Urinary tract infections (UTIs) are predominantly caused by uropathogenic Escherichia coli (UPEC). UPEC pathogenesis requires passage through a severe population bottleneck involving intracellular bacterial communities (IBCs) that are clonal expansions of a single invading UPEC bacterium in a urothelial superficial facet cell. IBCs occur only during acute pathogenesis. The bacteria in IBCs form the founder population that develops into persistent extracellular infections. Only a small fraction of UPEC organisms proceed through the IBC cycle, regardless of the inoculum size. This dramatic reduction in population size precludes the utility of genomic mutagenesis technologies for identifying genes important for persistence. To circumvent this bottleneck, we previously identified 29 positively selected genes (PSGs) within UPEC and hypothesized that they contribute to virulence. Here, we show that 8 of these 29 PSGs are required for fitness during persistent bacteriuria. Conversely, 7/8 of these PSG mutants showed essentially no phenotype in acute UTI. Deletion of the PSG argI leads to arginine auxotrophy. Relative to the other arg genes, argI in the B2 clade (which comprises most UPEC strains) of E. coli has diverged from argI in other E. coli clades. Replacement of argI in a UPEC strain with a non-UPEC argI allele complemented the arginine auxotrophy but not the persistent bacteriuria defect, showing that the UPEC argI allele contributes to persistent infection. These results highlight the complex roles of metabolic pathways during infection and demonstrate that evolutionary approaches can identify infection-specific gene functions downstream of population bottlenecks, shedding light on virulence and the genetic evolution of pathogenesis.
IMPORTANCE Uropathogenic Escherichia coli (UPEC) is the most common cause of human urinary tract infection (UTI). Population bottlenecks during early stages of UTI make high-throughput screens impractical for understanding clinically important later stages of UTI, such as persistence and recurrence. As UPEC is hypothesized to be adapted to these later pathogenic stages, we previously identified 29 genes evolving under positive selection in UPEC. Here, we found that 8 of these genes, including argI (which is involved in arginine biosynthesis), are important for persistence in a mouse model of UTI. Deletion of argI and other arginine synthesis genes resulted in (i) arginine auxotrophy and (ii) defects in persistent UTI. Replacement of a B2 clade argI with a non-B2 clade argI complemented arginine auxotrophy, but the resulting strain remained attenuated in its ability to cause persistent bacteriuria. Thus, argI may have a second function during UTI that is not related to simple arginine synthesis. This study demonstrates how variation in metabolic genes can impact virulence and provides insight into the mechanisms and evolution of bacterial virulence.