Journal of Molecular Modeling

, 22:262 | Cite as

Influence of microhydration on the structures and proton-induced charge transfer in RNA intermediates

Original Paper
Part of the following topical collections:
  1. Festschrift in Honor of Henry Chermette

Abstract

Solvation effects are of major interest in the context of radiation damage, due to their potential applications in cancer therapy. Reliable modeling of the solvent is, however, quite challenging, and numerous studies have been devoted to isolated biomolecules and stepwise-hydrated molecules in which the amount of solvent is controlled one molecule at a time. The influence of stepwise hydration on radiation damage is investigated here using the example of proton-induced charge transfer in two biomolecular targets. Uracil has been widely investigated both experimentally and theoretically in this context, and 2-aminooxazole was recently shown to be a potentially important intermediate in prebiotic chemistry. Focusing here on doubly hydrated biomolecules, stable structures and infrared spectra were obtained by combining the results of molecular dynamics simulations with those of quantum chemistry calculations performed at the density-functional theory level with the double hybrid M06-2X functional. The charge-transfer cross-sections upon proton impact were obtained from ab initio molecular calculations and after applying a semi-classical approach to investigate the collision. Our results suggest a significant relationship between the detailed hydration structure and the efficacy of proton-induced charge transfer, highlighting the competing roles of inter- and intramolecular hydrogen bonding.

Keywords

Water clusters Microhydration Charge transfer Proton collisions Molecular calculations RNA chemistry IR spectra Structure 

Introduction

Radiation damage in biological media has been widely investigated due to its applicability to cancer therapy treatments. Research efforts have increasingly been directed into understanding the different mechanisms at play at the molecular level—not only the interactions with the primary radiation but also those with secondary particles [1]. Among such secondary processes, low-energy electrons have been shown to cause single- or double-strand breaks through dissociative electron attachment [2, 3], and secondary ions may ionize the biomolecules, possibly causing them to fragment. These indirect processes can contribute significantly to the decay of the medium [4, 5, 6].

In the context of ion-induced damage, gas-phase experiments provide a wealth of information about single collision events, providing access to the fragmentation pattern associated with the biomolecular target for a given impinging ion [7, 8]. The fragmentation cross-sections may be deduced from mass spectra and interpreted theoretically by determining the molecular orbitals involved in the process. Theory and modeling are invaluable for unraveling the dynamical and chemical mechanisms of the ionized target [9, 10]. Although ionization occurs very rapidly compared to the subsequent molecular rearrangement, it drives the entire mechanism either directly or through delayed charge transfer from the incident ion to the biomolecular target. This charge-transfer process is analyzed here, using a molecular representation and focusing on proton impact collisions, which have been widely investigated for related systems in previous studies [11, 12, 13]. In terms of gas-phase experiments, this approach has been successfully applied to isolated biomolecules—mainly DNA and RNA building blocks, pyrimidine nucleobases [14, 15], sugars [16, 17], and 5-halouracil targets [18]. Other interesting molecules of prebiotic relevance such as 2-aminooxazole [19] have also been studied in this manner. Damage due to ion bombardment is a natural process in astrophysical environments, and charge transfer caused by the impact of protons can be an important source of decay in such environments, particularly in ionized clouds, where protons are particularly abundant. However, it is difficult to scale up the properties of isolated molecules to actual, incredibly complex biological media. In order to account for the presence of the water solvent, it is useful to disentangle the specific features of the solute from the solute–solvent interactions by hydrating the compound of interest one water molecule at a time. Mass spectrometry experiments have thus tried to bridge the gap between the gas and condensed phases by analyzing how hydration modifies the properties of the solute [20, 21]. In the context of radiation damage, where the contribution of the solvent may be very important, this bottom-up approach of controlled microhydration is of particular relevance. In the impact of low-energy electrons on uracil and thymine, dissociative electron attachment cross-sections have indeed been shown to be significantly enhanced if the molecule is embedded in a water cluster [22]. We have also demonstrated that, in the case of uracil, this effect even occurs when just one water molecule is present [23]. In contrast, the microhydration of 2-aminooxazole does not lead to such an enhancement [24], which we interpreted as resulting from the very different hydration patterns exhibited by the two molecules. These results suggest intriguing effects of the detailed hydrogen-bonding structure on the charge-transfer efficiency upon proton impact, which are all the more important because multiple isomers could be populated at a finite temperature—as in a real liquid solvent. In view of those results, we undertook in the present paper a systematic exploration of the influence of the conformer on the charge transfer in uracil and 2-aminooxazole following proton collisions. Doubly hydrated compounds were selected as they provide a wide variety of possible hydration patterns with competing solute–solvent and intrasolute hydrogen-bonding networks.

We followed a similar protocol to that used in our earlier works [23, 24] by first exploring the conformational diversity of the dihydrated biomolecules, using a well-established hierarchical protocol that relies on a force field to generate many candidate geometries which are subsequently refined at a more accurate level of quantum chemistry using an explicit description of electronic structure. The various theoretical approaches, including those used to compute the charge-transfer cross-sections, are detailed in the next section of the paper. The results for uracil and 2-aminooxazole are then presented and discussed in the third section, before some concluding remarks are made in the final section.

