Parallel evolution of non-homologous isofunctional enzymes in methionine biosynthesis

Abstract

Experimental validation of enzyme function is crucial for genome interpretation, but it remains challenging because it cannot be scaled up to accommodate the constant accumulation of genome sequences. We tackled this issue for the MetA and MetX enzyme families, phylogenetically unrelated families of acyl-L-homoserine transferases involved in L-methionine biosynthesis. Members of these families are prone to incorrect annotation because MetX and MetA enzymes are assumed to always use acetyl-CoA and succinyl-CoA, respectively. We determined the enzymatic activities of 100 enzymes from diverse species, and interpreted the results by structural classification of active sites based on protein structure modeling. We predict that >60% of the 10,000 sequences from these families currently present in databases are incorrectly annotated, and suggest that acetyl-CoA was originally the sole substrate of these isofunctional enzymes, which evolved to use exclusively succinyl-CoA in the most recent bacteria. We also uncovered a divergent subgroup of MetX enzymes in fungi that participate only in L-cysteine biosynthesis as O-succinyl-L-serine transferases.

Revisiting the methionine salvage pathway and its paralogues

“…Indeed, the known fate of methanethiol is highly variable, depending on the organism and on its particular niche (Weise et al, 2012). This highly volatile molecule may evade the cell as a gas, be oxidized to methane sulphonate (Vermeij and Kertesz, 1999), or, indeed, be recovered for methionine salvage via O-acyl-homoserine (O-acetyl-or O-succinyl-as preferred protecting groups) by cystathionine gamma-synthase or related enzymes (Kiene et al, 1999, with considerable expected variations in acylation, depending on the species, see Bastard et al, 2017). As an example, assimilation of methanethiol and dimethyldisulphide has been studied in some species, Corynebacterium glutamicum in particular (Bolten et al, 2010).…”

GapMind: Automated annotation of amino acid biosynthesis

“…There are many other features that could be used. In particular, we did not incorporate specificity-determining residues (i.e., (Bastard et al 2017;Shi et al 2007) ) into GapMind because this information is available for few enzyme families. GapMind also does not consider whether a candidate gene clusters with other proteins in the pathway.…”

No wisdom in the crowd: genome annotation in the era of big data – current status and future prospects

“…This idea was used in the identification of several unknown functions coded by a synthetic genome (Danchin and Fang, 2016). In the same way, the fact that the lysine biosynthetic pathway missed a key enzyme in B. subtilis, together with the observation that protection against misuse of non-proteinogenic amino acids is performed by acetylation in B. subtilis rather than succinylation as in E. coli (Bastard et al, 2017), led us to make the hypothesis that protein PatA, annotated as an aspartate aminotransferase (Berger et al, 2003), might be the missing N-acetyl-L,L-diaminopimelate aminotransferase DapX that we identified correctly by subsequent experiments (Borriss et al, 2018).…”