Measurement of 8-OHdG in DNA by HPLC/ECD: The Importance of DNA Purity

DNA has been considered to be pure when the ratio of absorbance at 260 nm over the absorbance at 280 nm (A260/A280) is 1.8 (8). This definition of purity, however, is based more upon a functional assessment (e.g., ability to cut the DNA with restriction enzymes) than upon more conventional analytical criteria. In a recent article, Glasel reexamined the significance of the A260/A280 (3). He noted that the method was originally and more aptly used to detect nucleic acid contamination in protein (10) and that the method was actually poor for estimating DNA purity. The A260/A280 for pure nucleic acid is about 2.0, while a solution with an A260/A280 of 1.8 theoretically contains about 40% nucleic acid and 60% protein (3). Assumptions about DNA purity may have important consequences depending upon the endpoint in question. The purity requirements of DNA for molecular biological applications are far less stringent than necessary for more analytical endpoints, such as quantification of DNA adducts. Recently, we demonstrated that protein contamination of DNA preparations must be taken into consideration when attempting to estimate covalent binding of a radioactive compound to DNA because of the potential of certain reactive metabolites to preferentially or exclusively bind to protein (1). Since protein absorbs approximately 8–10 times more light at 230 nm than at 280 nm, we, like Glasel, concluded that the A260/A280 was a poor indicator of DNA purity. We found that the A260/A230 was a superior indicator of protein contamination. In fact, addition of protein will shift the spectral minimum of DNA from 230 nm to longer wavelengths (1). Therefore, it is important to obtain the entire UV absorbance spectrum versus determining absorbances at only a few wavelengths. In this report, we show that quantification of 8-hydroxydeoxyguanosine (8OHdG, a promutagenic DNA adduct resulting from oxidative damage) by electrochemical detection (ECD) may be affected by the quality of DNA preparations. Furthermore, UV spectra of the DNA samples and the HPLC/UV profiles of hydrolyzed DNA provide a means to assess the functional purity of DNA necessary for this sensitive analysis. Using a relatively standard DNA isolation procedure (proteinase K, RNases, solvent extractions and alcohol precipitations), we observed that levels of 8-OHdG in DNA samples isolated from rat tissues were highly variable. In DNA samples that had very high 8-OHdG content, we noticed that the UV peak corresponding to 5methyldeoxycytosine (5mdC) was also unusually large (Figure 1). There was a good correlation (r = 0.98) between the occurrence of the exaggerated 5mdC peak and the large amount of 8-OHdG apparently present in the sample (Figure 2). Published values for 5mdC in normal rat liver are approximately 1 mol% (6,9). Our data suggested that when the amount of 5mdC in rat DNA quantified by our HPLC method exceeded about 1.4 mol%, the DNA contained sufficient amounts of contaminants to produce an overestimation of the 8-OHdG content. We further observed that most of the samples of DNA containing apparently high levels of 8-OHdG had A260/A280 ratios greater than 1.8, which has historically been considered indicative of “pure” DNA. Nevertheless, the samples did have abnormal UV spectra; the A260/A230 were less than 2.2 and the λmin were more than 231 nm. When DNA was purified from another piece of tissue taken from the original frozen sample, we found that the newly prepared DNA often had acceptable UV spectral values, normal 5mdC content and a reasonable 8OHdG content. This suggests that the initial DNA sample contained impurities, perhaps protein and other unidentified contaminants, that had not been completely removed. An impurity could have catalyzed the formation of 8-OHdG from dG in the initial samples. Alternatively, an electrochemically active impurity may have co-eluted with