Microhydration of Deprotonated Nucleobases

Hydration reactions of deprotonated nucleobases (uracil, thymine, 5-fluorouracil,2-thiouracil, cytosine, adenine, and hypoxanthine) produced by electrospray have been experimentally studied in the gas phase at 10 mbar using a pulsed ion-beam high-pressure mass spectrometer. The thermochemical data, ΔHo, ΔSo, and ΔGo, for the monohydrated systems were determined. The hydration enthalpies were found to be similar for all studied systems and varied between 39.4 and 44.8 kJ/mol. A linear correlation was found between water binding energies in the hydrated complexes and the corresponding acidities of the most acidic site of nucleobases. The structural and energetic aspects of the precursors for the hydrated complexes are discussed in conjunction with available literature data. Graphical Abstract ᅟ


Introduction
H ydrogen bonding plays a central role in biological structures and function, including protein and nucleic acid folding, molecular recognition, signal transduction, and enzymatic catalysis [1]. Hydrogen bonds in DNA and the interaction between two complementary nucleobases, which are held together by NH-O and NH-N hydrogen bonds, are dependent on the intrinsic basicity of the acceptor atoms as well as on the acidity of the donor groups [2,3]. The strength of these bonds is related to the pKa values of the components [4]. The hydrogen bonding between the nucleobases (NB) in DNA and RNA duplexes is very important for a greater understanding of their structure and function in vivo [5].
When ionizing radiation interacts with living organisms, the low-energy electrons (<15 eV) efficiently damage DNA by inducing single-and double-strand breaks [6]. These alterations are initiated by dissociative electron attachment (DEA) with the initial capture of an electron leading to a temporary negative ion, which may decompose by spontaneous ejection of the electron or by dissociation into neutral and anionic fragments [6,7]. Gas-phase studies have shown that the most abundant fragment anions formed via the DEA process of uracil [8], thymine [9,10], cytosine [11], 2-thiouracil [12], adenine [13], and hypoxanthine [14] are the deprotonated nucleobases [NB-H] -. The formation of these anions is energetically driven by the electron affinity of the [NB-H] • radicals, which lie in the range between 3 and 4.5 eV [9,11,15,16].
Although it is essential to characterize the properties of deprotonated forms of isolated nucleobases, it is equally important to examine their properties in environments that mimic some of the aspects of the biological world. Water is the natural medium of biological systems, and for that reason our investigations are focused on the hydration of different ionic forms of nucleobases. In our previous studies, we investigated the thermochemical properties for the gas-phase hydration of protonated nucleobases and protonated nucleosides [45], sodiated and potassiated nucleobases [46], and protonated and sodiated thiouracils [47].
In this paper, we present the experimental investigations of the interactions of one molecule of water with deprotonated uracil The five nucleobases (U, T, C, A, and G) are directly involved in the formation and the stability of the well-known double helix structure of DNA and RNA. We could not conduct measurements with G (guanine) as it is sparingly soluble in the electrospraying solution. H is a mutagenic purine base that most commonly arises from the oxidative deamination of A, and is associated with carcinogenesis and cell death [38]. Modified nucleobases, 5-FU and 2SU, are important and interesting compounds because of their biological and pharmacological properties. 5-FU is widely used in the treatment of a range of cancers, including colorectal and breast cancers, and cancers of aerodigestive tract [48,49]. 2-SU has found medical applications as antithyroid and anticancer drugs [50][51][52].
Several theoretical studies on the interaction of deprotonated nucleobases with water have been performed. Kryachko et al. [22] estimated the binding energies of water molecule with the N3anions of 2-SU, 4-SU, and 2,4-dSU. Wetmore and coworkers [29,30] computationally investigated the binding energies of neutral and the N1anionic uracil and its derivatives with small molecules (NH 3 , H 2 O, or HF) at the O 2 (N3), O 4 (N3), and O 4 (C5) binding positions. Their results showed that the binding strengths are relatively independent of the substituent. Furthermore, they reveal decrease in the deprotonation energy at N1 by about 20 kJ/mol with one associated water to uracil [29]. Computational studies by Bachrach and Dzierlenga [42] have indicated that the difference (54.4 kJ/mol) in deprotonation energy between the N1 and N3 sites of uracil decreases with each added water up to four. At this point, the energy difference has been halved, but addition of a fifth or sixth water has little effect on the energy difference. The Wetmore group [34] carried out density functional theory studies of the complexes between NH 3 , H 2 O, or HF molecules and four main binding sites in neutral and N1 deprotonated cytosine. They found that the trends in the effects of hydrogen bonds on the N1 acidity are similar for all pyrimidines. To the best of our knowledge, no experimental results on the gas-phase hydration of deprotonated nucleobases have been reported.