Methods

Structures and IR spectra

The complex energy landscapes of the doubly hydrated uracil and 2-aminooxazole molecules were first explored using the Amber force field (ff99 parameters) [25] in replica-exchange molecular dynamics (REMD) simulations, periodically saving and locally minimizing the configurations that occurred during the simulations. The REMD strategy is particularly well suited to moderately large systems, as it allows large energy barriers to be surmounted in a much shorter time than when a conventional molecular dynamics approach is used. Although reasonable, the Amber force field does not possess chemical accuracy, so the resulting structures were subsequently locally reoptimized using a more realistic method with an explicit description of electronic structure. Here, density-functional theory (DFT) with the global hybrid meta-GGA functional M06-2X [26] was employed as our method of choice, and the triple-zeta basis set aug-cc-pVTZ was applied; diffuse functions were included mainly to obtain better descriptions of the ionic systems.

For all stable structures, we considered the relative binding energies relative to the global minimum. Adiabatic and vertical ionization energies as well as adiabatic and vertical electron affinities were calculated at the same level of density-functional theory. Electronic structure calculations also provided information about the vibrational spectra of the microhydrated compounds for possible future experimental characterization. Infrared (IR) spectra were determined in the double harmonic approximation, and the frequencies in the mid-IR range were scaled by 0.95 to account for anharmonicities and nuclear delocalization corrections at the most basic level.

Charge transfer

The collisional system formed by the impinging proton and the biomolecular target was treated in the one-dimensional reaction coordinate framework conventionally used for complex systems [27, 28]. The collision dynamics were simulated with a semi-classical approach using the EIKONXS code [29] to an accuracy of 10−4 for the symmetry of the S matrix. As the collision process is very fast, the dynamics were developed using the sudden approximation hypothesis, which considers vibrational and rotational motions to be frozen during the collision. This approximation is valid for keV impact energies, for which the nuclear vibrational time and (a fortiori) the rotational periods are about two orders of magnitude longer than the collision time [30]. Building upon a recent analysis of time-dependent quantum wavepacket and semiclassical approaches to charge-transfer processes [31], the present semiclassical approach may be extended to lower collision energies. We developed the calculations in the 10 eV to 10 keV collision energy range. For all dihydrated compounds, the biomolecules exhibit an almost planar ring, and it has been widely shown that charge transfer is strongly enhanced perpendicular to the ring [14, 15]. To achieve a better comparison of the various isomers, the proton-induced charge transfer was only evaluated in the perpendicular z orientation, averaging the results for the two directions (+z and −z) when calculating the final value of the cross-section.

The molecular calculations needed for the dynamical part were carried out at the CASSCF/6-311G** level of theory using the MOLPRO code [32]. Similar active spaces were taken for each target, including the five highest orbitals on the bare biomolecule and the 1s orbital on hydrogen. ECP2sdf effective core potentials [33] were used for C, N, and O atoms, and the lowest-energy orbitals were frozen. The charge-transfer process is driven by nonadiabatic interactions between the proton–cluster entrance channel and the correlated ionized cluster levels. The nonadiabatic radial coupling matrix elements between all pairs of states were determined numerically using the finite difference technique:
$$ {g}_{KL}(R)=\left\langle {\psi}_K\left|\partial /\partial R\right|{\psi}_L\right\rangle =\left\langle {\psi}_K(R)\left|{\displaystyle \underset{\Delta \to 0}{ \lim }}\frac{1}{\Delta}\right|{\psi}_L\left(R+\Delta \right)-{\psi}_{\mathrm{L}}(R)\right\rangle, $$
(1)
where |ψK(R)〉 and |ψL(R)〉 are the entrance and charge-transfer molecular states.
If we take into account the orthogonality of the eigenfunctions |ψK(R)〉 and |ψL(R)〉 for K ≠ L, Eq. 1 reduces to
$$ {g}_{KL}(R)=\left\langle {\psi}_K\left|\partial /\partial R\right|{\psi}_L\right\rangle ={\displaystyle \underset{\Delta \to 0}{ \lim }}\frac{1}{\Delta}\left\langle {\psi}_K(R)\left|{\psi}_L\left(R+\Delta \right)\right.\right\rangle $$
(2)

In practice, a step size Δ = 0.0012 a.u. was found to lead to a stable differentiation procedure [34].

Results and discussion

Uracil@(H2O)2

The effects of hydration on the uracil nucleobase have been widely investigated using various experimental and theoretical approaches [35, 36, 37, 38, 39, 40]. In the context of radiation damage, interest has recently focused on reactions in which electrons or protons are generated along the radiation track. These particles drive biomolecule excitation, resulting in the ionization of the biomolecule and possibly its subsequent decay via fragmentation. The dissociative electron attachment cross-sections appear to be strongly modified by solvation [22]. Our earlier analysis [23] of charge transfer following the collision of protons with a series of uracil@(H2O)n = 1–5 clusters showed that the charge-transfer efficiency increases sharply when just one water molecule is added to the bare uracil. This enhancement can be associated with the polarization of the molecular orbital that is mainly involved in the charge-transfer mechanism in the presence of water. Considering the peripheral hydration patterns observed for uracil@(H2O)n clusters, such polarization effects remain largely unchanged as the microhydration level is increased, despite a decrease in the cross-section as the number of water molecules reaches n = 5 [23]. Therefore, a more detailed analysis of this microhydration effect is required to understand this process in depth. The case with n = 2 studied here provides a great diversity of hydration patterns, allowing us to gain further insight into the specific role of structure in the charge-transfer process.

