Oxidation pathways of adenine and guanine in aqueous solution from first principles electrochemistry

The 8-oxo-7,8-dehydropurine tautomers (8-oxoA and 8-oxoG) are mutagenic lesions found in DNA.

Two experimental pathways have been proposed for the formation of 8-oxoG: one initiated by deprotonation of the OH˙ radical adduct at the 8-position of guanine (G8OH˙) and the other initiated by a proton-coupled one-electron oxidation of G8OH˙.

We here report standard Gibbs energies of the above processes involving proton transfer (PT), electron transfer (ET), and proton-coupled electron transfer (PT–ET) reactions calculated from first principles using DFT (B3LYP) and a continuum solvent model (IEF-PCM).

The computed data show that the former pathway is unlikely to occur for A8OH˙ and G8OH˙ in neutral aqueous solution, because of the very low acidity of the hydrogen at the 8-position.

In contrast, the latter route involving proton-coupled one-electron oxidations of A8OH˙ and G8OH˙ are exergonic by about 25 kcal mol−1 in aqueous solution.

Energetically, adenine and guanine behave similarly toward oxidation to yield 8-oxoA and 8-oxoG.

However, the calculated standard Gibbs energetics confirms that the ease of ionization of the native and oxidized forms of nucleobases B to yield the radical cations B˙+ or their deprotonation products B(–H)˙ is 8-oxoG > G > 8-oxoA > A > C > T in aqueous solution.

Consequently, 8-oxoG will most readily trap radical cations and neutral radicals in DNA, since it can reduce any nucleobase radical cation B˙+ (via ET) or its deprotonation product B(–H)˙ (via PT–ET) back to the native form of the nucleobase.


Hydroxyl radicals are formed in the normal oxygen metabolism and as a result of the radiolysis of water when living cells are exposed to ionizing radiation.

About half of the DNA damage caused by OH radicals occurs on nucleobases, where these add to the double bonds of purine bases to yield C4, C5, and C8 radical adducts in adenine (A4OH˙, A5OH˙, and A8OH˙) and guanine (G4OH˙, G5OH˙, and G8OH˙).

Under oxidative conditions, the C8 radical adducts will react further to yield the 8-hydroxypurine derivatives (8-OHA and 8-OHG) and their 8-oxo-7,8-dehydropurine tautomers (8-oxoA and 8-oxoG).1–3

Addition of water to G˙+ is an alternative and feasible pathway to yield 8-oxoG.

This reaction seems to be important in DNA, but experimental evidence suggests that it does not occur in solutions of nucleotides.4

8-oxoA and 8-oxoG are the major stable purine products produced by oxidation and radiolysis.5–8

Even though A and G basically form the same kind of radiation-induced derivatives, 8-oxoA is only formed in about one third the amount of 8-oxoG during aerobic radiation of both naked DNA and chromatin.3

The 8-oxopurine derivatives are highly mutagenic lesions, as mispairing of 8-hydroxyguanine and 8-oxoguanine with adenine brings about G:C to T:A transversions.9–11

Two pathways for oxidation have been proposed (Fig. 1).

The first pathway (Fig. 1, PT branch) is initiated by deprotonation of the 8-hydroxypurine radical at C8 followed by subsequent electron loss and tautomerization,12 whereas the other (Fig. 1, PT–ET branch) is initiated by a proton-coupled one-electron oxidation of the 8-hydroxypurine radical (A8OH˙ and G8OH˙) to yield the 8-hydroxypurine derivative (8-OHA and 8-OHG) followed by tautomerizarion.4,13,14

Based on EPR measurements at 77 K, the PT-initiated pathway has been proposed to occur by photoionization (at 248 nm) of frozen aqueous solutions of DNA.

The experiment showed selectivity for guanine, but was unable to detect the initial radical adduct G8OH˙.

The PT–ET-initiated pathway has been suggested to take place during pulse radiolysis experiments in aqueous solution.

From these experiments the exact route for the formal loss of a hydrogen atom from B8OH˙ is however unclear, i.e., whether it takes place by way of an intermediate radical anion (8-OHA˙ and 8-OHG˙) or a neutral (8-OHA and 8-OHG) precursor.

