Interactions of Electrons with Bare and Hydrated Biomolecules: From

Review pubs.acs.org/CR

Interactions of Electrons with Bare and Hydrated Biomolecules: From Nucleic Acid Bases to DNA Segments Jiande Gu,*,† Jerzy Leszczynski,‡ and Henry F. Schaefer III*,§ †

Drug Design& DiscoveryCenter, State Key Laboratory of Drug Research, Shanghai Instituteof Materia Medica, Shanghai Institutes for BiologicalSciences, CAS, Shanghai 201203, P. R. China ‡ Interdisciplinary NanotoxicityCenter, Department of Chemistry and Biochemistry, JacksonState University, Jackson, Mississippi 39217, UnitedStates § Center for ComputationalQuantumChemistry, University of Georgia, Athens,Georgia 30602-2525, United States 5.2. Microhydrated Adenine−Thymine/Uracil Pair 5.3. Guanine−Cytosine Base Pair 5.4. Microhydrated Guanine−Cytosine Pairs 5.5. Other Base Pairs 5.6. Base Pair Binding with Metal Clusters 5.7. Hydrogen-Abstracted Radical Base Pairs 6. Electron Attachment to Nucleosides and Nucleotides 6.1. Nucleosides 6.2. Nucleoside Pairs 6.3. Nucleoside Monophosphates 6.4. Nucleoside Diphosphates 7. Electron Attachment to Single-Strand and Double-Strand Nucleotide Oligomers 7.1. dTpdA and dApdT 7.2. dGpdC and dCpdG 7.3. dGpdG 7.4. dGpdCpdG 7.5. [dGpdC]2 7.6. dGpdGpdG:dCpdCpdC 8. Electron Attachment Induced Bond Breaking in DNA 8.1. Mechanism of Electron Attachment Induced Bond Breaking in DNA 8.2. Reactions at the Thymidine Site 8.2.1. Gas Phase 8.2.2. Aqueous Solution 8.3. Reactions at the Cytidine Site 8.3.1. Gas Phase 8.3.2. Aqueous Solution 8.4. Reactions at the Adenosine Site 8.4.1. Gas Phase 8.4.2. Aqueous Solution 8.5. Reactions at the Guanosine Site 8.5.1. Gas Phase 8.5.2. Aqueous Solution 8.6. Comparison with Experiment 8.7. Successive Electron Attachment Induced Bond Breaks in Nucleotides 8.7.1. Bond Breaking of C3′−O3′

CONTENTS 1. Introduction 2. Methods Used in the Theoretical Studies 2.1. Semiempirical Methods 2.2. Wave Function Methods 2.2.1. Hartree−Fock Method 2.2.2. MP2 Method (Møller−Plesset Perturbation Theory Truncated at Second-Order) 2.2.3. Coupled-Cluster Theory 2.2.4. Multiconfigurational Perturbation Theory (CASPT2) 2.3. Density Functional Theory 2.4. Polarizable Continuum Model 3. Electron Attachment to Nucleic Acid Bases and the DNA Backbone 3.1. Guanine 3.2. Adenine 3.3. Cytosine 3.4. Thymine and Uracil 3.5. 5-Halouracils 3.6. DNA Backbone 3.7. Hydrogen Abstracted Nucleobases 4. Electron Attachment to Microsolvated Bases 4.1. Adenine-(H2O)n 4.2. Cytosine-(H2O)n 4.3. Thymine-(H2O)n 4.4. Uracil-(H2O)n 4.5. H-Bonded Nucleic Acid Bases 5. Electron Attachment to Nucleobase Pairs 5.1. Adenine−Thymine/Uracil Pair

B C D D D D D D E F F F G G H H I I J J J J J J M M

N O O P P P P P Q Q R U U U U V W X Y Y Z Z Z Z Z Z Z Z AA AB AB AB AB AC AC

Special Issue: 2012 The Solvated Electron Received: January 19, 2012

© XXXX American Chemical Society

A

dx.doi.org/10.1021/cr3000219 | Chem. Rev. XXXX, XXX, XXX−XXX

Chemical Reviews 8.7.2. Breaking of the N-Glycosidic Bond 9. New Results for Thymine: An Assessment of Methods 10. Conclusions and Outlook Author Information Corresponding Author Notes Biographies Acknowledgments Abbreviations References

Review

AC AD AF AG AG AG AG AH AH AI

Figure 1. Qualitative diagram of the potential energy surface for a neutral molecule M and the corresponding radical anion M•−. The energy differences shown represent the vertical attachment energy (VAE), vertical detachment energy (VDE), adiabatic electron affinity (AEA), zero-point vibrational energy (ZPE), and ZPE corrected AEA (AEAZPE). ΔETS represents the activation energy barrier for the possible reactions of the radical anion. Instead of chemical reaction, electron detachment should occur when ΔETS is larger than VDE.

1. INTRODUCTION The broadest motivation for exploring electron attachment to biological molecules comes from their important role in radiation damage. Large quantities of secondary electrons with subionization energies (below 20 eV) are produced during processes involving the interaction of high-energy particles with cellular constituents.1−5 These secondary electrons can damage biomolecules within cells, causing mutagenic, recombinogenic, and other potentially lethal DNA lesions.6−11 Experimental studies indicate that electrons with subionization energies play an important role in inducing damage in DNA.12−22 However, the detailed mechanisms of reactions involving electrons interacting with DNA are difficult to address experimentally, simply because too many factors affect the electron−DNA interactions. Formation of a stable radical anion is one of most probable initial events due to the interaction of a biomolecule with secondary electrons.23,24 In interacting with DNA, an electron can temporarily attach to a part of DNA to form a temporary electron-bound state. If this temporary state is the ground state of the radical anion, then the captured electron may be permanently localized on the DNA to form a stable anion. Electron trapping on the DNA bases is thought to be crucial in the cascade of events leading to DNA damage.25−30 An electron trapped by a neutral molecule results in the corresponding radical anion. The energy difference between the neutral molecule and the corresponding radical anion is referred to as the electron affinity (EA). The EA is an important physical measurable quantity often used in theoretical and experimental descriptions of electron attachment to a molecule. From a theoretical viewpoint, an electron residing in a molecule leads to changes in the nuclear configuration to form an equilibrium structure for the corresponding radical anion. In electronic structure theory, this provides two Born−Oppenheimer potential energy surfaces, specifying the geometries of both the neutral molecule and its radical anion (Figure 1). Three parameters described by energy differences between these potential energy surfaces are particularily valuable, i.e., the vertical attachment energy (VAE), adiabatic electron affinity (AEA), and vertical detachment energy (VDE). The VAE is the energy released from the instantaneous one electron attachment to a neutral species. No geometry relaxation takes place during this process. Therefore, both neutral and anionic molecules reflect the optimized geometry of the neutral species. Physically, the VDE is the energy required for the instantaneous detachment of an electron from the stable radical anion; both neutral and anionic compounds are thus placed at the optimized equilibrium nuclear configuration of the radical anion. The AEA represents

the difference between the total energies of the neutral and anionic species at their respective optimized geometries. These theoretical descriptions of VAE, VDE, and AEA are formulated as follows: VAE = En(optimized neutral) − Ea(optimized neutral) VDE = En(optimized anion) − Ea(optimized anion) AEA = En(optimized neutral) − Ea (optimized anion)

In the above, En is the total energy of a neutral species, and Ea is the total energy of the corresponding anion; the optimized neutral is the equilibrium geometry of the neutral molecule, while the optimized anion is the equilibrium geometry of the anionic species. Important factors affecting the AEA value of a molecule are its vibration degrees of freedom. Harmonic zero-point vibrational energies (ZPVE or ZPE) provide reasonable corrections for the AEA.27 ZPE corrections are small for many molecules.31 However, they can be important for DNA subunits, where ZPE corrections often increase the AEA of DNA fragments by ca. 0.1 eV or even more. Experimentally, the VDE corresponds to the peak maximum of the anion photoelectron spectra, while the AEA may be determined by the onset of structure in the anion photoelectron spectra.32 Two strands of the right-handed helix of a polynucleotide form a DNA molecule. Each strand incorporates repeated sugar−phosphate units (backbone). Four fundamental bases, adenine (A), guanine (G), cytosine (C), and thymine (T), are covalently bonded to the sugar moiety of the backbone. Two strands are bound antiparallely through H-bonds between bases G and C and bases A and T. Figure 2 illustrates the shortest double-stranded segment of DNA. It should be mentioned that the OH in the phosphate group is deprotonated in aqueous solutions and the negative charge on the corresponding oxygen atoms of the phosphate group is counterbalanced by a cation, such as Na+.33 The electron affinities of DNA and its subunits have been investigated experimentally by several different methods. Cyclic voltammetry has sometimes been used to determine the electron affinities of nucleic bases in aqueous solutions.34 Rydberg electron transfer (RET) spectroscopy35−39 and photodetachment−photoelectron spectroscopy (PDPES)40−54 have been applied to the study of gas phase electron trapping in biomolecules such as nucleic acid bases, water− B


Chemical Reviews

Review

5′-dCMP, 5′-dTMP),70−73 and the nucleoside-3′,5′-diphosphates (3′,5′-dGDP, 3′,5′-dADP, 3′,5′-dCDP, and 3′,5′dTDP)74 has been established. The predicted AEA and VDEs of the nucleosides have been confirmed by subsequent photoelectron spectroscopic experiments.54 Theoretical studies have been extended to explore the electron attachment to hydrogen-bonding paired DNA subunits, including nucleobase pairs,75−83 nucleoside pairs,84,85 and nucleotide−nucleobase pairs.86 The influences of microhydration on electron attachment to nucleobases have also been investigated extensively at different levels of theory.87−91 Thanks to the vast developments in computer science and technologies, the realistic description of electron attachment to the single-strand nucleotide oligomers and even double-strand nucleotide oligomers is now feasible. Dinucleoside phosphate deoxycytidylyl-3′,5′-deoxyguanosine (dCpdG), dinucleoside phosphate deoxyguanylyl-3′,5′-deoxycytidine (dGpdC),92 dinucleoside phosphate deoxythytidylyl-3′,5′-deoxyadenosine (dTpdA), dinucleoside phosphate deoxyadenylyl-3′,5′-deoxythytidine (dApdT),93 dinucleoside phosphate deoxyguanylyl3′,5′-deoxyguanosine (dGpdG),94 and trinucleoside phosphate dGpdCpdG,95 along with their corresponding radical anions, have been investigated using reliable DFT approaches. Recently, the electron affinities of a representative segment of DNA, the hydrated double helix (dGpdC)2, have also been predicted on the basis of density functional theory.96 Electron attachement to biomolecules through the formation of resonance states has been summarized in several excellent reviews.5,11,27−30 This review is focused exclusively on the formation of bound anionic states of the DNA related species. Results obtained from studies of electron attachment to DNA and its subunits at different levels of theory and from experiment are reviewed in the present article. The theoretical methods used in the studies of the electron attachment to DNA fragments and subunits are briefly reviewed in section 2. Electron attachment to nucleic acid bases and the DNA backbone is reviewed in section 3. Electron attachment to microhydrated bases is described in section 4. Electron attachment to nucleobase pairs is discussed in section 5. Electron attachment to nucleosides and nucleotides is summarized in section 6. Electron attachment to single-strand and double-strand nucleotide oligomers is discussed in section 7. Electron attachment induced bond breaking in nucleotides is introduced in section 8. Our conclusions are given in section 9.

Figure 2. Representative segment of DNA containing guanine (G), adenine (A), cytosine (C), and thymine (T), the four nucleobases. Hydrogen atoms in the backbones have been omitted for clarity. Color conventions: red for O, blue for N, gray for C, white for H, and orange for P.

nucleobase clusters, amino acid−nucleobase clusters, H-bonded nucleic acid base pairs, and nucleosides. Low-energy photoelectron transmission (LEPET) spectroscopy has been used to explore the electron capture capability of DNA oligomers in condensed matter.55,56 These experimental investigations provide valuable information concerning the electron capture abilities of biomolecules. However, details such as the electron density distribution of the captured electron may be difficult to obtain from experimental data. Parallel to experimental studies, theoretical investigations at various levels of theory have complemented the exploration of electron attachment to DNA and RNA. The beauty of the theoretical approach is that it can provide detailed physical pictures, such as where the excess electron might attach, how the geometry of the molecule changes as a result of the electron attachment, and how the excess electron is distributed in the radical anions. Although the first quantum chemical study of the electron affinity of nucleic acid bases was performed in 1976 by Younkin, Smith, and Compton, 57 systematic theoretical efforts made to explore electron attachment to DNA subunits and fragments started only about 20 years ago. A series of pioneering papers by Colson, Besler, and Sevilla58−62 applied the Hartree−Fock (HF) method in exploring electron capturing by DNA subunits. The first application of the post HF method MP2, by Oyler and Adamowicz, predicted the existence of dipole-bound states of nucleic acid base anions.63,64 Coupled cluster methods with single, double, and perturbative triple excitations (CCSD(T)) more recently have revealed tautomeric forms related to covalently bound radical anions of guanine and adenine, and this theoretical research has been successful in explaining the experimental observations of Bowen and co-workers.43,44 Ten years ago, experimental AEA values of individual nucleic acid bases began to be reproduced by density functional theory (DFT).65−68 The extensive studies of uracil have suggested that DFT predictions for the AEA are reasonably close to those predicted with high level conventional wave function methods such as CCSD(T), BD(T), and CASPT2 (see section 2.3). The successfully initiated applications of density functional theory to the investigation of the electron affinities of DNA and RNA bases opened the gate to the efficient description and comprehension of the interactions between electrons and DNA pieces of greater biological reality. Step by step, with the extensively calibrated B3LYP/DZP++ approach,31 a reliable data bank of the electron affinities of the 2′-deoxyribonucleosides,69 the nucleotides monophosphates (3′-dCMP, 3′-dTMP,

2. METHODS USED IN THE THEORETICAL STUDIES Direct determination of the electron affinities of small biomolecules such as nucleic acid bases is difficult. This is partly due to a combination of two factors: that the lowest unoccupied molecular orbitals of nucleobases have relatively high energies and, at the same time, that nucleobases have relative large dipole moments. Therefore, instead of occupying one of the empty MOs to form valence anions, an electron might be easily trapped by the dipole moment of the biomolecule to form a dipole-bound radical anion (Figure 3). Among the five nucleic acid bases, uracil is the one investigated most extensively in experiments for characterizations of its canonical anionic states. The capability to capture an excess electron in uracil has been studied by Rydberg electron transfer (RET)35,37−39 and by photodetachment−photoelectron spectroscopy (PD-PES).32,40−42,45 These experimental determinations of the AEA of uracil provide benchmarks for justifying and screening various theoretical approaches. C


Chemical Reviews

Review

Another important consideration for theoretical treatments of radical anions is that significant electron correlation energy contributions require methods that go beyond the Hartree− Fock level. 2.2.2. MP2 Method (Møller−Plesset Perturbation Theory Truncated at Second-Order). Applying the HF +MP2 approach (MP2 single point calculations based on the HF optimized structures)102−106 with a modified double-ζ basis set (6-31+G(d)X), Oyler and Adamowicz suggested the existence of dipole-bound states of the radical anions of uracil and thymine.63,64 These states were later confirmed by experiments using RET and PD-PE spectroscopy.32,35 The positive dipole bound AEA values (0.086 eV for U and 0.088 eV for T) predicted by the HF+MP2 level of theory are consistent with the PD-PES observations (0.054 ± 0.035 eV and 0.086 ± 0.008 eV for U; 0.068 ± 0.020 eV and 0.068 ± 0.008 eV for T).32,35 A negative AEA (−0.25 eV) for the formation of the valence anion of uracil was reported at the MP2/6-31+G(d) level of theory.63 The MP2 method has also been found to be heavily dependent on basis set. The MP2 results with aug-cc-pVXZ (X = D, T, Q, 5) basis sets for the AEA of uracil increase to −0.21 eV (aug-cc-pVDZ), −0.15 eV (aug-cc-pVTZ), −0.13 eV (augcc-pVQZ), and −0.12 eV (aug-cc-pV5Z).100 The double− triple−quadruple−quintuple ζ (DTQ5) extrapolated AEA based on the MP2 model is −0.10 eV,100 still significantly smaller than zero. 2.2.3. Coupled-Cluster Theory. Conventional coupledcluster theory with single, double, and perturbative triple excitations (CCSD(T))107−110 has been applied to the study of the electron attachment to uracil.100,101,111 On the basis of the MP2 optimized geometries, a positive AEA for the dipolebound state of the radical anion of uracil was reported at the CCSD(T)/aug-cc-pVDZ level of theory (0.071 eV). However, the AEA of the covalent anion of U was predicted to be −0.05 eV (−0.15 eV without zero-point energy corrections).111 The CCSD(T) method also has a strong dependence on the basis set applied. With the aug-cc-pVTZ basis set, the CCSD(T) method increases the EA by 0.043 eV (−0.12 eV without zeropoint-energy (ZPE) corrections).100 With a slightly modified version of CCSD(T), spin-adaption included (SA) CCSD(T), Dedikova et al. reported an 0.01 eV increase in the AEA of the valence anion of U when the aug-cc-pVQZ basis set is used.101 Therefore, the basis set limit AEA of U should be 0.001 eV at the SA CCSD(T)/aug-cc-pVQZ level with ZPE corrections. Recently, on the basis of density functional theory (DFT) optimized geometries and ZPE corrections, the CCSD(T)/augcc-pVTZ+2df method yields −0.001 eV for the uracil AEA corresponding to the valence anion.112 Similar studies using the Brueckner doubles method with a triples contribution (BD(T))113−115 predict a positive AEA (0.008 eV with the aug-ccpVTZ+2df basis set) for the valence anion of uracil.112 2.2.4. Multiconfigurational Perturbation Theory (CASPT2). Using ab initio coupled-cluster and multiconfigurational perturbation methods, Roca-Sanjuan et al. estimated that the AEA of the valence anion of U might be as large as 0.10 eV with the CASPT2/ANO-L//CASSCF/ANO-L method.116,117 2.2.5. Composite Schemes. 2.2.5.1. Scaling Methods. The first positive AEA for the valence anion of U was predicted by Sevilla et al.62 On the basis of HF and MP2 computations, they predicted that the adiabatic EAs of the DNA bases differ from the vertical EAs by a constant. With this constant, along with the scaled vertical EAs for the DNA bases, the adiabatic

Figure 3. Singly occupied molecular orbital (SOMO) of the dipolebound (BD) state of the radical anion of uracil (a) and the SOMO of the valence radical anion of uracil (b).

Applications of theoretical methods for electron attachment studies have been extensively reviewed by Rienstra-Kiracofe et al.31 Here we primarily focus on methods directly related to exploring the electron affinities of uracil. 2.1. Semiempirical Methods

Early molecular orbital studies attempting to evaluate the electron affinity (EA) of NBAs may be traced back to 1976. Using the Pariser−Parr−Pople (PPP) approximation to the Hartree−Fock equations, Younkin, Smith, and Compton57 predicted negative AEAs for adenine (−0.24 eV), cytosine (−0.03 eV), guanine (−0.86 eV), and uracil (−0.13 eV). Later, by applying an improved PPP version, Compton and coworkers modified their EA values for adenine (−0.82 eV), guanine (−0.94 eV), and uracil (−0.48 eV).97 Applying the AM198 approach, Zhang and Chen99 estimated very different AEA values for five nucleobases: 1.06 eV for adenine, 1.23− 1.48 eV for guanine, 0.82−1.19 eV for cytosine, 0.90−1.16 eV for thymine, and 0.87−1.14 eV for uracil. Thus, the electron affinities predicted by the semiempirical methods contradict each other. 2.2. Wave Function Methods

2.2.1. Hartree−Fock Method. Using the spin-restricted open-shell Hartree−Fock (ROHF) method with the standard Pople’s double-ζ basis set plus polarization functions and diffuse functions (6-31+G(d)), Colson, Besler, and Sevilla performed the first ab initio molecular orbital studies of electron attachment to the DNA bases.58 The negative AEAs of the bases (−0.99 eV and −1.07 eV for T, −1.00 eV for U) suggested that the electron attachment to thymine and uracil is unfavorable.58,62 One factor that affects the reliability of the MO method is the basis set used in the computations. The electron density spatial expansion of anion radicals requires larger basis sets with additional diffuse functions. Later ROHF studies with much larger basis sets, ranging from aug-cc-pVDZ, aug-cc-pVTZ, augcc-pVQZ, to aug-cc-pV5Z, produced similar AEA values for uracil (−0.86 eV to −0.91 eV).100,101 On the other hand, the spin-unrestricted open-shell Hartree−Fock (UHF) model with various basis sets somewhat improves the viability of the radical anion of uracil. The AEA of U varies from −0.67 eV at the UHF/aug-cc-pVDZ level to −0.72 eV at the UHF/aug-ccpV5Z level of theory.100 The puckered ring structure and the singly occupied molecular orbital (SOMO) of the radical anion of uracil suggest that it is a valence anion.100,101 D


Chemical Reviews

Review

predicted by the CCSD(T) and BD(T) methods when the basis sets are sufficiently large (at least aug-cc-pVTZ). Among the MO based approaches, CASPT2 is the only one which reproduces the AEA value of uracil within the experimental limits,117 although it is too early to conclude that this method will be generally reliable for biomolecule EAs. Other high level composite schemes, such as MP2+CCSD(T),100 G4, and G4MP2,112 appear capable of reproducing the positive AEA of uracil with reasonable accuracy.

