Pseudomonas aeruginosa est une bactérie à Gram négatif à caractère ubiquitaire que l’on retrouve dans une grande diversité d’environnements. C’est un pathogène opportuniste qui est responsable chez l’homme d’infections chroniques ou aigües qui peuvent être mortelles pour des patients immuno-déficients. L’établissement d’une infection chronique est généralement associé à la capacité de la bactérie à former un biofilm, qui se définit comme une population bactérienne attachée sur une surface et englobée par une matrice extracellulaire formée entre autre depolysaccharides. La formation du biofilm est un processus bien défini dans le temps et dans l’espace et qui implique la mise en jeu de nombreuses structures de surfaces dont l’assemblage est strictement contrôlé. Une des voies de régulation contrôlant cet assemblage est le système à 2composants Roc1 (« regulation of cup genes »). Les gènes cup codent des composants de la voie « chaperone-usher » qui permet le transport de sous-unités pilines et leur assemblage à la surface bactérienne sous forme de pili. Ces pili Cup sont important dans l’établissement du biofilm. Le système Roc1 est aussi impliqué dans la mise en place du système de sécrétion de type III, qui est communément associé aux infections aigues. De fait le système Roc1 peut être considéré comme un «interrupteur» décidant du mode d’infection associé à P. aeruginosa. Le système Roc1 est constitué d’un senseur non-orthodoxe (RocS1) et de deux régulateurs de réponse, RocA1 et RocR, dont le domaine effecteur est un domaine de liaison à l’ADN ou un domaine EAL à activité phosphodiesterase, respectivement. Il existe également d’autres gènes paralogues de Roc1 qui sont le système Roc2 avec RocS2 et RocA2 très similaire à RocS1 et RocA1, ainsi que RocS3 similaire à RocS1. Le travail réalisé au cours de ma thèse a montré qu’il existe une régulation croisée entre Roc1 etRoc2. Cependant, chacune des branches du réseau de régulation contrôle l’expression d’une série de gènes bien spécifiques. Nous avons montré que la signalisation via RocS2 et RocS1 lorsqu’elle converge sur RocA1 contrôle l’expression des gènes cupC et ce contrôle est totalement indépendantde RocA2. Par contre lorsque la signalisation RocS1 et RocS2 converge vers RocA2 alors ce sont les gènes mexAB-oprM, qui codent une pompe d’efflux impliquée dans la résistance aux antibiotiques, dont l’expression est alors réprimée.En conclusion, nous avons mis en évidence un modèle unique de régulation croisée qui résulte dans un effet antagoniste entre formation du biofilm et résistance aux antibiotiques. Si cela peut paraître inattendu, quelques données cliniques sont en faveur d’une telle balance. En effet, l’analyse de souches de P. aeruginosa, isolées à partir de patients atteints de mucoviscidose, révèle que dans ces isolats la pompe MexAB-OprM est inactive. La raison de cette adaptation n’est pas élucidée, mais l’absence de pompe fonctionnelle pourrait procurer un avantage, une meilleure aptitude à la souche à persister dans cet environnement. Il est également reconnu que dans les poumons de ces patients le mode préféré de développement pour P. aeruginosa est le biofilm. Mises bout à bout ces observations suggèrent donc que le système Roc pourrait être un système de régulation important pour percevoir l’environnement du poumon chez le patient mucoviscidosique et déclencher une réponse adaptée. / The opportunistic pathogen Pseudomonas aeruginosa is responsible for diverse chronic and acute infections in human. Chronic infections are associated with the capacity of P. aeruginosa to form biofilms. One of the pathways controlling biofilm formation is the Roc1 two-component system, involved in the regulation of cup genes allow the assembly of thin fimbriae at the surface of the bacterium. Cup fimbiae are important in biofilm formation. There exist paralogues of the Roc1 system - the Roc2 and Roc3 system. The work in this thesis has shown that cross-regulation occurs between Roc1 and Roc2. However, each branch in this network appears to control the expression of a specific subset of genes whose role and functions are striking in the context of an infection process. We characterized here a unique model of cross-regulation which results in the antagonistic regulation of biofilm formation and antibiotic resistance

