Selection of Bacterial Mutants in Late Infections: When Vector Transmission Trades Off against Growth Advantage in Stationary Phase.

Bacterial infections are often composed of cells with distinct phenotypes that can be produced by genetic or epigenetic mechanisms. This phenotypic heterogeneity has proved to be important in many pathogens, because it can alter both pathogenicity and transmission. We studied how and why it can emerge during infection in the bacterium Xenorhabdus nematophila, a pathogen that kills insects and multiplies in the cadaver before being transmitted by the soil nematode vector Steinernema carpocapsae. We found that phenotypic variants cluster in three groups, one of which is composed of lrp defective mutants. These mutants, together with variants of another group, have in common that they maintain high survival during late stationary phase. This probably explains why they increase in frequency: variants of X. nematophila with a growth advantage in stationary phase (GASP) are under strong positive selection both in prolonged culture and in late infections. We also found that the within-host advantage of these variants seems to trade off against transmission by nematode vectors: the variants that reach the highest load in insects are those that are the least transmitted. IMPORTANCE Pathogens can evolve inside their host, and the importance of this mutation-fueled process is increasingly recognized. A disease outcome may indeed depend in part on pathogen adaptations that emerge during infection. It is therefore important to document these adaptations and the conditions that drive them. In our study, we took advantage of the possibility to monitor within-host evolution in the insect pathogen X. nematophila. We demonstrated that selection occurring in aged infection favors lrp defective mutants, because these metabolic mutants benefit from a growth advantage in stationary phase (GASP). We also demonstrated that these mutants have reduced virulence and impaired transmission, modifying the infection outcome. Beyond the specific case of X. nematophila, we propose that metabolic mutants are to be found in other bacterial pathogens that stay for many generations inside their host.