EAs of DNA pyrimidine bases were scaled to be positive (0.40 eV for U, 0.30 eV for T, and 0.20 eV for C).62 2.2.5.2. MP2 and CCSD(T) Corrections. Bachorz et al. proposed a scheme in which both MP2 and CCSD(T) corrections for ROHF (with extrapolated basis sets) are included in electron affinity predictions. On the basis of this approach, they evaluated the AEA of the formation of the valence radical anion of uracil to be 0.040 ± 0.005 eV.100 2.2.5.3. Gaussian-n Theory. Recently, the Pople model chemistry methods have predicted a similar EA of 0.11 eV for uracil from Gaussian-4 theory (G4) and 0.16 eV by its variation, the G4MP2 method.112,118,119 Table 1 summarizes the predicted AEA of uracil in its canonical form at different levels of theory along with the best

2.3. Density Functional Theory

Over the past decade, density functional theory (DFT) has proven to be an affordable and reliable model for predicting electron affinities of relatively large molecular systems.31,65−68,120,121 Among various functionals, the correlation functional of Lee, Yang, and Parr (LYP),122,123 in conjunction with Becke’s three-parameter HF/DFT exchange functional (B3),124 B3LYP, has been the most widely used in the investigation of electron attachment to nucleic acid bases. Using the B3LYP functional, Russo, Toscano, and Grand65 predicted the AEA of the valence anion of U to be 0.14 eV with a TZVP basis set. This value increases to 0.22 eV when the 6-311+ +G(d,p) basis set is applied65 or to 0.18 eV when the 6311+G(2df,p) basis set is applied.66 A similar AEA was predicted by the Georgia group using B3LYP/TZ2P++ single point energies based on geometries optimized by the B3LYP/ DZP++ method (0.19 eV).68 It is interesting to note that, with the relatively small basis set 6-31+G(d), the B3LYP approach predicted a similar AEA for uracil (0.18 eV).67 In a comprehensive review of the atomic and molecular electron affinities, Rienstra-Kiracofe et al. summarize that the basis set dependence of DFT functionals is fairly small and a DZP++ quality basis set is sufficient for most molecules.31 Other functionals such as B3P86,124,125 BP86,125,126 BLYP,122,126 BHLYP,122,127 and B3PW91124,128,129 have also been applied to studies of AEAs of nucleic acid bases.65,66,68 The newly developed Minnesota density functionals, M052X130 and M06-2X,131 have also been tested for the prediction of the AEAs of nucleic acid bases112 (see Table 2). Overall, both the M06-2X and M05-2X methods provide AEAs of the five nucleic acid bases that best fit the G4 predictions, except for guanine, for which both methods underestimate its AEA by about 0.14 eV. On the other hand, the B3LYP/DZP++ approach overestimates the AEAs by ca. 0.1 eV (compared to the G4 values) systematically. The AEAs

Table 1. Adiabatic Electron Affinities (AEA) of Uracil at Different Levels of Theory in eV method

valence state

BD(T) CASPT2 scaling MP2+CCSD(T) G4 G4MP2

−0.13;57 −0.48;97 0.87−1.1499 −1.00;58 −0.86 ∼ −0.91;100,101 −0.67;100 −0.72100 −0.51;62−0.21; −0.15; −0.13; −0.12; −0.10100 −0.05;111 0.001;101 −0.001112 0.01112 0.10117 0.4062 0.04 ± 0.005100 0.11112 0.16112

PD-PES RET

0.15 ± 0.12;32 0.1642 0.06 ± 0.0337

semiempirical wave function

HF

MP2

CCSD(T)

composite

dipole bound state

0.0963

0.07111

experiment 0.05 ± 0.0432 0.09 ± 0.0135

experimental values. It is clear that HF methods severely underestimate the AEA of uracil, even using the complete basis set procedure. Although the MP2 method successfully predicts the AEA of uracil to form the anion as a dipole bound state, it fails to reproduce the positive AEA of uracil accompanying the formation of the valence anion. Near-zero AEAs of uracil are

Table 2. AEA Values for Nucleic Acid Bases Predicted by Different Density Functionals (ZPE Corrected, in eV) functional

U

T

C

A

G

B3LYPa B3LYPb B3P86b B3PW9c BP86b BHLYPb BLYPb M05-2Xd M06-2Xd G4d experiment28

0.18∼0.22 0.24 0.75 0.18 0.31 0.06 0.15 0.17 0.11 0.11 0.15 ± 0.12

0.14∼0.18 0.20 0.71 0.15 0.28 N6-dehydrogen > C8-dehydrogen > C2-dehydrogen species.164 Dehydrogenation of cytosine also leads to five different neutral radicals. The DFT study by Luo et al. predicted AEAs of these radicals ranging between 2.22 and 3.00 eV. The order of these AEAs follows N1-dehydrogen > N4-dehydrogen > C6dehydrogen > C5-dehydrogen neutral radicals.165 Electron attachment to the dehydrogenated thymine has been studied by Profeta et al.166 The AEAs of these neutral radicals are predicted to range between 1.04 and 3.74 eV. The radicals with a hydrogen atom removed from one of the nitrogen atoms present the largest AEAs of the six radicals investigated (3.22 eV for N1-dehydrogenated and 3.74 eV for N3-dehydrogenated species).166 Electron attachment to the dehydrogenated thymines has also been studied in their rare tautomeric forms. The largest AEA value for the dehydrogenated tautomers of thymine has been estimated to be 3.88 eV with B3LYP/DZP++.167 In general, the AEAs of the nitrogen-

Figure 11. SOMO of the radical anion of the sugar−phosphate-sugar (S−P−S) model for the dipole bound state (a) and the valence state (b).

Table 3. AEA of DNA Bases and Backbone in the Gas Phase and in Aqueous Solution (in eV)

canonical

PCM tautomeric

PCM a

b

G

A

C

T

backbone

−1.79 ∼ −0.61a −1.51 ∼ −0.12b −0.21c −0.07 ∼ −0.01d 1.27∼1.33e 0.00∼0.37f

−1.47 ∼ −0.73a −0.48 ∼ −0.26b −0.40,c 0.12h

−0.09,r −0.14,s 0.07∼0.20b

0.03,w 0.04x

1.39∼2.21g

1.72∼1.99k

−0.56l −0.08∼0.03b −0.01c 0.04∼0.06m 1.89∼2.01n 0.01∼0.02o −1.18p −0.06∼0.06q 1.88∼1.93q

1.44∼1.53i −0.38∼0.04j

c

d

0.08c 0.03t 1.85∼2.06n −0.25−0.12u

0.88,w 1.62x

1.77∼1.95v e

MP2, refs 62, 66. DFT, refs 65, 66, 68, 112, 150. G4, ref 112. DB state; DFT, refs 65, 68. DFT, refs 43, 152. fCCSD(T), ref 43. gDFT, ref 43. DB state N(7)H tautomer, MP2, ref 153. iDF, refs 44, 150. jCCSD(T) ref 44. kDFT, ref 44. lMP4, ref 154. mDB stat MP2, ref 36, MP4, ref 154. n DFT, ref 150, also this study. oDB state, MP4, ref 154. pMP4, ref 154 qDFT this study. rMP2, ref 155. sCCSD(T) ref 155. tMP2 ref 36. uCCSD(T) ref 155. vDFT ref 155. wDFT ref 158. xDFT ref 159. h

I


Chemical Reviews

Review

4.4. Uracil-(H2O)n

centered neutral radicals are higher than those of the carboncentered radicals. Electron affinities of hydrogen abstracted uracil have been predicted by Li, Sanche, and Sevilla using the DFT method (B3LYP with 6-31+G(d) and aug-cc-pVDZ basis sets).168 The AEAs of the corresponding radicals range from 2.34 eV to 3.78 eV. The order of these AEAs follows N3-dehydrogen > N1dehydrogen > C6-dehydrogen > C5-dehydrogen neutral radicals, which is similar to that of thymine. The largest AEAs of the neutral radicals of the dehydrogenated nucleobases follow the order (U−H)• > (T−H)• > (A−H)• > (C−H)• > (G−H)•, which is somewhat different from the order of that for the close-shell nucleobases.

Electron affinities of the uracil−water clusters have been extensively studied at different levels of theory.89−91,172−176 Dipole bound anions of monohydrated uracil have been identified by Smets, McCarthy, and Adamowicz by the MP2 method with a modified 6-31+G(d) basis set.173 Compared to the uracil monomer, the AEA of the monohydrated uracil is smaller (ranging from 0.01 to 0.04 eV). This is consistent with the reduced dipole moments of the hydrated complexes. A dipole bound radical anion of trihydrated uracil, U-(H2O)3, has also been predicted by the MP2 approach.174 On the basis of their modified 6-31++G(d) basis set, Smets et al. predicted the AEA for forming the dipole bound anion of U-(H2O)3 to be 0.01 eV. With the same method, the AEA associated with the formation of the valence anion of U-(H2O)3 is estimated to be near zero (−0.07 eV).174 Using the spin-projected unrestricted MP2 (PUMP2) method with a triple-ζ basis set (6-311++G(2df,2p)), Dolgounitcheva, Zakrzewski, and Ortiz located four valence anions of monohydrated uracil with positive AEAs. The corresponding AEAs range from 0.03 to 0.21 eV.175 Two of the dihydrated uracil complexes were found to form stable valence radical anions by the MP2 method. The corresponding AEAs have been reported to be 0.24 and 0.35 eV, respectively, by Morgado, Pichugin, and Adamowicz.172 A CCSD(T) study of electron attachment to the microhydrated uracil suggested that the AEA of the uracil−water clusters might be as high as 0.16 eV for a monohydrated cluster, 0.37 eV for a dihydrated cluster, and 0.42 eV for a trihydrated cluster, respectively.176 Systematic DFT studies of electron attachment to the microhydrated uracil (U-(H2O)n, n = 1−5) were performed by Kim89 and Bao.90,91 On the basis of the different conformations of the water−uracil clusters, the largest AEA of monohydrated U is estimated to be 0.58 eV.90 This value is about 0.42 eV higher than the CCSD(T) predictions (ca. 0.16 eV).176 For comparison, the photoelectron spectra of the uracil−water dimer anion suggest that the AEA of monohydrated U is between 0.3 and 0.4 eV (on the basis of the onset of the spectrum; see ref 48). The DFT approach seems to somewhat overestimate the experimental results even for the microhydrated bases. The EAs of U-(H2O)n have been found to increase as the number of waters in the clusters increases. The largest AEA reported at the B3LYP/DZP++ level of theory is as high as 0.96 eV.89 The search for the global minima on the potential energy surfaces of the radical anions of uracil−water clusters revealed that the formation of the water clusters is necessary in the lowest energy conformers of the tri-, tetra-, and pentahydrated uracil anions. The microhydration of anonic uracil is mainly through the H-bonding interaction between the O4 atom of the uracil and the tightly clustered water molecules (Figure 12).91 Electron attachment to the microhydrated uracil has been found to enhance the tautomeric process by reducing the activation energy.177 Table 4 summarizes the theoretical predictions for the AEAs of microhydrated bases.

4. ELECTRON ATTACHMENT TO MICROSOLVATED BASES 4.1. Adenine-(H2O)n

Dipole bound radical anions have been suggested for the canonical adenine−water clusters by Jalbout et al.169 For the monohydrated adenine, a dipole bound radical anion is found when the water is H-bonded to N7 and HN6 of adenine. The AEA value predicted at the MP2/6-31++G(d,p) level of theory is 0.02 eV. For the dihydrated adenine, at least six conformers of the adenine-(H2O)2 cluster are able to form dipole bound anions. The corresponding AEAs are evaluated to be from 0.00 to 0.09 eV by the MP2 method. Trihydrated adenine has also been predicted to have a dipole bound radical anion with nearzero AEA. 4.2. Cytosine-(H2O)n

The effects of microhydration on electron attachment to cytosine have been studied by Kim through explicit consideration of various structures of cytosine complexes with up to five water molecules.87 Microhydration has been found to increase the electron capture ability of cytosine significantly. The AEA of cytosine increases with the number of the solvating water molecules. The AEA of the pentahydrated complex of cytosine is predicted to be as large as 0.61 eV at the B3LYP/ DZP++ level of theory. It is noteworthy that microhydration with water molecules also improves the AEA for the formation of dipole bound radical anions of cytosine.170 4.3. Thymine-(H2O)n

Electron attachment to the monohydrated thymine has been studied by the MP2/CBS method with CCSD(T) corrections.172 Dipole bound anions of monohydrated thymine have been identified. The corresponding AEA of the monohydrated thymine is similar (ranging from 0.01 to 0.06 eV) to that for the thymine monomer (0.03 eV36). Valence radical anions of thymine−water clusters have been predicted to accommodate an excess electron. The AEAs predicted by the MP2/CBS method with CCSD(T) corrections range from 0.07 to 0.29 eV.171 Electron attachment to thymine−water clusters (T(H2O)n, n = 1−5) has been studied with DFT theory. Different hydration patterns of both neutral and anionic forms of the clusters have been explored at the B3LYP/DZP++ level of theory. Microhydration has been found to intensify the electron capturing ability of thymine significantly. The AEA of thymine increases as the number of waters in the clusters increases. The AEA of the pentahydrated complex of thymine is predicted to be as large as 0.91 eV at the B3LYP/DZP++ level of theory.88

4.5. H-Bonded Nucleic Acid Bases

Besides water, small molecules with an OH group, such as alcohols and acids, can also form H-bonded clusters with nucleobases. Figure 13 depicts the most stable structures of adenine−methanol, adenine−(methanol)2, and adenine− (methanol)3 complexes. The adenine−methanol cluster has been predicted to be able to capture an excess electron to form a dipole bound (DB) state anion. Figure 14 shows the SOMO J


Chemical Reviews

Review

Figure 12. Lowest energy valence anion structures of (a) tri-, (b) tetra-, and (c) pentahydrated uracil. Reproduced with permission from ref 91. Copyright 2007 American Chemical Society.

Figure 14. SOMO of the DB state of the radical anion resulting from electron attachment to the most favored neutral adenine−methanol cluster (a); SOMO of the DB state of the lowest energy radical anion (b). Based on the HF/6-311++G(d,p) computations.

Table 4. AEA of Microsolvated DNA Bases B-(H2O)n (in eV) n

A

1

0.016

a

0.12−0.30 0.02−0.12c

0.01−0.06 0.07−0.29 0.31−0.54f

2

0.00−0.09a

0.20−0.52b

0.26−0.75f

3

∼0.0a

0.19−0.62b

0.37−0.82f

4 5

C

T b

0.28−0.66b 0.28−0.61b

d

0.44−0.86f 0.74−0.91f

U e

0.01−0.04g 0.03−0.21h −0.05−0.16l 0.35−0.58m 0.35−0.58n −0.49−0.35i −0.05−0.37l 0.30−0.79m 0.30−0.80n 0.01,j −0.07k 0.12−0.42l 0.43−0.86m 0.24−0.90n 0.51−0.91m 0.64−0.96m

Figure 15. SOMO of the DB state of the radical anion of the adenine− (methanol)3 complex. Based on HF/6-311++G(d,p) computations.

over, attachment of an excess electron triggers barrier-free proton transfer from formic acid to adenine. Figure 16 illustrates the process of electron attachment to the H-bonded adenine−formic acid, triggering proton transfer from formic acid to adenine.

a

MP2, ref 169. bDFT, ref 87. cMP2, DB state, ref 170. dDB state, CCSD(T), ref 171. eCCSD(T), ref 171. fDFT, ref 88. gDB state, MP2 ref 173. hPUMP2, ref 175. iMP2, ref 172. jDB state, MP2 ref 174. k MP2, ref 174. lCCSD(T), ref 176. mDFT, ref 89. nDFT, ref 90.

Figure 13. Most stable structures of adenine−methanol, adenine− (methanol)2, and adenine−(methanol)3.

Figure 16. Electron attachment to the H-bonded adenine−formic acid triggers proton transfer from formic acid to adenine. The product lies 0.67 eV below the neutral complex. Reproduced with permission from ref 50. Copyright 2007 American Chemical Society.

of the DB state of the radical anion resulting from electron attachment to the most favored neutral adenine−methanol cluster. Also, the SOMO of the lowest energy anion in a DB state is depicted in Figure 14. The corresponding AEA is calculated to be 0.01 eV at the MP2/6-31++G(d,p) level of theory. Adenine−(methanol)2 is predicted by Jalbout and Adamowicz178 to be unable to form radical anions for all the structures explored. However, the adenine−(methanol)3 complex is found to form both DB (Figure 15) and valence anions with near-zero AEA.178 Electron attachment to H-bonded adenine−formic acid and 9-methyladenine−formic acid complexes results in the formation of viable adenine-centered radical anions.50 More-

The excess electron is located on the π* orbital of adenine or 9-methyladenine, while the negative charge resides on the formic acid moiety. These radical anion complexes are found to be genuine minima. Five conformers of the valence anions of H-bonded adenine−formic acid and 9-methyladenine−formic acid complexes have been located at the B3LYP/6-31++G(d,p) level of theory. The corresponding AEAs range from 0.36 to 0.67 eV.50 The VDEs of the corresponding anions are predicted to span the interval from 0.96 to 1.84 eV. The theoretical results nicely explain the photoelectron spectra of the adenine− formic acid and 9-methyladenine−formic acid anion complexes (Figure 17). K


Chemical Reviews

Review

Figure 19. Photoelectron spectrum of adenine−(formic acid)2 complexes measured with 2.54 eV electrons. Reproduced with permission from ref 179. Copyright 2007 Elsevier.

Figure 17. Photoelectron spectrum of adenine−formic acid (upper) and 9-methyladenine−formic acid anions (lower) measured with 2.54 eV electrons. Reproduced with permission from ref 50. Copyright 2007 American Chemical Society.

Electron attachment to adenine−(formic acid)n (n = 2, 3) was also examined by Bowen and co-workers.179 Double proton transfer from the formic acid to the adenine was found to be associated with electron attachment to the three adenine− (formic acid)2 complexes in the DFT (B3LYP/6-31++G(d,p)) studies (Figure 18). The AEAs for the corresponding

Figure 20. SOMOs of the single proton transferred anionic adenine− (formic acid)3 complex (a) and the double proton transferred anionic adenine−(formic acid)3 complex (b).

anionic cluster (2.86 eV) is much higher than that for the single proton transferred species (1.85 eV).179 The excess electron residing on the adenine is the driving force for this proton transfer. Electron attachment to the H-bonded uracil−glycine complex was studied by Gutowski et al.46 Fifteen conformers of the uracil−glycine complex were located by the DFT and MP2 approaches. In the presence of H-bonded glycine, electron attachment to the uracil leads to the valence anion. The excess electron is found to reside on the π* orbital of uracil. At least five conformers of uracil−glycine were predicted to undergo barrier-free proton transfer from the carboxylic group of glycine to the oxygen atom of uracil, forming the favored distonic radical anion complexes, in which the charge and radical centers are separated.180 Figure 21 depicts the formation of the lowest energy radical anion of the uracil−glycine complex. The VDE

Figure 18. Anionic adenine−(formic acid)2 complexes and the corresponding SOMOs.

complexes were predicted to range from 0.84 to 1.07 eV, based on the B3LYP/6-31++G(d,p) computations. The theoretical VDEs of these anionic adenine−(formic acid)2 clusters range from 1.86 to 2.72 eV. These values are confirmed by the anion photoelectron spectrum of adenine−(formic acid)2 in Figure 19. Electron attachment to adenine−(formic acid)3 was found to trigger either single or double proton transfer from the formic acid to the adenine. Figure 20 displays the SOMOs of the single proton transferred anionic adenine−(formic acid)3 and the double proton transferred anionic adenine−(formic acid)3. The AEA values associated with the single and double proton transferred anionic adenine−(formic acid)3 are close, 1.22 to 1.23 eV. However, the VDE of the double proton transferred

Figure 21. Formation of the lowest energy radical anion of the uracil− glycine complex upon electron attachment. The VDE of the resultant anion is 1.93 eV. Reproduced with permission from ref 46. Copyright 2002 EDP Sciences. L


Chemical Reviews

Review

have been located as the local minima on the potential energy surface (Figure 23). The corresponding AEAs are estimated to be −0.41 eV and the VDE to be 0.24 eV at the MP2/6-31+ +G(d,p) level of theory.182

of this radical anion is predicted to be 1.93 eV. The VDEs of other distonic radical anion complexes range from 1.41 to 2.18 eV. These VDEs match the photoelectron spectrum of the anionic complex of uracil−glycine (Figure 22).

Figure 22. Photoelectron spectrum of the uracil−glycine anion, recorded with 2.54 eV/photon. Redrawn with permission from ref 46. Copyright 2002 EDP Sciences.

