Postreplication repair of DNA in ultraviolet-irradiated mammalian cells

Abstract

DNA synthesized in ultraviolet-irradiated L5178Y mouse lymphoma cells is smaller than that made in unirradiated cells, as shown by pulse-labeling with tritiated thymidine, followed by sedimentation of the DNA in alkaline sucrose gradients. Hence, DNA synthesized on ultraviolet-irradiated DNA templates contains discontinuities. The estimated distance between these discontinuities is comparable to the average spacing between pyrimidine dimers in the parental strands, suggesting that the replication machinery leaves gaps opposite the dimers. On subsequent incubation these discontinuities are filled in by a process which is inhibited by hydroxyurea. In order to determine whether the gaps are filled in by a recombinational mechanism or by de novo synthesis, irradiated cells were pulse-labeled with radioactive thymidine, followed by incubation in bromodeoxyuridine and exposure to light of wavelength 313 nm. Bromuracil-containing sections of DNA are selectively disrupted by light of this wavelength. Thus if the gaps in the radioactively labeled daughter strands are filled in by exchange with thymidine-containing DNA from the parental strands (recombinational exchange), they will not contain bromuracil, and the molecular weight distributions of the radioactive DNA from initially irradiated and unirradiated cells will be similar after exposure to 313-nm light. On the other hand, if the gaps are filled in by bromuracil-substituted DNA (arising from de novo synthesis), the small pieces of radioactive DNA synthesized in the initially irradiated cells during the pulse-labeling will be reformed on exposure to 313-nm light, and the molecular weight distribution from those cells will have a lower average molecular weight than that from similarly treated unirradiated cells. The results are in accordance with the second alternative, indicating that the gaps in the daughter strands opposite the pyrimidine dimers are filled in with newly synthesized DNA. Furthermore parental DNA does not seem to be involved in the gap-filling process. This excludes many of the recombinational models which have been put forward to account for similar gap-filling phenomena observed in bacteria. The results give an approximate value of 800 nucleotides for the size of the filled-in gaps.