Electron-induced damage of DNA and its components: Experiments and theoretical models

Abstract

It is now over ten years since the seminal experiments of Leon Sanche’s group in Sherbrooke have compellingly shown that subexcitation electrons interacting with DNA could cause the occurrence of specific resonant processes which in turn would eventually lead to either single or double strand breaks in DNA materials, to the damaging of its molecular components and possibly to biological apoptosis.

Since then a great deal of activity has been spurred by that initial work, with experiments and computations being carried out in several laboratories around the world. Hence, several components of the DNA molecular structure and make-up, i.e. from the purinic and pyrimidinic bases to the sugar and phosphate fragments, have been analysed in detail in the gas phase, on thin-film deposits on noble metals, and in some form of condensed phase, in interaction with low energy electrons. Likewise, several theoretical and computational approaches have been directed at the study of the molecular processes deemed to be crucially involved in the various steps of the energy deposition by the impinging electron onto the molecular networks.

The aim of the present review is therefore to put together, after these ten years of intense activity, the major findings which have been consolidated from the broad variety of existing experiments and, at the same time, the main computational approaches which describe the extent of molecular damage following the initial electron attachment process. The present field, in fact, is becoming mature enough to profitably stand an overall evaluation of its experimental and theoretical/computational results and to further construct, from such a review, a starting point for the assessment of its future directions.

After a detailed analysis of the experimental data, in the gas phase and in other phases, we shall therefore report the main computational tools and theoretical concepts employed today for the interpretation of the measurements at the molecular level. An overall analysis of the subject will be attempted in the last Section of this review.

Introduction

It has been known since the discovery of X-rays and radioactivity more than a hundred years ago that the exposure of living beings to high energy radiation (particles or photons) may result in fatal effects. This problem became directly obvious after the dropping of nuclear bombs at Hiroshima and Nagasaki in 1945 or the accidents at the nuclear power plants at Tschernobyl (April 1986) and at Fukushima (March 2011). During the earlier of these disastrous events the local dosage of radiation received by people was at an extremely high level so that many exposed humans, also without immediate direct physical injuries, were killed within hours or days due to the effects of radiation. Many of the survivors still suffer today, but even for individuals who survived without obvious problems, their descendants may seriously be affected as a result of genetic alterations.

Apart from these fatal and extreme cases, everybody is exposed to some constant dosage of high energy radiation due to the presence of the so-called natural radioactivity but also in the case of specific medical treatments. A constant background of low intensity but high energy radiation is ubiquitous all over the Earth and it is essentially due to cosmic radiation and to the presence of radioactive Radon (²²²Rn) in the air, itself a product of the decay of radioactive uranium isotopes present on earth. On the energy scale on a much lower level is the problem of the increasing exposure of humans to ultraviolet background (UVB) radiation due to the decrease of the ozone concentration in the stratosphere. The energy quanta of UVB radiation (≈3.9–4.4 eV) are below the level of ionisation but also capable of inducing molecular changes when interacting with biomolecular systems. In the medical field radiation is applied in diagnostics via screening techniques, but also in cancer treatment (radiotherapy, proton therapy, etc.). The problem in radiotherapy is to expose only the cancerogenic material but to keep the other areas non-irradiated. One way of proceeding is to use radiosensitisers that accumulate in cancer cells with the effect that the sensitised cells will already decay at dosages which will leave the surrounding healthy material hopefully unaffected. It is thus the understanding of the molecular processes following radiation impact which is the necessary prerequisite for both a more effective and a more controlled radiotherapy strategy, a feature also important for the implementation of rational dose limits to human beings in terms of exposure time, intensity and energy of the impinging radiation.

The term radiation damage in this review shall chiefly deal with damage of biological material on a short time scale, i.e. reactions and processes occurring within ns or less following the interaction of a high energy quantum with a living cell. These initial events may ultimately lead to the collapse of such cells, resulting in death of the individual within hours or days but also to effects appearing on a much longer time scale. Alterations in the genetic expression of, e.g., DNA may result in the onset of diseases in the concerned individual and also in his descendants.