DFT studies of the anionic complexes of formic acid with uracil and thymine reveal that the excess electron residing on the base leads to the barrier-free proton transfer from the carboxylic group of formic acid to the valence radical anion in the complexes.47 Barrier-free proton transfer has also been reported for the H-bonded uracil−H2S and uracil−H2Se complexes.48 In the latter studies, the main proton acceptor for the barrier-free proton transfer is found to be the O4 atom of the anionic uracil. Due to the larger deprotonation energy of H2O, barrier-free proton transfer is not predicted for the uracil−water complexes by theory.48 The effects of electron attachment to the H-bonded uracil− alcohol complexes have been studied with both DFT and MP2 methods. Electron attachment is able to induce intermolecular proton transfer for a series of 18 uracil−alcohol complexes. The VDE values of the proton transferred anionic complexes are found to be systematically larger than those of the nonproton transferred complexes.49

Figure 23. An electron is trapped between A and U. Atom N1 of uracil and atom N9 of adenine take the same side in part a, while atom C6 of uracil and atom N9 of adenine take the same side in part b. These complexes may be viewed as electron bound anions. Reproduced with permission from ref 182. Copyright 2003 American Chemical Society.

Studies with density functional theory present a different scenario.75−79 DFT study of Reynisson and Steenken reported a adenine-centered radical anion of the AT base pair.75 The AEA is estimated to be −0.07 eV with B3LYP/6-31G(d,p), about 0.7 eV larger than that of A (−0.77 eV). The significantly reduced N1(A)···HN3(T) H-bond length (1.55 Å) in this anionic AT pair indicates that the negative charge is located on adenine.75 This adenine-centered anion is unlikely to be bound, although it might serve as an intermediate state. Electron redistribution should lead to a thymine-centered radical anion. The thymine-centered radical anion of the AT pair has been reported by various DFT studies. The molecular orbital analysis and the charge distribution analysis by Richardson et al. clearly demonstrate that the excess electron resides on the π* orbital of thymine.76 Figure 24 depicts the SOMO of the thymine-

5. ELECTRON ATTACHMENT TO NUCLEOBASE PAIRS 5.1. Adenine−Thymine/Uracil Pair

The first theoretical study of electron attachment to the adenine−thymine base pair was performed by Colson, Besler, and Sevilla59 at the HF SCF level of theory. Their study suggested that the electron affinity of T in the AT pair is only slightly affected because T and A both act as proton donors and acceptors. On the basis of an MP2 study, Al-Jihad, Smets, and Adamowicz181 located the excess electron on the thymine base in the radical anion of the Watson−Crick AT pair. The AEA of the AT pair was estimated to be −0.40 eV at the MP2/6-31+ +G(d,p) level of theory. Therefore, they concluded that the AT base pair is not an effective trap of an excess electron.181 It is interesting to note that the MP2 predicted AEA for the valence radical anion of T is less negative (−0.14 eV155). Thus, it was concluded that base pairing between T and A seems to reduce the electron accepting ability. On the basis of an MP2 study, Stepanian et al.182 suggested that the adenine−uracil dimer could be connected by an excess electron. In this anion pair the excess electron is found to be shared between the C5−C6−N1 side of uracil and the N7−C8 edge of adenine. The dipoles of the two bases point to the excess electron. Two conformers

Figure 24. SOMO of the thymine-centered radical anion of the Watson−Crick AT pair. Based on B3LYP/DZP++ computations.

centered radical anion of the AT pair. With the B3LYP/DZP++ approach, the corresponding AEA is predicted to be 0.36 eV, about 0.16 eV larger than that of the AEA of thymine. Electron attachment to the AT pair intensifies the pairing interaction between the bases by ca. 3.5 kcal/mol. Similarly, DFT studies of the adenine−uracil base pair report a uracil-centered radical M


Chemical Reviews

Review

Figure 25. Electron attachment triggered proton transfer from N9 of A to O4 of T.

Thus, DNA or RNA with 5XU (X = Br, Cl) species is expected to be vulnerable to electron attack. Using the MP2 method, Kobylecka, Leszczynski, and Rak184 studied the influence of sequence on the electron capture capability of the Watson−Crick AT pair. Their results suggest that the electron affinity of T in DNA strongly depends on the sequence. While the AEA of the 5′-CTC-3′ in its double strand is 0.26 eV (6.0 kcal/mol), the AEA of the 5′-GTG-3′ in its double strand is −0.35 eV(−8.0 kcal/mol). Interestingly, the AEA of sequence 5′-TTT-3′ (0.13 eV) is much larger than that for sequence 5′-ATA-3′ (−0.14 eV) in double strands. The stacking pattern between the bases seems to play a role, although the distribution of the excess electron in these sequences is not clear.

anion complex. H-bonding between A and U increases the electron capture capability of U. The AEA of the AU pair is estimated to be 0.32 eV, while that of uracil monomer is 0.18 eV at the B3LYP/6-31+G(d) level of theory.77 Different H-bonded AT pair anions and (9-methyl-A)(1methyl-T) pair (denoted as MAMT) anions have been explored by DFT and MP2 methods.79 Among ten examined anionic radical AT pairs, seven conformers have been found to be lower in energy than the Watson−Crick pair by 0.1−8.4 kcal/mol. These low energy complexes are expected to dominate the photoelectron experiments. Proton transfer from A to T has been observed in one anionic pair. The interbase proton transfer occurs between the N9 atom of the A and the O4 atom of the T (Figure 25). Although the AEAs for these two complexes are very close, the VDE for the anionic pair before proton transfer is 1.30 eV, while the VDE for the anionic pair after the proton transfer is 2.01 eV. Proton transfer thus favors the anion greatly. On the other hand, only one MAMT radical anion pair is found to lie energetically below the Watson−Crick pair (ca. 2 kcal/mol), and no barrier-free proton transfer has been found for these methylated AT pairs. It is important to note that this low energy pair is in the form of a Hoogsteen pair, which is of biological relevance (Figure 26).79

5.2. Microhydrated Adenine−Thymine/Uracil Pair

The effects of microhydration on electron attachment to AT and AU have been studied on the basis of various models.51,60,80,81 An HF SCF study reveals that the hydration by four water molecules increases the AEA of the Watson− Crick AT pair by ca. 0.43 eV. On the basis of AT-(H2O)5 and AT-(H2O)13, Kumar, Mishra, and Suhai80 estimated the AEA of the microhydrated AT pair to be 0.92−0.97 eV, approximately 0.62−0.67 eV higher than that of an isolated AT pair (0.30 eV77). Five structures of monohydrated Watson−Crick AU pairs have been explored by Kim.81 The AEAs of these AU−H2O clusters have been predicted to range from 0.49 to 0.68 eV, depending on the hydration positions. The electron affinities of dihydrated AU pairs have also been investigated. On the basis of ten different conformers of the AU−(H2O)2 cluster, the AEAs vary from 0.37 to 0.92 eV.81 The hydration patterns thus play key roles in the electron accepting abilities of microhydrated base pairs. Using the B3LYP/6-31+G(d,p) method, Bowen and coworkers studied the effects of electron attachment to the (9methyl-A)(1-methyl-T) base pair solvated by one formic acid molecule.51 Their analysis of the seven most stable anionic clusters reveals that electron attachment to the cluster has a strong tendency to trigger the reaction in which the proton from the carboxyl group of formic acid transfers to the oxygen atom of thymine. The AEAs for the formation of these protontransferred anions range between 0.62 and 0.77 eV. There are two nonproton transferred anions (the Hoogesteen AT pair and the Watson−Crick AT pair; see Figure 27) with AEAs of 0.59 and 0.64 eV. Proton transfer in these anionic clusters only slightly increases their electron affinities. Electron attachment to the Watson−Crick MAMT pair hydrogen-bonded to the side chains of the amino acids serine, asparagine/glutamine, threonine, and tyrosine has been studied by Leszczynski and co-workers using the B3LYP/6-31+ +G(d,p) method.185 The AEAs of the MAMT−asparagine/ glutamine side chain complexes are predicted to range from 0.2 to 0.6 eV. Similar AEAs (0.3−0.6 eV) are estimated for the

Figure 26. SOMO of the radical anion of the Hoogsteen AT pair. This anion pair lies energetically below the anionic Watson−Crick pair. B3LYP/DZP++ method.

The effects of electron attachment to adenine−halouracil base pairs A-5XU (X = Br, Cl, F, H) have been investigated with DFT.183 The AEAs of A-5XU are estimated by the B3LYP/6-31+G(d) approach to be slightly smaller (0.01−0.04 eV less) than those of the corresponding 5XU monomer in the gas phase (0.59 eV for A-5BrU; 0.56 eV for A-5ClU; 0.47 eV for A-5FU).156,183 The dehalogenation process of 5XU is slightly suppressed due to the pairing interaction with A. The activation barriers for the dehalogenation of A-5XU have been estimated to be 3.3 kcal/mol for A-5BrU, 5.6 kcal/mol for A5ClU, and 25.6 kcal/mol for A-5FU, respectively. Electron attachment to A-5XU can still effectively induce dehalogenation for X = Br and Cl. Using the PCM model, the AEAs of A-5XU are predicted to range from 1.92 eV (A-U) to 2.36 eV (A-5ClU). A-5XU pairs are good hosts for an excess electron in aqueous solutions. N


Chemical Reviews

Review

eV and −0.18 eV, respectively, by the MP2 method.186 HF studies suggest that electron attachment to the GC base pair leads to formation of a valence radical anion that could trigger interbase proton transfer.58,59 On the basis of the electron affinities of the isolated bases, cytosine should be the host of an excess electron in the radical anion GC•− pair. Therefore, the proton transfer from N1 of G to N3 of C is expected, forming a radical-charge center separated complex (G−H1)−(C+H3)•. Figure 28 illustrates this electron attachment triggered proton transfer. HF computations with the 3-21G basis set estimate the energy released from this proton transfer process to be 4.9 kcal/mol. The activation barrier for this proton transfer is computed to be 4.6 kcal/mol.59 The electron affinity of C in the GC pair is found to be −0.75 eV, ca. 0.52 eV larger than that in the isolated form (−1.27 eV).59 In contrast, DFT studies result in a positive electron affinity for the GC pair. The AEA of Watson−Crick paired GC is predicted to be 0.60 eV with the B3LYP/DZP++ method and 0.48 eV with the B3LYP/TZP++ approach. Charge distribution analyses support the previous assumption that the excess electron is primarily located on cytosine in the radical anion of the GC pair.82 A DFT prediction of the interbase proton transfer is 3.6 kcal/mol exothermic with the B3LYP/6-31+G(d) method. The total energy of (G−H1)−(C+H3)• is estimated to be 3.2 kcal/mol lower than that of the GC•− pair.83 The N7(H) tautomer of the guanine (G7H) cytosine base pair has been studied with the B3LYP/6-31++G(d,p) method. Electron attachment to the G7HC pair also induces the low energy barrier process of proton transfer from N1(G) to N3(C).52 Different forms of (9-methyl-G)(1-methyl-C) pairs have also been studied. The calculated AEA of these pairs span a range from 0.39 to 0.60 eV.53 Using the MP2 method, Kobylecka, Leszczynski, and Rak187 studied the influence of the sequence on the electron capture capability of the Watson−Crick GC pair. Their results suggest that the sequence is decisive for the AEA of C in DNA. While the AEA of the 5′-CCC-3′ in the double strand is 0.40 eV, the AEA of the 5′-GCG-3′ in the double strand is −0.19 eV. Three GC pairs stacked together in different orientations result in a significantly larger AEA. A detailed analysis of the distribution of the excess electron in these sequences is necessary to understand this phenomenon.

Figure 27. Radical anions of the formic acid solvated Watson−Crick MAMT pair (a) and the radical anion of the formic acid solvated Hoogsteen MAMT pair (b).

MAMT−serine side chain clusters and the MAMT−threonine side chain clusters. Slightly larger AEAs (0.5−0.7 eV) are predicted for the MAMT−tyrosine side chain compounds. The excess electron has been found to be primarily localized on the thymine in all of these complexes.185 In general, microsolvation increases the electron capture propensities of biomolecules. Although the effect of microhydration on the electron affinity is less significant than the influence predicted by the polarizable continuum model, the detailed changes in the local molecular structures, such as solvent−solute proton transfer, can only be revealed by the microsolvation approach. Table 5 summarizes the main results for the theoretical predictions involving AT/AU pairs. Table 5. AEAs of AT/AU Pairs and AT/AU−(H2O)n Clusters (in eV) n 0

1 2 5 13 PCM a

AT

AU

−0.40 0.36b, 0.25−0.63d 0.24−0.35e

c,f

a

0.32

A-5BrU 0.59

f

A-5ClU 0.56

f

A-5FU 0.47f

0.49−0.68g 0.37−0.92g 0.92h 0.97h b

2.02f

2.20f

c

d

2.36f

2.30f e

MP2, ref 181. DFT, ref 76. DFT, ref 77. DFT, ref 79. MAMT, DFT, ref 79. fDFT, ref 183. gDFT, ref 81. hDFT, ref 80.

5.3. Guanine−Cytosine Base Pair

5.4. Microhydrated Guanine−Cytosine Pairs

In the gas phase, electron attachment to the Watson−Crick GC pair could result in a dipole bound radical anion. Studies at the MP2 level of theory suggested that the excess is largely located in the area outside of the C5−C6 edge of cytosine in the GC dimer. The AEA for forming this dipole bound anion is estimated to be 0.09 eV at the MP2/6-31++G(d) level of theory.186 Valence anionic GC pairs in either the Watson− Crick or the anti-Watson−Crick form were thought to be less favorable. The corresponding AEAs are predicted to be −0.06

The Watson−Crick GC pair surrounded by 6 and 11 water molecules, respectively, has been studied by Kumar, Sevilla, and Suhai.188 Calculations with the B3LYP/6-31+G(d,p) method revealed that the electron capture ability increases with the number of waters in the clusters. The AEA for the hexahydrated GC pair is estimated to be 0.74 eV, and that of the hendecahydrated GC pair is estimated to be 0.95 eV. It is important to note that, using the PCM model, the AEAs of these two clusters are the same, 1.77 eV. Meanwhile, the PCM

Figure 28. SOMOs of the GC•− pair and the (G−H1)−(C+H3)• pair. Electron attachment triggers proton transfer in the Watson−Crick GC pair. O


Chemical Reviews

Review

prediction for the AEA of the isolated GC pair is 1.86 eV.188 The effect of microhydration on the AEA of the system is thus intensified by the PCM model.

6. ELECTRON ATTACHMENT TO NUCLEOSIDES AND NUCLEOTIDES

5.5. Other Base Pairs

DFT investigations of electron attachment to nucleosides reveal that the addition of the deoxyribose moiety improves the electron capture ability of the bases. The AEAs for the 2′deoxyribonucleosides are as follows: 0.09 eV for 2′deoxyriboguanosine (dG), 0.06 eV for 2′-deoxyriboadenosine (dA), 0.33 eV for 2′-deoxyribocytidine (dC), and 0.44 eV for 2′deoxyribothymidine (dT). These values are uniformly larger than those of the corresponding nucleobases.69 It should be noted that the AEA increase in dA is dramatic compared to that of the nucleic acid base A. On investigating the geometry of the radical anion dA•−, it is clear that the OH group at the 5′ position of the ribose forms an intramolecular H-bond with the negatively charged adenine moiety. It is this intramolecular Hbonding that increases the viability of the radical anion. Molecular orbital analysis suggests that, except for dG, all of these nucleosides become valence radical anions after capturing an excess electron (Figure 29). The SOMO of the radical anion

6.1. Nucleosides

Two valence anionic AC pairs have been studied using DFT. The corresponding AEAs are estimated to be 0.1−0.2 eV with B3LYP/6-311+G(2df,2p) single-point calculations based on B3LYP/6-31+G(d) optimized geometries.189 Electron attachment to the hydrogenated GC pairs has also been explored by the B3LYP/DZP++ method.190,191 Hydrogenated guanines (GH) are able to form (GH)C pairs through H-bonds. Eleven different forms of radical anions of (GH)C pairs were predicted to have AEAs from −0.16 to 2.85 eV.190 In comparison, nine isolated GH structures are predicted to have EAs in the range 0.07−3.12 eV.192 H-bonding with cytosine reduces the electron capture ability of hydrogenated G in general. An excess electron in the hydrogenated G is predicted by Kumar, Mishra, and Suhai to trigger interbase proton transfer.193 5.6. Base Pair Binding with Metal Clusters

Binding metal clusters to the AT and GC base pairs changes the electron capture abilities substantially. DFT computations (B3LYP/6-31+G(d,p)) of the AEA of AT binding to Au4 and Au8 clusters reveals EAs 1.78 eV for AT-Au4 and 2.47 eV for AT-Au8. Similarly, the AEAs of GC-Au4 and GC-Au8 are predicted to be 1.84 and 2.18 eV, respectively.193 5.7. Hydrogen-Abstracted Radical Base Pairs

Electron attachment to the AT base pair has been found to result in dehydrogenation on either adenine or thymine.194 Five dehydrogenated adenine radicals hydrogen bonded with uracil in the Watson−Crick form have been predicted to have AEA values between 1.86 and 3.52 eV and follow the order N9(A)-H > N6(A)-H > C8(A)-H > C2(A)-H species.195 Similarly, AEAs for the five hydrogen-abstracted AT radical pairs (1.84−3.49 eV) have been reported by Xie et al.196 Compared to the cases of the isolated dehydrogenated adenine radicals, base pairing increases AEAs by 0.24−0.90 eV. Dehydrogenation from the uracil of the AU pair leads to four different neutral radical pairs. The DFT study by Kim and Lind195 predicted AEAs of these radicals ranging between 2.36 and 3.75 eV (B3LYP/DZP++). The order of the EAs of the neutral radicals follows N3(U)-H > N1(U)-H > C6(U)-H > C5(U)-H.195 Using the CPCM polarizable conductor model,197,198 Xie et al. reported significant increases in the AEAs of hydrogen-abstracted AT radical pairs (up to 5.98 eV).196 Electron attachment to the hydrogen-abstracted GC radical pair has been studied by Lind et al. using the B3LYP/DZP++ approach. 199 Cytosine dehydrated GC radical pairs have the largest AEAs (3.65 eV for cytosine N1−H and 3.64 eV for cytosine N4−H complexes). The smallest AEA (1.93 eV) is found for the C8-dehydrogened guanine in the dehydrogenated GC pairs. For comparison, the C8-dehydrogened guanine in the isolated form has an AEA of 2.20 eV.163 Structural analysis of these hydrogen-abstracted anionic pairs reveals that some of these anionic pairs have highly twisted nonplanar structures, which might cause significant torsion in DNA helical structures.

Figure 29. SOMOs of the radical anions of nucleosides: (a) dG•−; (b) dA•−; (c) dT•−; (d) dC•−.

of dG reveals a dipole bound state. Analysis of the charge distributions of the anionic nucleosides shows that the negative charge populations on the bases are high in general. However, the excess electron density on the guanine moiety in dG is noticeably less as compared to the other bases.69 Relatively large vertical detachment energies for these valence anions are reported in the DFT study (0.91 eV for dA•−, 0.72 eV for dC•−, and 0.94 eV for dT•−),69 suggesting the possible detection of these radical anions. This prediction and the AEA values have been confirmed by the subsequent photoelectron spectroscopy experiments of Bowen and co-workers.54 Electron attachment to nucleosides may trigger base release.200−202 The possible mechanism of base release is postulated as follows: (1) an excess electron is trapped on the nucleobase and then (2) the excess electron shifts to the antibonding orbital of the glycosidic bond. The latter shift causes the heterolytic cleavage of the N1-glycosidic bond, along with the formation of base-yl anions and neutral 2′-deoxyriboseC1(H)-yl radicals; see Scheme 1. The process of the N1glycosidic bond breaking of anion radicals of pyrimidine nucleosides and dT and dC anions has been investigated at the P


Chemical Reviews

Review

Scheme 1. Possible Mechanism of N-Glycosidic Bond Break Induced by Electron Attachmenta

a

A stable dT radical anion is formed in the first stage, and the N1-glycosidic bond subsequently breaks to release the thymine anion.201

B3LYP/DZP++ level of theory. 201,202 The release of nucleobases by the capture of one electron depends on the formation of a stable anion radical of the nucleoside. The energy barriers for the glycosidic bond rupture have been predicted to be 18.0 kcal/mol for dT•− and 21.1 kcal/mol for dC•−. The low bond-breaking activation energy and the high vertical detachment energy for dT•− enable the heterolytic cleavage of the N1-glycosidic bond. However, the relatively high energy barrier of dC•− suggests that electron detachment should occur before glycosidic bond cleavage.201,202 Further study by Li, Sanche, and Sevilla on the N1-glycosidic bond breaking of dA•− predicted the bond rupture energy barrier to be 19.2 kcal/mol.202 Considering that the AEA of dA is near zero, electron attachment to dA is unlikely to be substantial. In aqueous solution, the electron capture abilities of the nucleosides are improved greatly. PCM model computations show that the AEAs of dA, dC, and dT are as high as 1.74 eV, 1.81 eV, and 1.90 eV, respectively.202 Thus, base release in aqueous solutions might be observable under appropriate conditions. Using the DFT method, Liang, Bao, and Gu explored possible breakings for the C−C and C−O bonds of the sugar moiety of dT•−.203 Their results suggest that electron attachment to dT weakens the C4′−O4′ bond of the sugar ring. The corresponding activation energy for this bond rupture is estimated to be 20.5 kcal/mol. Thus, electron attachment might lead to sugar ring-opening in thymidine.203 Hydrogen-abstracted 2′-deoxyriboadenosine (dA−H) radicals have been predicted to have substantial positive AEAs, in the range 0.99−3.47 eV by using the B3LYP/DZP++ approach. These neutral radicals are highly attractive to excess electrons. Note in particular that four of the five aliphatic-dehydrogenated dAs are subject to dissociative behavior upon electron attachment.204

Figure 30. SOMO of dAdT•−. The unpaired electron largely resides on the thymine moiety.

electron attachment to the dGdC pair may trigger N1(G) to N3(C) interstrand proton transfer, forming the more stable distonic radical anion (dG−−H1) (dC•+H3). Proton transfer from dG to dC does not significantly change the distribution of the unpaired electron in the radical pair (Figure 31). Interstrand

Figure 31. SOMOs of the dGdC•− anionic pair and the (dG−H1−) (dC+H3•) pair.

