The soil-dwelling gram-positive coccobacillus Rhodococcus equi is a well-known veterinary pathogen and emerging human pathogen. Although Rhodococcus infection is primarily associated with pyogranulomatous pneumonia in foals, these bacteria can also infect other animal species including humans. R. equi pathogenicity is mediated by the conjugative virulence plasmid (pVAP) which promotes intracellular proliferation in host macrophages. Currently R. equi is endemic in horse breeding farms worldwide. No R. equi vaccine is available and both treatment and prophylaxis rely on the administration of a prolonged course with a combination of macrolide antibiotics (typically erythromycin) and rifampicin. These antimicrobials were introduced in the therapy against R. equi in the 1980s and multiresistance has now emerged among foal isolates, increasing the risk of zoonotic transmission. In this thesis, the role of conjugal extrachromosomal replicons in the host adaptation of R. equi was explored. A previous epidemiological study indicated that different variants of the pVAP virulence plasmid are associated with different animal hosts. This PhD project provides experimental confirmation of a R. equi plasmid-driven host tropism. In vivo and in vitro competition assays were performed using a set of isogenic strains only different in the virulence plasmid type, in adapted (horse) and non-adapted (mouse) model species. The data obtained in the horse model provides clear evidence of a significant negative selection of the non-equine virulence plasmids both at a cell and at the animal level, while no selection was observed in the non-adapted mouse model. Furthermore, this project characterized the determinant responsible for macrolide resistance in R. equi, a novel erm methylase gene, erm(46). The erm(46) determinant was shown to be transferable between strains by conjugation and herein the underlying mechanism and how erm(46) becomes stabilized in R. equi is described. PacBio SMRT-sequencing based analysis revealed that the erm(46) gene is carried in a self-replicating conjugative plasmid of about 80 kb, that we designated pRErm46. The conjugation machinery of pRErm46 was hypothesized to be responsible for bringing the erm(46) determinant into R. equi. However, some erythromycin resistant isolates lack pRErm46 but erm(46) transfer is still observed. This reflects the observation that erm(46) is present in a mobile element that, upon acquisition with the pRErm46 replicon, transposes at a high frequency and to multiple locations of the host genome. If the erm(46) mobile element transposes to the chromosome, no further transfer of the resistance is observed at a detectable frequency in the absence of pRErm46. On the other hand, if the erm(46) element transposes to the R. equi virulence plasmid, the erm(46) determinant co-opts the pVAPA conjugal transfer machinery and gets transferred at the same high frequency as the virulence plasmid (10-2). This constitutes a unique example of efficient co-transfer, in the same genetic vehicle, of virulence and antimicrobial determinants, two key niche-adaptive traits required for within-host survival of bacterial pathogen.