A systematic exploration of the energy landscape of uracil@(H2O)2 clusters yields eight stable minima, as depicted in Fig. 1. Their relative energies, vertical and adiabatic electron affinities (EA), as well as the corresponding vertical and adiabatic ionization energies (IE) are presented in Table 1. While the force field and the DFT calculations show some differences when the structures are ranked in energy, both approaches predict the same global minimum, which validates our two-step exploration of the energy landscape based on Amber ff99. All isomers are almost planar and are labeled according to the different regions occupied by water molecules, following the notation introduced by Bachrach and Dzierlenga [40]. Most geometries correspond to locked-chain arrangements of the two water molecules on one side of the uracil, except for AC structures, in which the water molecules are localized in different regions around uracil and are not bonded to each other. For all isomers, the two water molecules are hydrogen bonded to uracil and they point their other OH bond either towards the same side of the uracil plane (A, B, C, AC structures) or towards the opposite side (A′, B′, C′, AC′ structures). The energy difference associated with pivoting one molecule to the other side is very small and amounts to about +0.43 kcal mol−1 relative to the same-side conformers. Within the simpler Amber ff99 description, these OH bonds rotate and fall into the plane, so the appearance of two minima for a given family is due to energetic contributions that are not included in the force field—most likely polarization forces.
Fig. 1

Structures of the uracil@(H2O)2 clusters obtained by optimization at the DFT/M06-2X/aug-cc-pVTZ level

Table 1

Relative energies, vertical and adiabatic electron affinities (EA), and vertical and adiabatic ionization energies (IE) of uracil@(H2O)2 compounds, as obtained from quantum chemical calculations performed at the DFT/M06-2X/aug-cc-pVTZ level. The relative values obtained with the Amber ff99 force field are also reproted

Structure

Energy (kcal mol−1) using Amber ff99

Energy (kcal mol−1) using DFT/M06-2X

Vertical EA (eV)

Adiabatic EA (eV)

Vertical IE (eV)

Adiabatic IE (eV)

A

0.0

0.0

−0.354

0.142

9.516

8.874

A′

0.0

0.432

−0.345

0.153

9.491

8.855

B

2.781

3.038

−0.498

0.421

9.605

9.108

B′

2.781

3.465

−0.489

0.342

9.598

9.089

C

2.282

2.449

−0.393

0.402

9.580

9.111

C′

2.282

2.885

−0.397

0.404

9.575

9.092

AC

3.470

1.352

−0.397

0.323

9.556

8.887

AC′

3.470

1.380

−0.339

0.323

9.554

8.885

The structures of the different families A, B, C, or AC are separated by larger energies, which confirms that the hydrogen-bonding topology is the factor that determines the stabilities of the various conformers. The most favorable hydration pattern (A and A′) is clearly associated with hydrogen bonds that include N–H as the proton donor and oxygen as the proton acceptor. It is worth noting an interesting evolution between structures A, AC, and C from structural and energetic aspects, the stepwise displacement of a water molecule from region A to region C destabilizing the cluster by about 1.0 kcal mol−1. Both A and C are, however, relatively similar regions, offering an N–H group as a proton donor and an oxygen as a proton acceptor, with the proximity of a free C=O bond acting as a destabilizing factor. A similar remark could be made for the B family, which is clearly energetically disfavored and shows a free C=O bond together with a free N–H bond, which is closely related to the locked-chain arrangement of water molecules.

For the different uracil@(H2O)2 clusters, the electron affinities and ionization energies are not very sensitive to the region occupied by the water molecules. However, noticeable variations occur between the vertical and adiabatic values for the ionization energies, and especially for the electron affinities. While the adiabatic EAs are positive and rather low, vertical values are markedly negative. This indicates that a major rearrangement occurs upon the addition of an electron to the cluster, leading to significant stabilization of the anion. Likewise, noticeable rearrangement occurs upon the removal of an electron, as evidenced by the difference of about 0.6 eV between the vertical and adiabatic values of the ionization energy. The planar uracil@(H2O)2 structures thus appear to very efficiently stabilize an extra electron and (albeit to a lesser extent) the cationic cluster.

The vibrational spectra obtained for the different isomers of uracil@(H2O)2 in the double harmonic approximation are shown in Fig. 2 and compared with the IR spectrum of bare uracil.
Fig. 2

Harmonic IR spectra of uracil@(H2O)2 compounds calculated at the DFT/M06-2X/aug-cc-pVTZ level of theory in the far- and mid-IR ranges. The spectrum of the bare uracil is shown at the bottom

In general, the two conformers of the same topological family A, B, C, or AC exhibit nearly identical spectra, and it is sufficient to discuss the differences among families. The spectrum for family A appears to be very similar to the bare uracil spectrum, with the N–H stretching mode at almost the same position (around 3100 cm−1) and the OH modes in the 3300–3400 cm−1 range. The hydrogen bonds created by the presence of the water chain in this region do not significantly influence the IR spectrum, which looks nearly identical to the spectrum of the bare biomolecule across the entire spectral domain. In contrast, for the locked-chain structures in the B and C regions, clear shifts are observed towards lower wavelengths for the N–H stretching mode and towards higher wavelengths for the O–H modes (due to the increased sensitivity upon pivoting one out-of-plane O–H mode to the other side of the uracil plane). The spectrum is quite different for the AC–AC′ species, in which both N–H bonds are connected to water molecules, and N–H stretching modes near 3000–3100 cm−1 vanish during the process. Structures from the AC family also show the strongest contrast with the bare molecule in the bending mode region, with the absence of intrawater bonding clearly demonstrated by the presence of the intense peak near 1600 cm−1, which is exclusive to these conformers.

