Mitochondrial endogenous oxidative damage ... - The FASEB Journal

Mitochondrial endogenous oxidative damage has been overestimated R. MICHAEL ANSON,1 EDGAR HUDSON,2 AND VILHELM A. BOHR3 Laboratory of Molecular Genetics, National Institute on Aging, Baltimore, Maryland 21224-6825, USA

The oxidatively induced DNA lesion 8-oxo-dG in mitochondrial DNA (mtDNA) is commonly used as a marker for oxidative damage to mitochondria, which in turn is thought to be a fundamental cause of aging. For years, mitochondrial levels of 8-oxo-dG were believed to be ⬃10-fold higher in mtDNA than in nuclear DNA even in normal, young animals. However, studies in our own and other laboratories have shown that this lesion is efficiently repaired. Also, mutational consequences specific to 8-oxo-dG (G to T transversions) are rarely reported. In the present study, we showed that the levels of damage measured using high-pressure liquid chromatography/electrochemical detection and an enzymatic/Southern blot assay were comparable. The latter assay does not require isolation of mitochondria, and so this assay was then used to determine the level of in vivo damage present in rat liver mtDNA both with and without organelle isolation. Levels of 8-oxo-dG are approximately threefold higher when measured in mtDNA purified from isolated mitochondria than when measured without prior mitochondrial isolation. Furthermore, most genomes were free of endogenous enzyme-sensitive sites (i.e., they did not contain 8-oxo-dG), and only after mitochondrial isolation were levels higher in mtDNA than in a nuclear sequence. Anson, R. M., Hudson, E., Bohr, V. A. Mitochondrial endogenous oxidative damage has been overestimated. FASEB J. 14, 355–360 (2000) ABSTRACT

Key Words: 8-hydroxy-2⬘-deoxyguanosine 䡠 8-hydroxydeoxyguanosine 䡠 8-oxoguanine 䡠 8-oxodG 䡠 aging 䡠 mitochondria

It is commonly believed that levels of the oxidatively induced DNA lesion 8-oxo-2⬘-deoxyguanosine (8-oxo-dG) are quite high in mitochondrial DNA (mtDNA). Many studies suggest that each genome contains several such lesions (1), and numbers as high as 163 per genome (4840 per 106 bases) have been reported (2). This damage was suggested to cause the eventual breakdown of proper mitochondrial function and to be a primary cause of aging (3, 4). However, the reported levels of 8-oxo-dG in mtDNA vary by five orders of magnitude, ranging 0892-6638/00/0014-0355/$02.25 © FASEB

from 0.08 per 106 bases in mtDNA isolated from cultured HeLa cells (5) to 4840 per 106 bases in mtDNA isolated from the hearts of 100-wk-old rats (2). Even within a single system, the values cover a wide range. Levels reported for mtDNA isolated from rat liver range from 4 per 106 bases (6) to 110 per 106 bases (2) (reviewed in refs 1, 7). At least two factors have contributed to this diversity. First, several methods are commonly used to measure the level of 8-oxo-dG in DNA, and very little information is available concerning their relative sensitivity. Second, in order to measure the level of damage in mtDNA it has been necessary to isolate the mitochondria. Mitochondrial isolation provides opportunity for the induction of oxidative damage. Also, since mitochondria are able to repair 8-oxo-dG (8 – 10) and a mitochondrial enzyme responsible for excision of 8-oxo-dG from double-stranded DNA (11) has been isolated, the possibility of post-isolation repair is an issue. Thus, measurement of oxidative damage to mtDNA after mitochondrial isolation may in part reflect the ability of the isolated mitochondria to resist or remove damage during the isolation procedure rather than the in vivo steadystate levels. To address these issues, we measured 8-oxo-dG in DNA using two methods: high-pressure liquid chromatography with electrochemical detection (HPLC/ ECD) (12, 13) and an enzymatic/Southern blot assay (14) using the bacterial repair enzyme formamidopyrimidine glycosylase (Fpg). The former method is the most widely used to detect endogenous 8-oxo-dG in DNA. The latter is used to measure the removal of exogenously induced 8-oxo-dG from DNA in studies of DNA repair. It has two major advantages over chemical methods such as HPLC/ECD: only a few micrograms of total DNA are required, as opposed to 50 or more micrograms of purified mtDNA; with the 1 Current affiliation: Laboratory of Cellular and Molecular Biology, NIA, NIH; Baltimore, Maryland 21224-6825, USA. 2 Current affiliation: IGEN International, Gaithersburg, MD 20877, USA. 3 Correspondence: Laboratory of Molecular Genetics, National Institute on Aging, GRC, NIH, 5600 Nathan Shock Dr., Baltimore, MD 21224-6825, USA. E-mail: [email protected]