The 8-oxopurine derivatives (8-oxoA and 8-oxoG) can be subjected to further oxidation (Fig. 1).

However, whereas 8-OHG and 8-oxoG may undergo one-electron oxidation, tautomerization, and deprotonation, depending on the reaction conditions,15 8-oxoA is not as easily oxidized, as seen from the standard reduction potentials of 0.94 and 0.60 V, respectively.3,16

It is claimed that 8-oxoG has a lower ionization potential than any of the nucleobases in their native state17,18 and that it does not distort the structure of the duplex DNA;19,20 as a result, it serves as a deep trap for a migrating electron hole.21,22

Owing to the current experimental uncertainties on the oxidation pathways of purine nucleobases, a theoretical approach may be useful in order to get insight into the redox processes for both A and G in solution.

A redox reaction can be studied theoretically using two different schemes: explicitly modelling the electron transfer reaction between oxidant and reductant, or by modelling the half-reactions independently.

The first method is convenient when examining the mechanistic details of specific redox reactions.

However, in cases in which the general redox features of a species are inferred from the behaviour toward different agents, it is more appropriate to draw conclusions from a half-reaction model.

In particular, the above-mentioned proposals for oxidation pathways of purine nucleobases are based on studies of reactions with different oxidants and experimental techniques that enable conclusions regarding the possible species acting as intermediates and products, as well as common features of the electron transfer processes.

We have previously introduced and applied a first principles scheme to compute Gibbs energies of reactions in which electrons and protons participate as independent ions in solution from absolute electrochemical potentials evaluated at given temperature and pressure.

In doing so, the Gibbs energies of reaction can be directly compared with the experimentally derived values.23,24

The approach has proved very useful for calculating standard Gibbs energies of PT, ET, and PT–ET reactions in aqueous solution involving e−(SHE), e−(aq), H+(aq), H˙(aq).

This work aims at applying DFT and continuum solvation models to determine the feasibility of oxidation of the 8-hydroxypurine radicals (A8OH˙ and G8OH˙) to yield the 8-oxopurine derivatives (8-oxoA and 8-oxoG) in aqueous solution by the two experimentally proposed pathways.

Further oxidation of the 8-oxopurine derivatives is examined in order to test their ability to trap electron holes in DNA.

A comparative analysis of the oxidation pathways for purine, adenine, and guanine is presented to explain the differences that render guanine more sensitive than adenine to oxidative damage.

First-principles standard Gibbs energies of electron transfer (ET), proton transfer (PT), and proton-coupled electron transfer (PT–ET) reactions involved in the oxidation pathways are reported.


Fully optimized geometries, harmonic vibrational frequencies, and zero-point energies of purine (Pu), adenine (A), and guanine (G) and their oxidation derivatives (Fig. 1) were computed with the B3LYP hybrid functional25 and the 6-31 G(d,p) basis set.

In order to obtain correct geometries in the optimizations of the anions,26–28i.e., puckered instead of planar anions,24,28 diffuse functions were added on the heavy atoms through the use of the 6-31+G(d,p) basis set.

However, for the non-anionic structures no diffuse functions were included as the changes in geometries in this case are insignificant in terms of the effect on total energies and vibrational frequencies, as already noted for other compounds.29

Single point energies for all structures were obtained at the B3LYP/6-311+G(2df,p) level in the gas phase and in aqueous solution by including the polarizable continuum model (PCM)30 in the integral equation formalism (IEF).31

The inclusion of diffuse functions in the calculation of Gibbs energies of solvation is a debatable point, since it is claimed that they give too low solvation energies.32,33

However, these effects are smaller than the overall accuracy of our estimates, and hence, exclusion of diffuse functions will not change the qualitative results.

We have thus for consistency chosen to retain the diffuse functions in the data reported throughout the study.

All calculations were performed with the Gaussian 94 and Gaussian 98 programs.34,35

The Gibbs energies of activation for tautomerization were calculated including a water molecule to assist the proton transfer toward ketonization.

This catalyzing water reduced the Gibbs energies of activation by about 30 kcal mol−1, without any drastic changes in the overall thermodynamics of the reactions.

