Kinetics of recombinational hybrid formation in X-irradiated mammalian cells: a possible first step in the repair of DNA double-strand breaks.

It is known that mammalian cells repair X-ray-induced double-strand breaks (DSB). The mechanism of this repair is, however, as yet unknown but it is thought that the repair may involve recombination between homologous DNA strands. We have investigated the presence and kinetics of occurrence of recombinational molecular intermediates or 'heavy-heavy' (HH) hybrids in DNA of X-irradiated mouse Ehrlich ascites tumour cells unifiliarly substituted with bromodeoxyuridine. Purified DNA from density-labelled cells was analysed using isopyknic CsCl density centrifugation. After rebanding of the high-density material, the relative amount of HH-hybrid material was determined. When the cells were incubated after X-ray exposure, hybrids accumulated with an apparently biphasic kinetic; first a rapid accumulation directly after irradiation followed by a second, more slowly appearing peak. When DSB repair was inhibited by the nucleoside analogue 9-beta-D-arabinofuranosyladenine (ara-A), a powerful DNA polymerase inhibitor, the kinetics of the HH-hybrid formation were similar to those during the incubation after X-irradiation without ara-A. We interpret this as indicating that the first step in repair of DSB i.e. hybrid formation, occurs in the absence of DNA synthesis as predicted by the Resnick model of DSB repair. In an experiment in which the ara-A was washed away from the irradiated cells after 8-h treatment with the drug and replaced by fresh growth medium, so allowing DSB to repair, a similar HH-hybrid kinetic response was found to that occurring directly after irradiation in the absence of ara-A. The reasons for this are not yet clear. In this case, however, the response of the ara-A-treated cells after X-irradiation was much stronger than that in the untreated cells where only approximately 20% of the DSB remained. These kinetic results which show a temporary appearance of HH hybrids, indicate that a net exchange of genetic material does not take place; they therefore lend support to the postulated recombinational model of Resnick in which a temporary exchange between homologous DNA strands takes place during DSB repair.