5'-(2-phosphoryl-1,4-dioxobutane) as a product of 5'-oxidation of deoxyribose in DNA: elimination as trans-1,4-dioxo-2-butene and approaches to analysis.

Oxidation of deoxyribose in DNA leads to the formation of a spectrum of electrophilic products unique to each position in the sugar. For example, chemical reactions following abstraction of the C5'-hydrogen atom partition to form either a nucleoside 5'-aldehyde residue attached to the 5'-end of the DNA strand or a 5'-formyl phosphate residue attached to the 3'-end of the DNA strand that is accompanied by a four-carbon fragment on the 5'-end. We now present two approaches that both identify the latter fragment as 5'-(2-phosphoryl-1,4-dioxobutane) and provide a means to quantify the formation of this residue by different oxidizing agents. The first approach involves oxidation of DNA followed by reaction with O-benzylhydroxylamine to form stable dioxime derivatives of the putative 5'-(2-phosphoryl-1,4-dioxobutane) residues. The beta-elimination product of this dioxime proved to be the expected trans-1,4-dioxo-2-butene, as judged by gas chromatographic and mass spectrometric (GC/MS) comparison to authentic dioximes of cis- and trans-1,4-dioxo-2-butene, which revealed a unique pattern of three signals for each isomer, and by X-ray crystallography. Using a benzylhydroxylamine dioxime derivative of [2H4]-labeled cis-1,4-dioxo-2-butene as an internal standard, the dose-response for the formation of 5'-(2-phosphoryl-1,4-dioxobutane) was determined to be linear for gamma-radiation, with approximately 6 lesions per 10(6) nt per Gy, and nonlinear for Fe2+-EDTA. A comparison of 5'-(2-phosphoryl-1,4-dioxobutane) formation to total deoxyribose oxidation suggests that gamma-radiation produces approximately 0.04 lesions per deoxyribose oxidation event. As a positive control for 5'-oxidation of deoxyribose, the enediyne calicheamicin was observed to produce 5'-(2-phosphoryl-1,4-dioxobutane) at the rate of approximately 9 lesions per 10(6) nt per microM. A second approach to identifying and quantifying the sugar residue involved derivatization with hydrazine and beta-elimination to form pyridazine followed by quantification of the pyridazine by GC/MS. Using this approach, it was observed that the enediyne, neocarzinostatin, produced a linear dose-response for pyridazine formation, as expected given the ability of this oxidant to cause 1'-, 4'-, and 5'-oxidation of deoxyribose in DNA. The antitumor antibiotic, bleomycin, on the other hand, produced pyridazine at a 10-fold lower rate, which is consistent with 4'-chemistry as the predominant mode of deoxyribose oxidation by this agent. These results provide novel insights into the chemistry of deoxyribose oxidation in DNA and two approaches to quantifying the 5'-(2-phosphoryl-1,4-dioxobutane) precursor of trans-1,4-dioxo-2-butene, an electrophile known to react with nucleobases to form novel DNA adducts.