Bacterial and plant HAD enzymes catalyse a missing phosphatase step in thiamin diphosphate biosynthesis.

The penultimate step of thiamin diphosphate (ThDP) synthesis in plants and many bacteria is dephosphorylation of thiamin monophosphate (ThMP). Non-specific phosphatases have been thought to mediate this step and no genes encoding specific ThMP phosphatases (ThMPases) are known. Comparative genomic analysis uncovered bacterial haloacid dehalogenase (HAD) phosphatase family genes (from subfamilies IA and IB) that cluster on the chromosome with, or are fused to, thiamin synthesis genes and are thus candidates for the missing phosphatase (ThMPase). Three typical candidates (from Anaerotruncus colihominis, Dorea longicatena and Syntrophomonas wolfei) were shown to have efficient in vivo ThMPase activity by expressing them in an Escherichia coli strain engineered to require an active ThMPase for growth. In vitro assays confirmed that these candidates all preferred ThMP to any of 45 other phosphate ester substrates tested. An Arabidopsis thaliana ThMPase homologue (At4g29530) of unknown function whose expression pattern and compartmentation fit with a role in ThDP synthesis was shown to have in vivo ThMPase activity in E. coli and to prefer ThMP to any other substrate tested. However, insertional inactivation of the At4g29530 gene did not affect growth or the levels of thiamin or its phosphates, indicating that Arabidopsis has at least one other ThMPase gene. The Zea mays orthologue of At4g29530 (GRMZM2G035134) was also shown to have ThMPase activity. These data identify HAD genes specifying the elusive ThMPase activity, indicate that ThMPases are substrate-specific rather than general phosphatases and suggest that different evolutionary lineages have recruited ThMPases independently from different branches of the HAD family.