Experimental
The experiments were performed with a high-pressure mass spectrometer using a pulsed ion-beam ESI ion source, which has been described in detail elsewhere [53]. Briefly, the reactant ions were produced by electrospraying water/acetonitrile (20%:80%) solutions containing~2.0 mM nucleobase to which a few drops of ammonium hydroxide were added. The pH value of solution measured with Schott CG 837 (Mainz, Germany) instrument was~10.5. Each solution was supplied to a silica capillary (15 μm i.d., 150 μm o.d) by a syringe pump at a rate of 0.8 μL/min, and a negative voltage was held at approximately 4 kV.
The clustered ions were desolvated by a dry nitrogen gas counter current and in a heated (~80°C) pressure-reducing capillary through which they were introduced into the forechamber, and then deflected toward a 3-mm orifice in the interface plate leading to the reaction chamber (RC). Ions drifting across the RC toward the exit slit under the influence of a weak electric field (2 V/cm at 10 mbar) were hydrated and reached equilibrium prior to being sampled to the mass analysis section of the mass spectrometer. Ion detection was provided by a channeltron equipped with a conversion dynode. The output pulses of the multiplier were counted using a multichannel scaler with dwell-time per channel of 1 μs. Mass spectra were registered with continuous ion sampling, while for equilibrium determination the ion beam was injected into the RC in a pulsing mode by applying short pulses (-52 V, 200 μs) to the deflection electrode. The latter mode of operation allows for measurements of the arrival time distribution (ATD) of the ions across the RC.
The reagent gas mixture consisting of pure N 2 as the carrier gas at about 10 mbar and a known partial pressure of water vapor (0.1-0.25 mbar) was supplied to the RC via the heated reactant gas inlet (RGI) at a flow rate of~100 mL/min. The pressure was measured with an MKS capacitance manometer attached near the inlet of the RGI. The amount of water introduced into the N 2 gas flow was kept constant throughout the temperature-dependent measurements of the equilibrium constants. Water concentrations were controlled continuously with a calibrated temperature and humidity transmitter (Delta OHM, Type DO 9861T; Casselle di Selazzano, Italy). The RC temperature was monitored by an iron-constantan thermocouple, which was embedded close to the ion exit slit; the temperature can be varied from ambient to~300°C by electrical heaters.
The gas-phase hydration energies of deprotonated nucleobases were determined by measurement of the equilibria described by the general reaction (1) for which the thermodynamic equilibrium constant is where I n and I n-1 are recorded ATD peak areas of [NB-H] -٠(H 2 O) n and [NB-H] -٠(H 2 O) n-1 respectively, and P is the known partial pressure of water (in mbar). The standard pressure P o is 1000 mbar. Equilibrium attainment in the RC was verified by comparing the ATDs of the reactant and product ions, and testing that the I n /I n-1 ratio was independent of ion residence time. A typical example of such tests is shown in Figure 1 for Measuring K n-1,n as a function of temperature T and using the thermodynamic relationships (3) and (4) [30]. f Ref. [37]. g Ref. [20]. h Ref. [31]. i Ref. [39]. j Ref. [32]. k Ref. [38]. l Ref. [54]. During these experiments, we determined thermochemical data for the hydration Reaction 5 to support the validity of the present results and provide bases for comparison with the data obtained in previous studies [54] (see Table 1).

Results and Discussion
The van't Hoff plots for the temperature studies of the hydration reactions of [NB-H]are shown in Figure 2 and the results are summarized in Table 1, along with related literature data.
The results show that the hydration enthalpies, ΔH o , for all anions are essentially the same, and the small differences can be attributed to the correlation with the gas-phase acidities of nucleobases. The data will be presented elsewhere. In this work, the term "gas-phase acidity" is used to refer to the enthalpy change, ΔH o ac , associated with deprotonation. Table 1 shows the gas-phase acidities of the most acidic and the less acidic site of nucleobases. For all these nucleobases, more than one site in the molecule can be deprotonated. Similarly to the neutral nucleobases, their deprotonated forms can exist in several tautomeric structures, and the measured hydration enthalpy changes for [NB-H]may represent an average over several contributing structures. The formation of [NB-H]by ESI could occur from different locations. The anions produced from aqueous solution may be different from those formed in the gas-phase region, in which changes can occur either in the transition of the ion from the charged droplet to the gas phase or in the gas phase due to ion-molecule reactions [55], where catalyzed isomerization can occur in the presence of neutral nucleobase [20]. The possible anionic structures of [NB-H]created by ESI that might be involved in the hydration equilibrium 1 are characterized in the following discussion.

