Oxidative Damage to a Supercoiled DNA by Water Soluble Peroxyl

Oxidative Damage to a Supercoiled DNA by Water. Soluble Peroxyl Radicals Characterized with DNA. Repair Enzymes. Cristina Sanchez, R. Adam Shane, ...
1 downloads 0 Views 139KB Size
Download PDF
1118

Chem. Res. Toxicol. 2003, 16, 1118-1123

Oxidative Damage to a Supercoiled DNA by Water Soluble Peroxyl Radicals Characterized with DNA Repair Enzymes Cristina Sanchez, R. Adam Shane, Thomas Paul, and Keith U. Ingold* National Research Council, Ottawa, Ontario, Canada K1A 0R6 Received May 12, 2003

Earlier work (Paul, T., et al. (2000) Biochemistry 39, 4129-4135) has demonstrated that the water soluble positively charged peroxyl radical, (H2N)2+CC(CH3)2OO• (+AOO•), caused direct strand scission of the Escherichia coli plasmid supercoiled DNA, pBR 322, with ca. 50% scission occurring at a +AOO•/base pair (bp) ratio of 0.2. There was no measurable direct scission with a negatively charged peroxyl (-BOO•) at -BOO•/bp ) 24, nor with a neutral peroxyl (COO•) at COO•/bp ) 5. Base modification (BM) of the same DNA by the same peroxyls has now been investigated using four base excision repair (BER) glycosylases. At +AOO•/bp ) 0.04, there is 10% direct strand scission, and the Fpg protein recognized an additional 25% BM, while endonuclease (Endo) IV recognized an additional 20% BM and the other two BER enzymes did not give statistically significant BMs. None of the BER enzymes showed BMs in the DNA treated with -BOO•. However, Fpg and Endo IV showed that at COO•/bp ) 3.4 there was a BM comparable to that observed at +AOO•/bp ) 0.04. Thus, COO• radicals are only ca. 1.2% as reactive toward the DNA’s bases as +AOO•. These results underline the importance of Coulombic forces in DNA reactions. It is also proposed that +AOO• has a higher intrinsic reactivity in H-atom abstractions and electron transfer processes than -BOO• or COO• radicals.

For example, the positively charged alkylperoxyl radical (H2N)2+CC(CH3)2OO• (+AOO•) caused ca. 50% of the plasmid SC DNA derived from Escherichia coli pBR 322 (2.9 × 106 Da, 4361 bps) to suffer a single strand break

to afford R DNA at a +AOO/bp ratio of 0.2 (1). Strand cleavage by a different positively charged alkylperoxyl occurred with a similar efficiency (1). In sharp contrast, the negatively charged alkylperoxyl, -O3SCH2CH2C(CH3)(CN)OO• (-BOO•), generally produced no detectable strand scission, (i.e., no detectable decrease in SC DNA and concomitant increase in R DNA), at peroxyl radical/ bp ratios 100-fold greater, viz., -BOO•/bp ) 24:1 (1). Furthermore, the uncharged alkylperoxyl, COO•, generally produced no detectable strand scission at the highest readily achievable COO•/bp ratio of 5:1 (1). A subsequent claim (3) that SC pBR 322 does undergo strand cleavage when subjected to attack by uncharged alkylperoxyl radicals (purportedly generated from the reactions of various alkyl hydroperoxides with various peroxidases) has been shown to be in error for at least one specific hydroperoxide/peroxidase couple (4) and has been corrected (5). Nevertheless, claims for direct DNA strand cleavage by neutral peroxyl radicals continue to be made (6).

* To whom correspondence should be addressed. Tel: 1-613-9900938. Fax: 1-613-941-8447. E-mail: [email protected]. 1 Abbreviations: SC, supercoiled; R, relaxed; L, linear; +AOO• (from 2,2′-azobis(amidinopropane), ABAP, or +ANNA+), (H2N)+2CC(CH3)2OO•; -BOO• (from 3,3′-azobis(3-cyano-1-butanesulfonic acid, -BNNB-), -O SC H C(CH )(CN)OO•; COO• (from azobis[2-methyl-N-(2-hydroxy3 2 4 3 ethyl)propionamide], CNNC), HOCH2CH2NHC(O)C(CH3)2OO•; BM, base modification; bp, base pair; BER, base excision repair; Endo, endonuclease; Fpg, Fpg protein; Exo, exonuclease; AP, apurinic/ apyrimidinic. 2 In this equation, e is the efficiency of escape of the geminate pair of R• radicals from the solvent cage in which they are produced and k1 is the rate constant for decomposition of the azo compound at 37 °C. For the positively charged, negatively charged, and neutral peroxyl radicals employed in the present study, the values of e and 106k1 s-1 are, respectively (2): +AOO•, 0.5, 1.3; -BOO•, 0.5, 0.35; COO•, 0.1, 0.18. Note that the e value for COO• was measured in ref 2 and found to be only one-fifth of the value of 0.5 assumed in ref 1.

