12738
J. Phys. Chem. B 2006, 110, 12738-12748
Repair Reactions of Pyrimidine-Derived Radicals by Aliphatic Thiols Aleksandra Wo´ jcik,† Sergej Naumov,‡ Bronislaw Marciniak,§ and Ortwin Brede*,† Interdisciplinary Group of Time-ResolVed Spectroscopy, UniVersity of Leipzig, Permoserstrasse 15, 04303 Leipzig, Germany, Institute of Surface Modification, 04303 Leipzig, Germany, and Faculty of Chemistry, Adam Mickiewicz UniVersity, Grunwaldzka 6, 60-780 Poznan˜ , Poland ReceiVed: March 14, 2006; In Final Form: April 25, 2006
Pyrimidinyl radicals of various structures (Pyr•) were generated in aqueous and alcohol-containing solutions by means of pulse radiolysis to determine the rate constants of their repair reactions by different thiols (RSH ) cysteamine, 2-mercaptoethanol, cysteine, and penicillamine): Pyr• + RSH f PyrH + RS•. C5-OH and C6-OH adduct radicals of the pyrimidines react with thiols with k9 ) (1.2-10.0) × 106 dm3 mol-1 s-1. Similar repair rate constants were found for uracil- and thymine-derived N1-centered radicals, k31 ) (1.56.1) × 106 dm3 mol-1 s-1. However, pyrimidine radical anions protonated at their C6 position and C6uracilyl radicals, with carbonyl groups at their C5 position, react with thiols faster, with k24 ) (0.5-7.6) × 107 dm3 mol-1 s-1 and k14 ) (1.4-4.8) × 107 dm3 mol-1 s-1, respectively. Quantum chemical calculations, at the B3LYP/6-31G(d,p) and self-consistent reaction field polarizable continuum model level point to the combined effects of the energy gap between interacting molecular orbitals, charge distribution within different pyrimidine-derived radicals, and the coefficients of the atomic orbitals as the possible reasons for the differences in the rate constants of repair.
1. Introduction Thiols (RSH) act as antioxidants. Therefore they are capable of protecting cells, in a certain sense, against the damages induced by ionizing radiation.1,2 In model systems, one can speak about radiation protection based on preventing the attack of the reactive species, for example, •OH radicals, on the molecule of interest according to the competition between reactions 1 and 23
DNA + •OH f DNA•OH •
•
RSH + OH f RS + H2O
(1) (2)
In living cells, however, this kind of competition is unlikely to occur on a large scale since it requires extremely high concentrations of thiols, greatly exceeding their physiological levels. However, thiols can act through “chemical repair” of damage already induced in biomolecules, for example
DNA•OH + RSH f DNA(H)OH + RS•
(3)
This kind of “repair” brings no restoration of the original structure of DNA. One can speak about “protection” or “repair” only in the sense of converting more reactive C-centered radicals into less reactive, although not completely harmless, S-centered radicals.4-7 Accordingly, protection of DNA by thiols can occur also by electron transfer8
DNA•+ + RSH f DNA + RS• + H+
(4)
Although investigated in the past,9,10 the mechanism of DNA protection by thiols still attracts attention.11-13 The rate constants for repair reactions are of particular interest as far as under* Author to whom correspondence should be addressed. E-mail: brede@ mpgag.uni-leipzig.de. † University of Leipzig. ‡ Institute of Surface Modification. § Adam Mickiewicz University.