6.2. Nucleoside Pairs

Electron attachment to the dAdT nucleoside pair has been studied by using the B3LYP/DZP++ method.84 The AEA evaluated (0.60 eV in the gas phase) for the Watson−Crick dAdT pair indicates the strong electron capture tendency of the nucleoside pair. Charge distributions and molecular orbital analysis suggest that both the negative charge and radical center are predominantly located on the thymine moiety in dAdT•− (Figure 30). Electron attachment to the dAdT nucleoside pair is found to increase the dAdT dissociation energy by ca. 3.0 kcal/mol.84 DFT studies of the dGdC nucleoside pair suggest that the Watson−Crick dGdC pair is an even better electron trapping center. The AEA of dGdC is predicted to be 0.83 eV in the gas phase. The cytosine fragment is found to be the excess electron host in dGdC•−. Similar to the Watson−Crick GC base pair,

proton transfer further stabilizes the radical anion by ca. 2.0 kcal/mol. The activation energy barrier for the proton transfer process is estimated to be 2.1 kcal/mol (after ZPVE correction) in the dGdC•− anionic pair, slightly lower than that of the anionic GC pair (2.6 kcal/mol at the same level of theory). The dissociation energy of the dGdC•− pair increases (up to 11 kcal/mol, as compared to the neutral pair) due to electron attachment.85 6.3. Nucleoside Monophosphates

HF level studies on the pyrimidine nucleoside monophosphates, 2′-deoxycytidine-3′-monophosphate (3′dCMP), 205−207 2′-deoxycytidine-5′-monophosphate (5′dCMP),207 and 2′-deoxythymidine-5′-monophosphate (5′Q


Chemical Reviews

Review

dTMP),208 in the gas phase suggested that pyrimidine nucleotides are not able to host an excess electron stably in the gas phase. The EAs of the pyrimidine nucleotides were estimated to be negative in the gas phase, even in the presence of the stacking interaction of the neighboring bases. However, these HF studies concluded that the same systems have positive AEA in aqueous solution.205−209 DFT investigations of electron attachment to nucleotides reveal that the presence of a phosphate either at the 3′- or 5′position of the ribose moiety of nucleosides increases the electron capture ability. In the gas phase, the AEAs of cytidine monophosphate are predicted to be as high as 0.44 eV for 3′dCMP (base-centered radical)72 and 0.34 eV for 5′-dCMP.70 An analogous study for thymidine monophosphate led to predictions that the AEAs of 3′-dTMP and 5′-dTMP are 0.56 and 0.44 eV, respectively.70,71 Using the PCM model, the AEAs of pyrimidine nucleoside monophosphates are estimated to range from 1.89 to 2.18 eV in aqueous solution.70,72,73 A study of electron attachment to 5′-dTMP in aqueous solutions also suggested that the AEAs associated with formation of the basecentered radical anions might be independent of the existence of counterions in aqueous solution.73 On the basis of an 11 water microsolvated model with one sodium cation, phosphatedeprotonated 5′-dTMP, with the PCM model, Kumar and Sevilla derived a similar AEA for 5′-dTMP in aqueous solution (2.17 eV).210 Another important feature of the radical anions of pyrimidine nucleoside monophosphates is that the bases are the preferred accepting site for an excess electron, either in the gas phase or in aqueous solution. DFT computations suggest that, compared to the phosphate-centered radical anion (Figure 32),

by both HF196−200 and DFT level studies.70,71 The low activation energy barriers predicted by the DFT investigations (6−7 kcal/mol for C3′−O3′ bond breaking and ca. 18 kcal/mol for C5′−O5′ bond breaking in the gas phase) suggest that the C−O σ bonds are vulnerable to electron attack. Recent DFT studies by Rak, Kobylecka, and Storoniak suggest that electron attachment might induce P−O bond cleavage in 3′-dTMP and 5′-dTMP.211 Similar studies of 3′-dGMP show that the C3′−O3′ bond is also weak in the radical anion of 3′-dGMP.212 The activation energy barrier for C3′−O3′ bond rupture is estimated to be only 10.3 kcal/mol in the gas phase. This energy barrier is further reduced (to 5.3 kcal/mol) in the PCM treatment for modeling aqueous solution. Application of the microhydration approach (with 21 water molecules) to mimic the effects of aqueous solution results in an energy barrier (6.5 kcal/mol) similar to that found using the PCM model. DFT studies of electron attachment to 2′-deoxyadenosine-5′monophosphate (5′-dAMP) by Kobylecka et al.213 suggest that an excess electron can be captured by 5′-dAMP to form a dipole bound radical anion when the base adenine adopts the anticonformation. The corresponding AEAs are predicted to be 0.13 and 0.25 eV. The valence anion formed by the adenine base accepting an excess electron is predicted to induce intramolecular proton transfer from the phosphate group to the anionic adenine base, forming a deprotonated anionic phosphate moiety and a protonated neutral radical base. As expected, these radical anion structures lie lower in energy by 0.73 to 1.03 eV than the neutral complex. 213 The corresponding VDE values are predicted to range from 1.62 to 1.89 eV. Figure 33 depicts the structure and SOMO of the

Figure 33. Structure and SOMO of the lowest energy radical anion of 5′-dAMP. One of the protons in the phosphate is transferred to atom N3 of adenine. Two intramolecular H-bonds stabilize this anion.

Figure 32. SOMO of the phosphate-centered radical anion (a) and the base-centered radical anion (b) of 3′-dCMP.

the base-centered radical anion of 3′-dCMP lies 0.5 eV lower in the gas phase. Similarly, the base-centered radical anions of 3′dCMP are about 1.0 eV lower in energy than the corresponding phosphate-centered radical anion in aqueous solution, irrespective of whether the phosphate of 3′-dCMP is deprotonated.72 Base-pairing has been found to increase the electron capture capability of 5′-dTMP significantly in the gas phase. A DFT study of the 5′-dTMP-A Watson−Crick pair predicted that the AEA of this complex is as high as 0.8 eV in the gas phase. However, the influence of base-pairing on the EA is less important in aqueous solution. Also, the presence of counterions has little influence on the EA value of the 5′-dTMP-A pair. Studies of electron attachment to the neutral 5′-dTMP-A pair and the phosphate-deprotonated 5′-dTMP-A pair in aqueous solution result in virtually the same AEA (2.04 eV vs 2.01 eV).86 Electron attachment to nucleotides has a strong tendency to trigger either C3′−O3′ or C5′−O5′ σ bond cleavage, as suggested

most stable anion. The AEA and VDE of this anion are 1.03 and 1.81 eV. The largest AEA and VDE values are close to the experimental determinations (AEA = 1.2 eV and VDE = 1.81 eV) of Bowen and co-workers from photoelectron spectroscopy.214 6.4. Nucleoside Diphosphates

Electron attachment to the nucleotides of pyrimidines has been studied on the basis of several models. The most sophisticated models are 2′-deoxycytidine-3′,5′-diphosphate (3′,5′-dCDP) and 2′-deoxythymidine-3′,5′-diphosphate (3′,5′-dTDP).74 Such models allow examination of the cooperative influence of the phosphate group at both the 3′- and 5′-positions of the nucleosides. The AEAs of 0.44 eV for 3′,5′-dCDP and 0.52 eV for 3′,5′-dTDP favor the formation of the corresponding radical anions. Meanwhile, the large values of the vertical detachment energy (VDE) for these two radical anions (0.71 eV for 3′,5′dCDP•− and 0.67 eV for 3′,5′-dTDP•−) ensure that, in the gas phase, electron detachment will not compete with subsequent R


Chemical Reviews

Review

Figure 34. SOMO of the radical anions of 3′,5′-dCDP•−, 3′,5′-dTDP•−, 3′,5′-dADP•−, and 3′,5′-dGDP•−.

reactions with activation energy barriers less than 16.4 kcal/mol (0.71 eV) for 3′,5′-dCDP•− and 15.5 kcal/mol (0.67 eV) for 3′,5′-dTDP•−. Charge distribution analysis and molecular orbitals reveal that the excess electron is largely located on the base moiety in 3′,5′-dCDP•− and 3′,5′-dTDP•− . However, the contribution of the phosphate group at the 3′-position in hosting the excess electron is of vital importance. About 20− 30% of the negative charge is found on the phosphate group at the 3′-position in 3′,5′-dCDP•− and 3′,5′-dTDP•−, while about 63% of the negative charge is on the bases. The SOMOs of the radical anions also demonstrate the unpaired electron to be partly distributed around the 3′-phosphate (see Figure 34). Solvent effects remarkably increase the electron capturing ability of the nucleoside diphosphates. The AEAs are 1.99 and 1.98 eV for 3′,5′-dCDP•− and 3′,5′-dTDP•−, respectively, in the PCM treatment. Moreover, the increased VDEs of 3′,5′dCDP•− (2.22 eV) and 3′,5′-dCTP•− (2.17 eV) due to the solvent interactions suggest that in aqueous solution the reactions with energy barriers less than 50 kcal/mol might proceed without electron detachment from this radical anion. Due to the interaction with the polarizable medium, the excess electron is found to bind tightly to the pyrimidine bases in aqueous solution. The phosphate group at the 3′-position no longer accommodates the unpaired electron to a large degree (see Figure 35). Let us next consider electron attachment to 2′-deoxyguanosine-3′,5′-diphosphate (3′,5′-dGDP). The AEA of 3′,5′-dGDP reaches 0.24 eV (0.36 eV with ZPE corrections), and the VDE of the corresponding radical anion (3′,5′-dGDP•−) is 0.32 eV. This suggests that 3′,5′-dGDP•− is able to endure reactions with activation energies less than 7.4 kcal/mol (0.32 eV) in the gas phase. Molecular orbital analysis shows that the excess electron is partly bound to the phosphate group at the 3′-position and partly bond to the guanine moiety in 3′,5′-dGDP•−. The SOMO of 3′,5′-dGDP•− in the gas phase presents the typical dipole bound feature for the guanine moiety (Figure 34). The

Figure 35. SOMOs of the radical anions of 3′,5′-dCDP•−, 3′,5′dTDP•−, and 3′,5′-dADP•− in aqueous solution.

charge distribution analysis reveals that about 46% of the negative charge is located near the 3′-phosphate group and 46% of the negative charge is localized near the guanine moiety. The influence of the polarizable surroundings remarkably alters the electron capture pattern of 3′,5′-dGDP. In aqueous solutions, 3′,5′-dGDP can accept one excess electron to form either base-centered (AEA is 1.31 eV) or phosphate-centered (AEA is 0.93 eV) radical anions (Figure 36).149 Similar basecentered radical anions have also been predicted for 3′-dGMP (AEA = 1.36 eV) and 5′-dGMP (AEA = 1.35 eV) with the PCM model. The excess electron in the phosphate-centered radical anion is found to reside on the 3′-phosphate group. This phosphate-centered radical anion is attributed to the cooperative influence of the two phosphate groups and the polarizable medium. It is interesting to note that this phosphate-centered radical anion could be substantially favored by successive S


Chemical Reviews

Review

Figure 36. SOMO of two radical anions of 3′,5′-dGDP•− in aqueous solutions: (a) phosphate-centered radical anion; (b) guanine-centered radical anion.

Scheme 2. Dinucleoside Phosphates Deoxythymidylyl-3′,5′-deoxyadenosine (dTpdA) and Deoxyadenylyl-3′,5′-deoxythymidine (dApdT)

Figure 37. SOMOs of dT•−pdA (a) and dApdT•− (b).

negative charge is located near the 3′-phosphate group, while about 42% is on the adenine base. Interaction with the polarizable surroundings (represented by water as solvent) remarkably improves the electron capture ability of 3′,5′-dADP (AEA = 1.59 eV in the PCM model calculation). Moreover, due to the solvent interactions, the VDE of 3′,5′-dADP•− increases to 1.59 eV. This large VDE value implies that, in the presence of polarizable surroundings, reactions with energy barriers less than 23 kcal/mol might occur before electron detachment occurs for this radical anion. The molecular orbital analysis illustrates the valence anion characteristics. The excess electron occupies the π* orbital of the adenine base.

riboses in nucleotide oligomers. In a study of the DNA backbone, it has been shown that the AEA of the phosphatecentered S−P−S (sugar−phosphate−sugar) model is as high as 1.51−1.62 eV in aqueous solution.159 The electron attachment and detachment energies of 2′deoxyadenosine-3′,5′-diphosphate (3′,5′-dADP) are summarized as follows: the AEA of 3′,5′-dADP is 0.22 eV, and the VDE of the corresponding 3′,5′-dADP•− radical anion is 0.26 eV. The molecular orbitals show that 3′,5′-dADP•− represents a valence radical anion. The excess electron is partly bound to the 3′phosphate group and partly located on the adenine base. The charge distribution analysis suggests that about 46% of the T


Chemical Reviews

Review

7. ELECTRON ATTACHMENT TO SINGLE-STRAND AND DOUBLE-STRAND NUCLEOTIDE OLIGOMERS

Table 7. Theoretical Predictions for the Electron Affinities of T Related DNA Fragments in Aqueous Solution (PCM Model) (in eV)

7.1. dTpdA and dApdT

The dinucleoside phosphates deoxythymidylyl-3′,5′-deoxyadenosine (dTpdA) and deoxyadenylyl-3′,5′-deoxythymidine (dApdT) (Scheme 2), each microhydrated by four water molecules, have been predicted by DFT studies to have a strong tendency to capture an electron to form anionic radical complexes.93 The AEAs of these microhydrated oligonucleotides are predicted to be 1.0−1.1 eV. The electron capture capability of thymine seems largely unaffected by its position in the sequence. The excess electron is located on the thymine base moiety in both dT•−pdA and dApdT•−, as revealed by their SOMOs (Figure 37). About 80% of the negative charge is found to reside on the thymine base in both radical anions. Table 6 summarizes the theoretical predictions for the electron

AEA •−

dApdT → [dApdT] dTpdA → [dTpdA]•− 3′,5′-dTDP → [3′,5′dTDP]•− 5′-dTMP → [5′-dTMP]•− 3′-dTMP → [3′-dTMP]•− dT → dT•− T → T•− dAdT → dAdT•− AT → AT•− AT5w → [AT5w]•− AT13w → [AT13w]•−

a

VEA a

thymine, the position of cytosine in the oligomer affects the electron capture ability of cytosine. PCM model calculations result in a roughly 1 eV increase in the EAs of both dGpdC and dCpdG. The AEAs are 1.65 eV for dGpdC and 1.91 eV for dCpdG, respectively, in aqueous solution.

VDE

1.04 1.09a 0.35b

0.47 0.57a 0.17b

1.72a 1.93a 0.67b

0.28c 0.44d 0.31e 0.06f 0.43g 0.19,h 0.11i 0.87j 0.73j

0.01c 0.26d 0.03k −0.30k 0.20g −0.03,h −0.16h

0.99c 1.53d 0.94e 0.68k 1.14g 0.64,h 0.60i

AEA 2.16a 2.27a 1.98b 1.96,c 2.00,d 2.17e 1.90f 2.06g

a Reference 93. bReference 74. cReference 70. dReference 73; IPCM model, ref 215. eReference 210, phosphate deprotonated, charge balanced with Na+, microsolvated with 11 water molecules. fReference 202. gReference 150.

Table 6. Electron Affinities of Thymine Containing Nucleobases, Nucleosides, Nucleotides, Oligonucleotides, and the Corresponding Microhydrated Complexes (in eV, non-ZPE Corrected) process

process dApdT → [dApdT]•− dTpdA → [dTpdA]•− 3′,5′-dTDP → [3′,5′-dTDP]•− 5′-dTMP → [5′-dTMP]•− 5′-dTMP-Na → [5′-dTMP-Na]•− dT → dT•− T → T•−

7.3. dGpdG

DFT studies of electron attachment to the dinucleoside phosphate deoxyguanylyl-3′,5′-deoxyguanosine (dGpdG, Scheme 4) in aqueous solutions (PCM model) reveal that there are four different electron distribution patterns for the aqueous radical anions of dGpdG; all four are found to be local minima on the potential energy surface.94 The excess electron is found to mainly reside (a) on the proton of the phosphate group (dGpH−dG) or (b) on the phosphate group (dGp•−dG) or (c) on the nucleobase at the 5′ position (dG•−pdG) or (d) on the nucleobase at the 3′ position (dGpdG•−) (Figures 39 and 40). These four radical anions are all expected to be electronically viable species under the influence of the polarizable medium. The corresponding aqueous AEAs range between 1.23 and 1.60 eV. The predicted energetics of the radical anions follow the order dGp•−dG > dG•−pdG > dGpdG•− > dGpH−dG. Both phosphate-centered and basecentered radical anions are important in electron attachment to DNA. The electron affinities of guanine related DNA subunits in aqueous solution are summarized in Table 8. It is important to note that the proton of the phosphate group does not exist in aqueous solution. Consequently, the anionic dGpH−dG should be viewed as an electron accommodated by the microsolvated cation near the phosphate. Also, the deprotonation of the phosphate moiety is expected to reduce the electron accepting ability for forming phosphatecentered radical anions. However, this deprotonation of the phosphate moiety should not significantly alter the electron accepting ability of the base-centered radical anions in aqueous solution, since the AEAs of base-centered radical anions of nucleotides are independent of the counterions in aqueous solution.72,73,86 From comparison with electron attachment to the guanine related DNA fragments, pdGp, pdG, dGp, dG, and G, the electron affinity for the formation of the normal guaninecentered radical anion in DNA is estimated to be around 1.5 eV. In comparison, the AEA associated with the formation of the phosphate-centered radical anion in DNA is estimated to be 1.6 eV in aqueous solution. The base−base stacking pattern in DNA single strands seems to be unaffected by electron

a

Reference 93. bReference 74. cReference 70. dReference 71. Reference 69. fReference 70. gReference 84. hReference 76. i Reference 77. jReference 80. kB3LYP/DZP++, present research. e

affinities of thymine containing DNA subunits and the corresponding microhydrated complexes. The electron affinity of thymine in DNA single strands strongly depends on its surroundings. Although the AEA of dTpdA is close to that of dApdT, it is still not clear if the electron capture ability of thymine in DNA is sequence dependent. Full aqueous solvation increases the electron-capturing ability of the oligonucleotides, by up to 1 eV.93 It is important to note that the effects of the polarizable medium increase the AEA of all the thymine related complexes to around 1.90−2.16 eV (Table 7). This seems to demonstrate that the effects of full solvation on the AEA of DNA single strands cannot be described fully by a few microhydration water molecules. 7.2. dGpdC and dCpdG

The four water molecules microhydrated dinucleoside phosphates deoxyguanylyl-3′,5′-deoxycytidine (dGpdC) and deoxycytidylyl-3′,5′-deoxyguanosine (dCpdG) (Scheme 3) have been predicted to have substantial electron affinities: the AEA is 0.66 eV for dGpdC and 0.90 eV for dCpdG with B3LYP/DZP++.92 Similar results have been obtained at the M05-2X/DZP++ level of theory. The excess electron is found to largely reside on the cytosine base in both radical anions, as shown in their SOMOs (Figure 38). In contrast to the case for U


Chemical Reviews

Review

Scheme 3. Dinucleoside Phosphates Deoxyguanylyl-3′,5′-deoxycytidine (dGpdC) and Deoxycytidylyl-3′,5′-deoxyguanosine (dCpdG)

Figure 38. SOMOs of [dCpdG]•− (a) and [dGpdC]•− (b).

strands.95 The study of the electron attached dGpdCpdG species demonstrates that cytosine containing DNA single strands have a strong tendency to capture low-energy electrons and to form electronically viable cytosine-centered radical anions. The AEA of dGpdCpdG is predicted to be 0.9 eV in the gas phase (M05-2X/DZP++). The excess electron has been identified primarily on the π* orbital of cytosine, as can be viewed from the SOMO of the radical anion (Figure 41). Comparative studies of dGpdCp and pdCpdG models (Scheme 6) suggest that stacking interactions between the bases make little contribution to the AEA of cytosine in DNA single strands. Also, the base−base stacking does not affect the electronic stability of the cytosine-centered radicals. Electron attachment induces and intensifies the intrastrand H-bonding between the successive bases. Intrastrand H-bonding is found to be critical in increasing the AEAs and VDEs of the cytosinecentered DNA fragments. Table 9 summarizes the theoretical electron affinities of cytosine containing nucleotides, oligonucleotides, and the corresponding microhydrated complexes. It is clear that the AEA increases as the system size increases. The AEAs of the tetrahydrated dCpdG (0.90 eV) and that of dGpdCpdG (0.90 eV) seem to suggest that hydrogen-bonding between cytosine and its surroundings is an important contributor to the viability of cytosine-centered radical anions. The polarizable medium has been found to have profound effects on the phenomena considered here. It enhances significantly the electron-capture ability of different DNA segments. Using a PCM model, the AEA of dGpdCpdG is estimated to be 2.06 eV. Under the analogous conditions, the

Scheme 4. Qualitative Structure of the Dinucleoside Phosphate Deoxyguanylyl-3′,5′-deoxyguanosine (dGpdG)

attachment. On the contrary, intrastrand H-bonding is found to be greatly influenced by electron attachment, especially for the base-centered radicals, dG•−pdG and dGpdG•−. The intrastrand H-bonding patterns revealed by theory also suggest that electron attachment might induce intrastrand proton transfer between successive guanines in DNA single strands. 7.4. dGpdCpdG

The trimer of the deoxynucleoside phosphate dGpdCpdG (see Scheme 5) has been used as a model to investigate electron attachment to the cytosine moiety in guanine rich DNA single V


Chemical Reviews

Review

Figure 39. SOMOs of the phosphate-centered radical anions of dGpdG: (a) dGp•−dG; (b) dGpH−dG. Reprinted with permission from ref 94. Copyright 2012 John Wiley and Sons.

