Decomposition of Thymidine by Low‐Energy Electrons: Implications for the Molecular Mechanisms of Single‐Strand Breaks in DNA

Abstract

The damage to biological systems by high-energy quanta occurs mainly through secondary species created in the ionization path. Interaction of the low-energy (0–3 eV) electrons with thymidine (Td) in the gas phase was investigated. These electrons induce loss of hydrogen from the N3 position and scission of the N1C1 bond. The transfer of excess electrons, initially localized on thymine, to the sugar moiety was not observed. Low-energy electrons decompose gas-phase thymidine (a thymine unit coupled to a sugar unit) through dissociative electron attachment (DEA) involving two low-lying resonances at 1.2 and 1.8 eV and a broad resonant feature located between 5.5 and 12 eV. The peak at 1.2 eV arises from the rupture of the glycosidic N1C1 bond (observable through the anion of the sugar moiety), whereas the resonance located at 1.8 eV results from the loss of a neutral hydrogen atom from the N3 position of the thymine moiety within thymidine. The broad resonant feature in the energy range 5.5–12 eV is associated with the rupture of the N1C1 bond, in this case observable through the thymine anion generated. From the analysis of the DEA signatures it follows that excess-electron transfer from the thymine to the sugar unit within thymidine and vice versa is not operative. These findings have significant consequences for the molecular description of DNA damage caused by low-energy electrons. It excludes the possibility of electron migration from low-lying π* MOs of a nucleobase to low-lying σ* MOs of the DNA backbone leading to strand breaks. The investigation of damage in biological systems, more specifically, in living cells and DNA, induced by high-energy quanta has been the subject of countless investigations over the past decades.1–3 Such alterations can appear, for example, when human beings are exposed to radioactivity or other sources of high-energy radiation. On the other hand, analogous questions arise in the application of radiosensitizers in tumor therapy. It is well accepted that the main biological effect is usually not produced by the primary interaction of the high-energy quanta with the complex molecular network in a living cell, but rather by the action of the secondary species which are generated along the ionization track.2 The interaction of these secondary species (ions, electron, or radicals) with DNA and its surroundings can cause mutagenic, genotoxic, and other potentially lethal DNA lesions,3 such as single-strand breaks (SSBs) and double-strand breaks (DSBs). Electrons are the most abundant of these secondary species having an initial energy distribution extending to about 20 eV.4 For the understanding of the effects of radiation in cells, it is therefore essential to investigate the action induced by these electrons on vital cellular components such as water and DNA. So far the effects from high-energy radiation has been investigated for two systems that vary greatly in complexity, namely plasmid DNA and building blocks of DNA in the gas-phase. Experiments on plasmid DNA demonstrated that electrons at subionization energies induce SSBs and DSBs.5 Recent experiments showed the potential of electrons at even subexcitation energy (0–4 eV) to induce SSBs in plasmid DNA.1 Although SSBs within double-stranded DNA is not likely to be lethal for the cell, the situation may dramatically change during the process of replication of the nucleic acid, that is, when the double strand separates into two individual strands. NB−# is the transitory negative ion (TNI) generated by resonant electron capture and (NB−H)− is the closed-shell anion formed by ejection of a neutral hydrogen radical from NB−#. The reaction is energetically driven by the appreciable electron affinity of the (NB−H) radicals that lie in the range between 3 and 4 eV.6, 10 Further experiments with partly deuterated thymine9, 11 demonstrated that hydrogen abstraction is operative exclusively from the N sites. Moreover, by exactly tuning the electron energy, the loss of H can even be made site selective: although electrons at 1 eV induce H loss at the N1 position (N1H), the process can be switched, at 1.8 eV, to N3-H.12 These last findings were obtained by using thymine and uracil, methylated at one of the N positions. The observation that neither the loss of H from CH3, nor the loss of the entire methyl group takes place when at the corresponding N position H is replaced by CH3. These findings have significant consequences for the molecular description of strand breaks by low-energy electrons. A recent theoretical study13 (modeling a section of DNA that contained cytosine, the sugar ring, and the (neutralized) phosphate group) predicted a transfer of the electron from the initial π* anion state of the base to a σ* state in the backbone, ultimately leading to rupture of the CO bond between the phosphate and the sugar. Within DNA the N1 position of T is coupled to the sugar moiety, and according to the results obtained in the methylated compounds, transfer of energy via the N1 position is inhibited once N1H is replaced by N1CH3. Although the inhibition of energy transfer should also apply if CH3 is replaced by the sugar, one cannot necessarily exclude the transfer of excess charge, initially localized on thymine, through the N1C1 glycosidic bond to the sugar molecule, particularly when low-lying σ* MOs are accessible. To track this problem, we have studied herein low-energy electron impact to gas-phase thymidine (Td), which represents a thymine coupled to the sugar 2-deoxyribose (DRB)1 through a condensation reaction. For simplicity, the deoxyribose moiety in Td (DRB−OH) will be assigned as dR throughout this manuscript. The present investigations were performed in a crossed electron/molecule beams device at the Innsbruck laboratory with a method previously described in detail.6, 14 The electron beam is formed in a custom-built hemispherical electron monochromator, operated at an energy resolution between 110 and 130 meV (full width at half maximum) and an electron current of 5–8 nA. The molecular beam emanates from a source that consists of a temperature-regulated oven and a capillary. Experiments have been performed in the range of 390–426 K to obtain information on the thermal-decomposition behavior of Td. This was an inherent problem in the herein described experiments and as such is discussed below. Negative ions formed in the collision zone of the crossed beams are extracted by a weak electric field towards the entrance of the quadrupole mass spectrometer. The mass-selected negative ions are detected by a channeltron by using a single-pulse counting technique. The intensity of a particular mass-selected negative ion is then recorded as a function of the electron energy. The electron-energy scale was calibrated by using the known SF6− signal near 0 eV. Absolute calibration of the DEA cross sections were established by using the established cross section of Cl−/CCl4,15 which yielded measures for the cross sections at an estimated accuracy within one order of magnitude. Thymidine was purchased from Sigma–Aldrich at a stated purity of 99.5 %. Yields of product ions observed from electron attachment to thymidine (Td, 242 amu) obtained at a temperature of 398 K. T represents thymine and dR the 2-deoxyribose molecule after loss of OH. The numbers on the intensity scale (I) correspond to the estimated absolute DEA cross sections in units of 10−22 m2. The dotted portions of the ion-yield curves are identified to arise from thermally decomposed Td (see the text). The signal at 125 amu can be assigned to (T−H)− and that at 115 amu to (dR−2 H)−. Under the assumption that the entire signal arises from reactions of intact Td targets, the signals can then be attributed to the scission of the N1C1 bond with the excess electrons localized on the thymine and sugar unit. The sugar unit is then further subjected to the loss of two hydrogen atoms. These results are in agreement with a previous study performed at the Berlin Laboratory16 in which questions were raised with respect to a possible thermal decomposition of Td. A detailed study of that problem revealed that the vast majority of the beam in fact consists of nondecomposed Td molecules. As shown in Figure 1, the dashed portions in the ion yields arise from electron attachment to products of thermally decomposed Td. Irrespective of the presence of thermal decomposition, the (Td−H)− signal (241 amu) must arise from intact Td molecules. The site from which the hydrogen atom is abstracted can immediately be identified through comparison with the signal arising from hydrogen loss from thymine methylated at the N1 position (m1T; Figure 2). In m1T, hydrogen loss from the N1 position is blocked,12 and the ion signal (M−H)− (M corresponds to the target molecule) is consequently owing to hydrogen loss from the N3 position. As the shapes of the two signals in Figure 2 nearly coincide, we can conclude that Td hydrogen abstraction predominantly occurs at the thymine moiety, more precisely, at the N3 position. The small contributions near 1 and 0 eV may be due to H loss from the sugar moiety. Comparison of the yield of anion (M−H)− resulting from hydrogen loss in thymidine (M=Td) with that of thymine methylated at the N1 position (M=m1T). Figure 3 shows the 125-amu signal and the 115-amu signal on an extended energy scale. Although the shape of the low-energy 125-amu signal (1) is very similar to that from isolated T, it differs completely at higher energies (2) in that an additional pronounced contribution in the range 5.5 to 10 eV is observed only from Td. Superficially, from this additional (T−H)− signal, one could conclude that the beam consists only of nondecomposed Td molecules. The additional signal would then also be evidence of the different probabilities to distribute excess energy, that is, in isolated T, the excess energy must be shared between H and (T−H)−, as the negative ion is unstable with respect to dissociation and autodetachment above 3–4 eV. This therefore makes DEA unlikely. In contrast, in Td the excess energy can be shared between polyatomic fragments of approximately equal size with a correspondingly large number of degrees of freedom resulting in the formation of (T−H)− at energies above 5 eV. Comparison of negative-ion formation from Td and the isolated building blocks T6 and 2-deoxyribose (DRB) 17 on an extended energy scale. A careful analysis of the 125-amu signal at different sublimation temperatures, however, revealed that the intensity ratio between features 1 and 2 changes with increasing temperature in favor of the low-energy contribution 1, see Figure 4. This directly indicates that the low-energy (T−H)− signal arises from thermally decomposed Td (in part or completely). Moreover, the shape of the 115-amu fragment also shows a variation with temperature (not shown here) in that the peak near 0 eV is relatively enhanced until, at 425 K, the shape of the ion yield becomes very close to that from isolated DRB. This also indicates thermal decomposition leading to the 115-amu signal located close to 0 eV (in part or completely). The yield of (T−H)− from Td at an elevated temperature of 426 K. The likely thermal-decomposition products are (T−H) and dR (rupture of the glycosidic N1C1 bond). Electron attachment to neutral (T−H), however, would never lead to a (T−H)− signal