Accumulating evidence indicates that homologous recombination plays an important role in mammalian cells, primarily in the reactivation of stalled replication forks. During a normal round of replication, stalling of the replication fork occurs frequently when a nick or a lesion, which cannot be by-passed, is encountered. Progression of the replication fork can also be hindered by a protein bound to or secondary structures in the DNA. It is vital for the cell to restart stalled replication forks efficiently since nucleases might otherwise damage the unprotected DNA and since failure to reinitiate replication leads ultimately to cell death. The mechanism by which homologous recombination reactivates stalled replication forks and the substrates that trigger this response in mammalian cells have not yet been elucidated.

The present thesis focuses on the function of homologous recombination as well as on the mechanism for induction of this process in mammalian cells. Our major interest in this context was the repair of lesions encountered during replication and that interfere with replication fork progression. The role of the RAD51 protein in this process was also examined. The main experimental system we employed consisted of Chinese hamster cell lines deficient in different DNA repair pathways, as well as the V79 Chinese hamster SPD8 cell line, which contains an endogenous locus at which homologous recombination events can be monitored.

Our findings demonstrate that homologous recombination can be employed by mammalian cells to repair several different types of lesions encountered during replication. The mechanism underlying such homologous recombinational repair appears to depend on the type of lesion involved. A replication fork encountering a single-strand break was shown to collapse into a double-strand break and subsequently be repaired by homologous recombination involving a mechanism referred to as break-induced replication. A stalled replication fork was shown to induce a substrate for homologous recombination both with and without the induction of a double-strand break. This suggests that homologous recombination could repair lesions other than double-strand breaks during replication. Non-homologous end-joining was shown to play only a minor role in the repair of DNA damage associated with replication, being employed only in situations when the replication fork is processed into a double-strand break.

Experiments concerning the role of RAD51 in repair of lesions encountered during replication, employing a cell line that overexpresses this protein, indicated that RAD51 is rate limiting for the repair of DNA damage associated with the replication fork. The level of RAD51 was also measured in lung cancer cells and found to determine both the survival and the level of double-strand breaks formed in response to etoposide treatment, suggesting RAD51 to be involved in the resistance of tumours to chemotherapeutic drugs.

In summary, our findings demonstrate that RAD51-dependent homologous recombination is an important repair pathway during replication in mammalian cells.