Monophosphate Anion - American Chemical Society

Jan 30, 2009 - Instituto de Ciencia Molecular, UniVersitat de Vale`ncia, Apartado 22085, ... Fundació General de la UniVersitat de Vale`ncia, Plaza d...
1 downloads 0 Views 322KB Size
J. Phys. Chem. B 2009, 113, 2451–2457

2451

Determination of the Electron-Detachment Energies of 2′-Deoxyguanosine 5′-Monophosphate Anion: Influence of the Conformation Mercedes Rubio,*,†,‡ Daniel Roca-Sanjua´n,† Luis Serrano-Andre´s,† and Manuela Mercha´n† Instituto de Ciencia Molecular, UniVersitat de Vale`ncia, Apartado 22085, ES-46071 Valencia, Spain, and Fundacio´ General de la UniVersitat de Vale`ncia, Plaza del Patriarca, 4-1, ES-46002 Valencia, Spain ReceiVed: July 10, 2008; ReVised Manuscript ReceiVed: December 18, 2008

The vertical electron-detachment energies (VDEs) of the singly charged 2′-deoxyguanosine 5′-monophosphate anion (dGMP-) are determined by using the multiconfigurational second-order perturbation CASPT2 method at the MP2 ground-state equilibrium geometry of relevant conformers. The origin of the unique low-energy band in the gas phase photoelectron spectrum of dGMP-, with maximum at around 5.05 eV, is unambiguously assigned to electron detachment from the highest occupied molecular orbital of π-character belonging to guanine fragment of a syn conformation. The presence of a short H-bond linking the 2-amino and phosphate groups, the guanine moiety acting as proton donor, is precisely responsible for the pronounced decrease of the computed VDE with respect to that obtained in other conformations. As a whole, the present research supports the nucleobase as the site with the lowest ionization potential in negatively charged (deprotonated) nucleotides at the most stable conformations as well as for B-DNA-like type arrangements, in agreement with experimental evidence. 1. Introduction Determination of the ionization energies of nucleic acid components is a fundamental task to understand the initial effect of ionizing radiation and oxidizing agents on DNA as well as charge-transfer processes (electron-hole migration) following DNA irradiation.1 The guanine nucleobase was long identified as the lowest-energy ionization site of DNA.2 Analysis of the relative distribution of the ion radicals formed in γ-irradiated DNA at low temperatures showed the predominance of the radical cation of guanine in the generated species.3-5 The result is consistent with the ionization potentials determined for isolated nucleobases, in which guanine displayed the lowest value according to experimental and theoretical data.6 A study performed by Yang et al.7 also provided experimental evidence that guanine is the site with the lowest ionization potential (IP) in gas-phase oligonucleotides. These authors recorded the photodetachment photoelectron spectra (PD-PES) of singly charged mono-, di-, and trinucleotide anions in the gas phase and used density functional theory (DFT) to interpret the results. A weak and well-separated band appears in the photoelectron spectrum of 2′-deoxyguanosine 5′-monophosphate anion (dGMP-) at electron-binding energies much lower than for the other mononucleotides.7 A similar low-energy band is recorded in all the spectra of di- and trinucleotides containing guanine nucleobase.7 As we have shown in a recent paper,8 a reliable interpretation of the observed spectra requires performing calculations at higher theoretical level than that employed by Yang et al.7 In particular, we recently applied the multiconfigurational second-order perturbation method known as CASPT29 to compute the vertical electrondetachment energies (VDEs) of 2′-deoxythymidine 5′-monophosphate anion (dTMP-). According to the computed VDEs, the removal of an electron from the highest occupied π-orbital of * Towhomcorrespondenceshouldbeaddressed.E-mail:Mercedes.Rubio@ uv.es. † Universitat de Vale`ncia. ‡ Fundacio´ General de la Universitat de Vale`ncia.

thymine moiety has been characterized as the lowest-energy oxidation process in dTMP-,8 in contrast to the results reported by Yang et al.7 which ascribe such process to the phosphate group of the nucleotide in all species except for systems containing guanine. Considering that thymine has the highest vertical ionization potential of all nucleic acid bases,6 we concluded that in all DNA nucleotides the most favorable oxidation site is related to the nucleobase.8 The current research aims to confirm such conclusion through the theoretical determination of the low-lying vertical electron-detachment energies of dGMP-, which additionally displays specific spectral features. Assignments of the recorded PD-PES of deoxyribonucleotide anions have subsequently been proposed by Ortiz and coworkers10 on the basis of the electron-detachment energies computed using the partial third-order (P3) self-energy approximation of the electron propagator method. This study10 has pointed out that the molecular conformation of the nucleotide can have a decisive role in the interpretation of the gas-phase experimental data. In fact, the low-lying energy band measured in the gas-phase PD-PES of dGMP- has been assigned to the removal of the electron from a π-orbital of guanine with a syn orientation in dGMP-, which was predicted to be the most stable conformation according to DFT geometry optimizations.7,10-12 Such conformation is different from those characterized for the remaining deoxyribonucleotides anions10 and from that occurring in the B-form of DNA, the form prevalent under physiological conditions. Several conformers of purinic deoxyribonucleotide anions have been considered in a recently reported determination of VDEs using the P3 approach.11 The influence of the nucleotide conformation on the electron-donating properties is therefore an important aspect to treat which will be deeply analyzed here for the particular case of dGMP-. As far as we know, and apart from the already mentioned studies by Ortiz and co-workers with the P3 method,10,11 only some theoretical studies at the DFT level of calculation have been reported in the past years on the ionization energies of DNA nucleosides13 and deprotonated nucleotides,14 focusing just on the lowest

