Cleaving DNA with DNA: Cooperative Tuning of Structure and Reactivity Driven by Copper Ions

Abstract A copper‐dependent self‐cleaving DNA (DNAzyme or deoyxyribozyme) previously isolated by in vitro selection has been analyzed by a combination of Molecular Dynamics (MD) simulations and advanced Electron Paramagnetic Resonance (Electron Spin Resonance) EPR/ESR spectroscopy, providing insights on the structural and mechanistic features of the cleavage reaction. The modeled 46‐nucleotide deoxyribozyme in MD simulations forms duplex and triplex sub‐structures that flank a highly conserved catalytic core. The DNA self‐cleaving construct can also form a bimolecular complex that has a distinct substrate and enzyme domains. The highly dynamic structure combined with an oxidative site‐specific cleavage of the substrate are two key‐aspects to elucidate. By combining EPR/ESR spectroscopy with selectively isotopically labeled nucleotides it has been possible to overcome the major drawback related to the “metal‐soup” scenario, also known as “super‐stoichiometric” ratios of cofactors versus substrate, conventionally required for the DNA cleavage reaction within those nucleic acids‐based enzymes. The focus on the endogenous paramagnetic center (Cu2+) here described paves the way for analysis on mixtures where several different cofactors are involved. Furthermore, the insertion of cleavage reaction within more complex architectures is now a realistic perspective towards the applicability of EPR/ESR spectroscopic studies.

and Table S4 summarizing g-factor and hyperfine data  EPR experiments.Continuous Wave (CW) X-Band measurements were carried out using an X-band Bruker E500 instrument (9.4 GHz, TE012 resonator) equipped with a nitrogen flow cryostat.All CW experiments were recorded at 120 K and with a shot repetition rate of 100 Hz, unless stated otherwise.Pulsed EPR experiments at X-band were performed on a Bruker ELEXYS E-580 X-band spectrometer with a SuperX-FT microwave bridge and a Bruker ER EN4118X-MD4 dielectric resonator.Cryogenic temperatures (20 K) were obtained by the use of an Oxford flow cryostat.The field-swept EPR spectra were recorded by electron spin echo (ESE) detection; electron-spin-echo (ESE)-detected EPR experiments were carried out with the pulse sequence: π/2-τ-π-τ-echo.For the X-band experiments the mw pulse lengths tπ/2 = 16 ns and tπ = 32 ns and a τ value of 200 ns were used.A two-step phase-cycle was applied to remove all unwanted echoes.The Hyperfine Sublevel Correlation (HYSCORE) experiments were carried out using the pulse sequence π/2-τ-π/2-t1-π-t2-π/2-τ-echo.The time traces of the HYSCORE spectra were baseline corrected using a third-order polynomial, apodized with a Hamming window and zero-filled.After two-dimensional Fourier transformation, the absolute value spectra were calculated.A four-step phase cycle (for X-band experiments) was used to remove unwanted echoes.The pulse sequence for the four-pulse DEER experiment was π/2obs − τ1 − πobs − t1 − πpump − (τ1 + τ2 − t1) − πobs − τ2.The pump pulse was applied on the spectral maximum and the observer pulses were applied at a frequency offset of 55 MHz.
Measurements on the Bruker Elexsys E580 spectrometer were acquired with a pulse delay τ1 of 200 ns and a dead time delay of 100 ns.All DEER data were analyzed using DeerAnalysis2022 based on MATLAB.The distance distribution Pdis(r) was fitted by Tikhonov regularization (using residual method for regularization parameter selection).The capture of active radicals generated during the reaction was examined by EPR spectra (Bruker ELEXSYS 500 spectrometer) using DMPO as a spin trapping agent (Supplementary methods).DMPO (0.8 M) was immediately added to the c4s4 solution and transferred to a glass capillary tube.
Then, the capillary tube was placed into a quartz EPR tube and EPR spectra were recorded.
Modelling and MD simulations.The 3D structure of the 46-mer (Fig. 1a) was modelled using x3dna package (v2) 42 using 1JVE pdb 43 as template for the "GGA hairpin".Parmbsc1 forcefield 44 was used for the MD simulation of the 46mer using gromacs2016 package 45 46mer was placed in a dodecahedron box with 1.0 nm distance between the box walls and the DNA molecular and was solvated with tip3p 46 water molecules (MD-46mer).For simulations with Cu2+ ions, 10 mM of Cu2+ ions were randomly placed in the simulation box (MD-46mer-Cu 2+ ) (Fig. S2).Simulation system was energy minimized using steepest descent algorithm until the largest force was smaller than 1000 kJ/mol/nm, followed by temperature equilibration to 300 K in 100ps using Berendsen thermostat with a tau-t of 0.2 ps.Pressure was equilibrated to 1 atm in 1 ns using Berendsen barostat and temperature was regulated using Berendsen thermostat at 300 K. 47 Production run simulations were started using the equilibrated structures and ten replicas were simulated for 500 ns.For production run simulations, temperature was regulated usi velocity-rescaling thermostat 48 and pressure with Parrinello-Rahman barostat 49 at 300 K and 1 atm using tau-t of 0.1ps and tau-p of 2ps.Structures were saved every 10 ps.
Clustering of the MD simulations was done using gromos algorithm 50 with snapshots sampled at 200 ps intervals using only phosphate backbone atoms with a RMSD cut-off of 0.2 nm.All the analysis was done using in-house python scripts and data visualization was done using Pymol. 51R experiments.NMR experiments [52][53][54][55][56][57] were recorded with a 3-mm NMR sample tube at 298°K using the following Bruker spectrometers with z-axis pulsed field gradients: Avance NEO at 900-MHz (21 Tesla) proton frequency with CPTCI CryoProbe is a proton-optimized triple resonance NMR 'inverse' probe.A standard pulse sequence lebpgp2s from Topspin 4.1, (Bruker) was used for diffusion experiments on the 900-MHz (21 Tesla) instrument.In total, 65 536 points with 16 scans were recorded in the proton dimension for each one dimension with variable diffusion gradient strength ranging between 2 and 95% in various steps.The following parameters were used: diffusion time (Δ) 0.085 s, gradient pulse (δ) 2500 µs smoothed rectangular-shaped gradients SMSQ10.100,relaxation delay (d1) 10 s.NMR experiments were recorded with a 5-mm NMR sample tube at 298°K Avance NEO at 400 MHz (9.4Tesla) equipped with a TBI probe.All spectra were processed with Topspin 4.1(Bruker) and DOSY treatment Software Dynamic center (Bruker) MALDI-TOF experiments.The MALDI-TOF [58][59][60] mass spectra measurements were performed in the negative mode on a Microflex mass spectrometer (Bruker, Wissembourg, France).Prior to mass analysis, oligonucleotides solutions were purified and concentrated using Zip Tip pipette tips (Merck Millipore) filled with 0.6 µL C18 resin.A mixture of the purified DNA sample (10 pmol, 1 µL) was added to the matrix (3-hydroxypicolinic acid in 10 mM ammonium citrate buffer) and spotted on a polished stainless target plate using the dried droplet method.Spectra were calibrated using reference oligonucleotides of known masses.Each spectrum were obtained by summing 300 shots by the use of a 337 nm pulsed nitrogen laser beam (60 Hz).
Linear mode was run with optimized voltages for ion sources (IonSource-1: 20 kV, IonSource-2: 18.5 kV) and pulsed ion extraction delay was fixed at 100 ns.In order to eliminate the intense low masses of the spectra (matrix peaks, solvents clusters) which normally saturates microchannel plate detectors, and with the aim to enhance the ratio signal to noise, all ions with less mass than 1800 Daltons were deflected.Spectra were accumulated by FlexControl Software (v.3.3.108.0) and processed with FlexAnalysis using Savitsky-Golay algorithm (with 0.2 m/z, one cycle) and baseline subtraction (Top Hat).