standing of this mechanism is concerned. It appears that the net charge on the thiols strongly influences the repair rate constants observed for DNA (investigated both in aqueous solution and in prokaryotic cells) and for its model system, polyuracil. This effect is explained by counterion condensation and the co-ion depletion phenomenon.14-16 The fundamental mechanisms of direct and indirect DNA damage seem to be known.1 Furthermore, the probable structures of the radicals involved are known (i.e., radical cation centers at guanine, radical anion centers at thymine or cytosine,17 and •OH radical and H• atom adducts to unsaturated bonds of nucleobases1). Despite these advances in understanding, it is still difficult to state precisely which radical structure is actually being repaired. There are four different nucleobases in the DNA molecule. Their primary radicals can undergo further reactions, including protonation or reaction with the sugar moiety.18 Moreover, in double-stranded DNA, considerable attack of reactive species at the sugar moiety itself can occur due to the hindered accessibility of the nucleobases. Considering how thoroughly investigated the thiol-repair effect has been in systems mimicking biological ones as well as in cells themselves (see previous paragraph), surprisingly little is known about the rate constants for the thiol-repair reactions with pyrimidine-derived radicals having well-defined structures (Pyr•)
Pyr• + RSH f PyrH + RS•
(5)
The source of the difficulty is that direct observation of the repair reaction is complicated by many competing reactions of thiols. By monitoring the absorption of the disulfide radical anion, Adams et al.19 determined, in an indirect manner, the upper limit of the repair rate constant krepair < 1 × 107 dm3 mol-1 s-1 for •OH radical adducts of uracil and deoxythymidine being repaired by cysteamine. Nucifora et al.20 used pulse radiolysis with electron paramagnetic resonance (EPR) detection to investigate the repair reaction of •OH radical adducts of thymine and uracil by
10.1021/jp061574e CCC: $33.50 © 2006 American Chemical Society Published on Web 06/02/2006
Repair Reactions of Pyr-Derived Radicals CHART 1: Structures of the Pyrimidines and Their Derivatives and the Thiols Used in the Experiments along with the pKa Values of the Compounds
cysteamine and cysteine. Here the estimated rate constants of repair (reaction 5) range from krepair e 3 × 105 dm3 mol-1 s-1 to krepair e 1 × 106 dm3 mol-1 s-1 at 5-10 °C. The aim of the present paper is the direct determination of the precise repair rate constants by time-resolved techniques. We generated various pyrimidinyl radicals (Pyr•), having defined structures, by pulse radiolysis and investigated their behavior in the presence of different aliphatic thiols. In this manner, we determined the repair rate constants for a variety of pyrimidinyl radicals being repaired by selected thiols. These rate constants are discussed on the basis of quantum chemical calculations. 2. Materials and Methods 2.1. Chemicals. Thiols were purchased from Merck (2mercaptoethanol, MerSH, and cysteine, CysSH), from Sigma (cysteamine hydrochloride, CyAmSH) and from Fluka Chemika (penicillamine, PenSH); they had purities of 98-99%. Pyrimidines (thymine, T, uracil, U, cytosine, C; purity 97-98%) were obtained from Aldrich. Their nucleosides (uridine, 2′-deoxythymidine, and cytidine) and nucleotides (5′-monophosphate disodium salts of uridine, UMP, 2′-deoxythymidine, dTMP, and cytidine, CMP) were from Fluka Biochemica (purity g 99%). 5-Bromouracil (BrU) was purchased from Sigma (98% purity).
