The Structure of Fitness Landscapes in Antibiotic Resistant Bacteria : : Molecular Origins and Evolutionary Consequences
- Author(s): Deris, John Barrett
- et al.
Antibiotic use is so ingrained in modern healthcare and agriculture that it can be difficult to imagine life in the pre-antibiotic era of the twentieth century. Many privileges we currently take for granted as rights, e.g., modern surgery, would not be possible without these drugs. But the rapid rise of antibiotic resistance may soon thrust the world back into this era. In order to predict (and ultimately prevent) the emergence of antibiotic resistance, it is crucial to establish quantitative, predictive links between the fitness of drug resistant organisms and the molecular mechanisms conferring resistance. Although the resistance mechanisms are often well characterized in vitro, their contributions to microbial fitness may depend critically on the environment and on the internal state of bacteria, which are often unknown in quantitative terms. To bridge this gap I investigate E. coli strains constitutively expressing resistance to translation-inhibiting antibiotics. The results show that in the presence of drugs, genes providing drug resistance are subject to an innate, positive feedback due to the global effect of drug- inhibited growth on gene expression. This feedback results in complex behaviors for isogenic populations of cells, including an abrupt drop in the growth rate of cultures at a threshold drug concentration. At drug concentrations below this threshold, cells exhibit growth bistability-- the coexistence of large populations of non-growing cells among otherwise identical, but growing, cells. This work demonstrates for the first time that many bacteria remain susceptible to an antibiotic even as they carry resistance to it. These behaviors do not appear in strains that lack drug resistance, and a quantitative characterization of drug-drug resistance interactions reveals a whole that is surprisingly richer than its parts. A mathematical model of bacterial growth based on the innate feedback predicts the onset of bistability and the growth rates of growing sub-populations, without invoking any ad hoc fitting parameters. Furthermore, the model describes a fitness landscape for bacterial drug resistance in different environments, allowing me to characterize the factors that determine the evolvability of resistance. The approach I use can be generalized to study resistance against other classes of antibiotics, besides the translation inhibitors studied here