Engineering is a critical process in synthetic biology research. An engineering cycle consists of four stages: design, build, test and learn. In our project, we continuously executed eight cycles. The accumulation of engineering results eventually led us to achieve the final goal of this project.
Our project aims to enhance the binding and degradation capabilities of PETase toward PET microplastics by the fusion of PET-binding peptides.
In the case of insufficient binding ability of PETase towards PET, improving the binding efficiency becomes a key goal. Given the complexity, inefficiency, and time-consuming nature of traditional experimental screening processes, we introduced deep learning models to accelerate PET-binding peptide screening.
We constructed two models: LSTM and GCN. The LSTM model was used for the preliminary screening of peptides, focusing on capturing long-range dependencies within the sequence. Next, the GCN model was applied to analyze the three-dimensional structural information of the peptides and to build node features, thereby improving the accuracy of screening and the reliability of predicting biological binding capabilities.
We trained and tested the LSTM and GCN models, evaluating their ability to screen peptides, respectively. Preliminary experiments were conducted to validate the effectiveness of the models in predicting the binding affinity between peptides and PET plastic. Based on the experimental results, we continuously adjusted the model parameters to improve the recognition rate of target peptides.
Based on the evaluation of model performance, the screening accuracy of the LSTM model gradually increased to 93.4%, while the accuracy of the GCN model reached 90.4%. The GCN model successfully identified 16 target peptides with potential high PET-binding affinity. Based on the obtained peptide data, we further optimized the screening accuracy and enhanced the performance of the deep learning models through dataset expansion and model structure optimization. And we conducted experimental validation using the predicted peptides.
Since the attachment of short peptides to the amino and carboxyl termini can lead to differences in their position and functional characteristics within the peptide chain or protein, as well as potentially affecting the overall properties of the protein, we first selected PET-binding peptides 50-1, 50-3, 50-4 and 50-5 obtained from deep learning screening, as examples, to fuse with the N-terminus and C-terminus of PETase, respectively. We constructed fusion proteins for validation and characterization.
We fused the aforementioned short peptides to the N-terminus or C-terminus of PETase using the linker G4S, and their encoding genes were linked to pET-21b to construct recombinant plasmids, respectively. The correctly verified recombinant plasmids were then transformed into Escherichia coli BL21 (DE3) for shake-flask fermentation to obtain the fusion proteins.
We used pNPB as the substrate to test the catalytic activity of the fusion protein under conditions of 60°C and pH 8.0.
Fig. 1 Enzymatic activity of fusion proteins with peptides 50-1 to 50-5
(a). Enzymatic activity of C-terminal fusion proteins with peptides 50-1 to 50-5. (b). Enzymatic activity of N-terminal fusion proteins with peptides 50-1 to 50-5.
We compared the enzymatic activity of fusion proteins with the same peptide linked to the N-terminus and C-terminus of PETase and found that there are differences depending on the position of the peptide attachment. Therefore, attaching the peptides to different ends of PETase does indeed affect its degradation efficiency. However, we still face the problem of low fusion protein expression.
Since the mined peptides are derived from non-E. coli hosts and possess strong hydrophobicity, they may have a significant negative effect on the expression of the fusion protein. We took 50-4-N-G4S-PETase (BBa_K5157051) as an example to optimize the types of culture medium, inducer concentration, fermentation temperature, and fermentation time.
We conducted optimization of fermentation conditions at the shake-flask level: we selected TB, LB, and ZYM as culture media to explore the most suitable medium type; we tested IPTG final concentrations of 0.025 mmol/L, 0.050 mmol/L, 0.075 mmol/L, and 0.100 mmol/L to find the optimal inducer concentration; we conducted fermentation at 16°C and 25°C to determine the most suitable fermentation temperature; and we tested fermentation durations of 24 h and 48 h to identify the optimal fermentation time.
We used pNPB as the substrate to test the catalytic activity of the fusion protein obtained from fermentation under conditions of 60°C and pH 8.0.
Fig. 2 Optimization of fermentation conditions for 50-4-N-G4S-PETase
(a). Enzymatic activity of 50-4-N-G4S-PETase in different culture media. (b). Enzymatic activity of 50-4-N-G4S-PETase at various inducer concentrations. (c). Enzymatic activity of 50-4-N-G4S-PETase at different fermentation temperatures. (d). Enzymatic activity of 50-4-N-G4S-PETase at different fermentation times.