We have therefore included the corrected water-catalyzed Gibbs energies of activation for ketonization to the Gibbs energy diagrams of oxidation of Pu, A, and G.

According to our first principles approach to electrochemistry and ion thermochemistry,23 the absolute standard chemical potential of species B in an ideal-dilute solution μ,aq(B) is split up into the absolute standard chemical potential in the gas phase μ,g(B) plus the bulk Gibbs energy of solvation ΔsG(B):μ,g(B) = μ,g(B) + ΔsG(B).

The absolute standard chemical potential in the ideal gas phase is estimated according to quantum and statistical mechanics as the molar contribution of the translational, rotational, vibrational, electronic, and zero-point vibrational energy of a molecular species B, plus the thermal contributions of bringing the molecules together in the ideal gas phase at given temperature and pressure.23

The Gibbs energies of proton, electron, and proton-coupled electron transfer reactions can then be written in terms of absolute standard chemical potentials at a given temperature and pressure:

(i) For PT: HX(aq) = X−(aq) + H+(aq)2.303RT × pKa = ΔPTG0T = μ0,aqT(X) + μ0,aqT(H+) − μ0,aqT(HX)

(ii) For ET: Ox(aq) + e−(β) = Red(aq), where e−(β) = e−(vac), e−(SHE), e−(aq)ΔETG = μ,aq(Red) − μ,aq(Ox) − μβ(e)

(iii) For PT–ET: Ox(aq) + H+(aq) e−(β) = Red(aq), where e−(β) = e−(vac), e−(SHE), e−(aq)ΔPT–ETG = μ,aq(Red) − μ,aq(Ox) − μ,aq(H+) − μβ(e)In the case of ET and PT–ET reactions, the constant μβ(e) defines the Gibbs energy level of a conventional reference state for electrons.

Electrons in a vacuum at infinity e−(vac) form a reference state with μβ(e) = μvac0K(e) = 0.

It is thereby possible to define a reduced absolute electrode potential, Eabs,r (Ox/Red).23,36,37

Electrons in a single reference electrode, e.g., the standard hydrogen electrode SHE with e−(SHE), form a reference state with μβ(e) = μ,SHE(e), such that the standard electrode potential can be defined in the hydrogen scale, ESHE(Ox/Red), for cells without liquid junctions.23

Using Fermi–Dirac and Boltzmann statistics, experimental solvation data, and the absolute potential of the SHE, we have estimated internally consistent values for μ,SHE(e) and μ,aq(H+), which are listed in Table .123

Moreover, in order for the Gibbs energies of reaction to agree with the bond dissociation energy scale, we correct μ,aq(H+) with the difference in ZPE’s of the reduced and oxidized forms for PT–ET or, in the case of PT reactions, of the conjugate base and the acid.23

The Gibbs energy diagrams given below contain the standard Gibbs energy of each species relative to that of the purine nucleobase and the hydroxyl radical.

In order for the Gibbs energetics to reflect that the number of particles is conserved along the reaction profile, μ,SHE(e) and μ,aq(H+) are added cumulatively to the relative standard Gibbs energy of each species.

The standard Gibbs energy change associated with a given step is calculated as the difference between the listed values for the species in question.

Results and discussion

Tautomerism in the 8-oxopurine derivatives

The neutral 8-oxo-7,8-dehydropurine derivatives, 8-oxoB (B = Pu, A, G), as well as their radical cations and anions can occur in the oxidation pathways.

Albeit the derivatives have been reported to occur primarily in their keto forms, the oxidation seems to go by way of the minor enol forms in aqueous solution.

The standard Gibbs energy profiles of ketonization of 8-hydroxypurine (8-OHPu) and the radical ions (Figs. 2 and 3) show that the catalytic effect of water significantly lowers the activation Gibbs energy to less than 10 kcal mol−1 in aqueous solution.

The activation Gibbs energy for the water-catalyzed ketonization of neutral 8-OHPu is 1.8 kcal mol−1 higher than that of the radical anion and 2.2 kcal mol−1 lower than that of the radical cation.

