REPROGRAMMING ENZYMES FOR THE FUTURE: PROTEIN ENGINEERING FOR MICROPLASTIC DEGRADATION

The inadequate management of plastic waste has become a major environmental challenge, contributing to the widespread occurrence of microplastics and their adverse effects on diverse organisms, including humans. In this context, structure–function studies and molecular modeling have contributed to the identification and characterization of enzymes capable of degrading synthetic polymers, the primary constituents of most microplastics. Among these enzymes, the polyethylene terephthalate (PET) hydrolase from Ideonella sakaiensis (IsPETase) has attracted considerable attention due to its ability to degrade PET. Building upon these advances, protein engineering enables the development of enzyme variants with enhanced stability, catalytic efficiency, and adaptability to different environmental conditions, thereby expanding their potential for biotechnological applications aimed at plastic biodegradation. This study aims to identify and optimize enzymes with potential for microplastic degradation through protein engineering strategies, thereby expanding their biotechnological applications for mitigating the environmental impacts of plastic pollution. Molecular information for IsPETase was retrieved from the UniProt database and used as a reference for sequence searches in metagenomic datasets available in the MGnify and WGS databases. Polymer structures were built using MarvinSketch, and three-dimensional structural analyses were performed in PyMOL. Protein engineering was conducted using ProtGPT2, with loss, perplexity and sequence identity values applied as filtering criteria. Protein structures were protonated using PDB2PQR and Open Babel. Molecular docking simulations were carried out with GOLD employing the ChemPLP, ChemScore, GoldScore, and ASP scoring functions. Enzyme–ligand interactions were subsequently analyzed and visualized using LigPlot and PLIP. A total of 2,399 sequences from database MGnify and 73 sequences from database WGS were initially identified. Following filtering steps based on sequence completeness, identity, coverage, RMSD, and source organism, six sequences were selected for subsequent analyses. Protein engineering approaches generated 54 variants, of which nine met the length and sequence identity criteria (≥50%) relative to the reference enzymes. The selected sequences exhibited variability in loss, perplexity, and sequence identity values, indicating different degrees of similarity to naturally occurring proteins. Subsequent analyses will focus on molecular docking simulations of the engineered proteins and the characterization of interactions between active-site residues and polymer substrates.