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12739 The structures of the pyrimidines and thiols used in the experiments together with their pKa values are presented in Chart 1. Potassium peroxodisulfate (K2S2O8) came from Merck (purity 99%). 2-Propanol (iPrOH), tert-butyl alcohol (t-BuOH), and acetone with 99.9% purity were also from Merck. The distilled water used in all the experiments was treated with the Millipore Milli-Q Plus system. 2.2. Pulse Radiolysis. Pulse radiolysis experiments were carried out using a pulse-transformer-type accelerator ELIT (Institute of Nuclear Physics, Novosibirsk, Russia) generating 10 ns pulses of 1 MeV electrons with 20-100 Gy doses.24 The absorbed dose per pulse was measured with an electron dosimeter (absorption of a hydrated electron at 580 nm in slightly alkaline solution). Transient species were detected with an optical absorption setup consisting of a pulsed xenon lamp (XBO 450, Osram), a Spectra Pro-500 monochromator (Acton Research Corporation), a R9220 photomultiplier (Hamamatsu Photonics), and a 1 GHz digitizing oscilloscope (TDS 640, Tektronix). All experiments were performed at room temperature either in pure water or in water/alcohol mixtures to scavenge •OH radicals (2 mol dm-3 iPrOH and 0.5 mol dm-3 t-BuOH). The pH was adjusted with HCl; H2SO4 was used only for the measurements concerning Pyr•(N1) radicals. The pH was measured with a 540GLP pH meter (WTW). Only freshly prepared solutions were used. During irradiation, the samples flowed continuously through a quartz cell with an optical path length of 1 cm. Approximately, 15 min prior to the irradiations and also for the duration of the measurements, solutions were bubbled with the highest available purity N2O (in the experiments requiring the absence of a hydrated electron) or with N2 in all of the other experiments. For one set of measurements, 1 mol dm-3 acetone was used instead of N2O to scavenge hydrated electrons while the solution was bubbled with N2. In most of the measurements, a cut-off filter (320 nm) was used for minimizing the steadystate photolysis of the pyrimidines by the UV light. 3. Results and Discussion Experimental results presented herein were obtained from the direct observation of time profiles of the pyrimidine radicals decaying according to reaction 5. For the investigations, pyrimidinederived radicals with known structures were chosen that exhibit well-defined absorption bands in the visible spectral region. In the pulse radiolysis of aqueous solutions of the pyrimidines, the corresponding Pyr• radicals were formed by the reactions of the pyrimidines with the primary radiolytic species such as hydroxyl radicals, solvated electrons, and hydrogen atoms. In some cases also secondary reactive transients (2-hydroxy-2propyl radicals) were used. The crucial problem of our studies is that the oxidizing agent that generates the pyrimidine radicals whose repair is to be studied will competitively oxidize the thiols to form the thiyl radicals, which are the products of the repair mechanism itself. This competition for the oxidizing agent lowers the yield of the pyrimidine radicals, which are the main species being monitored. Furthermore, the repair reaction, under study, is the second of two consecutive reactions whose initial reaction is the reaction that is directly in competition with the oxidation of thiols by hydroxyl radicals. The kinetic scheme of these competing and consecutive reactions is given in a simplified manner as follows
12740 J. Phys. Chem. B, Vol. 110, No. 25, 2006 CHART 2: Structures of the Investigated Radicals
Wo´jcik et al. cytosine, •OH attack at the C5 position occurs almost exclusively (87%) 29 while this percentage is slightly lower for uracil and thymine (80% and 65%, respectively).30 Considering that •OH radical attack at C5 prevails, we treated the species as one, described by the symbol Pyr•OH. Pulse radiolysis experiments were carried out in 5 × 10-3 mol dm-3 aqueous solutions of the nucleobases, their nucleosides, and their nucleotides at pH ) 2.8-3.0. The nucleobases were also irradiated at pH ) 5.0-5.5. In nucleosides and nucleotides, more than 80% of the •OH radicals attack the base not the sugar moiety.29 In water radiolysis, hydrated electrons (eaq-), hydroxyl radicals (•OH), and, to a minor extent, hydrogen atoms (H•) are formed. When •OH reactions are studied, hydrated electrons are converted into •OH radicals through the reaction with N2O
N2O + eaq- + H2O f N2 + OH- + •OH
(6)
Under our experimental conditions (pH ≈ 2.8; [H+] ) 1.6 × 10-3 mol dm3) around 12% of the hydrated electrons react also with protons. As the result of such a reaction hydrogen atoms are formed (cf. reaction 21). As mentioned above, •OH radicals add to the C5-C6 double bonds of pyrimidines1 a
Radicals of this structure appear also as moieties in nucleosides and nucleotides.