10.1021/jp806105h CCC: $40.75  2009 American Chemical Society Published on Web 01/30/2009

2452 J. Phys. Chem. B, Vol. 113, No. 8, 2009 oxidized state. It is also worth mentioning the early research performed by Le Breton and co-workers who used an approach based on combining experimental and theoretical data to determine the ionization energies of isolated15,16 and solvated17,18 nucleotides. In the present paper, we report results from high-level ab initio calculations on dGMP-. The vertical electron-detachment energies have been computed for several dGMP- conformers by using the CASPT2 method.9 In addition to the most stable conformations in vacuo determined from ab initio calculations including electron correlation, a B-DNA-like conformation has been considered, as we did in our previous study on dTMP-.8 In the light of the present results, a reliable interpretation of the gas-phase PD-PES data is provided. Moreover, the effect of the negatively charged sugar-phosphate group in the ionization properties of guanine nucleobase, and specially how the relative orientation of both nucleotide building blocks affects these properties, are examined here through the comparison of the computed data for the different conformers. As shown below, the relative stability of the dGMP- conformers considered is greatly determined by the hydrogen-bond interactions occurring in the nucleotide, which can account for the differences found in the computed VDEs of the conformers. Recent experimental and theoretical studies on the redox properties of the Watson-Crick base pairs13,19,20 and DNA nucleosides13 have revealed that the oxidation potential of guanine and adenine decreases upon pairing with the complementary nucleobase and that the ionization energies are dependent on the nucleoside conformation. The present study is expected to give additional insight into the different factors affecting the ionization energies of the nucleobases in DNA. In line with the conclusions reached in our study on dTMP-,8 the current research also supports the nucleobase as the lowest-energy oxidation site in gas-phase negatively charged nucleotides, refuting therefore the location in the phosphate group previously proposed.7,16 2. Calculation Details As a first step of the study, the geometry of the ground state of dGMP- was fully optimized at the MP2 level of theory using Dunning’s correlation consistent cc-pVDZ basis set with the Gaussian 03 program.21 It should be remarked here that it is extraordinarily difficult and out of the scope of the current study to perform an exhaustive conformational search for dGMP- at the selected computational level because of the large number of possible conformers; thus, for instance, 92 conformers have been characterized at the DFT level for a nucleoside as thymidine.22 Accordingly, several initial conformations were employed for the geometry optimizations of dGMP- taking into account the results from previous conformational studies10,12,23,24 and the geometric parameters in DNA. Figure 1 displays the chemical structure and atom numbering of dGMP- indicating also the torsion angles mainly responsible for the conformational flexibility of the nucleotide together with the inherent conformational behavior of the sugar molecule.25 From the set of optimized geometries, four conformers were selected to perform the calculation of the corresponding VDEs: a B-DNA-like conformation which was obtained by using as starting point in the geometry optimization the values for the torsion angles χ, γ, and β of B-DNA reported by Leulliot et al.,26 and the three lowest-energy conformers characterized, which have a syn orientation of the nucleobase with respect to the sugar, unlike the B-DNA conformer. A more detailed analysis of the conformations considered will be given in the next section. The VDEs of each obtained dGMP- conformer were computed at the CASPT2 level of theory9 using the MOLCAS,

Rubio et al.

Figure 1. Chemical structure and atom numbering for dGMP-. Torsion angles are also shown.

version 6.3, quantum chemistry software.27-29 The CASPT2 method has been successfully applied to a large number of chemical problems;30-35 the determination of the ionization potentials and electron affinities of DNA and RNA nucleobases6,36 and dTMP-8 are particularly relevant for the present work. The complete active space self-consistent field (CASSCF) wave functions, which constitute the reference function in the secondorder perturbation CASPT2 treatment,9 were based on an active space of 13 active orbitals with 16 active electrons for dGMPand 15 for neutral dGMP. The active orbitals correspond to seven π-orbitals (4π + 3π*) of guanine fragment and six orbitals localized on the phosphate group which include the four lonepair orbitals of the oxygen atoms (O6 and O7) (see Figure 1). Test calculations found that the relative values of the computed VDEs were stable within 0.1 eV with respect to a reduction of the active space. An exception was conformer sConf2, which, in order to include the four lone-pair orbitals, required the extension of the active space to 14 orbitals and 18/17 electrons, adding one more occupied π orbital. The reason is the lower energy in sConf2 of the transitions from the nucleobase π structure. The 1s electrons of C, N, O, and P atoms were kept frozen at the SCF level of theory in all calculations. A singleroot CASSCF calculation was carried out for the ground state of each dGMP- conformer while the state-average procedure was used for computing the low-lying states of the corresponding neutral dGMP. All the electrons except the core ones were correlated in the CASPT2 calculations. The newly developed IPEA modified zeroth-order Hamiltonian37 was used to perform CASPT2 calculations. This new approximation corrects a systematic error of the CASPT2 method affecting the processes where the number of paired electrons is changed. Furthermore, in order to minimize the effect of weakly interacting intruder states, the imaginary level-shift technique (0.2 au) was also applied.38 The CASPT2 calculations employed a generally contracted basis set of atomic natural orbitals (ANOs) built from P(13s10p4d)/C,N,O(10s6p3d)/H(7s) primitive sets39 with a P[4s3p1d]/C,N,O[3s2p1d]/H[2s] contraction scheme, resulting in 352 contracted functions. This basis set represents a good compromise between accuracy and computational cost, as shown by the previous study on dTMP-.8 Additional enlargement of the basis set for the calculation of the lowest IP of guanine6 showed differences smaller than 0.05 eV. 3. Results and Discussion 3.1. Conformational Analysis. The optimized MP2/cc-pVDZ conformations for dGMP- considered in the present study are shown in Figure 2. For an easier characterization, the relevant optimized parameters are collected in Table 1 together with the relative energies computed at the MP2/cc-pVDZ level. As stated

