Radicals Derived from Histone Hydroperoxides Damage Nucleobases

Chem. Res. Toxicol. 2000, 13, 665-672

665

Radicals Derived from Histone Hydroperoxides Damage Nucleobases in RNA and DNA Catherine Luxford, Roger T. Dean, and Michael J. Davies* The Heart Research Institute, 145 Missenden Road, Camperdown, Sydney, NSW 2050, Australia Received March 6, 2000

Exposure of individual histone proteins (H1, H2A, H2B, H3, or H4) and histone octamers (consisting of two molecules each of H2A, H2B, H3, and H4) to hydroxyl radicals, generated by γ-irradiation, in the presence of O2 generates protein-bound hydroperoxides in a dosedependent fashion; this is in accord with previous studies with other proteins. These histone hydroperoxides are stable in the absence of exogenous catalysts (e.g., heat, light, and transition metal ions), but in the presence of these agents decompose rapidly to give a variety of radicals which have been identified by EPR spin trapping. Histone hydroperoxide-derived radicals generated on decomposition of the hydroperoxides with Cu+ react with both pyrimidine and purine nucleobases. Thus, with uridine the histone hydroperoxide-derived radicals undergo addition across the C5-C6 double bond of the pyrimidine ring to give cross-linked adduct species which have been identified by EPR spectroscopy. HPLC analysis of the products generated on reaction of histone hydroperoxide-derived radicals with 2′-deoxyguanosine, or intact calf thymus DNA, has shown that significant levels of the mutagenic oxidized DNA base 8-oxo-7,8-dihydro2′-deoxyguanosine (8-oxodG) are formed, with the yield dependent on the individual histone protein, the presence of hydroperoxide functions, and the concentration of metal ion. These studies demonstrate that initial oxidative damage to individual histone proteins or histone octamers can result in the transfer of oxidative damage to associated DNA via the formation and subsequent decomposition of protein hydroperoxides to reactive radicals, and provide a novel route for the formation of mutagenic lesions in DNA.

Introduction The generation of free radicals within biological systems results in a wide spectrum of oxidative damage encompassing all major components of cells and their surrounding milieu (1-3). Proteins comprise a major, if not the most significant, cellular target for radicals as a result of their abundance and high rate constants for reaction with radicals (reviewed in refs 2 and 4). Radical attack on proteins results in oxidative damage, which can include structural modification (e.g., fragmentation, crosslinking, and side chain alterations), loss of enzymatic or structural function, and altered cellular handling and turnover (reviewed in refs 2 and 4). Radical-mediated damage to both intra- and extracellular proteins has been demonstrated in many diseases (reviewed in refs 2 and 4). Thus, evidence has been presented for the formation of elevated levels of protein oxidation products in a number of chronic human diseases, including cataract formation, chronic inflammation, cancer, atherosclerosis, and diabetes (reviewed in refs 2, 4, and 5); in some of these cases, protein oxidation appears to be the causative agent. In various cancers and HIV infection, radical damage has also been shown to cause disruption of cell cycle control via activation and/or inactivation of oncogenes, including p53, and transcription factors such as NF-κB (reviewed in ref 3). Oxidative damage has also been implicated in the formation of lipofuschin and other * To whom correspondence should be addressed: The Heart Research Institute, 145 Missenden Rd., Camperdown, Sydney, NSW 2050, Australia. Telephone: +61 2 9550 3560. Fax: +61 2 9550 3302. E-mail: [email protected].

modified protein components during aging (reviewed in refs 2, 4, 6, and 7). It has been demonstrated that oxidative damage to protein side chains can give rise to the formation of further reactive intermediates (8-15), as well as unreactive products such as hydroxylated materials and carbonyl groups (reviewed in refs 2, 4, 5, and 7). The reactive materials include both reducing agents [principally protein-bound 3,4-dihydroxyphenylalanine (8, 9)] and oxidizing species [primarily hydroperoxides (8, 10)]. The protein hydroperoxides, which are long-lived in vitro in the absence of exogenous catalysts (including light, heat, and redox-active transition metal ions) (8, 10, 12, 13), decompose rapidly in the presence of these agents with the formation of a variety of reactive radicals (1315); these include protein-derived alkoxyl (protein-O•), peroxyl (protein-OO•), and carbon-centered species (proteinC•). The formation of these hydroperoxide-derived species allows the propagation of oxidative damage and the occurrence of protein chain reactions (2, 14, 16), as well as the generation of oxidative damage on other biomolecules, including lipids, proteins, antioxidants, and DNA (13, 17-19).1 Protein-bound 3,4-dihydroxyphenylalanine has been shown to catalyze DNA damage in the presence of redox-active metal ions (11). Large numbers of proteins are associated with, or in proximity to, DNA within the cell nucleus as a result of their enzymatic and structural functions, and it has been demonstrated that such proteins are subject to oxidative 1

P. E. Morgan, C. L. Hawkins, and M. J. Davies, unpublished data.

10.1021/tx000053u CCC: $19.00 © 2000 American Chemical Society Published on Web 06/16/2000

666

Chem. Res. Toxicol., Vol. 13, No. 7, 2000

damage (1). Furthermore, it is well established that within cell nuclei, histone and other nuclear proteins are major targets for radiation-induced damage (20-24). Thus, irradiation of mixtures of isolated histone proteins results in levels (within 5%) of amino acid consumption almost identical to those observed with deoxyribonucleohistone complexes irradiated under identical conditions (20, 21), implying that the majority of damage occurs on nuclear proteins, rather than DNA, under the conditions that were employed. We therefore hypothesized that the formation and subsequent decomposition of hydroperoxides on individual nuclear histone proteins, and the histone octamer core, might provide a potential source of oxidative damage to DNA. This process would be expected to be facilitated by transition metal ions (especially Cu2+) which are known to be present within the nucleus (25). This proposition has been tested in the study reported here and complements recent studies on the reactions of other peptide and protein hydroperoxides (13, 19).

