The reaction mechanism of the *Ideonella sakaiensis* PETase enzyme

The study addresses how Ideonella sakaiensis PETase catalyzes PET hydrolysis at the atomistic level and resolves open mechanistic questions in serine hydrolases. PETase initiates PET depolymerization in a two-enzyme system, producing BHET and MHET; MHET is further processed by MHETase. PETase contains a canonical Ser-His-Asp catalytic triad, suggesting a two-step acylation–deacylation mechanism. However, conflicting hypotheses from crystallography, docking, and MD concern substrate positioning relative to Trp185 (edge-to-face vs parallel-displaced π-π interactions), and the histidine-mediated proton transfer mechanism (flipping vs moving histidine, and potential double-proton transfer/charge-relay with Asp). Prior QM/MM studies report inconsistent free energy barriers, rate-limiting steps, and whether tetrahedral geometries are transition states or metastable intermediates. This work aims to elucidate the unbiased reaction coordinates, energetics, and kinetics of both acylation and deacylation, clarify the proton transfer route via His and its interaction with Asp, define the role of the oxyanion hole, and determine how Trp185 flexibility contributes to catalysis.

Previous structural and computational studies proposed differing PETase mechanisms. Crystal structures and flexible docking suggested multiple Trp185 conformations and π-π interaction modes with PET (parallel-displaced vs edge-to-face), with MD revealing high Trp185 flexibility. Guo et al. hypothesized an edge-to-face interaction during catalysis shifting to parallel-displaced upon product release. In serine hydrolases, NMR for subtilisin suggested a flipping histidine mechanism, whereas other work supports a moving histidine that laterally shuttles a proton via Nε2. A double-proton transfer charge-relay between Asp and His was proposed but remains contested. Prior QM/MM and ONIOM studies on PETase yielded conflicting barriers and mechanistic features (concerted vs stepwise, existence of metastable tetrahedral intermediates) and often assumed reaction coordinates a priori. Adaptive string methods improved pathway identification but lacked verified reaction coordinates. Collectively, these studies motivated an unbiased approach that discovers and validates reaction coordinates to resolve these controversies.

The authors performed QM/MM molecular dynamics with DFTB3 (3ob-3-1) to study acylation and deacylation using transition path sampling (TPS) with flexible-length aimless shooting (ATESA) and inertial likelihood maximization (iLMax) to extract optimal reaction coordinates (RCs). System setup: starting from a Michaelis complex of PETase bound to a hydroxyethyl-capped PET dimer (from prior docking). Protein modeled with CHARMM36 and ligand with CGenFF; solvated TIP3 waters, 0.10 M NaCl, 310 K, 1 atm. QM regions: acylation included the PET dimer; side chains of Ser160, His237, Asp206, Trp159, Ser238 (104 QM atoms; net charge −1); deacylation included the acylated PET fragment, same catalytic side chains, Trp159, Ser238, plus an active-site water (75 QM atoms). Following equilibration, putative TSs were identified using ATESA find_ts by restraining forming/breaking bonds. Aimless shooting generated unbiased trajectory ensembles: acylation 13,750 trajectories (~40.5% reactive); deacylation 19,000 (~39.6% reactive). iLMax screened CVs (acylation: 299 CVs; deacylation: 182 CVs) including distances, angles, dihedrals, and distance differences, and identified linear RCs maximizing transmission. RC validation used committor probability histograms (PB) centered at 0.5 and reactive flux correlation functions to obtain transmission coefficients. Free energy profiles (potentials of mean force) along discovered RCs were obtained by umbrella sampling (windows across RC values; MBAR analysis). Rate constants were computed via the Eyring equation using ΔG‡ from PMFs, κ from reactive flux/committor analyses, at 310 K. Additional analyses included umbrella sampling of Trp185 χ1 dihedral in relevant states, monitoring His237–Asp206 hydrogen bonding, proton localization on His nitrogens, oxyanion hole H-bond distances (Tyr87, Met161), Mulliken charges at key states, aromatic interaction geometries (Trp185/Tyr87 with PET), and hybridization at the scissile carbon via improper dihedrals. Classical MD of deacylation product states probed MHET release and Trp185 dihedral behavior.

