Antagonistic interactions between competing species of bacteria are an important driver of bacterial community composition in the human gut microbiota. Of particular significance is the role of the type six secretion system (T6SS), which many species of Gram-negative bacteria use to kill competitor bacteria in a contact-dependent manner. T6SSs are syringe-like nanomachines that function to deliver antibacterial toxins into susceptible competitors. Many bacteria present in the human gut microbiota possess an extremely potent T6SS that is capable of rapidly eradicating nearby bacteria. Remarkably, however, species of beneficial bacteria that coexist in the gut are often resistant to T6SS attack by their neighbours. This resistance is mediated by bacterial immunity proteins that block the activity of the antibacterial toxins delivered by the T6SS. Intriguingly, past studies have shown that the widespread T6SS-mediated competition in the gut has led to the acquisition of repertoires of immunity genes across different bacterial strains. By examining available human gut metagenomes, I identified a putative immunity locus, named I2, in a species of gut bacteria. This locus is located downstream of its cognate T6SS toxin-encoding locus, E2, and I show when co-expressed with E2 in E. coli, it protects against E2 mediated-toxicity. Additionally, I show that four gut-derived I2 homologues bearing sequence identity levels to I2 ranging from 38% to 75% are equally capable of abrogating E2 toxicity. Using quantitative biophysical measurements, I also show that these I2 homologues physically bind E2 equally tightly pointing to the potential molecular mechanism of toxin neutralization. Lastly, through mutagenesis experiments, I found that the E2-I2 interaction is likely mediated by electrostatic forces between a small number of residues found in the interaction interface of the two proteins. Overall, these findings demonstrate that a human gut microbiome encoded type VI secretion system effector can be neutralized by divergent immunity proteins. / Thesis / Master of Science (MSc)

Clostridium perfringens is a spore-forming anaerobic Gram-positive rod which has gliding motility through type IV Pili (TFP). Since the discovery of TFP in Gram-positive bacteria is relatively new, more studies are required to understand the mechanism and interaction of the proteins of this machinery. Moreover, the similarities between TFP and type 2 secretion system (T2SS) suggest that C. perfringens has also a T2SS. We studied the localization of TFP ATPases, PilB1, PilB2 and PilT in Bacillus subtilis to compare the localization in an organism other than C. perfringens and which lacks any known genes similar to TFP. Unlike the case in C. perfringens, PilB1 in B. subtilis localized to the poles in the absence of PilT, with some central foci at the future division sites. Colocalization of PilB1 was also studied with PilT and the results suggested that PilB1 needs PilT to migrate from the poles to the center. Localization of PilB2 in B. subtilis, was similar to the results in C. perfringens and to the localization of PilB1 in B. subtilis. We have not been able to co-express PilB2 with PilT yet. Succeeding in this study will help us better understand the interactions between PilB proteins and PilT. In another project, we studied a von Willebrand factor Type A-Domain Containing protein (vWA) which is secreted from C. perfringens strain 13. We overexpressed and purified this protein and tested the effects on mammalian cells. We found that the vWA is probably not a toxin but since it seems to bind to macrophage membranes, we propose that the vWA could be part of a toxin complex, probably the subunit of the complex that binds to the host cells. / Master of Science

The Type III secretion system (T3SS) of Pseudomonas aeruginosa uses a needle-like protein apparatus to detect eukaryotic host cells and translocate effectors directly into the host cell. The effectors are also known as cytotoxins, which cause disruption of a series of signaling events in the host cell, facilitating the infection by P. aeruginosa. As the T3SS is antigenic and the expression of T3SS is energy-consuming, it is highly regulated where several regulatory proteins interact with each other and control the expression of T3SS genes. Among these proteins, ExsA, the master regulator of T3SS in P. aeruginosa, is of great importance as it is a transcriptional activator that activates the expression of all T3SS genes. Also, as ExsA belongs to the AraC protein family which only exists in bacteria and fungi, it makes an excellent potential target for drugs against P. aeruginosa related infections. With a combination of molecular biology tools and structural biology methods, we solved the N-terminal domain structure of the ExsA protein in P. aeruginosa. The model of the ExsA N-terminal domain has enriched our knowledge about ExsA dimerization and can serve as the base for mapping the interaction