Salmonella is a major cause of food-borne outbreaks. Since the early 1990s, isolation of multidrug-resistant Salmonella has increased with the emergence of epidemic Salmonella Typhimurium DT104 strains which are mostly resistant to ampicilin, chloramphenicol, streptomycin, sulfonamides and tetracyclines. The genes coding these resistances are clustered on a complex class 1 integron (MDR region) located on a genomic island of 43 kb designated SGI1. Since its first identification in S. Typhimurium DT104, SGI1 has been identified worldwide in other Salmonella serotypes, and more recently in Proteus mirabilis. For these strains, multidrug resistance is conferred, either to the classical structure of SGI1, or to related variants of SGI1 corresponding to the initial structure of SGI1 with modification of the complexe integron In104. The Salmonella Genomic Island 1 (SGI1) constitutes a great concern since it confers a multidrug resistance phenotype often responsible of therapeutic failures which may cause important complications, or even death. In this context, the work has been focused on the major health issue represented by the epidemic diffusion of the Salmonella Typhimurium clone in the course of 1990s in human and cattle. As a first objective, the work allowed to appreciate the molecular evolution of SGI1 in the course of time and to assess the diffusion of SGI1 to other bacterial strains than Salmonella in natural conditions. Therefore, an overview of the multidrug resistance in Salmonella Typhimurium strains isolated from diseased cattle in France from 2002 to 2007 was carried out and a screening of natural strains from other bacterial species than Salmonella that may harbor SGI1 was undertaken. The results indicated weak molecular evolutions of SGI1, which seems in contradiction with the great capability of SGI1 to recombine and transfer, as attested in vitro as in vivo. Nevertheless, this study allowed the first description of a new SGI1 variant, named SGI1-T, which is the result of intra-molecular recombination events. Another second objective of the thesis was to contribute to a better knowledge of the role of SGI1 in bacterial virulence. A strategy of experimental modeling (murine systemic salmonellosis) was first set up to compare the levels of virulence conferred by isogenic strains differing only by the presence or the absence of SGI1. A second approach was also carried out to evaluate the role of SGI1 in biofilm formation. Indeed, the organization in biofilm facilitates bacterial colonization, which constitutes in turn an element of effectiveness of the final virulence. A positive role of SGI1 in biofilm formation was demonstrated in the framework of this study, and more generally, questions the role of SGI1 in the Quorum Sensing regulation system.
