Biochemical and structural analyses of purified proteins are essential for the understanding of their properties. However, many proteins are unstable and difficult to purify, hindering their characterization. The B2 proteins of the lasso peptide biosynthetic pathways are cysteine proteases that cleave precursor peptides during the maturation process. The B2 proteins are poorly soluble, and no experimentally solved structures are available. Here, we performed a rapid semicomprehensive mutational analysis of the B2 protein from the thermophilic actinobacterium, Thermobifida f usca (FusB2), using a cell-free transcription/translation system, and compared the results with the structure prediction by AlphaFold2. Analysis of 34 FusB2 mutants with substitutions of hydrophobic residues confirmed the accuracy of the predicted structure, and revealed a hydrophobic patch on the protein surface, which likely serves as the binding site of the partner protein, FusB1. Our results suggest that the combination of rapid cell-free mutant analyses with precise structure predictions can greatly accelerate structure−function research of proteins for which no structures are available.

Biochemical and structural analyses of purified proteins are essential for the understanding of their properties. However, many proteins are unstable and difficult to purify, hindering their characterization. The B2 proteins of the lasso peptide biosynthetic pathways are cysteine proteases that cleave precursor peptides during the maturation process. The B2 proteins are poorly soluble and no experimentally-solved structures are available. Here, we performed a rabid semi-comprehensive mutational analysis of the B2 protein from the thermophilic actinobacterium, Thermobifida fusca (TfuB2) using a cell-free transcription/translation system, and compared the results with the structure prediction by AlphaFold2. Analysis of 34 TfuB2 mutants with substitutions of hydrophobic residues confirmed the accuracy of the predicted structure, and revealed a hydrophobic patch on the protein surface, which likely serves as the binding site of the partner protein, TfuB1. Our results suggest that the combination of rapid cell-free mutant analyses with precise structure predictions can greatly accelerate structure-function research of proteins for which no structures are available.