Alginate overproduction by Pseudomonas aeruginosa, which causes a mucoid phenotype, is a major virulence factor associated with chronic pulmonary infections in cystic fibrosis patients. Expression of the algD operon for alginate biosynthesis requires three major regulators in association with the ECF sigma factor, σ22, in mucoid strains that are typically defective in anti-sigma factor, MucA. One such algD regulator is AlgB, a member of the NtrC family of two-component systems, which typically utilize σ54. However, neither σ54 nor the cognate sensor kinase (KinB) of AlgB are required for algD expression in such mucoid strains. I hypothesized that KinB-phosphorylated AlgB must play some role in gene regulation, and so I sought to construct a constitutively active AlgB that simulated kinase-phosphorylation. I took a predictive approach and genetically introduced substitutions in AlgB that had been shown to activate DctD, a close homologue of AlgB in Rhizobium (52). When one such substitution, AlgBE125K, was transferred to a nonmucoid P. aeruginosa PAO ΔalgB-kinB (JK159) strain, alginate overproduction was observed. Interestingly, introduction of an algT mutation to remove σ22 did not block alginate production induced by AlgBE125K; although, it did stimulate the production of alginate in the presence of AlgBwt in trans to similar levels induced by the constitutive mutant. In contrast, introduction of an rpoN mutation showed that alginate production mediated by AlgBwt and AlgBE125K was σ54 dependent. The increase in expression of alginate by AlgBwt in the presence of σ54 and the absence of σ22 suggested a competition between the sigma factors for binding to PalgD. Biochemical assays were conducted to assess the constitutive property of AlgBE125K. For the ATPase assay, an equivalent amount of ATP hydrolysis was observed between the mutant and the wild type AlgB proteins. Slight differences seen for the EMSA data suggested possible higher order complex formation for AlgBE125K compared to AlgBwt. Collectively, these results suggested that in wild-type (MucA+) P. aeruginosa, expression of the algD operon is dependent on the phosphorylation of AlgB by KinB in a typical two-component fashion that is triggered by some as yet unknown environmental stimulus.

Two component systems (TCSs) are regulatory pathways in bacteria and lower eukaryotes that integrate multiple stimuli and bring about appropriate responses to promote adaptation of the bacteria to their niches. They are commonly insulated from cross-talk and form discrete regulatory systems where the sensor histidine kinase (SK) and the response regulator (RR) share high fidelity for one another. The GacS network controls the switch between acute and chronic virulence of P. aeruginosa. The network is unusual in having a 'core' SK, GacS, which is modulated directly by one other SK, RetS. Here the complex relationship between GacS and RetS is dissected to reveal three distinct mechanisms by which they interact. Two of these mechanisms involve the dephosphorylation of GacS-P by RetS and it is these mechanisms that are important in vivo for the regulation of biofilm formation, rsmY and rsmZ expression, swarming, and virulence in both Galleria mellonella and an acute model of infection in mice. This study reveals an unprecedented level of complexity in the ability of RetS to interact with GacS and suggests that RetS has a number of mechanisms by which it can downregulate the GacS network output. Furthermore, the interactions of additional SKs that have previously been linked to the GacS network were investigated. Here I demonstrate that many of these kinases can interact with one another but that RetS remained the only kinase tested that could directly interact with GacS. The interactions observed between kinases could be either stimulatory, having a synergistic impact on phosphorylation levels, or inhibitory. I also show that kinase-kinase interactions allow for the regulation of phosphorylation of downstream proteins. Finally, we searched for additional SKs that may be able to interact with the GacS network. Here I identify three new kinases, which show differing interactions with the kinases of the GacS network. The discovery of additional SKs in the GacS network indicates that the network is likely to respond to a far greater number of different signals than previously realised as it decides between acute and chronic virulence.

