|Home | About | Journals | Submit | Contact Us | Français|
Oxidative damage to DNA has long been associated with aging and disease, with guanine serving as the primary target for oxidation owing to its low ionization potential. Emerging evidence points to a critical role for sequence context as a determinant of the guanine ionization potential and the associated chemical reactivity of the guanine, as well as the spectrum of damage products that arise from oxidation. Recent studies also suggest that the generally accepted model of oxidation hotspots in runs of guanine bases may not hold for biologically relevant oxidants. One of the primary methods used to address these important problems of sequence context utilize gel electrophoresis to identify the location and quantity of base damage arising in model oligonucleotides. However, this approach has limited study to those agents that produce few strand breaks arising from deoxyribose oxidation, while ionizing radiation, Fenton chemistry and other biologically relevant oxidants produce sizeable proportions of both base and sugar damage. To this end, we have developed a universal method to quantify sequence context effects on nucleobase damage without interference by strand breaks from deoxyribose oxidation.
DNA damage resulting from oxidative stress has been strongly associated with cancer, chronic degenerative diseases and aging (reviewed in refs. 1,2). While both the nucleobase and deoxyribose moieties of DNA are targets for oxidation, recent interest in charge transfer and sequence context effects on the location and quantity of damage have focused attention on the bases, with particular attention paid to guanine due to its low ionization potential (3) and the myriad products arising from its primary and secondary oxidation (4). Sequence context has been shown to play a significant role in modulating the ionization potential of guanine in duplex DNA and, hence, the reactivity of guanine with oxidizing agents. For example, it has been demonstrated that many one-electron oxidants, such as anthraquinones (5), rhodium complexes (6) and riboflavin-mediated photooxidation, selectively damage guanine when the base is located adjacent to other guanines (e.g., GG, GGG). This reactivity has been rationalized on the basis of the low ionization potential conferred to guanine in these sequence contexts and the migration of cationic holes to these sites from guanine radical cations located in sequence contexts conferring higher ionization potentials (7). On the other hand, we have recently demonstrated that nitrosoperoxycarbonate, an oxidant formed by reactive oxygen species during chronic inflammation, is selective for oxidizing guanines with the highest ionization potentials (8), while hydroxyl radical generated by Fe+2-EDTA and γ-radiation are equally reactive with guanines irrespective of sequence context (Margolin et al., manuscript in preparation). We have, therefore, shown that sequence selectivity of guanine oxidation in double-stranded DNA is not only a function of sequence context, as has been previously thought, but also depends on the oxidant identity and its interactions with the DNA. Determination of sequence effects in nucleobase oxidation by various agents can thus provide valuable information on their mechanism of damage induction in DNA and on the relationship between reactivity and the potential to cause mutations.
The most widely employed approach to studying sequence context effects on DNA damage involves gel electrophoretic analysis of damage in model oligodeoxynucleotides exposed to oxidizing agents. Base damage in the oligos is converted to strand breaks by treating the DNA with either hot piperidine or with DNA repair enzymes such as E. coli formamidopyrimidine DNA glycosylase (Fpg) for oxidized purines and E. coli endonuclease IV (Nth) for oxidized pyrimidines (10). The strand breaks are then localized on sequencing gels and quantified by autoradiography or phosphorimager analysis. The problem inherent with this approach is that it is limited to oxidizing agents that produce only base damage, since oxidation of deoxyribose results in the formation of direct strand breaks and easily hydrolysable abasic sites that create a background of strand breaks that interfere with quantification of base-derived strand breaks. Such is the case with the biologically relevant oxidizing agent such as ionizing radiation, peroxynitrite, and Fenton chemistry arising with iron and copper (11).
We have developed a method that obviates the background of deoxyribose oxidation-induced strand breaks. Using relatively inexpensive 3’-phosphorothioate-protected oligodeoxynucleotides, the background of strand breaks is removed from the analysis by digestion of the oxidized oligos with E. coli exonuclease III (ExoIII). Subsequent treatment with hot piperidine or DNA repair enzymes exposes the base damage as strand breaks that can be localized and quantified in sequencing gels. This approach provides a nearly universal method for defining the sequence context effects on oxidative damage to DNA.