355

enzymatic/Southern blot method, isolation of mitochondria and mtDNA is not required. However, its sensitivity relative to HPLC/ECD had not been established. In the present paper we used photoactivated methylene blue (MB) to create a series of DNA standards that contain linearly increasing amounts of 8-oxodG. This technique permits determination of the sensitivity and precision of a measurement system across a range of DNA damage levels. These standards were used to compare the HPLC/ECD and enzymatic/Southern blot methods for measurement of 8-oxo-dG in DNA. The results were linear and there was close agreement between the two methods. This showed that both measurement systems provide precise determinations of the level of damage, and the agreement between the two methods argues that the actual levels detected are accurate. We then applied the enzymatic/Southern blot method to the measurement of oxidative damage in mtDNA both with and without isolation of mitochondria. The level of damage seen in the two samples, from the same animals, was threefold higher in mtDNA when the mitochondria had been subjected to isolation. In the absence of organelle isolation, most genomes did not contain enzyme-sensitive sites (that is, they were lesion free). The implications of this finding, as well as the advantages and disadvantages of each method and questions that can best be addressed by the application of both methods simultaneously, will be discussed.

MATERIALS AND METHODS Cells and animals The rapidly growing Burkitt’s lymphoma cell line BL2 (15) was used as a DNA source for in vitro work. Six- to 7-month-old outbred male Wistar rats (Gerontology Research Center Colony) were used in this study. They were maintained on a 12/12 L-D cycle with lights on at 07:00 h and food (NIH Open Formula 079) and water available ad libitum. Animals were killed by cervical dislocation between 8:30 and 9:30 a.m.). Mitochondrial isolation Mannitol, EGTA, spermine, and spermidine were obtained from Sigma Chemical Company (St. Louis, Mo.). EDTA was from Quality Biological, Inc. (Gaithersburg, Md.). HEPES was obtained from Advanced Biotechnologies, Inc. (Columbia, Md.) and sucrose from ICN Biomedicals, Inc. (Aurora, Ohio). Livers were removed and placed immediately in ice-cold MSHE buffer (0.21 M mannitol, 0.07 M sucrose, 10 mM HEPES (pH 7.4), 1 mM EDTA, 1 mM EGTA, 0.15 mM spermine, 0.75 mM spermidine), where they were minced. The MSHE was replaced with fresh buffer and the livers were homogenized with a glass and Teflon homogenizer. The unbroken cells and nuclei were pelleted by centrifugation at 356

