Mining bacterial (meta)genomes for enzymes active in aerobic, mesophilic conditions

Enzyme biochemistry can now draw on hundreds of thousands of prokaryotic genomes and metagenomes to identify orthologous genes for research, biocatalysis, and metabolic engineering. In many applications, adaptation to O₂ and mild temperatures are essential. But as organism lifestyle information can be poor or absent (especially for metagenomes), it is challenging to avoid orthologous genes from anaerobes and extremophiles. Taking bacterial sulfide-dependent THI4 thiazole synthases as test-cases, we built computational pipelines that use only DNA sequence inputs to explore (i) the average oxidation state of carbon (Z_C) in orthologous enzymes and (ii) the presence of O₂-metabolism genes in the corresponding (meta)genomes. Z_C has been proposed to be highest (least negative) in proteins of organisms from O₂-rich, mesophilic environments. We found that Z_C values of 2300 THI4s ranged from -0.107 (relatively oxidized) to -0.302 (strongly reduced). As predicted, genes specifying cytochrome c or o oxidases (supporting respiration at relatively high O₂ levels) and, to a lesser extent, cytochrome bd oxidases (which can function to scavenge O₂) were more frequent in genomes encoding THI4s with high Z_C values. Eight THI4s with Z_C values in the top 5 % and from (meta)genomes having cytochrome oxidases were tested for ability to complement a THI4Δ yeast strain in aerobic conditions. Three THI4 genes from a metagenome with cytochrome c/o oxidase (but without cytochrome bd) were active. These results support the feasibility of combining Z_C and cytochrome oxidase profiles to identify bacterial orthologous enzymes that work in aerobic, mild temperature conditions.