Materials and Methods Materials. Calf thymus histones (H1, subgroup f1, very lysine-rich fraction; H3, subgroup f3, arginine-rich fraction; H4, subgroup f2a1, slighty arginine-rich fraction; and nucleosome core octamers), catalase (bovine liver, 15 000 units mg-1), ebselen, reduced glutathione, xylenol orange, MNP,2 and calf thymus DNA (lyophilized) were purchased from Sigma (St. Louis, MO). Calf thymus histones (H2A, subgroup f2a2, lysine-rich fraction; and H2B, subgroup f2b, slighty lysine-rich fraction) were purchased from Fluka. Authentic 8-oxodG was purchased from Sapphire Bioscience (Cayman Chemical). DMPO was obtained from Sigma/Aldrich Chemical Co. (Milwaukee, WI) and purified before being used by filtration through activated charcoal. Trace metal ions were removed from the calf thymus DNA samples by dissolution (500 µg mL-1) and incubation overnight at 4 °C in the presence of Chelex 100 resin. The resin was subsequently removed by filtration through 5 µm filters. Solutions were subsequently aliquoted and stored at -20 °C before being used. All buffers and aqueous solutions were made up with Nanopure water purified using a four-stage Milli Q system (MilliporeWaters). Microcon microconcentrators were supplied by Millipore Corp. (Bedford, MA). Chelex 100 resin was purchased from Bio-Rad Laboratories (Hercules, CA). All other chemicals were ACS or AR grades and purchased from Sigma, Merck (Darmstadt, Germany), or ICN (Costa Mesa, CA). Solvents were HPLC grade and supplied through Mallinckrodt (St. Louis, MO). Formation of Histone Protein Hydroperoxides via γ-Irradiation. Individual histone proteins, or histone octamers (2-4 mg of protein mL-1), were exposed to γ-radiation in the presence of O2 using a 60Co source to final doses of 1000 or 2000 Gy at a dose rate of ca. 19 Gy min-1 as described previously (13). Immediately postirradiation, catalase (100 units mg-1 histone protein) was added to remove H2O2 formed during the irradiation and the solutions incubated at room temperature for 5-10 min. The yield of protein hydroperoxides is not affected by such treatment with catalase (10). The histone hydroperoxide solutions were subsequently aliquoted and stored at -80 °C in the absence of light. Quantification of Histone Protein Hydroperoxide Concentrations. Hydroperoxide concentrations were determined by either an iodometric assay (26) or a modified FOX (FeSO4/ xylenol orange) assay (27, 28). 2 Abbreviations: dG, 2′-deoxyguanosine; DMPO, 5,5-dimethyl-1pyrroline N-oxide; DMPO-OH, the radical adduct formed by addition of HO• to DMPO; EC, electrochemical; G, radiation yield (number of species formed per 100 eV of absorbed energy); histone-C•, histone carbon-centred radical; histone-O•, histone alkoxyl radical; histoneOO•, histone peroxyl radical; histone-OOH, histone hydroperoxides; HO•, hydroxyl radical; MNP, 2-methyl-2-nitrosopropane; 8-oxodG, 8-oxo-7,8-dihydro-2′-deoxyguanosine.

Luxford et al. EPR Spectroscopy. Radical species were detected using the spin traps DMPO and MNP and electron paramagnetic resonance (EPR) spectroscopy using a Bruker EMX X-band spectrometer with 100 kHz modulation at 20 °C (13). Instrumental settings and substrate concentrations are described in detail in the figure legends. Hyperfine coupling constants were measured directly from individual EPR spectra field scans and confirmed by computer simulation using the WINSIM program (29), with correlation coefficients between experimental data and simulations typically greater than 0.95. All solutions used for the EPR spin trapping experiments were thoroughly deoxygenated by bubbling with N2 gas. Reduction of Histone Hydroperoxides. Histone H1 hydroperoxides were reduced to corresponding alcohols by incubation with 83 µM ebselen (10 mM stock in 100% ethanol) and 3.3 mM reduced glutathione (GSH) at 37 °C for 15 min. Ebselen and GSH were subsequently removed via centrifugation through Microcon-3 (3 kDa molecular mass cutoff) microconcentrators at 7500g for 30 min. The eluates were discarded, and the retentates containing the histone H1 alcohols were washed by addition of 100 µL of H2O with gentle mixing and the columns respun at 7500g for a further 30 min. These final retentate solutions containing the histone H1 alcohols were collected and stored at -20 °C until they were uses. Residual hydroperoxide concentrations were determined using the FOX assay; >90% loss of hydroperoxides was routinely achieved. Oxidation of 2′-Deoxyguanosine and DNA by Histone Hydroperoxides. In final sample volumes of 400 µL, individual histone, or octamer, hydroperoxide solutions (110-300 µM hydroperoxide) were added to solutions of 2′-deoxyguanosine (200 µM) or calf thymus DNA (50 µg). Control samples were held on ice prior to purification. Reactions were initiated by the in situ generation of Cu+ [formed by sequential addition of deoxygenated solutions of CuSO4 (150 µM) and TiCl3 (100 µM) which produces Cu+ in a stoichiometric manner (30)]. Samples were incubated at 20 °C for either 5 min (2′-dG experiments) or 10 min (DNA experiments). The 2′-dG/histone hydroperoxide reactions were halted by addition of 1 volume of chloroform, and subsequent mixing and centrifugation at 15000g for 2 min. The 2′-dG samples were collected from upper (aqueous) layers and further purified of any residual histone proteins by centrifugation through Microcon-10 microconcentrators (10 kDa molecular mass cutoff) at 9000g for 15 min. The eluates were then concentrated to dryness under vacuum, resuspended in 100 µL of H2O, and stored at -80 °C prior to HPLC analysis. Calf thymus DNA/histone hydroperoxide reactions were stopped by the addition of 0.1 volume of 3 M sodium acetate and 2 volumes of ice-cold ethanol to precipitate the DNA. The DNA was collected after storage at -80 °C for >30 min and subsequent centrifugation at 15000g for 30 min at 4 °C. The DNA pellets were then rinsed with 500 µL of 70% ice-cold ethanol and recentrifuged at 15000g for 2 min. The supernatants were discarded, and the pellets were allowed to dry in air, then resuspended in 100 µL of H2O, and hydrolyzed as described previously (11). HPLC Analysis of 8-OxodG Formation. Separation and quantification of 8-oxodG were performed by HPLC as described previously (11), except for the following modifications. Samples were separated with a Zorbax ODS column (analytical, 4.6 mm × 250 mm, 5 µm particle size; Hewlett-Packard, Santa Clarita, CA) at 30 °C, using isocratic elution with degassed solutions of 50 mM potassium phosphate (pH 5.5) and methanol (81:19 v/v) at a flow rate of 1 mL min-1 for 30 min. Electrochemical detection potentials were set at 50 and 400 mV for the first and second electrodes, respectively. Peak identification was made on the basis of retention time, comparison with authentic 8-oxodG standards, and sample spiking; authentic 8-oxodG standards eluted at ca. 8.2 min using these conditions. Calibration of the electrochemical detector peak area response was carried out for each experiment using a range of pure 8-oxodG standards with a sensitivity limit of approximately 1 pmol. Unmodified nucleosides and 2′-dG standards (retention time of

