Work

The Company They Keep: How Pathogens Interact, Compete and Persist

Year: 2026
Type: dissertation
Abstract: Bacteria are ubiquituous in the environment and play important roles in many of the processes that are necessary for human life. For example, they are involved in nutrient cycling by fixation of nitrogen, they aid food production by fermentation, and they support human health by shaping our gut microbiota. Such processes are rarely carried out by single species in isolation but happen often within complex communities in which many different bacteria interact with each other and their environment. Though most bacteria are benign or even beneficial, some bacteria can also be pathogenic. For example, in cystic fibrosis (CF) patients, bacterial infections of the lung are the leading cause of morbidity and mortality. These infections are typically polymicrobial with Pseudomonas aeruginosa, Staphylococcus aureus and Haemophilus influenzae predominating. In the CF lung, bacteria compete for limited resources and colonization sites. This leads bacteria to produce toxins and antimicrobial proteins to inhibit competitors and release nutrient-scavenging molecules such as siderophores to access iron more efficiently than competitors. At the same time, they can also exhibit cooperative behaviors that promote survival, for instance by establishing biofilms that protect the community from the host's immune system and from antibacterial treatment. Over extended time, the pathogens adapt in response to pressure from the host's immune system, interactions with other species and the environment within the lungs. This often leads to the transition from acute to chronic infection, in which virulence is reduced while persistence is increased. Although the evolutionary paths of pathogenic bacteria in adult CF patients have been well studied, much less is known about when such changes first emerge and how much they influence interactions within the bacterial community. To better understand the evolutionary trajectories observed in adults, it is essential to examine what occurs earlier in life during infections. Accordingly, my thesis aimed to track how the strain composition and phenotypic traits of P. aeruginosa, S. aureus, and H. influenzae isolated from pediatric CF patients vary over the first eight years of life, and how this variation affects competitiveness. These shifts can arise through two non-exclusive processes: (i) within-host evolution, where resident strains persist but adapt genetically and phenotypically; and (ii) strain turnover, where resident strains are replaced by newly acquired strains. It is often assumed that P. aeruginosa displaces early colonizers such as S. aureus and H. influenzae. However, with age only H. influenzae declines, while S. aureus frequently persists or even prevails, indicating that changing species abundance cannot be explained by simple ecological displacement. Because infection dynamics in cystic fibrosis are complex and remain poorly understood, I use two complementary approaches in my dissertation. First, I investigate what happens in patients by tracking how P. aeruginosa, S. aureus, and H. influenzae interactions change during early patient life. This approach is challenging but reflects real clinical conditions. Second, I conduct in vitro studies under controlled conditions to examine how S. aureus adapts to P. aeruginosa, which enables a deeper molecular analysis but relies on more artificial settings. By combining these two approaches, the thesis offers an integrative perspective. In my first project (Chapter 2), I analyzed clinical isolates of P. aeruginosa, S. aureus and H. influenzae obtained from pediatric CF patients. These isolates were collected as part of the Swiss CF Infant Lung Development (SCILD) cohort from oropharyngeal swabs of children aged 0–8 years. To identify changes in community composition and traits, I combined whole-genome sequencing, phenotypic screening and supernatant assays to assess their competitiveness. Although most phenotypes were highly heterogeneous across time, I found that some P. aeruginosa isolates from children older than five years exhibited significantly reduced protease activity. Whole-genome data further revealed that these isolates carried single nucleotide polymorphisms (SNPs) in one or more quorum-sensing genes. These results indicate that selection for quorum-sensing mutants, previously observed in adult CF isolates, is already occurring in pediatric patients. As P. aeruginosa changes, the interaction network shifts, with reduced protease production correlating with weaker inhibition of S. aureus and H. influenzae. Together, these findings indicate that the selection for quorum-sensing mutants in P. aeruginosa begins early in cystic fibrosis infection and that this change alters interactions with other strains, while traits such as biofilm formation and siderophore production show no consistent changes during early childhood. In my second project (Chapter 3), I contributed to a study initiated by Dr. Selina Niggli, a former