like that shown in Figures 1 and 3.18 A likely scenario is that the thymine radical (T−H), formed by thermal decomposition, is transformed into T by hydrogen pickup in the course of collisions with the walls of the capillary. Accordingly, the complement dR (117 amu, C5H9O3) is transformed into a closed-shell compound of the stoichiometric composition C5H10O3 and subjected to DEA. The alternate (and probably more likely) scenario is that thymidine decomposes into T + C5H8O3 (116 amu); that is, the cleavage of the N1C1 glycosidic bond is accompanied by hydrogen transfer generating two closed-shell fragments which are subjected to DEA. Irrespective of the thermal-decomposition products, our data indicate that near 400 K (the temperature at which the spectra of Figure 1 and Figure 3 have been recorded), the vast majority of the molecular beam consists of intact Td molecules. This can be concluded directly from the intensity ratio between feature 1 and feature 2 in the (Td−H)− signal in Figure 3. As the cross section for DEA has, normally, a general reciprocal dependence with energy,18 the density of the target creating feature 2 (Td) must be appreciably larger than that creating feature 1. This also follows from the absolute intensity of the (T−H)− signal, which, from T is more than two orders of magnitude larger than from Td (at a comparable gas pressure). It is in fact likely that the entire low-energy feature of the (T−H)− signal originates from T (present as a thermal-decomposition product) and not from intact Td. This directly follows from the shape of the ion yield, (M−H)−, owing to the loss of hydrogen from T that is methylated at the N1 position (m1T) (see Figure 2). Methylation at N1 or coupling to the sugar should have the same effect on T with respect to hydrogen loss. We therefore conclude that the low-energy (T−H)− signal arises from thermal decomposition. Similar arguments apply for the 115-amu signal near 1.2 eV. This leads to the conclusion that the structure near 0 eV is a result of decomposition, whereas the majority of the signal intensity at 1.2 eV can be attributed to the DEA reaction of intact Td. In Figure 1 the dotted part of the ion-yield curve corresponds to contribution from thermal-decomposition products. To summarize the present findings: At energies above 5.5 eV, electron attachment to thymidine occurs through localization of the excess charge on the thymine part. This is supported by the observation of a similar feature in isolated T (yielding a variety of DEA fragments and characterized as a core excited resonance6). The excited transient anion decomposes into (T−H)− through rupture of the glycosidic N1C1 bond. The neutral counterpart is C5H9O3 (dR), which may further decompose (by loss of H2, H2O, etc.) due to the presence of appreciable excess energy. At subexcitation energies (<3 eV), Td captures electrons through resonances located at 1.8 eV and at 1.2 eV. The 1.8-eV feature ((Td−H)− at 241 amu) can be assigned as a shape resonance with π* character, with the excess charge localized on the thymine moiety leading the loss of H from the N3 position. The 1.2-eV resonance ((dR−2 H)− at 115 amu) is linked to electron localization on the sugar moiety (shape resonance of σ* character) leading to the rupture of the glycosidic N1C1 bond. We have hence a situation wherein initial capture into σ* MOs at the sugar unit leads to rupture of the σ (N-C) bond, whereas capture into π* MOs of the thymine unit exclusively leads to loss of H from the N3 site. This last reaction is expected to occur through vibronic coupling, that is, a mixing of π* states with repulsive valence σ* states through vibrational motion.19 Only core excited states of the thymine unit induce cleavage of the glycosidic NC bond. We shall now briefly consider the implications of these findings for the molecular mechanism of DNA damage by low-energy electrons. The fact that the 115-amu signal does not carry any signature of electron attachment to the thymine moiety and, accordingly, neither the (Td−H)− signal (loss of hydrogen from N3) nor the (T−H)− signal carries any signature of electron attachment to the sugar moiety. It therefore follows that electron transfer between the two units (associated with fragmentation) is inhibited. Rupture of the N1C1 bond in fact takes place, but only through electron localization on either of the two units. The present findings directly demonstrate that migration of the excess charge from the π* anions of the nucleobases to the DNA backbone is inhibited and may hence not contribute to SSBs as previously proposed.13 Instead, attachment of low-energy electrons can lead to loss of H from the N3 position (1.8 eV) and to the rupture of the N1C1 bond at either 1.2 eV or in the energy range between 5.5 eV and 10 eV. All three processes can be considered as primary events contributing to strand breaks. The first creates a highly mobile and reactive H radical (with its potential to induce further damage in DNA), whereas the latter two induce rupture of the glycosidic bond that can lead to excision of thymine from DNA. It has to be noted that the presence of the phosphate group could in fact modify the charge-transfer behavior with respect to that presently observed in Td. So far there is little information on the role of the phosphate group. In a very recent study20 on self-assembled monolayers of DNA with orientations perpendicular and parallel to the surface, the desorption of OH− following electron bombardment was studied. From the analysis of the data the authors concluded that in the energy range 2–5 eV direct DEA reactions to the phospate unit take place, whereas at higher energy reactive scattering may also contribute to the OH− desorption signal.