Vol. 14

February 2000

500 ⫻ g for 7 min. This pellet was resuspended in cold MSHE buffer and an aliquot was taken for isolation of nuclear DNA. The supernatant, which contained the mitochondria, was centrifuged again for 7 min at 9500 ⫻ g. The mitochondrial pellet was washed twice by resuspension in MSHE buffer, followed by centrifugation for 7 min at 9500 ⫻ g before the final resuspension in MSHE buffer. DNA isolation Total DNA for use in experiments involving in vitro damage was isolated from BL-2 human transformed lymphoblasts (15) by salt extraction (16). Rat liver total and mtDNA was isolated as follows: sodium dodecyl sulfate (SDS) (Quality Biological, Inc.) was added to the mitochondria or homogenate to 1% and the lysate incubated for 1 h at 37°C to promote complete lysis. After this treatment, 1/4 volume of 55°C saturated NaCl was added and the mixture was cooled to 4°C. The resulting protein and SDS precipitate was removed by centrifugation at 10,000 ⫻ g for 5 min and the supernatant was purified using freshly prepared phenol (Life Technologies, Gaithersburg, Md.), phenol/ chloroform, and chloroform (Mallinckrodt, Phillipsburg, N.J.). DNA was precipitated overnight and resuspended at ⬃500 ␮g/ml in TE, pH 8.0, prior to treatment with 100 ␮g/ml RNAse for 1 h at 37°C. The RNAase was removed by addition of SDS to 1%, followed by precipitation with NaCl as described above. The precipitate was removed by centrifugation at 10,000 ⫻ g for 5 min and the DNA in the supernatant was precipitated with 2 volumes of ethanol. Fpg digestion Reaction conditions for digestion with Fpg (which for this study was obtained from Dr. Arthur Grollman, State University of New York, Stony Brook, N.Y.) were as follows: concentrated reaction buffer was mixed with DNA dissolved in TE (final 1⫻ concentrations: 0.5 mg/ml bovine serum albumin, 50 mM Tris-Cl, pH 7.5, 50 mM KCl, 1 mM EDTA). The Fpg:DNA ratio was held between 0.01 to 0.1 ␮g Fpg:␮g DNA. All incubations were at 37°C. The Fpg/Southern blot assay has been described previously (10). In brief, a master mix of DNA was prepared in 1⫻ reaction buffer for each sample. One aliquot of this mix was then treated with Fpg and a second with 1⫻ reaction buffer. Between 0.5 ␮g and 2 ␮g of DNA was used for each reaction, using digestion conditions as described above. After Fpg treatment, the DNA was fully denatured by a 37°C incubation in 0.06 M NaOH for 15 min, then the entire reaction run on an alkaline gel and analyzed by Southern blot hybridization. The principle behind the assay is that Fpg generates a single-strand nick at the site of damage in a restricted DNA fragment. In a denaturing gel, the cleaved strand migrates faster than the full-length strand and thus moves forward, away from the main band. After blotting and probing, this loss of damaged DNA from the main band allows quantitation of the undamaged DNA. Southern blotting and probing Southern blotting was performed essentially as described previously (10). Membranes were hybridized to 32P-labeled probes generated as riboprobes (SP6/T7 Transcription Kit, Boehringer Mannheim Corporation, Indianapolis, Ind.) or by random primed labeling (Random Primed DNA Labeling Kit, Boehringer Mannheim). Two probes were used, one for the mitochondrial genome and one for nuclear DNA. The mitochondrial probe was directed against the heavy strand of

The FASEB Journal

ANSON ET AL.

the mitochondrial ribosomal sequence and the nuclear probe was directed against the nuclear ribosomal sequence. Both probes have been described previously (10). Radioactivity in the bands was quantitated using a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, Calif.) and ImageQuant software (Molecular Dynamics). Damaging DNA in vitro After treatment with Fpg and restriction with PvuII (Boehringer Mannheim), BL-2 DNA was diluted in 10 mM Tris-HCl, 1 mM EDTA, pH 8 (TE) to 150 ␮g/ml. One volume of 6 ␮M MB (Ricca Chemical Company, Arlington, Tex.: ␭max ⫽ 655 nm, ε⫽89125) was added to the DNA in the dark, mixed rapidly, and the mixture was aliquoted into 60 mm plastic tissue culture plates. The solution depth was ⬃2 mm. The dish was exposed to visible light [0.187 kW/m2, measured with an IL1400A Radiometer/Photometer (International Light, Inc., Newburyport, Mass.)] from a 300 W tungsten bulb situated beneath the plate. Heat buildup was prevented by use of a fan beneath the Plexiglas surface on which the dish rested. To remove MB, SDS was added to 1%, the solution was extracted three times with butanol, and the DNA was then ethanol precipitated, washed with 70% ethanol, and resuspended in TE. One-half of each sample was digested with Fpg at this stage. All samples were repurified with SDS and salt (16) and the DNA was stored as an ethanol precipitate until ready for analysis. All work was done under a dim blue light while MB was present in solution with DNA. HPLC/ECD Nucleosides were prepared enzymatically from DNA using nuclease P1 and alkaline phosphatase (Boehringer Mannheim Corporation), then filtered through a 0.22-micron filter and a 30 kDa cutoff spin column (12). Desferal (1 mM) was included to chelate iron and reduce the risk of artifactual oxidative damage. The nucleosides were separated isocratically using a C-8 column (YMC, Inc., Wilmington, N.C.) at a flow rate of 1 ml/min for 30 min to avoid spillover to the successive run. The mobile phase was composed of 100 mM sodium acetate (Sigma Chemical Company), pH 5.15, and 5% methanol (Fisher Scientific International Inc., Hampton, N.H.). Both 8-oxo-dG and dG were detected electrochemically using a four-channel coularray (ESA, Inc., Chelmsford, Mass.). Two channels were set at low potentials (285, 400 mV) for detection of 8-oxodG and two channels at higher potentials for detection of 2-deoxyguanosine (800 –900 mV). The peaks were identified in the samples according to the retention time and ratio accuracy. All samples were analyzed twice. A 2⬘deoxyguanosine standard was purchased from Sigma Chemical Company and an 8-oxo-dG standard from ESA, Inc. The concentrations of the standards were determined spectrophotometrically. For 8-oxo-dG, we used ε ⫽ 11300 at 295 nm in mobile phase and for 2⬘-deoxyguanosine, ε ⫽ 13000 at 254 nm. Data analysis Damage, as measured by HPLC/ECD, is normalized to unmodified dG. For conversion to standard units (lesions per 106 dN), 8-oxo-dG per 105 dG was multiplied by 2.2 based on 22% dG for mammalian DNA (for a list of nucleic acid compositions by species, see ref 17). To calculate the amount of damage in standard units using the Fpg/Southern blot assay, the number of incisions per fragment was calculated using the Poisson distribution: InciENDOGENOUS mtDNA OXIDATIVE DAMAGE