DFT calculations.
All calculations were performed using the ORCA program package. 61Full geometry optimization was carried out for using the GGA functional BP86 62 with the def2-TZVP basis sets. 63For the Coulomb fitting the def2/J auxiliary basis sets were used. 64For according to the experimental conditions, all calculations were performed using an implicit solvation model (epsilon = 20, refractive index = 1.33) 65 by invoking the Control of the conductor-like polarizable continuum model (CPCM). 66EPR parameters, namely g-tensors and metal hyperfine coupling constant, were obtained from single-point calculations using the previously defined aug-cc-pVTZ-Jmod basis set 67 for Cu, the EPR-II basis for nitrogens 68 while the def2-TZVP basis sets 63 were used for all other atoms.Hyperfine tensors were computed with the B3PW91 functional, 69 which was shown to be the best functional for copper hyperfine coupling constants 67 while the g-tensor was computed using a modified version of B3PW91 with 40% exact (Hartree-Fock) exchange. 70 S6.
Table S6.Diffusion coefficient and calculated molecular weights for the different fragments obtained from the cleavage reaction, including the references analyzed by DOSY (H2O, c4, s4 and histamidine monomer).

Figure S2 .Figure S4 .Figure S5 .
Figure S2.Modelled structure of 46-mer with Cu2+ randomly placed around the simulation box Figure S3.Cluster representatives and the respective population percentage of the cluster from the MD ensemble of 46-mer without Cu 2+ Figure S4.Cluster representatives and the respective population percentage of the cluster from the MD ensemble of 46-mer with Cu 2+ Figure S5.Distance distributions profiles obtained from MD ensembles with (MD-46mer-Cu 2+ ) and without Cu 2+ (MD-46mer) between the phosphate atoms of C6 and A21 and G19 and C46.

Figure S6 .Figure S7 .
Figure S6.Interaction between Cu2 + and the 46mer with a distance cut-off of 0.4nm was defined as residence time (RTA) and the contact frequency as percentage of simulation time is shown for each independent MD trajectories below Figure S7.Time series distance profiles between the phosphate atoms of A14 and A41 calculated from the independent trajectories of MD simulations with Cu2 + (MD46mer-Cu2 + ) Figure S8.Spin labeled c4s4 oligomer and structures of nitroxide spin labels Figure S9.Distance distributions from DEER/ PELDOR Figure S10.CW EPR spectra and corresponding EasySpin fit; Peisach Plot and Table S1 summarizing EPR parameters Figure S11.Full HYSCORE spectrum for the c4s4 Figure S12.HYSCORE and Hyscorean fit, TableS2

Figure S7 .
Figure S7.Time series distance profiles between the phosphate atoms of A14 and A41 calculated from the independent trajectories of MD simulations with Cu2 + (MD46mer-Cu2 + ).

Figure S8 .
Figure S8.Spin labeled c4s4 oligomer and structures of nitroxide spin labels.On the left side the oligomers purchased from Eurogentec®, while on the right side the oligomers obtained with the procedure published elsewhere (reference 26-27 of the main text).

Figure S13 .
Figure S13.DFT-model for CopperG13G31 obtained from geometry optimization with 1 H (Top) and without 1 H (Bottom) removed for clarity.The numbering for the monomeric deoxy-guanosine is reported: N7 and N9 correspond to N46/N15 and N43/N12, respectively.

Table of contents S1 .
Materials and Methods

Table S4 .
Calculated nitrogen hyperfine coupling constants (individual components A, MHz) of DFT-optimized model for the complex Cu 2+ /G13/G31 (Numbering of deoxy-guanosine is reported in Supporting Information S13).