abstract: Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis, is the 10th leading cause of death, worldwide. The prevalence of drug-resistant clinical isolates and the paucity of newly-approved antituberculosis drugs impedes the successful eradication of Mtb. Bacteria commonly use two-component systems (TCS) to sense their environment and genetically modulate adaptive responses. The prrAB TCS is essential in Mtb, thus representing an auspicious drug target; however, the inability to generate an Mtb ΔprrAB mutant complicates investigating how this TCS contributes to pathogenesis. Mycobacterium smegmatis, a commonly used M. tuberculosis genetic surrogate was used here. This work shows that prrAB is not essential in M. smegmatis. During ammonium stress, the ΔprrAB mutant excessively accumulates triacylglycerol lipids, a phenotype associated with M. tuberculosis dormancy and chronic infection. Additionally, triacylglycerol biosynthetic genes were induced in the ΔprrAB mutant relative to the wild-type and complementation strains during ammonium stress. Next, RNA-seq was used to define the M. smegmatis PrrAB regulon. PrrAB regulates genes participating in respiration, metabolism, redox balance, and oxidative phosphorylation. The M. smegmatis ΔprrAB mutant is compromised for growth under hypoxia, is hypersensitive to cyanide, and fails to induce high-affinity respiratory genes during hypoxia. Furthermore, PrrAB positively regulates the hypoxia-responsive dosR TCS response regulator, potentially explaining the hypoxia-mediated growth defects in the ΔprrAB mutant. Despite inducing genes encoding the F1F0 ATP synthase, the ΔprrAB mutant accumulates significantly less ATP during aerobic, exponential growth compared to the wild-type and complementation strains. Finally, the M. smegmatis ΔprrAB mutant exhibited growth impairment in media containing gluconeogenic carbon sources. M. tuberculosis mutants unable to utilize these substrates fail to establish chronic infection, suggesting that PrrAB may regulate Mtb central carbon metabolism in response to chronic infection. In conclusion, 1) prrAB is not universally essential in mycobacteria; 2) M. smegmatis PrrAB regulates genetic responsiveness to nutrient and oxygen stress; and 3) PrrAB may provide feed-forward control of the DosRS TCS and dormancy phenotypes. The data generated in these studies provide insight into the mycobacterial PrrAB TCS transcriptional regulon, PrrAB essentiality in Mtb, and how PrrAB may mediate stresses encountered by Mtb during the transition to chronic infection. / Dissertation/Thesis / Doctoral Dissertation Microbiology 2019

Phosphate is an essential compound for life. Escherichia coli employs a signal transduction pathway that controls the expression of genes that are required for the high-affinity acquisition of phosphate and the utilization of alternate sources of phosphorous. These genes are only expressed when environmental phosphate is limiting. The seven genes for this signaling pathway encode the two-component regulatory proteins PhoB and PhoR, as well as the high-affinity phosphate transporter PstSCAB and an auxiliary protein called PhoU. As the sensor kinase PhoR has no periplasmic sensory domain, the mechanism by which these cells sense environmental phosphate is not known. This paper explores the hypothesis that it is the alternating conformations of the PstSCAB transporter which are formed as part of the normal phosphate transport cycle that signal phosphate sufficiency or phosphate limitation. We tested two variants of PstB that are predicted to lock the protein in either of two conformations for their signaling output. We observed that the pstBQ160K mutant, predicted to reside in an inward facing, open conformation signaled phosphate sufficiency whereas the pstBE179Q mutant, predicted to reside in an outward facing, closed conformation signaled phosphate starvation. Neither mutant showed phosphate transport.