VDEs of 2′-Deoxyguanosine 5′-Monophosphate Anion

Figure 2. Optimized MP2/cc-pVDZ conformations for dGMP-. See text for labeling.

above, and it can be noted in Figure 2, the most striking difference between the B-DNA-like conformation (denoted as aConf1) and the three characterized lowest-energy conformers concerns the relative orientation of the guanine and deoxyribose moieties, which is determined by the value of the torsion angle around the glycosyl bond, χ (see Table 1). The so-called anti orientation, present in the A and B forms of DNA and consequently in aConf1, situates the nucleobase away from the sugar and it is, in general, preferentially adopted (see, for instance, refs 24 and 40 and references therein). However, the three lowest-energy conformers of dGMP- have the nucleobase in a syn orientation, namely, placed over the sugar. The preference of guanine for a syn orientation in 5′-nucleotides is well documented7,10,12,24 and it has been related to a hydrogen bond type interaction between the 2-amino group of guanine and the phosphate group.24 Indeed, the anti conformation has been predicted to be the most stable in guanine nucleosides23,24 as well as in all deoxyribonucleotides anions except for dGMP-.10 Several studies have reported the lowest-energy conformer of dGMP- obtained from DFT geometry optimizations.7,10-12 In line with the present results, all these geometries agree with the syn orientation of guanine but differ in the conformation of the sugar ring. The earlier reported structures were considered in our MP2 geometry optimizations. Thus, the conformer named as sConf2 is closely related to that reported by Zakjevskii et al.10 Differences less than 8° are, for instance, found between the torsion angles determining the type of conformer characterized for both structures.

J. Phys. Chem. B, Vol. 113, No. 8, 2009 2453 The pseudorotation phase angle (P) and the puckering amplitude (τm) are used to describe the conformation of the furanose ring, which is also related to the value of the torsion angle δ (see Table 1).25 A similar conformation for the sugar, corresponding to the south (S) sugar puckering, is found in the four dGMP- conformers considered, as shown by the δ and P values included in Table 1. In particular, the P value for sConf1 and sConf2 belongs to the known as C2′-endo conformation (144° e P e 180°), one of the preferred sugar conformational states according to crystallographic data in nucleosides, nucleotides, and related compounds.25,40 On the other hand, the four conformers have also in common a gauche+ (g+) orientation (30° e γ e 80°)40 of the oxygen atom O5′ with respect to the sugar, a conformational feature depending on the value of the torsion angle γ (see Table 1). It has been suggested that a syn orientation of the nucleobase favors a value of the torsion angle γ around 180° (trans orientation), as observed in the Z-form of DNA (see, for instance, refs 23 and 24). However, all our attempts to obtain a conformer with a χ ) syn/γ ) trans orientation failed and led to one of the χ ) syn/γ ) g+ conformers as compiled in Table 1. Furthermore, we could not find either a structure with a C3′-endo sugar conformation,25 as described in other studies.7,12 In accordance with the current research, the south sugar puckering and the γ ) g+ orientation have been also found to yield conformations lower in energy in guanine nucleosides with a syn orientation of the nucleobase.24 As shown in Figure 2, the most stable conformer characterized, sConf1, has an orientation of the phosphate group which enables the formation of two hydrogen bonds with the guanine moiety: one involving the 2-amino group, as expected from previous considerations,24 and the other connecting the hydrogen atom of the phosphate group and the N3 atom of guanine. The latter H-bond is probably quite weak considering the computed distance between the H and N3 atoms (2.288 Å) and the deviation from a linear arrangement of the three atoms involved; however, its presence can explain the lower energy computed for sConf1 compared to sConf2 which lacks this interaction (see Figure 2). It should also be noted that the hydrogen bond involving the 2-amino group is expected to be stronger in sConf1, as suggested by the slightly shorter O6 · · · H distance (see Table 1). Thus, sConf2 is predicted to be 1.31 kcal/mol higher in energy than sConf1. As can be seen in Table 1 and Figure 2, both conformers, sConf1 and sConf2, differ mainly in the orientation of the phosphate group, which is defined by the torsion angles R and β. The conformation sConf3, computed 3.04 kcal/mol less stable than sConf1 has, however, the same orientation of the phosphate moiety as sConf1 (see β and R values in Table 1), but the torsion angle χ is significantly different (109.44° in sConf3 vs 78.20° in sConf1). Consequently, the two H-bond type interactions described in sConf1 are also found in sConf3, although that involving the 2-amino group is clearly weaker in sConf3, as indicated by the O · · · H distance (0.2 Å longer) and the angular deformation (∠N2sH · · · O ) 147.8°). Accordingly, the lengthening noted for the corresponding N2-H distance is smaller in sConf3 compared to sConf1 and sConf2 (see Table 1). On the other hand, the C2-N2 bond distance in the guanine fragment seems to show also the weakness of the interaction since its value in sConf3 is practically the same as in aConf1 (1.4088 vs 1.4083 Å, respectively), where the 2-amino group is in the opposite direction with respect to the phosphate fragment of the molecule (see Figure 2), and different from the bond length optimized for sConf1 and sConf2. Moreover, the 2-amino group is strongly pyramidalized in sConf3, almost perpendicular to the plane of