The uniquely strong DNA strand-cleaving abilities of the positively charged alkylperoxyl radicals have been attributed to their Coulombic attraction by the negatively charged DNA polyanion (1). However, the biological relevance of DNA strand cleavage by positively charged alkylperoxyls is doubtful. Small, and hence free to diffuse, positively charged peroxyls are less likely to be formed in vivo than negatively charged peroxyls (such as those formed during the peroxidation of free polyunsaturated fatty acids, RCO-2 f -O2CR-HOO•) and uncharged peroxyls (such as those formed during the peroxidation of polyunsaturated fatty acid esters and superoxide’s conjugate acid, HOO•). Thus, the peroxyl radicals most likely to be produced in vivo do not directly cleave double-

Introduction We have recently demonstrated that direct single strand cleavage of (double-stranded) SC1 plasmid DNAs by water soluble alkylperoxyl radicals at 37 °C is relatively facile when the peroxyl carries a positive charge but is generally below the level of detection when the peroxyl is without charge or carries a negative charge (1). This work was quantitative in that the alkylperoxyl radicals were generated at known rates and for defined periods of time by the thermal decomposition of azo compounds in aerated Tris buffer (pH 7.4, eq 1)2 containing known quantities of a SC DNA. k1

RNtNR 9 8 2e ROO• + N2 H O, 37 °C, O 2

2

(1)

10.1021/tx030024u CCC: $25.00 Published 2003 by the American Chemical Society Published on Web 08/15/2003

Oxidative Damage to a Supercoiled DNA

stranded DNA. However, this does not necessarily mean that they do not react with and damage DNA by oxidatively modifying the DNA’s bases. Indeed, this has been shown to be the case for the reaction of +AOO• radicals with a double-stranded plasmid DNA, pSP 189 (7) and with human fibroblast (hf) DNA (8, 9). In the most thorough studies of any +AOO•/DNA system, Termini and co-workers (8, 9) measured both direct strand scission of hf DNA by +AOO• and the concomitant BMs. The BMs were determined by measuring the additional strand scission that occurred following treatment of the oxidized hf DNA with a mixture of two BER glycosylases, the Fpg and Nth proteins (recent reviews, 10 and 11). The BER glycosylases are sensitive probes for oxidatively modified DNA bases (see, e.g., 7-9 and 12-14). Although the specificity of BER glycosylases is rather broad and has not yet been completely defined for any member of this family (9-11, 14, 15), it is well-established that Fpg excises a broad spectrum of oxidatively modified purines and the Nth protein excises a broad spectrum of oxidatively modified pyrimidines. Termini and co-workers (8, 9) used a mixture of these two enzymes to ensure that the widest possible extent of DNA base damage was detected and discovered that BM occurred on average 4.7 ( 1.2 times more frequently than strand scission (9). With this background, we decided to carry out a preliminary and strictly limited investigation as to whether negatively charged and uncharged alkylperoxyls could induce BER glycosylase detectable BMs in DNA. Because +AOO• radicals have been demonstrated to produce BMs (7-9), as a control, we subjected SC pBR 322 to attack by these radicals using a very low ratio of + AOO•/bp ) 0.04, i.e., only ca. one-fifth of the quantity of +AOO• radicals that produce a 50% yield of the R DNA single strand break product (1). The peroxyl radical source, +ANNA+, was removed, and the oxidized DNA was then incubated with each of four different BER glycosylases, and the “additional” strand scission (above that due to the +AOO• alone and to the glycosylase alone) was determined. Despite the rather broad specificity of BER glycosylases, we chose to use individual glycosylases in the hope of obtaining a somewhat better “footprint” of base damage by peroxyl radicals than would be obtained with a mixture. Because even high concentrations of -BOO• and COO• give no detectable direct strand cleavage of SC pBR 322, these two radicals were employed at peroxyl/bp ratios that were roughly 100 times greater than that used with + AOO•. Removal of the parent azo compounds and incubation with the BER glycosylases revealed that COO• produced extensive BMs while -BOO• produced no more than minor damage to the DNA. These results are fully consistent with the electrostatic arguments previously advanced (1) to explain the uniquely strong DNA-cleaving activity of positively charged alkylperoxyl radicals.