It should be noted that the explicit inclusion of a single water molecule in the model allows us to retrieve the main catalytic contribution of water to the activation Gibbs energy.

It is well known that water molecules from the first and second solvation sheaths reorganize so that their dipole moment distribution assists the proton transfer by electrostatic stabilization of the transition state.

Hence, a further lowering of the activation energy can be expected.

A priori, the number of water molecules in the first, second, and third solvation shells needed to reproduce the experimental Gibbs energetics is uncertain if we use a cluster model.

The reorganization of water dipoles in the solvation shells brings about both electrostatic and entropic stabilization.

The assessment of enthalpies, entropies, and Gibbs energies can be very inaccurate when using a cluster model, owing to problems in convergence of these quantities with increasing cluster size.

However, the average electrostatic contribution can be assessed with a continuum solvent model approach, even though it cannot reproduce the local microscopic rearrangements of water molecules, which create the local electrostatic inhomogeneity that is averaged out in the macroscopic dielectric constant of the solvent.

In short, we take into account two main solvent effects, i.e., the catalytic role of a single water molecule in the tautomerization mechanism and the average electrostatic stabilization described by a self-consistent reaction field model.

We thus neglect the local electrostatic inhomogeneity created by the water dipoles around the tautomerization center.

The structures of the neutral 8-hydroxy- and 8-oxopurine derivatives and their radical cations are planar at the B3LYP/6-31G(d,p) level as well as 8-oxoPu˙ at the B3LYP/6-31+G(d,p) level (Fig. 3).

However, 8-OHPu˙ is puckered with the OH group in an equatorial position.

A difference of only 0.6 kcal mol−1 is observed in the standard Gibbs energies of ketonization of 8-OHPu when including a water molecule or not (Fig. 2).

In the case of 8-OHPu˙, however, the inclusion of a water molecule in the model lowers the standard Gibbs energy of ketonization by 4 kcal mol−1 because the complex reduces the ring puckering and the destabilizing strain seems larger than the explicit hydrogen-bonding stabilization in this radical anion.

In the case of 8-OHPu˙+, the standard Gibbs energy of ketonization is about 10 kcal mol−1 less exergonic with the inclusion of the explicit water molecule, owing to an unfavourable hydrogen-bonding pattern in the complex with the radical cation.

The water molecule forms only one strong hydrogen bond with either radical cation tautomer as the bond distances shown in Fig. 3

PT-initiated oxidation pathway in the 8-hydroxypurine radicals

The 8-deprotonation of the C8 radical adducts B8OH˙ (Fig. 1, left branch) is endergonic by more than 25 kcal mol−1 in aqueous solution (Fig. 4).

This standard Gibbs energy of deprotonation suggests that a very strong base is needed to remove the proton from the 8-position in aqueous solution.

The 8-deprotonation of the C8 radical adducts is therefore very unlikely in neutral or acidic medium, and in turn, the PT-initiated oxidation seem thermodynamically unfavourable.

Moreover, the standard Gibbs energies of deprotonation increase in the order Pu8OH˙ < A8OH˙ < G8OH˙.

In particular, the 8-deprotonation of G8OH˙ is far more disfavoured than that of A8OH˙.

The radical anions, 8-OHB˙, may follow two pathways to give the major 8-oxo-7,8-dehydropurine tautomers, 8-oxoB (Fig. 4).

Since the difference in the barriers of ketonization in the radical anions (5.2 kcal mol−1) and in the neutral enol forms (7.0 kcal mol−1) is small, the thermodynamic effect owing to the electron loss could determine the preference for a particular pathway.

Inspection of the standard Gibbs energies of one-electron oxidation of the enol and keto forms indicates that, whereas the one-electron oxidation of 8-OHPu˙ and 8-OHG˙ (−55.6 and −72.5 kcal mol−1, respectively) is favoured over that of 8-oxoPu˙ and 8-oxoG˙ (−47.3 and −64.6 kcal mol−1, respectively) by about 8 kcal mol−1, the one-electron oxidation of 8-oxoA˙ (−65.1 kcal mol−1) is 3 kcal mol−1 more favoured than that of 8-OHA˙ (−62.1 kcal mol−1).

