Increasing concern over the environmental persistence and toxicity of per- and polyfluoroalkyl substances (PFAS), often referred to as “forever chemicals”, has necessitated technologies for their remediation [1]. Long-chain PFAS often degrade into smaller chain ones, such as tri-fluoroacetate (TFA), which extensively accumulate in the environment. No enzymes have yet been discovered to achieve the full degradation of short-chain PFAS, one reason that could explain their extraordinary persistence. It is therefore urgent to design a biomimetic system offering a cost-effective and sustainable solution to mitigate the environmental disaster of PFAS pollution.

The only known enzyme capable of breaking a carbon-fluorine bond is fluoroacetate dehalogenase (FAcD), achieving the conversion of mono-fluoroacetate (MFA) into glycolate [2]. Yet, the enzyme fails at cleaving the C-F bonds of TFA, an analogue molecule to MFA where two hydrogens are replaced by fluorines. To shift the degradation ability of FAcD towards polyfluorinated compounds, the first step was to better understand the reason for its current inactivity. To this end, classical molecular dynamics (MD), binding affinity MM/PB(GB)SA, and hybrid quantum mechanics/molecular mechanics (QM/MM) techniques were employed. Classical thermodynamic integration results showed that there was close to no thermodynamic barrier for MFA and TFA to enter the binding pocket of FAcD. However, both substrates appeared to exhibit high conformational sampling and instability within the binding pocket, which seemed to hinder enzyme efficiency. Further, the QM/MM study of the reaction step with MFA and TFA highlighted key residues involved in the difference in reactivity between the two substrates.

A genetic algorithm (GA) model, based on random mutations from a library of amino acid conformers, was then used to optimize the wild-type FAcD [3]. Based on our previous findings, the conformational flexibility of MFA and TFA in the binding pocket should be reduced to enhance reactivity. Therefore, multi-objective optimization was employed to optimize both the stability (force field-based descriptor) and binding (MM/PBSA calculation) of the enzyme. Best-performing mutants will be selected at the pareto front, tested with QM/MM and validated experimentally. The aim is also to now incorporate a reactivity descriptor in the multi-objective optimization, to lower the current activation barrier for C-F bond cleavage with TFA. Efforts to develop a fast yet accurate reactivity descriptor for the GA optimization will be discussed.

[1]  E.M. Suderland et al.. J Expo Sci Environ Epidemiol, 2029, 29, 131-147.

[2]  S. Farajollahi et al., ACS Omega, 2024, 26, 28546-28555.

[3]  N. Browning et al., ChemRxiv, 2023.