Emerging evidence points to the importance of deoxyribose oxidation in the toxicity of oxidative DNA damage, including the formation of protein-DNA crosslinks and base adducts. With the goal of understanding the differences in deoxyribose oxidation chemistry known to occur with different oxidants, we have compared the formation of one product of 3'-oxidation of deoxyribose in DNA, 3'-phosphoglycolaldehyde (PGA) residues, in isolated DNA and cells exposed to ionizing radiations. A recently developed gas chromatography/negative chemical ionization mass spectrometry method was used to quantify PGA residues in purified DNA and in human TK6 lymphoblastoid cells exposed to gamma radiation (60Co) and alpha particles (241Am). The level of PGA residues was then correlated with the total quantity of deoxyribose oxidation determined by plasmid topoisomer analysis. Alpha-particle irradiation (0-100 Gy) of purified DNA in 50 mM potassium phosphate (pH 7.4) produced a linear dose response of 0.13 PGA residues per 10(6) nucleotides per gray. When normalized to an estimate of the total number of deoxyribose oxidation events (2.0 per 10(6) nucleotides per gray), PGA formation occurred in 7% (+/-0.5) of deoxyribose oxidation events produced by alpha-particle radiation. In contrast, the efficiency of PGA formation in gamma-irradiated DNA was found to be 1% (+/-0.02), which indicates a shift in the chemistry of deoxyribose oxidation, possibly as a result of the different track structures of the two types of ionizing radiation. Studies with gamma radiation were extended to TK6 cells, in which it was observed that gamma radiation produced a linear dose response of 0.0019 PGA residues per 10(6) nucleotides per gray. This is consistent with an approximately 1000-fold quenching effect in cells, similar to the results of other published studies of oxidative DNA damage in vivo.

1881 Deoxyribose oxidation plays a critical role in the genetic toxicology of oxidative stress, including involvement in complex DNA lesions, cross-linking with DNA repair proteins and the formation of endogenous DNA adducts. The latter is illustrated by our discovery that 3′-phosphoglycolaldehyde (PGA), arising from 3′-oxidation of deoxyribose in DNA, reacts to form glyoxal and the mutagenic glyoxal adduct of dG. Given the lack of sensitive methods to quantify deoxyribose oxidation products, we have used PGA as a model to develop a sensitive GC/MS method to identify and quantify deoxyribose oxidation products containing aldehydes and ketones. In previous studies, we demonstrated that gamma-radiation produces 1.5 PGA residues/10e6 nt/Gy in isolated DNA, with PGA representing 1% of deoxyribose oxidation events. Here, we focus on peroxynitrite (ONOO-), a chemical mediator of inflammation that arises from the reaction of superoxide and NO produced by macrophages. Under biological conditions, ONOO- reacts with carbon dioxide to form nitrosoperoxycarbonate (ONOOCOO-), a reaction that shifts the oxidative DNA damage chemistry from predominantly deoxyribose with ONOO- to guanine with ONOOCOO-. To further investigate this issue, we quantified the yield of PGA in isolated DNA in 50 mM phosphate buffer, pH 7.4, with bicarbonate concentrations ranging from 0 to 100 mM and ONOO- concentrations from 0 to 1000 μM. At biologically relevant concentration of 25 mM bicarbonate (∼1 mM carbon dioxide), the production of PGA was below the detection limit of 2 PGA/10e6 nt. The yield of PGA with ONOO- was inversely proportional to the bicarbonate concentration over the range of 0 to 10 mM. This leads us to conclude that ONOOCOO- is either not capable of oxidizing deoxyribose to produce PGA, perhaps due to the low oxidation potential of the carbonate radical anion active intermediate in ONOOCOO-induced oxidations, or that the 3′-oxidation chemistry follows a different path with this oxidant than with ONOO-. Another factor affecting the yield of PGA is the reaction of ONOO- with PGA to form a 3′-phosphoglycolate residue, a less genotoxic DNA oxidation product. To address this question we exposed synthetic PGA and an oligonucleotide containing a 3′-PGA residue to ONOO- in the presence of 1 mM calf thymus DNA in 50 mM phosphate buffer in the presence or absence of 25 mM bicarbonate. While 50% of isolated PGA was oxidized by 1 mM ONOO-, PGA bound to the oligonucleotide was unreactive toward ONOO-. We propose that negative charge repulsion with the sugar-phosphate backbone accounts for the protective effects of the oligonucleotide. These observations suggest that carboin dioxide tension may play a role in defining the spectrum of oxidative DNA lesions in cells and that the reactions by negatively charged oxidants in DNA are limited by charge repulsion.

Deoxyribose oxidation in DNA represents a biologically important facet of oxidative DNA damage that gives rise to protein-DNA cross-links and base adducts. Toward the goal of quantifying deoxyribose oxidation chemistry in cells, we report a method for the quantification of 3'-phosphoglycolaldehyde (PGA) residues, which likely arise from 3'-oxidation of deoxyribose in DNA. The method exploits the aldehyde moiety in PGA by derivatization as a stable oxime with pentafluorobenzylhydroxylamine, followed by solvent extraction and gas chromatography/negative chemical ionization/mass spectrometry. A stable isotopically labeled [(13)C(2)]PGA was synthesized and used as an internal standard. The assay showed a linear response over the range of 30 fmol to 300 pmol, and its precision was verified by analysis of a synthetic, PGA-containing oligodeoxynucleotide. The limit of detection in the presence of DNA was 30 fmol per sample, corresponding to two molecules of PGA in 10(6) nucleotides for 170 microg of DNA. Samples were exposed to 0-100 Gy of (60)Co gamma-radiation, which resulted in a linear dose-response of 1.5 PGA residues per 10(6) nucleotides per Gy and a radiation chemical yield (G-value) of 0.0016 micromol/J. When compared to the total quantity of deoxyribose oxidation occurring under the same conditions (141 oxidation events per 10(6) nucleotides per Gy; determined by plasmid topoisomer analysis), PGA formation occurs in 1% of deoxyribose oxidation events. This small fraction is consistent with current models of limited solvent accessibility of the 3'-position of deoxyribose, although partitioning of 3'-chemistry could lead to other damage products that would increase the fraction of oxidation at this site in deoxyribose.