As a result, it is likely that for G, one-electron oxidation of 8-OHG˙ is followed by subsequent tautomerization to yield 8-oxoG; whereas for A, one-electron oxidation and tautomerization could compete.

Clearly, though, the oxidation reactions are thermodynamically favoured for all systems discussed above.

The calculated standard Gibbs energies of reduction of the 8-hydroxypurine derivatives 8-OHB and their major 8-oxo-7,8-dehydropurine tautomers 8-oxoB (more than ca. 45 kcal mol−1 endergonic) suggest that the respective anions are powerful reductants.

PT–ET-initiated oxidation pathway in the 8-hydroxypurine radicals

The 8-proton-coupled one-electron oxidation of the C8 radical adducts B8OH˙ (Fig. 1, right branch) is exergonic by about 25 kcal mol−1 in aqueous solution–Fig. 5, ΔG (8-OHB) − ΔG(B8OH˙).

Accordingly, a moderate to strong oxidant may be able to yield the 8-hydroxypurine derivatives, 8-OHB.

The standard Gibbs energy of PT–ET becomes less negative in the order Pu8OH˙ > A8OH˙ > G8OH˙ (−27.9, −26.8, and −25.1 kcal mol−1, respectively).

An alternative stepwise oxidation of the C8 radical adducts B8OH˙ will be driven by a strongly exergonic deprotonation.

The one-electron oxidation of the C8 radical adduct to yield the cation B8OH+ is endergonic and the ease of oxidation increases in the series Pu8OH˙ < A8OH˙ < G8OH˙, according to the standard Gibbs energies of one-electron oxidation of the C8 radical adducts (29.2, 11.2, and 0.2 kcal mol−1, respectively).

The subsequent deprotonation of the C8 cation adducts B8OH+ is very favoured but decreases in the order Pu8OH+ > A8OH+ > G8OH+ when comparing the standard Gibbs energies of deprotonation (−57.1, −38.0, and −25.3 kcal mol−1, respectively).

In short, the increasing reducing character of the C8 radical adducts B8OH˙ offsets the decreasing deprotonation tendency of the C8 cation adducts B8OH+ to give 8-OHB.

The explicit route (reduction followed by deprotonation vs. proton coupled one-electron oxidation of B8OH˙) will depend on the physiological conditions such as pH.

Once 8-OHB is formed, the small barrier to ketonization (7.0 kcal mol−1, Fig. 2) and the standard Gibbs energies of ketonization (−11.2, −10.0, and −12.3 kcal mol−1 for Pu, A, and G, respectively) readily enable a shift of the equilibrium toward the major 8-oxoB product.

Oxidation of the 8-oxopurine derivatives

Unlike the radical anions 8-OHB˙ and 8-oxoB˙ (Fig. 4), the neutral keto form 8-oxoB is as easily oxidized as the neutral enol form 8-OHB (Fig. 5).

In both cases, the standard Gibbs energies of oxidation for the enol and keto forms are very close: 46.0 and 46.2 kcal mol−1 for Pu, 31.2 and 30.4 kcal mol−1 for A, and 20.1 and 21.5 kcal mol−1 for G, respectively.

For both pathways oxidation is hence endergonic and the standard Gibbs energy changes decrease in the order Pu > A > G. Of particular interest is that 8-oxoG is about 10 kcal mol−1 more easily oxidized than 8-oxoA.

Consequently, the Gibbs energies indicate that the major 8-oxoB product can be converted into 8-oxoB˙+ by reaction with a suitable oxidant.

The energetics also suggest that in the presence of the appropriate oxidant, the enol form 8-OHB could either be oxidized once formed or go through the keto form 8-oxoB (activation barrier 7 kcal mol−1; Fig. 2), if required by the oxidation mechanism.

However, in absence of appropriate oxidants, the equilibrium would shift toward the major keto form 8-oxoB.

The calculated standard Gibbs energies of oxidation of the 8-hydroxypurine derivatives 8-OHB and their major 8-oxo-7,8-dehydropurine tautomers 8-oxoB (more than ca. 20 kcal mol−1 endergonic) indicate that the respective radical cations are good oxidants.