2454 J. Phys. Chem. B, Vol. 113, No. 8, 2009

Rubio et al.

TABLE 1: Relevant Geometrical Parameters Optimized at the MP2/cc-pVDZ Level for the Conformations of dGMPConsidered

torsion angles/deg χ (O4′-C1′-N9-C4) γ (C3′-C4′-C5′-O5′) δ (O3′-C3′-C4′-C5′) β (P-O5′-C5′-C4′) R (O8-P-O5′-C5′) sugar conformation P/deg τmax/deg selected distances/Å N9-C1′ C4-C5 C2-N2 N2-H O5′-N3 H-bond interactions C8-H8 · · · O6-(P) N2-H · · · O6-(P) O8-H · · · N3 ∆EMP2/kcal/mol ∆ECASPT2/kcal/mol

aConf1

sConf3

sConf2

sConf1

anti/S/g+

syn/S/g+

syn/C2′-endo/g+

syn/C2′-endo/g+

-81.29 55.94 155.34 -102.00 -100.64

109.44 50.08 153.74 -115.99 105.35

70.69 39.69 150.59 166.58 79.10

78.20 60.85 149.89 -114.31 104.41

207.52 37.09

221.58 38.70

163.16 40.73

170.58 38.39

1.4610 1.4047 1.4083 1.0210 5.42

1.4718 1.4067 1.4088 1.0517 2.87

1.4474 1.4054 1.3680 1.0578 3.24

1.4576 1.4082 1.3694 1.0692 2.94

1.807 Å, 147.8° 2.073 Å, 155.1° 3.04 3.60

1.644 Å, 175.6° 5.423 Å 1.31 0.87

1.603 Å, 175.8° 2.288 Å, 144.3° 0.0 0.0

2.107 Å, 173.2° 5.28 6.63

TABLE 2: Computed CASPT2 Vertical Electron-Detachment Energies (eV) for the dGMPConformers aConf1 +

sConf3 +

sConf2

sConf1 +

state

anti/S/g

syn/S/g

syn/C2′/g

syn/C2′/g+

π (G) n (phosphate) n (phosphate) n (phosphate) π

5.81 6.34 6.51 6.84 7.29

5.60 6.25 6.40 7.04 7.28

5.01 6.42 6.63 6.89 6.95

5.29 6.31 6.42 6.99 7.20

the guanine rings, probably as a way to increase the interaction with the phosphate moiety. This pyramidal character is also present in all remaining conformers but in a lesser extent. As a whole, the comparison among the four conformers characterized reveals differences at most of 0.02 Å in the optimized bond distances of the guanine moiety, the largest deviations taking place in the vicinity of the C2 atom. The same result is found when comparing the bond lengths of the phosphate fragment of the different conformers. As regards the aConf1 structure, the most relevant in the DNA context because of its similarity to B-DNA, it displays a C8sH · · · OsP hydrogen bond type interaction,41 which is, however, weaker than the hydrogen bond involving the 2-amino group. This conformer is thus found 5.28 kcal/mol higher in energy than sConf1. The comparison of the present MP2/cc-pVDZ findings with the DFT/B3LYP/6-311G(d,p) results from the conformational study of dGMP- performed by Zakjevskii et al.11 shows that, despite the similarity of some of the conformers characterized, the relative energies are strongly affected by the level of theory. For instance, the computed MP2/ cc-pVDZ energy difference between aConf1 and sConf2 (3.96 kcal/mol) is less than half of the value obtained at the DFT level for the equivalent structures (1 and 3 in ref 11). 3.2. Vertical Electron-Detachment Energies. The computed vertical electron-detachment energies for the four considered conformers of dGMP- are collected in Table 2. In all of them, four states of neutral dGMP are calculated vertically in the energy range 5.0-7.0 eV above the ground state of the corresponding dGMP- conformer. Moreover, as shown in Table

