Pseudomonas aeruginosa is a ubiquitous Gram-negative bacterium that causes devastating acute and chronic infections in a wide range of tissues. One site of major public health importance is in the cornea, where it is the most common causative pathogen for contact lens-associated infections, causing sight-threatening keratitis. The cornea is normally supremely capable at resisting infections by would-be pathogens, but contact lens wear, scratching injuries, or immunodeficiency may render it susceptible. The identities and functions of corneal defenses during health, and how they are impacted by lens wear, are not fully understood. When exposed to a susceptible cornea, however, P. aeruginosa is well-equipped to withstand the remaining defenses and traverse the multilayered epithelium, reaching the underlying stroma and triggering the inflammatory keratitis. Many studies on keratitis focus on the immune response to stromal bacteria, ignoring interactions between host and pathogen required for the microbe to reach that site. My project centered these interactions, after developing image analysis methods to explore what occurs when Pseudomonas aeruginosa encounters the corneal epithelium.The Fleiszig Lab has published on the use of an ex vivo murine model of infection, where eyes were tissue paper blotted, enucleated, and treated with calcium chelator EGTA, which permitted P. aeruginosa to adhere to and then traverse the otherwise resistant epithelium. This model differed from others used as it retained the natural structure of living tissue, forcing P. aeruginosa to interact with the corneal epithelial cells in order to progress infection. Using this model, the lab went on to show that P. aeruginosa wildtype strain PAO1 could fully traverse the corneal epithelium but not bypass the underlying basal lamina to enter the stroma. A 3D image analysis method was then developed for quantifying tissue penetration, which was used to show that ExsA, the transcriptional activator for the Type 3 Secretion System (T3SS), was required for bacteria to effectively traverse the susceptible ex vivo cornea. The T3SS assembles as a multimeric needle across the bacterial membranes, allowing translocation of effector proteins (exotoxins) through pores formed on target cell membranes. This allows bacteria to deliver these exotoxins directly to the cytoplasm of host cells, drastically altering their physiologies. PAO1 encodes three such proteins: ExoS, ExoT, and ExoY, which collectively act to disrupt cell polarity and tight junctions, promote intracellular survival and replication of the bacteria, and delay host cell death, among other activities. Although many of the findings for the relevance of T3SS factors in pathogenesis came from in vitro studies, several in vivo studies have shown their significance in the eye, lungs, and other sites. Still, which T3SS factors contributed to traversal of intact live corneal epithelium, or any other epithelial layer in the context of surrounding tissues, was unknown before my work.I first advanced the confocal image analysis techniques used previously to increase the granularity and accuracy of measurements of bacterial traversal. Using this method, I tested the traversal of mutants lacking the T3SS activator, translocon pore, needle, or all known effectors. I also studied mutants lacking a repressor of T3SS expression that are therefore constitutively T3SS-on. Surprisingly, I found that needle-deficient mutants could traverse as well as wildtype, while discovering roles for T3SS translocon pores, effectors, and regulation in promoting traversal. Expression of the effectors alone by ExsA complementation was found to be sufficient to increase median traversal depth of a mutant lacking all other T3SS components. Presence of ExoS was detected in the supernatant of this strain, suggesting a novel, T3SS needle-independent mechanism for expression and release of effector proteins. Possible mechanisms discussed include via packaging of T3SS components into OMVs (Outer Membrane Vesicles) or via host factor- or stress-induced bacteriolysis.Using this same model and analytical method, I also performed experiments building on observations of T3SS bistability within PAO1. A recent publication from the lab showed that, among an isogenic population of PAO1 and after induction of the T3SS, only half of the bacteria became T3SSHigh, the rest remaining with low expression of the T3SS. During traversal, we observed that the wildtype bacteria which were T3SSLow had traversed significantly deeper than the T3SSHigh wildtype subpopulation. Hypothesizing cooperation between these two phenotypes, I confirmed that co-infection in equal amounts with wildtype permitted ΔexsA mutants to effectively traverse the corneal epithelium. Further experiments revealed that, for a given strain to promote traversal of ΔexsA mutants ‘in trans’, the co-infecting strain must encode the translocon pore; the needle and exotoxins are not required. It remains to be tested if this factor is released by a similar needle-independent fashion as ExoS, and how the translocon proteins could affect the host epithelium in such a way to promote traversal by ΔexsA mutants in trans. In addition to the advancement and utilization of methods examining P. aeruginosa interactions with in situ corneal epithelium, I performed experiments to study these host-pathogen dynamics in vitro. As mentioned previously, it has been shown that PAO1 invades cells, and