L’évaporation du carburant injecté dans une chambre de combustion est un phénomènecrucial dans un foyer aéronautique car elle détermine la quantité de vapeur qui sera ensuite brûléepour fournir de l’énergie au moteur. Cependant, ce phénomène reste mal décrit du fait desdifficultés de mesurer expérimentalement les taux d’évaporation des gouttes appartenant à unbrouillard. D’autre part, les hypothèses des modèles théoriques ne sont toujours pas représentativesdes conditions rencontrées dans les foyers de combustion.La thèse s’inscrit dans une démarche visant à étudier l’évaporation d’un groupe de gouttesdans un écoulement turbulent au moyen de la Simulation Numérique Directe (SND). En effet, lorsde l’évaporation de groupes de gouttes, des effets collectifs peuvent influer sur le taux d’évaporationde chaque goutte ou sur le taux d’évaporation global du nuage de gouttes. L’approche SNDpermet de quantifier précisément ces effets afin d’améliorer les modèles actuels d’évaporation.Dans un premier temps, des algorithmes ont été développés et utilisés dans une configuration1D sphérique pour décrire l’évaporation d’une goutte statique isolée et sans gravité. Puisdans un second temps, l’évaporation d’une goutte a été étudiée dans un écoulement laminaire.Une analyse des échanges de chaleur entre la goutte et le milieu extérieur ainsi que de la force detraînée exercée par l’écoulement laminaire sur la goutte a été effectuée. Dans cette partie, il a étémis en évidence que l’évaporation induit une diminution des échanges thermiques et de la traînée.Il a notamment été observé que dans certains cas de forte évaporation, la traînée de la gouttepeut devenir négative. Cela implique que l’évaporation peut être à l’origine d’un phénomène depropulsion de la goutte. Une analyse théorique permet de lier ce comportement à une asymétriedu débit d’évaporation. Dans un troisième temps, l’influence de la turbulence sur l’évaporationd’une goutte a été étudiée. Pour cela, un générateur de fluctuations turbulentes a été implémentéet des techniques de calculs parallèles ont été introduites pour réduire le temps des calculs. Celaa permis d’analyser les échanges thermiques et le comportement de la traînée d’une goutte eninteraction avec un écoulement turbulent. Il a été montré que ces deux grandeurs ont tendanceà être amplifiées par la turbulence. Enfin, dans un dernier temps, l’évaporation de groupes degouttes a été étudiée. Pour trois groupes de gouttes différents, les déplacements des gouttes ontété analysés avec les échanges de chaleur lorsque ceux-ci sont placés dans un écoulement laminaireou turbulent avec ou sans changement de phase. En présence d’évaporation, il a été mis enévidence que les déplacements sont différents des cas sans évaporation et donc que le changementde phase modifie les effets collectifs. De plus, ces effets de groupes ont aussi été observés sur leséchanges thermiques. / The vaporisation of injected fuel in a combustion chamber is a crucial phenomenon inan aeronautical motor because it determines the vapour quantity which will be burned to bringenergy to the motor. Still, this phenomenon is not well understood due to the difficulties tomeasure on experiments vaporisation rates of injected sprays. Moreover, hypothesis of theoriticalmodels are not representatives of conditions encountered in combustion furnaces.The thesis take place in an effort to analyse the evaporation of droplet groups in a turbulentflow by mean of Direct Numerical Simulation (DNS). Indeed, during droplet group evaporation,collective effects can modify single droplet rates of vaporisation and the group global rate ofvaporisation. The DNS approach should allows to quantify precisely this effect and leads to animprovement of actual models of evaporation.Firstly, algorithms are developped and used in a 1D spherical configuration to describe theevaporation of a single static droplet without gravity. Secondly, the vaporistion of a droplet in alaminar flow has been studied. The analysis focus on heat exchanges between the droplet and theexternal environment as well as the force exerced on the droplet by the laminar flow. In this part,it has been highlighted that the evaporation induced a decrease in thermal exchanges and drag.In some cases of strong evaporation, the drag of the droplet has been observed to be negative.It means that the evaporation can cause a propulsion phenomenon of the droplet. A theoriticalanalysis allows to link this behaviour to an asymetry of the vaporisation rate. Thirdly, a studyof the turbulence influence on the evaporation of a droplet has been carried out. A generator ofturbulent fluctuations has been implemented and parallel approaches have been introduced toreduced computational time. It allowes to analyse thermal exchanges and drag behaviour of adroplet interacting with a turbulent flow. The analysis showed that theses two variables increasewith turbulence. Lastly, the evaporation of groups of droplets has been studied. For three differentgroups of droplets, trajectories of droplets have been analysed with heat exchanges when they areput in a laminar or a turbulent flow with or without phase change. In presence of evaporation,the analysis pointed out that trajectories were different from cases whitout evaporation and sothat phase change modifies collective effects. Moreover, these collective effects have also beenobserved on thermal exchanges.

Staphylococcus aureus is defined by its ability to agglutinate during exposure to human blood plasma. Although agglutination has long correlated with disease severity, the function of agglutination during infection remains unclear. Increasing evidence suggests the mechanisms of agglutination are highly complex and poorly understood. The goal of this dissertation was to characterize the mechanisms required for S. aureus agglutination in vitro and determine how these factors contribute to pathogenesis. Chapter II focuses on the development of two in vitro agglutination assays, which allow the process to be measured quantitatively. Through these assays, we confirmed the major factors contributing to agglutination are human fibrinogen and the bacterial surface protein, ClfA. Productive interactions between these two factors are required for agglutination to proceed. Surprisingly, we also identified a novel regulatory system that significantly contributed to agglutination. Inactivation of the ArlRS two-component system (TCS) prevents agglutination in both of the developed assays. Studies in Chapter III focused on characterizing the mechanism by which ArlRS inhibits agglutination. To examine regulation, quantitative PCR identified the major output of the ArlRS system as the gene ebh. Surprisingly, transcript levels of known extracellular matrix (ECM) binding proteins did not change. Characterization of ebh indicated that overexpression in an arlRS mutant is the major factor responsible for preventing agglutination. Deletion of ebh restores the ability of the arlRS mutant to agglutinate in both gravity and flow-based agglutination assays. Fluorescence microscopy of clumps indicates wildtype cells bind and incorporate fluorescently labeled human fibrinogen (Fg) displaying co-localization with the clumps. Surprisingly, arlRS mutants also bound human Fg, but these interactions were not productive for clumping, suggesting successful agglutination is more complex than binding ECM proteins. These studies indicate that ArlRS regulates agglutination through a unique mechanism that depends on the surface protein Ebh. Studies in Chapter IV were performed to determine the role ArlRS played in pathogenesis. A rabbit model of infective endocarditis and sepsis was employed to assess ArlRS virulence because this model has been shown to require agglutination for disease progression. Mutants in arlRS displayed reduced virulence in the rabbit model of infective endocarditis, which correlated with the mutant's inability to form a vegetation of the heart valve. These studies provide further insight into the importance of S. aureus agglutination during infection and define a mechanism of regulation through a novel surface protein.