Uracil and Its Derivatives
For uracil and its derivatives, the possible deprotonation sites are N1 and N3. In the gas phase, N1 is more acidic than N3, by about 45-60 kJ/mol (see Table 1), while in aqueous solution the N1 and N3 acidities of uracil are indistinguishable, and the N1monoanion is in equilibrium with that of N3in ca. 1:1 ratio [43]. A similar proportion also holds for the mixtures of the monoanions N1and N3in aqueous medium of thymine [56] and 2-thiouracil [57]. For 5-fluorouracil, the spectral data [58] show the predominance of N3in the N1and N3monoanionic mixture in aqueous solution. However, in alkaline aqueous solution, the situation can be different. Theoretical studies [59] show that in alkaline aqueous media, the deprotonation at N1, with equilibrium constant, K eq (N1) , should be the dominant path of uracil ionization. This result is supported by the reaction field calculations with the isodensity polarizable continuum (IPC) model, with the equilibrium constant ratio, K eq (N1) / K eq (N3) = 5 × 10 4 . In the case of 5FU, the N1 -/ N3anion fraction ratio in aqueous alkaline solution was found to be 0.61 [60]. The N3anion, if formed in aqueous solution, in the gas phase can isomerize to N1in the presence of neutral nucleobase [20]. According to the in vacuo ab initio calculations, the N1anion of [U-H]is more stable than N3by 58.5 kJ/mol [59]; for [5FU-H]this difference is 49.9 kJ/mol [60]. The energy barrier (185.4 kJ/mol) calculated [61] for the uracil N1 --> N3conversion is too high to be overcome at thermal energies in our instrument. Therefore, it is reasonable to assume that the N1would be the predominant form of the [NB-H]anions of uracil and its derivatives (structure 1 in Scheme 2) formed by ESI in the present study and these species are the most favorable precursors for hydrated complexes. Calculations [42] for the uracil N1predict that the most stable complex with water, 1a, is formed when water is attached to the anion in a bidentate fashion between the deprotonated N1 and the adjacent carbonyl oxygen. Configuration 1b and the complexes with water binding at the O4(C5) and O4(N3) positions in uracil (not shown in Scheme 2), are significantly (at least 12.6 kJ/mol) higher in energy than 1a [42] and would be expected to be minor in abundance under the present experiments. It is very likely that the 1a and 1b structures are also formed from the hydrated structure As can be seen in

Cytosine
According to the calculations [39], the canonical tautomer of cytosine, 2, is the most stable and the three other most stable tautomers are higher in energy by 7.1 (2a), 10.5 (2b), and 9.2 kJ/mol (2c). The next most stable tautomer is predicted to be lying 16.7 kJ/mol higher in energy than 2 (Scheme 3).
As it has been shown [39] that the cytosine formed by electrospray of a methanol aqueous solution adopts predominantly the 2 form, where the most acidic site is N1. Thus, it might be expected that the N1anion of the tautomer 2 should be the dominant precursor of the [C-H] -٠(H 2 O) complex formed in the present experiments. The measured hydration energy for this complex (44.8 ± 2 kJ/mol, Table 1), is significantly lower than the water binding strengths calculated for the 2e (57.1 kJ/mol) and 2f (51.1 kJ/mol) complexes [34].These results imply that the 2d complex dominates in the equilibrium reaction 1.

Adenine
In the gas phase, the canonical tautomer of adenine, 3, is the most stable and predominant species. The next two tautomers, 3a and 3b, are higher in energy by~34 kJ/mol [41,62], Scheme 4.
Tautomerization 3 -> 3a and 3 -> 3b is predicted [63] to occur with a very large activation barrier (250-293 kJ/mol), indicating that the processes may not occur in the gas phase. In water, however, the energy difference between the canonical and these two tautomers is reduced to 4.7 kJ/mol (3a) and 18.0 kJ/mol (3b) [64]. The experimental measurements [65][66][67] and calculations [68] show that only the 3 and 3a tautomers might be present in an aqueous solutions, and their population ratio, 3/3a, was estimated to be in the range of 3.6-4.9 at 293 K.
In our experiments, the formation of [A-H]by ESI can occur from different locations of the parent molecule. In aqueous solution, these anions may result from the dominant tautomer 3 with possibly up to 20% of the 3a tautomer. In the atmospheric pressure region, the ion formation predominantly from 3 may be expected. Therefore, it is very likely that the 3c anion (Scheme 4) formed from a mixture of 3 and 3a should be the precursor for the hydrated complexes.
The negative Mulliken charges predicted by theoretical studies [69] for the N atoms of the adenine N9are equal to 0.25e (N1), 0.25e (N3), 0.23e (N9), 0.23e (N7), and 0.20e (N10). These results suggest that the negative charge in 3c is uniformly distributed, and a possibility exists that the resonance structures of this anion, 3d, 3e, 3f, and 3 g, can interact with the water molecule leading to the hydrated complexes 3 h, 3i, and 3j (Scheme 4). It is also possible that we have a mixture of these complexes, and the hydration energies measured for these systems represent an average of their contribution. However, a comparison of the calculated [33] gas-phase acidities for

