Abstract:[Objective] The aim of this study was to identify a novel L-cysteine sulfinate decarboxylase from alkaline polluted soil metagenome and to use non-rational design method to improve the enzyme. [Methods] L-cysteine sulfinate decarboxylase gene undec1A was cloned into pETBlue-2 vector and expressed in Escherichia coli Tuner (DE3) pLacI. The recombinant Undec1A protein was purified to homogeneity. The original Undec1A protein was characterized and a mutagenesis library was constructed with sequential error-prone PCR method, then an interesting variant was identified. [Results] Multiple sequence alignment analysis showed that Undec1A protein shared the similar pyridoxal 5'-phosphate binding sites and the substrate recognition motif with the other known L-cysteine sulfinate decarboxylases. Molecular docking results indicated that amino acid residues Val237, Asp239, Asp266, Ile267, Ala268 and Lys298 contributed to the decarboxylation of L-cysteine sulfinate acid. Recombinant Undec1A protein had an apparent Km of (1.557±0.015) mmol/L, a Vmax of (49.07±3.19) μmol/(L·min), and a kcat of (45.80 ±1.32)/min at the optimal reaction conditions of 35℃ and pH 7.0 when using L-cysteine sulfinate acid as the substrate. Furthermore, the protein engineering approach of random mutagenesis via sequential error-prone PCR was applied on the original Undec1A protein. Compared with the original Undec1A protein, the best variant of Undec1A-1180 in the random mutagenesis library, exhibited 5.62-folds at the optimal reaction conditions of 35℃ and pH 7.0. [Conclusion] These results are the first step towards a better understanding of the properties of Undec1A protein. Protein engineering with error-prone PCR paves the way toward the metagenome-derived genes for biotechnological applications.