Their experimental counterparts amount to 21.7 and 13.8 kcal mol−1 for adenine and guanine derivatives at pH 8, respectively.16

It should be remarked that the calculated standard Gibbs energy of reduction for the 8-oxoG˙+/8-oxoG couple (−21.4 kcal mol−1) is slightly less exergonic than that of the parent nucleobase couple G˙+/G (−25.8 kcal mol−1).24

The pH-dependent absorption spectra of the one-electron oxidized 8-hydroxy-2’deoxyguanosine (8-OHdG) yields a pKa of 6.6 for 8-oxodG˙+.15

Our calculations show that the 8-oxopurine-derived radical cations 8-oxoB˙+ are acidic.

The ease of 7-deprotonation of 8-oxoB˙+ into 8-oxoB(-H7)˙ is 8-oxoPu˙+ > 8-oxoA˙+ > 8-oxoG˙+, since the standard Gibbs energies of deprotonation are −15.9, −10.8, and −5.1 kcal mol−1, respectively.

Accordingly, 8-oxoG˙+ is a more stable radical cation than 8-oxoA˙+, inasmuch as the latter shows a greater tendency to yield 8-oxoA(–H7)˙.

The same experimental data15 for 8-OHdG suggest that the 1-deprotonation of 8-oxoG˙+ to yield 8-oxoG(–H1)˙ is as likely to occur as the 7-deprotonation to yield 8-oxoG(–H7)˙ (Fig. 6).

The computed standard Gibbs energy change favours 8-oxoG(–H7)˙ over 8-oxoG(–H1)˙ by only 3.6 kcal mol−1.

Subsequent deprotonation of 8-oxoG(–H7)˙ or 8-oxoG(–H1)˙ corresponds to a experimental standard Gibbs energy change of 16.8 kcal mol−115 and yields the radical anion 8-oxoG(–H1, –H7)˙.

Our calculated standard Gibbs energies of deprotonation of 8-oxoG(–H7)˙ and 8-oxoG(–H1)˙ are halfway and amount 8.3 and 4.7 kcal mol−1, respectively.

Both the proton-coupled one-electron oxidation of the 8-oxo-7, 8-dehydropurine tautomers (8-oxoA and 8-oxoG) and the deprotonation of their radical cations (8-oxoA˙+ and 8-oxoG˙+) yield dehydrogenated nucleobase radicals (8-oxoA(–H7)˙, 8-oxoG(–H7)˙, 8-oxoG(–H1)˙) in neutral aqueous solution.

The calculated standard Gibbs energy of reduction of 8-oxoG(–H7)˙ to 8-oxoG in aqueous solution (–16.4 kcal mol−1) agrees very well with the experimental one (−17.0 kcal mol−1).

Hence, also in the case of the dehydrogenated radicals (B˙+ = B(−H)˙ + H+) the calculated standard Gibbs energy of reduction of the 8-oxoG(–H7)˙/8-oxoG couple is lower than that of the G(–H1)˙/G couple (−25.6 kcal mol−1).24

The calculated standard Gibbs energy of reduction of the 8-oxoA(–H7)˙/8-oxoA couple (−19.6 kcal mol−1) is greater than that of the 8-oxoG(–H7)˙/8-oxoG couple but 8-oxoA could still reduce G(–H1)˙.

As also noted for the nucleobase radical cations (A˙+, G˙+, T˙+, C˙+),24 the counterparts of the 8-oxo-7, 8-dehydropurine tautomers, 8-oxoB˙+, seem to be strong acids, according to the standard Gibbs energy of deprotonation for the equilibria 8-oxoB˙+(aq) = 8-oxoB(−H)˙(aq) + H+(aq).

In particular, 8-oxoA˙+PTG = −10.7 kcal mol−1) is more acidic than 8-oxoG˙+PTG = −5.1 kcal mol−1, compared to the experimental value15 of 9.0 kcal mol−1) when yielding 8-oxoB(–H7)˙, but the H7 proton of 8-oxoG˙+ is more acidic than the H1 proton (ΔPTG = −5.1 vs. −1.5 kcal mol−1).