Histone Hydroperoxide Radicals Oxidize RNA and DNA

Chem. Res. Toxicol., Vol. 13, No. 7, 2000 667

Table 1. Yields of Histone Hydroperoxides Formed by γ-Irradiation in the Presence of O2a [hydroperoxide] (µM) ( SD histone

1000 Gy irradiation

2000 Gy irradiation

H1 H2A H2B H3 H4 core octamers

136 ( 12 77 ( 3 172 ( 16 102 ( 23 114 ( 7 244 ( 17

432 ( 104 119 ( 9 413 ( 83 174 ( 8 267 ( 79 319 ( 61

a Hydroperoxides were generated by 60Co γ-irradiation of individual histone solutions (2-4 mg/mL) in the presence of O2 to final doses of 1000 and 2000 Gy at a dose rate of ca. 19 Gy/min. Immediately postirradiation, catalase (100 units/mg of histone protein) was added to eliminate H2O2 generated during the irradiation (see Materials and Methods). Yields of hydroperoxides were determined by either iodometric or FOX assays (see Materials and Methods). Values are means ( SD of 40 determinations.

ca. 6.4 min) were monitored with UV detection at 254 nm, and the amount of injected base or DNA was calculated from the peak areas using a range of authentic 2′-dG standards. Statistical Analysis. Statistical analyses were carried out using unpaired t tests within the MYSTAT statistical applications program.

Results Formation of Histone Hydroperoxides. γ-Irradiation of individual histones, or histone octamers, in the presence of O2 resulted in the dose-dependent formation of hydroperoxides as observed with other peptides and proteins (8, 10, 12, 13); the concentrations of hydroperoxides formed at total irradiation doses of 1000 and 2000 Gy are given in Table 1. Stability of Histone Hydroperoxides. The histone hydroperoxides were found to be stable in the absence of exogenous catalysts, with no discernible loss of hydroperoxide content at temperatures below -20 °C over the course of 12 months (data not shown). Incubation at 37 °C greatly increased the rate of hydroperoxide breakdown (Figure 1); less rapid loss was detected at room temperature or 4 °C (data not shown). With each histone, there was an initial rapid loss of hydroperoxide, with a subsequent slow rate of decay, with the rate and extent of overall loss being dependent on the histone under examination. This is presumably due to the progressive breakdown of different subsets of hydroperoxides on the proteins, as observed with other materials (8, 10, 12, 13). Residual hydroperoxide levels after incubation for 21 h at 37 °C ranged from 17.5% (histone H3) to 62% (histone H4). Formation and Identification of Histone Hydroperoxide-Derived Radicals. Addition of Cu+ [formed by the in situ reduction of Cu2+ (150 µM) by Ti3+ (100 µM) as described previously (13)] to deoxygenated histone hydroperoxide solutions (110-350 µM), containing catalase (300 units) and the spin trap DMPO (150 mM) at 20 °C, resulted in the detection of a number of radical adduct signals by EPR spectroscopy. Omission of either the histone hydroperoxides (by substitution with nonirradiated histone solutions, or reduction to the alcohol using ebselen/GSH) or the Cu+ redox couple resulted in the loss of these radical adduct signals, and the detection of low concentrations of the spin adduct DMPO-OH [believed to arise from autoxidation of Cu+, or via a nucleophilic reaction of H2O with the trap (31)]. Omission of either the Cu2+ or Ti3+ from the redox couple resulted in much weaker, but otherwise similar, EPR signals. The

Figure 1. Decomposition of histone hydroperoxides over time on incubation at 37 °C. The initial yields and the subsequent loss of hydroperoxides were determined using iodometric or modified FOX assays (see Materials and Methods) on (a) histones H1, H2A, and H2B and (b) histones H3 and H4 and core octamers. Hydroperoxides were generated as described in Materials and Methods. Values are expressed as a percentage of the initial (t ) 0) hydroperoxide concentrations and are means ( SD of five determinations.

failure to detect substrate-derived signals in the absence of the hydroperoxides confirms that the adduct signals were from radicals produced as a result of the Cu+catalyzed histone protein hydroperoxide decomposition. The radical adducts were stable over the period required (several minutes) to acquire their EPR spectra. Representative spectra obtained on Cu+-catalyzed decomposition of each of the individual histone hydroperoxides plus histone octamer hydroperoxides in the presence of DMPO are shown in Figure 2. These signals have been assigned to the adducts to DMPO of various histone-derived radicals on the basis of data obtained in previous studies (13-15). These assignments, together with their hyperfine coupling constants and relative concentrations (from computer simulation of the experimental spectra), are summarized in Table 2. Reactions of the Histone Hydroperoxide-Derived Radicals with DNA Nucleosides. Reaction of the radicals generated during the Cu+-catalyzed decay of individual histone hydroperoxides and histone octamers with nucleobases was investigated in EPR experiments utilizing the spin trap 2-methyl-2-nitrosopropane (MNP). Saturated solutions of the various nucleosides were employed to obtain maximal signal intensities. Incubation of such solutions with histone hydroperoxides (110300 µM hydroperoxide) was carried out under anoxic conditions in the presence of catalase (300 units) and the spin trap MNP (7.5 mM final concentration, from a 100 mM stock in CH3CN). Reactions were initiated by the