Experimental Section Materials. All materials were available from previous studies (1, 4) except for the four BER glycosylases, which were purchased from Trevigen (MD). These BER glycosylases were E. coli Endo III (activity 25 U/µg), E. coli Fpg (activity 50 U/µg), E. coli Endo IV (activity 10 U/µg), and E. coli Exo III (activity 151 U/µg). Incubation of pBR 322 with Peroxyl Radicals Followed by Incubation with a BER Glycosylase. The experimental procedure for treatment of the DNA with a known total

Chem. Res. Toxicol., Vol. 16, No. 9, 2003 1119

Figure 1. Agarose gel (1.5%) electrophoresis of SC pBR 322. Even numbered lanes: 1.92 µg of DNA in 90 µL of buffer was incubated for 1 h at 37 °C with 2.76 × 10-4 M +ANNA+. The DNA was then separated from the azo compound and subjected to a second incubation for 1 h at 37 °C with various concentrations of Fpg. The odd numbered control lanes were incubated for 1 h at 37 °C with no +ANNA+ and then a further hour with Fpg. The total volume in each well was ca. 20 µL, and the approximate amount of DNA in each lane was 320 ng. Band areas given below are in arbitrary units. No attempt was made to quantify the L DNA. (a) Of the total SC + R DNA. concentration of peroxyl radicals formed by the thermal decomposition of a water soluble azo compound has been described previously (1). The reactions were carried out at 37 °C in oxygensaturated, Chelex 100-treated Tris buffer (pH 7.4), with 140 mM KCl and 30 mM MgCl2 added immediately before the experiment. The duration of the incubation of 1.92 µg of DNA in 90 µL of buffer (equivalent to the 320 ng of DNA in 15 µL of buffer employed in earlier work (1)) with the peroxyl radical-generating azo compounds was 1 h with +ANNA+ and -BNNB- but was 24 h with the thermally more stable (1) and much less water soluble CNNC (which also generates radicals with a much lower efficiency than the two charged azo compounds2 (2)). If the (potentially) oxidized DNA was to be treated with a BER glycosylase, it was separated from the azo compound using a QIAprep Spin Miniprep Kit (Qiagen, Germany) prior to incubation with the enzyme for 1 h at 37 °C. Samples of the completely untreated (i.e., “fresh”) DNA, the DNA subjected only to peroxyl radical attack, fresh DNA incubated for 1 h at 37 °C with the BER glycosylase, and DNA subjected to peroxyl radical attack and then treated with the glycosylase were analyzed by gel electrophoresis (1, 4). The DNA bands were detected by ethidium bromide excitation/fluorescence with a UVP UV transilluminator and photographed, as described previously (1, 4).

Results A photograph of one of our many gels showing the results of incubating fresh pBR 322 DNA and DNA treated with positively charged peroxyl radicals (at + AOO•/bp ratio ) 0.04) with various concentrations of Fpg is shown in Figure 1. In each lane, the lower band is due to SC DNA, the upper band is due to R DNA, and between is a much weaker band (which does not always appear) due to L DNA.3 This experiment was repeated two more times. Triplicate experiments were also carried

1120 Chem. Res. Toxicol., Vol. 16, No. 9, 2003

Sanchez et al.

