Computational design of carbohydrate binding modules-fused PETase like enzyme for enhanced polycaprolactone (PCL)...

A novel PET hydrolase-like enzyme identified from metagenomic databases using HMMR search was computationally fused with five different carbohydrate-binding modules (CBMs). AlphaFold3 predicted the 3D structures of the fused enzyme-CBM, which were validated using ERRAT, Verify3D, and PROCHECK. Molecular docking was performed with polycaprolactone triol using AutoDock Vina, followed by 100 ns molecular dynamics (MD) simulations using AMBER. Trajectory analyses and binding free energy calculations (QM/MM-GBSA) were conducted. The putative PET hydrolase-like enzyme shared 49.62% similarity with Ideonella sakaiensis PETase (5XJH). The fused models exhibited the best stability, with an instability index of <40 and a thermostability aliphatic index between 58.83 and 68.27. Structure validation confirmed high-quality 3D models, with >90% of the residues in the allowed Ramachandran regions. All the fused models showed favourable binding to PCL-triol, exhibiting strong interactions. In MD simulations, BlCBM5 and TrCBM complexes displayed a minimal fluctuation: all-atom RMSD ∼0.35 and ∼0.45 nm, backbone RMSD ∼0.48, ∼0.41 nm, atom contacts ∼4.2-5, ∼2-6, and H-bonds ∼2-5, ∼1-2, respectively. The BlCBM5 and TrCBM complexes showed the lowest binding energies, with MM-GBSA values of -36.66 ± 0.12 and -21.48 ± 0.11 kcal/mol, and QM/MM-GBSA values of -37.36 ± 0.13 and -21.70 ± 0.11 kcal/mol, respectively. Residue-level analysis identified key contributors (M133, W157, and F62) in both models. BlCBM5 and TrCBM complexes were the top candidates for enhancing PCL plastic degradation. The findings of this study were based on predictive insights, and experimental validation is required in the future.