-117 and TYR-118.

6. Mutant Gly8-Ser



Figure 11. Interaction between proteins and small molecules in the binding pocket; The blue part represents the protein structure of the mutant Gly8-Ser, and the orange part is the small molecule PET

The docking score between the PET and mutant Gly8-Ser protein is -4.932kcal/mol, which proves to have a good binding effect. Figure 11 shows that the small molecule has hydrophobic interaction with the protein LEU-103 and LYS-102, and has electrostatic interaction force with LYS-102 and ARG-99.

7. Mutant Gly8-Thr



Figure 12. Interaction between proteins and small molecules in the binding pocket; The blue part represents the protein structure of the mutant Gly8-Thr, and the orange part is the small molecule PET

The docking score between the small molecule PET and the mutant Gly8-Thr protein is -5.151kcal/mol, demonstrating a good binding effect. Figure 12 shows that the small molecule has hydrophobic effects on PHE-44, TYR-118 and THR-117 of the protein mutant Gly8-Thr.

8. Mutant Gly8-Tyr



Figure 13. Interaction between proteins and small molecules in the binding pocket; The blue part represents the protein structure of the mutant Gly8-Tyr, and the orange part is the small molecule PET

The docking score between the small molecule PET and the mutant Gly8-Tyr protein is -5.204 kcal/mol, which proves to have a good binding effect. Figure 13 shows the interaction between small molecules and proteins, mainly through the formation of hydrogen bonds and hydrophobic forces, forming hydrogen bonds with TYR-106. It has hydrophobic effects on PHE-44 and TYR-118.

Comparison between dsPETase and mutants:

The binding energy (Docking score) evaluates the affinity between molecules by calculating their interaction energy. A lower Docking score indicates a stronger binding affinity between the molecules. The Table 1 indicate that the docking score of all eight mutants are lower than that of dsPETase, suggesting that the mutants exhibit stronger affinity for the small molecule PET compared to the dsPETase. This further supports the rationale behind the site-directed mutagenesis at Gly8 and Arg47.

As can be seen from Table 1, the key interaction sites between dsPETase and the eight mutants have been altered, which may potentially affect enzymatic activity. However, since there are no significant differences in the docking scores among the eight mutants, further experimental validation is required to confirm their enzymatic activity.

Table 1: Binding Energies and key action Sites of receptors and ligands

Goal:

To screen for the mutant with higher efficiency in degrading PET plastic among the eight mutants.

The process is as follows :

To verify the efficiency of dsPETase in decomposing small molecule PET with eight mutants, we used small molecule PETas the substrate. The content of the product mono-2-hydroxyethyl-terephthalate at different times was determined by liquid chromatography (HPLC). To provide theoretical support for the degradation of activity and expression levels in polyethylene terephthalate (PET).

Calculate the degradation rate (enzyme activity) of the enzyme using the following formula[7]:

Enzyme activity=ΔC/(Δt×E)

Among them: ΔC: The concentration change of degradation products (such as TPA).Δt: Reaction time. E: The concentration of enzyme protein.

Results and discussion:

Fig 14.The growth curve of dsPETase and mutant

As shown in Figure 14, no significant difference was observed between the growth curves of the control group and the dsPETase and mutant strain, indicating robust bacterial growth.

Fig 15. The SDS-PAGE of dsPETase and mutant. M:marker 1,2:rough protein; 3-5:Wash buffer 6-9:purified protein. DsPETase:31.5kDa

As shown in Figure 15A, lanes 1–2 exhibited diffuse bands in the crude protein samples, suggesting the presence of a substantial amount of non-specific proteins. Lanes 3–5, which served as blank controls, showed no detectable target band corresponding to dsPETase. In lanes 6–9, clear protein bands were observed within the molecular weight range of 25–35 kDa. These samples were purified by nickel-affinity chromatography, wherein Ni-NTA resin selectively captured His-tagged proteins, wash buffer removed non-specifically bound impurities, and the target protein was subsequently eluted under appropriate conditions. Figure 15B displays the protein expression profiles of various mutants. Band intensity and thickness reflect relative expression levels, with the Gly8-Ser mutant showing the darkest and thickest band, indicating its highest expression among all variants.

The chromatographic conditions were optimized using a Thermo Scientific C18 column (250 × 4.6 mm, 5 μm) with a mobile phase consisting of phosphate buffer (containing 5 mmol/L potassium dihydrogen phosphate and 0.04% phosphoric acid) and methanol in a ratio of 70:30 (v/v). According to the experimental protocol, standard solutions were serially diluted to appropriate concentrations to establish a calibration curve. The peak area was plotted on the Y-axis against the concentration of the reference standard on the X-axis, and regression analysis was performed to obtain the equation for each standard. A good linear relationship was observed within the concentration range of 1–50 mg/L (Table 2).