Figure 2. Percentage of SC pBR 322 DNA left after treatment with different amounts of BER glycosylases for 1 h at 37 °C. DNA was untreated (0) or incubated with +ANNA+ (0.27 mM) for 1 h at 37 °C (9). The BER glycosylase/AP lyase enzymes used were Exo III (A), Fpg (B), Endo III (C), and Endo IV (D). The error bars indicate the SD of at least three independent measurements.

out using the same +AOO•/bp ratio of 0.04 and various concentrations of the other three BER glycosylases. The results of these four sets of triplicate experiments are shown graphically in Figure 2. It is clear that +AOO• radicals cause some direct strand scission of SC pBR 322 (as previously reported (1)) even at a +AOO•/bp ratio as low as 0.04. As expected (7-9), these radicals produce extensive base damage. We therefore turned to the neutral (COO•) and negatively charged (-BOO•) radicals, neither of which cause measurable direct strand scission of SC pBR 322 at readily attainable concentrations (1). We chose to use radical/bp ratios of 3.4 (COO•) and 3.9 (-BOO•) and the four BER glycosylases at the highest concentrations that had been employed in the described +AOO• work. The same experimental procedure was employed, and the photographs of typical gels are shown in Figures 3 and 4. The results of triplicate experiments with all three radicals at the highest BER glycosylase concentrations are presented in bar graph form in Figure 5.

Discussion Treatment of pBR 322 with +AOO• at a radical/bp ratio of 0.2 produced ca. a 50% decrease in SC DNA (1). Treatment with an +AOO•/bp ratio of 0.04, as in the 3 A single strand scission event converts SC DNA into R DNA (1, 4, 16, 17). R DNA is always present (in variable amounts) in SC pBR 322 (1, 4). A second strand scission event converts R DNA into L DNA provided this event occurs on the uncut strand and probably within about five bp of the break in the first strand and, for random strand scission, a 50% conversion of R pBR 322 into L DNA requires 28 separate events (1). Thus, L DNA never becomes a major product because ring opening of R DNA will generally yield only small DNA fragments.

Figure 3. Agarose gel (1.5%) electrophoresis of SC pBR 322. Even numbered lanes: 1.92 µg of DNA in 90 µL of buffer was incubated for 25 h at 37 °C with 3.5 × 10-2 M CNNC. The DNA was then separated from the azo compound and subjected to a second incubation for 1 h at 37 °C with a high concentration of a BER glycosylase/AP lyase enzyme. The odd numbered control lanes were incubated for 24 h at 37 °C with no CNNC and then for 1 h with the BER glycosylase. The total volume in each well was ca. 20 µL, and the approximate amount of DNA was 320 ng. Band areas given below are in arbitrary units. No attempt was made to quantify the L DNA. (a) Of the total SC + R DNA.

present study, would therefore be expected to produce ca. a 10% decrease in the SC DNA, and such would

Oxidative Damage to a Supercoiled DNA

Figure 4. Agarose gel (1.5%) electrophoresis of SC pBR 322. Even numbered lanes: 1.92 µg of DNA in 90 µL of buffer was incubated for 1 h at 37 °C with 0.10 M -BNNB-. The DNA was then separated from the azo compound and subjected to a second incubation for 1 h at 37 °C with a high concentration of a BER glycosylase. The odd numbered control lanes were incubated for 1 h at 37 °C with the BER glycosylase. The total volume in each well was ca. 20 µL, and the approximate amount of DNA was 320 ng. Band areas given below are in arbitrary units. No attempt was made to quantify the L DNA. (a) Of the total SC + R DNA.

appear to be the case, see Figure 1 where SC constitutes 61% of the fresh DNA (lane 1), but constitutes only 51% of the +AOO•-treated DNA (lane 2). Treatment of pBR 322 (used as received) with each of the four BER glycosylases reveals that even this fresh DNA contains modified bases (see Figure 2). This result is unsurprising in view of the relatively large content of R DNA (ca. 30-50%) in fresh pBR 322. Of the four glycosylases examined, the most active toward fresh DNA (by a small margin) would generally appear to be Fpg (see Figure 2B), which is recognized to be an important enzyme in the prevention of mutations resulting from oxidative stress (18-21). Fpg is best known for its ability to excise 8-oxo-deoxyguanosine (one of the most common oxidative lesions in DNA (21, 22)) and 2,6-diamino-4hydroxy-5-(N-methyl)formamidopyrimidine. Fpg also excises a broad spectrum of modified purines, many of which have imidazole ring-opened structures (23-27), and it would appear to have a small activity toward modified pyrimidines (28). Fpg is also endowed with an AP lyase activity that incises DNA at abasic sites by a β-δ elimination mechanism causing the observed strand scission (29). The other three glycosylases employed generally induce rather similar amounts of destruction of fresh SC pBR 322 (see Figures 2 and 5). Much larger differences between the effects of the four glycosylases were found with pBR 322, which had been subjected to reaction with +AOO• at a radical/bp ratio of 0.04 (see Figures 2 and 5+A). Fpg still induces the largest decrease in SC DNA, followed by Endo IV, an enzyme known to be induced by oxidative stress (30). Endo IV shows a very similar AP lyase activity toward abasic sites as Exo III (17, 31), which was the least active BER glycosylase toward the +AOO•-damaged DNA. There is,