sions ⫽ -ln(Fpg treated/untreated). This number was divided by the number of dGs present in the fragment and multiplied by 2.2 ⫻ 105. For the ribosomal fragment, for which the exact sequence was not known, an approximate conversion was obtained by assuming that the fragment has an average dG content and multiplying the number of lesions in the fragment by 106/fragment length. (In actuality, the ribosomal region is known to be dG-rich, and so this procedure probably leads to a slight overestimation of the true level in the bulk DNA.) A single-factor ANOVA was used to determine whether the probability there was a difference between the levels of 8-oxo-dG measured in mtDNA from isolated mitochondria, mtDNA from the same animals measured without mitochondrial isolation, and nuclear ribosomal DNA from the same animals. When a significant difference was observed, a T test assuming unequal variance was used to determine the differences between pairs.

RESULTS Induction of 8-oxo-dG in DNA with MB was a linear function of light exposure (Fig. 1A, filled circles). Aliquots of DNA from each time point were treated with Fpg after damage, and residual 8-oxo-dG was measured by HPLC/ECD. Less than 10% of the induced 8-oxo-dG remained after Fpg treatment (Fig. 1A, open circles). This percentage was not dependent on the amount of damage present prior to Fpg digestion within the range measured, suggesting that 10% of the induced 8-oxo-dG may have been inaccessible to the Fpg enzyme. DNA damaged in vitro with MB was analyzed by the Fpg/Southern blot method. Both nuclear ribosomal and mitochondrial sequences were probed, and the number of Fpg-sensitive sites was determined using the Poisson equation. This number was converted to lesions/106 dN based on the approximate fragment length (for nuclear ribosomal DNA) or the number of guanines present in the sequence (for mtDNA). The results were compared with the results of an HPLC/ECD analysis of the same DNA (Fig. 1B). The methods gave virtually identical answers. This was true regardless of which sequence was probed (data not shown). In addition to 8-oxo-dG, photoactivation of MB also causes DNA strand breaks. This can be seen as a shift in the molecular weight of the DNA that has not been treated with Fpg, as viewed on a denaturing agarose gel stained with Sybrgreen (Molecular Probes) (data not shown), as well as a weakened signal in the lanes not treated with Fpg on the Southern blot (Fig. 1C). DNA from mitochondria isolated from rat liver or from the crude homogenate were subjected to Fpg/ Southern blot analysis (Fig. 2). The endogenous levels based on DNA from the crude homogenate were between 5 and 10, 8-oxo-dG per 106 dN for both nuclear and mtDNA, and were not significantly 357

in a double-stranded mitochondrial genome. Thus, on average, only 16 of every 100 mitochondrial genomes would contain the lesion.) The endogenous levels based on DNA from isolated mitochondria from the same animals were approximately threefold higher (P⬍0.01).

DISCUSSION To improve the confidence in the levels of 8-oxo-dG detected in a biological sample, it is helpful to compare several samples known to contain varying amounts of 8-oxo-dG in the range of interest and show that the detection system is responsive in that range. The MB light model is particularly useful in

Figure 1. A comparison of HPLC/ECD and the Fpg/Southern blot Assay for the detection of 8-oxo-dG in randomly damaged DNA. A) DNA was pretreated with Fpg to remove endogenous damage and exposed to photoactivated MB to induce randomly distributed damage. Induction of 8-oxo-dG in DNA was a linear function of light exposure (filled circles), and Fpg post-treatment removed 90% of the damage that had been induced (open circles). B) Quantitation of damage induced in vitro and detected by the Fpg/Southern blot method (open squares) and by HPLC/ECD (filled circles) in the same samples. The data shown represents damage based on probing with a mitochondrial probe and does not differ from data based on probing with a nuclear ribosomal probe (not shown). C) Representative data showing one of the Fpg/Southern blot assays on which the graph in panel B is based. (⫺) and (⫹) refer to Fpg digestion.

