Extended Data Fig. 3: In the absence of magnesium, the open/closed rate exchange rate is in the slow exchange regime, in agreement with earlier reports²⁵.

(a) [¹H, ¹⁵N]-HSQC spectra for Zn²⁺ and Co²⁺ samples with either 20 mM ADP or 20 mM Mg²⁺ADP. Residues that were broadened in the Mg²⁺ADP sample show reduced linewidths in the ADP sample, indicating a shift in exchange timescales from intermediate to slow exchange. Black lines trace the PCSs between diamagnetic and paramagnetic samples. For residue 38, the corresponding diamagnetic peaks are at 8.46 ppm (¹H) and 123.5 ppm (¹⁵N) and are not shown for illustration purposes. (b) ¹H_N CPMG dispersion profiles for Co²⁺ Adk with 20 mM ADP and 0 mM Mg²⁺. Representative traces show slow exchange that is fit to the Tollinger equation⁷⁵ (F-statistics were used to determine whether the description by a slow exchange model compared to a “no-exchange” model was justified at the 95% confidence interval; p>0.05). (c) Representative CPMG relaxation dispersion profiles for residues in the presence on 20 mM Mg²⁺ADP. Notably, the paramagnetic chemical shift differences in the absence and presence of magnesium are similar, whereas the timescale is significantly altered (k_{open, ADP} = 2.6 ± 0.3 s⁻¹ vs k_{open, ADP} = 180 ± 36 s⁻¹). Uncertainties (s.d.) in R_2,eff are determined from the rmsd in the intensities of duplicate points (n = 3) according to the definition of pooled relative standard deviation. Uncertainties (s.d.) in chemical shift differences were calculated from the covariance matrix.