Chem. Res. Toxicol., Vol. 16, No. 9, 2003 1121

Figure 5. Percentage of untreated (white bars) and peroxyl radical-treated (black bars) SC pBR 322 DNA remaining after treatment with a BER glycosylase/AP lyase enzyme for 1 h at 37 °C. The enzyme concentrations were used as follows: 600 U/mL Exo III, 600 U/mL Fpg, 600 U/mL Endo III, and 200 U/mL Endo IV. SC pBR 322 DNA (1.92 µg/90 µL) was incubated with 0.27 mM +ANNA+ for 1 h at 37 °C (panel +A), with 100 mM -BNNB- for 1 h at 37 °C (panel -B) and with 10 mM CNNC for 24 h at 37 °C (panel C). Error bars are SD of three independent measurements.

however, one important difference between these two enzymes: Endo IV has the novel ability to hydrolyze the phosphodiester bond 5′ to an R-deoxyadenosine residue (32), which, at first sight, suggests that R-deoxyadenosine might be a significant product of DNA damage by +AOO• radicals (and also by COO• radicals; see below and Figure 5). However, the oxidation of adenine in pBR 322 by + AOO• radicals is unlikely because in the hf DNA/AOO• reaction the order of reactivity of the bases was G . C > T with adenine essentially unreactive (9). It would appear that the difference between Endo IV and Exo III may be more than the ability of the former to cleave DNA next to an R-deoxyadenosine residue. This would be consistent with the fact that Exo III, unlike Endo IV, does not produce a statistically significant greater loss of SC DNA from pBR 322 treated with 0.04 +AOO•/bp than from fresh pBR 322 (see Figures 2A and 5+A). The same is true for pBR 322 treated with COO• radicals (Figure 5C). Only in the case of the -BOO•-treated DNA do neither the Exo III nor the Endo IV produce a statistically significant greater loss of SC DNA than for the untreated DNA (Figure 5-B). The extent of peroxyl radical-induced BM of SC pBR 322 recognized by each of the BER glycosylases can be approximately estimated from the differences in the percentage of SC DNA remaining in fresh DNA and in the radical-treated DNA. This method for estimating peroxyl radical-induced damage to the bases can be justified by the fact that the summed total of SC + R

