Many insect species are associated with endosymbiotic bacteria which have the particularity of living within host tissues. Endosymbionts benefit from this stable and nutritious environment, while providing ecological advantages to their host, such as protection against parasites or thermal tolerance. Spiroplasma poulsonii is an endosymbiotic bacterium that infects natural populations of Drosophila melanogaster. Spiroplasma poulsonii lacks a cell wall, a fact that renders it invisible to the host immune system and allows it to thrive in the host hemolymph. It invades the female germline by co-opting the host’s yolk transport and uptake machinery, which ensures its vertical transmission. Its efficient transmission is associated with a phenotype called male-killing, whereby infected male embryos die during their early development while infected females survive. The mechanisms that ensure the stability of Drosophila-Spiroplasma symbiosis are increasingly well understood, but the bacterial genes involved remain poorly known because of the intractability of Spiroplasma. This project aims at better characterizing the Drosophila-Spiroplasma interaction, with particular focus on the bacterial side. In the first part, I developed a method to cultivate Spiroplasma poulsonii in vitro by optimizing a commercially available medium. This culture method allowed comparing the transcriptome of in vitro grown versus host-grown Spiroplasma, enabling us to identify putative genes involved in the interaction with the host. Interestingly, inside its insect host, S. poulsonii up-regulates genes coding for toxins of the Ribosomal Inactivating Protein (RIP) family. RIPs were previously known for their role in host protection against macro-parasites, such as wasps and nematodes. Their up-regulation in unparasitized hosts compared to culture was thus peculiar, raising the question what effect these RIPs have on host biology. Thus, in the second part, I studied the function of S. poulsonii RIPs in the absence of parasites. I showed that two of them are constantly expressed within the host and provide evidence that these toxins shorten host life span and increase embryonic mortality. Interestingly, the expression of RIPs was more toxic to male embryos than females, suggesting that RIPs contribute to S. poulsonii-induced male-killing. Last, I studied the D. melanogaster response to RIPs, and how the host mitigates the deleterious effects of these toxins by up-regulating the cytosolic chaperone Heat-Shock-Protein 70B (HSP70B). This protein carries out essential functions in protein homeostasis under normal and stressful conditions such as folding, refolding, or increasing the half-life of proteins. Interestingly, up-regulation of Hsp70B in the presence of RIPs results in an increased lifespan and in a better tolerance to heat stress, which may be ecologically advantageous. Altogether, this work illustrates how Spiroplasma-derived RIP toxins can differentially affect Drosophila depending on the ecological context, ranging from beneficial upon parasite infection or heat shock to detrimental in the absence of such environmental pressures.