Unprecedented hour-long residence time of a cation in a left-handed G-quadruplex†

Cations are critical for the folding and assembly of nucleic acids. In G-quadruplex structures, cations can bind between stacked G-tetrads and coordinate with negatively charged guanine carbonyl oxygens. They usually exchange between binding sites and with the bulk in solution with time constants ranging from sub-millisecond to seconds. Here we report the first observation of extremely long-lived K+ and NH4+ ions, with an exchange time constant on the order of an hour, when coordinated at the center of a left-handed G-quadruplex DNA. A single-base mutation, that switched one half of the structure from left- to right-handed conformation resulting in a right–left hybrid G-quadruplex, was shown to remove this long-lived behaviour of the central cation.


D 2 O exchange NMR spectroscopy
Previously prepared NMR samples were flash frozen in liquid nitrogen and subsequently lyophilized. Just before the start of the experiment, an amount of D 2 O was added that was equal to the previous sample volume and the sample was immediately measured on a NMR spectrometer.

Tracking of 15 NH 4 + using 1 H NMR spectroscopy
A sample of Z-G4 was folded in either 100 mM 15 NH 4 + /K + and subsequently dialyzed against H 2 O for 10 minutes in order to remove most of the 15 NH 4 + /K + ions in solution. It was quickly flash frozen in liquid nitrogen and lyophilized. Just before the start of the experiment, the sample was re-suspended in a buffer containing the opposite ion, either 100 mM K + / 15 NH 4 + ions. The sample was immediately tracked with NMR where we interchangeably recorded both 1 H and 15 N filtered 1 H NMR spectra over 3 hours.

Tracking of dissociation of K + and NH 4 + ion using using 1 H NMR spectroscopy
A sample of Z-G4 was folded in either 100 mM 15 NH 4 + /K + and subsequently dialyzed against H 2 O for 10 minutes in order to remove most of the 15 NH 4 + /K + ions in solution. It was quickly flash frozen in liquid nitrogen and lyophilized. Just before the start of the experiment, the sample was re-suspended in deionized H 2 O.

Sample preparation for ESI-MS experiments
A sample of Z-G4 was folded in 90 mM TMAA pH = 7, 10 mM KCl to a concentration of 0.1 mM, heated to 90 °, slowly cooled overnight and left for 7 days. Before measurements the sample was diluted 10x by either 100 mM TMAA pH = 7 or 100 mM NH 4 OAc pH = 7. In both cases, after the 10x dilution, the effective concentration of K + in solution was 1 mM, while the sample concentration was 0.01 mM. 100 mM TMAA buffer was used when measuring the sample behavior in K + as TMAA does not coordinate between tetrads, due to its bulky nature, and is added to maintain ionic strength.
100 mM NH 4 OAc buffer was used when tracking the conversion from K + to NH 4 + . The effective concentration of the NH 4 + ion after the 10x dilution is 90 mM making the K + /NH 4 + ratio 1/90. Samples were injected in the negative-ion ESI-MS with softest possible experimental tuning, that preserved the most NH 4 + ion species. Samples were tracked through time by periodical injections and the recorded data was cleaned of non-specific adducts.