1122 Chem. Res. Toxicol., Vol. 16, No. 9, 2003

DNA in each lane is generally fairly constant (i.e., generally within ca. (10-15% of the mean value for each gel; see tables accompanying Figures 1, 3, and 4). This indicates that more complete degradation of the R DNA to L DNA and small fragments is relatively unimportant. As an example of this process, treatment of fresh DNA with Fpg left 70% of unreacted SC DNA while treatment with 0.04 +AOO•/bp and then Fpg left 35% of unreacted SC DNA (Figure 5+A). Thus, 70 - 35 ) 35% of the SC DNA suffered +AOO•-induced direct strand breaks and BM recognized by Fpg. At this +AOO•/bp ratio of 0.04, the +AOO• radicals cause only a 10% loss of SC DNA by direct strand scission (vide supra). Therefore, and as Termini and co-workers (9) have previously noted, BM dominates the +AOO• radical-induced DNA damage spectrum. For the SC pBR 322/+AOO• system, the Fpgidentified BM/strand break ratio is ((70 - 35%) - 10%)/ 10% ) 2.5. Endo IV was the only other enzyme examined in the present work that gave a statistically significant difference in the percentage of SC pBR 322 remaining between fresh- and +AOO•-treated material (see Figures 2D and 5+A). From Figure 5+A, the Endo IV-identified BM/strand break ratio ) ((75 - 45%) - 10%)/10% ) 2.0. These results using pBR 322 and two individual BER glycosylases are congruent with Termini and co-workers’ (9) BM/strand break ratio of 4.7 ( 1.2 using hf DNA and two combined glycosylases (see Introduction). By comparing Figures 3 and 4 or Figure 5C,-B. it can be seen that base damage induced by neutral COO• radicals (at a COO•/bp ratio of 3.4) is much greater than base damage induced by the negatively charged -BOO• radicals (at a -BOO•/bp ratio of 3.9). Indeed, for the negatively charged radicals, none of the four BER glycosylases gave statistically significant differences in the loss of SC DNA as compared with the DNA treated only with the glycosylases (Figure 5-B). However, the neutral peroxyls did induce base damage that was recognized by both Fpg and Endo IV, the same two BER glycosylases that recognized base damage by the positively charged peroxyls (cf. Figure 5+A,C). Comparable extents of base damage were induced using +AOO•/bp ) 0.04 and COO•/ bp ) 3.4, which indicates that the neutral peroxyl radicals are ca. 1.2% as reactive toward the bases in double-stranded SC DNA as the positively charged peroxyls. Because direct strand scission of SC pBR322 was not observed at a COO•/bp ratio of 5:1 (1),2 it seems probable that COO• radicals have even less than 1% of the ability of +AOO• radicals to induce strand scission. Our preliminary survey of base damage to doublestranded SC pBR 322 DNA by charged and neutral water soluble alkylperoxyl radicals has shown that (i) positively charged +AOO• radicals are much more active than neutral and negatively charged peroxyl radicals. (ii) BM dominates the +AOO• damage spectrum; the Fpg- and Endo IV-detected BM/strand break ratios are 2.5 and 2.0, respectively. (iii) Neutral COO• radicals cause base damage at a COO•/bp ratio of 3.4 comparable to that produced by +AOO• at a +AOO•/bp ratio of 0.04. There is no measurable direct strand scission by COO• at a COO•/ bp ratio of 5:1 (1).2 (iv) Negatively charged -BOO• radicals are ineffective at inducing BMs. It seems highly probable that these results primarily reflect electrostatic forces between the DNA polyanion and the peroxyl radicals (and their precursor azo compounds). For the positively charged peroxyl radicals, there may also be a second factor that makes them

Sanchez et al.

particularly reactive. It is well-known that peroxyl radicals with neutral electron-withdrawing substituents (such as trichloromethylperoxyl, Cl3COO•) have a higher unpaired spin density on the terminal oxygen atom (33) and an enhanced reactivity, as compared with alkylperoxyls such as tert-butylperoxyl or methylperoxyl, both in H-atom abstractions, e.g., k(Cl3COO• + c-C6H12) ) 1 × 103 M-1 s-1 (34) whereas k(Me3COO• + c-C6H12) ) 3 × 10-3 M-1 s-1 (35), and in electron transfer reactions (i.e., electron abstraction), e.g., k(Cl3COO• + ascorbate) ) 9 × 108 M-1 s-1 but k(H3COO• + ascorbate) ) 2 × 106 M-1 s-1 (36). The positive charge on +AOO• peroxyls is also expected to enhance their reactivity relative to COO•, but this factor has not, to our knowledge, been considered previously in any biorelated work with +ANNA+. A logical extension of these ideas implies that the negatively charged -BOO• radicals may be less reactive than neutral COO• radicals. To explore the reactivities of +AOO•, BOO•, and COO•, we plan to measure the absolute rate constants for reactions of these three radicals with a variety of substrates.