The correlation between the various conformers of uracil@(H2O)2 and their charge-transfer responses to proton impact is now discussed. The charge-transfer cross-sections that were obtained for the impingement of protons on the clusters perpendicular to the uracil plane averaged over the +z and −z directions are represented in Fig. 3. In the specific case of isomer A (which has the lowest energy), the two contributions are shown separately along with the resulting average, indicating relatively minor variations for these near-planar conformers. Similar uncertainties were found between B and B′ and between AC and AC′, while marginally larger variations were observed between C and C′.
Fig. 3

Charge-transfer cross-sections for the collisions of protons with uracil@(H2O)2 in the perpendicular orientation, averaged over the +z and −z directions. Collisions with the isomers A, B, C, and AC are shown by empty symbols, whereas collisions with A′, B′, C′, and AC′ are shown by full symbols. For the A conformer, the specific contributions of the +z and –z directions are shown as a colored zone together with the average value

Although they are all planar, the different conformers exhibit significant variations in charge transfer depending on the hydration region. The results obtained for the various structures can be grouped into two main sets depending on the magnitude of the charge-transfer cross-section. All conformers of the types A, AC, and C, irrespective of the relative orientations of the two water molecules away from the uracil plane, show charge-transfer cross-sections of the same order of magnitude as bare uracil. The charge-transfer process is somewhat more efficient for the A and A′ compounds, in particular at lower collision energies, but the mean cross-section values are roughly similar for A′, C–C′, and bare uracil across the entire collision energy domain, with the cross-section increasing slightly with increasing energy. For the AC conformers, the cross-section values are between those of A and C. Interestingly, the successive jumping of water molecules from A to C through AC, which energetically destabilizes the cluster, also decreases the charge-transfer cross-section in a stepwise fashion. From a chemical point of view, the A and C regions are very similar, and both can induce efficient hydrogen bonds with an N–H group as proton donor and an oxygen as proton acceptor. The main difference comes from the presence of a free C=O bond in the C geometry, which happens to be a destabilizing factor and is found to slightly reduce the charge-transfer efficiency. However, the effect is quite minor. Conversely, isomers from the B family display approximately one magnitude higher cross-sections than that of bare uracil, which remain almost constant with increasing collision energy. In these higher-energy conformers, the presence of equidistant C=O and N–H bonds thus seems to favor charge transfer despite destabilizing the system. The process is clearly very sensitive to the structure of the target, and may be employed as a tool for structural analysis.

2-Aminooxazole@(H2O)2

Interest has recently focused on 2-aminooxazole in the context of prebiotic chemistry. This molecule has been suggested as a possible key intermediate in an efficient reaction sequence leading to the formation of RNA under prebiotic conditions [41]. The chemical mechanism—which initially involves simple molecules such as cyanamide and glycolaldehyde—could be related to the chemistry of HCN, which is pivotal to our understanding of the processes that led to the development of life [42]. The microwave spectrum of 2-aminooxazole has been investigated theoretically [43], and photochemical processes have been proposed to explain the formation of the isolated molecule as well as hydrated clusters in astrophysical environments [44, 45, 46]. In this respect, we recently studied the evolution of 2-aminooxazole in microhydrated environments [24] and, notably, its response to proton impact, focusing on the effects of stepwise hydration rather than the details of the microhydrated structures. Microhydration effects are particularly relevant for prebiotic compounds which have been hypothesized to form in extraterrestrial environments—a theory supported by the discovery of aminoacids on meteorites [47]. However, it is unclear how they could survive in space and, in particular, how they could resist from proton damage. This motivated us to investigate radiation damage on aminooxazole in bare [19] and microhydrated [24] forms. Our results indicate that 2-aminooxazole@(H2O)n clusters exhibit a segregated structure with a water cluster on the amino side group, which leads to a weaker effect on the charge-transfer efficiency than observed for uracil@(H2O)n [23]. Smaller 2-aminooxazole@(H2O)n structures were found to be very similar to the corresponding microhydrated uracil compounds, with a locked-chain arrangement on the side of the biomolecule. However, these results are in significant conflict with those obtained in another recent theoretical study by Szabla and coworkers [46], who, despite predicting a clear segregation between the biomolecule and the cluster, found that this cluster preferred to be on the oxygen side instead of the nitrogen side. This provides additional justification for investigating the detailed microhydrated structures of dihydrated 2-aminooxazole and discriminating the various hydration patterns.

For the present system, nine low-energy structures were located using our combined force field/DFT approach. They are presented in Fig. 4. Following the notation already employed for uracil, we sorted them into the A, B, C, and AC families according to the relative positions of the two water molecules around the aminooxazole cycle. Most of these structures have the two water molecules near the plane of the biomolecule, except for A2, where they lie on both sides of the plane with an H-bond between them. For conformers with the water molecules near the plane of the oxazole, varying the orientations of the free hydrogen bonds on either the same side or opposite sides leads to the structures A–A′, B–B′, and AC–AC′. Finally, the conformer A1 has only one water that is hydrogen bonded to the biomolecule; the other water is bound to that water molecule. Quantum chemical data for these structures, including relative energies, electron affinities, and ionization energies, are given in Table 2. As also seen for uracil, the force field predicts remarkably similar relative binding energies and, more importantly, the same global minimum as indicated by the DFT calculations.
Fig. 4