utilizes its T3SS to establish membrane bleb niches and promote intracellular survival, though the impact of this on traversal is unclear. Recent observations using infections of human corneal epithelial cells (hTCEpis) suggested that i.c. PAO1 may also express biofilm components. To confirm this, I developed an immunofluorescence technique for quantifying biofilm-reporting bacteria amongst the entire i.c. bacterial population. In the ensuing publication, we discovered that after invasion, wildtype diversified into the T3SSHigh subpopulation known to replicate in the cytoplasm and membrane bleb niches, but also into a T3SSLow, biofilm-producing, chronic state-like subpopulation which predominantly resided in vacuoles. This work established that these two phenotypes could co-exist within the same host cell and that the vacuolar population resisted antibiotics at a higher-than-clinical dose. We also showed that both subpopulations were found inside epithelial cells during in vivo murine infections. Separately, I contributed to work investigating host cell death to ΔexsA mutants (vacuole-restricted) or wildtype bacteria. Using a hTCEpi cell line I made with CRISPR-mediated knockout of Caspase-4, a pro-inflammatory immune sensor, we discovered that host cells can trigger pyroptosis via this non-canonical inflammasome to limit the growth of i.c. PAO1. Unpublished work included in this dissertation also includes studying the traversal of T3SS mutants through an in vitro multilayer of epithelial cells, a model the lab has previously used to show the requirement of bacterial type IV pili for traversal. Here, I present preliminary data showing that mutants lacking exsA, the T3SS needle, exotoxins, or translocon proteins all traverse significantly less than wildtype. Further exploration could explore the apparent differences between traversal of T3SS mutants through multilayer human corneal epithelial cells in vitro and the murine corneal epithelium ex vivo. I also developed an automated analysis pipeline on ImageJ to analyze the intracellular localization of P. aeruginosa during time-lapse in vitro infections. As discussed, PAO1 can diversify into vacuolar or cytoplasmic-resident subpopulations in epithelial cells. Previously, these two populations were manually separated by visual cues such as the small, round appearance of vacuolar bacteria and disorganized spreading of rod-like T3SSHigh bacteria in the cytoplasm. The new method allowed for the quantitative, reproducible distinction and measurement of both populations within live cells while also providing cell death and infection rates over time. Several manuscripts utilizing this approach are in preparation. I then used this method to study the importance of a secondary molecule known to control the acute-to-chronic state switch, cyclic di-GMP (cdG). Results showed that mutants with higher cdG triggered less cell death than wildtype and were found to spread more often in the cytoplasm than wildtype. Since the switch between T3SSHigh and T3SSLow involves complex regulation at many levels, more work is needed to characterize factors from both subpopulations contributing to relevant i.c. phenotypes.In summary, the results of my work provide novel insights into how P. aeruginosa interacts with the corneal epithelium and roles played by the T3SS. Throughout this dissertation, I have also helped develop a suite of imaging and analytical tools to answer research questions and allow others to continue along these lines of investigation. I showed that ExsA-mediated expression of the T3SS is required for bacterial traversal of live intact corneal epithelium, and involved both the exotoxins or translocon proteins, their contributions being additive and possible even in the absence of a functional needle. Bacterial traversal was also found to require regulation of the T3SS and to be possible through cooperation between diverse populations. Mutants with T3SS expression “Locked-On” were unable to traverse efficiently, and T3SSHigh bacteria promoted deeper penetration of the epithelium by T3SSLow bacteria through a yet-undescribed mechanism of the translocon proteins. Building on the importance of diversification for pathogenesis, I contributed work to in vitro studies showing that intracellular P. aeruginosa can exist in both a T3SSHigh state in the cytoplasm and a T3SSLow state in the vacuoles of host cells. Again, both of these subpopulations played distinct but complementary roles in virulence. The overlap in roles played by T3SS components and bacteria in different T3SS expression states shown for both intracellular and traversing P. aeruginosa strongly suggests the relatedness of these phenotypes, and helps build a theoretical model for how traversal occurs. The research discussed here, studying P. aeruginosa interactions with the corneal epithelium, ultimately highlights lessons of diversity and cooperation. The ability of the host tissue to defend against its myriad invaders requires a diverse number of defense mechanisms that must work in unison to maintain health and visual clarity. P. aeruginosa is such a successful pathogen because the infecting population adapts to nearly any condition, utilizes virulence factors in various manners, and splits into distinct phenotypes which all contribute some advantage. Ultimately, conducting this dissertation has required integrating a multidisciplinary approach with expertise, guidance, and methods from diverse sources.