Bioinformatic screening of natural polyphenol oxidases for theaflavins synthesis

We developed a bioinformatic approach to screen natural active PPOs. Using the PPO core sequence (pfam PF00264), 529 plant PPOs were blasted in the Uniprot database. After deducted incomplete, synthetic and repeated sequences, 178 independent plant PPOs with higher than 30% identity to tea PPO (A6N8J4) were obtained. The 3-D structures of 163 PPOs were constructed based on the PPO crystal structure of sweet potato (PDB 1BUG), optimized by molprobity and evaluated using the Procheck and Verify3D programs. The similarities of spatial structures of 163 PPO models were compared to tea PPO using the TM-score and RMSD systems. The 163 PPOs were docked with tea catechins as ligands using Autodock-Vina programs, and their potential enzyme activities were quantitatively evaluated by their binding affinity. Analysis indicated that PPO activity may mainly rely on some specific amino acids at the reaction center, but not on its homology of amino acids and spatial structure. The enzyme activities of 10 top potential plants were analyzed. The results showed that Chinese pear and sweet potato had high PPO activity, higher than tea PPO. The rich resource of sweet potato allows large production of the enzyme for industrial application. This study indicated the bioinformatic screening is an efficient method in finding out highly active enzymes from a large number of sequences.