These short or ultrashort reactions and processes on which we shall focus here can be considered as the initial, crucial and decisive steps for the eventual radiation damage. It has been commonly accepted that the high energy tracks formed by $\alpha$, $\beta$ and $\gamma$ radiation can in the first step ionise cell components along the track, thus leading to various dissociation/ionisation channels which are indeed responsible for the production of damaging radicals. Such radicals can in turn act chemically and physically on the cell material, thereby becoming the primary cause of cell apoptosis processes. At the same time, it is also accepted by now that the primary impact of the energetic components along the tracks produces a great deal of secondary electrons emitted from several different sources present in the biological environment. The theoretical analysis and the microscopic mechanisms that we shall discuss in detail within the present review will therefore deal almost exclusively with the consequences of having the secondary electrons produced by the above primary processes. In other words, we shall endeavour to present the rather compelling connection existing between the known appearance of a large number of secondary electrons and the cellular apoptosis observed as the end product from the primary impact.

Before considering in more detail the action of high energy radiation to a living cell, we first recall some basic facts on the composition of DNA representing the most important and vital component of the cell nuclei. DNA is a biopolymer consisting of two chains (strands) containing the four heterocyclic bases thymine (T), adenine (A), cytosine (C) and guanine (G), each of them bound to the DNA backbone which is itself composed of phosphate and sugar units. Both strands are connected through reciprocal hydrogen bonding between base pairs in opposite positions in the two strands (Fig. 1). The structure is such that adenine pairs with thymine (A–T) and guanine with cytosine (G–C) resulting in the well-recognised right handed double helix form. The DNA itself is surrounded by proteins and embedded in water, which is by far the most abundant component of a cell.

While the amount of energy deposited per volume of condensed biological material can be determined quite precisely [1], [2], there is so far no consistent understanding and knowledge of the subsequent sequence of processes and reactions which result in a chemical modification or damage of the biomolecular system and the response (repair and misrepair) of the system to these alterations. Among this sequence of reactions the fast and ultrafast initial processes (driven to a large extent by low energy electrons (LEEs)) are important and decisive for the subsequent evolution of the system and hence for the chemical and biological effects occurring on a much longer time scale. These initial processes are briefly outlined in Section 1.1.

We may first consider the sequence of processes and reactions which follow the impact of a high energy quantum on a living cell. Consider, for example, a beam of photons at energies near the MeV range (but below the level of electron–positron pair production at 1.02 MeV) entering a living cell. The primary photon process is either absorption or scattering (Compton scattering), the latter involving one or more components of the cell. Absorption (i.e. annihilation of the photon) removes electrons from essentially any level, from valence molecular orbitals (MOs) to core atomic orbitals (AOs). Depending on the energy of the created photoelectrons they induce further ionisation and excitation events, etc. while being eventually slowed down within the medium. Similarly, Compton scattering by the primary photons generates scattered photons at lower energy and energetic electrons, which generate in turn a sequence of electrons evolving via subsequent multiple scattering events. All these electrons originating from the primary interaction of the high energy photon with the molecular network of the cell are usually labelled as secondary electrons although they are the result of primary, secondary, tertiary, etc. interactions, including electrons from other events like Auger processes, i.e., where electrons are emitted via relaxation of the core holes. In that direction, the so-called intermolecular Coulombic decay (ICD) following core ionisation has been very recently proposed as being an effective source for generating low energy secondary electrons in water [3]. The estimated quantity arising from all these processes is 10⁴ secondary electrons per 1 MeV primary quantum. Taking a snapshot at a few femtoseconds after the primary interaction we have then (multiply) charged sites within the complex molecular network of the cell (eventually undergoing Coulomb explosion), electronically excited sites and, last but not least, an exceeding number of low energy secondary electrons at an energy distribution extending to a few tens of eV [4]. Although the ionised as well as electronically excited sites can result in the rupture of chemical bonds (when repulsive potential energy surfaces are accessed), the major effects are those which become induced by the large number of secondary electrons. In the course of several inelastic collisions within the medium they are thermalised within picoseconds before they reach some stage of solvation, chiefly becoming by then chemically inactive species.

