Initial Primer Synthesis of a DNA Primase Monitored by Real-Time NMR Spectroscopy

Primases are crucial enzymes for DNA replication, as they synthesize a short primer required for initiating DNA replication. We herein present time-resolved nuclear magnetic resonance (NMR) spectroscopy in solution and in the solid state to study the initial dinucleotide formation reaction of archaeal pRN1 primase. Our findings show that the helix-bundle domain (HBD) of pRN1 primase prepares the two substrates and then hands them over to the catalytic domain to initiate the reaction. By using nucleotide triphosphate analogues, the reaction is substantially slowed down, allowing us to study the initial dinucleotide formation in real time. We show that the sedimented protein–DNA complex remains active in the solid-state NMR rotor and that time-resolved 31P-detected cross-polarization experiments allow monitoring the kinetics of dinucleotide formation. The kinetics in the sedimented protein sample are comparable to those determined by solution-state NMR. Protein conformational changes during primer synthesis are observed in time-resolved 1H-detected experiments at fast magic-angle spinning frequencies (100 kHz). A significant number of spectral changes cluster in the HBD pointing to the importance of the HBD for positioning the nucleotides and the dinucleotide.

Figure S3: 1 H-31 P CP-MAS spectra of primase-DNA CT -ATP C -dGTP C with different ATP C and dGTP C ratios.The spectra on top were recorded on a sample using a primase:dGTP C :ATP C 1:1:1 molar ratio and the bottom spectra on a sample with a primase:dGTP C :ATP C 1:2:2 molar ratio.The color of the spectra reflects the time points after ultracentrifugation at which the spectra have been taken.In the 1:1:1 sample the reaction has completed already during ultracentrifugation (only the dinucleotide-bound state is detected), which is not the case in the 1:2:2 sample.In the latter, an increase in dinucleotide binding is observed over time (increasing resonance at around -3 ppm), whereas the constant triphosphate resonances indicate the continuous replacement of bound triphosphates from the excess in the supernatant (pointing to an unfinished reaction).Time point zero represents the end of the ultracentrifugation after spinning up the MAS rotor.The Pγ integral of dGTP C has been normalized to 1(right panel), which assumes that the reaction that is occurring before taking the first spectrum is negligible.For PCP the leftmost pair of resonances has been integrated (left panel).The integrals have been normalized with the last data point to 0.68 (which corresponds to the integral of the dGTP C Pγ resonance after ~400 h).This has been done to compensate for the incomplete reaction after more than 400 h.The assumption that the amount of PCP formation before the first measurement is negligible and that the process of normalization is therefore valid, is based on the missing PCP resonances in the first 1 H-31 P CP-MAS spectrum after rotor filling and has been confirmed by analyzing the ratio of the absolute integrals of dGTP C (t= 0 h)/ PCP (t= 400 h), which is 0.69.With this method, we observe an increase in PCP with an initial rate constant of 0.35 ± 0.17 d -1 and a decrease in dGPT C with an initial rate constant of 0.28 ± 0.09 d -1 , purple linear regressions in both plots.Table S1: CSP data extracted from 2D hNH between primase-H145A-DNA CT -ATP N -dGTP C and primase-DNA CT -ATP N -dGTP C (spectra are shown in Figure 6

Figure S1 :
Figure S1: Overlay of 1 H-15 N solution-state HSQC spectra of HBD-DNA CT -2ATP in the absence (shown in blue) and presence of dATP at pH 7.0 (a), dCTP at pH 5.5 (b), or dTTP at pH 5.5 (c) (shown in red), respectively.

Figure S2 :
Figure S2: Close-up view of K366 and N348 in a 1 H-15 N solution-state HSQC experiment with different ATP and dGTP analogues.The cross peak corresponding to N348 becomes only visible, if both nucleotides are bound to the protein, as the loop K340-N348 only rigidifies upon nucleotide binding.This experiment has been used to identify suitable ATP and dGTP analogues for this study.

