Abstract
More than 10 % of global solid waste consists of poly(ethyleneterephthalate) (PET). Among other techniques, PET hydrolases (PETases) can be used to depolymerize this plastic. However, wildtype PETases exhibit poor specific activities and insufficient thermostability, limiting their use in depolymerization processes which require high temperatures. In 2022, machine learning-aided enzyme engineering of a PETase stemming from the bacterium Ideonella sakaiensis (IsPETase) resulted in a more functional, active, stable, and tolerant variant (FAST-PETase). To rationalize the molecular basis of FAST-PETase's improved thermal stability, we performed comparative Constraint Network Analysis (CNAnalysis) and Molecular Dynamics (MD) simulations of wildtype IsPETase (WT-PETase) and FAST-PETase at 30°C and 50°C identifying thermolabile sequence stretches in the wildtype enzyme. Further analysis of the backbone flexibility revealed that all mutations of FAST-PETase affected these critical regions. Counterintuitively, the in-silico analyses additionally highlighted that the flexibility of these regions decreased at 50°C in FAST-PETase, instead of exhibiting increased flexibility at higher temperature as would be expected from thermodynamic considerations. This effect was confirmed by physical energy calculations, which suggest that temperature-dependent conformational changes of FAST-PETase decrease the free energy of unfolding (ΔG(stability)) and rigidify the enzyme at elevated temperatures enhancing stability. Looking forward, these findings might help guide the rational engineering of protein thermostability and contribute to our understanding of the thermal adaptation of thermophilic enzymes.