Bacterial tolerance is a phenomenon observable in all known bacterial species, in which a sub-population exhibits tolerance to multiple stresses and hence high-level survival fitness under adverse environmental conditions, such as high concentration of antibiotics, starvation, oxidative stress and heat stress. In recent years, stress tolerant bacterial sub-populations have been shown to be responsible for the onset and exacerbation of recurrent and chronic bacterial infections, yet cellular mechanisms underlying onset of stress tolerance in bacteria remain undefined. This thesis describes research works aimed at investigating the range of bacterial physiological and genetic changes that may be elicited in response to starvation stress, and whether such responses confer antibiotic tolerance and enhance survival fitness in bacteria under adverse growth conditions. The thesis is depicting four projects I undertook during my doctorate study.

Briefly, our works showed that bacteria require proton motive force (PMF) to maintain a tolerance phenotype for a long period and that bacteria did not simply become dormant and shut down all physiological functions to withstand starvation. Our works confirm that bacteria actually over-express a range of proteins to maintain and generate PMF. Importantly, we also found that specific genetic changes occur during long term starvation, and that such changes enabled the tolerant sub-population, which are also known as persisters, to survive under extreme conditions. These findings suggest that active physiological and genetic changes in bacteria elicited in response to starvation are meaningful targets for development of strategies to combat recurrent and chronic infection diseases.

In Chapter 2, the tolerance phenotypes of starvation-induced persisters of different bacterial species subjected to treatment with different antibiotics were characterized. A study in our laboratory previously showed that physiological dormancy is not the only reason by which starvation could induce tolerance, and that a range of active tolerance responses are elicited in the E. coli strain BW25113 during tolerance formation. Tolerant sub-populations of this strain were found to survive the antimicrobial action of beta-lactam (ampicillin), aminoglycoside (gentamicin) and fluoroquinolone (ciprofloxacin) at up to ten times of MIC. Based on these observations, we also characterized the tolerance phenotypes in other bacterial pathogens such as Salmonella, Klebsiella pneumoniae, Acinetobacter baumannii and Pseudomonas aeruginosa. The results showed that long-term starvation induced formation of persisters of these bacterial persisters, but persisters of different species exhibited different levels of tolerance to the test antibiotics, indicating that they were susceptible to antibiotics to some extent, especially during long term treatment. All in all, out phenotypic data confirm that persisters are not physiologically dormant, otherwise they should be unresponsive to antibiotics throughout the course of treatment. As β-lactam antibiotics inhibit cell wall synthesis, we investigated whether cell wall synthesis activities were detectable in persisters. Using fluorescent labels to monitor cell wall synthesis, protein synthesis and DNA synthesis activities, our results showed that cell wall synthesis still occurred in persisters which had experienced starvation for six days. However, the fluorescence signals decreased gradually on the 6th day. This finding indicates that both of cell wall synthesis and DNA synthesis are active at the beginning of starvation but slows down gradually. Interestingly, the cell wall synthesis, protein synthesis and DNA synthesis activities of persisters of different kinds of bacteria were not identical. Fluorescent signals of strains of several bacterial species started to significantly decrease on the 3rd day, while that of other species decreased on the 6th day. Besides these, energy provision and efflux pumps were also proved to decrease under 6-day starvation, therefore, although antibiotic targets, cell wall synthesis, protein synthesis and DNA synthesis decreased, the antibiotic accumulation increased to help kill persisters under 6-day starvation.

In Chapter 3, we showed that maintenance and active generation of PMF are both important for tolerance formation. Since the phenomenon of bacterial tolerance was first discovered in 1944, the cellular mechanisms underlying tolerance development have been intensively studied but the key mechanism has not been identified. As our laboratory recently showed that a set of genes related to maintenance of proton motive force were up- regulated under starvation, we further investigated the active tolerance mechanisms and found that, apart from the pspA gene identified in the previous study, the rcsB gene and the osmC / bdm genes regulated by the rcsB gene product also played a role in maintenance of proton motive force and contributed to tolerance formation. Another group of proteins, namely the NuoL, Ndh, AppC, CyoB and NuoF proteins which are electron transport chain components, were also found to actively generate proton motive force in persisters to help maintain tolerance. As pspA is the most upregulated gene, double and triple knockout strains in which different combinations of pspA and the other tolerance genes were deleted were tested for changes in the tolerance phenotype. Tolerance assay, membrane potential assay and antibiotic accumulation assay were performed, with results confirming that these proteins act in combination to maintain and generate PMF in persisters. Therefore, PMF maintenance is regarded as an important mechanism for tolerance formation, especially during long term starvation.