Further deprotonation (Fig. 6) is thus more likely to occur on 8-oxoG(–H1)˙ than on 8-oxoG(–H7)˙, due to the larger ΔPTG when forming 8-oxoG(–H1, –H7)˙.

It should be noted that the IEF-PCM/B3LYP/6-311+G(2df,p)//B3LYP/6-31G(d,p) standard Gibbs energies of deprotonation are more exergonic than the few experimental values available.

The main source of error in the calculations comes from the insufficient solvation energies obtained by IEF-PCM on these radicals, and further studies on this aspect are thus needed.

Nonetheless, the qualitative results provide the correct trends.


The 8-oxo-7,8-dehydropurine tautomers are highly mutagenic lesions in DNA.

We have in the present work applied a first principles electrochemical approach to explore mechanisms for their formation and explain the observed preference for 8-oxoG over 8-oxoA.

The PT-initiated oxidation pathway of the 8-hydroxypurine radicals (A8OH˙ and G8OH˙) followed by electron loss12 seems to be very unlikely in aqueous solution because of the very low acidity of the hydrogen at the 8-position.

Instead, the PT–ET-initiated oxidation pathway4,13,14 of A8OH˙ and G8OH˙ seems to be the favoured route in aqueous solution.

In the case of the PT–ET-initiated oxidation, G8OH˙ is only slightly more sensitive than A8OH˙ to proton-coupled one-electron oxidation followed by tautomerization to yield 8-oxoG and 8-oxoA.

Likewise, 8-oxoG and 8-oxoA behave similarly toward one-electron oxidation to yield the radical cations 8-oxoG˙+ and 8-oxoA˙+.

The difference in reactivity does not appear to be related to the 8-hydroxylation step either, since the barriers of adenine and guanine are close to the diffusion-controlled limit and both reactions are exergonic by about 15 kcal mol−1.24

However, the proton-coupled one-electron oxidation of 8-oxoG is preferred over that of 8-oxoA by about 5 kcal mol−1, and hence, a higher yield of 8-oxoG(–H7)˙ should be expected over 8-oxoA(–H7)˙ in DNA.

Our calculations furthermore show that the ease of ionization of nucleobases in their native24 and oxidized forms to yield B˙+ (via ET) or B(–H)˙ (via PT–ET) is 8-oxoG > G > 8-oxoA > A > C > T in aqueous solution.

Accordingly, 8-oxoG can reduce any nucleobase radical cation B˙+ (via ET) or their deprotonation products B(–H)˙ (via PT–ET) to the nucleobase native forms.

The smallest driving force thus corresponds to the repair of lesions in G (G˙+ or G(–H1)˙):


G˙+(aq) + 8-oxoG(aq) = G(aq) + 8-oxoG˙+(aq)

ΔrG0298K = −4.4 kcal mol−1


G(−H1)˙(aq) + 8-oxoG(aq) = G(aq) + 8-oxoG(−H7)˙(aq)

ΔrG0298K = −9.3 kcal mol−1 (exp1: −12.7 kcal mol−1)

Since the driving force for the PT–ET repair is about twice as strong as that for the ET repair, the presence of 8-oxoG could also contribute to shift the equilibrium at neutral pH toward the deprotonation of the remaining G˙+, which behaves as a weak acid.

Likewise, 8-oxoG can reduce both 8-oxoA˙+ (via ET) and 8-oxoA(–H7)˙ (via PT–ET).

Hence, albeit the conditions for damages to occur on A and G are essentially identical, 8-oxoG is more prone to reduce/repair damages on other sites (including G˙+ and 8-oxoA) and to function as a sink for oxidative damage occurring in the base stack.

Thus, the detection of 8-oxo-7, 8-dehydropurine derived radicals would yield higher amounts of 8-oxoG than 8-oxoA.

This provides an explanation as to why the observed ratio 8-oxoA:8-oxoG is 1:3 when both naked DNA and chromatin are irradiated, and could be a likely way in which 8-oxoG acts as a trap of radical cations and neutral radicals in DNA.

We emphasize, however, that experimental and computational studies on the addition of water to G˙+ yielding 8-oxoG 4 may provide additional insights into this problem.