Structures of the 2-aminooxazole@(H2O)2 clusters obtained by optimization at the DFT/M06-2X/aug-cc-pVTZ level

Table 2

Relative energies, vertical and adiabatic electron affinities (EA), and vertical and adiabatic ionization energies (IE) of 2-aminooxazole@(H2O)2, as obtained from quantum chemical calculations performed at the DFT/M06-2X/aug-cc-pVTZ level. The relative values obtained with the Amber ff99 force field are also reported

Structure

Energy (kcal mol−1) using Amber ff99

Energy (kcal mol−1) using DFT/M06-2X

Vertical EA (eV)

Adiabatic EA (eV)

Vertical IE (eV)

Adiabatic IE (eV)

A

0.0

0.0

−0.494

−0.487

8.550

7.870

A′

0.0

0.494

−0.498

−0.424

8.550

7.800

A1

3.993

6.208

−0.376

−0.336

8.362

7.548

A2

3.887

4.844

−0.409

−0.256

7.997

7.745

B

3.637

5.617

−0.412

−0.271

8.837

7.848

B′

3.637

5.833

−0.359

−0.306

8.834

7.838

C

3.751

6.029

−0.459

−0.288

8.515

7.620

AC

4.058

4.700

−0.427

−0.304

8.473

7.541

AC′

4.058

4.673

−0.429

−0.343

8.482

7.542

Clearly the conformers A and A′ are much lower in energy than all the other local minima, a result which supports the data we obtained in a previous investigation in which we explored a broader range of cluster sizes but fewer competing structures [24]. Placing the water molecules on the oxygen side is significantly energetically disfavored (by more than 5 kcal mol−1). As previously observed for uracil@(H2O)2, the cluster is stepwise destabilized when each water molecule jumps from A to C. More generally, the formation of simultaneous hydrogen bonds with the nitrogen of the ring and the amino group appears to be energetically preferred, although we note that the near-planar configurations (A and A′) are much more favorable than the out-of-plane arrangement A2, with the single-bonded conformer A1 being (unsurprisingly) even less stable. Finally, moving the water molecule connected to the amino group in A–A′ to the opposite carbon in the cycle in B–B′ destabilizes the cluster by more than 5.5 kcal mol−1, which again emphasizes the importance of the nitrogen component of the biomolecule.

The ionization energies given in Table 2 show similar behavior to that presented in Table 1 for uracil@(H2O)2 clusters. Despite being about 1 eV lower than for dihydrated uracil, the ionization energies of 2-aminooxazole@(H2O)2 also show rather low sensitivity to the conformer, as well as a comparable difference of about 0.7 eV between the vertical and adiabatic IEs. These results suggest that there is rearrangement upon the removal of an electron, as also noted for uracil@(H2O)2 clusters. The electron affinity depends also only weakly on the conformer of 2-aminooxazole@(H2O)2, and the adiabatic and vertical values are about the same. The addition of one electron therefore does not induce major rearrangement, which indicates the presence of a diffuse electron cloud at the periphery of the cluster. This behavior contrasts markedly with our findings for uracil@(H2O)2 clusters, which were found to accommodate the extra electron very efficiently. This difference is likely related to the presence of the amino group on the five-membered ring of oxazole. Indeed, since the presence of this chemical group was found to lead to segregated structures with a water cluster on the side of the amino group [24], it may also stabilize a peripheral extra electron, which would explain the contrasting trends.

The IR spectra obtained for dihydrated 2-aminooxazole in the double harmonic approximation are represented in Fig. 5. As also seen for uracil, spectra from the same structural families generally have similar patterns, especially in the stretching region near 3000 cm−1. The far-IR domain does not appear to be much useful for distinguishing among the present structures, especially as we anticipate that finite temperature anharmonicities will greatly alter the spectra that are actually observed for these low-frequency modes. As expected, the markedly different hydrogen-bonding networks of conformers A1 and A2 show greater differences from A than from A′. The spectra of conformers B and B′ are the most similar to the spectrum of the bare biomolecule, which is consistent with the absence of any hydrogen bond involving the amino group from the three systems. The spectrum of conformer C definitely differs from the spectra of A and A′; these spectral differences could be used to discriminate the various conformers and, especially, solve the discrepancy between our calculations and the results of Szabla and coworkers [46].
Fig. 5

Harmonic IR spectra of 2-aminooxazole@(H2O)2 compounds calculated at the DFT/M06-2X/aug-cc-pVTZ level of theory in the far- and mid-IR ranges. The spectrum of the bare 2-aminooxazole molecule is shown at the bottom

A comparative analysis of the 2-aminooxazole@(H2O)2 conformers was also carried out in relation to proton-induced charge transfer, focusing only (as also done for dihydrated uracil) on collisions perpendicular to the biomolecular cycle and averaging the cross-sections in the two directions +z and −z. It is more important to account for both orientations here than for uracil, because the amino side group leads now to a nonplanar molecule.

