Mining and rational design of psychrophilic catalases using metagenomics and deep learning models

Shuning Wu,Guoshun Xu,Yongping Su,Huoqing Huang,Xinxin Xu,Yuhong Zhang,Jian Tian,Wei Zhang,Zhiwei Zhang,Bo Liu
DOI: https://doi.org/10.1007/s00253-023-12926-1
IF: 5
2024-01-04
Applied Microbiology and Biotechnology
Abstract:A complete catalase-encoding gene, designated soiCat1, was obtained from soil samples via metagenomic sequencing, assembly, and gene prediction. soiCat1 showed 73% identity to a catalase-encoding gene of Mucilaginibacter rubeus strain P1, and the amino acid sequence of soiCAT1 showed 99% similarity to the catalase of a psychrophilic bacterium, Pedobacter cryoconitis. soiCAT1 was identified as a psychrophilic enzyme due to the low optimum temperature predicted by the deep learning model Preoptem, which was subsequently validated through analysis of enzymatic properties. Experimental results showed that soiCAT1 has a very narrow range of optimum temperature, with maximal specific activity occurring at the lowest test temperature (4 °C) and decreasing with increasing reaction temperature from 4 to 50 °C. To rationally design soiCAT1 with an improved temperature range, soiCAT1 was engineered through site-directed mutagenesis based on molecular evolution data analyzed through position-specific amino acid possibility calculation. Compared with the wild type, one mutant, soiCAT1S205K, exhibited an extended range of optimum temperature ranging from 4 to 20 °C. The strategies used in this study may shed light on the mining of genes of interest and rational design of desirable proteins.Key points• Numerous putative catalases were mined from soil samples via metagenomics.• A complete sequence encoding a psychrophilic catalase was obtained.• A mutant psychrophilic catalase with an extended range of optimum temperature was engineered through site-directed mutagenesis.
biotechnology & applied microbiology
What problem does this paper attempt to address?