In order to further investigate why PMF is required for tolerance formation, a series of membrane transporters, especially transporters of the major facilitator superfamily, were studied in experiments described in Chapter 4. The relative functional role of known transporters and efflux pumps in tolerance development was tested by assessing the effect of deletion of specific efflux pump and transporter-encoding genes on long-term maintenance of antibiotic tolerance in an Escherichia coli population under starvation. We identified eight specific efflux pumps and transporters and two known efflux pump components, namely, ChaA, EmrK, EmrY, SsuA, NhaA, GadC, YdjK, YphD, TolC, and ChaB, that play a key role in tolerance development and maintenance. In particular, deletion of each of the nhaA and chaB gene is sufficient to totally abolish the tolerance phenotype when the test population was subjected to prolonged antimicrobial treatment. These findings therefore depict existence of active, efflux-mediated bacterial tolerance mechanisms, and facilitate design of intervention strategies to prevent and treat chronic and recurrent infections due to persistence of antibiotic-tolerant sub-populations in the human body. This work also explains why antibiotic persisters need to maintain PMF. The reason is that PMF is required to support a range of efflux or transportation functions. Intriguingly, we found that tolerance-maintaining efflux activities were encoded by as many as ten efflux or transporter genes, yet deletion of only one of these genes could cause a significant reduction in tolerance level; we hypothesize that the product of each of these genes plays an essential role in enhancing the survival fitness of bacteria during starvation or adverse environmental conditions.

Apart from the active PMF maintenance mechanisms and specific efflux activities, we also discovered that unique genetic changes also occur in bacteria which exhibit antibiotic tolerance upon encountering starvation for a prolonged period. Such genetic changes allow persisters to exhibit significantly higher survival fitness. As described in Chapter 5 of this thesis, bacterial populations which exhibit stress tolerance are not expected to mutate; this is the key difference between tolerance and resistance, the latter usually involve genetic mutations or acquisition of exogenous resistance-encoding genetic elements. In this study, we found that under short term starvation bacteria did not undergo mutation, but upon encountering starvation for an extended period of time, the tolerant population started to mutate. This finding is unprecedented and novel. Importantly, mutations that occurred at specific sites in several genes were repeatedly found when bacteria were subjected to 6-days starvation. These genes, namely yhhI, rhsC and insH, have been poorly characterized and their functions are not well understood. Apart from a synonymous mutation in the insH gene, all detectable mutations resulted in amino changes. To test the effects of these mutations on bacterial survival fitness, three single mutants as well as double mutants were created, with the results showing that specific mutations in these three genes render bacteria to survive at a significantly higher rate under a range of adverse conditions.

In conclusion, bacteria actively express a range of cellular mechanisms to protect themselves against environmental stress and enhance survival fitness. These include express of proteins which help maintain PMF and some hitherto unknown site-specific mutagenesis mechanisms that cause mutational changes in a number of genes when bacteria encounter a prolonged period of starvation. These findings confirm that bacterial persisters are not merely physiologically dormant, and that cellular mechanisms that support phenotypic antibiotic tolerance are actively expressed by bacterial antibiotic persisters. Consistently, our data showed that persisters exhibit a certain degree of cell wall synthesis, protein synthesis and DNA replication activities, so that they become susceptible to various antibiotics unless PMF is well maintained to allow efflux activities to pump antibiotics out of the bacterial cells. These findings provide novel insight into the physiological and genetic basis of bacterial antibiotic tolerance and suggest that inhibition of active tolerance mechanisms is a novel and effective approach to eradicate antibiotic persisters.