2 and revealed by the analysis of the CASSCF wave functions, the lowest VDE is related in all cases to the removal of an electron from the highest occupied π orbital of the guanine moiety, in agreement with the P3 results reported by Zakjevskii et al.11 for the equivalent structures. Therefore, and as an important conclusion, the present results indicate that the characterization of the nucleobase as the lowest-energy oxidation site in dGMP- does not concern only the syn conformers, as it was suggested from the DFT calculations performed by Yang et al.,7 but it is a general feature of the nucleotides. As also noted in Table 2, in the studied conformers there is an energy gap between 0.53 and 1.41 eV between the lowest VDE event and the immediately higher one, which is ascribed to the phosphate group in the nucleotide. In accordance with the conclusions derived from our study of dTMP-,8 the present findings clearly establish that the π-orbital of the guanine fragment is the more easily oxidized site in isolated dGMP- in the most stable characterized syn conformers, as well as in the B-DNA like conformation. As can be deduced from the results listed in Table 2, the computed VDEs for each conformer differ mainly in the value obtained for the lowest state of the corresponding neutral dGMP conformer which is of π character, as stated above. The largest value, 5.81 eV, is found in the B-DNA-like conformation, aConf1, whereas the lowest VDEs for the most stable conformers, sConf1 and sConf2, in isolated conditions shows a decrease of 0.8 and ∼0.5 eV, respectively, leading to the values 5.01 eV (sConf2) and 5.29 eV (sConf1). These changes can be qualitatively rationalized by the relative strengths of the intramolecular hydrogen-bond interactions existing in the different conformers, as determined from the nature and the geometrical parameters of the atoms involved. As shown in Table 1 and discussed above, the syn conformers have an H-bond involving the 2-amino group of the guanine moiety which acts as proton donor to the phosphate fragment. Such interaction is therefore expected to increase the electronic charge density of the guanine moiety, making easier in the syn conformers the electrondetachment process from the π-orbital of the nucleobase in comparison with aConf1. Such effect can be illustrated by a Mulliken population analysis of the CASSCF charge distribution

VDEs of 2′-Deoxyguanosine 5′-Monophosphate Anion for the ground state of dGMP-, which provides the following results on the guanine moiety: -0.006 for aConf1, -0.208 for sConf3, -0.243 for sConf2, and -0.256 for sConf1. These values correlate well with the relative strength of the H-bond qualitatively inferred from the N2sH · · · O6 distance. It is noted, however, that the lowest VDE computed for sConf1 is 0.28 eV larger than the one calculated for sConf2, despite the somewhat shorter N2sH · · · O6 distance present in the former (1.603 vs 1.644 Å). The presence of an H-bond interaction in sConf1 where the guanine moiety also acts as proton acceptor through the N3 atom could account for the increase obtained for the lowest VDE of sConf1 compared to that of sConf2. Similarly, since this type of hydrogen bond is also present in sConf3, the value computed for the lowest VDE of this conformer, 5.60 eV, is expected to be larger than the one we would have obtained in the absence of the O8sH · · · N3 interaction. Furthermore, the fact that the H-bond involving the 2-amino group is weaker in sConf3 than in the other two syn conformers considered (see above) also explains the larger VDE ascribed to the π-orbital for sConf3 compared to the corresponding values calculated for the remaining syn conformers. As a whole, the present results for the vertical electron-detachment energy related to the highest occupied π-orbital of the guanine moiety for the different conformations show that the ionization energy of guanine is modulated by the strength and type of hydrogen-bond interactions present in the conformer, in agreement with the results from recent theoretical studies on nucleosides13 and hydrogenbonded amino acid-guanine complexes.42 As shown in Table 2, and according to the present CASPT2 calculations, the three next states of the neutral dGMP conformers correspond to electron detachment from the oxygen (O6 and O7) lone-pair orbitals of the phosphate group. In each conformer they can be found within a 0.5-0.8 eV energy range (the largest difference occurring for the sConf3 conformer, 6.25-7.04 eV), whereas their positions differ by less than 0.1-0.2 eV in the different conformations. Moreover, the present values are close to the previous computed VDEs for the corresponding states of dTMP- in a B-DNA-like conformation (6.21, 6.58, and 6.94 eV).8 It should be noted here that, compared to the gas-phase experimentally determined VDEs of the H2PO4- molecule (5.06, 5.42, and 5.82 eV),43 the electron removal from the phosphate group of singly charged mononucleotide anions requires larger energies8,10,11 because of the different intramolecular interactions established between the building blocks of the nucleotide. For sConf2, the CASPT2 results situate the two lowest states obtained by electron removal from the phosphate group slightly higher in energy than in the remaining conformers. In principle, this energy increase could be attributed to the short hydrogen bond (H-bond) (1.644 Å) existing in sConf2 between the oxygen atom of the phosphate moiety and the 2-amino group of the guanine fragment. Consequently, we should also expect somewhat higher VDEs related to the phosphate group for sConf1 since this conformer has also the same short H-bond as sConf2, as shown in Table 1. According to the present CASPT2 calculations, such result is, however, not obtained. Instead, there is a noticeable similarity between the VDEs corresponding to the phosphate fragment computed for sConf1 and sConf3. A possible explanation could be the role played by the second H-bond interaction present in sConf1, which also occurs in sConf3, connecting the O-H part of the phosphate and the N3 atom of guanine. We can compare the present CASPT2 values with the P3 results reported by Zakjevskii et al.11 for structures equivalent to aConf1 and sConf2. Although a qualitative agreement is noted between the two types of calculations and

