Deep Learning-driven synergistic engineering of PET hydrolase for post-consumer PET depolymerization | JoVE Visualize

Area of Science:

Biotechnology and Synthetic Biology
Computational Chemistry and Molecular Modeling
Environmental Science and Engineering

Background:

Plastic pollution, particularly from polyethylene terephthalate (PET), is a severe global environmental issue.
Enzymatic depolymerization offers a sustainable solution, but current PET hydrolases lack sufficient efficiency and stability for industrial application.
Artificial intelligence (AI) presents new avenues for accelerating enzyme engineering to overcome these limitations.

Purpose of the Study:

To systematically computationally redesign the PET hydrolase NI using a deep-learning framework (EITLEM-Kinetics).
To simultaneously optimize catalytic activity and thermostability by integrating mutation free-energy constraints with kinetic parameter prediction.
To develop a highly efficient and stable PET hydrolase for processing post-consumer PET waste.

Main Methods:

Employed the deep-learning framework EITLEM-Kinetics for computational redesign of PET hydrolase NI.
Integrated mutation free-energy calculations with kinetic parameter predictions to guide enzyme engineering.
Experimentally validated identified beneficial mutations and performed combinatorial iteration to optimize the enzyme variant.

Main Results:

Identified nine beneficial substitution sites, overcoming the activity-stability trade-off.
Developed an optimal variant, NI-KYRF, showing an 80-90% increase in depolymerization activity towards PET films and powders, and a 2.61°C increase in melting temperature.
NI-KYRF achieved a specific activity of 665 μmol$_{TPAeq}$ h$^{-1}$ mg$_{enzyme}^{-1}$, significantly outperforming existing PET hydrolases, and demonstrated high conversion rates (95.5% for pc-PET powder, 80.1% for untreated film) within 24 hours.

Conclusions:

The deep learning-guided computational design framework (EITLEM-Kinetics) is effective for engineering high-performance PET hydrolases.
The engineered NI-KYRF variant shows enhanced depolymerization efficiency and expanded substrate scope, including PBAT and PBT.
This approach holds broad potential for advancing enzymatic recycling of plastics and mitigating environmental pollution.