Because the reaction of interest is the second reaction in the branch of the reaction scheme containing the consecutive reactions (bold faced), the thiol concentration should be kept to a level as low as possible. This strategy mitigates the competition with the other competitive branch and allows pyrimidine radicals to build up to an adequate level for ease of observation and analysis. For our experiments, cysteamine in the form of a hydrochloride salt was used. •OH radicals are known to react with chloride ions generating chlorine atoms.25 Also because such a reaction is strongly reversible, under our experimental conditions its contribution to the observed effects can be neglected. The Pyr• radicals used in the experiments were chosen on the basis of (i) whether they were easily generated, (ii) whether they had clearly defined structures and spectra, and (iii) whether they had reasonably large extinction coefficients. The structures of the Pyr• radicals chosen with these three criteria are presented in Chart 2. 3.1. •OH Radical Adducts to the Pyrimidine C5-C6 Double Bond (I). First, •OH radical adducts to the pyrimidine C5-C6 double bonds (Pyr•OH) were chosen for monitoring their repair reactions (reaction 5). At neutral pH, •OH adducts to the pyrimidine C5-C6 double bonds (Pyr•OH) show a pronounced optical absorption band at λ ≈ 380 nm for T•OH and U•OH and at λ ≈ 340 nm and λ ≈ 440 nm for C•OH.26 At pH ) 3, all the spectra of these •OH adducts look similar; only the maximum at λ ≈ 340 nm for C•OH is shifted to λ ≈ 310 nm.27 The observed bands are superpositions of the optical absorptions coming from two isomers of the •OH radical adduct: Pyr•(C5OH) and Pyr•(C6OH) (Chart 2). They differ slightly with respect to the shape of the spectra (the C6 adducts have maxima blue-shifted by about 20 nm relative to the spectral maxima of the C5 adducts), but the two isomers have almost identical extinction coefficients.28 There are differences in the redox properties of the two isomers with the C6 adduct being slightly oxidizing and the C5 adduct being reducing. These differences have been exploited to determine the ratio of •OH radical addition to the C5 and C6 positions. In the case of
Pyr + •OH f Pyr•OH k7 ) (4.5-6.0) × 109 dm3 mol-1 s-1 (7) Analogous addition of this kind, although considerably slower (k ≈ 3 × 108 dm3 mol-1 s-1), occurs also for the hydrogen atoms.1 However, due to the differences in the H• and •OH reactivities and low extinction coefficients of the C5 hydrogen atom adducts to pyrimidines,31 the contribution of the hydrogen atoms to the observed effects is only of minor importance. To avoid the formation of the strongly light-absorbing disulfide radical anions (from thiolate anions and RS•), the pH of the irradiated solutions was kept well below the pKa of the thiols. The observed transient spectra were in accordance with data available in the literature.26,27 The addition of thiols (cysteamine, 2-mercaptoethanol, and cysteine) in increasing concentrations (1 × 10-3, 2 × 10-3, and 3 × 10-3 mol dm-3) did not cause any marked changes in the shape of the spectra although increasing the concentration of the thiols led to a lower initial optical density change (∆O.D.). Only for penicillamine did an additional absorption appear around 330 nm. However, for all the thiols under study, accelerations in the decays of the time profiles of the Pyr•OH radicals were registered; cf. Figure 1. The lowering of the absorbance amplitude of the pyrimidine radicals as well as a slightly faster formation of the observed species in the presence of thiols were both caused by the competition between reactions 7 and 832
RSH + •OH f RS• + H2O k8 ) (3.7-8.6) × 109 dm3 mol-1 s-1 (8) Thus, at the highest thiol concentrations (3 × 10-3 mol dm-3) used in these experiments, the Pyr•OH absorptions were reduced by about 50%. However, the acceleration of the decay of the pyrimidine radicals is the result of the chemical repair reaction of Pyr•OH by the added thiols
Pyr•OH + RSH f Pyr(H)OH + RS•
(9)
The chemical repair reaction represents an additional channel
Repair Reactions of Pyr-Derived Radicals
J. Phys. Chem. B, Vol. 110, No. 25, 2006 12741
Figure 2. (A) Time profiles of the U•OH radical decaying in the presence of various concentrations of cysteamine together with the firstorder fits: (a) 0, (b) 1 × 10-3, (c) 2 × 10-3, (d) 3 × 10-3 mol dm-3 CyAmSH. (B) Stern-Volmer-type plot.