J. Phys. Chem. B, Vol. 113, No. 8, 2009 2455 they both predict an increase of the VDEs related to the phosphate fragment in sConf2 in comparison with their values in aConf1, the P3 results are systematically lower up to 0.4 eV, except for the lowest π(G) state of sConf2 which is placed at the same energy, 5.01 eV. The VDE corresponding to the next occupied π orbital of guanine moiety was estimated at the CASPT2 level to be situated in the energy range 7.0-7.3 eV for the different conformers, as shown in Table 2. This result suggests that the electron detachment from the lone-pair orbitals of the guanine fragment could also require energies lower than 7.3 eV since calculations on the guanine nucleobase yield vertical ionization potentials (VIPs) for these orbitals ∼0.5 eV lower than that corresponding to the second VIP of π character.6 Most probably the presence of states related to the lone-pair orbitals of guanine contributes to increase the density of states inferred from the experimental spectrum where a broad and continuous band is observed between 6.0 and 7.5 eV.7 We shall finally discuss the present results in relation to the available experimental data. As stated in the Introduction, Yang et al.7 recorded the gas-phase photodetachment photoelectron spectrum of dGMP-, yielding thus spectroscopic data directly comparable with the present computations in vacuo provided that the most stable conformers had been characterized. Therefore, such comparison also enables us to gain insight about the geometry of the lowest-energy conformers in the gas phase. The photoelectron spectrum of dGMP- shows a weak and distinct low-energy band with maximum at 5.05 ( 0.10 eV and a broad and continuous band, similar to the one recorded for the remaining deoxyribonucleotide anions, approximately in the energy range 6.0-7.5 eV.7 As can be deduced from the present findings, these spectroscopic features are consistent with the computed VDEs for the conformers sConf1 and sConf2, in line with the lowest-energy states determined for both structures and supporting therefore a similar conformation to sConf1 and/ or sConf2 as the preferred one in the gas phase. For deprotonated mononucleotides, the ion mobility data reported by Gidden and Bowers12 indicate that only one family of conformers is present in the gas phase. On the basis of molecular mechanics and DFT calculations, these authors have characterized the family of dGMP- conformers as having a C3′-endo conformation for the sugar part and a syn orientation of the nucleobase with the amino group hydrogen-bonded to the deprotonated phosphate. Except for the sugar conformation (see above), the present results are in accordance with the conclusions derived by Gidden and Bowers.12 Thus, the conformers sConf1 and sConf2, which differ mainly in the orientation of the phosphate group (see Figure 2 and Table 1), can be classified as belonging to the same family. It should also be reminded here that the MP2/cc-pVDZ computed energy difference between them is only 1.31 kcal/ mol (0.87 kcal/mol at the CASPT2 level) without including the zero-point vibrational energy correction, a term which could reduce such difference, approaching it to the thermal energy at room temperature (∼0.6 kcal/mol), as used in the experiments, where both conformers can be easily accessed.7,12 As it is clearly established from the present computed VDEs for the four conformers considered, the band peaking at ∼5.0 eV in the gasphase photoelectron spectrum of dGMP- arises from electron removal from the highest occupied π-orbital of guanine moiety in a syn orientation in the nucleotide and acting as proton donor in a short H-bond connecting the amino and phosphate groups. It is worth remarking that the latter feature is indispensable for decreasing the VDE ascribed to this π-orbital to a value close to 5.0 eV (in particular, 5.01 eV for sConf2 and 5.29 eV for