interfaces on ExsA and ExsD. Further, we have found two homologues of ExsA by structural alignment, which share a lot of similarities and have conserved amino acid residues that are important for ligand binding. The fact that both of these two proteins are regulated by small ligands rather than proteins also raises the possibility that ExsA may have a second regulatory mechanism under which ExsA is regulated by a small ligand, which so far has not been observed or reported by researchers. In order to map the binding site of ExsA on its anti-activator ExsD, we removed the coiled-coil region (amino acid residue 138-202, the potential binding site) of ExsD, based on the  structure of ExsD. We surprisingly found that the ExsD variant without the coiled-coil region readily inhibits ExsA-dependent in vitro transcription. This result rules out other possibilities and makes us focus on the N-terminus and adjacent regions of ExsD for the interface with ExsA. Moreover, in order to gain a comprehensive understanding of the dynamics of the regulation of T3SS in P. aeruginosa, we have begun to build a mathematical model of the T3SS regulatory pathways. We are measuring the cellular concentrations of T3SS regulatory proteins with quantitative molecular biology methods such as quantitative western blot, quantitative PCR and quantitative mass spectrometry. We have determined the cellular level of ExsA and ExsD proteins under different physiological conditions, and found that some factors such as temperature have a significant impact on the levels of ExsA and ExsD. This study has thus unveiled some unknown features of the T3SS of P. aeruginosa and its related infections. / Master of Science

Pseudomonas aeruginosa is a ubiquitous opportunistic pathogen that is a substantial threat, particularly in hospital settings, causing severe infections in immunocompromised patients that may lead to death. Pseudomonas aeruginosa harbors a multitude of virulence factors that enable this pathogen to establish both acute and chronic infections in humans. A key determinant of acute infections is a hollow molecular needle structure used for injecting toxins into a host cell, called the type three secretion system (T3SS). The secretion machinery itself is highly complex and, together with the specific secreted factors, requires expression of more than 30 genes. Due to the high energy cost of its synthesis to the organism this system is highly regulated to finely time gene expression to coincide with host contact. ExsA, a member of the AraC-type transcription factor family, is the main transcriptional activator of all the genes necessary for expression of the T3SS. Members of the AraC family are characterized by the presence of two helix-turn-helix (HTH) motifs, which bind to the promoter DNA and activate transcription. ExsA uses its HTH containing C-terminal domain (CTD) to regulate gene expression from 10 different promoters. The N-terminal domain (NTD) of ExsA mediates dimerization and regulation of ExsA-activity. While most AraC-type activators are regulated by a small molecule ligands, ExsA is regulated by another protein, ExsD. As part of a four-protein signaling cascade, ExsD interacts directly with ExsA to prevent transcription of T3SS-associated genes under non-inducing conditions prior to host cell contact. The entire regulatory cascade includes of two additional proteins, ExsC and ExsE. ExsA, ExsC, ExsD, and ExsE follow a partner-switching mechanism to link expression of the secretion system with host cell contact. Our laboratory is working to understand this unique signaling mechanism by determining the molecular basis for the regulation of this important virulence factor. Previous studies in the laboratory have solved the structures of ExsE, ExsC and ExsD, and shed light on how these proteins interact and compete for overlapping binding sites. However, it is still unclear as to how the ExsA and ExsD interact and thus how regulation is mediated at the molecular level. In the presented study, we sought to map the molecular interface between ExsA and ExsD. First, the crystal structure of ExsA-NTD is presented wherein the dimerization interface of the protein was identified. Two of the well-studied AraC-type proteins, AraC and ToxT crystal structures have been solved by others in the presence of their respective ligands. Residues that were involved in ligand binding in AraC and ToxT were aligned with the residues in ExsA and analyzed for interaction with ExsD. However, this canonical binding pocket appeared to be not involved in the interaction between ExsA and ExsD. Structure directed site-specific mutagenesis was carried out to construct many different variants of ExsD and ExsA. Thus constructed variants were purified and analyzed in a functional assay. Using this approach, we were able to identify regions on ExsD and ExsA that are crucial for the interaction and for the regulation of ExsA-dependent transcription. It turns