Population genetics of rifampicin-resistant Pseudomonas aeruginosa

Antibiotic resistance is generally associated with a cost in terms of reduced competitive fitness in the absence of antibiotics. Despite this 'cost of resistance', the cessation of antibiotic treatment does not result in significant reductions in the prevalence of resistance. The maintenance of resistance, in spite of the costs, has been attributed to the rarity of reversion mutations, relative to compensatory mutations at other loci in the genome. However, the large size of bacteria populations, and the potential for migration, suggest that reversion mutations should occasionally be introduced to resistant populations. In this thesis, I show that additional mechanisms can prevent fixation of reversion mutations even if they do occur. Using an experimental evolution approach, with rifampicin resistance in Pseudomonas aeruginosa as a model system, I measured the costs of resistance in several environments and followed the adaptive dynamics of resistant populations where a sensitive lineage had invaded by migration. The results suggest that several additional mechanisms contribute to the maintenance of antibiotic resistance. Most rifampicin resistance mutations are not unconditionally costly in all environments, suggesting that migration between environments could maintain a resistant reservoir population. In environments where resistance is initially costly, the fixation of a revertant is not guaranteed, even if introduced through migration. Revertant fixation was impeded or prevented by clonal interference from adaptation in the resistant strain. Revertants that did successfully replace the resistant strain were forced to adapt to do so. Contrary to assumptions in the existing literature, fitness in the resistant strains was not recovered by general compensatory mutations, but instead by adaptive mutations specific to the environment. The data challenge several assumptions about the maintenance of antibiotic resistance: that resistance mutations are always costly, that the rarity of back mutations prevents the reversion of resistance, and that resistant strains recover fitness by compensatory mutations.