The charge-transfer cross-sections for the nine structures are displayed in Fig. 6, together with the corresponding results for bare 2-aminooxazole. Most of the conformers exhibit the same behavior already noted for the bare molecule [19], with a similar smooth increase observed throughout the entire energy domain. The A2 isomer is an exception—the cross-section decreases at higher energies—but this structure is significantly less stable and has a very specific H-bond topology that causes this behavior. Conformers of the A and AC clusters exhibit charge-transfer cross-sections that do not depend significantly on the orientation of the O–H bond; the values between A and A′ and between AC and AC′ are very similar. For the B family on the C=C side, markedly stronger variations are seen between the B and B′ conformers despite them being very close in energy. The magnitudes of the charge-transfer cross-sections for the bare molecule appear to lie roughly in between the values obtained for the various 2-aminooxazole@(H2O)2 conformers. Therefore, on average, the water chain appears to only weakly influence the response of the molecule to proton impact, as it is mostly localized on the side of the amino group.
Fig. 6

Charge-transfer cross-sections for collisions of protons with 2-aminooxazole@(H2O)2 in the perpendicular orientation. Full (empty) circles refer to conformers of the families A, B, C, and AC (A′, B′, C′ and AC′). All results are averaged over the values obtained for collisions in the +z and –z directions

As the temperature increases, populations of the higher-energy conformers should become more prominent. As previously noted for uracil clusters, stepwise jumping of a water molecule from region A to region C should therefore occur for 2-aminooxazole. Also similar to uracil, this thermal excitation leads to a decrease in the charge-transfer cross-section. However, in contrast to uracil, the two-water chain on the oxygen side promotes charge transfer. The presence of the hydrogen-bond chain between the ring oxygen and the amine group thus appears to be a favorable destabilizing factor. This is consistent with the mechanism of electron-driven proton transfer along H2O chains proposed by Szabla and coworkers [46].

Concluding remarks

Microhydrated biomolecules bridge the gap between isolated compounds in the gas phase and their counterparts in a biological medium. The insight attained by studying stepwise hydration with a controlled number of water molecules is particularly useful for disentangling the roles of solute–solvent and solvent–solvent interactions, although extrapolation to the bulk limit is never straightforward. The present work was performed to theoretically investigate two small biomolecules of relevance to RNA chemistry, namely uracil and 2-aminooxazole (recently suggested to be a prebiotic precursor), both of which were placed in contact with just two water molecules so that a near-exhaustive exploration of their energy landscapes could be achieved. The stable structures generated using density-functional theory with the M06-2X functional and the triple-zeta basis set aug-cc-pVTZ have the water chain localized at the side of each molecule, and contrast with earlier findings present in the literature, at least in the case of 2-aminooxazole. There is generally a marked correlation between the hydration pattern, the electronic properties (the electron affinity and ionization potential in particular), and the IR spectrum, although it is unclear whether the vibrational response could be used to experimentally characterize the most stable conformers. In many cases, the localization of the water molecules in the biomolecular plane should also lead to some partial disorder, as the dangling water O–H bonds that are not connected to the biomolecule are relatively free to move between the sides of the plane.

The charge-transfer cross-sections following the impact of energetic protons were also calculated for collisions perpendicular to the biomolecular cycle and along the two possible directions. We generally found that the cross-section significantly depended on the hydrogen-bonding pattern, but that there were interesting differences between the two biomolecular examples chosen. In uracil, the addition of a couple of water molecules markedly increases the charge-transfer efficiency for all conformers. Conversely, 2-aminooxazole is more chemically heterogeneous, and charge transfer can be enhanced or disfavored depending on the conformer, with the net effect of the addition of water molecules being much weaker than for uracil.

The contrasting behaviors of the two systems considered here highlight the complexity of the charge-transfer processes involved in radiation damage and their interplay with the chemical structure of the complex involved. They also emphasize the need to carefully evaluate the diversity of conformers in the energy landscape, given that the most stable structure is essentially impossible to predict based on intuition only, and hydrogen-bonding patterns involving both the solute and solvent have similar orders of magnitude. The present work could be extended along several possible lines, starting with investigations of larger sizes and bulk icy surfaces [24], which are particularly relevant in the context of interstellar grains [48].

Notes

Acknowledgments

We acknowledge computational support from the CCIN2P3 in Villeurbanne and the CCRT/CINES/IDRIS under the allocation x2016087662 made by GENCI. The authors also acknowledge also support from the COST actions TD1308 Life-Origins, CM1204 XLIC, and CM1401 Astrochemical History.

