The production of guanine radicals in DNA via the flash−quench technique is shown to cause the formation of covalent adducts between DNA and histone protein. In the flash−quench experiment, the DNA-bound intercalator Ru(phen)2dppz2+ (phen = 1,10-phenanthroline, dppz = dipyridophenazine) is excited with 442 nm light and quenched oxidatively by Co(NH3)5Cl2+, methyl viologen (MV2+), or Ru(NH3)6 3+ to produce Ru(phen)2dppz3+, a strong oxidant (+1.6 V) that can oxidize a nearby guanine base (+1.3 V). The guanine radical thus produced is vulnerable to nucleophilic attack and can react with amino acid side chains to form DNA−protein cross-links. Evidence for DNA−protein cross-linking was provided by the chloroform extraction assay, a filter binding assay, and gel electrophoretic analysis. After flash−quench treatment, pUC19 plasmid DNA undergoes a dramatic decrease in mobility that is reversed upon digestion with proteinase K, as seen by agarose gel electrophoresis. In polyacrylamide gel electrophoresis (SDS-PAGE) experiments, the histone protein shows similar mobility shifts. Cross-linking is observed with poly(dG-dC) and mixed sequence DNA, but not with poly(dA-dT), indicating that the reaction requires guanine bases. Measurements of emission quenching indicate that for a given quencher, the amount of cross-linking is correlated to the amount of quenching. When comparing different quenchers, however, the amount of cross-linking is inversely related to the amount of quenching and decreases in the order Co(NH3)5Cl2+ > MV2+ > Ru(NH3)6 3+. This trend in cross-linking correlates instead with the lifetime of the guanine radical measured by transient absorption spectroscopy, and suggests that the cross-linking reaction requires > 100 μs. These results demonstrate that the flash−quench technique is an effective approach for the study of covalent adducts between DNA and protein formed as a result of guanine oxidation, and suggest one possible fate for oxidatively damaged DNA in vivo.

DNA-protein cross-links form when guanine undergoes a 1-electron oxidation in a flash-quench experiment, and the importance of reactive oxygen species, protein, and photosensitizer is examined here. In these experiments, a strong oxidant produced by oxidative quenching of a DNA-bound photosensitizer generates an oxidized guanine base that reacts with protein to form the covalent adduct. These cross-links are cleaved by hot piperidine and are not the result of reactive oxygen species, since neither a hydroxyl radical scavenger (mannitol) nor oxygen affects the yield of DNA-histone cross-linking, as determined via a chloroform extraction assay. The cross-linking yield depends on protein, decreasing as histone > cytochrome c > bovine serum albumin. The yield does not depend on the cytochrome oxidation state, suggesting that reduction of the guanine radical by ferrocytochrome c does not compete effectively with cross-linking. The photosensitizer strongly influences the cross-linking yield, which decreases in the order Ru(phen)(2)dppz(2+) [phen = 1,10-phenanthroline; dppz = dipyridophenazine] > Ru(bpy)(3)(2+) [bpy = 2,2'-bipyridine] > acridine orange > ethidium, in accordance with measured oxidation potentials. A long-lived transient absorption signal for ethidium dication in poly(dG-dC) confirms that guanine oxidation is inefficient for this photosensitizer. From a polyacrylamide sequencing gel of a (32)P-labeled 40-mer, all of these photosensitizers are shown to damage guanines preferentially at the 5' G of 5'-GG-3' steps, consistent with a 1-electron oxidation. Additional examination of ethidium shows that it can generate cross-links between histone and plasmid DNA (pUC19) and that the yield depends on the quencher. Altogether, these results illustrate the versatility of the flash-quench technique as a way to generate physiologically relevant DNA-protein adducts via the oxidation of guanine and expand the scope of such cross-linking reactions to include proteins that may associate only transiently with DNA.