2456 J. Phys. Chem. B, Vol. 113, No. 8, 2009 sConf1), as shown by the results obtained for sConf3. On the other hand, according to the present calculations for sConf1 and sConf2, the broadband recorded at ∼6.0-7.5 eV in the gas phase PD-PES is assigned to electron removal from the lonepair orbitals of the oxygen atoms of the phosphate group. As pointed out above, the possibility that states related to electron detachment from the lone-pair orbitals of guanine fragment also appear in this energy range cannot be excluded. 4. Summary and Conclusions The vertical electron-detachment energies of the singly charged 2′-deoxyguanosine 5′-monophosphate anion have been computed at the CASPT2 level9 using a ANO-type basis set of split valence quality, including polarization functions in all heavy atoms. Calculations have been carried out for four different dGMPconformers which were determined from MP2/cc-pVDZ geometry optimizations and following experience gained in previous conformational studies.10,12,23,24 The relative stability of the conformers characterized has been rationalized on the basis of the intramolecular hydrogen-bond interactions present in the structure. In line with previous studies on nucleosides13 and amino acid-guanine complexes,42 the present results have also shown that such intramolecular H-bond interactions have an important effect on the electron-detachment energy ascribed to the HOMO π-orbital of the guanine moiety. Thus, this energy has been found to vary within 0.8 eV in the four dGMP- conformers examined. The VDEs corresponding to the phosphate fragment of the nucleotide are, however, less influenced by the strength and type (H donor or acceptor) of the H-bond interactions existing in the nucleotide, according to the present calculations. The origin of the low-energy band, peaking at 5.05 ( 0.10 eV and absent for the remaining deoxyribonucleotide anions,7 in the gas phase photoelectron spectrum of dGMP- has been thus understood: it is related to syn conformations having a short H-bond connecting the 2-amino and phosphate groups where the guanine moiety acts as proton donor. This H-bond gives rise to a considerable decrease of the VDE corresponding to the π-orbital of the guanine fragment compared to its value in the B-DNA-like conformer (aConf1) or in other syn conformations. It has been clearly established from the present results that the lowest-energy oxidation site in isolated dGMP- is the nucleobase fragment and not only for the most stable conformers, as proposed by Yang et al.,7 but also for a B-DNA-like conformation. Therefore, the present findings, together with our previous study on dTMP-,8 support the nucleobase as the site with the lowest IP in negatively charged (deprotonated) nucleotides with a B-DNA-like conformation, in line with the experimental data from γ-irradiated DNA.3-5 As expected from previous theoretical determinations,10,14-16 the CASPT2 VDEs for dTMP-8 and dGMP- also show that the negatively charged sugar-phosphate group reduces considerably the ionization potentials of the nucleobases (computed to have the vertical values 9.07 and 8.09 eV for thymine and guanine, respectively6), making them more similar (within 0.2 eV) in the nucleotides. This result is of relevance for understating the chargetransport mechanism in DNA, which may strongly depend on the differences between the oxidative and reductive properties of each of its building blocks.44 Acknowledgment. Financial support is acknowledged from projects CTQ2007-61260 and CSD2007-0010 Consolider-Ingenio in Molecular Nanoscience of the Spanish MEC/FEDER, Ajudes Investigacio´ UVEG 2006, and GV+FEDER funds, project GVAINF2007-051.

Rubio et al. Supporting Information Available: Cartesian coordinates for the optimized conformers. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Colson, A.-O.; Sevilla, M. D. In Computational Molecular Biology; Leszczynski, J., Ed.; Elsevier: Amsterdam, 1999; Vol. 8, pp 245-270. (2) Gra¨slund, A.; Ehrenberg, A.; Rupprecht, A.; Stro¨m, G. Biochim. Biophys. Acta 1971, 254, 172. (3) Boon, P. J.; Cullis, P. M.; Symons, M. C. R.; Wren, B. W. J. Chem. Soc., Perkin Trans 2 1984, 1393. (4) Sevilla, M. D.; Becker, D.; Yan, M.; Summerfield, S. R. J. Phys. Chem. 1991, 95, 3409. (5) Yan, M.; Becker, D.; Summerfield, S.; Renke, P.; Sevilla, M. D. J. Phys. Chem. 1992, 96, 1983. (6) Roca-Sanjua´n, D.; Rubio, M.; Mercha´n, M.; Serrano-Andre´s, L. J. Chem. Phys. 2006, 125, 084302, and references therein. (7) Yang, X.; Wang, X.-B.; Vorpagel, E. R.; Wang, L.-S. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 17588. (8) Rubio, M.; Roca-Sanjua´n, D.; Mercha´n, M.; Serrano-Andre´s, L. J. Phys. Chem. B 2006, 110, 10234. (9) Andersson, K.; Malmqvist, P.-Å.; Roos, B. O. J. Chem. Phys. 1992, 96, 1218. (10) Zakjevskii, V. V.; King, S. J.; Dolgounitcheva, O.; Zakrewski, V. G.; Ortiz, J. V. J. Am. Chem. Soc. 2006, 128, 13350. (11) Zakjevskii, V. V.; Dolgounitcheva, O.; Zakrewski, V. G.; Ortiz, J. V. Int. J. Quantum Chem. 2007, 107, 2266. (12) Gidden, J.; Bowers, M. T. J. Phys. Chem. B 2003, 107, 12829. (13) Crespo-Herna´ndez, C. E.; Close, D. M.; Gorb, L.; Leszczynski, J. J. Phys. Chem. B 2007, 111, 5386. (14) Hou, R.; Gu, J.; Xie, Y.; Yi, X.; Schaefer, H. F., III J. Phys. Chem. B 2005, 109, 22053. (15) Tasaki, K.; Yang, X.; Urano, S.; Fetzer, S.; LeBreton, P. R. J. Am. Chem. Soc. 1990, 112, 538. (16) Kim, H. S.; Yu, M.; Jiang, Q.; LeBreton, P. R. J. Am. Chem. Soc. 1993, 115, 6169. (17) Fernando, H.; Papadantonakis, G. A.; Kim, N. S.; LeBreton, P. R. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 5550. (18) Kim, N. S.; LeBreton, P. R. J. Am. Chem. Soc. 1996, 118, 3694. (19) Caruso, T.; Carotenuto, M.; Vasca, E.; Peluso, A. J. Am. Chem. Soc. 2005, 127, 15040. (20) Caruso, T.; Capobianco, A.; Peluso, A. J. Am. Chem. Soc. 2007, 129, 15347. (21) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, ReVision C.02; Gaussian, Inc.: Wallingford, CT, 2004. (22) Yurenko, Y. P.; Zhurakivsky, R. O.; Ghomi, M.; Samijlenko, S. P.; Hovorun, D. M. J. Phys. Chem. B 2007, 111, 9655. (23) Hocquet, A.; Leilliot, N.; Ghomi, M. J. Phys. Chem. B 2000, 104, 4560. (24) Foloppe, N.; Hartmann, B.; Nilsson, L.; MacKerell, A. D., Jr. Biophys. J. 2002, 82, 1554. (25) Altona, C.; Sundaralingam, M. J. J. Am. Chem. Soc. 1972, 94, 8205. (26) Leulliot, N.; Ghomi, M.; Scalmani, G.; Bertier, G. J. Phys. Chem. A 1999, 103, 8716. (27) Andersson, K.; Barysz, M.; Bernhardsson, A.; Blomberg, M. R. A.; Carissan, Y.; Cooper, D. L.; Cossi, M.; Fu¨lscher, M. P.; Gagliardi, L.; de Graaf, C.; Hess, B.; Hagberg, G.; Karlstro¨m, G.; Lindh, R.; Malmqvist, P.-Å.; Nakajima, T.; Neogra´dy, P.; Olsen, J.; Raab, J.; Roos, B. O.; Ryde, U.; Schimmelpfennig, B.; Schu¨tz, M.; Seijo, L.; Serrano-Andre´s, L.; Siegbahn, P. E. M.; Stålring, J.; Thorsteinsson, T.; Veryazov, V.; Widmark, P.-O. MOLCAS, Version 6.0; Department of Theoretical Chemistry, Chemical Centre, University of Lund: P.O.B. 124, S-221 00 Lund, Sweden, 2004. (28) Karlstro¨m, G.; Lindh, R.; Malmqvist, P.-Å.; Roos, B. O.; Ryde, U.; Veryazov, V.; Widmark, P.-O.; Cossi, M.; Schimmelpfenning, B.; Neogra´dy, P.; Seijo, L. Comput. Mater. Sci. 2003, 28, 222.