References

  1. 1.
    Michael BD, O’Neill PD (2000) A sting in the tail of electron tracks. Science 287:1603–1604CrossRefGoogle Scholar
  2. 2.
    Boudaiffa B, Cloutier P, Hunting D, Huels MA, Sanche L (2000) Resonant formation of DNA strand breaks by low-energy (3 to 20 eV) electrons. Science 287:1658–1660CrossRefGoogle Scholar
  3. 3.
    Martin F, Burrow PD, Cai Z, Cloutier P, Hunting D, Sanche L (2004) DNA strand breaks induced by 0–4 eV electrons: the role of shape resonances. Phys Rev Lett 93:068101/4Google Scholar
  4. 4.
    de Vries J, Hoekstra R, Morgenstern R, Schlathölter T (2003) Charge driven fragmentation of nucleobases. Phys Rev Lett 91:053401/4CrossRefGoogle Scholar
  5. 5.
    Deng Z, Bald I, Illenberger E, Huels MA (2005) Beyond the Bragg peak: hyperthermal heavy ion damage to DNA components. Phys Rev Lett 95:153201/4CrossRefGoogle Scholar
  6. 6.
    González-Magaña O, Tiemens M, Reitsma G, Boschman L, Door M, Bari S, Lahaie PO, Wagner JR, Huels MA, Hoekstra R, Schlathölter T (2013) Fragmentation of protonated oligonucleotides by energetic photons and Cq+ ions. Phys Rev A 87:032702/13CrossRefGoogle Scholar
  7. 7.
    de Vries J, Hoekstra R, Morgenstern R, Schlathölter T (2002) Cq+-induced excitation and fragmentation of uracil: effects of the projectile electronic structure. J Phys B 35:4373–4382CrossRefGoogle Scholar
  8. 8.
    Sadr-Arani L, Mignon P, Chermette H, Douki T (2014) Theoretical and experimental strudy of the fragmentation of protonated uracil. Chem Phys Lett 605:108–114CrossRefGoogle Scholar
  9. 9.
    López-Tarifa P, Hervé du Penhoat MA, Vuilleumier R, Gaigeot MP, Tavernelli I, Le Padellec A, Champeaux JP, Alcamí M, Moretto-Capelle P, Martín F, Politis MF (2011) Ultrafast nonadiabatic fragmentation dynamics of doubly charged uracil in a gas phase. Phys Rev Lett 107:023202/4CrossRefGoogle Scholar
  10. 10.
    Sadr-Arani L, Mignon P, Chermette H, Abdoul-Carime H, Farizon B, Farizon M (2015) Fragmentation mechanisms of cytosine, adenine and guanine ionized bases. Phys Chem Chem Phys 17:11813–11826CrossRefGoogle Scholar
  11. 11.
    Moretto-Capelle P, Le Padellec A (2006) Electron spectroscopy in proton collisions with dry gas-phase uracil base. Phys Rev A 74:062705/5CrossRefGoogle Scholar
  12. 12.
    Coupier B, Farizon B, Farizon M, Gaillard MJ, Gobet F, de Castro Faria NV, Jalbert G, Ouaskit S, Carré M, Gstir B et al (2002) Inelastic interactions of protons and electrons with biologically relevant molecules. Eur Phys J D 20:459–468CrossRefGoogle Scholar
  13. 13.
    Alvarado F, Bernard J, Lin B, Brédy R, Chen L, Hoekstra R, Martin S, Schlathölter T (2008) Precise determination of 2-deoxy-D-ribose internal energies after keV proton collisions. ChemPhysChem 9:1254–1258Google Scholar
  14. 14.
    Bacchus-Montabonel MC, Tergiman YS (2012) Radiation damage on biomolecular systems: dynamics of ion induced collision processes. Comput Theor Chem 990:177–184CrossRefGoogle Scholar
  15. 15.
    Bacchus-Montabonel MC, Tergiman YS (2011) An ab-initio study of ion induced charge transfer dynamics in collision of carbon ions with thymine. Phys Chem Chem Phys 13:9761–9767CrossRefGoogle Scholar
  16. 16.
    Bacchus-Montabonel MC (2013) Ab-initio treatment of ion-induced charge transfer dynamics of isolated 2-deoxy-D-ribose. J Phys Chem A 118:6326–6332Google Scholar
  17. 17.
    Bacchus-Montabonel MC (2013) Looking at radiation damage on prebiotic building blocks. J Phys Chem A 117:14169–14175CrossRefGoogle Scholar
  18. 18.
    Bacchus-Montabonel MC, Tergiman YS, Talbi D (2009) Ab-initio molecular treatment of charge-transfer processes induced by collision of carbon ions with 5-halouracil molecules. Phys Rev A 79:012710/7CrossRefGoogle Scholar
  19. 19.
    Bacchus-Montabonel MC (2015) Proton-induced damage on 2-aminooxazole, a potential prebiotic compound. J Phys Chem A 119:728–734CrossRefGoogle Scholar
  20. 20.
    Schermann JP (2008) Spectroscopy and modeling of biomolecular building blocks. Elsevier, AmsterdamGoogle Scholar
  21. 21.
    Wyttenbach T, Bowers MT (2009) Hydration of biomolecules. Chem Phys Lett 480:1–16CrossRefGoogle Scholar
  22. 22.
    Smyth M, Kohanoff J, Fabrikant II (2014) Electron-induced hydrogen loss in uracil in a water cluster environment. J Chem Phys 140:184313/7CrossRefGoogle Scholar
  23. 23.
    Bacchus-Montabonel MC, Calvo F (2015) Nanohydration of uracil: emergence of three-dimensional structures and proton-induced charge transfer. Phys Chem Chem Phys 17:9629–9633CrossRefGoogle Scholar
  24. 24.
    Calvo F, Bacchus-Montabonel MC, Clavaguéra C (2016) Stepwise hydration of 2-aminooxazole: theoretical insight into the structure, finite temperature behavior and proton-induced charge transfer. J Phys Chem A 120:2380–2389CrossRefGoogle Scholar
  25. 25.
    Wang J, Cieplak P, Kollman KA (2000) How well does a Restrained Electrostatic Potential (RESP) model perform in calculating conformational energies of organic and biological molecules? J Comput Chem 21:1049–1074CrossRefGoogle Scholar