Obviously, the sequence described above can involve any of the cell components (DNA, water, proteins), separately or in combination. Water as the most abundant component of a cell generates the reactive OH radical which in turn can attack any cell component in its surrounding. It is assumed that damage of the genome in a living cell by ionising radiation is about one third direct and two third indirect [5]. Direct damage concerns energy deposition and subsequent reactions directly in the DNA and its closely bound water molecules. Indirect damage results from energy deposition in water molecules and the other biomolecules in the vicinity of the DNA. It is believed that most of the indirect damage is due to the attack by the highly reactive hydroxyl radical OH.

The sequence of processes and reactions occurring along this track in the complex molecular network of a living cell may be divided into three major stages (i) the physical stage, that refers to processes and reactions in the time window from fs up to ps after the primary interaction. In this time window electronic excitation and ionisation, with subsequent bond ruptures along repulsive potential energy surfaces, create radicals like OH as well as abundant numbers of secondary electrons (ii) the chemical stage which includes molecular relaxation and reorganisation involving multiple bond cleavages and the subsequent formation of new molecules. Such reactions typically occur in the time window between picoseconds (10⁻¹² s) to microseconds (10⁻⁶ s) (metastable decays). (iii) The biological stage finally refers to the alterations occurring on a longer time scale, including the overall response of the system which could extend over a time range of several years.

As mentioned above, high energy quanta generate large amounts of ballistic electrons at initial energies up to some tens of eV. It appears that these ballistic electrons are responsible for triggering the crucial reactions relevant to direct damage of DNA. This became more evident after a series of recent landmark experiments performed by Sanche and coworkers [6] who irradiated plasmid DNA by a well-defined electron beam at variable energy in the range 3–20 eV. They observed the appearance of single strand breaks (SSBs) and double strand breaks (DSBs) at energies above 15 eV (as expected), the efficiency for strand breaks continuously increasing with electron energy. Surprisingly, they also observed appreciable SSBs and DSBs in the range below the ionisation threshold (<10 eV) with the efficiency showing a resonant behaviour with respect to the electron energy. In subsequent experiments the electron energy was brought down to lower energies with the result that strand breaks were also detected within structured resonances below 3.0 eV and peaking at 0.8 eV [7]. The low energy resonance, however, was associated solely with SSBs. These were unexpected findings, since it was believed till then that only electrons at energies above the ionisation threshold could contribute to DNA damage. From the resonant structure of the damage profile it was proposed that resonant electron captures could occur at particular molecular components and sites of the DNA and thus may be the initial step towards causing the observed strand breaks [6]. It is well known that at low energies, preferentially below the level of electronic excitation, electrons can be captured by molecules thereby creating transient negative ions (TNIs) which could subsequently decompose into several fragments (dissociative electron attachment, DEA) [8], [9], [10]. This means that a chemical bond can be cleaved by electrons at energies well below the corresponding bond dissociation energy and sometimes already at threshold (0 eV). Energetically, such a low energy bond-breaking process becomes accessible provided that the electron binding energy on the negatively charged fragment is sufficiently large to locally increase the available energy. By that mechanism, therefore, the secondary ballistic electrons, on their way to be slowed down (but before getting solvated) can be captured by specific molecular sites within DNA and thereby trigger the chemical reactions that can eventually lead to cell apoptosis.

The landmark experiments by Sanche and coworkers initiated much activity concerning the detailed investigation of the interaction of LEEs with DNA and its building blocks both experimentally and theoretically. It has been shown by gas phase experiments, in fact, that single DNA bases possess low energy resonances at subexcitation energies [11], [12] and that they are subject to subsequent decomposition (DEA) [12], [13], [14]. Hence, in the following we shall first review experiments dealing with the interaction of low energy electrons with the isolated building blocks of DNA in the gas phase, namely the DNA bases, the sugar unit and the phosphate unit. The question will therefore become how the intrinsic properties of the individual building blocks evolve when they are realistically coupled into the molecular network of DNA. Equally important is to understand to which degree these properties are still preserved in solution when the molecular system is embedded in a dissipative environment like that provided by structured polar solvents.