VDEs of 2′-Deoxyguanosine 5′-Monophosphate Anion (29) Veryazov, V.; Widmark, P.-O.; Serrano-Andre´s, L.; Lindh, R.; Roos, B. O. Int. J. Quantum Chem. 2004, 100, 626. (30) Serrano-Andre´s, L.; Mercha´n, M.; Nebot-Gil, I.; Lindh, R.; Roos, B. O. J. Chem. Phys. 1993, 98, 3151. (31) Roos, B. O.; Andersson, K.; Fu¨lscher, M. P.; Malmqvist, P.-Å.; Serrano-Andre´s, L.; Pierloot, K.; Mercha´n, M. AdV. Chem. Phys. 1996, 93, 219. (32) Borin, A. C.; Serrano-Andre´s, L. Chem. Phys. 2000, 262, 253. (33) Serrano-Andre´s, L.; Mercha´n, M.; Borin, A. C. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 8691. (34) Serrano-Andre´s, L.; Mercha´n, M.; Borin, A. C. J. Am. Chem. Soc. 2008, 108, 2473. (35) Serrano-Andre´s, L.; Mercha´n, M. In Photostability and PhotoreactiVity in Biomolecules: Quantum Chemistry of Nucleic Acid Base Monomers and Dimers; Leszczynski, J., Shukla, M., Eds.; Springer: New York, 2008.

J. Phys. Chem. B, Vol. 113, No. 8, 2009 2457 (36) Roca-Sanjua´n, D.; Mercha´n, M.; Serrano-Andre´s, L.; Rubio, M. J. Chem. Phys., 2008, 129, 095104. (37) Ghigo, G.; Roos, B. O.; Malmqvist, P-Å. Chem. Phys. Lett. 2004, 396, 142. (38) Forsberg, N.; Malmqvist, P-Å. Chem. Phys. Lett. 1997, 274, 196. (39) Pierloot, K.; Dumez, B.; Widmark, P.-O.; Roos, B. O. Theor. Chim. Acta 1995, 90, 87. (40) Gelbin, A.; Schneider, B.; Clowney, L.; Hsieh, S.-H.; Olson, W. K.; Berman, H. M. J. Am. Chem. Soc. 1996, 118, 519. (41) Gu, Y.; Kar, T.; Scheiner, S. J. Am. Chem. Soc. 1999, 121, 9411. (42) Wang, J.; Gu, J.; Leszczynski, J. Chem. Phys. Lett. 2007, 442, 124. (43) Wang, X.-B.; Vorpagel, E. R.; Yang, X.; Wang, L.-S. J. Phys. Chem. A 2001, 105, 10468. (44) Roca-Sanjua´n, D.; Serrano-Andre´s, L.; Mercha´n, M. Chem. Phys. 2008, 349, 188.

JP806105H