References (1) Paul, T., Young, M. J., Hill, I. E., and Ingold, K. U. (2000) Strand cleavage of supercoiled DNA by water soluble peroxyl radicals. The overlooked importance of peroxyl radical charge. Biochemistry 39, 4129-4135. (2) Bedard, L., Young, M. J., Hall, D., Paul, T., and Ingold, K. U. (2001) Quantitative studies on the peroxidation of human lowdensity lipoprotein initiated by superoxide and by charged and neutral alkylperoxyl radicals. J. Am. Chem. Soc. 123, 1243912448. (3) Adam, W., Kurz, A., and Saha-Mo¨ller, C. R. (2000) Peroxidasecatalyzed oxidative damage of DNA and 2′-deoxyguanosine by model compounds of lipid hydroperoxides: involvement of peroxyl radicals. Chem. Res. Toxicol. 13, 1199-1207. (4) Shane, R. A., and Ingold, K. U. (2002) Cleavage of supercoiled DNA by horseradish peroxidase plus tert-butyl hydroperoxide is not due to tert-butylperoxyl radicals. Chem. Res. Toxicol. 15, 1324-1329. (5) Adam, W., Kurz, A., and Saha-Mo¨ller, L. R. (2002) Peroxidasecatalyzed oxidative damage of DNA and 2′-deoxyguanosine by model compounds of lipid hydroperoxides: involvement of peroxyl radicals. Chem. Res. Toxicol. 15, 1330. (6) Adam, W., Hartung, J., Okamoto, H., Marquardt, S., Nau, W. M., Pischel, U., Saha-Mo¨ller, C. R., and Spehar, K. (2002) Photochemistry of N-isopropoxy-substituted 2(1H)-pyridone and 4-ptolylthiazole-2(3H)-thione: alkoxyl radical release (spin-trapping, EPR, and transient spectroscopy) and its significance in the photooxidative induction of DNA strand breaks. J. Org. Chem. 67, 6041-6049. (7) Harkin, L. A., and Burcham, P. C. (1997) Formation of novel C1oxidized abasic sites in alkylperoxyl radical-damaged plasmid DNA. Biochem. Biophys. Res. Commun. 237, 1-5. (8) 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. (9) Rodriguez, H., Valentine, M. R., Holmquist, G. P., Akman, S. A., and Termini, J. (1999) Mapping of peroxyl radical induced damage to genomic DNA. Biochemistry 38, 16578-16588. (10) David, S. S., and Williams, S. D. (1998) Chemistry of glycosylases and endonucleases involved in base-excision repair. Chem. Rev. 98, 1221-1261. (11) Mol, C. D., Parikh, S. S., Putnam, C. D., Lo, T. P., and Tainer, J. A. (1999) DNA repair mechanisms for the recognition and removal of damaged DNA bases. Annu. Rev. Biophys. Biomol. Struct. 28, 101-128. (12) Epe, B., Ballmaier, D., Roussyn, I., Briviba, K., and Sies, H. (1996) DNA damage by peroxynitrite characterized with DNA repair enzymes. Nucleic Acids Res. 24, 4105-4110. (13) Laval, J., Jurado, J., Saparbaev, M., and Sidorkina, O. (1998) Antimutagenic role of base-excision repair enzymes upon free radical-induced DNA damage. Mutat. Res. 402, 93-102.

Oxidative Damage to a Supercoiled DNA (14) Cadet, J., Bourdat, A.-G., D’Ham, C., Duarte, V., Gasparutto, D., Romieu, A., and Ravanat, J.-L. (2000) Oxidative base damage to DNA: specificity of base excision repair enzymes. Mutat. Res. 462, 121-128. (15) Wallace, S. S. (1994) DNA damages processed by base excision repair: biological consequences. Int. J. Radiat. Biol. 66, 579589. (16) Inouye, S. (1984) Site-specific cleavage of double-strand DNA by hydroperoxide of linoleic acid. FEBS Lett. 172, 231-234. (17) Ha¨ring, M., Ru¨diger, M., Demple, B., Boiteux, S., and Epe, B. (1994) Recognition of oxidized abasic sites by repair endonucleases. Nucleic Acids Res. 22, 2010-2015. (18) Tchou, J., Kasai, H., Shibutani, S., Chung, M.-H., Laval, J., Grollman, A. P., and Nishimura, S. (1991) 8-Oxoguanine (8hydroxyguanine) DNA glycosylase and its substrate specificity. Proc. Natl. Acad. Sci. U.S.A. 88, 4690-4694. (19) Michaels, M. L., and Miller, J. H. (1992) The GO system protects organisms from the mutagenic effect of the spontaneous lesion 8-hydroxyguanine (7,8-dihydro-8-oxoguanine). J. Bacteriol. 174, 6321-6325. (20) Michaels, M. L., Tchou, J., Grollman, A. P., and Miller, J. H. (1992) A repair system for 8-oxo-7,8-dihydrodeoxyguanine. Biochemistry 31, 10964-10968. (21) Boiteux, S. (1997) Properties and biological functions of the Nth and Fpg proteins of Escherichia coli: two DNA glycosylases that repair oxidative damage in DNA. J. Photochem. Photobiol. 19, 87-96. (22) Chung, M. H., Kasai, H., Jones, D. S., Inoue, H., Ishikawa, H., Ohtsuka, E., and Nishimura, S. (1991). An endonuclease activity of Escherichia coli that specifically removes 8-hydroxyguanine residues from DNA. Mutat. Res. 254, 1-12. (23) Graves, R. J., Felzenszwalb, I., Laval, J., and O’Connor, T. R. (1992) Excision of 5′-terminal deoxyribose phosphate from damaged DNA is catalyzed by the Fpg protein of Escherichia coli. J. Biol. Chem. 267, 14429-14435. (24) Boiteux, S., Gajewski, E., Laval. J., and Dizdaroglu, M. (1992) Substrate specificity of the Escherichia coli Fpg protein (formamidopyrimidine-DNA glycosylase): Excision of purine lesions in DNA produced by ionizing radiation or photosensitization. Biochemistry 31, 106-110. (25) Dizdaroglu, M., Zastawny, T. H., Carmical, J. R., and Lloyd, R. S. (1996) A novel DNA N-glycosylase activity of E. coli T4 endonuclease V that excises 4,6-diamino-5-formamidopyrimidine from DNA, a UV-radiation- and hydroxyl radical-induced product of adenine. Mutat. Res. 363, 1-8.