different (P⫽0.36). (This is equivalent to 0.16, 8-oxodG per 32 kdN, which is the number of nucleotides 358

Vol. 14

February 2000

Figure 2. 8-oxo-dG levels in a nuclear sequence and in mtDNA, with and without mitochondrial isolation DNA from mitochondria isolated from rat liver, or from the crude homogenate was subjected to Fpg/Southern blot analysis. Levels of damage detected for nuclear (A, using a ribosomal probe) and mitochondrial (B) sequences in DNA isolated from the crude homogenate were not significantly different (P⫽0.36). The endogenous levels based on DNA from isolated mitochondria (C) were approximately two- to threefold higher (P⬍0.01). Also shown are representative Fpg/Southern blots on which the graph is based. The letters on the blots (A, B, C) correspond to those on the graph. ⫺ and ⫹ refer to Fpg digestion.

The FASEB Journal

ANSON ET AL.

this regard because the linearity of the lesion induction across a tremendous range allows methods to be compared not only as to their accuracy, but also as to their precision—that is, their ability to detect differences between several samples that differ by a constant factor. Both HPLC/ECD and the Fpg/Southern blot method were able to detect the linear induction of 8-oxo-dG in MB treated DNA as a function of light exposure (Fig. 1A). Prior to treatment of DNA with photoactivated MB, endogenous 8-oxo-dG was removed from the DNA enzymatically (18). This was done for two reasons: 1) to generate samples that had less damage than would be found in vivo, and 2) to correct for the possibility that endogenous damage would be more extensive in some sequences than in others. This is necessary because the HPLC/ECD and Fpg/Southern blot methods measure different parameters to determine the amount of damage present in a sample. The HPLC/ECD system measures damage directly, and a random distribution is assumed when the number of lesions detected is divided by the amount of DNA assayed. The Fpg/Southern blot method, in contrast, measures undamaged DNA. A random distribution is assumed when the number of lesion-free DNA strands is used in the Poisson equation (14) to calculate the average number of lesions in a given amount of DNA. If damage is distributed randomly, the distinction is not important. If it is not, however, HPLC/ECD will overestimate the number of DNA strands that actually contained damage, whereas the Fpg/Southern blot method will underestimate the number of lesions present in the sample. Although there may be some minor effect of sequence, DNA in vitro is much more homogenous than it is in situ, where protein binding and cellular differences make it likely that some cells and sequences will contain more damage than others. The efficacy of this removal of endogenous damage and the enzymatic/Southern blot method both depend on the efficiency of the enzymatic reaction. To determine this efficiency, we treated aliquots of damaged DNA from each time point with Fpg and measured residual 8-oxo-dG using HPLC/ECD. Greater than 90% of the induced 8-oxo-dG was removed by Fpg treatment (Fig. 1A). This percentage was not dependent on the amount of damage present prior to Fpg digestion within the range measured. Next we analyzed the damaged samples using the Fpg/Southern blot method and compared the results with those of the HPLC/ECD analysis of the same DNA. The methods give virtually identical answers (Fig. 1B). However, at least one FapyGua is produced by treatment of DNA with photoactivated MB for every 5 to 10 8-oxo-dGs (19). FapyGua is not alkali sensitive and is excised by Fpg, and so contribENDOGENOUS mtDNA OXIDATIVE DAMAGE

