Measuring energy levels in bacterial dormancy

Bacteria are evolving strategies to survive antibiotic treatments at a pace that is not matched by the one at which new drugs are discovered. Beyond the more notorious antimicrobial resistance, other survival mechanisms such as tolerance and persistence are today thought to play a major role in infections. Whether resistant, tolerant or persistent, cells that stop growing seem to have a survival advantage over replicating ones. These cells are conventionally referred to as dormant and very little is known about their physiology and the mechanistic reasons behind their remarkable survival capabilities. Because cell growth is intimately linked to cell physiological traits such as energy availability, this work seeks to investigate cellular energetics at the single cell level using E. coli as a model system. In particular the focus is addressed to two of the most important energy parameters in all life forms: ATP concentration and proton motive force (PMF). The PMF, further than participating in ATP synthesis, fuels a number of cellular processes that play prominent roles in cellular homeostasis. Regulation of each of its two components, the pH difference across the plasma membrane and the membrane voltage, is in turn essential for cell survival. Because cytoplasmic pH can be assayed at the single cell level with the genetically encoded fluorescent sensor pHluorin and PMF as a whole can be quantified from bacterial flagellar motor speed, the focus of this work was first addressed to membrane voltage estimation techniques. Nernstian reporters have in the past been used for the purpose, but their characterization never reached the depth of detail necessary for the measurement of membrane voltage of cells in physiological conditions. Using both an experimental and mathematical approach, I explored and described the parameter landscape in which these reporters can be used as sensors and when instead they influence cell physiology. Having built and validated such a preliminary interpretative framework, I formulated an algorithm for the characterization of novel dyes with respect to their interactions with the physiology of the cell. I applied the workflow to the characterisation of a Nernstian dye that had never been used before in E. coli. Although, in my conditions, none of the Nernstian dyes available were found suitable for (Vm) estimation, the workflow I developed is in the position to offer a simple and robust method to benchmark novel dyes and test the results obtained with old ones. ATP dynamics represent another fundamental aspect of cellular energetics and measurements at the single cell level have been sought for more than a decade. The most promising approach published suffered of low signal intensities that were not compatible with the exposure times required for time series measurements. By optimizing sensor expression and performing structural modifications, I obtained improvements in the signal intensity which rendered the sensor available to time lapse measurements. To further improve signal-to-noise ratio I installed a laser in our custom-built microscope. Coupling the ATP sensor with measurements of bacterial flagellar motor speed, which correlates to PMF, and the pH sensor pHluorin, I could investigate the main energy parameters of E. coli cells in vivo, in real time and with single cell resolution. To study these physiological traits in dormant cells I first established a definition of dormancy that lies on simple axioms such as viability and growth halt. I then individuated conditions capable to sharply induce dormancy, such the presence of bacteriostatic antibiotics, quorum sensing molecules or starvation. Opposed to the classical view that sees dormancy as an energetically poor state, my results show that the observation of growth arrest alone is scarcely informative on the physiological state of the cell. Dormant cells can be both high and low in energy, depending on the conditions that induced growth halt. While highly energetical cells might be better suited at surviving antibiotics via active means, scarcely energetical ones might have less targets to offer to antibiotics to carry out their function. In any of these scenarios, this work suggests that the environmental cues that lead to dormancy are likely to dramatically alter bacteria's susceptibility to the different antibiotic classes.