668


Luxford et al. Table 2. EPR Parameters of the Histone Radical Adducts Detected from the Cu+-Catalyzed Breakdown of Histone Hydroperoxides in the Presence of the Spin Trap DMPO histone protein spin adduct hydroperoxide assignment H1c H2A H2B H3 H4 core octamers

C-centeredd C-centerede C-centerede C-centeredd C-centerede C-centerede C-centeredd C-centerede C-centerede C-centeredd C-centerede C-centeredd C-centerede C-centeredd C-centerede

hyperfine coupling relative radical constant (mT)a concentration a(N) a(H) (% of total)b 1.536 1.565 1.570 1.536 1.572 1.600 1.538 1.558 1.599 1.535 1.573 1.534 1.571 1.535 1.569

2.034 2.268 2.374 2.029 2.237 2.425 2.035 2.223 2.450 2.032 2.252 2.028 2.270 2.033 2.268

47 ( 9 27 ( 12 17 ( 11 55 ( 16 31 ( 15 9(4 56 ( 4 35 ( 8 10 ( 3 51 ( 1 42 ( 1 58 ( 11 40 ( 13 56 ( 12 36 ( 8

a Typically (0.01 mT. b Mean percentage values (n ) 9 determinations, from computer simulations of experimental spectra) of total radical adduct concentrations detected. Values do not sum to 100% due to slight contributions from spin trap degradation products (typically ∼7%). c From ref 13; included for comparison. d Assigned to adduct species with partial structure •C(O)NHR, thought to arise from backbone cleavage; see the text for further details. e Assigned to side chain-derived radicals; see the text for further details.

Figure 2. EPR spectra detected on reaction of histone hydroperoxides with Cu+ in the presence of the spin trap DMPO. Histone protein hydroperoxides were generated by 60Co γ-irradiation of individual histone protein solutions as described in Figure 1 and Materials and Methods. Reactions were initiated by the addition of Cu+ [formed in situ by the sequential addition of deoxygenated Cu2+ (150 µM) and Ti3+ (100 µM) solutions] to histone hydroperoxide solutions (typically between 100 and 360 µM) containing catalase (300 units) and the spin trap DMPO (150 mM) at 20 °C. Hydroperoxides formed on (a) histone H1, (b) histone H2A, (c) histone H2B, (d) histone H3, (e) histone H4, and (f) core octamers. Typical EPR spectrometer settings were as follows: gain, 1 × 106; modulation amplitude, 0.025 mT; time constant, 81 ms; scan time, 83 s; center field, 347.5 mT; field scan, 8 mT; and with eight scans accumulated. Lines marked with O are assigned to DMPO-OH and lines marked with ( to •C(O)NHR radical adducts. Other lines are assigned to carboncentered radical adducts; see the text and Table 2 for further details.

addition of Cu+ (formed in situ as described above). In experiments where pyrimidine nucleosides were employed, signals which have been assigned to adducts to MNP of base-derived radicals were detected; Figure 3 shows representative EPR spectra obtained with uridine as the nucleoside and histone octamer-derived hydroperoxides. Omission of the spin trap resulted in the loss of all signals, whereas omission of the hydroperoxide or Cu2+-Ti3+ redox couple resulted in the detection solely of the well-characterized MNP breakdown product di-tertbutyl nitroxide, (tBu)2NO•. Omission of uridine resulted in the detection of very weak signals from the trapping of histone hydroperoxide-derived radicals (data not shown). The assignment of the signals detected in the presence of uridine and the histone octamer hydroperoxides (see the legend of Figure 3) has been made on the basis of previously reported data with other attacking radicals

(13, 32, 33). In each case, the observed spectra are consistent with the presence of two uridine-derived radicals, the uridine C5-yl and C6-yl adducts formed by addition of the hydroperoxide-derived radicals across the C5-C6 double bond on the uridine ring (cf. Scheme 1). This assignment is supported by the broad nature of the spectral lines in Figure 3, which is consistent with the addition of a large, presumably protein-derived, radical across the double bond of the pyrimidine ring. Only weak, unanalyzable signals were detected when purine bases were employed as the target. In light of the product analyses described in detail below, this is thought to be due to a slow rate of trapping of the purine radicals resulting from reaction with the hydroperoxides, or the instability of the resulting radical adducts, rather than a lack of reaction. Formation of 8-OxodG on Oxidation of 2′-dG and DNA by Histone Hydroperoxide-Derived Radicals. The formation of 8-oxodG on reaction of histone hydroperoxides with Cu+ in the presence of free 2′-dG or calf thymus DNA (after appropriate digestion; see Materials and Methods) was quantified by HPLC separation and EC and UV detection, with the (dual-channel) EC detector employed to quantitate 8-oxodG, and the UV channel the unoxidized parent. This approach allows the yield of 8-oxodG to be expressed per unoxidized parent, and compensates for any general loss of material during the digestion and purification procedures. Incubations typically involved addition of the histone hydroperoxides (110-300 µM hydroperoxide) to calf thymus DNA (50 µg) or 2′-dG (200 µM), with the reaction initiated by the in situ formation of Cu+ (see Materials and Methods). Figure 4 summarizes the data that were obtained. Reaction of 2′-dG (Figure 4a) with histone hydroperoxidederived radicals generated by Cu+ produced statistically significant (p < 0.05 compared to nonirradiated controls) elevated levels of 8-oxodG with all the irradiated histone proteins with the exception of histone H2A. The lack of significance with this histone probably reflects, at least



of O2 has been shown to generate high yields of hydroperoxides in a dose-dependent manner; this is in accord with studies on other proteins (including histone H1) and amino acids (8, 10, 12-14). These hydroperoxides are believed to arise via reactions 1, 2, and either 3 or 4 and 5. The hydroperoxides once formed are stable at low temperatures in the frozen state, but decompose slowly at 37 °C, in the absence of added metal ions, via complex kinetics; this is consistent with previous reports (8, 10, 12, 13) and presumably reflects the heterogeneous nature of the hydroperoxide formed at different sites on these proteins and their differing stabilities. The hydroperoxides decompose rapidly in the presence of Cu+, and generate a number of carbon-centered radicals, presumably via the formation of an initial undetected alkoxyl radical, as a result of the occurrence of a pseudo-Fenton reaction (reaction 6), and subsequent rearrangement and/ or fragmentation of these species [reaction 7; cf. studies with other proteins and amino acids (13-15)]. Previous studies with histone H1 have shown that replacement of the hydroperoxides with the corresponding alcohols formed by reduction of hydroperoxides does not yield these species in the presence of Cu+, confirming that radical formation is hydroperoxide-dependent (13).