Figure S4 :
Figure S4: 2D 31 P-31 P 200 ms PDSD spectra of primase-DNA CT -ATP C -dGTP C before (blue) and after the dinucleotide formation reaction (red).The P  and P  chemical shifts of ATP C and dGTP C overlap.The formation of the PCP:Mg 2+ complex and the increase of the peak at around -3 ppm (see black arrow) indicates the formation of the dinucleotide.The P α resonances only appear rather weak on the diagonal with the current signal-to-noise ratio.

Figure S5 :
Figure S5: Real-time solution state 31 P NMR spectra of primase-DNA CT -ATP N -dGTP C with a 1:1:50:50 ratio.Due to the excess of nucleotides used, the peak corresponding to the 4 th phosphate of the dinucleotide (the phosphodiester group) can be distinguished from the DNA template.

Figure S6 :
Figure S6: 2D 31 P-31 P 150 ms DARR spectra of primase-DNA CT -ATP N -dGTP C (blue) and PCP:Mg 2+ (grey).The latter has been prepared from a solution of medronic acid and MgCl2 (ratio 1:10).A minor fraction of the polymorph of the complex PCP:Mg 2+ is shown with red dashed lines.Note that the individual PCP resonances are not resolved in the 2D spectrum due to a too short acquisition time in the direct dimension.The additional resonances observed in the grey spectrum are attributed to different polymorphs of PCP:Mg 2+ formed from the medronic acid solution.

Figure S7 :
Figure S7: 2D 31 P-31 P 50 ms DARR spectrum of primase-H145A-DNA CT -ATP N -dGTP C .Dashed lines show the correlation between the neighboring 31 P nuclei for ATP N , dGTP C and the PCP:Mg 2+ complex, in the latter clearly showing the set of two Mg 2+ :PCP resonances.

Figure S8 :
Figure S8: Temperature-dependent 1 H-31 P CP-MAS and direct-pulsed 31 P MAS NMR spectra of primase-DNA CT -ATP C -dGTP C .The bound nucleotides are only detected at low temperature in the CP spectra, whereas at higher temperature only the Mg 2+ :PCP resonances are observed in the CP spectra.

Figure S9 :
Figure S9:Changes in the peak integrals of the 1 H-31 P CP-MAS spectra of primase-DNACT -ATP N -dGTP C as a function of time.The Pγ integral of dGTP C has been normalized to 1(right panel), which assumes that the reaction that is occurring before taking the first spectrum is negligible.For PCP the leftmost pair of resonances has been integrated (left panel).The integrals have been normalized with the last data point to 0.68 (which corresponds to the integral of the dGTP C Pγ resonance after ~400 h).This has been done to compensate for the incomplete reaction after more than 400 h.The assumption that the amount of PCP formation before the first measurement is negligible and that the process of normalization is therefore valid, is based on the missing PCP resonances in the first 1 H-31 P CP-MAS spectrum after rotor filling and has been confirmed by analyzing the ratio of the absolute integrals of dGTP C (t= 0 h)/ PCP (t= 400 h), which is 0.69.With this method, we observe an increase in PCP with an initial rate constant of 0.35 ± 0.17 d -1 and a decrease in dGPT C with an initial rate constant of 0.28 ± 0.09 d -1 , purple linear regressions in both plots.

Figure S10 :
Figure S10: 2D 31 P-31 P 150 ms DARR spectra of primase-DNA CT -ATP N -dGTP C after 3 days (blue) and 12 days (red) of sample filling.Red arrows show the side products (AMP and Pi) of ATP N hydrolysis.

Figure S11 :
Figure S11: Intensity change as a function of time determined from the 1 H-31 P CPMAS spectra of primase-DNA CT -ATP N -dGTP C (a), from the 1 H-31 P CPMAS spectra of primase-H145A-DNA CT -ATP N -dGTP C (b) and from the solution-state 31 P spectra of primase-DNA CT -ATP N -dGTP C (c).(d) Solution-state NMR tube of primase-DNA CT -ATP N -dGTP C showing the formed Mg 2+ :PCP complex as white precipitate.

Table S2 :
Overview about experimental parameters of the performed solid-state NMR experiments.For more details about the used adiabatic CP steps and the tangential shapes used see reference 2 .