Figure 40. SOMOs of the base-centered radical anions of dGpdG: (a) dGpdG•−; (b) dG•−pdG. Reprinted with permission from ref 94. Copyright 2012 John Wiley and Sons.

Table 8. Adiabatic Electron Affinities of G, dG, pdG, dGp, pdGp, and dGpdG in Aqueous Solution (in eV) process

AEA

VDE

dGpdG → dGpH−dG dGpdG → dGp•−dG dGpdG → dG•−pdG dGpdG → dGpdG•− pdGp → pdGp•− pdGp → pdG•−p pdG → pdG•− dGp → dG•−p dG → dG•− G → G•−

1.23a 1.60a 1.48a 1.38a 0.93,b 0.95,c 1.20d 1.31b, 1.27d 1.35b 1.36b 1.30b 1.27,b (1.33)e

1.67a 4.64a 2.37a 1.91a 1.01b 1.88b 1.91b 1.90b 1.90b 1.88b

Scheme 5. Structure of a Trimer of the Deoxynucleoside Phosphate dGpdCpdG

a

Reference 94. bReference 149. cReference 74: with the PCM model, based on gas phase optimized structures. dM06-2X/DZP++ approach, structures optimized in solution. Reference 94. eReferences 43, 152.

interactions between the bases should not significantly increase the electron affinity of the system in aqueous solution.

AEAs of dGpdCp and pdCpdG are predicted to be 1.96 and 2.07 eV, respectively. It is important to note that the AEA values for forming C-centered radical anions of DNA segments (ranging from nucleotide dCMP to dinucleotide dimer [dGpdC]2)71,72,74,92,95,96 are close to 2 eV in aqueous solution (see Table 10). It is suggested that the ultimate value of the AEA of cytosine inside DNA in aqueous solutions is less than 2.5 eV.95 Moreover, the AEA values of pdCpdG and dGpdCpdG are so close that one may conclude that stacking

7.5. [dGpdC]2

The representative minimal unit of the DNA double helix, the dinucleoside phosphate deoxygaunylyl-3′,5′-deoxycytidine dimer octahydrate, [dGpdC]2, has been constructed and its geometry fully optimized (B3LYP/6-31+G(d,p) level) for the first time by Gu and co-workers.96 DFT studies of electron attachment to this complex suggest that dGdC containing DNA double strands are excellent electron acceptors. The AEAs of W


Chemical Reviews

Review

Table 9. Electron Affinities of Cytosine Containing Nucleotides, Oligonucleotides, and the Corresponding Microhydrated Complexes (ZPE Corrections Not Included (in eV))a process

Figure 41. SOMO of [dGpdCpdG]•−. Reprinted with permission from ref 95. Copyright 2012 American Chemical Society.

AEA

VEA

VDE

dGpdCpdG → [dGpdCpdG]•− dGpdCp → [dGpdCp]•− pdCpdG → [pdCpdG]•− dGpdC → [dGpdC]•− dCpdG → [dCpdG]•− 3′,5′-dCDP → [3′,5′-dCDP]•−

0.90b

0.23b

1.65b

0.73b 0.64b 0.64, 0.66c 0.90, 0.90c 0.24,b 0.27d

5′-dCMP → [5′-dCMP]•− 3′-dCMP → [3′-dCMP]•− dC → dC•− C → C•− C1w → [C1w]•−

0.34e 0.44e 0.21f −0.11,c −0.09g 0.16(exptl),h 0.07−0.18i 0.29(exptl),h 0.03−0.41i

0.01b −0.12b −0.11, 0.25c −0.17, 0.16c −0.60,b 0.03d −0.11e 0.15e −0.09f −0.25j

1.42b 1.41b 1.50, 1.42c 1.61, 1.64c 1.07,b 0.71d 0.85e 1.28e 0.72f 0.48j 0.15−0.93i

C2w → [C2w]•−

0.22−1.21i

a Plain print is used for results obtained with M05/DZP++; italics, with B3LYP/DZP++. bReference 95. cReference 82, microhydrated with four water molecules. dReference 74. eReferences 71 and 72. f Reference 69. gReferences 68 and 69. hReference 32. iReference 87. j Present research.

[dGpdC]2 are predicted to be 1.14−1.19 eV in the microhydrated clusters. Orbital analysis of the electron attached [dGpdC]2 reveals that the excess electron is primarily located on the π* orbital of one of the cytosine bases, resulting in geometric distortions of the electron resided cytosine and the H-bonding patterns around it. The optimized structure of the radical anion of [dGpdC]2 suggests a strong tendency toward interstrand proton transfer (from the guanine to the electron attached cytosine). This interstrand proton transfer leads to an even more stable radical anion, consisting of one deprotonated anionic guanine and one protonated neutral cytosine radical, [dG−HpdC:dC+HpdG]•− (see Figure 42). Table 11 summarizes the predicted electron affinities of GC-centered DNA fragments in the gas phase. PCM model estimations of the AEA associated with [dGpdC]2•− yield 2.03 eV. This value matches other assessments of cytosine-centered radical anions for different nucleotides in aqueous solution (see Tables 10 and 12).71,72,74,92,95 The AEA associated with the formation of [dG−HpdC:dC+HpdG]•− is predicted to be 2.26 eV in aqueous solution. This suggests that unless there is direct reaction under the influence of its surroundings, such as proton transfer or tautomerism, the AEA of the cytosine site in DNA in aqueous solutions should be around 2 eV.

7.6. dGpdGpdG:dCpdCpdC

Electron attachment induced interstrand proton transfer in the H-bonding paired guanine and cytosine has been studied by Chen, Kao, and Hsu216 using a two-layer ONIOM (B3LYP/631+G(d):PM3) model following developments by Morokuma’s group.217−223 The activation energy for proton transfer is evaluated to be 8.0 kcal/mol for the electron attached dGpdGpdG:dCpdCpdC complex. Deprotonation at the phosphate groups in the complex has only limited influence on the proton transfer rate. The corresponding activation energy is 9.0 kcal/mol. Using a microhydration (namely, complex hydrated with five water molecules) model, the above energy barrier is reduced to 3.7 kcal/mol, which is close to that predicted by the PCM model (3.4 kcal/mol) and the combined microhydration−PCM model (2.8 kcal/mol). Thus, the PCM model is a reasonable approximation for an aqueous solution, as long as there are no direct reactions (such as solvent−solute proton transfer) taking place.

Scheme 6. Models of pdCpdG and dGpdCp

X


Chemical Reviews

Review

Table 10. Theoretical Predictions for the AEAs of Cytosine Containing Nucleosides, Nucleotides, Oligonucleotides, and Oligonucleotide Dimers in Aqueous Solution (in eV) process

AEA

dGpdCpdG → [dGpdCpdG]•− dGpdCp → [dGpdCp]•− pdCpdG → [pdCpdG]•− [dGpdC]2 → [dGpdC]2•− dGpdC → [dGpdC]•− dCpdG → [dCpdG]•− 3′,5′-dCDP → [3′,5′-dCDP]•− 5′-dCMP → [5′-dCMP]•− 3′-dCMP → [3′-dCMP]•− dC → dC•− C → C•−

2.06a 1.96a 2.07a 2.03b 1.65c 1.91c 1.99d 1.89e 2.18e 1.81f 1.89g

Table 11. Theoretical Predictions for the AEAs of GC Centered DNA Segments in the Gas Phase (in eV, without ZPE Corrections) process

AEA

[dGpdC]2 → [dGpdC]2•−

1.13−1.19a

[dGpdC]2 → [dG−HpdC:dC+HpdG]•− dGdC → [dGdC]•− dGdC → [dG−H:dC+H]•− GC → GC•−

1.66−1.71a

GC → [G−H:C+H]•− GC-(H2O)6 → [GC-(H2O)6]•− GC-(H2O)11 → [GC-(H2O)11]•−

0.68b 0.77b 0.44,b,c 0.36d 0.57,b 0.50d 0.61e 0.85e

VAE 0.38a

0.16b 0.03,b −0.15d e

0.06 −0.19e

VDE 2.20− 2.21a 2.74− 2.76a 0.68b 1.16d 2.09d 1.84e 2.28e

a

Reference 96, microhydrated with eight water molecules. bReference 85. cReference 82. dReference 83. eReference 188.

a

Reference 95. bReference 96, microhydrated with eight water molecules. cReference 92, microhydrated with four water molecules. d Reference 74. eReferences 71 and 72. fReference 202. gB3LYP/DZP+ +, present research.

Table 12. Theoretical Predictions for the AEA of GC Centered DNA Segments in Aqueous Solution (in eV, without ZPE Corrections)

Chen, Kao, and Hsu also suggest that the GC pair in aqueous DNA might be able to host two excess electrons. The resultant G−H−:C+H− dianion is expected to exist long enough to trigger DNA damage.216,224,225

process

AEA

[dGpdC]2•−

2.03a 2.26a 1.86b 1.77b 1.77b

[dGpdC]2 → [dGpdC]2 → [dG−HpdC:dC+HpdG]2•− GC → GC•− GC-(H2O)6 → [GC-(H2O)6]•− GC-(H2O)11 → [GC-(H2O)11]•−

8. ELECTRON ATTACHMENT INDUCED BOND BREAKING IN DNA Electron attachment caused DNA damage is well-known from experiments.5−11,27,28 Such damage can occur either on the nucleic acid bases or on the DNA backbone. Theoretical studies of electron attachment induced damage on nucleobases have been reviewed by Lyngdoh.226 Investigation of electron attachment induced damage to the DNA backbone is a very active field of research.5,12−21,29,30,70,71,201,205−209,212,226−234

a

Reference 96, microhydrated with eight water molecules. bReference 188.

phosphate-sugar moiety lengthens the distance at which the π*/σ* configuration interaction takes place.207,227 This process results in a σ*-attached anion that promptly breaks to yield a phosphate site anion and a sugar site radical. Electrons with energy above 2 eV might attach to PO π* orbitals and result in C−O σ bond rupture.227 In aqueous solution, HF SCF studies predict base-centered radical anions with energies lower than that of the analogous neutral species. C−O σ bond breaking might follow in the gas phase after the formation of a base-centered radical anion.205−209 A similar mechanism has been suggested by Kumar and Sevilla228,233 On the basis of time-dependent functional theory (TD-DFT), they suggest that the excess electron is located on the σ* orbital of the phosphate, corresponding to the first excited state of the radical anion (π to σ* transition with excitation energy 1.42 eV, based on TD-BH&H computations).228 The excess electron then shifts from the phosphate

8.1. Mechanism of Electron Attachment Induced Bond Breaking in DNA

Studies of electron attachment to nucleoside monophosphates have suggested that an excess electron residing on a base of the nucleotides might trigger either C−O bond cleavage or Nglycosidic bond rupture.70−72,205−209,212,227,232−234 On the basis of HF level studies,205−209 Simons227 summarizes his findings on electron attachment induced bond breaking in DNA as follows: In the gas phase, electron attachment to a pyrimidine nucleotide requires specific amounts of energy. Electrons with energy below ca. 2 eV can only attach to the π* orbitals of a nucleobase, to yield a shape resonance anion. A through-bond electron transfer event occurs when the C−O σ bond of the

Figure 42. Interstrand proton transfer leads [dGpdC]2•− to [dG−HpdC:dC+HpdG]•−. Y


Chemical Reviews

Review

Scheme 7. Possible Mechanism for Electron Attachment Induced DNA Single-Strand Breaks at Pyrimidine Sites

mol). The C−O σ bonds are vulnerable to the 3′,5′-dTDP•−. NGlycosidic bond breaking is unlikely to take place via the mechanism suggested above. It is clear from the energy profile of the reaction pathways (Figure 44) that C3′−O3′ bond breaking is the main route to electron attachment induced DNA damage. 8.2.2. Aqueous Solution. Solvent effects increase the energy barriers for C−O σ bond rupture. The corresponding energy barriers in the PCM model are high, up to 18.8 and 14.2 kcal/mol for C5′−O5′ and C3′−O3′, respectively (Table 14). However, these energy barriers are still lower than the VDE of 3′,5′-dTDP•− in aqueous solution. On the other hand, base release seems unlikely due to the high activation energy for Nglycosidic bond rupture (28.8 kcal/mol).

toward the antibonding orbital associated with the C−O σ bond, causing bond breaking. Other density functional theory studies of electron attachment induced DNA damage provide slightly different pictures.70−72,212,228−230,232,234 For pyrimidine nucleotides, based on studies for nucleosides and nucleoside monophosphates,70−72,232,234 a possible reaction mechanism for electron attachment induced σ bond breaking in a DNA single strand has been speculated as follows: at the nascent stage the excess electron resides on the π* orbital of the nucleobase, forming an electronically stable radical anion; and during the thermal movement of the atoms in the molecule, redistribution of the excess electron leads to bond breaking at either the C−O σ bonds or the N-glycosidic bond (Scheme 7). Reactions at the guanosine site should be somewhat different. Usually, electron attachment therein leads to the DB state for dGMP and dGDP in the gas phase,74,212,230 while either phosphate-centered or base-centered radical anions are possible in aqueous solution.74,149,230 Detailed and comprehensive reviews for the C− O σ and N-glycosidic bond breaks based on nucleoside monophosphates models have been given by Simons227b and by Caron and Sanche recently.235 More realistic descriptions of electron attachment induced damage in DNA single strands require more complex and sophisticated models. Such models allow simultaneously examining both C5′−O5′ and C3′−O3′ bond cleavages and Nglycosidic bond rupture processes.229−231

8.3. Reactions at the Cytidine Site

8.3.1. Gas Phase. Similar to the case of thymidine diphosphate, several models have been used to study reactions at the cytidine site.70,71,201,202,229 Electron attachment to the cytidine site results in a viable base-centered valence radical anion (Table 15). The transition state structures for the bond cleavage processes in 3′,5′-dCDP •− may be recognized by the corresponding lengthened bond distances (Figure 45). The corresponding activation energy barriers for the C−O σ bond cleavage process (Table 16) are lower than the VDE (0.71 eV or 16.4 kcal/mol) of 3′,5′-dCDP•−. The C−O σ bonds are weak in 3′,5′-dCDP•−, especially for the C3′−O3′ bond. N-glycosidic bond breaking is unlikely to happen. It is clear from the energy profile of the reaction pathways (Figure 46) that, as for 3′,5′dTDP•−, the C3′−O3′ bond breaking is the main pathway in electron attachment induced DNA damage at the cytidine site. 8.3.2. Aqueous Solution. Solvent effects increase the energy barriers for C−O σ bond breakage. The corresponding energy barriers in the PCM model are high, 18.7 and 13.4 kcal/ mol for C5′−O5′ and C3′−O3′, respectively (Table 16). However, these energy barriers are still lower than the VDE of 3′,5′dCDP•− in aqueous solution (51.2 kcal/mol or 2.2 eV). On the other hand, base release seems unlikely due to the high activation energy barrier for N-glycosidic bond rupture (16.2 kcal/mol).

8.2. Reactions at the Thymidine Site

8.2.1. Gas Phase. Different models have been used to study reactions at the thymidine site.70,71,201,202,229 Electron attachment to the thymidine site results in a viable base-centered valence radical anion (Table 13). The transition state structures for bond cleavage processes in 3′,5′-dTDP•− display correspondingly elongated bond distances (Figure 43). The activation energies for the C−O σ bond cleavage process (Table 14) are lower than the energy needed for vertical electron detachment (VDE = 0.67 eV or 15.5 kcal/ Table 13. Electron Attachment and Detachment Energies (in eV)a229 process

AEA

VEA

VDE

3′,5′-dTDP → 3′,5′-dTDP•− (gas phase) 3′,5′-dTDP → 3′,5′-dTDP•− (PCM model)

0.35 (0.52)b 1.98b

0.17b 1.57c

0.67b 2.17c

8.4. Reactions at the Adenosine Site

8.4.1. Gas Phase. Electron attachment to 3′,5′-dADP leads to the formation of a valence anion in which the excess electron is partly located on the base and partially resides on the phosphate group.74,231 However, bond breaking is still possible in the radical anion of 3′,5′-dADP. The structures of the three

a

Results within parentheses are zero-point vibrational energy corrected. bReference 74. cReference 229 Z


Chemical Reviews

Review

Figure 43. Transition states for bond breaking reactions of 3′,5′-dTDP•−. Reproduced with permission from ref 229. Copyright 2010 Oxford University Press.

Table 14. Relative Energies of Transition States for Bond Breaking Pathways of 3′,5′-dTDP•− in the Gas Phase (kcal/mol)229 ΔETSa

bond breaking C5′−O5′ bond C3′−O3′ bond N-glycosidic bond

ΔETS(PCM)b

e

i

e

13.4 (13.8 , 18.7 ) 6.0 (7.1f) 19.2 (18.9g, 20.9h)

18.8 (17.9 ) 14.2 (13.7f) 28.8

ΔE°TSc

ΔG°TSd e

11.6 (11.9 ) 5.7 (5.3f) 18.8 (17.6g)

11.5 (11.8e) 6.9 (4.4f) 21.1 (18.0g, 18.0h)

a ΔETS = E(transition state) − E(radical anion). bΔETS(PCM), using PCM model with ε = 78.39. cWith the ZPE correction. dFree energy at T = 298 K. eReference 70, using 2′-deoxypyrimidine-5′-monophosphate as the model. fReference 71, using 2′-deoxypyrimidine-3′-monophosphate as the model. gReference 201, using 2′-deoxypyrimidine nucleoside as the model. hReference 202, using 2′-deoxypyrimidine nucleoside as the model. i Reference 228, using 2′-deoxypyrimidine-5′-monophosphate as the model.

Figure 44. Energy profiles of the C5′−O5′, C3′−O3′, and N-glycosidic bond breaking processes for 3′,5′-dTDP•− in the gas phase and in aqueous solution. Reproduced with permission from ref 229. Copyright 2010 Oxford University Press.

transition states signify three correspondingly different bond breakage phenomena (Figure 47). The activation energies for these bond rupture processes (Tables 17 and 18) are larger than the VDE of 3′,5′-dADP•− in the gas phase (6.0 kcal/mol or 0.26 eV). No bond breaking should be expected to follow the proposed reaction mechanism when adenosine is placed in DNA in the gas phase. 8.4.2. Aqueous Solution. 3′,5′-dADP does form a viable base-centered radical anion in the presence of the polarizable

Table 15. Electron Attachment and Detachment Energies (in eV) of 3′,5′-dCDPa process

AEA •−

3′,5′-dCDP → 3′,5′-dCDP (gas phase) 3′,5′-dCDP → 3′,5′-dCDP•− (PCM model) a c

0.27 (0.44) 1.99b

b

Results within parentheses are ZPE corrected. Reference 229.

VEA

VDE

b

0.71b 2.22c

0.03 1.45c b

Reference 74.

Figure 45. Transition states for reactions involving bond breaking in 3′,5′-dCDP•−. Reproduced with permission from ref 229. Copyright 2010 Oxford University Press. AA


Chemical Reviews

Review

energy for the C1′−N9 glycosidic bond breaking is predicted to be 24.1 kcal/mol. However, these energy barriers are noticeably higher than the VDE of 3′,5′-dGDP•− (0.32 eV or 7.4 kcal/mol; see Table 19). Thus, no bond breaking should be expected in the gas phase from the energy profiles of the reaction pathways (Figure 50). Instead, electron detachment should dominate. This is consistent with the fact that 3′,5′-dGDP•− is basically a dipole-bound anion. 8.5.2. Aqueous Solution. The effects of the polarizable surroundings stabilize the radical anion 3′,5′-dGDP•− significantly.73 Moreover, 3′,5′-dGDP can accept one excess electron to form either base-centered or phosphate-centered radical anions in aqueous solution.149 Solvent effects not only greatly decrease the activation energies for the C−O bond cleavage processes (1.1−3.6 kcal/mol for the phosphate-centered radical anion) but also significantly increase the vertical detachment energy (VDE = 1.01 eV or 23.3 kcal/mol for the phosphatecentered radical anion) of 3′,5′-dGDP•−. Consequently, C−O σ bond cleavage processes should be facile. The presence of the polarizable surroundings also lowers the activation energy barrier of N-glycosidic bond breaking to 10.0 kcal/mol; nevertheless, this process is expected to be suppressed by the C−O σ bond cleavage processes (Figure 51 and Table 20). The above predictions nicely explain the fact revealed in recent experiments that guanosine is one of the vulnerable sites in DNA in aqueous solution under bombardment by low energy electrons.137,138 On the basis of the energy difference between the phosphatecentered anion and the guanine base-centered anion, we can estimate the corresponding bond rapture activation energies for the base-centered radical anion of 3′,5′-dGDP. The basecentered radical anion lies ca. 0.36 eV (8.3 kcal/mol) below the phosphate-centered anion. Thus, the corresponding ΔETS(PCM) for the base-centered anion of 3′,5′-dGDP should be around 9.4 kcal/mol for C5′−O5′, 11.9 kcal/mol for C3′−O3′, and 18.3 kcal/mol for the N-glycosidic bond breaking processes, respectively. The high bond breaking rate observed at the guanosine site in aqueous solution137,138 seems to suggest that the phosphate group could be an important electron acceptor in guanine-rich DNA.