utes to the damage measured by the Fpg/Southern blot method. Since the two assays agree exactly, and yet roughly 10% of the damage detected by the Fpg/Southern blot assay is not the lesion of interest, then by subtraction the Fpg assay is detecting only ⬃90% of the 8-oxo-dG detected by HPLC/ECD. This agrees with the 90% enzymatic efficiency calculated above based on HPLC/ECD detection of residual 8-oxo-dG in Fpg treated DNA (Fig. 1A). The linearity of the results shows that both measurement systems provide precise determinations of the level of damage, and close agreement between the two methods argues strongly for the precision and accuracy of both assays. To address the issue of mitochondrial isolation, we applied the Fpg/Southern blot analysis (Fig. 2) to the measurement of oxidative damage in mtDNA both with and without isolation of mitochondria. Nonisolated mtDNA contained ⬃5 lesions per 106 bases, whereas levels of damage were approximately threefold higher in the isolated mtDNA from the same animals (P⬍0.01). Since the DNA isolation protocol was the same for all samples, it is likely that the damage is occurring during the mitochondrial isolation. Published values for the ratio of mitochondrial to nuclear levels of 8-oxo-dG range from 2.0 (20) to 16.0 (21). To calculate such a ratio in the present study, it must be assumed that the nuclear sequence probed is representative of the remainder of the genome. This may not be the case: as noted above, there are theoretical reasons to suspect that nuclear ribosomal DNA may contain higher levels of damage than other nuclear sequences. However, this assumption does allow a rough comparison with earlier work in the field. Basing the ratio on mtDNA damage present without isolation of organelles, the ratio of mitochondrial to nuclear levels of 8-oxo-dG is less than 1. Basing the calculation on the level of damage in mtDNA from isolated mitochondria, the ratio is 2.5. Many previous studies have shown that oxidative DNA damage in isolated mtDNA increases with age (22) and in certain disease states, such as Alzheimer’s Disease (23) and atherosclerosis (24). Although the present work suggests that the absolute numbers reported in earlier studies may have been too high, the changes attributed to age or disease remain valid, since both young and old (or diseased and healthy) mitochondria were subjected to the same analysis. The reason for the observed changes must now be revisited: did the change in the DNA truly occur in vivo, or was there a change in other mitochondrial components caused by age or disease that led the mtDNA to be more readily oxidized during isolation? There are many opportunities to oxidize DNA as it is prepared for analysis. This is 359

especially true for mtDNA, which cannot be isolated from the organelles until the mitochondria themselves have been isolated and manipulated extensively. Since an increase in vulnerability during isolation is likely to reflect an increased vulnerability to challenge in vivo, the underlying reason for the vulnerability would be of great importance. If the increase in damage is not due to sensitivity during isolation but instead occurs in vivo, questions concerning the cause of the increased damage, and the cellular and subcellular distribution of the damage, become even more pressing. In summary, using an enzymatic/Southern blot method that does not require mitochondrial isolation, the present study shows that damage in mtDNA is not extensive in rat liver from young rats. Indeed, in the absence of organelle isolation, most genomes did not contain enzyme-sensitive sites (were lesionfree). Rather, low levels of oxidative damage are induced during mitochondrial isolation. It has been proposed in the past that it is high steady-state levels of damage that lead to functional consequences, and that the increase seen with age is due to a vicious cycle of damage causing dysfunction, which in turn causes more damage (25). In view of the present results, this idea must be reevaluated. Steady-state levels of damage in young, normal mtDNA are maintained at low levels, and the cause for increases commonly seen with age and in some disease states must be investigated. The increases may be due to an age-related failure of anti-oxidative pathways or of mtDNA repair, to increased susceptibility to oxidation during isolation, or to some other factor as yet not considered.

8. 9.

10.

11.

12.

13.

14.

15.

16. 17. 18.

19.

20.

REFERENCES 1. 2. 3. 4. 5.

6. 7.

360

Beckman, K. B., and Ames, B. N. (1996) Detection and quantification of oxidative adducts of mitochondrial DNA. Methods Enzymol. 264, 442– 453 Takasawa, M., Hayakawa, M., Sugiyama, S., Hattori, K., Ito, T., and Ozawa, T. (1993) Age-associated damage in mitochondrial function in rat hearts. Exp. Gerontol. 28, 269 –280 Miquel, J., Economos, A. C., Fleming, J., and Johnson, J. E., Jr. (1980) Mitochondrial role in cell aging. Exp. Gerontol 15, 575– 591 Harman, D. (1972) The biologic clock: the mitochondria? J. Am. Geriatr. Soc. 20, 145–147 Higuchi, Y., and Linn, S. (1995) Purification of all forms of HeLa cell mitochondrial DNA and assessment of damage to it caused by hydrogen peroxide treatment of mitochondria or cells. J. Biol. Chem. 270, 7950 –7956 Hegler, J., Bittner, D., Boiteux, S., and Epe, B. (1993) Quantification of oxidative DNA modifications in mitochondria. Carcinogenesis 14, 2309 –2312 Beckman, K. B., and Ames, B. N. (1999) Endogenous oxidative damage of mtDNA. Mutat. Res. 424, 51–58

Vol. 14

February 2000

21. 22. 23. 24.