  26. 26.
    Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, non-covalent interactions, excited states, and transition elements: two new functionals and systematic testing for four M06-class functionals and 12 other functionals. Theor Chem Accounts 120:215–241CrossRefGoogle Scholar
  27. 27.
    Salem L (1982) Electrons in chemical reactions: first principles. Wiley, New YorkGoogle Scholar
  28. 28.
    Bacchus-Montabonel MC, Talbi D, Persico M (2000) Quantum chemical determination of the rate coefficients for radiative association of CH3 + and H2. J Phys B 33:955–959CrossRefGoogle Scholar
  29. 29.
    Allan RJ, Courbin C, Salas P, Wahnon P (1990) State-selective effects in the differential cross section for electron capture from laser-excited sodium atoms by proton. J Phys B 23:L461–L466CrossRefGoogle Scholar
  30. 30.
    Bacchus-Montabonel MC, Tergiman YS (2006) Anisotropic effect in the charge transfer of Cq+ ions with uracil. Phys Rev A 74:054702/4CrossRefGoogle Scholar
  31. 31.
    Linguerri R, Hochlaf M, Bacchus-Montabonel MC, Desouter-Lecomte M (2013) Characterization of the MgO2+ dication in the gas phase: electronic states, spectroscopy and atmospheric implications. Phys Chem Chem Phys 15:824–831CrossRefGoogle Scholar
  32. 32.
    Werner HJ, Knowles PJ (2012) MOLPRO: a package of ab initio programs (version 2012.1). H.J. Werner/P.J. Knowles, Univ. Stuttgart/Cardiff Univ.Google Scholar
  33. 33.
    Nicklass A, Dolg M, Stoll H, Preuss H (1995) Ab initio energy-adjusted pseudopotentials for the noble gases Ne through Xe: calculation of atomic dipole and quadrupole polarizabilities. J Chem Phys 102:8942–8952Google Scholar
  34. 34.
    Bene E, Vibók Á, Halász G, Bacchus-Montabonel MC (2008) Ab initio molecular treatment of charge transfer processes induced by collision of C2+ ions with the OH radical: a linear approach. Chem Phys Lett 455:159–163CrossRefGoogle Scholar
  35. 35.
    Shishkin OV, Gorb L, Leszczynski J (2000) Modeling of the hydration shell of uracil and thymine. Int J Mol Sci 1:17–27CrossRefGoogle Scholar
  36. 36.
    Gaigeot MP, Ghomi M (2001) Geometrical and vibrational properties of nucleic acid constituents interacting with explicit water molecules as analyzed by density functional theory calculations. 1. Uracil + n wH2O (n w = 1,7). J Phys Chem B 105:5007–5017Google Scholar
  37. 37.
    Close DM, Crespo-Hernandez CE, Gorb L, Leszczynski J (2005) The influence of microhydration on the ionization energy thresholds of uracil and thymine. J Phys Chem A 109:9279–9283CrossRefGoogle Scholar
  38. 38.
    Kim S, Schaefer J III (2006) Effects of microsolvation on uracil and its radical anion: uracil. (H2O)n (n = 1-5). J Chem Phys 125:144305/9Google Scholar
  39. 39.
    Li X, Wang H, Bowen KH (2010) Photoelectron spectroscopic study of the hydrated nucleoside anions: uridine-(H2O)n = 0–2, cytidine-(H2O)n = 0–2, and thymidine-(H2O)n = 0,1. J Chem Phys 133:144304/5Google Scholar
  40. 40.
    Bachrach SM, Dzierlenga MW (2011) Microsolvation of uracil and ist conjugate bases: a DFT study of the role of solvation on acidity. J Phys Chem A 115:5674–5683Google Scholar
  41. 41.
    Powner MW, Gerland B, Sutherland JD (2009) Synthesis of activated pyrimidine ribonucleotides in prebiotically plausible conditions. Nature 459:239–242CrossRefGoogle Scholar
  42. 42.
    Sutherland JD (2016) The origin of life—out of the blue. Angew Chem Int Ed 55:104–121CrossRefGoogle Scholar
  43. 43.
    Møllendal H, Konovalov A (2010) Microwave spectrum of 2-aminooxazole, a compound of potential prebiotic and astrochemical interest. J Phys Chem A 114:2151–2156CrossRefGoogle Scholar
  44. 44.
    Szabla R, Šponer JE, Šponer J, Góra RW (2013) Theoretical studies of the mechanism of 2-aminooxazole formation under prebiotically plausible conditions. Phys Chem Chem Phys 15:7812–7818Google Scholar
  45. 45.
    Szabla R, Tuna D, Góra RW, Šponer J, Sobolewski AL, Domke W (2013) Photochemistry of 2-aminooxazole, a hypothetical prebiotic precursor of RNA nucleotides. J Phys Chem Lett 4:2785–2788CrossRefGoogle Scholar
  46. 46.
    Szabla R, Šponer J, Góra RW (2015) Electron-driven proton transfer along H2O wires enables photorelaxation of πσ* states in chromophore-water clusters. J Phys Chem Lett 6:1467–1471CrossRefGoogle Scholar
  47. 47.
    Cooper GW, Onwo WM, Cronin JR (1992) Alkyl phosphoric acids and sulfonic acids in the Murchison meteorite. Geochim Cosmochim Acta 56:4109–4115CrossRefGoogle Scholar
  48. 48.
    Karssemeijer LJ, Ioppolo S, van Hemert MC, van der Avoird A, Allodi MA, Blake GA, Cuppen HM (2014) Dynamics of CO in amorphous water ice environments. Astrophys J 781:16/15CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2016

Authors and Affiliations

  1. 1.Univ Lyon 1, Université Claude Bernard Lyon 1, CNRS, Institut Lumière MatièreVilleurbanneFrance
  2. 2.Laboratoire Interdisciplinaire de Physique, Université Grenoble 1 and CNRS UMR5588GrenobleFrance

Personalised recommendations