Figure 1. Normalized time profiles of the Pyr•OH radicals as a function of cysteamine hydrochloride concentration: (a) 0, (b) 1 × 10-3, (c) 2 × 10-3, (d) 3 × 10-3 mol dm-3 CyAmSH. All experiments were carried out with N2O-saturated aqueous solutions of 5 × 10-3 mol dm-3 nucleobase at pH ) 2.8, dose per pulse 50 Gy: (A) uracil, (B) thymine, (C) cytosine.
for the decay for the Pyr•OH radicals that, in the absence of thiols, decay mainly by recombination27
2 Pyr•OH f products k10 ) (2.2-7.0) × 108 dm3 mol-1 s-1 (10) On the basis of the changes registered in the decay kinetics of the pyrimidinyl radicals in the presence of thiols, the repair rate constants k9 were calculated. These rate constants, together with the wavelengths used for monitoring changes in the kinetics, are shown in Table 1. Under our experimental conditions, it was possible to fit all the time profiles with firstorder exponential decay functions to estimate the acceleration of the decays (shortening lifetimes τ) with increasing thiol concentrations. From the slope of the curves representing changes of the observed rate constants of decay (kobs ) 1/τ) versus concentrations of thiols (Stern-Volmer-type plots), we determined the repair rate constants k9 (Table 1). In the fitting
procedure, only the first part of the decay (up to 200 µs) was taken into account since it represented the observed changes in the most reliable way. On longer time scales, additional absorptions from stable products could interfere with the observation of the decays of the radicals.26 An example of the first-order fitting together with the Stern-Volmer-type plot based on it is shown in Figure 2. Recombination reaction 10 of the observed Pyr•OH radicals competes with the reaction of interest, i.e., the repair reaction of Pyr•OH by thiols (reaction 9). Therefore, the observed time profiles have a contribution from second-order decay kinetics. This contribution depends on the dose. The first part of the second-order decay can be well described with first-order kinetics. As mentioned in the previous paragraph, during the data evaluation process this first part of the time profile was fitted with a first-order exponential function. Data sets for various doses (20 and 50 Gy) were treated separately. Although the contribution from the second-order decay was larger at 50 Gy than that at 20 Gy, the general tendency observed after thiol addition, reflected by the shortening of the lifetimes of Pyr•OH radicals, was the same at both doses. Such an approach enabled determination of the rate constants without applying mixedorder kinetic analysis. In the experiments the doses of 20 or 50 Gy were used. Such large doses were required because of the low extinction coefficients of the observed species (380 ) 700 dm3 mol-1 cm-1 and 380 ) 750 dm3 mol-1 cm-1 for U•OH and T•OH, respectively, and 440 ) 550 dm3 mol-1 cm-1 for C•OH 27). Variations of the dose (20 or 50 Gy) and pH (3.0 or 5.5) resulted, nevertheless, in similar rate constants. Therefore, in Table 1, only one set of data obtained under standard conditions (pH ) 2.8-3.0, 50 Gy) is presented. The rate constants, so determined, are somewhat higher than those previously estimated by Nucifora et al.20 From Table 1,
12742 J. Phys. Chem. B, Vol. 110, No. 25, 2006
Wo´jcik et al.
TABLE 1: Repair Rate Constants k9 Determined at pH ) 2.8-3.0 with 50 Gy Dose Per Pulsea
radical (I)
λmonitoring
CyAmSH k9 × 10-6 (dm3 mol-1 s-1)
OH T•OH C•OH uridine•OH deoxythymidine•OH cytidine•OH UMP•OH dTMP•OH CMP•OH
390 nm 380 nm 440 nm 390 nm 380 nm 440 nm 380 nm 380 nm 440 nm
2.6 (
