The distribution of DNA excision-repair sites in human diploid fibroblasts following ultraviolet irradiation.

Using the technique for separating DNA fragments containing excision-repair sites from total genomic DNA as described in the previous paper (Cohn, S. M., and Lieberman, M. W. (1984) J. Biol. Chem. 259, 12456-12462), we have developed a method for directly determining the distribution of excision-repair sites in the genome. DNA was prepared from confluent, diploid human fibroblasts which had been irradiated with ultraviolet light and incubated in the presence of 5-bromo-2'-deoxyuridine (BrdUrd), repaired fragments were isolated, and the dependence of the fraction of total DNA fragments containing excision-repair sites on DNA fragment length was determined by electrophoretic analysis. The observed dependence was compared to the relationship expected for a random distribution of repair sites. At 36 h following 3 J/m2 UV, the distribution of repair sites was indistinguishable from a random distribution; however, at doses of UV above 6 J/m2, the observed dependence indicated that the distribution of repair sites was nonrandom. A time course of the distribution of repair sites following 12 J/m2 UV was clearly nonrandom from 4 h after irradiation until at least 36 h following irradiation. By 72 h, however, the distribution had become random. In cells treated with hydroxyurea, a reduced number of excision-repair sites were present, but the distribution of repair sites was also nonrandom. Autoradiographic analysis of the amount of unscheduled DNA synthesis in individual nuclei suggested that the nonrandom distribution of repair sites did not result from variable extents of repair synthesis in different cell populations or from cell death.