Hypoxanthine
Theoretical and experimental studies [38,[70][71][72] indicate that in the gas phase hypoxanthine can exist mainly in two keto tautomeric forms, 4 and 4a, (Scheme 5). The canonical structure 4 is calculated to be less stable than the 4a by 3.5 kJ/mol; the next most stable tautomer, 4b, is 22.6 kJ/mol higher in energy than 4a [38].
The calculations [70] show that the 4a tautomer represents about 80% of the population in the gas phase. The predicted concentration for 4b would be less than 0.1%. Hydration shifts in the tautomeric equilibria toward the 4 form; in the case of the dihydrated species, the populations of the 4 and 4a tautomers would be about 50% [70]. Also, quantum chemical and Monte-Carlo calculations [73] indicate that both species might be coexisting under similar tautomeric populations in neutral hypoxanthine aqueous solution. The resonance Raman spectroscopy and quantum chemical calculations study [74] reported that in solution the hypoxanthine anion is formed only via deprotonation of the N7H and N9H sites. Thus, based on these results, one may assume that [H-H]formed from 4 and 4a by ESI, either in solution or within the droplets, represent a mixture of the deprotonated tautomers of similar populations. The negative Mulliken charge distribution predicted by the calculations [74] for the N3, N7, N9, and O10 atoms of the 4c anion are equal to 0.509, 0.543, 0.541, and 0.589e, respectively. The charges at the N3 and N9 atoms are comparable with those of N7 and O10, and both these positions could be the reactive sites for water interaction with resonance structures 4d and 4e leading to the complexes 4f and 4g, as schematically depicted in Scheme 5. The acidity values calculated [38] for the most

Correlation Between Water Binding Energies and Acidities
A plot of the binding energies (-ΔH o ) of water molecule in the [NB-H] -٠(H 2 O) complexes versus the corresponding gas phase acidities of the most acidic site of NB is shown in Figure 3. The gas-phase acidity values used for this figure and also quoted in Table 1, except for 2SU [35], are obtained experimentally and reported in the literature [20,31,[37][38][39]. A fair linear relation is observed in Figure 3. The correlation coefficient is 0.98. Changes in hydration enthalpies of [NB-H]can be thermochemically analyzed on the basis of the gasphase acidity enthalpy, ΔH o ac , for deprotonation given by Equation (6) ΔH

Comparison to Neutral and Protonated Nucleobase
The hydration enthalpies obtained in this work for [NB-H] -٠(H 2 O) along with the literature values calculated [19,26,67,72] for the neutral, [NB]٠(H 2 O), and those measured [45] previously for protonated forms, [NB+H] + ٠(H 2 O), using the same methods employed here are compared in Table 2 and Figure 5. For all anionic complexes, the water binding energies are larger than those for the corresponding neutral complexes. This confirms the electrostatic nature of water interaction with the anionic forms of nucleobases. The stronger H-bonding interactions in the cationic complexes than those in anionic   Table 1. The D(NB-H) and EA(NB-H) values for 5FU, A,U,T, and C are taken from Ref. [37]; for H, Ref. [14]. For 2SU, the EA(NB-H)=365 kJ/ mol was estimated as the difference of D(NB-H) -E(DEA) based on E(DEA)=52.9 kJ/mol [12] and using for D(NB-H) the same value as for U can be attributed to higher positive charge density concentrated on the site of [NB+H] + protonation compared with a delocalized negative charge in the anionic nucleobases. For example, in the N1anions of [U-H]and [2SU-H] -, the large negative charge is located on the O2(S2) and O4 atoms [22,42]. In [A-H]and [H-H] -, as discussed above, the negative charge is uniformly distributed on the N atoms. The electrostatic potentials calculated [76] for the deprotonated A, C, and T indicate that the negative charge is "spread" throughout the

Conclusions
In the present work, we have investigated the monohydration of deprotonated nucleobases produced by electrospray from alkaline solutions (pH~10.5

Open Access
This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.  Table 2