For the systematic study of bacteri al virulence factors, we initially planned to screen the 12,000 mutant strains of the miniTn5-Sm tranposon-induced mutant bank in a wild-type S. marcescens strain Db10. Phagocytosis-deficient eater mutant flies (Kocks et al., 2005) were used in this screen to isolate the bacterial strains mutated for virulence factors and the genes responsible for crossing the gut barrier. In the eater mutant background, flies succumb to septicemia caused by the rapid proliferation of the bacteria in the hemolymph. Out of 1348 mutant strains screened, 58 candidate mutants have been isolated. Only 20% percent of the potential mutant strains displayed an increased virulence indicating that there are very few factor(s) that negatively control the virulence program of the bacterium. The fly survival phenotypes induced by the candidate mutants isolated in the first round of screen were retested. Only those bacterial strains that were consistent with the phenotype were chosen for the molecular identification of the transposon insertion sites using one primer PCR (Karlyshev et al., 2000). Once the genes impaired in each case had been identified, they were knocked off in the S. marcescens Db10 by site specific plasmid insertion mutagenesis. A mutant strain with the transposon inserted into the fliR gene, a component of the type III flagellar protein export system, exhibited attenuation of virulence. The plasmid insertionmutant strain generated to interrupt the gene fliR reproduced the fly survival phenotyp, indicating that the fliR gene is important for the virulence of S. marcescens. The fliR mutants are able to cross the peritrophic matrix, functionally similar to the human mucus. The bacteria were found in the vicinity of the epithelial cells but were not able to efficiently invade the intestinal epithelium as compared to the wild-type strain. Consequently lower titer of FliR mutants was found in the hemolymph. The inefficiency of the FliR mutants to invade cells was also confirmed in ex-vivo assay using insect cells.I thus demonstrated that the fliR gene which is important in the motility apparatus is also required by S. marcescens for the crossing of the epithelial barrier of D. melanogaster.[...]A strong oxidative response is triggered by D. melanogaster in the midgut against commensals and pathogens (Ha et al., 2009). In order to check whether the strong oxidative immune response is eventually killing the flies themselves, hydrogen peroxide was chemically neutralized in the midgut during the S. xylosus A. oral infection. No difference inthe fly survivals was observed with or without neutralization of the oxidative response indicating that over-production of reactive oxygen species (ROS) does not seem to be responsible for the fly death caused by a very low number of bacteria. Flies could efficiently survive to killed bacteria and filtered supernatant solution from overnight bacterial culture indicating that they do not die to the toxins released by the bacteria. Most surprisingly MyD88-, the Toll pathway-, mutant flies were surviving better to S. xylosus A. oral infection. A series of experiments lead us to the finding that the flies actually succumbed to starvation when orally infected with S. xylosus and that the MyD88 is required for the starvation susceptibility in microbiota-mediated manner. In conclusion my work has lead us to the better understanding of the host-bacterial interactions in the intestine.