Electrospray ion mobility-mass spectrometry measurements
The experiments were carried out on an Agilent 6560 IMS-Q-TOF (Agilent Technologies, Santa Clara, CA), with its drift tube ion mobility cell operated in helium. The helium pressure in the drift tube was 3.89 ± 0.01 Torr, and the pressure in the trapping funnel is 3.63 ± 0.01 Torr. The pressure differential between the drift tube and the trapping funnel ensures only helium is present in the drift tube. Injection was in negative ion mode, using the standard electrospray source and a syringe pump at 4 µL/min. The acquisition software version was B.07.00 build 7.00.7008. The arrival time distributions were extracted from the entire isotopic distribution of each adduct, using IM-MS browser Step-field experiments (five drift tube voltages for each samples) were performed to determine the CCS. The arrival time distributions (ATDs) for each charge state of the complexes were fitted with one gaussian peak using OriginPro 2016, to determine the arrival time t A of the center of the peak. The arrival time t A is related to V (voltage difference between the entrance and the exit of the drift tube region) by: t 0 is the time spent outside the drift tube region and before detection. A graph of t A vs. 1/V provides K 0 from the slope and t 0 as the intercept. The drift tube length is L = 78.1 cm, the temperature is measured accurately by a thermocouple (here, T = 297 ± 1 K), and the pressure is measured by a capacitance gauge (p = 3.89 ± 0.01 Torr). The CCS is then determined using Equation (S2): The reconstruction of the experimental CCS distributions from the arrival time distributions at the lowest voltage is then performed using Equation (S3), where the factor a is determined from the of the peak center at the lowest voltage and the CCS calculated from the regression described above, from the peak centers.
We checked with a method described elsewhere 1 whether diffusion outside the IMS would significantly contribute to the peak width, but we found that this contribution would contribute to only 1-2% of the peak width a the lowest voltage. Hence using Equation (S3) renders well the actual width of the CCS distribution.

Generation of gas-phase structures and calculation of theoretical collision cross sections
The gas-phase modeling was started from the published X-ray structure (PDB code 4U5M). Protons were added, and all but eight phosphate groups (arbitrarily chosen) were neutralized by protons, to attain a total charge state of 5-with 3 K + ions inside. We optimized the structure at the PM7 semi-empirical level 2 using Gaussian 16 rev. B.01 3 . Then, atom Centered Density Matrix Propagation molecular dynamics (ADMP, 1000 fs, 296 K) at the semi-empirical level (PM7) was also performed using Gaussian 16 rev. B01. The theoretical CCS values were calculated for a structure every 6 fs (every 30 steps), using the trajectory model (Mobcal 4 , original parameters for helium, N and O parameterized as C, P and K parameterized as Si).

Isotope-exchange mass spectroscopy
The 41 K/ 39 K isotope exchange experiments were performed on a Bruker 7T SolariX XR ESI-Q-FTICRMS (Bruker, Bremen, Germany) with the ESI source operated in negative ion mode. The injection flow rate was 180 µL/h. All spectra were acquired in soft conditions. 5 The capillary exit voltage was -180 V, and skimmer 1 voltage was -5 V. The collision energy (entrance of the hexapole collision cell) was set to 1 V.
The stock solutions of the quadruplexes (50 µM strand concentration) were prepared in 100 mM trimethylammonium acetate (TMAA) with 5 mM KCl enriched to 96.5% 41 K. The solution was then diluted to 2.5 µM in quadruplex using a solution of 100 mM TMAA and 1 mM KCl (natural isotope ratio, i.e. 93.26% 39 K and 6.73% 41 K). The potassium exchange inside the quadruplex was probed as a function of time (25 min and 3 days). The end point was obtained upon annealing of the solutions at high temperature to disrupt the quadruplexes, cooling them down and leaving them at room temperature for 3 days prior to the mass analysis. Following dilution, the effective isotopic abundance of K + for the mixture were 78.8% in 39 K and 21.2% in 41 K.      The red arrow shows the appearance of the inner G-tetrad imino proton peak with time (species with K + ) while the blue arrow shows the disappearance of the inner G-tetrad imino proton peak with time (species with NH 4 + ). All growing imino proton peaks correspond to K +containing Z-G4, while all diminishing imino proton peaks correspond to NH 4 + -containing Z-G4 (B) Graphs of the corresponding peak areas versus time.        The guidelines indicate the average centroid predicted for 0, 1, 2 and 3 cations exchanged. Z-G4 exchanges 2 K + ions faster than 25 mins, while the last K + is still preserved after 3 days. The ZG4-T4mod exchanges nearly all three cations faster than 25 mins. Figure S17. (A) Superimposition of the PM7 optimized structure (guanines in cyan, backbone in green, thymines in brown) and the X-ray structure (PDB: 4U5M) (guanines in pink, backbone in light blue and thymines in dark blue). The RMSD between the two structures is 1.43 Å 2 , the main contribution coming from rearrangment of thymines (B) and backbone (C).