Table 16. Relative Energies of the Transition States for Bond Breaking in 3′,5′-dCDP•− in the Gas Phase (kcal/mol) bond breaking

ΔETSa

ΔETS(PCM)b

ΔE°TSc 12.3 (12.5e) 5.2 (4.7f) 25.0 (20.4g)

C5′−O5′ bond

14.2 (14.3e)

18.7 (18.0e)

C3′−O3′ bond N-glycosidic bond

6.0 (6.2f) 26.2 (21.6g, 22.7h)

13.4 (12.8f) 26.3

ΔG°TSd 13.5 (12.8e) 7.6 (4.5f) 26.6 (21.2g, 20.1h)

ΔETS = E(transition state) − E(radical anion). bΔETS(PCM), using the PCM model with ε = 78.39. cWith zero point energy (ZPE) corrections. dFree energy at T = 298 K. eReference 70, using 2′deoxypyrimidine-5′-monophosphate as the model. fReference 71, using 2′-deoxypyrimidine-3′-monophosphate as the model. gReference 201, using 2′-deoxypyrimidine nucleoside as the model. hReference 202, using 2′-deoxypyrimidine nucleoside as the model. a

medium. The effects of the polarizable surroundings are found to greatly increase the activation energies for the C−O σ bond cleavage processes (13.2 kcal/mol for C3′−O3′ σ bond rupture and 22.5 kcal/mol for C5′−O5′ σ bond rupture). However, these energy barriers are still below the value of the electron vertical detachment energy (VDE = 1.59 eV or 36.7 kcal/mol) of 3′,5′dADP•−. Consequently, C3′−O3′ σ bond rupture is expected to occur for the adenine-centered radical anion in the presence of water. The energy profiles for the reaction pathways suggest that C3′−O3′ σ bond cleavage processes should dominate the electron attachment induced DNA single strand dissociation in adenine-rich DNA (Figure 48). 8.5. Reactions at the Guanosine Site

8.5.1. Gas Phase. Different from the other bases, electron attachment to 3′,5′-dGDP does not lead to the formation of base-centered valence radical anion.74 However, the electron affinities of 3′,5′-dGDP in the gas phase suggest that it is possible to form a dipole-bound radical anion as a local minimum on the potential energy surface. This DB state of the radical anion might serve as the precursor of the electron attchment induced bond breaking. The transition state for the C5′−O5′ σ bond breaking can be identified by the elongated C5′−O5′ atomic distance of 1.78 Å (Figure 49).230 The corresponding activation energy for the σ bond cleavage process has been predicted to be 13.0 kcal/mol. Meanwhile, the transition state for C3′−O3′ σ bond breaking is characterized by the lengthened C3′−O3′ atomic distance of 1.74 Å (Figure 49), with an activation barrier of 11.2 kcal/mol. The activation

8.6. Comparison with Experiment

For cytidine, experiments concerning LEE-induced bond breaks of oligonucleotide tetramer GCAT in thin solid films revealed a ratio of 5:11 for C5′−O5′ bond breaks to C3′−O3′ bond breaks (at the cytidine site) induced by incident electrons with the

Figure 46. Energy profiles of the C5′−O5′, C3′−O3′, and N-glycosidic bond breaking process for 3′,5′-dCDP•− in the gas phase and in aqueous solution. Reproduced with permission from ref 229. Copyright 2010 Oxford University Press. AB


Chemical Reviews

Review

Figure 47. Three transition states for bond breaking reactions in 3′,5′-dADP•−. Reproduced with permission from ref 231. Copyright 2011 American Chemical Society.

ratio is also qualitatively consistent with theoretical predictions of the low activation energies for C3′−O3′ bond breakage.229 Considering the oligonucleotide GCAT under study in the thin film experiments,19 the influence of the surroundings on the LEE-induced DNA damage is greater than that revealed by the gas-phase simulations; however, the outcome is less than that suggested by the PCM model. The consistency between the theoretical prediction and the experimental observations related to the reaction pathway ratio provides strong supportive evidence for the base-centered radical anion mechanism for low energy electron-induced single strand bond breaking around the pyrimidine sites of the DNA single strands. Recent experiments have demonstrated that guanosine is one of the vulnerable sites of DNA in aqueous solution under bombardment by low energy electrons.137,138 In the presence of the polarizable surroundings, the significant increase in the electron affinities of 3′,5′-dGDP and in the vertical detachment energy of 3′,5′-dGDP•− ensures the formation of the electronically viable radical anion. Furthermore, the surrounding-solute interactions greatly reduce the activation barriers of the C−O bond cleavage, to approximately 1.1−3.6 kcal/mol.230 These low energy barriers suggest that either C5′−O5′ or C3′−O3′ bond rupture takes place at the guanosine site in DNA single strands. The activation energies of these C−O bond cleavages indicate that C5′−O5′ bond breaking is favored over C3′−O3′.

Table 17. Electron Attachment and Detachment Energies (in eV)231 process

AEA •−

3′,5′-dADP → 3′,5′-dADP (gas phase) 3′,5′-dADP → 3′,5′-dADP•− (PCM model) a

a

0.10 (0.22) 1.59a

a

VEA

VDE

a

0.26a 1.59b

0.04 1.37b

Reference 74. bReference 231.

Table 18. Relative Energies of Transition States for Bond Breaking in 3′,5′-dADP•− (kcal/mol)231 bond breaking process

ΔETSa

C5′−O5′ bond C3′−O3′ bond N-glycosidic bond

10.0 8.9 21.3 (20.3e)

ΔETS(PCM)b ΔE°TSc 22.5 13.2 20.9

9.3 7.1 20.0

ΔG°TSd 11.1 7.3 21.0 (19.2e)

a ΔETS = E(transition state) − E(radical anion). bΔETS(PCM) = E(transition state) − E(radical anion); using PCM model with ε = 78. c With the ZPE correction. dFree energy at T = 298 K. eReference 202, using 2′-deoxyadenosine as the model.

8.7. Successive Electron Attachment Induced Bond Breaks in Nucleotides

8.7.1. Bond Breaking of C3′−O3′. Protonation at the N3 position of the radical cytidine anion results in formation of a viable neutral radical. This neutral radical is a good electron acceptor, with an AEA of 0.52 eV in the gas phase (B3LYP/631++G(d,p) computation for N3-protonated-3′-dCMP radical).225 The DFT study of electron attachment to this neutral radical by Dabkowska, Rak, and Gutowski225 suggests that the resultant anion might undergo intramolecule proton transfer (from C2′ of ribose to C6 of cytosine; see Scheme 8), causing C3′−O3′ bond breaking. The energy profile along this reaction pathway (Figure 52) suggests that the proton transfer (PT) step is followed by barrier-free C3′−O3′ bond breaking. The activation energy barrier for the PT step is estimated to be 6.5 kcal/mol (B3LYP/ 6-31++G(d,p) level).225 8.7.2. Breaking of the N-Glycosidic Bond. Studies of the N3-protonated-5′-dCMP radical (denoted here as [5′-dC(N3H)MP]•) demonstrate that electron attachment to this neutral anion could trigger C2′ to C6 proton transfer and result in the formation of an abasic nucleotide through N-glycosidic bond cleavage (Scheme 9).224 The AEA for electron attachment to [5′-dC(N3H)MP]• is predicted to be 0.40 eV (0.32 eV without ZPE correction) with the B3LYP/DZP++ method.

Figure 48. Energy profile of the C5′−O5′, C3′−O3′, and N-glycosidic bond breaking processes of 3′,5′-dADP•− in PCM modeled aqueous solution. Reproduced with permission from ref 231. Copyright 2011 American Chemical Society.

energy of 15 eV. This ratio decreases to 3:8 (10 eV) and 4:21 (6 eV) as the energy of the incident electrons diminishes.19,27 Therefore, one should expect that the ratio of bond breaks of C5′−O5′ to those for C3′−O3′ induced by the near-zero electron attachment will be even smaller. On the other hand, the fraction of cytosine base release is negligible. The ratio observed in the experiments clearly follows the theoretical sequence of bond breaking reaction pathways, either in water or in the gas phase.229 For thymidine, the analogous experiments reveal bond breaks of the oligonucleotide trimer TTT (TpTpT),21 with a ratio of 2.5:2.9 for C5′−O5′ to C3′−O3′ with relatively high energy incident electrons (11 eV). This AC


Chemical Reviews

Review

Figure 49. Three transition states for bond breaking reactions in 3′,5′-dGDP•−. Reproduced with permission from ref 230. Copyright 2009 John Wiley and Sons.

Table 19. Electron Attachment and Detachment Energies for 3′,5′-dGDP (in eV) process

AEA •−

3′,5′-dGDP → 3′,5′-dGDP (gas phase) 3′,5′-dGDP → 3′,5′-dGDP•− (PCM model)

a

0.24 (0.36) 1.31,a 0.95a

VEA a

0.14

a

0.89c

Table 20. Relative Energies of Bond Breaking Transition States for 3′,5′-dGDP•− (kcal/mol) (Ref 230)

VDE 0.32

bond breaking process

ΔETSa

ΔETS(PCM)b

ΔE°TSc

ΔG°TSd

C5′−O5′ bond C3′−O3′ bond N-glycosidic bond

13.0 11.3 (10.3e) 24.1

1.1 3.6 (5.3e) 10.0

11.1 9.3 (8.6e) 21.6

10.4 8.6 21.1

a

1.88,b 1.01a

a

Reference 74. Phosphate-centered radical anion. ZPE corrected AEA in parentheses. bReference 149. Base-centered radical anion. cPresent study.

ΔETS = E(transition state) − E(radical anion). bΔETS (PCM) = E(transition state) − E(radical anion); using the PCM model with ε = 78. cWith the zero point energy (ZPE) correction. dFree energy at T = 298 K. eReference 212, using 2′-deoxyguanosine-3′-monophosphate as the model. a

3.2 kcal/mol below the anion [5′-dC(N3H)MP]− (Figure 53). The activation energy barrier for the C2′ to C6 proton transfer is predicted to be 13.2 kcal/mol. N-Glycosidic bond breaking is found to be easy; the corresponding activation energy barrier is 2.8 kcal/mol. Interestingly, this PT energy barrier in the N3protonated-3′-dCMP anion is only 6.5 kcal/mol (free energy).224 Therefore, the pathway for electron attachment induced C2′ to C6 proton transfer should be strongly dependent on its environment. The N3-protonated cytidine neutral radical is a stable product of electron attachment to the Watson−Crick GC nucleotide pair. In aqueous solution, the formation of the G−H−:C+H− dianion is possible, as suggested by the study of electron attachment to dGpdGpdG:dCpdCpdC.216 The C2′ to C6 proton transfer mechanism seems to suggest DNA double strands might also be damaged as a result of electron attachment.

Figure 50. Energy profiles for the C5′−O5′, C3′−O3′, and N-glycosidic bond breaking processes of 3′,5′-dGDP•− in the gas phase. Reproduced with permission from ref 230. Copyright 2009 John Wiley and Sons.

9. NEW RESULTS FOR THYMINE: AN ASSESSMENT OF METHODS Uracil and thymine are the nucleobases most likely to have bound valence anions. An urgently important goal for new experiments is to measure the vertical detachment energies (VDEs) and then the adiabatic electron affinities (AEAs), should the latter be positive. The Hartree−Fock method is now known to poorly predict the energetics of nucleoside anions. The valence nucleoside anions are predicted to be too high in energy compared to the analogous neutral species. Although to a lesser degree, secondorder perturbation theory (MP2) suffers from the same problem for the valence nucleoside anions. For thymine in particular, the MP2 AEA is −0.31 eV with both the aug-ccpVTZ and aug-ccpVQZ basis sets. Higher level methods suggest that the thymine EA is either zero or slightly positive.100,112,117

Figure 51. Energy profiles for the C5′−O5′, C3′−O3′, and N-glycosidic bond breaking processes of 3′,5′-dGDP•− in the PCM modeled aqueous solutions. Reproduced with permission from ref 230. Copyright 2009 John Wiley and Sons.

In contrast to the case of the N3-protonated-3′-dCMP anion, C2′ to C6 proton transfer in the 5′-dC(N3H)MP anion results in the intermediate [5′-dC(N3H,C6H)MP]−, which lies about AD


Chemical Reviews

Review

Scheme 8. Proposed Mechanism for C3′−O3′ Bond Breaking Due to Successive Electron Attachment

Figure 52. Energy profile along the proton transfer pathway followed by barrier free C3′−O3′ bond breaking for the anionic N3-protonated3′-dCMP. Redrawn following ref 225. Figure 53. Energy profile for the electron attachment triggered proton transfer followed by N-glycosic bond breaking in the N3-protonated5′-dCMP anion (in kcal/mol). Redrawn following ref 224.

Density functional methods are difficult to assess theoretically because there is no practical, straightforward sequence of DFT methods leading to the exact solution of the Schrödinger equation. Thus, the ultimate reference for the reliability of DFT predictions is experiment. In Table 21 we provide four standard measures of the viability of the valence anion of thymine. These mostly new results (41 DFT methods) were all obtained with the standard DZP++ basis set. The results are ordered sequentially by the values of the vertical detachment energies (VDEs). All methods except B3P86 agree that the vertical electron affinity of thymine is negative; that is, the electron is not bound at the neutral thymine equilibrium geometry. Most of the DFT methods yield adiabatic electron affinities near zero, when zero point vibrational energies are not considered. It seems clear that the B3P86 method is an outlier, and this method is not recommended for study of DNA subunits. Avoiding controversy for the moment, we turn next to the vertical detachment energies (VDEs), in principle observable by photoelectron detachment experiments of the Bowen type. All theoretical VDEs are positive, with the popular B3LYP method (VDE = 0.68 eV) near the middle. Although some would disagree, we consider the DZP++ B3LYP predictions to be reliable to ±0.2 eV for this family of DNA subunits. Why? Our

answer to this question is that Bowen’s experiments confirm the DZP++ results for several related systems.54,214,236 For thymidine, the DZP++ B3LYP VDE is 0.94 eV, only 0.05 eV greater than Bowen’s experimental result. For cytidine, the theoretical VDE is 0.72 eV, 0.15 eV less than Bowen’s experiment. Finally, for adenosine, the DZP++ B3LYP prediction is 0.91 eV, 0.23 eV below experiment, which may be slightly on the high side. In all three cases, theory69 preceded experiment,54 so there was no possibility of adjusting theory to match experimental results. Furthermore, theory predicted no valence bound gaunosine anion (2004),69 and no such anion was observed in the 2007 experiments of Bowen.54 Another example is provided by the thymidine−water complex, for which DZP++ B3LYP predicts VDE = 1.32 eV,236 in unexpectedly close agreement with experiment, 1.3 eV.237 A final example is the AT base pair, for which theory76 predicted (2003) VDE = 0.63 eV and the best experiments (2004)79 provide 0.7 eV. From all available facts at hand, it appears that the simple DZP++ B3LYP method does a good

Scheme 9. Mechanism of Electron Attachment Induced Formation of an Abasic Site in DNA

AE


Chemical Reviews

Review

AEA of 0.20 eV. Unfortunately, 0.20 eV is about the uncertainty in the DFT predictions. In preparing this review, we consulted a number of leading DNA anion theorists. Without mentioning names, the field was roughly split between (a) those of the opinion that the vibrationally bound thymine has a small electron affinity (perhaps 0.1 eV) and (b) those who think T•− is not adiabatically bound. Where does this leave us? In disagreement over a very small energy difference. In the absence of much more exhaustive convergent quantum mechanical methods (e.g., CCSD(T) aug-cc-pVQZ) than those yet attempted, we must await the definitive experiment.

Table 21. Electron Affinities of Thymine Predicted by Different Density Functionals with the DZP++ Basis Set functional

VEA

AEA

AEA(ZPE)

VDE

HF MP2 B3P86 HSE2PBE LC-wPBE M06HF M052x CAM-B3LYP wB97 M062x tHCTHhyb tHCTH wB97X HCTH HCTH407 B3LYP mPW3PBE mPW1PW91 BP86 X3LYP B3PW91 BH&HLYP HCTH147 mPW1PBE PBEh1PBE PBE1PBE HSEh1PBE B98 wB97XD TPSSh mPW1LYP BMK B971 VSXC B1LYP B97D M06 B972 M05 O3LYP B1B95 BLYP HCTH93 M06L BH&H

−0.69 −0.53 0.19 −0.16 −0.42

−0.77 −0.50 0.58 0.24 0.06 0.09 0.02 0.05 −0.06 −0.01 0.05 0.10 −0.06 0.13 0.13 0.06 0.05 0.03 0.14 0.03 0.02 −0.13 0.08 −0.01 0.00 −0.01 0.00 −0.04 −0.09 −0.04 −0.07 −0.16 −0.08 −0.07 −0.09 −0.04 −0.07 −0.14 −0.10 −0.09 −0.16 −0.02 −0.09 −0.24 −0.25

−0.63 −0.42 0.71 0.37 0.19 0.23 0.16 0.18 0.07 0.12 0.18 0.23 0.07 0.26 0.26 0.20 0.18 0.14 0.28 0.16 0.15 0.00 0.21 0.12 0.13 0.12 0.13 0.09 0.04 0.09 0.08 −0.03 0.05 0.06 0.05 0.09 0.06 −0.01 0.03 0.04 −0.02 0.12 0.04 −0.11 −0.12

0.16 0.21 1.20 0.88 0.78 0.78 0.75 0.73 0.72 0.70 0.70 0.69 0.69 0.68 0.68 0.68 0.66 0.65 0.65 0.64 0.64 0.64 0.64 0.64 0.63 0.62 0.62 0.62 0.62 0.61 0.59 0.58 0.57 0.57 0.56 0.55 0.54 0.52 0.50 0.49 0.48 0.47 0.47 0.44 0.44

−0.42 −0.41 −0.54 −0.45 −0.33 −0.15 −0.59

−0.30 −0.33 −0.36 −0.34 −0.35 −0.60 −0.40 −0.39 −0.39 −0.24 −0.43 −0.52 −0.26 −0.44 −0.60 −0.46 −0.47

−0.53 −0.45 −0.26 −0.54 −0.17 −0.22 −0.68

10. CONCLUSIONS AND OUTLOOK Theoretical descriptions of the interactions between electrons and DNA subunits are approaching reliability for DNA segments, including biological entities such as nucleotide oligomers and duplexes of nucleotide oligomers. Electron affinities of nucleic acid bases predicted by state of art wave function methods (such as G4, SPT2, CCSD(T), and BD(T) with large basis sets) and the carefully calibrated density functionals (such as B3LYP, M05-2X, M06-2X, with medium size basis sets) match reliable photoelectron spectroscopy observations with reasonable accuracy. The predictions of the electron affinities of nucleosides and nucleotides by the DFT have been successfully reproduced in subsequent photoelectron experiments. These studies validate the applications of DFT methods in exploring the electrons interacting with yet larger biomolecules in DNA. Formation of dipole-bound states of radical anion is common for nucleic acid bases in the gas phase, due to their relatively large dipole moments (with the exception of adenine). Valence anions are found to be less favored than the neutral species for guanine and adenine (perhaps even cytosine, where the VDE is positive). Except for thymine and uracil, the electron affinities of the nucleobases in the canonical forms are small or zero. Various tautomeric forms of the nucleic acid bases, especially guanine, have been found to be important electron acceptors in the gas phase. Interactions with their surroundings, such as water clusters, significantly increase the electron capture abilities of the bases. The effects of the polarizable medium on electron capture by the bases are critically important. When there is no direct interaction between solvent and solute, such as proton transfer between solvent and solute, the effects of the polarizable continuum override the effects of microsolvation. The electron affinities of nucleosides are larger than those of nucleobases, due to the ability to delocalize the negative charge over a larger molecular space. The excess electron is primarily located near the π* orbital of the nucleobases in dT•−, dC•−, and dA•−. In dG•−, the excess electron is largely dipole-bound to the guanine moiety. Electron attachment to nucleoside monophosphates results in viable base-centered radical anions for the pyrimidines. Although it is possible to form the phosphate-centered radical anions in dCMP or dTMP, they are less favorable. Electron attachment to dAMP may induce proton transfer from the phosphate group to the base moiety, resulting in a stable basecentered radical and phosphate-centered anion. Electron attachment to dGMP leads to a viable dipole bound state of the radical anion, with the excess electron mainly distributed around the base moiety. Microsolvation with discrete water molecules around guanine does not change the DB state of this anion.

job of predicting vertical detachment energies. By extension, the same method should reliably predict VEA and AEA values. Finally, we move to the sticky question of whether thymine has an adiabatic electron affinity (AEA). Comparison of the second and third columns of Table 21 shows that zero-point vibrational energies always increase the classical (equilibrium geometry of T•− to equilibrium geometry of T) AEAs. This is because molecular anions are typically more floppy than the corresponding neutrals. The consequence is that the ZPVE of the anion is less than that of the neutral. As a result, all but five of the 40 DFT methods yield a positive AEA for thymine. The B3LYP method used so frequently predicts a ZPVE corrected AF


Chemical Reviews

Review

may be the initial states for many electron attachment induced chemical processes, such as bond breakage in DNA, as suggested by Kumar and Sevilla.228,233 Study of the excited states of radical anion DNA subunits could lead to a better understanding of the formation of transient negative ions (TNI) of DNA constituents238−245 during dissociative electron attachment.