• Reaction coordinates and kinetics:
  • Acylation RC (linear combination of three CVs capturing nucleophilic attack, scissile ester cleavage, and proton transfer geometry) passed committor tests (PB ~ 0.5) and showed κ = 1.00. Free energy barrier ΔG‡ = 16.1 ± 0.6 kcal/mol. Eyring rate constant k = 28.8 ± 17.9 s−1 at 310 K.
  • Deacylation RC similarly validated (PB ~ 0.5), κ = 0.89, ΔG‡ ≈ 18.2 kcal/mol (from Discussion), with k = 0.82 ± 0.10 s−1 at 310 K. Thus, deacylation is rate-limiting by ~35-fold.
  • Computed overall rate for the rate-limiting step agrees with experimental kcat values reported by Erickson et al. (1.5 ± 0.5 s−1 for amorphous PET film; 0.8 ± 0.0 s−1 for crystalline PET powder at 30 °C).
• Mechanism:
  • Both acylation and deacylation proceed via concerted mechanisms with a single transition state featuring a tetrahedral geometry at the scissile carbon; no metastable tetrahedral intermediate was observed.
  • Proton transfer occurs via a moving histidine mechanism: His237 shuttles a proton via Nε2 from nucleophile to leaving group without ring flipping.
  • No evidence for a double-proton transfer (charge-relay) between Asp206 and His237: His237 Nδ1 maintains its proton, and the Asp206–His237 H-bond persists throughout both steps.
• Active-site interactions:
  • Oxyanion hole (Tyr87, Met161 backbones) stabilizes negative charge on the carbonyl oxygen: Tyr87 H-bonds throughout; Met161 engages upon nucleophilic attack and remains until product state of deacylation.
  • Charge analysis: carbonyl oxygen charge ~ −0.7 at reactants, ~ −0.9 at the acylation TS; returns to ~ −0.7 at AEI and deacylation products. In deacylation TS, the carbonyl oxygen charge is less negative (~ −0.6) with greater negative charge on the attacking water oxygen.
• Trp185 role:
  • Trp185 exhibits two χ1 conformations (−60° and +60°). A spontaneous conformational change to +60° was observed in the AEI during equilibration for deacylation.
  • Free energy profiles show minimal preference in acylation states but a strong preference for +60° in deacylation reactants/products, facilitating product release and inhibiting reverse reaction.
  • Aromatic interactions: During acylation, PET–Trp185 interactions are edge-to-face; during deacylation, they shift toward parallel-displaced π-π, with Tyr87 also engaging in parallel-displaced interactions, forming an aromatic clamp.
  • Flexibility of Trp185 promotes catalysis, consistent with decreased activity in mutations restricting its motion (e.g., S214H/S238F-W159H context).
• Overall, unbiased RC discovery with TPS+iLMax yields high-transmission RCs that resolve prior mechanistic ambiguities.

The unbiased QM/MM TPS approach identified concise, validated reaction coordinates for both acylation and deacylation, enabling robust free energy and kinetic estimates. The results address long-standing questions in serine hydrolase catalysis: the PETase mechanism uses a moving histidine proton shuttle (not ring flipping), there is no charge-relay double-proton transfer between Asp and His, and the tetrahedral geometry corresponds to the transition state rather than a metastable intermediate. The oxyanion hole stabilizes charge consistently across steps, with dynamic engagement of Met161 upon attack. Mechanistically, deacylation has a slightly higher barrier and lower transmission-corrected rate, making it the rate-limiting step; the computed rate constant agrees with experimental kcat values. Importantly, Trp185 conformational flexibility between steps modulates active-site packing and π-π interactions, lowering the free energy of the AEI configuration favorable for deacylation and likely facilitating product egress while disfavoring reversal. These insights refine previous structural and QM/MM proposals that assumed RCs, demonstrating the value of discovering and validating RCs directly from unbiased trajectory ensembles. They also provide concrete targets for engineering PET hydrolases: maintaining or enhancing Trp185 flexibility, preserving oxyanion hole geometry, and modulating interactions that stabilize the deacylation transition state.

Using unbiased QM/MM transition path sampling with inertial likelihood maximization, the study elucidates the PETase catalytic mechanism with validated reaction coordinates for acylation and deacylation. PETase employs a moving histidine mechanism for proton transfer, maintains a stable Asp–His hydrogen bond without a double-proton transfer, and stabilizes the developing oxyanion via Tyr87 and Met161 backbones. Both steps are concerted with a single tetrahedral transition state; deacylation is rate-limiting (k ≈ 0.82 s−1 at 310 K), in agreement with experiment. Trp185’s conformational reorientation between steps promotes catalysis and product release, rationalizing decreased activity when its motion is restricted. These mechanistic insights are likely extensible to other PET-degrading cutinases with conserved active sites and can inform protein engineering strategies focused on modulating Trp flexibility, oxyanion stabilization, and deacylation barrier reduction. Future work should investigate substrate context effects (full polymer chains and surfaces), temperature and pretreatment dependencies, endo/exo cleavage modes, and how Trp185 resets its χ1 conformation across catalytic cycles, particularly at solid–enzyme interfaces.

The simulations used a hydroxyethyl-capped PET dimer as a substrate analog in solution, whereas PETase acts on PET surfaces in practice; binding modes to solid PET and polymer chain conformations may differ. Conformational distributions and kinetics can depend on temperature and substrate pretreatment. The mode of action (exo vs endo) on polymer chains remains system-dependent. The precise mechanism for resetting Trp185 χ1 to its initial basin between catalytic cycles is unresolved. These factors may affect quantitative energetics and dynamics relative to in situ conditions.