25.

Driggers, W. J., LeDoux, S. P., and Wilson, G. L. (1993) Repair of oxidative damage within the mitochondrial DNA of RINr 38 cells. J. Biol. Chem. 268, 22042–22045 Taffe, B. G., Larminat, F., Laval, J., Croteau, D. L., Anson, R. M., and Bohr, V. A. (1996) Gene-specific nuclear and mitochondrial repair of formamidopyrimidine DNA glycosylase-sensitive sites in Chinese hamster ovary cells. Mutat. Res. 364, 183–192 Anson, R. M., Croteau, D. L., Stierum, R. H., Filburn, F., Parsell, R., and Bohr, V. A. (1998) Homogenous repair of singlet oxygen-induced DNA damage in differentially transcribed regions and strands of human mitochondrial DNA. Nucleic Acids Res. 26, 662– 668 Croteau, D. L., ap Rhys, C. M., Hudson, E. K., Dianov, G. L., Hansford, R. G., and Bohr, V. A. (1997) An oxidative damagespecific endonuclease from rat liver mitochondria. J. Biol. Chem. 272, 27338 –27344 Shigenaga, M. K., Aboujaoude, E. N., Chen, Q., and Ames, B. N. (1994) Assays of oxidative DNA damage biomarkers 8-oxo-2⬘deoxyguanosine and 8-oxoguanine in nuclear DNA and biological fluids by high-performance liquid chromatography with electrochemical detection. Methods Enzymol. 234, 16 –33 Floyd, R. A., Watson, J. J., Wong, P. K., Altmiller, D. H., and Rickard, R. C. (1986) Hydroxyl free radical adduct of deoxyguanosine:-sensitive detection and mechanisms of formation. Free Rad. Res. Commun. 1, 163–172 Bohr, V. A., and Okumoto, D. S. (1988) Analysis of pyrimidine dimers in defined genes. In DNA Repair: A Laboratory Manual of Research Procedures (Friedberg, E. C., and Hanawalt, P. C., eds) Vol. III, pp. 347–366, Marcel Dekker, Inc., New York Denepoux, S., Razanajaona, D., Blanchard, D., Meffre, G., Capra, J. D., Banchereau, J., and Lebecque, S. (1997) Induction of somatic mutation in a human B cell line in vitro. Immunity 6, 35– 46 Miller, S. A., Dykes, D. D., and Polesky, H. F. (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16, 1215 Fasman, G. D. (1975) Handbook of Biochemistry and Molecular Biology: Nucleic Acids, Vol. II, pp. 268 –275, CRC Press, Inc., Boca Raton, Florida Saljoughi, S., Beckman, K. B., and Ames, B. N. (1997) The use of Fapy glycosylase-treated ‘clean’ DNA as a control in the measurement in DNA oxidation by HPLC-EC. FASEB J. 11, A1250 (abstr.) 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 Mecocci, P., MacGarvey, U., Kaufman, A. E., Koontz, D., Shoffner, J. M., Wallace, D. C., and Beal, M. F. (1993) Oxidative damage to mitochondrial DNA shows marked age-dependent increases in human brain. Ann. Neurol. 34, 609 – 616 Richter, C., Park, J. W., and Ames, B. N. (1988) Normal oxidative damage to mitochondrial and nuclear DNA is extensive. Proc. Natl. Acad. Sci. USA 85, 6465– 6467 Richter, C. (1995) Oxidative damage to mitochondrial DNA and its relationship to ageing. Int. J. Biochem. Cell Biol. 27, 647– 653 Mecocci, P., MacGarvey, U., and Beal, M. F. (1994) Oxidative damage to mitochondrial DNA is increased in Alzheimer’s disease. Ann. Neurol. 36, 747–751 Corral-Debrinski, M., Shoffner, J. M., Lott, M. T., and Wallace, D. C. (1992) Association of mitochondrial DNA damage with aging and coronary atherosclerotic heart disease. Mutat. Res. 275, 169 –180 Bandy, B., and Davison, A. J. (1990) Mitochondrial mutations may increase oxidative stress: implications for carcinogenesis and aging? Free Rad. Biol. Med, 8, 523–539

The FASEB Journal

Received for publication July 2, 1999. Revised for publication September 15, 1999.

ANSON ET AL.