Figure 3. EPR spectra detected on reaction of core octamer hydroperoxides with Cu+ in the presence and absence of uridine and the spin trap MNP. Hydroperoxides were generated by 60Co γ-irradiation of core octamer solutions, in the presence of O2, and assayed as reported in Materials and Methods. Reactions were initiated by addition of Cu+ (formed in situ as outlined in Materials and Methods) to deoxygenated, aqueous solutions of the substrates in the presence of the spin trap MNP (7.5 mM from a 100 mM stock in acetonitrile; final acetonitrile concentration of 7.5% v/v) and catalase (300 units). (a) Nonirradiated core octamers in the presence of uridine (saturated solutions) at 20 °C, (b) core octamer hydroperoxides in the absence of uridine, and (c) core octamer hydroperoxides in the presence of uridine (saturated solution) at 20 °C. Typical EPR spectrometer settings were as described in the legend of Figure 2 except for a field scan of 6 mT. Features marked with O are assigned to the spin trap degradation product (tBu)2NO•, b to (one or more) C5-yl radical adduct [a(N) ) 1.508 mT, a(H) ) 0.313 mT], and * to (one or more) C6-yl radical adduct [a(N) ) 1.556 mT, a(H) ) 0.242 mT, a(ring nitrogen) ) 0.293 mT], with the latter two adducts present at a concentration ratio of ca. 70:30 (determined by computer simulation of the experimental spectra, n ) 7).

in part, the lower levels of hydroperoxides obtained on irradiation of this particular protein (see Table 1). With calf thymus DNA (Figure 4b), decomposition of all the histone hydroperoxides with Cu+ yielded statistically significant (p < 0.05) elevated levels of 8-oxodG compared to those of nonirradiated and Cu+-free controls. The lower, but still statistically elevated, levels of 8-oxodG formed from both 2′-dG and DNA with some of the irradiated histones in the absence of added Cu+ are thought to be due to a slow rate of thermal decomposition of the histone hydroperoxides; similar effects have been observed previously with histone H1 hydroperoxides (13). The low yield of 8-oxodG observed with Cu+ with the nonirradiated histone proteins is ascribed to the formation of low levels of radical formation (O2-•/HO•) arising from the autoxidation of Cu+ (see the EPR results above).

Discussion Exposure of each of the individual histone proteins as well as intact core octamers to γ-radiation in the presence

histone + HO• f histone• + H2O

(1)

histone• + O2 f histone-OO•

(2)

histone-OO• + RH f histone-OOH + R•

(3)

histone-OO• + e- f histone-OO-

(4)

-

+

histone-OO + H f histone-OOH

(5)

histone-OOH + Cu+ f histone-O• + HO- + Cu2+ (6) histone-O• f histone-C•

(7)

The residues on which these hydroperoxide groups are situated, and their precise locations within the protein structures, remain to be elucidated. It has been established that irradiation of histone proteins gives rise to a much larger extent of loss of lysine residues (G ∼ 0.31) than of other residues (G e 0.16) (20, 21), and it is known that lysine residues are abundant on most of the histone proteins [e.g., ca. 30 mol % of the amino acids of histone H1, and >10 mol % of the residues in the other histones (34, 35)]. Furthermore, lysine residues are known to give high yields of hydroperoxides when irradiated in free solution (10, 12), and are therefore likely to be a major, but not the only, sites of the hydroperoxides detected on irradiated histones. This conclusion is supported by our previous product studies which have demonstrated the formation of significant yields of lysine hydroxides (at C-3, C-4, and C-5 on the side chain) on histone H1 after γ-irradiation in the presence of O2 (36). The detection of radical adduct species assigned to backbone radicals by EPR is consistent with R-carbon sites on the protein backbone also being a major site of hydroperoxide formation (15). It has been shown that the radicals derived from the histone octamer hydroperoxides react readily with uridine to give adduct species (i.e., protein/amino acid-RNA base cross-links); this is in accord with other recent studies where model peptides, proteins, and histone H1

670


Luxford et al. Scheme 1

were examined (13). The uridine-derived radical adducts detected in the study presented here are likewise assigned to adduct species formed by addition of the octamer-derived radicals to the C5-C6 double bond of the pyrimidine ring, to give a protein-nucleoside adduct, which then reacts with the spin trap to give the observed radical (cf. Scheme 1). The 70:30 ratio in spin adduct concentrations of the C5-yl over C6-yl adducts may be due to steric factors [cf. studies with free amino acids, peptides, and other proteins (13)]. The exact nature of the initial octamer-derived radicals which participate in this reaction, and hence the nature of the cross-link, remains to be determined, as does the significance of protein hydroperoxide-mediated damage to RNA. Covalent binding of a large protein molecule to RNA would however be expected to result in loss of structural or functional activity. Evidence has been previously presented for the formation of lysine-thymine cross-links both in γ-irradiated mixtures of lysine and thymine and in intact nucleohistone (37); this type of cross-link may be similar to that detected in the current study with uridine, though these previous workers postulated an alternative mechanism of formation. Furthermore, lysinethymine cross-links have also been reported after UV irradiation of nuclei and nucleosomes (38, 39). The detection of 8-oxodG as a product of oxidation of both 2′-dG and DNA is consistent with the purine bases also being significant targets for the initial histone hydroperoxide-derived radicals; this is in accord with our previous studies with histone H1 (13). Though there is controversy in the literature as to the absolute levels of this material formed in DNA and the role of artifactual oxidation (see, for example, ref 40), this is of minor importance in the current study as all of the analyses have been made with respect to control samples taken through identical sample handling and analysis procedures. The formation of significant concentrations of 8-oxodG has been shown to require the presence of the hydroperoxide, with only control levels detected with either nonirradiated, or irradiated and reduced, proteins when 2′-dG was employed as a target. The low, but statistically significant, levels detected with reduced histone H1 and DNA as a target are probably due to incomplete reduction of the initial histone hydroperoxide and/or the catalysis of radical formation by the low levels of protein-bound 3,4-dihydroxyphenylalanine formed on the protein by the irradiation process (9, 11). The latter material would not