The electron affinities of the pyrimidine nucleoside diphosphates are close to those of dCMP and dTMP. The excess electron is largely located on the base moiety in 3′,5′dCDP•− and 3′,5′-dTDP•− . However, the contribution of the phosphate group at the 3′-position in hosting the excess electron is of vital importance. About 20−30% of the negative charge is found on the phosphate group at the 3′-position in 3′,5′-dCDP•− and 3′,5′-dTDP•−, while ca. 63% of the negative charge is on the bases. The unpaired electron density is partly spread around the 3′-phosphate. The important contribution of the phosphate at the 3′-position for hosting the excess electron is also found in 3′,5′-dADP•−, in which about 46% of the negative charge is located near the 3′-phosphate group. The radical anion 3′,5′-dGDP•− is dipole bound in the gas phase. In aqueous solution, it can be either a viable guanine basecentered radical anion or a higher energy 3′-phosphate centered valence radical anion. However, in guanine rich aqueous nucleotide oligomers, the phosphate-centered valence anion might compete with the base-centered anion. Of course, the latter assumes there is a counterion near the deprotonated phosphate group. In aqueous solution, the phosphate group is deprotonated and bears a negative charge. One important finding of the studies of electron attachment to nucleotides is that, in aqueous solution, the electron capture ability of the bases is almost independent of the existence of a counterion. The position of cytosine in the sequence of an oligomer affects the cytosine electron capture ability. This is mainly due to the formation of intrastrand H-bonds between the neighboring bases. Stacking between G and C in nucleotide oligomers does not have an important influence on the electron capture ability of cytosine. The ultimate electron affinity of cytosine in DNA single strands in aqueous solutions should be ca. 2 eV. The excess electron in the radical anions of pyrimidine− purine pairs is located mainly on the pyrimidine base. Electron attachment to the Watson−Crick GC pair has strong potential to trigger the interstrand proton transfer from G to C. This tendency has been found in either paired nucleic acid bases, or paired nucleosides, or paired nucleotide oligomers, forming the more favorable charge-radical center separated distonic radical anions. In the GC paired nucleotide oligomer, this distonic radical anion could capture another electron to form a stable dianion in aqueous solution. However, electron attachment to the Watson−Crick AT pair does not trigger such interstrand proton transfer. Electron attachment to nucleotides may trigger either C―O bond cleavage or N-glycosidic bond rupture. The mechanisms for the electron attachment induced bond breaks in DNA have been examined from different viewpoints. These mechanisms explain the results of experimental studies of low energy electron induced DNA damages. There are two final points that require our attention: (1) the similar electron affinities of phosphate (in sugar−phosphate− sugar models) and the purine bases (A and G) in aqueous solution suggest that these three sites might be equally good candidates for electron acceptors in nucleotide oligomers. This implies that the excess electron might locate and relocate on different sites in nucleotide oligomers.90 The study of the distributions of different centered radical anions and the relocation of the excess electron from one site to another could be important in understanding electron attachment related charge transfer and signal transfer through DNA. (2) Electron attachment to DNA subunit induced excited electronic states

AUTHOR INFORMATION Corresponding Author

*E-mail addresses: J.G., [email protected]; H.F.S., qc@uga. edu. Notes

The authors declare no competing financial interest. Biographies

Jiande Gu received his B.S. and Ph.D. (1995) degrees in chemistry from Sichuan University, Chengdu, China. He was a postdoctoral research fellow with Professor Kaixian Chen at Shanghai Institute of Materia Medica, Chinese Academy of Sciences, from 1995 to 1997. Between 1998 and 2000, he was a postdoctoral research fellow with Professor F. Albert Cotton at Texas A&M University. Dr. Gu has been a professor of chemistry at the Shanghai Institute of Materia Medica, Chinese Academy of Sciences, since 1998. Dr. Gu is a frequent visiting professor at the Center for Computational Chemistry, University of Georgia, and the Computational Center for Molecular Structure and Interactions (Now Interdisciplinary Nanotoxicity Center), Jackson State University. His primary research interests lie in the area of theoretical biochemistry.

Jerzy Leszczynski, Professor of Chemistry and President’s Distinguished Fellow at the Jackson State University (JSU), joined the AG


Chemical Reviews

Review

PD-PES

photodetachment−photoelectron spectroscopy LEPET low-energy photoelectron transmission HF Hartree−Fock UHF spin-unrestricted open-shell Hartree− Fock SCF self-consistent-field MP2 Møller−Plesset correlation energy truncated at second-order perturbation theory PMP2 projected MP2 MP4 Moller−Plesset fourth-order perturbation theory with single, double, triple, and quadruple excitations DFT density functional theory B3LYP the correlation functional of Lee, Yang, and Parr in conjunction with Becke’s three-parameter HF/DFT exchange functional CCSD(T) coupled-cluster theory with single, double, and perturbative triple excitations SA CCSD(T) spin-adaption included CCSD(T) PPP Pariser−Parr−Pople AM1 Austin Model 1 MO molecular orbital SOMO singly occupied molecular orbital CASPT2 multiconfigurational perturbation methods G4 Gaussian-4 theory DB dipole bound BD(T) Brueckner doubles method with a triples contribution PCM Barone−Tomasi polarizable continuum model IPCM PCM uses a static isodensity surface for the cavity DTQ5 double−triple−quadruple−quintuple ζ 5XU 5-halouracil S−P−S sugar−phosphate−sugar AT pair adenine−thymine pair AU pair adenine−uracil pair GC pair guanine−cytosine pair MAMT (9-methyladenine)(1-methylthymine) pair MGMC (9-methylguanine)(1-methylcytosine) pair GH hydrogenated guanine dG 2′-deoxyriboguanosine dA 2′-deoxyriboadenosine dT 2′-deoxyribothymidine dC 2′-deoxyribocytidine dAdT nucleoside pair 2′-deoxyriboadenosine-2′-deoxyribothymidine pair dGdC nucleoside pair 2′-deoxyriboguanosine-2′-deoxyribocytidine pair 3′,5′-dCDP 2′-deoxycytidine-3′,5′-diphosphate 3′,5′-dTDP 2′-deoxythymidine-3′,5′-diphosphate 3′,5′-dGDP 2′-deoxyguanosine-3′,5′-diphosphate 3′,5′-dADP 2′-deoxyadenosine-3′,5′-diphosphate dApdT dinucleoside phosphate deoxyadenylyl3′,5′-deoxythymidine

faculty of the JSU Department of Chemistry in 1990. Dr. Leszczynski attended the Technical University of Wroclaw (TUW) in Wroclaw, Poland, obtaining his M.S. (1972) and Ph.D. (1975) degrees. He had served as the director for the Computational Center for Molecular Structure and Interactions (NSF-CREST Center). Since October 2008, Dr. Leszczynski has directed the Interdisciplinary Nanotoxicity CREST Center at JSU. He is the recipient of the White House Millennium Award for Teaching and Research Excellence in Mathematics, Science, and Engineering. Other selected awards include the following: the Maria Sklodowska-Curie’s Medal, Polish Chemical Society, 2007, and the USA Presidential Award for Excellence in Science, Mathematics, and Engineering Mentoring, 2009. He is a member of the European Academy of Sciences and the European Academy of Sciences, Arts and Humanities, 2004.

Henry F. Schaefer III received his B.S. degree in Chemical Physics from M.I.T. in 1966 and his Ph.D. from Stanford University in 1969. The same year, he joined the faculty of the University of California at Berkeley, where he was a professor for 18 years. Since 1987, Dr. Schaefer has been Graham Perdue Professor at the University of Georgia and Director of the Center for Computational Quantum Chemistry. He has received four awards from the American Chemical Society (Pure Chemistry, Baekeland, Remsen, and Theoretical Chemistry), plus the Centenary Medal from London’s Royal Society of Chemistry. He is a Fellow of the American Academy of Arts and Sciences.

ACKNOWLEDGMENTS This research was supported by the U.S. National Science Foundation, Grant CHE-1054286, and CREST Grant No. HRD-08. In Shanghai the work was supported by the National Science & Technology Major Project ‘Key New Drug Creation and Manufacturing Program’, China (Number: 2009ZX09301001). ABBREVIATIONS EA AEA VDE VAE, VEA ZPVE, ZPE TS A G C T U RET

electron affinity adiabatic electron affinity vertical detachment energy vertical attachment energy zero-point vibrational energy transition state adenine guanine cytosine thymine uracil Rydberg electron transfer AH


Chemical Reviews dTpdA dGpdC dCpdG dGpdG (G−H1)−(C+H3)• LEE [dGpdC]2 TD-DFT TNI

Review

J., Eds.; A Comprehensive Theoretical and Experimental Analysis (Challenges and Advances in Computational Chemistry and Physics); Springer: 2008; pp 531−575. (29) Kumar, A.; Sevilla, M. D. Radiation effects on DNA: Theoretical Investigations of Electron, Hole and Excitation Pathways to DNA Damage. In Radiation Induced Molecular Phenomena in Nucleic Acids; Shukla, M., Leszczynski, J., Eds.; A Comprehensive Theoretical and Experimental Analysis (Challenges and Advances in Computational Chemistry and Physics); Springer: 2008; pp 577−617. (30) Kumar, A.; Sevilla, M. D. In Low Energy Electron (LEE) Induced DNA Damage: Theoretical Approaches to Modeling Experiment; Shukla, M., Leszczynski, J., Eds.; Handbook of Computational Chemistry Vol. III: ApplicationsBiomolecules; Springer: 2012. (31) Rienstra-Kiracofe, J. C.; Tschumper, G. S.; Schaefer, H. F.; Nandi, S.; Ellison, G. B. Chem. Rev. 2002, 102, 231. (32) Schiedt, J.; Weinkauf, R.; Neumark, D. M.; Schlag, E. W. Chem. Phys. 1998, 239, 511. (33) Saenger, W. Principles of Nucleic Acid Structure; Springer-Verlag: New York, 1984. (34) Wiley, J. R.; Robinson, J. M.; Ehdaie, S.; Chen, E. C. M.; Chen, E. S. D.; Wentworth, W. E. Biochem. Biophys. Res. Commun. 1991, 180, 841. (35) Desfrancois, C.; Abdoul-Carime, H.; Schermann, J. P. J. Chem. Phys. 1996, 104, 7792. (36) Desfrancois, C.; Abdoul-Carime, H.; Carles, S.; Periquet, V.; Schermann, J. P.; Smith, D. M. A.; Adamowicz, L. J. Chem. Phys. 1999, 110, 17876. (37) Desfrancois, C.; Periquet, V.; Bouteiller, Y.; Schermann, J. P. J. Phys. Chem. A 1998, 102, 1274. (38) Desfrancois, C.; Abdoul-Carime, H.; Schermann, J. P. Int. J. Mod. Phys. B 1996, 10, 1339. (39) Desfrancois, C.; Carles, S.; Schermann, J. P. Chem. Rev. 2000, 100, 3943. (40) Hendricks, J. H.; Lyapustina, S. A.; de Clercq, H. L.; Snodgrass, J. T.; Bowen, K. H. J. Chem. Phys. 1996, 104, 7788. (41) Hendricks, J. H.; Lyapustina, S. A.; de Clercq, H. L.; Bowen, K. H. J. Chem. Phys. 1998, 108, 8. (42) Eustis, S.; Wang, D.; Lyapustina, S.; Bowen, K. H. J. Chem. Phys. 2007, 127, 224309. (43) Haranczyk, M.; Gutowski, M.; Li, X.; Bowen, K. H. J. Phys. Chem. B 2007, 111, 14073. (44) Haranczyk, M.; Gutowski, M.; Li, X.; Bowen, K. H. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 4804. (45) Li, X.; Bowen, K. H.; Haranczyk, M.; Mazurkiewicz, K.; Rak, J.; Gutowski, M. J. Chem. Phys. 2007, 127, 174309. (46) Gutowski, M.; Dabkowska, I.; Rak, J.; Xu, S.; Nilles, J. M.; Radisic, D.; Bowen, K. H. Eur. Phys. J. D 2002, 20, 431. (47) Haranczyk, M.; Dabkowska, I.; Rak, J.; Gutowski, M.; Nilles, J. M.; Stokes, S.; Radisic, D.; Bowen, K. H. J. Phys. Chem. B 2004, 108, 6919. (48) Haranczyk, M.; Bachorz, R.; Rak, J.; Gutowski, M.; Radisic, D.; Stokes, S. T.; Nilles, J. M.; Bowen, K. H. J. Phys. Chem. B 2003, 107, 7889. (49) Haranczyk, M.; Rak, J.; Gutowski, M.; Radisic, D.; Stokes, S. T.; Nilles, J. M.; Bowen, K. H. J. Phys. Chem. B 2005, 109, 13383. (50) Mazurkiewicz, K.; Haranczyk, M.; Gutowski, M.; Rak, J.; Radisic, D.; Eustis, S. N.; Wang, D.; Bowen, K. H. J. Am. Chem. Soc. 2007, 129, 1216. (51) Storoniak, P.; Mazurkiewicz, K.; Haranczyk, M.; Gutowski, M.; Rak, J.; Eustis, S. N.; Ko, Y. J.; Wang, H.; Bowen, K. H. J. Phys. Chem. B 2010, 114, 11353. (52) Szyperska, A.; Rak, J.; Leszczynski, J.; Li, X.; Ko, Y. J.; Wang, H.; Bowen, K. H. ChemPhysChem 2010, 11, 880. (53) Szyperska, A.; Rak, J.; Leszczynski, J.; Li, X.; Ko, Y. J.; Wang, H.; Bowen, K. H. J. Am. Chem. Soc. 2009, 131, 2663. (54) Stokes, S. T.; Li, X.; Grubisic, A.; Ko, Y. J.; Bowen, K. H. J. Chem. Phys. 2007, 127, 84321. (55) Ray, S. G.; Daube, S. S.; Naaman, R. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 15.

dinucleoside phosphate deoxythymidylyl-3′,5′-deoxyadenosine dinucleoside phosphate deoxyguanylyl3′,5′-deoxycytidine dinucleoside phosphate deoxycytidylyl3′,5′-deoxyguanosine dinucleoside phosphate deoxyguanylyl3′,5′-deoxyguanosine radical anion of N1-deprotonated guanine-N3-protonated cytosine pair low energy electron dinucleoside phosphate deoxygaunylyl3′,5′-deoxycytidine dimer time-dependent DFT method transient negative ions

REFERENCES (1) Sanche, L. Radiat. Phys. Chem. 1989, 34, 15. (2) Uehara, U.; Nikjoo, H.; Goodhead, D. T. Radiat. Res. 1999, 152, 202. (3) LaVerne, J. A.; Pimblott, S. M. J. Phys. Chem. 1995, 99, 10540. (4) LaVerne, J. A.; Pimblott, S. M. Radiat. Res. 1995, 141, 208. (5) Sanche, L. Eur. Phys. J. D 2005, 35, 367. (6) Ward, J. F. In Advances in Radiation Biology 5; Lett, J. T., Adler, H., Eds.; Academic Press: New York, 1977; pp 181−239. (7) Yamanoto, O. In Aging, Carcinogenesis and Radiation Biology; Smith, K., Ed.; Plenum: New York, 1976; pp 165−192. (8) Fucarelli, A. F., Zimbrick, J. D., Eds.; Radiation Damage in DNA: Structure/Function Relationships at Early Times; Battelle: Columbus, OH, 1995. (9) von Sonntag, C. The Chemical Basis for Radiation Biology; Taylor and Francis: London, 1987. (10) von Sonntag, C. Adv. Quantum Chem. 2007, 52, 5. (11) Becker, D.; Sevilla, M. D. The Chemical Consequences of Radiation Damage to DNA. In Advances in Radiation Biology, Vol. 17; Lett, J., Ed.; Academic Press: New York, 1993; pp 121−180. (12) Boudaiffa, B.; Cloutier, P.; Hunting, D.; Huels, M. A.; Sanche, L. Science 2000, 287, 1658. (13) Martin, F.; Burrow, P. D.; Cai, Z.; Coultier, P.; Hunting, D.; Sanche, L. Phys. Rev. Lett. 2004, 93, 068101. (14) Pan, X.; Cloutier, P.; Hunting, D.; Sanche, L. Phys. Rev. Lett. 2003, 90, 208102. (15) Caron, L. G.; Sanche, L. Phys. Rev. Lett. 2003, 91, 113201. (16) Zheng, Y.; Cloutier, P.; Hunting, D.; Wagner, J. R.; Sanche, L. J. Am. Chem. Soc. 2004, 126, 1002. (17) Huels, M. A.; Boudaiffa, B.; Cloutier, P.; Hunting, D.; Sanche, L. J. Am. Chem. Soc. 2003, 125, 4467. (18) Abdoul-Carime, H.; Gohlke, S.; Fischbach, E.; Scheike, J.; Illenberger, E. Chem. Phys. Lett. 2004, 387, 267. (19) Zheng, Y.; Cloutier, P.; Hunting, D. J.; Sanche, L.; Wagner, J. R. J. Am. Chem. Soc. 2005, 127, 16592. (20) Zheng, Y.; Cloutier, P.; Hunting, D. J.; Wagner, J. R.; Sanche, L. J. Chem. Phys. 2006, 124, 064710. (21) Li, Z.; Zheng, Y.; Cloutier, P.; Sanche, L.; Wagner, J. R. J. Am. Chem. Soc. 2008, 130, 5612. (22) Abdoub-Carime, H.; Sanche, L. Int. J. Radiat. Biol. 2002, 78, 89. (23) Sanche, L. Scanning Microsc. 1995, 9, 619. (24) Sanche, L. Mass Spectrom. Rev. 2002, 21, 349. (25) Steenken, S.; Telo, J. P.; Novais, H. M.; Candeias, L. P. J. Am. Chem. Soc. 1992, 114, 4701. (26) Colson, A. O.; Sevilla, M. D. Int. J. Radiat. Biol. 1995, 67, 627. (27) Sanche, L. Low-Energy Electron Interaction with DNA: Bond Dissociation and Formation of Transient Anions, Radicals, and Radical Anions. In Radical and Radical Ion Reactivity in Nucleic Acid Chemistry; Greenberg, M. M., Ed.; John Wiley & Sons: 2010; pp 239−293. (28) Sanche, L. Low Energy Electron Damage to DNA. In Radiation Induced Molecular Phenomena in Nucleic Acids; Shukla, M., Leszczynski, AI


