Resistance is futile: The evolution of antibiotic resistance in Escherichia coli and Mycobacterium tuberculosis.

摘要

Antibiotic resistance is a growing worldwide problem. It has been suggested that by removing the selective pressure of the antibiotic, sensitive strains may re-establish themselves in the population, and now-obsolete antibiotics could be useful again. To understand this process, we must explore the "cost of resistance"---the reduction in competitive ability when antibiotic-resistant bacteria are co-cultured with their antibiotic-sensitive ancestors. This idea is complicated by the fact that the cost can be ameliorated through subsequent compensatory evolution. This dissertation examines the cost of antibiotic resistance and subsequent compensatory evolution of that cost using three bacteria-antibiotic systems, in four chapters. E.coli and triclosan is used to quantify the cost of resistance and determine if compensatory evolution is possible in the case of a compound different from most antibiotics---one used widely in the community and stable in the environment. E.coli and rifampin enables an in-depth examination of the cost of resistance and the process of compensatory evolution in a well-studied system. M. tuberculosis and rifampin is used to examine the cost of resistance and the possibilities of compensatory evolution in a human pathogen, and the relevance of laboratory competition assays to the real world experienced by bacteria. Chapter One demonstrates that the cost of resistance to triclosan is variable, is lowest in fatty acid biosynthesis mutations, and can be ameliorated over 200 generations. Chapter Two establishes a genotype-by-environment interaction for rifampin resistance, which extends from the specific mutation to the tertiary structure of the RNA polymerase. Chapter Three explores an interaction between the specific rifampin-resistance conferring mutation and population bottlenecks on the process of compensatory evolution. Chapter Four determines that the least-costly resistance mutation for M. tuberculosis in the laboratory is the most commonly found mutation clinically, indicating that the laboratory assay captures essential elements of the competitive environment in humans. For clinical isolates, these same mutations carry an even lower cost than observed in strains selected in vitro. This may be the result of compensatory evolution occurring in a single patient. This group of chapters brings together both the unifying themes and points out unique issues in each system studied.