ImageQuant software (GE)
This method for studying the sequence effects on nucleobase oxidation uses small, synthetic 5’-32P-labeled, double-stranded oligodeoxynucleotides containing guanines in defined sequence contexts. After treatment with a damaging agent, a strand break is introduced at the sites of guanine oxidation by treatment with either hot piperidine or Fpg glycosylase (reviewed in ref. 10). The relative instability of most of the primary and secondary guanine oxidation products to treatment by either one or both of these agents ensures the complete conversion of most guanine oxidation products to strand breaks. The treated oligodeoxynucleotides are then resolved on a DNA sequencing gel and the strand breaks formed at each oxidized guanine are quantified using standard image analysis software.
For analysis of guanine oxidation by an agent that produces significant amounts of deoxyribose oxidation, a protocol modification is introduced that allows the removal of the background of direct strand breaks. This is accomplished preparing double-stranded oligodeoxynucleotides that contain exonuclease-resistant phosphorothioate linkages at their 3’ ends and treating these oligodeoxynucleotides with ExoIII after the damage reaction is complete (See Note 3). Phosphorothioate linkages protect the 3’ ends of the parent oligodeoxynucleotides and oligodeoxynucleotides containing only base lesions from digestion by ExoIII (12). Oligodeoxynucleotides containing strand breaks now have exposed 3’ ends that are substrates for the enzyme. ExoIII recognizes substrates with 3’-hydroxyl, 3’-phosphate and 3’-phosphoglycolate termini (13), as well as substrates containing abasic sites that are cleaved endonucleolytically (14) and thus removed from base damage analysis. After the ExoIII reaction is complete, the only [32P]-labeled oligodeoxynucleotides remaining in solution are the parent molecules or those containing damaged bases that can be revealed as strand breaks following reaction with hot piperidine or Fpg treatment and gel electrophoresis (see Figure 1). Our control experiments have shown that ExoIII treatment does not alter the sequence selectivity of guanine oxidation observed in sequence damage experiments (i.e., the presence or absence of ExoIII does not affect the quantity and location of base damage, as shown in Figure 2 for the guanine-specific oxidant, nitrosoperoxycarbonate).
The first part of this section describes the steps necessary to determine sequence selectivity of guanine oxidation by agents that selectively oxidize guanines in duplex DNA. The second part describes the modifications of the method that provide for removal of the background of direct strand breaks induced by agents capable of significant deoxyribose oxidation. When working with a new damaging agent it is necessary to measure the amount of direct strand breaks that it causes, in order to determine if the ExoIII treatment described in the second part of the method should be used. This can often be accomplished with a single dose-response experiment with and without hot piperidine treatment. The same dose-response experiment should be conducted to determine a dose of the damaging agent that will be used in all subsequent experiments. A dose that is typically chosen should be high enough to induce statistically significant damage at every guanine of interest, as measured by a paired Student’s T-test. However, it should be low enough to damage less than 30% of the parent oligodeoxynucleotide. According to a Poisson distribution, this low level of damage ensures that each oligodeoxynucleotide sustains an average of one or fewer damage reactions. If these single-hit conditions are violated, sequencing gel quantification of DNA damage becomes impossible due to the inability to quantify a second damage event.
All synthetic oligodeoxynucleotides should be purified before use in sequence damage experiments to remove failure sequences and damaged molecules. Gel electrophoresis is the most efficient and reliable method for oligodeoxynucleotide purification (See Note 4). Due to the high frequency of nucleobase damage occurring in synthesis, all oligodeoxynucleotides should be treated with hot piperidine prior to purification (See Note 5).
All oligodeoxynucleotides are dissolved in TE buffer (10 mM Tris, 0.5 mM EDTA, pH 8.0) to a final concentration of 100–500 pmol/µl. An equal volume of 2 M piperidine solution in distilled and deionized water is added and the oligodeoxynucleotides are incubated at 90 °C for 20 min. Following drying under vacuum (e.g., Speedvac). the samples are again dissolved in TE containing bromophenol blue dye and 20–25% glycerol to a DNA concentration of 100–500 pmol/µl.