Microbial adaptation to antibiotic treatment, both in the lab and the clinic

Despite the use of antibiotics in modern medicine for nearly a century, there is still much to learn about how bacteria respond and adapt to these commonly used drugs. Furthermore, antibiotic treatments in patients can span weeks or even months, therefore it is crucial to understand how pathogens evolve to antibiotics within the human body over time. This thesis comprises two parts, with the first exploring the impact that antibiotics have on bacterial growth and viability during nutrient starvation through experimental evolution and the second investigating the genomes of multidrug-resistant pathogens acquired from patients suffering with chronic infection. Initially in chapter two we demonstrate how ribosome-targeting antibiotics such as doxycycline and erythromycin can actually be beneficial for Escherichia coli, both through the stimulation of growth and improved long-term viability in an environment starved of nutrients. This is not true for all antibiotics though, as antibiotics with alternative cellular targets (rifampicin and penicillin) do not confer the same benefits. Given that antibiotics are primarily associated with negative impacts on bacterial growth, this is a surprising finding and we sought to identify the mechanism. Whole genome sequencing was performed on doxycycline-exposed populations of E.coli as they went through starvation, as well as the use of a GFP-tagged promoter library to study short-term metabolic changes occurring as nutrients are depleted. Following this, we demonstrate the influence that ribosome production has on long-term viability through the use of E.coli rrn knockout strains. Through doing so, we show that interference with ribosome functioning can improve long-term viability in nutrient starved environments, mirroring the effects of ribosome-targeting antibiotics. Chapter four is dedicated to understanding the effects of genomic background on doxycycline-induced benefits in E.coli. Through the use of the keio library, over 2,500 individual growth curves were generated, consisting of growth data in both the presence and absence of doxycycline. Principle component analysis was subsequently carried out, and groups of strains associated with particular phenotypes were assigned into functional categories. Through doing so we uncover some unexpected phenotypes, for example E.coli knockout strains that are only able to grow in media containing doxycycline. Finally, chapters five and six explore the genomes of pathogenic bacteria during long-term antibiotic treatment. Bacteria evolve resistance towards antibiotics within laboratory experiments in a matter of days, but how do pathogens adapt to repeated antibiotic treatment within the human body? Klebsiella pneumoniae and E.coli isolates spanning 18 and 26 months respectively were acquired from two different patients and characterised through the use of nanopore sequencing and phenotypic studies. Nanopore sequencing allowed the resistomes of these pathogens to be elucidated, as well as genomic variations and structural changes. Whilst the K.pneumoniae isolates were found to be clonal, thereby allowing genomic changes to be tracked over time, the E.coli isolates were found to belong to different clonal groups. Comparative genomics was therefore carried out on these isolates to assess the genomic variability between isolates and determine the genetic basis for antibiotic resistance. In summary, this thesis describes a novel, beneficial effect of ribosome-targeting antibiotics on long-term bacterial viability. Whilst an antibiotic may be effective in the short-term, over long periods it may actually stimulate bacterial growth. This has implications not only on the clinical use of antibiotics, but also on our understanding of natural antibiotic production in the environment. Additionally, this thesis builds upon our knowledge of within-host evolution of pathogens and demonstrates the ability of nanopore sequencing to elucidate resistomes. Alongside phenotypic data, genomics can be a powerful tool in determining the antimicrobial resistance profile of pathogens, particularly in complex clinical cases such as those presented here.