out that backbone interactions between the amino-terminal residues of ExsD and the beta-barrel region of the ExsA-NTD are pivotal. This result explains how ExsA and ExsC compete for ExsD binding, since both target the same regions on ExsD. / PHD / Pseudomonas aeruginosa is an opportunistic pathogen that is notorious for causing severe infections in immunocompromised individuals. Acute Pseudomonas aeruginosa infections are characterized by immediate adverse effects. An initial acute infection may become chronic, leading to long-term morbidity and mortality in affected individuals. During the initial stages of infection P. aeruginosa uses the type three secretion system, a syringe-like structure, to puncture the host cell and inject potent toxins. The activation of the genes required for forming this structure is tightly controlled by an activator protein, ExsA. When P.aeruginosa is not invading a host, ExsA is inhibited by another protein called ExsD, to prevent the needless production of the secretion apparatus. The presented work explores the mechanism of how ExsD achieves this inhibition of ExsA. This information is of potential biomedical interest because a clear understanding of the molecular basis for the interaction could inform the development of a small-molecule mimic of ExsD to be used in therapy. In Chapter 2 we report the structure of the domain of ExsA that is known to bind ExsD. Also, in this chapter and more so in Chapter 3, we performed a detailed analysis of potential interacting regions and ultimately succeeded in identifying key interacting regions in both ExsA and ExsD.

Chlamydia trachomatis est une bactérie intracellulaire obligatoire. Ce pathogène de l’Homme est la première cause infectieuse de cécité ainsi que de maladies sexuellement transmissible d’origine bacterienne.Utilisant une souche de C. trachomatis L2 exprimant une protéine fluorescente, nous avons développé des méthodes de microscopie et de cytométrie en flux permettant de suivre les différentes étapes du développement de la bactérie. Ces méthodes faciliteront les futures études de l’infection par Chlamydia.Chlamydia interagit avec différents processus cellulaires, et plus particulièrement via la sécrétion d’effecteurs par le système de sécrétion de type 3 (ST3). Nous avons identifié une famille de protéines possédant un signal de ST3 qui partagent un domaine, le DUF582, présent uniquement chez les Chlamydia pathogènes.Nous avons montré que les 5 protéines DUF582 de C. trachomatis sont exprimées à partir du milieu du cycle infectieux. Nous avons démontré que la protéine Hrs interagit avec le DUF582 et que la protéine DUF582 CT619 interagit avec Tsg101. Hrs et Tsg101 sont d’importants composants de la machinerie ESCRT impliquée dans de nombreux processus de fission membranaire.Utilisant l’interférence ARN, nous avons montré que Hrs et Tsg101 ne sont requis ni pour l’entrée, ni pour le développement de la bactérie. Ceci suggère que les protéines DUF582 bloquent des processus dépendant de Hrs/Tsg101. A l’inverse, la bactérie pourrait utiliser la machinerie ESCRT mais l’existence de mécanismes redondants expliquerait l’absence de phénotype dans les expériences d’interférence. Nous discutons trois hypothèses concernant le rôle des protéines DUF582 dans l’infection. / Chlamydia trachomatis is an obligate intracellular human pathogen. It is the first infectious cause of blindness and the most common cause of sexually transmitted diseases of bacterial origin. Using a strain of C. trachomatis serovar L2 expressing a fluorescent protein we developed microscopy and flow cytometry based methods to quantify several steps of its developmental cycle. These methods will facilitate future studies aimed at testing anti-bacterial compounds or various culture conditions. Chlamydiae interfere with many cellular processes, in particular via the secretion of bacterial proteins through a type 3 secretion (T3S) system. We identified a family of proteins that possess T3S signals. They share a domain designated as DUF582, which is only found in pathogenic chlamydiae. We showed that the five DUF582 proteins of C. trachomatis are expressed from the mid phase of infection. We demonstrated that the protein Hrs is a common interactor for the DUF582. In addition the N-terminal part of the DUF582 protein CT619 interacts with Tsg101. Hrs and Tsg101 are both important components of the ESCRT machinery, which is an ancient machinery required for several processes involving membrane fission.Using RNA interference we showed that Hrs and Tsg101 are dispensable for bacterial entry and growth. This last result suggest that DUF582 proteins actually prevent Hrs and/or Tsg101 driven processes. Alternatively, the bacteria might highjack the ESCRT machinery but redundant mechanisms would explain the absence of phenotype on bacterial development observed in the silencing experiments. We discuss three hypotheses as to the possible role of the DUF582 proteins in infection.