be expected to be removed by the ebselen/GSH hydroperoxide reduction process (see Materials and Methods). The nature of the histone-derived radicals responsible for the formation of the 8-oxodG from either free 2′-dG or DNA is unclear, though it is likely to be either the initial alkoxyl radical formed from the hydroperoxides by Cu+, or peroxyl radicals, as previous studies have shown that carbon-centered radicals react with purine bases primarily via addition reactions (41, 42). Peroxyl radicals may be formed in these systems either as a result of addition of O2 to the carbon-centered radicals formed via the fragmentation and/or rearrangement of the initial alkoxyl radicals or via direct reaction of Cu2+ with the hydroperoxides. There is controversy in the literature as to whether alkoxyl radicals and/or peroxyl radicals can generate 8-oxodG from 2′-dG. Thus, it has been recently reported that peroxyl (see, for example, refs 43 and 44) and acyloxyl (45) radicals do not generate 8-oxodG from 2′-dG, and it has also been reported that peroxyl radicals are not capable of oxidizing 2′-dG to the corresponding base radical cation (which would be expected to undergo rapid hydration to give 8-oxodG) (46). In light of these reports, and the experimental data obtained in this study which show that 8-oxodG is generated in significant amounts, we propose that the 8-oxodG may arise via initial reaction of the alkoxyl radical with 2′-dG and subsequent conversion (possibly via nucleophilic reaction with water) of the adduct or base radical cation to 8-oxodG. Further mechanistic studies aimed at identifying the species involved in these reactions are underway. The yield of 8-oxodG formed from free 2′-dG as a result of oxidation by histone hydroperoxide-derived radicals in the presence of Cu+ is low, with between 0.07% (66 molecules of 8-oxodG per 105 dG molecules) to 1.14% (1137 molecules of 8-oxodG per 105 dG molecules) conversion with histone H2A and H1 hydroperoxides, respectively. This variation is believed to arise, at least in part, from the varied hydroperoxide concentrations that are generated (cf. Table 1) and confirms previous results showing a hydroperoxide concentration dependence of 8-oxodG formation (13). The yields of 8-oxodG obtained from DNA, when expressed per 105 molecules of parent base, are in general higher than those obtained with 2′-dG (compare panels a and b of Figure 4). This is particularly noticeable with histone H2A, and is somewhat surprising since significant reaction with the other DNA bases (and possibly the sugar-phosphate backbone) might be expected to occur,


Figure 4. Formation of 7,8-dihydro-8-oxo-2′-deoxyguanosine (8-oxodG) from dG (a) and calf thymus DNA (b) on reaction with histone and core octamer hydroperoxides in the presence of Cu+. Hydroperoxides were generated by 60Co γ-irradiation of individual histone protein solutions or core octamers (2-4 mg/mL) in the presence of O2 (see Materials and Methods and Table 1). Histone H1 hydroperoxides were reduced to their corresponding alcohols as described in Materials and Methods. Reactions were initiated by addition of Cu+ [formed in situ by the sequential addition of deoxygenated solutions of Cu2+ (150 µM) and Ti3+ (100 µM)] to the hydroperoxide solutions (85-300 µM) containing either 2′-deoxyguanosine (200 µM) or calf thymus DNA (ca. 50 µg) at 20 °C. Subsequent sample processing and quantification of 8-oxodG formation (by HPLC with UV and EC detection) was performed as outlined in Materials and Methods. Values quoted are means ( SD of 15 determinations. Values statistically significant above the nonirradiated (*), reduced (O), and Cu+-free (() controls at the p < 0.05 level are indicated. ND means not determined.

given the observed reaction with uridine (see above) and previous studies with histone H1 where lower yields of 8-oxodG were detected with DNA than with 2′-dG (13). This difference is thought to reflect a more efficient transfer of damage from the core histones (H2A, H2B, H3, and H4, either individually or as the octamer) to DNA, compared to that with histone H1 or other proteins and peptides, due to the intimate and strong electrostatic interactions between these (positively charged) proteins and (negatively charged) DNA (47). These interactions may be particularly favorable with histone H2A, and would be expected to maximize the potential for interaction between the radical species formed on the histone


protein and the DNA bases, and minimize the occurrence of other competing reactions of the protein-derived radicals (e.g., radical-radical termination and radical rearrangement reactions). This result is in accord with previous studies which have reported a preponderance of protein-DNA cross-links with the core histones in γ-irradiated chromatin (22). The high yield of 8-oxodG from 2′-dG in DNA, compared to that of free 2′-dG, may also be facilitated by damage transfer (positive hole migration) from other oxidized bases (e.g., adenosine) to guanosine as a result of the lower oxidation potential of the latter base (48). Other competing reactions of the protein-derived radicals must, however, still be major pathways for removal of the histone hydroperoxidederived radicals as the overall conversion of hydroperoxide into 8-oxodG in these systems can be calculated, under the conditions employed, as being on the order of 0.5% using a value of ca. 1100 molecules of 8-oxodG formed per 105 molecules of parent base in DNA with a hydroperoxide concentration of ca. 200 µM. The physiological and pathological occurrence of such damage transfer reactions from protein to RNA or DNA in intact nuclei remains to be established, as does the absolute requirement for added transition metal ions such as Cu+. Previous studies have reported the presence of redox-active transition metals in the nucleus [e.g., Cu2+ bound to DNA or nuclear matrix proteins (25, 49-51)]. However, some of the data obtained in the current study suggest that the requirement for metal ions in catalyzing the decomposition of the histone hydroperoxides may not be absolute, given that there appears to be a slow spontaneous (thermal) decomposition of these hydroperoxides at 37 °C, and that this process can result in the formation of statistically significant levels of 8-oxodG (cf. data for the formation of 8-oxodG from 2′-dG and DNA in Figure 4 in the absence of added Cu+). The data in Figure 4 are likely to underestimate the significance of this process given the short incubation times employed, during which only partial decomposition of the hydroperoxides will have occurred (cf. data in Figure 1). The presence of such transition metal ions might not therefore be an absolute prerequisite for DNA oxidation. This process may be important in vivo if the histone hydroperoxides are not rapidly repaired; the rate of removal or reduction of such hydroperoxides in vivo is unknown, and might be expected to be slower for the histone octamer proteins than for many other cellular proteins, due to the relative rigidity and inaccessibility of the octamer unit to reducing agents and protective enzymes (e.g., GSH, and other, peroxidases). Thus, the formation of hydroperoxides on nuclear proteins, such as the histones, and their subsequent decomposition to reactive radicals, may provide a novel and important route to the formation of both mutagenic (8-oxodG) lesions in DNA and protein-DNA cross-links.

