Analysis of DNA Recombination Proteins in DNA Damage Response

Chromosomes are the carriers of our genome. All the information for a cell's survival and propagation is stored there in the base sequence of the DNA. Unfortunately, our DNA is under continuous attack from DNA damaging agents, of which some are produced during a cell's own metabolic processes, while others may be of exogenous origin. DNA damage leads to mutations if not (or incorrectly) repaired and, depending on the nature of the mutation, can lead to cancer or other diseases. Our cells suffers 10,000-1,000,000 DNA lesions a day, but still many of us do not develop cancer or, if so, at a relatively late age. This illustrates the importance of the various DNA repair mechanisms in our body: without it we would not be able to survive for long. The research described in this thesis is about the response of proteins involved in homologous recombination, one of the DNA repair pathways, to DNA damage in mammalian cells. Recombinational repair is involved in the repair of a special class of DNA lesions: double strand breaks (DSBs) and interstrand cross-links (ICLs). These are a very toxic class of lesions, i.e. even one DSB could be lethal to a cell (if left unrepaired). DSBs can also be repaired by another pathway, called non-homologous end-joining. The critical differences between end-joining and recombination is that during end-joining the broken ends of a DNA molecule are stuck together, which is not necessarily error-free. Homologous recombination on the other hand, uses an undamaged, identical piece of DNA as a template, makes a copy of it in order to repair the damaged molecule and is error-free. However, this is not an easy task, as the template, usually the sister chromatid or homologous chromosome, must be found. Then the broken molecule must pair and one (or both) of the recessed ends must invade the template molecule. These steps, named homologous pairing and strand exchange/invasion are considered to be the critical steps of homologous recombination. Key proteins involved in these processes are the in chapter one described Rad52 group of proteins, which includes Rad51 and Rad54, the central proteins of this thesis. Equally, the nature and importance of DNA lesions, the various DNA repair pathways and the two other functions of recombination are described in this introductory chapter. To study the proper functioning of Rad51 after DNA damage, we made use of one particular feature of the protein which is described in chapter two: it redistributes into nuclear foci as a response to treatment with DNA damaging agents that cause DSBs and ICLs. Since Rad52 and the Rad51 paralogues XRCC2 and XRCC3 are also involved in recombinational repair we screened the respective mutant cell lines to assess the hierarchy and order of these proteins relative to Rad51. As the XRCC2 and XRCC3 mutants were unable to form Rad51 foci in response to induced DNA damage, we concluded that the XRCC2 and XRCC3 proteins are essential for proper Rad51 functioning after DSB and ICL induction. A Rad52KO mutation did not lead to ablation of Rad51 foci, but using a cell line expressing a Rad52GFP construct we could demonstrate that both Rad51 and Rad52 form damage inducible foci and partially co-localize. Thus, though Rad52 is involved, it is not essential for Rad51-associated DSB repair in mammalian cells. We also screened other mutant cell lines, known for their sensitivity for DSB or ICL inducing agents to detect whether the gene products they were mutated in were essential for Rad51-associated repair. The Snm1 mutant and V-H4 cell line were able to form Rad51 foci, but the VC-8 cell line was unable to form foci. Complementation studies revealed that the mutation involved the breast cancer associated gene Brca2. We also looked for interactions between Rad51 and Rad54, as described in chapter three. We found that Rad54 also forms foci as a response to DSB inducing ionizing radiation (IR). In addition, we tested different types of DNA damaging agents to see which ones could induce Rad51 and Rad54 foci. Not only do these damage-inducible foci co-localize, but the proteins do physically interact, which we could only detect after inducing DSBs. We provided evidence that DNA damage-induced Rad51 foci are less stable in the absence of Rad54. From results of a topological assay we conclude that Rad54 is capable of introducing supercoiling in dsDNA by translocating along it, which could be important for the formation or stabilization of Rad51-mediated joint molecule formation. The mechanistic and biochemical functions of Rad54 are further discussed in chapter four. In this chapter, our own results and recent biochemical data of others are put in perspective and argue that Rad54 is a much more versatile protein than previously thought. Originally considered as an accessory protein whose mechanistic functions had not been properly clarified we discuss that it could have important functions throughout the three main stages of recombination; during pre-synapsis, when it stabilizes Rad51 nucleoprotein filaments, during joint formation or synapsis, and post-synapsis when it could remove the recombination proteins after the reaction. DSBs can be formed by treating cells with DNA damaging agents. However, DSBs also occur during DNA replication, another process in which homologous recombination is involved (see chapter one, section 1.5). The DSBs could occur in the vicinity of stalled replication forks and we investigated this in closer detail in chapter five. Using hydroxy-urea (HU), a replication inhibitor, and looking for Rad54 foci formation as an indicator for DNA damage, we found that Rad54 foci are formed indeed when replication is stalled. Formation of these foci was biologically relevant, as we observed that Rad54KO icells are sensitive to HU. However, the number of foci positive cells disappears much faster in HU-treated cell populations than in IR-treated populations, indicating that the HU-induced foci probably reflect a process different from the usual DSB repair. Although we found accumulation of HU-induced DSBs in the absence of Rad54, we think that it is not due to a defect in HR-mediated DSB repai,r because this process is likely to be very inefficient in the presence of HU, even in DSB-proficient cells. Possibly, a function of Rad54 is to regre