[en] Deinococcus radiodurans is a Gram-positive bacterium known for its extreme resistance to a broad variety of DNA damaging agents. Among these, Ionizing Radiations and desiccation are the most harmful for the cell, since they introduce breaks in the genome. Double Strand Breaks (DSB) are particularly hazardous for the cell and they need to be repaired very efficiently, in order to avoid mutations leading to altered, if not lethal, phenotypes. Homologous Recombination (HR) is the most efficient mechanism by which DSBs are repaired. D. radiodurans is able to completely restore its genome in only 3 hours, and it accomplishes the entire process through the RecFOR pathway. In order to be repaired, DSBs first need to be recognized. The protein believed to be responsible for this important step that takes place soon after the damage occurs in the cell, is RecN. RecN is recruited at the early stages of DNA repair and in vivo studies have demonstrated its propensity to localize to discrete foci. In vitro studies also suggest that RecN possesses a DNA end-joining activity previously observed for SMC proteins (such as cohesin), which are structurally related to RecN. Several structural studies have been carried out on the SMC-like protein, Rad50, but so far no structural information is available for RecN. The work presented here focused on the structural characterization of RecN and its constitutive domains. We obtained crystal structures of three partially overlapping constructs of RecN and Small Angle X-ray Scattering was performed on the individual domains and the full-length protein. The study of RecN in solution complemented our crystallographic study and enabled us to build a reliable, atomic model of the full-length protein. Mutations were designed and the mutant RecN proteins were produced in order to characterize the ATP hydrolysis activity of RecN, which is a conserved feature of this family of proteins. Extensive biochemical studies were carried out on wild-type and mutants of both the full-length protein and the single domains, in order to determine the role and function of each of the domains. Our results led us to propose a model for how RecN might recognize DSBs, tether two broken DNA ends and prepare the DNA for subsequent repair by the RecFOR machinery. (author)