Chem. Res. Toxicol., Vol. 16, No. 9, 2003 1123 (26) Karakaya, A., Jaruga, P., Bohr, V. A., Grollman, A. P., and Dizdaroglu, M. (1997) Kinetics of excision of purine lesions from DNA by Escherichia coli Fpg protein. Nucleic Acids Res. 25, 474479. (27) Li, Q., Laval, J., and Ludhum, D. B. (1997) Fpg protein releases a ring-opened N-7 guanine adduct from DNA that has been modified by sulfur mustard. Carcinogenesis 18, 1035-1038. (28) Frosina, G., Fortini, P., Rossi, O., Carrozzino, F., Raspaglio, G., Cox, L. S., Lane, D. P., Abbondandolo, A., and Dogliotti, E. (1996) Two pathways for base excision repair in mamalian cells. J. Biol. Chem. 271, 9573-9578. (29) Boiteux, S., O’Connor, T. R., Lederer, F., Gouyette, A., and Laval, J. (1990) Homogeneous Escherichia coli Fpg protein: a DNA glycosylase which excises imidazole ring-opened purines and nicks DNA at apurinic/apyrimidinic sites. J. Biol. Chem. 265, 39163922. (30) Demple, B. (1996) Redox signaling and gene control in the Escherichia coli soxRS oxidative stress regulonsa review. Gene 179, 53-57. (31) Takeuchi, M., Lillis, R., Demple, B., and Takeshita, M. (1994) Interactions of Escherichia coli endonuclease IV and exonuclease III abasic sites in DNA. J. Biol. Chem. 269, 21907-21914. (32) Ide, H., Tedzuka, K., Shimzu, H., Kimura, Y., Purmal, A. A., Wallace, S. S., and Kow, Y. W. (1994) R-Deoxyadenosine, a major anoxic radiolysis product of adenine in DNA, is a substrate for Escherichia coli endonuclease IV. Biochemistry 33, 7842-7847. (33) Sevilla, M. D., Becker, D., and Yan, M. (1990) Structure and reactivity of peroxyl and sulphoxyl radicals from measurement of oxygen-17 hyperfine couplings: relationship with Taft substituent parameters. J. Chem. Soc., Faraday Trans. 86, 32793286. (34) Mosseri, S., Alfassi, Z. B., and Neta, P. (1987) Absolute rate constants for hydrogen abstraction from hydrocarbons by the trichloromethylperoxyl radical. Int. J. Chem. Kinet. 19, 309-317. (35) Howard, J. A. (1984) Measurement of absolute propagation and termination rate constants for alkylperoxyls in solution by the hydroperoxide method. Isr. J. Chem. 24, 33-37. (36) Alfassi, Z. B., Huie, R. E., and Neta, P. (1997) Kinetic studies of organic peroxyl radicals in aqueous solutions and mixed solvents. In Peroxyl Radicals (Alfassi, Z. B., Ed.) Chapter 9, Wiley, New York.

TX030024U