Cystic fibrosis (CF) is a fatal, incurable genetic disease that affects over 30,000 people in the United States alone. People with this disease have a homozygous mutation in the cystic fibrosis transmembrane conductance regulator (CFTR) which causes defects in chloride transport and leads to build up of mucus in the lungs and disruption of function in various organs. CF patients often suffer from chronic bacterial infections within the lungs, wherein the bacteria persist as a biofilm, leading to poor prognosis. Two of these pathogens, <i>Stenotrophomonas maltophilia</i> and <i>Pseudomonas aeruginosa</i>, are often found in the lungs of patients with CF and are an increasing medical concerns due to their intrinsic antimicrobial resistance. Both species can readily form biofilms on biotic and abiotic surfaces such as intravascular devices, glass, plastic, and host tissue. Biofilm formation starts with bacterial attachment to a surface and/or adjacent cells, initiating the acute infection stage. Chronic, long-term infection involves subsequent or concurrent altered genetic regulation, including a downregulation of virulence factors, resulting in the bacteria committing to a sessile lifestyle, markedly different from the planktonic one. Many of these genetic switches from an acute to chronic lifestyle are due to pressures from the host immune system and lead to permanently mutated strains, most likely an adaptive strategy to evade host immune responses. Biofilms are extremely problematic in a clinical setting because they lead to nosocomial infections and persist inside the host causing long-term chronic infections due to their heightened tolerance to almost all antibiotics. Understanding the genetic networks governing biofilm initiation and maintenance would greatly reduce consequences for CF and other biofilm-related infections and could lead to the development of treatments and cures for affected patients. This study showed that in<i> S. maltophilia</i>, isogenic deletion of phosphoglycerate mutase (<i>gpmA</i>) and two chaperone-usher pilin subunits, <i>S. maltophilia</i> fimbrae-1 (<i>smf-1</i>) and<i> cblA</i>, lead to defects in attachment on abiotic surfaces and cystic fibrosis derived bronchial epithelial cells (CFBE). Furthermore, Δ<i>smf-1</i> and Δ<i>cblA</i> showed defects in long-term biofilm formation, mimicking that of a chronic infection lifestyle, on abiotic surfaces and CFBE as well as stimulating less of an immune response through TNF-α production. This study also showed that in <i>P. aeruginosa</i>, the Type III secretion system (T3SS), an important virulence factor activated during the acute stage of infection, is downregulated when <i>polB</i>, a stress-induced alternate DNA polymerase, is overexpressed. This downregulation is due to post-transcriptional inhibition of the master regulatory protein, ExsA. Taken together, this project highlights important genes involved in the acute and chronic infection lifestyle and biofilm formation in <i>S. maltophilia</i> and genetic switches during the acute infection lifestyle in <i>P. aeruginosa</i>.

