Predicting Hydrolase Mechanism for Sustainable Polyester Degradation

Enzyme catalyzed degradation is a promising approach to mitigating plastic pollution. However, finding suitable enzyme catalysts remains a challenge. Aside from sufficient substrate compatibility, one key consideration is enzyme mechanism, which the environmental impact depends on. An enzyme performing random scission leads to incomplete degradation from crystallinity induced recalcitrance as well as faster microplastic generation, whereas chain end scission results in degradation to naturally decomposable small molecules. Furthermore, when enzymes are nanoscopically embedded within the plastic matrix, a processive mechanism generates a chain slide motion that can lead to the degradation of amorphous and crystalline regions alike. As such, finding enzymes with chain-end, processive mechanism for a given polymer system can be the key to unlocking complete polymer degradation. Here, we investigated the catalytic activity and mechanism of 10 hydrolases on polycaprolactone (PCL). Surface and enzyme-embedded PCL degradation studies and degradation product analysis yielded 4 enzymes with no activity towards PCL, 3 with random PCL scission, and 3 with chain-end PCL scission. Through binding pocket analysis, each hydrolase was labeled with key geometry, chemistry, and flexibility features, hypothesized to influence both degradation capability and mechanism. Molecular modeling was done in parallel to predict the potential for processive chain slide of polyesters in enzyme binding pockets. We found binding pocket polarity, binding pocket secondary structure composition, and evidence of chain-slide potential from molecular modeling to be highly correlated to experimentally determined degradation mechanism. A computational pipeline was developed to automate binding pocket analysis and molecular docking, and a database of hydrolases with reported PCL compatibility was screened <i>in silico</i> for degradation mechanism. The model and key features highlighted in this work are generalizable to other commercial polyesters. In addition, this model can also guide future works with applications in rapid screening, enzyme engineering, and general applicability to binding pocket analysis beyond plastic degradation.