The Phosphatase RosC from Streptomyces davaonensis is Used for Roseoflavin Biosynthesis and has Evolved to Largely Prevent Dephosphorylation of the Important Cofactor Riboflavin-5′-phosphate.

[Display omitted] • RosC is the first roseoflavin specific phosphatase whose structure has been solved. • The reaction mechanism of AFP dephosphorylation is now clear. • RosC has evolved to contain specially designed strand-to-helix sections and a unique N-terminal helical extension to carry out its functions. • Distinction between AFP and FMN is supported by the aspartate residue D166. • The crystal structure of RosC gives an idea how cell-damaging dephosphorylation of FMN may be avoided. The antibiotic roseoflavin is a riboflavin (vitamin B 2) analog. One step of the roseoflavin biosynthetic pathway is catalyzed by the phosphatase RosC, which dephosphorylates 8-demethyl-8-amino-riboflavin-5′-phosphate (AFP) to 8-demethyl-8-amino-riboflavin (AF). RosC also catalyzes the potentially cell-damaging dephosphorylation of the AFP analog riboflavin-5′-phosphate also called "flavin mononucleotide" (FMN), however, with a lower efficiency. We performed X-ray structural analyses and mutagenesis studies on RosC from Streptomyces davaonensis to understand binding of the flavin substrates, the distinction between AFP and FMN and the catalytic mechanism of this enzyme. This work is the first structural analysis of an AFP phosphatase. Each monomer of the RosC dimer consists of an α/β-fold core, which is extended by three specific elongated strand-to-helix sections and a specific N-terminal helix. Altogether these segments envelope the flavin thereby forming a novel flavin-binding site. We propose that distinction between AFP and FMN is provided by substrate-induced rigidification of the four RosC specific supplementary segments mentioned above and by an interaction between the amino group at C8 of AFP and the β-carboxylate of D166. This key amino acid is involved in binding the ring system of AFP and positioning its ribitol phosphate part. Accordingly, site-specific exchanges at D166 disturbed the active site geometry of the enzyme and drastically reduced the catalytic activity. Based on the structure of the catalytic core we constructed a whole series of RosC variants but a disturbing, FMN dephosphorylating "killer enzyme", was not generated. [ABSTRACT FROM AUTHOR]