Two-component systems (TCSs) are highly conserved across bacteria and are used to rapidly sense and respond to changing environmental conditions. The human pathogen Staphylococcus aureus uses the S. aureus exoprotein expression (sae) TCS to sense host signals and activate transcription of virulence factors essential to pathogenesis. Despite its importance, the mechanism by which the sae sensor kinase SaeS recognizes specific host stimuli is unknown. This thesis describes topology and mutagenesis studies of the sensing domain of SaeS, including basal expression and inducer-dependent phenotypes. Meanwhile, investigation of the sae auxiliary protein SaeP has identified a novel DNA binding function for this surface expressed lipoprotein that may be involved in fine-tuning the activity of the sae system. Overall, these structure-function studies provide insight into the sae signal transduction mechanism and raise some new questions regarding the role the sae system plays in the larger regulatory network S. aureus uses to control expression of its secreted virulence factors.

Biofilms are a crucial adaptation for bacterial survival against stresses from external environments. Biofilms are adherent colonies of sessile bacteria embedded within a self-produced matrix. Bacterial control over formation, maintenance, and response to external stresses are strictly regulated. However, complexities of intracellular signaling for biofilm regulation are still not fully understood. In this thesis, I report on two distinct regulatory systems important for biofilm formation in the opportunistic pathogen Pseudomonas aeruginosa. The first regulatory system I report on is the two-component system TctD-TctE. This system is involved in regulating the uptake of tricarboxylic acids such as citric acid and is involved in biofilm-specific susceptibility to aminoglycoside antibiotics. Here I describe work I performed characterizing the involvement of TctD-TctE in biofilm development when citric acid is present as a carbon source in nutrient media. In further characterizing a previously observed aminoglycoside susceptibility, I found that a strain with a deletion of TctD-TctE (ΔtctED) has a heightened accumulation of tobramycin in its biofilms when grown in the presence of citric acid. In ΔtctED, I determined that there was an inhibition of overall cell growth when citric acid was present in nutrient media. Additionally, in the presence of citric acid, ΔtctED displayed high levels of biofilm formation. This contrasted with normal biofilm development observed in the PA14 wild type strain where biofilm mass was reduced in the presence citric acid. The second project of this thesis reports on a novel regulatory small RNA, the Small RNA Regulator of Biofilms (SrbA). SrbA was found to be unique to P. aeruginosa and displayed no homology with any other sequenced bacterial species. I found that loss of SrbA resulted in a significant reduction in biofilm mass. Subsequently, loss of SrbA also leads to attenuation of P. aeruginosa pathogenicity in Caenorhabditis elegans nematodes. Bacterial biofilms possess specific regulatory programs that are still just being appreciated for their complexity. This thesis work adds to our understanding of biofilm regulation by studying roles of the two-component system TctD-TctE and the small RNA SrbA in P. aeruginosa.

In this thesis we develop a two component mixture model to perform a Bayesian regression. We implement our model computationally using the Gibbs sampler algorithm and apply it to a dataset of differences in time measurement between two clocks. The dataset has ``good" time measurements and ``bad" time measurements that were associated with the two components of our mixture model. From our theoretical work we show that latent variables are a useful tool to implement our Bayesian normal mixture model with two components. After applying our model to the data we found that the model reasonably assigned probabilities of occurrence to the two states of the phenomenon of study; it also identified two processes with the same slope, different intercepts and different variances. / McAnulty College and Graduate School of Liberal Arts; / Computational Mathematics / MS; / Thesis;