While the single building blocks can directly be transferred as intact compounds into the gas phase by thermal evaporation, larger units like thymidine (thymine bound to sugar), or entire nucleotides will instead undergo decomposition during evaporation [15]. The study on such gas phase compounds therefore will require more sophisticated evaporation techniques like the recently introduced LIAD (laser-induced acoustic desorption) [16]. A further technique that may be able to handle samples of larger complexity uses oligonucleotides (as model DNA) immobilised on a metal surface and irradiated by LEEs [17]. It has recently been proposed that ultracold He nanodroplets can act as an interesting nano-laboratory to generate aggregates of biomolecules and to study the reactions triggered by LEEs at temperatures close to 0 K [18].

When searching for a viable theoretical description it is well known that the impinging electrons, due to their light mass, couple rather inefficiently with the heavier network of the molecular nuclei and, in any case, they do so by various types of interactions which are mediated by virtual excitations of the bound electrons. Since, at the energies under consideration, the direct interaction times of impinging electrons with each molecular unit are usually of the orders of attoseconds or less, we immediately see that some other mechanism has to come into action in order to maximise the energy redistribution process during the impact of the secondary electrons. This additional mechanism is provided by the occurrence of a resonance, i.e., of the dynamical trapping of the impinging electron by a local potential barrier generated within the complicated interaction of the scattered electron with the electronuclear network of the target molecule.

Thus, we begin to see that a nanoscopic description of the sequence of quantum events which lead to the ultimate cellular damage after the impact of the secondary electrons involves several types of steps and of possible mechanisms:

• the description, as accurately as possible, of the interaction forces between the impinging electron and the target at its equilibrium geometry as is known within the relevant medium;

• the quantum formulation of the multichannel scattering problem in order to provide realistic estimates of the total cross section sizes as a function of energy from such, usually polar, species. The general shape of the cross sections at low energy will also be a preliminary indicator of possible compound resonances (see below);

• the efficient search of dynamical resonances, from single-channel or multichannel mechanisms, through the analysis of the scattering matrix properties or via an analysis of the multichannel delay matrices in the energy regions of interest [19];

• the adiabatic or non-adiabatic following of the behaviour of specific resonant complexes (the Transient Negative Ions, TNIs) in terms of energy locations and resonance widths that become modified by nuclear rearrangement processes in one or more dimensions [20];

• the mapping of the excess electron (the scattering electron) density distributions over the molecular volume, in order to relate the actual coupling of the trapped particle with specific collocations of the metastable electrons over the molecular bonding networks [21];

• the analysis of the cooperative energy redistributions after electron attachments in terms of the various potential energy surfaces associated to each of the TNIs and containing the features of the extra electron.

The complete modelling of the overall process from electron impact to damaged molecular fragments therefore requires the usage of a broad range of computational tools and their balanced employment for each of the biospecies for which experimental data are available. The role which the biological environment, i.e., the surrounding water molecules of the most important solvent medium, has on the evolution of the damaging process is still a matter of debate, although some of the recent experiments seem to indicate that the overall swiftness of the attachments and intramolecular rearrangements (IVRs) suggest, even when fairly narrow resonances are present, that the water molecules might have a “static” effect on the initial energetics but would then simply act as “spectators” during the occurrence of the initial resonant scattering and therefore may not influence the ensuing energy redistributions that guide the final apoptosis of the cell. All the above aspects will be analysed and reviewed in the following sections dealing with the theoretical and computational aspects of the damaging induced by secondary electrons.

The importance of reactions of presolvated electrons with amino acids and nucleotides has already been pointed out more than 2 decades ago by time resolved (picosecond) pulse radiolysis experiments [22]. Since then some information in the field of Radiation Chemistry has been accumulated concerning the scavenging properties of electrons in the presolvated state [23], [24], [25], [26]. While such experiments provide valuable insight into time dependence of electron scavenging processes, they do not allow direct information on the chemical reactions induced by those electrons. In Chapter 2.2.4 we shall briefly consider very recent experiments using ultrashort laser spectroscopy which, as further reported in our Conclusions, provide detailed information on reactions of prehydrated electrons with biomolecular systems in aqueous solutions.

In the following sections we shall therefore analyse in some detail both the outcome of several experimental studies and the results from quantum dynamical models describing the primary electron attachment process and the subsequent reactions.