Extended Data Fig. 10: Time-resolved atomic-level insights into the catalysis reaction.

a) Series of 1D ¹H spectra recorded after addition of Mg²⁺ to pre-formed non-stabilized Dz^5C–RNA complexes revealing clear time-dependent changes of the peak intensities. b) Extract of 1D spectra recorded on Dz^5C–RNA before (black), directly after (blue), and at indicated time points (up to about 2 h) after Mg²⁺ addition. A spectrum recorded on stabilized Dz^5C–RNA^2ʹF, preincubated with Mg²⁺ for three days, serves as a reference for the Mg²⁺-equilibrated precatalytic complex (red). c) Contour plot of time-resolved 1D NMR data following the cleavage reaction. A magnification of data shown in (a) highlighting two peaks representing the educt (right) and product (left) state is displayed. A clear shift of the peak maximum during the initial stage of the reaction is apparent for the educt state, which is not present for the product state (dashed lines are given as guides). The direction of the observed frequency change follows the CSP induced by Mg²⁺ binding (visible in (b)) and therefore is in line with an increasing effective Mg²⁺:Dz ratio. The data strongly suggest that the post-catalytic complex has a lower affinity for Mg²⁺ ions than the precatalytic complex leading to an effective Mg²⁺ release before product release. d,e) Series of 2D [¹H-¹H] TOCSY spectra recorded before (black) and successively after adding 1 mM MgCl₂. The acquisition time of each spectrum was 3 h. Clear peak position, peak shape, and/or intensity changes are present between the first (d, cyan) and second spectrum (e, brown). Only very weak intensity changes are present between second and third spectrum (e, orange). Indicated magnifications compare peak positions before (black) and during Mg²⁺-induced transition (d, cyan/blue denote positive/negative contour levels) and after full cleavage (e, orange). f) Magnification of signal for rU₊₆ shown in (d). The peak positions from the Mg²⁺ titration obtained on Dz^5C–RNA^2ʹF at indicated Mg²⁺ concentrations are shown as purple dashed lines. The regions representing the educt and product peaks in the real-time spectrum are highlighted in brown and orange. The region of the Mg²⁺-free state is highlighted in grey. The data reveal that the observed transition does not start from the Mg²⁺-free state but instead from a state that matches the frequencies of the Mg²⁺-equilibrated state at Mg²⁺ levels between 0.5 and 1 mM. g) Overlay of time-resolved experimental data for one cross peak (T8-H6/H7; positive/negative contours are denoted in blue/cyan, respectively) with best-fit simulated spectrum (red/yellow denote positive/negative contours). Characteristic features of the initial and final states are indicated. The simulations can reproduce the experimental data well. Note that data shown in panels a, b, d and g show enlarged versions of the respective data shown in Fig. 3. h) Difference between experimental and simulated real-time NMR data as a function of different rate constants applied in the simulations. Results are shown individually for each resolved atom. The minimum represents the best fit condition. The nuclei can be divided into two groups, which either show rates that are slower than the FRET rate determined under matching conditions (brown dashed line) and are plotted in blue, or faster rates (plotted in red). Respective nucleotides are mapped on the structure in Fig. 3d) 1D slice of T8(H6-H7) obtained from the indirect dimension of the 2D spectrum recorded during the cleavage reaction (d, cyan). Experimental data (black, representing state C₁) is compared to peak shapes simulated using either a 2-state transition model (blue) or a 3-state transition model (red). Both simulated spectra represent the best fit for the respective model. Note that negative shoulders are better represented by the 3-state model (dotted arrows). j) Difference between the experimentally obtained data for indicated nucleotides and the simulated data as a function of the rate constant used in the simulation. While for dC₊₅ (grey) the overall fit is close to the experimental data (as visible by the rather low overall values of the difference), the peak is insensitive to changes in the rate constants (as visible by the low variations over the range of applied rate constants). The respective peak, therefore, is not a good sensor of the transition kinetics and was excluded from data interpretation. On the contrary, the data for C3 show a clear minimum and display good sensitivity in respect to changes in rate constants. The respective peak was therefore included in further data analysis. k) Likelihood of interactions of G14 or hexa-hydrated Mg²⁺ with the O2ʹ atom of rG₀ or the O5ʹ atom of rU_-1 at the RNA cleavage site, when in any conformation of the cleavage site (left) and when the cleavage site is in the in-line attack conformation (angle O2ʹ-P-O5ʹ: 130–180°, right). An interaction is considered present if a hetero-atom of G14 or Mg²⁺ is within 5 Å of the target atom. l) Effect of Mg²⁺ binding to the metal ion-binding site II on the frequency of structural features of the RNA cleavage reaction during MD simulations at 20 mM Mg²⁺. m) FRET-based activity assay in the presence of Mg²⁺ (red data points, 1 mM) or Mn²⁺ (blue data points, 0.5 mM) of Dz^5C with a 6-thio-modification at either G14 (left) or at G6 (right). While the measured behaviour of 6-thio-G6 is well in line with previous reports²³, the unusual behaviour of 6-thio-G14 would be in line with the acid-base mechanism shown in panel k (right, X₁ = G14).