Acknowledgment. We thank Dr. J. M. Gebicki for use of the 60Co source, Drs. J. M. Gebicki, S. Gebicki, C. Hawkins, and W. Jessup for helpful comments and suggestions, and the Association for International Cancer Research (U.K.) and the Australian Research Council for financial support.

References (1) von Sonntag, C. (1987) The chemical basis of radiation biology, pp 1-515, Taylor and Francis, London.

672


(2) Davies, M. J., and Dean, R. T. (1997) Radical-mediated protein oxidation: from chemistry to medicine, pp 1-443, Oxford University Press, Oxford. (3) Halliwell, B., and Gutteridge, J. M. C. (1999) Free Radicals in Biology and Medicine, 3rd ed., pp 1-936, Oxford University Press, Oxford. (4) Dean, R. T., Fu, S., Stocker, R., and Davies, M. J. (1997) The biochemistry and pathology of radical-mediated protein oxidation. Biochem. J. 324, 1-18. (5) Davies, M. J., Fu, S., Wang, H., and Dean, R. T. (1999) Stable markers of oxidant damage to proteins and their application in the study of human disease. Free Radical Biol. Med. 27, 11511163. (6) Sohal, R. S., and Weindruch, R. (1996) Oxidative stress, caloric restriction, and aging. Science 273, 59-63. (7) Stadtman, E. R., and Berlett, B. S. (1998) Reactive oxygenmediated protein oxidation in aging and disease. Drug Metab. Rev. 30, 225-243. (8) Simpson, J. A., Narita, S., Gieseg, S., Gebicki, S., Gebicki, J. M., and Dean, R. T. (1992) Long-lived reactive species on free-radicaldamaged proteins. Biochem. J. 282, 621-624. (9) Gieseg, S. P., Simpson, J. A., Charlton, T. S., Duncan, M. W., and Dean, R. T. (1993) Protein-bound 3,4-dihydroxyphenylalanine is a major reductant formed during hydroxyl radical damage to proteins. Biochemistry 32, 4780-4786. (10) Gebicki, S., and Gebicki, J. M. (1993) Formation of peroxides in amino acids and proteins exposed to oxygen free radicals. Biochem. J. 289, 743-749. (11) Morin, B., Davies, M. J., and Dean, R. T. (1998) The protein oxidation product 3,4-dihydroxyphenylalanine (DOPA) mediates oxidative DNA damage. Biochem. J. 330, 1059-1067. (12) Gebicki, J. M. (1997) Protein hydroperoxides as new reactive oxygen species. Redox Rep. 3, 99-110. (13) Luxford, C., Morin, B., Dean, R. T., and Davies, M. J. (1999) Histone H1- and other protein- and amino acid-hydroperoxides can give rise to free radicals which oxidise DNA. Biochem. J. 344, 125-134. (14) Davies, M. J., Fu, S., and Dean, R. T. (1995) Protein hydroperoxides can give rise to reactive free radicals. Biochem. J. 305, 643-649. (15) Davies, M. J. (1996) Protein and peptide alkoxyl radicals can give rise to C-terminal decarboxylation and backbone cleavage. Arch. Biochem. Biophys. 336, 163-172. (16) Neuzil, J., Gebicki, J. M., and Stocker, R. (1993) Radical-induced chain oxidation of proteins and its inhibition by chain-breaking antioxidants. Biochem. J. 293, 601-606. (17) Fu, S., Gebicki, S., Jessup, W., Gebicki, J. M., and Dean, R. T. (1995) Biological fate of amino acid, peptide and protein hydroperoxides. Biochem. J. 311, 821-827. (18) Gebicki, S., Dean, R. T., and Gebicki, J. M. (1996) in Oxidative Stress and Redox Regulation: Cellular Signaling, AIDS, Cancer and Other Diseases, p 139, Institute Pasteur, Paris. (19) Gebicki, S., and Gebicki, J. M. (1999) Crosslinking of DNA and proteins induced by protein hydroperoxides. Biochem. J. 338, 629-636. (20) Robinson, M. G., Weiss, J. J., and Wheeler, C. M. (1966) Irradiation of deoxyribonucleohistone solutions. I. Amino acid destruction. Biochim. Biophys. Acta 124, 176-180. (21) Robinson, M. G., Weiss, J. J., and Wheeler, C. M. (1966) Irradiation of deoxyribonucleohistone solutions. II. Labilization of the DNA-histone linkage. Biochim. Biophys. Acta 124, 181186. (22) Mee, L. K., and Adelstein, S. J. (1981) Predominance of core histones in formation of DNA-protein crosslinks in γ-irradiated chromatin. Proc. Natl. Acad. Sci. U.S.A. 78, 2194-2198. (23) Elia, M. C., and Bradley, M. O. (1992) Influence of chromatin structure on the induction of DNA double strand breaks by ionizing radiation. Cancer Res. 52, 1580-1586. (24) Xue, L. Y., Friedman, L. R., Oleinick, N. L., and Chiu, S. M. (1994) Induction of DNA damage in γ-irradiated nuclei stripped of nuclear protein classes: differential modulation of double-strand break and DNA-protein crosslink formation. Int. J. Radiat. Biol. 66, 11-21. (25) Dijkwel, P. A., and Wenink, P. W. (1986) Structural integrity of the nuclear matrix: differential effects of thiol agents and metal chelators. J. Cell Sci. 84, 53-67. (26) Jessup, W., Dean, R. T., and Gebicki, J. M. (1994) Iodometric determination of hydroperoxides in lipids and proteins. Methods Enzymol. 233, 289-303. (27) Wolff, S. P. (1994) Ferrous ion oxidation in the presence of ferric ion indicator xylenol orange for measurement of hydroperoxides. Methods Enzymol. 233, 182-189.

