Dmm019901 473..486

Studying infectious diseases requires suitable hosts for experimental in vivo infections.Recent yearshaveseen theadvent ofmanyalternatives to murine infection models. However, the use of non-mammalian models is still controversial because it is often unclear howwell findings from these systems predict virulence potential in humans or other mammals. Here, we compare the commonly used models, fruit fly and mouse (representing invertebrate and mammalian hosts), for their similarities and degree of correlation upon infection with a library of mutants of an important fungal pathogen, the yeast Candida glabrata. Using two indices, for fly survival time and for mouse fungal burden in specific organs, we show a good agreement between the models. We providea suitable predictivemodel forestimating the virulence potential ofC. glabratamutants in themouse from fly survival data. Asexamples, we found cell wall integrity mutants attenuated in flies, andmutants of a MAP kinase pathway had defective virulence in flies and reduced relative pathogen fitness in mice. In addition, mutants with strongly reduced in vitrogrowth generally, but not always, had reduced virulence in flies. Overall, we demonstrate that surveying Drosophila survival after infection is a suitable model to predict the outcome of murine infections, especially for severely attenuated C. glabrata mutants. Pre-screening of mutants in an invertebrate Drosophila model can, thus, provide a good estimate of the probability of finding a strain with reduced microbial burden in the mouse host.