Chemical Reviews

Review

(56) Naaman, R.; Sanche, L. Chem. Rev. 2007, 107, 1553. (57) Younkin, J. M.; Smith, L. J.; Compton, R. N. Theor. Chim. Acta (Berlin) 1976, 41, 157. (58) Colson, A.; Besler, B.; Close, D. M.; Sevilla, M. D. J. Phys. Chem. 1992, 96, 661. (59) Colson, A.; Besler, B.; Sevilla, M. D. J. Phys. Chem. 1992, 96, 9787. (60) Colson, A.; Besler, B.; Sevilla, M. D. J. Phys. Chem. 1993, 97, 13852. (61) Sevilla, M. D.; Brent, B.; Colson, A. J. Phys. Chem. 1994, 98, 2215. (62) Sevella, M. D.; Besler, B.; Colson, A. J. Phys. Chem. 1995, 99, 1060. (63) Oyler, N. A.; Adamowicz, L. J. Phys. Chem. 1993, 97, 11122. (64) Oyler, N. A.; Adamowicz, L. Chem. Phys. Lett. 1994, 219, 223. (65) Russo, N.; Toscano, M.; Grand, A. J. Comput. Chem. 2000, 21, 1243. (66) Wetmore, S. D.; Boyd, R. J.; Eriksson, L. A. Chem. Phys. Lett. 2000, 322, 129. (67) Saettel, N. J.; Wiest, O. J. Am. Chem. Soc. 2001, 123, 2693. (68) Wesolowski, S. S.; Leininger, M. L.; Pentchev, P. N.; Schaefer, H. F. J. Am. Chem. Soc. 2001, 123, 4023. (69) Richardson, N. A.; Gu, J.; Wang, S.; Xie, Y.; Schaefer, H. F. J. Am. Chem. Soc. 2004, 126, 4404. (70) Bao, X.; Wang, J.; Gu, J.; Leszczynski, J. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 5658. (71) Gu, J.; Wang, J.; Leszczynski, J. J. Am. Chem. Soc. 2006, 128, 9322. (72) Gu, J.; Xie, Y.; Schaefer, H. F. J. Am. Chem. Soc. 2006, 128, 1250. (73) Gu, J.; Xie, Y.; Schaefer, H. F. ChemPhysChem 2006, 7, 1885. (74) Gu, J.; Xie, Y.; Schaefer, H. F. Nucleic Acids Res. 2007, 35, 5165. (75) Reynisson, J.; Steenken, S. Phys. Chem. Chem. Phys. 2002, 4, 5353. (76) Richardson, N. A.; Wesolowski, S. S.; Schaefer, H. F. J. Phys. Chem. B 2003, 107, 848. (77) Li, X.; Cai, Z.; Sevilla, M. D. J. Phys. Chem. A 2002, 106, 9345. (78) Kumar, A.; Knapp-Mohammady, M.; Mishra, P. C.; Suhai, S. J. Comput. Chem. 2004, 25, 1047. (79) Radisic, D.; Bowen, K. H.; Dakowska, I.; Storoniak, P.; Rak, J.; Gutowski, M. J. Am. Chem. Soc. 2005, 127, 6443. (80) Kumar, A.; Mishra, P. C.; Suhai, S. J. Phys. Chem. A 2005, 109, 3971. (81) Kim, S.; Schaefer, H. F. J. Phys. Chem. A 2007, 111, 10381. (82) Richardson, N. A.; Wesolowski, S. S.; Schaefer, H. F. J. Am. Chem. Soc. 2002, 124, 10163. (83) Li, X.; Cai, Z.; Sevilla, M. D. J. Phys. Chem. B 2001, 105, 10115. (84) Gu, J.; Xie, Y.; Schaefer, H. F. J. Phys. Chem. B 2005, 109, 13067. (85) Gu, J.; Xie, Y.; Schaefer, H. F. J. Chem. Phys. 2007, 127, 155107. (86) Gu, J.; Xie, Y.; Schaefer, H. F. J. Phys. Chem. B 2006, 110, 19696. (87) Kim, S.; Schaefer, H. F. J. Chem. Phys. 2007, 126, 64301. (88) Kim, S.; Wheeler, S. E.; Schaefer, H. F. J. Chem. Phys. 2006, 124, 204310. (89) Kim, S.; Schaefer, H. F. J. Chem. Phys. 2006, 125, 144305. (90) Bao, X.; Sun, H.; Wong, N.-B.; Gu, J. J. Phys. Chem. B 2006, 110, 5865. (91) Bao, X.; Liang, G.; Wong, N.-B.; Gu, J. J. Phys. Chem. A 2007, 111, 666. (92) Gu, J.; Xie, Y.; Schaefer, H. F. Chem.Eur. J. 2010, 16, 5089. (93) Gu, J.; Xie, Y.; Schaefer, H. F. Chem. Phys. Lett. 2009, 473, 213. (94) Gu, J.; Xie, Y.; Schaefer, H. F. Chem. Eur. J. 2012, 18, 5232. (95) Gu, J.; Wang, J.; Leszczynski, J. J. Phys. Chem. B 2012, 116, 1458. (96) Gu, J.; Wong, N.-B.; Xie, Y.; Schaefer, H. F. Chem.Eur. J. 2010, 16, 13155. (97) Compton, R. N.; Yoshioka, Y.; Jordan, K. D. Theor. Chim. Acta (Berlin) 1980, 54, 259. (98) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F. J. Am. Chem. Soc. 1985, 107, 3902. (99) Zhang, Q.; Chen, E. C. M. Biochem. Biophys. Res. Commun. 1995, 217, 755.

(100) Bachorz, R. A.; Klopper, W.; Gutowski, M. J. Chem. Phys. 2007, 126, 085101. (101) Dedikova, P.; Demovic, L.; Pitonak, M.; Neogrady, P.; Urban, M. Chem. Phys. Lett. 2009, 481, 107. (102) Head-Gordon, M.; Pople, J. A.; Frisch, M. J. Chem. Phys. Lett. 1988, 153, 503. (103) Head-Gordon, M.; Head-Gordon, T. Chem. Phys. Lett. 1994, 220, 122. (104) Saebø, S.; Almlöf, J. Chem. Phys. Lett. 1989, 154, 83. (105) Frisch, M. J.; Head-Gordon, M.; Pople, J. A. Chem. Phys. Lett. 1990, 166, 275. (106) Frisch, M. J.; Head-Gordon, M.; Pople, J. A. Chem. Phys. Lett. 1990, 166, 281. (107) Cizek, J. In Advances in Chemical Physics, Vol. 14; LeFebvre, R., Moser, C., Eds.; Wiley Interscience: New York, 1969; p 35. (108) Purvis, G. D.; Bartlett, R. J. J. Chem. Phys. 1982, 76, 1910. (109) Scuseria, G. E.; Janssen, C. L.; Schaefer, H. F. J. Chem. Phys. 1988, 89, 7382. (110) Scuseria, G. E.; Schaefer, H. F. J. Chem. Phys. 1989, 90, 3700. (111) Bachorz, R. A.; Rak, J.; Gutowski, M. Phys. Chem. Chem. Phys. 2005, 7, 2116. (112) Gu, J.; et al. Unpublished. (113) Dykstra, C. E. Chem. Phys. Lett. 1977, 45, 466. (114) Handy, N. C.; Pople, J. A.; Head-Gordon, M.; Raghavachari, K.; Trucks, G. W. Chem. Phys. Lett. 1989, 164, 185. (115) Kobayashi, R.; Handy, N. C.; Amos, R. D.; Trucks, G. W.; Frisch, M. J.; Pople, J. A. J. Chem. Phys. 1991, 95, 6723. (116) Andersson, K.; Malmqvist, P.-A.; Roos, B. O. J. Chem. Phys. 1992, 96, 1218. (117) Roca-Sanjuan, D.; Merchan, M.; Serrano-Andres, L.; Rubio, M. J. Chem. Phys. 2008, 129, 095104. (118) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 2007, 126, 084108. (119) Curtiss, L. A.; Redfern, P. C.; Raghavachari, K. J. Chem. Phys. 2007, 127, 124105. (120) Galbraith, J. M.; Schaefer, H. F. J. Chem. Phys. 1996, 105, 862. (121) Tschumper, G. S.; Schaefer, H. F. J. Chem. Phys. 1997, 107, 2529. (122) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785. (123) Miehlich, B.; Savin, A.; Stoll, H.; Preuss, H. Chem. Phys. Lett. 1989, 157, 200. (124) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (125) Perdew, J. P. Phys. Rev. B 1986, 33, 8822. (126) Backe, A. D. Phys. Rev. A 1988, 38, 3098. (127) Backe, A. D. J. Chem. Phys. 1993, 98, 1372. (128) Perdew, J. P. Phys. Rev. B 1986, 34, 7406. (129) Perdew, J. P.; Wang, Y. Phys. Rev. B 1992, 45, 13244. (130) Zhao, Y.; Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput. 2006, 2, 364. (131) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215. (132) Zhao, Y.; Truhlar, D. G. Chem. Phys. Lett. 2011, 502, 1. (133) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157. (134) Gu, J.; Wang, J.; Leszczynski, J.; Xie, Y.; Schaefer, H. F. Chem. Phys. Lett. 2008, 459, 164. (135) Gu, J.; Wang, J.; Leszczynski, J. Chem. Phys. Lett. 2011, 512, 108. (136) Siefermann, K. R.; Abel, B. Agew. Chem., Int. Ed. 2011, 50, 5264. (137) Sanche, L. Nature 2009, 461, 358. (138) Wang, C.-R.; Nguyen, J.; Lu, Q.-B. J. Am. Chem. Soc. 2009, 131, 11320. (139) Wang, C.-R.; Lu, Q.-B. J. Am. Chem. Soc. 2010, 132, 14710. (140) Migus, A.; Gauduel, Y.; Martin, J. L.; Antonetti, A. Phys. Rev. Lett. 1987, 58, 1559. (141) Rossky, P. J.; Schnitker, J. J. Phys. Chem. 1988, 92, 4277. (142) Long, F. H.; Lu, H.; Eisenthal, K. B. Phys. Rev. Lett. 1990, 64, 1469. (143) Laenen, R.; Roth, T.; Laubereau, A. Phys. Rev. Lett. 2000, 85, 50. AJ


Chemical Reviews

Review

(144) Borgis, D.; Rossky, P. J.; Turi, L. J. Chem. Phys. 2007, 127, 174508. (145) Wang, C.-R.; Luo, T.; Lu, Q.-B. Phys. Chem. Chem. Phys. 2008, 10, 4463. (146) Orozco, M.; Luque, F. J. Chem. Rev. 2000, 100, 4187. (147) Cossi, M.; Barone, V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996, 255, 327. (148) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105, 2999. (149) Gu, J.; Xie, Y.; Schaefer, H. F. J. Phys. Chem. B 2010, 114, 1221. (150) Li, X.; Cai, Z.; Sevilla, M. D. J. Phys. Chem. A 2002, 106, 1596. (151) Wetmore, S. D.; Boyd, R. J.; Eriksson, L. A. J. Phys. Chem. B 1998, 102, 9332. (152) Haranczyk, M.; Gutowski, M. Angew. Chem., Int. Ed. 2005, 44, 6585. (153) Roehrig, G. J.; Oyler, N. A.; Adamowicz, L. J. Phys. Chem. 1995, 99, 14285. (154) Smith, D. M. A.; Jalbout, A. F.; Smets, J.; Adamowicz, L. Chem. Phys. 2000, 260, 45. (155) Mazurkiewicz, K.; Bachorz, R. A.; Gutowski, M.; Rak, J. J. Phys. Chem. B 2006, 110, 24696. (156) Li, X.; Sanche, L.; Sevilla, M. D. J. Phys. Chem. A 2002, 106, 11248. (157) Wetmore, S. D.; Boyd, R. J.; Eriksson, L. Chem. Phys. Lett. 2001, 343, 151. (158) Li, X.; Sevilla, M. D.; Sanche, L. J. Am. Chem. Soc. 2003, 125, 13688. (159) Liang, G.; Gu, J.; Xie, Y.; Schaefer, H. F. Unpublished. (160) Kumar, A.; Sevilla, M. D. Chem. Rev. 2010, 110, 7002. (161) Steenken, S. Chem. Rev. 1989, 89, 503. (162) Jena, N. R.; Mishra, P. C. J. Phys. Chem. B 2007, 111, 5418. (163) (a) Luo, Q.; Li, Q.; Xie, Y.; Schaefer, H. F. Collect. Czech. Chem. Commun. 2005, 70, 826. (b) Xie, H.; Cao, Z. Int. J. Quantum Chem. 2007, 107, 1261. (164) Evangelista, F. A.; Paul, A.; Schaefer, H. F. J. Phys. Chem. A 2004, 108, 3565. (165) Luo, Q.; Li, Q.; Kim, S.; Wheeler, S. E.; Xie, Y.; Schaefer, H. F. Phys. Chem. Chem. Phys. 2005, 7, 861. (166) Profeta, L. T. M.; Larkin, J. D.; Schaefer, H. F. Mol. Phys. 2003, 101, 3277. (167) Jiao, D.; Wang, H. Mol. Phys. 2008, 106, 2653. (168) Li, X.; Sanche, L.; Sevilla, M. D. J. Phys. Chem. B 2004, 108, 5472. (169) Jalbout, A. F.; Adamowicz, L. J. Phys. Chem. A 2001, 105, 1033. (170) Smets, J.; McCarthy, W. J.; Adamowicz, L. Chem. Phys. Lett. 1996, 256, 360. (171) Frigato, T.; Svozil, D.; Jungwirth, P. J. Phys. Chem. A 2006, 110, 2916. (172) Morgado, C. A.; Pichugin, K. Y.; Adamowicz, L. Phys. Chem. Chem. Phys. 2004, 6, 2758. (173) Smets, J.; McCarthy, W. J.; Adamowicz, L. J. Phys. Chem. 1996, 100, 14655. (174) Smets, J.; Smith, D. M. A.; Elkadi, Y.; Adamowicz, L. J. Phys. Chem. 1997, 101, 9152. (175) Dolgounitcheva, O.; Zakrzewski, V. G.; Ortiz, J. V. J. Phys. Chem. A 1999, 103, 7912. (176) Dedikova, P.; Neogrady, P.; Urban, M. J. Phys. Chem. A 2011, 115, 2350. (177) Li, D.; Ai, H. J. Phys. Chem. B 2009, 113, 11732. (178) Jalbout, A. F.; Adamowicz, L. J. Mol. Struct. 2002, 605, 93. (179) Mazurkiewicz, K.; Haranczyk, M.; Storoniak, R.; Gutowski, M.; Rak, J.; Radisic, D.; Eustis, S. N.; Wang, D.; Bowen, K. H. Chem. Phys. 2007, 342, 215. (180) Yates, B. F.; Bouma, W. J.; Radom, L. J. Am. Chem. Soc. 1984, 106, 5805. (181) Al-Jihad, I.; Smets, J.; Adamowicz, L. J. Phys. Chem. A 2000, 104, 2994. (182) Stepanian, S. G.; Jalbout, A. F.; Hall, C. S.; Adamowicz, L. J. Phys. Chem. A 2003, 107, 7911.

(183) Li, X.; Sevilla, M. D.; Sanche, L. J. Am. Chem. Soc. 2003, 125, 8916. (184) Kobylecka, M.; Leszczynski, J.; Rak, J. J. Am. Chem. Soc. 2008, 130, 15683. (185) Szyperska, A.; Gajewicz, A.; Mazurkiewicz, K.; Leszczynski, J.; Rak, J. Phys. Chem. Chem. Phys. 2011, 13, 19499. (186) Smets, J.; Jalbout, A. F.; Adamowicz, L. Chem. Phys. Lett. 2001, 342, 342. (187) Kobylecka, M.; Leszczynski, J.; Rak, J. J. Chem. Phys. 2009, 131, 085103. (188) Kumar, A.; Sevilla, M. D.; Suhai, S. J. Phys. Chem. B 2008, 112, 5189. (189) Tian, S. X. J. Phys. Chem. A 2005, 109, 5153. (190) Zhang, J. D.; Chen, Z.; Schaefer, H. F. J. Phys. Chem. A 2008, 112, 6217. (191) Zhang, J. D.; Schaefer, H. F. J. Chem. Theory Comput. 2007, 3, 115. (192) Zhang, J. D.; Xie, Y.; Schaefer, H. F. J. Phys. Chem. A 2006, 110, 12010. (193) Kumar, A.; Mishra, P. C.; Suhai, S. J. Phys. Chem. A 2006, 110, 7719. (194) Lind, M. C.; Richardson, N. A.; Wheeler, S. E.; Schaefer, H. F. J. Phys. Chem. B 2007, 111, 5525. (195) Kim, S.; Lind, M. C.; Schaefer, H. F. J. Phys. Chem. B 2008, 112, 3545. (196) Xie, H.; Xia, F.; Cao, Z. J. Phys. Chem. A 2007, 111, 4384. (197) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem. 2003, 24, 669. (198) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995. (199) Lind, M. C.; Bera, P. P.; Richardson, N. C.; Wheeler, S. E.; Schaefer, H. F. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 7554. (200) Zheng, Y.; Cloutier, P.; Hunting, D. J.; Wagner, J. R.; Sanche, L. J. Am. Chem. Soc. 2004, 126, 1002. (201) Gu, J.; Xie, Y.; Schaefer, H. F. J. Am. Chem. Soc. 2005, 127, 1053. (202) Li, X.; Sanche, L.; Sevilla, M. D. Radiat. Res. 2006, 165, 721. (203) Liang, G.; Bao, X.; Gu, J. J. Comput. Chem. 2008, 29, 2648. (204) Evangelista, F. A.; Schaefer, H. F. J. Phys. Chem. A 2004, 108, 10258. (205) Barrios, R.; Skurski, P.; Simons, J. J. Phys. Chem. B 2002, 106, 7991. (206) Barrios, R.; Anusiewicz, I.; Skurski, P.; Simons, J. J. Phys. Chem. A 2004, 108, 2999. (207) Berdys, J.; Anusiewicz, I.; Skurski, P.; Simons, J. J. Am. Chem. Soc. 2004, 126, 6441. (208) Berdys, J.; Skurski, P.; Simons, J. J. Phys. Chem. B 2004, 108, 5800. (209) Anusiewicz, I.; Berdys, J.; Sobczyk, M.; Skurski, P.; Simons, J. J. Phys. Chem. A 2004, 108, 11381. (210) Kumar, A.; Sevilla, M. D. J. Phys. Chem. B 2007, 111, 5464. (211) Rak, J.; Kobylecka, M.; Storoniak, P. J. Phys. Chem. B 2011, 115, 1911. (212) Schyman, P.; Laaksonen, A. J. Am. Chem. Soc. 2008, 130, 12254. (213) Kobylecka, M.; Gu, J.; Rak, J.; Leszczynski, J. J. Chem. Phys. 2008, 128, 44315. (214) Stokes, S. T.; Grubisic, A.; Li, X.; Ko, Y. J.; Bowen, K. H. J. Chem. Phys. 2008, 128, 44314. (215) Foresman, J. B.; Keith, T. A.; Wiberg, K. B.; Snoonian, J.; Frisch, M. J. J. Phys. Chem. 1996, 100, 16098. (216) Chen, H.-Y.; Kao, C.-L.; Hsu, S. C. N. J. Am. Chem. Soc. 2009, 131, 15930. (217) Maseras, F.; Morokuma, K. J. Comput. Chem. 1995, 16, 1170. (218) Humbel, S.; Sieber, S.; Morokuma, K. J. Chem. Phys. 1996, 105, 1959. (219) Matsubara, T.; Sieber, S.; Morokuma, K. Int. J. Quantum Chem. 1996, 60, 1101. (220) Svensson, M.; Humbel, S.; Froese, R. D. J.; Matsubara, T.; Sieber, S.; Morokuma, K. J. Phys. Chem. 1996, 100, 19357. AK


Chemical Reviews

Review

(221) Svensson, M.; Humbel, S.; Morokuma, K. J. Chem. Phys. 1996, 105, 3654. (222) Dapprich, S.; Komaromi, I.; Byun, K. S.; Morokuma, K.; Frisch, M. J. J. Mol. Struct. 1999, 462, 1. (223) Vreven, T.; Morokuma, K. J. Comput. Chem. 2000, 21, 1419. (224) Gu, J.; Wang, J.; Rak, J.; Leszczynski, J. Angew. Chem., Int. Ed. 2007, 46, 3479. (225) Dabkowska, I.; Rak, J.; Gutowski, M. Eur. Phys. J. D 2005, 35, 429. (226) Lyngdoh, R. H. D.; Schaefer, H. F. Acc. Chem. Res. 2009, 42, 563. (227) (a) Simons, J. Acc. Chem. Res. 2006, 39, 772. (b) Simons, J. J. Phys. Chem. A 2008, 112, 6401. (228) Kumar, A.; Sevilla, M. D. J. Am. Chem. Soc. 2008, 130, 2130. (229) Gu, J.; Wang, J.; Leszczynski, J. Nucleic Acids Res. 2010, 38, 5280. (230) Gu, J.; Wang, J.; Leszczynski, J. ChemPhysChem 2010, 11, 175. (231) Gu, J.; Wang, J.; Leszczynski, J. J. Phys. Chem. B 2011, 115, 14831. (232) Xie, H.; Wu, R.; Xia, F.; Cao, Z. J. Comput. Chem. 2008, 29, 2025. (233) Kumar, A.; Sevilla, M. D. ChemPhysChem 2009, 10, 1426. (234) Zhang, R.; Zhang, K.; Eriksson, L. A. Chem.Eur. J. 2008, 14, 2850. (235) Caron, L. G.; Sanche, L. Theoretical Studies of Electron Interactions with DNA and its Subunits: from Tetrahydrofuran to Plasmid DNA. In Low-Energy Electron Scattering from Molecules, Biomolecules and Surfaces; Carsky, P., Curik, R., Eds.; CRC Press: Boca Raton, FL, 2011; pp 171−247. (236) Kim, S.; Schaefer, H. F. J. Chem. Phys. 2010, 133, 144305. (237) Li, X.; Wang, H.; Bowen, K. H. J. Chem. Phys. 2010, 133, 144304. (238) Gianturco, F. A.; Lucchese, R. R. J. Chem. Phys. 2004, 120, 7446. (239) Gianturco, F. A.; Sebastianelli, F.; Lucchese, R. R.; Baccarelli, I.; Sanna, N. J. Chem. Phys. 2008, 128, 174302. (240) Winstead, C.; McKoy, V. J. Chem. Phys. 2006, 125, 174304. (241) Winstead, C.; McKoy, V. J. Chem. Phys. 2006, 125, 244302. (242) Winstead, C.; McKoy, V.; Sanchez, S. d’A. J. Chem. Phys. 2007, 127, 85105. (243) Winstead, C.; McKoy, V. J. Chem. Phys. 2006, 125, 74302. (244) Winstead, C.; McKoy, V. Radiat. Phys. Chem. 2008, 77, 1258. (245) Wang, Y.; Tian, S. X. Phys. Chem. Chem. Phys. 2011, 13, 6169.

AL


Interactions of Electrons with Bare and Hydrated Biomolecules: From

Recommend Documents