基于序列统计分析的酶pH适应性改造

Enzymatic pH adaptation is a significant concern for protein engineers due to its great importance in ensuring the maximum activity of enzymes. It is of great insterst for scientists to develop an efficient strategy to shift enzymatic pH optimum. The optimal pH is primarily governed by the the dissociation state of key catalytic residues, which are further influenced by complex protein microenvironment and amino acid interactions. Due to the insufficient knowledge about the relationship between the enzyme structure and pH adaptation, it is difficult to accurately locate the key mutagenesis residues using traditional rational protein design. In this study, xylanses from GH11 family were selected as the model to study the pH adaptation. 115 well-characterized xylanases from more than 1500 sequences were selected, and then a new sequence library with annotated pH optimum was constructed. The pI value and hydrophobic index were used to descript the 115 sequences by artificial neural networks and Lasso linear regression analysis. Eight potential sites which were related to pH adaption were identified. Among them, the pH optimum of F54W，S55D，A165E，D175Y and Q176E shifted 0.5-0.75 units towards acidic area compared with the wild-type enzyme, meanwhile these mutants maintained more than 80% of the wild-type activity.