Luxford et al. (28) Gay, C., Collins, J., and Gebicki, J. M. (1999) Hydroperoxide assay with the ferric-xylenol orange complex. Anal. Biochem. 273, 149155. (29) Duling, D. R. (1994) Simulation of multiple isotropic spin trap EPR spectra. J. Magn. Reson. 104B, 105-110. (30) Gilbert, B. C., Stell, J. K., Whitwood, A. C., Halliwell, C., and Sanderson, W. R. (1991) Mechanisms of peroxide decomposition. An EPR investigation of the reactions between some transition metal ions (TiIII, FeII, CuI) and monoperoxyphthalic acid and its anion. J. Chem. Soc., Perkin Trans. 2, 629-634. (31) Hanna, P. M., Chamulitrat, W., and Mason, R. P. (1992) When are metal ion-dependent hydroxyl and alkoxyl radical adducts of 5,5-dimethyl-1-pyrroline N-oxide artifacts? Arch. Biochem. Biophys. 296, 640-644. (32) Hazlewood, C., and Davies, M. J. (1995) EPR spin trapping studies of the reaction of radicals derived from hydroperoxide tumourpromoters with nucleic acids and their components. J. Chem. Soc., Perkin Trans. 2, 895-901. (33) Ho, W. F., Gilbert, B. C., and Davies, M. J. (1997) EPR spin trapping studies of the reaction of the hydroxyl and other electrophilic radicals with uridine and related compounds. Isotopic substitution to refine analyses and aid quantification. J. Chem. Soc., Perkin Trans. 2, 2533-2538. (34) DeLange, R. J., and Smith, E. L. (1972) The structure of histones. Acc. Chem. Res. 5, 368-373. (35) Wolffe, A. (1992) Chromatin: Structure and Function, Academic Press, London. (36) Morin, B., Bubb, W. A., Davies, M. J., Dean, R. T., and Fu, S. (1998) 3-Hydroxylysine, a potential marker for studying radicalinduced protein oxidation. Chem. Res. Toxicol. 11, 1265-1273. (37) Dizdaroglu, M., and Gajewski, E. (1989) Structure and mechanism of hydroxyl radical-induced formation of a DNA-protein crosslink involving thymine and lysine in nucleohistone. Cancer Res. 49, 3463-3467. (38) Mandel, R., Kolomijtseva, G., and Brahms, J. G. (1979) DNAprotein interactions in nucleosomes and in chromatin. Structural studies of chromatin stabilized by ultraviolet-light induced crosslinking. Eur. J. Biochem. 96, 257-265. (39) Saito, I., Sugiyama, H., and Matsuura, T. (1982) Mechanism of DNA-histone crosslinks in UV-irradiated nuclei. Nucleic Acids Symp. Ser. 11, 225-228. (40) Cadet, J., D’Ham, C., Douki, T., Pouget, J. P., Ravanat, J. L., and Sauvaigo, S. (1998) Facts and artifacts in the measurement of oxidative base damage to DNA. Free Radical Res. 29, 541-550. (41) Augusto, O., Netto, L. E., and Gomes, L. F. (1992) DNA alkylation by carbon-centered radicals. Braz. J. Med. Biol. Res. 25, 11711183. (42) Augusto, O. (1993) Alkylation and cleavage of DNA by carboncentered radical metabolites. Free Radical Biol. Med. 15, 329336. (43) Harkin, L. A., Butler, L. M., and Burcham, P. C. (1997) Role of G f T transversions in the mutagenicity of alkylperoxyl radicals: induction of alkali-labile sites in bacteriophage M13mp19. Chem. Res. Toxicol. 10, 575-581. (44) Valentine, M. R., Rodriguez, H., and Termini, J. (1998) Mutagenesis by peroxyl radical is dominated by transversions at deoxyguanosine: evidence for the lack of involvement of 8-oxo-dG and/ or abasic site formation. Biochemistry 37, 7030-7038. (45) Kawanishi, S., Oikawa, S., Murata, M., Tsukitome, H., and Saito, I. (1999) Site-specific oxidation at GG and GGG sequences in double-stranded DNA by benzoyl peroxide as a tumour promoter. Biochemistry 38, 16733-16739. (46) Steenken, S., and Jovanovic, S. V. (1997) How easily oxidizable is DNA: one-electron reduction potentials of adenosine and guanosine radicals in aqueous solution. J. Am. Chem. Soc. 119, 617-618. (47) Luger, K., Mader, A. W., Richmond, R. K., Sargent, D. F., and Richmond, T. J. (1997) Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature 389, 251-260. (48) Symons, M. C. R. (1997) Electron movement through proteins and DNA. Free Radical Biol. Med. 22, 1271-1276. (49) Lebkowski, J. S., and Laemmli, U. K. (1982) Non-histone proteins and long-range organization of HeLa interphase DNA. J. Mol. Biol. 156, 325-344. (50) Chiu, S. M., Xue, L. Y., Friedman, L. R., and Oleinick, N. L. (1993) Copper ion-mediated sensitization of nuclear matrix attachment sites to ionizing radiation. Biochemistry 32, 6214-6219. (51) Chiu, S., Xue, L., Friedman, L. R., and Oleinick, N. L. (1995) Differential dependence on chromatin structure for copper and iron ion induction of DNA double-strand breaks. Biochemistry 34, 2653-2661.

TX000053U

Radicals Derived from Histone Hydroperoxides Damage Nucleobases

Recommend Documents