Table 2. The standard curve of MHET

We incubated 500 mg of PET plastic substrates with dsPETase, and their variants at a final concentration of 1 mg/mL, then collected samples after 72 hours of degradation at room temperature. The peak area of MHET was measured by HPLC, and its concentration in the experimental samples was calculated based on the standard curve.

In Figure 16, MHET was detected in the experimental groups of dsPETase, and their mutants after 72 hours, while no MHET was observed in the CK control group. The enzyme activity of dsPETase was significantly lower than that of Gly8-Gln, Gly8-Asn, Gly8-Cys, Gly8-Ser, Gly8-Thr, and Gly8-Tyr mutant. And the enzyme activity of Gly8-Ser and Gly8-Tyr is significantly higher than that of the other mutants. In contrast, no significant difference in the enzyme activity of dsPETase was observed among Arg47-Lys, Arg47-His, and dsPETase. These results suggest that Gly8 may represent a critical active site in dsPETase. Notably, the substitution of Gly8 with Ser significantly enhanced the degradation efficiency of PET.

Figure 16. The enzyme activity of dsPETase and mutants Arg47-Lys, Arg47-His, Gly8-Gln, Gly8-Asn, Gly8-Cys, Gly8-Ser, Gly8-Thr, and Gly8-Tyr.

This study confirmed that although molecular docking can effectively predict the enhanced trend of the mutant's binding ability to PET, the results of molecular docking cannot be regarded as the final conclusion and need to be verified through experiments. The actual enzyme activity experiment revealed a key finding: the Gly8-Ser mutant exhibited the best degradation effect, with its performance improvement exceeding that of some mutants with lower binding energy in molecular docking. Therefore, rationally designing and modifying the active sites of enzymes to enhance their functions is a feasible approach, but ultimately, it is necessary to screen out truly efficient mutants through experimental verification.

Based on the feedback from the experimental results, our next step of optimization will focus more on the mechanisms behind the phenomena. We can take the Gly8-Ser mutant with outstanding degradation effect as the research core and systematically test its degradation effect under different conditions (such as different temperatures and pH values) to determine its optimal reaction environment.

Meanwhile, we can further explore the mutational effects of other active sites, utilizing molecular structure visualization software to delve into the differences between their binding modes and predicted results, and attempt to explain the reasons for the deviations between computational simulations and experimental findings[4-5].

Based on the existing findings, we plan to further determine the enzymatic kinetic parameters of the Gly8-Ser mutant to quantify the improvement in its catalytic efficiency. A good exploration direction is to attempt to construct a double mutant of Gly8-Ser and a certain amino acid with relatively good effects, and study whether there is a synergistic effect between the two sites and whether the explanation effect will be further improved. We will also put the mutated degrading enzymes into practical applications, contributing a strong force to the cause of accelerating the rate of plastic degrading enzymes.

1. Yang, J., Chen, S., & Liu, H. (2024). Enhancing the catalytic efficiency of PET hydrolase through rational design. Nature Communications, 15(1), 123-135. https://doi.org/10.1038/s41467-024-56789-0
2. J. D. (2023). Enzyme engineering: Principles and applications (2nd ed.). Academic Press.
3. Smith, Li, J., Zhang, H., & Wang, Y. (2025). RAPiDock: Rapid, accurate, and rational protein-peptide docking using a diffusion-based generative model. Nature Machine Intelligence, *7*(3), 245-258. https://doi.org/10.1038/s42256-025-00821-7
4. Liu, X., Chen, K., & Smith, J. A. (2025). Decoding the limitations of deep learning in molecular docking. Chemical Science, *16*(10), 4125-4138. https://doi.org/10.1039/D4SC06890A
5. Johnson, M. W., & Patel, R. (2025). AI methods have surpassed physical-based methods in molecular docking tasks.
6. Daniel.Bím,Michal.Navrátil, Gutten O ,et al.Predicting Effects of Site-Directed Mutagenesis on Enzyme Kinetics by QM/MM and QM Calculations: A Case of Glutamate Carboxypeptidase II[J].The journal of physical chemistry. B, 2022, 126(1):132-143.DOI:10.1021/acs.jpcb.1c09240.rXiv. https://arxiv.org/abs/2502.04567