Catalytic promiscuity is defined as the ability of an enzyme to catalyse several chemically distinct reactions besides its native activity. This latent promiscuity appears to be more widespread than originally thought and may have played an important role in adaptive evolution by providing possible evolutionary starting points for the emergence of new functions [1-2]. The Alkaline Phosphatase (AP) superfamily provides an ideal framework to explore this hypothesis, as it encompasses a large set of evolutionary related metalloenzymes frequently exhibiting crosswise catalytic promiscuity [3-4]. Based on these observations, our work aims at addressing the following question: which mechanistic and structural features determine specificity or promiscuity among AP superfamily members?

We performed directed evolution to investigate how P. aeruginosa arylsulfatase (PAS) can be turned into a phosphonate monoester hydrolase (PMH), one of its secondary activities. Three rounds of neutral drift [5], followed by six rounds of selection for improved promiscuous activity resulted in a highly generalist enzyme with a kcat/KM of 103-104 M-1 s-1 for four chemically distinct substrates:  phosphate diesters, phosphate, phosphonate, and sulfate monoesters. Furthermore, the analysis of evolutionary intermediates revealed the clustering of substitutions in loops that are absent in PMH structures. Semi-rationally designed deletion libraries of these regions produced mutants tolerating up to 18 AA deletions and displaying compelling changes in specificity (>103-fold). Detailed kinetics, LFER, mutational and structural analysis of evolutionary intermediates shed new light on the fine molecular changes accompanying the respecialization of a promiscuous scaffold. Taken together, our results provide new insights into the respective contribution of binding and catalysis, as well as multi-functional trade-offs in the evolution of new functions. It also emphasizes the role of deletions in opening alternative evolutionary trajectories.