Guanosine Dianions Hydrated by One to Four Water Molecules

Intermolecular interactions such as those present in molecule···water complexes may profoundly influence the physicochemical properties of molecules. Here, we carried out an experimental–computational study on doubly deprotonated guanosine monophosphate···water clusters, [dGMP – 2H]2–·nH2O (n = 1–4), using a combination of negative anion photoelectron spectroscopy (NIPES) with molecular dynamics (MD) and quantum chemical (QM) calculations. Successive addition of water molecules to [dGMP – 2H]2– increases the experimental adiabatic detachment (ADE) and vertical detachment energy (VDE) by 0.5–0.1 eV, depending on the cluster size. In order to choose the representative conformations, we combined MD simulations with a clustering procedure to identify low energy geometries for which ADEs and VDEs were computed at the CAM-B3LYP/6-31++G(d,p) level. Our results demonstrate that the assumed approach leads to sound geometries and energetics of the studied microsolvates since the calculated ADEs and VDEs are in pretty good agreement with the experimental characteristics. The evolution of hydrogen bonding with cluster size indicates the possibility of the occurrence of proton transfer for clusters comprising a larger number of water molecules.


Experimental methods
The NIPE spectra were recorded using a home-made cryogenic magnetic-bottle time-offlight (TOF) photoelectron spectrometer, coupled with an electrospray ionization (ESI) source and a temperature-variable cryogenic ion trap. 1 The [dGMP-2H] 2− dianion and its solvated clusters were produced by spraying 0.1 mM solutions of the corresponding sodium salt in H 2 O/CH 3 CN (1:3) under a humidity-controlled environment. The clusters were guided by an RF quadrupole ion guide and detected by a quadrupole mass spectrometer to optimize ESI conditions to ensure stable and intense solvated cluster beams. Then, the anions were directed into the cryogenic 3D ion trap, where they were accumulated and cooled by collisions with a cold buffer gas (20% H 2 balanced in helium). In the present work, the temperature inside the ion trap was set at 20 K. The cryogenic anions were then pulsed out into the extraction zone of a TOF mass spectrometer at a 10 Hz repetition rate. Each of the desired clusters was massselected and maximally decelerated before being photodetached in the interaction zone with 157 nm (7.867 eV) light from an F 2 excimer laser. The laser was operated at a 20 Hz repetition rate with the ion beam off at alternating laser shots for shot-by-shot background subtraction.
The distribution of the detached electrons was analyzed using a 5.2 m-long magnetic-bottle electron flight tube, and then converted into kinetic energy spectrum after calibration with the known data of I − and OsCl 6 2− . The electron binding energy (EBE) spectra were obtained by subtracting the kinetic energy spectra from the photon energy used. The electron energy resolution (∆E/E) was about 2%, i.e., ~20 meV for 1 eV kinetic energy electrons.

Computational methods
To interpret the obtained PES spectra of the [dGMP-2H] 2− •nH 2 O anions, quantum chemistry calculations preceded by molecular dynamics based conformational search were conducted for all solvated systems. The initial [dGMP-2H] 2− geometry was taken from our previous study. 2 In order to identify the most important structures for the geometries generated in the MD simulation, a clustering against each of four microsolvated groups was used. Then the contributions of particular structures were computed based on the differences in their Gibbs free energies calculated quantum-chemically. Finally, the VDE and ADE values for the most represented anionic geometries were calculated and compared to the experimental data.

Molecular dynamics conformational search
The geometries of complexes for QM calculations were obtained from clustering the MD simulations performed with Gromacs18. [3][4][5][6][7][8] The nucleoside dianion starting structure was taken from our previous study and parametrized with parmbsc1 9 forcefield (DG3) and phosaa10 S3 modifications (for the additional hydrogen atom) with charges calculated explicitly with Gaussian16, 10 B3LYP 11 /6-31++G(d,p). [12][13][14] One to four waters were placed around the nucleoside dianion randomly and the complex put into a geometric center of cubic box with side length 40 Å. Three replicas with different starting points (water orientations) were prepared.
After a short steepest descent minimization (up to 5000 steps) 1 microsecond NVT simulations were performed with a timestep of 2 fs for each replica in 200 and 300 K resulting in 6 simulations for every n=1 to 4 waters of the [dGMP-2H] 2− •nH 2 O systems. Note that the targeted hydrated clusters were generated from ESI solutions at ambient conditions and the rapid solvent evaporation and cryogenic cooling afford to freeze and retain conformational spaces accessible at near room temperatures. 15 A Verlet cut-off and a twin range cut-off 16 of 18 Å was used along with PBC conditions. LINear Constraint Solver (LINCS) 17,18 was used to constraint hydrogen bonds and Berendsen thermostat. 19 The center of mass translation and rotation around the center of mass were removed. Next, we clustered the last 900 ns of each of the generated MD trajectories using the gromos algorithm implemented in Gromacs 18. For that root mean square deviation (RMSD) was computed based on all atoms and the cut-offs of 0.1, 0.15, 0.2 and 0.25 Å were applied for 1, 2, 3 and 4 waters, respectively. The representative structures of each cluster were further transferred to QM calculations.

Quantum chemistry methodology adjustment
In order to choose an appropriate DFT level, four XC functionals were tested on the set of ten manually prepared singly solvated [dGMP-2H] 2− •H 2 O dianions (manual-1 to manual-10).
We compared the experimental PES spectrum with the VDE and ADE values calculated for those ten systems using the following four methods: i) B3LYP 11 22 as performing similarly well as the CAM-B3LYP functional, was tested. Finally, we also considered iv) the HSE06 23 functional combined with the aug-cc-pvdz basis set, which was found to be the best method for reproducing the experimental values of VDE in anionic silicon clusters. 24 Considering only the structures of x M ≥0.01 for further discussion the number of conformations is reduced to 13, 7, 9 and 11 for the system solvated with 1, 2, 3 and 4 water molecules, respectively. Those the most represented conformations were discussed in the main discussion, while for full data see Table S1-S4.

VDE and ADE definitions
Vertical detachment energy (VDE) of [dGMP-2H] 2− •nH 2 O is defined as the difference in electronic energy of dianionic system and its corresponding anion radical, both in the optimized dianion geometry, while adiabatic detachment energy (ADE) is the difference in Gibbs free energy between dianionic and anion radical systems, both in their optimal geometries.
To reduce the set of VDE/ADE values to a single number, we also calculated a weighted average VDE (VDE avg ) and ADE (ADE avg ), taking into account the contribution of each conformer to the equilibrated mixture of conformers (for the complete rather than reduced data sets):

Hydration energies
For particular hydrates, hydration energies were calculated as the difference between the sum of the free energies (ΔG hyd ) of non-interacting monomers and that of the cluster, all calculated at the optimal geometries: where G(X) denote the free energy of species X obtained by correcting its electronic energy for zero-point vibration, thermal contributions to energy from vibrations, rotations, and translations, entropy and the pV term. These terms were determined in the harmonic oscillator-rigid rotor approximation for T = 298 K and p = 1 atm.  ΔG hyd-BSSE /n (kcal/mol) Figure S1. Water binding energy per water molecule (ΔG hyd-BSSE /n) dependence on experimental VDE increment (ΔVDE exp ).