Fig. 2: Ni^II-RcnR refined mid-range metal availabilities decode Mn^II as the cognate MncA metal.

a Apo-subtracted difference spectra of RcnR (17.2 µM RcnR monomer, ɛ calculated from total protein) titrated with Ni^II, inset showing peak wavelength confirming 1:1 stoichiometry of Ni^II to RcnR monomer, or 4:1 to RcnR₄ (n = 1). RcnR (20 μM monomer) also migrated with one equivalent of N^iII by SEC (Supplementary Fig. 3b). b Representative Ni^II titration of RcnR (31.5 μM monomer) in EGTA (464 μM), solid line representing calculated K_D for Ni^II from simultaneously fitted RncR-EGTA competitions (n = 4 independent experiments) at varied EGTA concentrations (Supplementary Fig. 3c, fitting models in Supplementary Software). Dashed lines, simulations with affinities 10-fold tighter and weaker than calculated K_D. c RcnR binding to hexachlorofluorescein-labelled rcnRA operator-promoter (10 nM) by fluorescence anisotropy. Solid line, best simultaneous fit to n = 3 experimental replicates (circles, triangles, squares) for Ni^II-RcnR (fitting model in Supplementary Software), dashed line simulates apo-RcnR using published K_D 1.5 × 10⁻⁷ M and maximum ∆r_obs 0.1115³⁹. d Ni^II and DNA affinities determined above used with previously measured RcnR molecules cell⁻¹ to calculate (via Supplementary Data 2) relationship between intracellular Ni^II availability and RcnR DNA occupancy (θ_D), as for Co^II (circles show θ_D 0.99, 0.90, 0.10 and 0.01). Combined mid-point of ranges for Ni^II-RcnR and Ni^II-NikR also shown (blue arrow). e Metal availabilities (squares and inset text) as activities/concentrations and free energies, ∆G_M, at mid-points (50% DNA occupancies, representing idealised cells) of metal sensors (bars are sensor ranges, 10% to 90%), now including Ni^II-RcnR. Pale bars show individual ranges where two cognate sensors. MncA metal preferences (pale blue circles, Cu^I triangles, from Fig. 1d) as free energies (∆G_MP) from pseudo-affinities giving 99% Mn^II metalation at mid-range Mn^II availability (∆G_M) without competing metals (Supplementary Fig. 5 simulates alternative values for Mn^II-MncA ∆G_MP). Inset shows occupancies of MncA predicted from free energy differences between MncA and labile metal (∆∆G = ∆G_MP− ∆G_M). Mn^II has the largest favourable gradient annotated ∆∆G_max. Supplementary Data 3 enables similar predictions of cognate metals for other proteins. Source data are provided as a Source Data file.