Docking interactions determine substrate specificity of members of a widespread family of protein phosphatases
How protein phosphatases achieve specificity for their substrates is a major outstanding question. PPM family serine/threonine phosphatases are widespread in bacteria and eukaryotes, where they dephosphorylate target proteins with a high degree of specificity. In bacteria, PPM phosphatases control diverse transcriptional responses by dephosphorylating anti-anti-sigma factors of the STAS domain family, exemplified by B. subtilis phosphatases SpoIIE, which controls cell-fate during endospore formation, and RsbU, which initiates the General Stress Response. Using a combination of forward genetics, biochemical reconstitution, and AlphaFold2 structure prediction, we identified a conserved, tripartite substrate docking interface comprised of three variable loops on the surface of the PPM phosphatase domain that recognize the three-dimensional structure of the substrate protein. Non-conserved amino acids in these loops facilitate the accommodation of the cognate substrate and prevent dephosphorylation of the non-cognate substrate. Together, single-amino acid substitutions in these three elements cause an over five-hundred fold change in specificity. Our data additionally suggest that substrate-docking interactions regulate phosphatase specificity through a conserved allosteric switch element that controls the catalytic efficiency of the phosphatase by positioning the metal cofactor and substrate. We hypothesize that this is a generalizable mechanistic model for PPM family phosphatase substrate specificity. Importantly, the substrate docking interface with the phosphatase is only partially overlapping with the much more extensive interface with the upstream kinase, suggesting the possibility that kinase and phosphatase specificity evolved independently.