The experimental results indicate that using TB medium with an IPTG final concentration of 0.05 mmol/L, fermentation at 16°C for 48 h is most beneficial for the expression of the fusion protein. Therefore, we used this shake flask fermentation conditions in subsequent experiments.
We designed to use G4S to connect 40-1, 40-2, 40-3, 40-4, 50-1, 50-3, 50-4, 50-5, 60-1, 60-2, 60-3, 60-4, 70-1, 70-2, 70-3, 70-4 to the C-terminus and N-terminus of PETase, respectively, to construct a fusion protein prokaryotic expression system.
We linked the fusion protein gene to pET-21b to construct recombinant plasmids. The correctly verified recombinant plasmids (BBa_K5157051~BBa_K5157069) were then transformed into E. coli BL21 (DE3) for shake-flask fermentation to obtain the fusion proteins.
We used pNPB as the substrate to test the catalytic activity of the fusion protein under conditions of 60°C and pH 8.0, while also assessing the degradation capability of the purified fusion protein on PET microplastics.
Fig. 3 Identification of PET-binding peptides
(a). Enzymatic activity of C-terminal PET-binding peptides. (b). Enzymatic activity of N-terminal PET-binding peptides. (c). Degradation of PET microplastics by C-terminal PET-binding peptides. (d). Degradation of PET microplastics by N-terminal PET-binding peptides.
The experimental results indicated that 19 fusion proteins were successfully expressed and 50-4-N-G4S-PETase (BBa_K5157051), PETase-G4S-50-5-C (BBa_K5157056), and 60-2-N-G4S-PETase (BBa_K5157067) exhibited both excellent expression and PET degradation capabilities. We confirmed the feasibility of peptides in enhancing microplastic degradation, but the binding capacity still requires further investigation.
To further verify whether the obtained PET-binding peptides can effectively bind to PET microplastics, we designed fusion proteins by linking these aforementioned peptides with good degradation effects, 50-4, 50-5, and 60-2, to enhanced green fluorescent protein (eGFP) through the linker G4S.
We linked the fusion protein gene to pET-20b to construct
recombinant plasmids (BBa_K5157070~BBa_K5157072).
The
correctly verified recombinant plasmids were then transformed into
E. coli BL21 (DE3) for
shake-flask fermentation to obtain the fusion proteins.
The eGFP solution and eGFP fusion protein solution with equal fluorescence intensity were taken and incubated with PET thin film, protected from light at 37°C for 24 h. Then, free proteins were washed and eluted, the PET film was observed under a fluorescence microscope, and the remaining fluorescence adsorption intensity of PET films was measured at 488 nm excitation wave.
Fig. 4 Fluorescence microscopy results
of
50-4-N-G4S-eGFP, 50-5-C -G4S-eGFP,
60-2-N-G4S-eGFP, and their controls
(a). eGFP as a control for 50-4-N-G4S-eGFP. (b). eGFP as a control for 50-5-C-G4S-eGFP. (c). eGFP as a control for 60-2-N-G4S-eGFP. (d). Binding of 50-4-N-G4S-eGFP to PET film. (e). Binding of 50-5-C-G4S-eGFP to PET film. (f). Binding of 60-2-N-G4S-eGFP to PET film.
The experimental results confirmed that these PET-binding peptides could bind to PET films, suggesting that the PET affinity of these binding peptides led to the improved degradation efficiency of PETase-PET-binding peptide.
As a segment connecting the enzyme and the peptide, the choice and composition of the linker can influence the property of the fusion protein. Therefore, we took 70-1-C as an example to attempt replacing the linker G4S with other linkers to optimize the spatial conformation between the enzyme and the peptide, reduce the mutual interference between the two protein components, and enhance the substrate catalytic efficiency of the fusion protein.
We replaced the gene of linker of the fusion protein in pET-21b to construct recombinant plasmids (BBa_K5157073~BBa_K5157077), respectively. The correctly verified recombinant plasmids were then transformed into E. coli BL21 (DE3) for shake-flask fermentation to obtain the fusion proteins, respectively.
