ATPase assays suggest different activation states for RimK.

This plot shows the experimental data from Fig 2E as points and the curves are computed from kinetic models based on RimA and cdG binding to RimK (S1 Table). Under the assumption that RimK can exist in two states, RimK and RimK*, with different activities and that RimK* achieved either by binding cdG or RimA, we simultaneously optimised all the parameters in the system (percentage of RimK*, k_cat* and K_m*). This plot shows that a simple two state model is consistent with the data. However, under this hypothesis and with [RimK] = 1 μM, [RimA] = 1 μM, [cdG] = 25 μM, the best fit to the ATPase data is achieved for an equilibrium dissociation constant (K_d) of 83.2 nM between RimK and RimA and 15.1 μM between RimK and cdG. This it at odds with independent measurements that estimate K_d between RimK and cdG to be about 1 μM [19]. This suggests that the simple two state system is unlikely and that RimK can exist in at least four different activation states (RimK, RimK.RimA, RimK.cdG and RimK.cdG.RimA, S1 Table). This model of RimK ATPase activity results in an insignificantly small better fit to the ATPase activity data but with a K_d between RimK and cdG of 1 μM and a K_d between RimK and RimA of 0.2 μM.

(TIFF)