PhD Defense by Mikkel S. Svenningensen

Title: Stochastic Bacterial Growth and Survival

Abstract: An isogenic bacterial population exhibits huge cell-to-cell variation in survival times during treatment with antibiotics. Most cells are eradicated within a generation time, whereas a few cells survive surprisingly long. These long-term surviving cells are referred to as bacterial persisters. The phenomenon of bacterial persistence is investigated in this thesis from different angles with four different projects. The approach is both theoretical and experimental.

The first project dealt experimentally with persisters measured by fluorescent timelapse microscopy. The persisters were triggered by a temperature-sensitive allele of the valS gene, raising the intracellular (p)ppGpp level as a function of the temperature. The intracellular levels of (p)ppGpp and the RelE toxin were measured with fluorescent reporters. We tried to establish whether the stochastic formation of persisters originated from fluctuations in either the (p)ppGpp level or the RelE level. We showed a correlation between fluctuations in the RelE reporter and persistence. However, removing ten type II mRNase toxin-antitoxin systems, including the RelBE system, from our wildtype strain, did not change the persister fraction in batch cultures.

The second project was a theoretical study of protein distributions. In general, the tails of probability distributions are very sensitive to small changes in the distribution mean. We showed this tail-effect in cell-to-cell protein distributions, by showing that the probability for rare protein concentrations depended strongly on small changes in the population mean. We demonstrated how the left-tail was very dependent on whether the mean changed due to the transcription or the translation rate. We proposed how the tail-effect could be related to persistence.

In the third project, we experimentally showed the presence of week-long survivors formed during balanced exponential growth. We then perturbed the killing dynamics by changing the carbon source in the growth medium and by removing the gene relA, which had a surprisingly small effect on the long-term killing dynamics. However, a glucose downshift prior to the addition of antibiotics significantly increased the persister fraction for at least four days. This effect was shown to be dependent on relA.

The fourth project was a theoretical investigation of the origins of heterogeneity in doubling time distributions. We proposed a simple growth model that produces a subpopulation of extremely slow-growing cells. The slow growth corresponds to collapses in the central metabolism and it could be interpreted as a persister state.


Meeting ID: 626 9407 9045
Passcode: 014247