We used pNPB as the substrate to test the catalytic activity of the fusion protein under conditions of 60°C and pH 8.0.
Fig. 5 Enzymatic activity of 70-1-C fusion proteins connected by different linkers
The experimental results indicate that the flexible linker G4S and the rigid linker SLE performed better. In subsequent experiments, we needed to explore the effect of rigid linker SLE in fusion proteins.
Flexible linkers offer greater flexibility, while rigid linkers emphasize structural stability and precision. From the previous screening cycle, we obtained a well-performing rigid linker SLE, and we attempted to replace the flexible linker G4S in all fusion proteins with the rigid linker SLE.
We replaced the gene of original linkers of the fusion proteins with SLE encoding gene to pET-21b to construct recombinant plasmids (BBa_K5157077~BBa_K5157095), respectively. The correctly verified recombinant plasmids were then transformed into E. coli BL21 (DE3) for shake-flask fermentation to obtain the fusion proteins, respectively.
We used pNPB as the substrate to test the catalytic activity of the fusion protein under conditions of 60°C and pH 8.0, while also assessing the degradation capability of the purified fusion protein on PET microplastics.
Fig. 6 Replacing the linker with SLE
(a). Enzymatic activity of C-terminal PET-binding peptides. (b). Enzymatic activity of N-terminal PET-binding peptides. (c). Degradation of PET microplastics by C-terminal PET-binding peptides. (d). Degradation of PET microplastics by N-terminal PET-binding peptides.
The experimental results indicate that both the enzymatic activity and degradation efficiency of 60-2-N-SLE-PETase (BBa_K5157093) were excellent. After the above cycles, we had obtained some fusion proteins with promising results, but they still fell short of expectations.
We obtained peptides 50-4-N, 50-5-C, and 60-2-N that effectively enhanced the PETase degradation efficiency towards PET microplastics; however, there is still room for improvement in the expression levels and binding ability to PET microplastic of the fusion proteins. Therefore, using molecular design and structural biology tools, we further modified the sites of these PET-binding peptides. To enhance the precision and efficiency of the mutations, we introduced Mutcompute-super model, which can predict the impact of different mutations on protein stability and microplastic binding capability based on the three-dimensional structure of the protein. Through this tool, we optimized the protein folding conformation, thereby improving the expression levels and microplastic substrate binding capability of the fusion proteins.
Based on the optimal mutation sites predicted by the 3D deconvolution model, we designed specific primers and introduced site-directed mutations on the existing recombinant plasmids to construct recombinant plasmids (BBa_K5157096~BBa_K5157114) containing the mutated fusion proteins. The correctly verified recombinant plasmids were then transformed into E. coli BL21 (DE3) for shake-flask fermentation to obtain the fusion proteins.
We used pNPB as the substrate to test the catalytic activity of the fusion protein under conditions of 60°C and pH 8.0. At the same time, we utilized experiments to verify the effects of the predicted mutations from the 3D deconvolution model and assessed the degradation capability of the purified fusion protein on PET microplastics.
Fig. 7 The expression and PET degradation results of fusion protein mutants
(a). Enzymatic activity of the mutants 50-4-N-G4S-PETase. (b). Enzymatic activity of the mutants PETase-G4S-50-5-C. (c). Enzymatic activity of the mutants 60-2-N-SLE-PETase. (d). Degradation of PET microplastics by the mutants 50-4-N-G4S-PETase. (e). Degradation of PET microplastics by the mutants PETase-G4S-50-5-C. (f). Degradation of PET microplastics by the mutants 60-2-N-SLE-PETase.
The experimental results indicate that the mutants V19I (BBa_K5157098), L23K, and S61A of 50-4-N-G4S-PETase exhibit good PET microplastic degradation effects, while the mutants R287V (BBa_K5157103), R301V, and A325S of PETase-G4S-50-5-C also show favorable PET microplastic degradation. Additionally, the mutants A13S, N29D, and L60I (BBa_K5157114) of 60-2-N-SLE-PETase demonstrate effective PET microplastic degradation.
To Top
© 2024 - Content on this site is licensed under a Creative Commons Attribution 4.0 International License .
The repository used to create this website is available at gitlab.igem.org/2024/jiangnan-China .