PhD student in the lab of Prof. Dr. Rolf Kümmerli. We investigated how S. aureus strains from different genetic backgrounds adapt to growth inhibitory secreted toxins of P. aeruginosa. To examine how S. aureus adapts to these toxins over time, we conducted an evolution experiment in which the three S. aureus strains (JE2, Cowan I and 6850) were serially passaged for one month in the presence or absence of P. aeruginosa supernatant. Independent populations from all three strains rapidly evolved resistance and no longer exhibited growth inhibition, indicating that adaptation happens quickly. Interestingly, S. aureus further showed strain-specific phenotypic changes, suggesting that each genetic background followed an independent adaptive trajectory. To investigate the genetic changes of these adaptations, we then performed whole-genome sequencing on 150 phenotypically characterized clones. Most mutations were specific to the treatment condition and unique to each strain background. Notably, many of the recurrently mutated genes were associated with membrane transporters. Given that S. aureus membrane transporters are known to regulate virulence and antibiotic resistance, our finding that P. aeruginosa secreted inhibitory compounds select for mutations in these genes is of particular clinical relevance. This highlights how interspecies interactions may directly shape the evolutionary trajectories of pathogens. In my third project (Chapter 4), I analyzed an evolved S. aureus JE2 clone from Chapter 3, which was selected in P. aeruginosa supernatant and carried a single mutation in the gene tcyA, a cysteine transporter subunit. This locus was of particular interest because tcyA mutations emerged in four out of five independently evolved populations, suggesting a key adaptive role in the interaction with P. aeruginosa. I first compared the evolved clone with the ancestor using proteomics and found significantly reduced TcyA and TcyB synthesis, which provided moderate protection against PQS and selenocystine (genotypic effect). I then analyzed the proteome of JE2 grown in P. aeruginosa supernatant versus TSB control and observed much broader changes (phenotypic effect). Virulence factors and multiple transporter pathways were downregulated, while reactive oxygen species scavengers and a multidrug efflux pump were upregulated. Exposure to P. aeruginosa supernatant increased antibiotic tolerance, including a higher IC90 to ampicillin. These results support the hypothesis that resistance arises from both genetic change and phenotypic plasticity. The tcyA mutation contributes, but the larger phenotypic response provides rapid protection in fluctuating environments and reduces the need for extensive hard-wired genetic adaptation. This work links a specific genetic change to increased survival while showing that environment-driven proteomic shifts account for most of the adaptive benefit, and it may generalize to other bacterial competitions where temporary stresses favor rapid responses over fixed genomic mutations. In the projects described above, I primarily used supernatant assays to investigate the inhibitory effects that one bacterial strain can have on another one. While this approach is useful for determining whether a strain produces inhibitory compounds, it does not capture the dynamics of interactions, since the strains cannot respond to each other in real time. To overcome this limitation, in project 4 (Chapter 5) I developed a novel co-culture plate that connects two to twelve wells through a semi-permeable membrane. This allows metabolite diffusion whilst keeping the strains physically separated. In this way, the plate enables controlled interaction studies without cross-contamination risk. An additional benefit is that individual cultures, although connected, can be easily tracked by optical density (OD), fluorescence, or luminescence since they are not mixed. The plate is easy to integrate into standard laboratory workflows and provides a versatile platform for studying bacterial interactions. Beyond this, we demonstrated that it can be applied to a wide range of experimental setups, including phage–bacteria interactions and cultures with mammalian cell lines. In summary, I showed that combining controlled in vitro experiments with work on clinical isolates is a powerful way to study pathogen interactions. Clinical isolates allowed me to study the real changes happening in pediatric cystic fibrosis patients, whereas the in vitro studies revealed specific mechanisms that can drive those changes. By analyzing these isolates, I found that P. aeruginosa already carries quorum-sensing mutations and shows reduced protease activity in early CF patient life, changes that alter its interactions with co-occurring bacteria. In parallel, experimental evolution demonstrated that S. aureus adapts rapidly to P. aeruginosa’s inhibitory metabolites through both strain-specific genetic changes and phenotypic plasticity. Building o
Language: English
Open Access status: green
OpenAlex ID: W7138118263
Cites:
Cited by:
Related to: 10