Polyethylene terephthalate (PET), which is used in drinking bottles, fibers, and many other applications, is one of a few plastics that can be broken down to its constituent monomers by naturally occurring enzymes. This study developed a landscape profiling method to identify and characterize the potential of microbial enzymes to degrade these plastics. Two enzymes were engineered with sequential mutagenesis and exhibited excellent performance relative to benchmarks, especially under the harsh conditions that are ideal for use in recycling applications..
Research Background and Objectives
PET (polyethylene terephthalate) is a representative general-purpose plastic widely used in various fields such as PET bottles, clothing, seat belts, takeout cups, and car mats. While most PET waste is separately collected and mechanically recycled into intermediate products, the recycled materials often degrade in quality, ultimately leading to incineration or landfill disposal. As a method to address this issue, chemical recycling technology has been developed to break down the PET polymer bonds using chemical catalysts and return them to the original raw materials. However, it has not been a perfect alternative due to the limitations of applying the method, which is caused by high temperature and high-pressure conditions. Therefore, the scientific community has turned to biological/biocatalytic recycling to solve these problems through enzymes. With complex bonding structures, enzymes react selectively with PET at low temperatures and in water solvent conditions to produce pure reactants. Thus, they are excellent at converting contaminated raw materials. There has been a fierce competition worldwide to develop PET-degrading enzymes using advanced technologies in various fields such as synthetic biology, computational chemistry, and AI-driven protein design.
Research Approach
The research team attempted to experimentally determine the fitness landscape of various enzyme protein sequences. Since conducting experiments on all sequences was physically impossible, it was necessary to use a statistical sampling method through a landscape. To construct a landscape of the Polyesterase-Lipase-Cutinase Family, a neighborhood analysis module was devised to control the network’s rigidity using distance histogram data for each protein sequence. This analysis generated a two-dimensional semantic network. Based on this semantic network, the research team proposed an innovative approach to experimentally measure the fitness for PET degradation activity and thermal stability using hierarchization and cluster sampling. Also, to improve the selected enzymes, the team attempted a unique strategy of applying cross-template engineering to reflect natural diversity and fitness information in a rational design based on the protein’s 3D structural information.
Results and Discussion
The new approach identified the most promising enzymes, Mipa-P and Kubu-P, among 158 nodes, which showed a superior PET-degradation rate and durability compared to other benchmarks. Cross-template engineering created heat-resistant variants MipaM19 (Mipa-PM19) and KubuM12 (Kubu-PM12) with melting temperatures exceeding 92 and 99°C, respectively. Surprisingly, Kubu-M12 withstood the condition of a minimum enzyme dosage of 0.58 g/mg and high PET loading of 20% and 30%, degrading more than 90% of the PET substrate within 8 hours. It showed overwhelming performance compared to other engineered benchmark enzymes. Moreover, Kubu-M12 withstood 99% ethylene glycol solvent and produced 30 mM level bis(2-hydroxyethyl) and terephthalic acid. For the first time in the world, the enzymatic catalytic glycolysis reaction was achieved at a significant level.
Read more on PAL website