Escherichia coli enteroinvasora (EIEC) é um dos agentes etiológicos da disenteria bacilar. Seu processo fisiopatológico é desencadeado pela expressão de fatores de virulência, que proporcionam sua invasão e sobrevivência nas células do hospedeiro, ativando o sistema imune inato e adaptativo da mucosa intestinal. Trabalhos recentes têm salientado a importância do sistema de secreção e da flagelina bacteriana como agonista de receptores da imuninade inata dos macrófagos, em especial alguns dos receptores do tipo NLR. Uma vez que esta espécie de E. coli também é capaz de expressar flagelina e fazer a montagem completa do flagelo e do sistema de secreção do tipo III, a nossa proposta foi avaliar o papel da flagelina e do sistema de secreção de EIEC na resposta imune dos macrófagos murinos. Para isso, utilizamos três cepas de EIEC: a cepa selvagem; a cepa mutante no gene responsável pela síntese da flagelina; e a cepa sem o plasmídio de virulência plnv, deficiente no sistema de secreção, para a infecção de macrófagos peritoniais de camundongos C57BI/6, caspase-1-/-, IPAF-/- e ASC-/-. Neste estudo foi possível observar que o escape bacteriano e a morte dos macrófagos infectados por EIEC, assim como a ativação da caspase-1 e posterior secreção de IL-1&#946; é independente da flagelina bacteriana, mas dependente do sistema de secreção, além disso, a ativação da caspase-1 de macrófagos infectados por EIEC é dependente do receptor IPAF e parcialmente da proteína adaptadora ASC. Assim, no nosso modelo, a ativação da caspase-1 dos macrófagos infectados por EIEC parece estar envolvida com o processamento e secreção de IL-1&#946; e, possivelmente na secreção de IL-18, mas não na morte celular. No modelo de infecção in vivo, o sistema de secreção bacteriano foi importante para a sobrevivência bacteriana no hospedeiro, assim como para a indução de uma resposta inflamatória no local da infecção. Ainda, a caspase-1 parece ter um papel importante para o controle da infecção in vivo por EIEC, podendo assim contribuir para uma resposta imune protetora do hospedeiro. / Enteroinvasive Escherichia coli (EIEC) is one of the etiologic agents responsible for bacillary dysentery. The pathophysiological process induced by this bacteria is triggered by the expression of virulence factors that provide the invasion and survival in host cells, resulting in activation of innate and adaptive immune system present on intestinal mucosa. Recent studies have emphasized the importance of the secretion system and bacterial flagellin as agonist of innate immune receptors present in macrophage, especially NLR (Nod like receptors). Then, our proposal was evaluate the role of flagellin (f1iC) and secretion system of EIEC in the induction of immune response of murine macrophages using the EIEC strains wild type (WT), mutant flagellin gene (f1iC), and a strain deficient in secretion system (DSS) for infection of peritoneal macrophages of C57Bl/6, caspase-1-/-, IPAF-/- and ASC-/-- mice. In this study we observed that the bacterial escape and death of infected macrophages with EIEC, the caspase-1 activation and subsequent IL-1&#946; secretion is independent of bacterial flagellin, but dependent of secretion system, moreover, the caspase-1 activation in infected macrophages is IPAF-dependent and partially dependent of the adapter protein ASC. Thus, in our model, the caspase-1 activation in EIEC infected macrophages seems to be involved with the processing and secretion of IL-1&#946; and possibly with the secretion of IL-18, but not involved with cell death. In the infection model in vivo, bacterial secretion system was important for bacterial survival in the host, as well as for the inflammatory response induction at the infection site. In addition, caspase-1 seems to have an important role to the control of in vivo infection by EIEC and can contribute to a protective immune response of the host.

A Transferência Horizontal de Genes (THG) é um dos principais mecanismos de evolução bacterianos, impactando a evolução de praticamente todas famílias gênicas. Neste trabalho identificamos e avaliamos padrões de possíveis transferências horizontais de genes pertencentes a duas classes funcionais de dois níveis taxonômicos distintos. Caracterizamos a ocorrência e evolução de 45 genes importantes para a fixação de N2 em 479 genomas de Proteobacteria. Identificamos cinco potenciais aquisições de genes ligados a fixação de N2 por linhagens de Proteobacteria, as quais foram identificadas consistentemente em 36 dos genes analisados. Realizamos predições de transferências horizontais dos 45 entre todos os 479 genomas de Proteobacteria e identificamos possíveis enriquecimentos de THG, provavelmente ligados à sinais filogenéticos e ecológicos. Desenvolvemos um pipeline para identificação semi-automática de efetores do Sistema Secretor do Tipo III em Aeromonas, o qual reportou 21 famílias de potenciais efetores presentes em 105 genomas. Entre os 21 efetores identificados 17 foram descritos pela 1º vez em Aeromonas, corroborando a sensibilidade de nosso pipeline. Com o auxílio de nossos colaboradores foram realizados testes de citotoxidade para efetores identificados in silico, e apenas quatro não inibiram o crescimento de Saccharomyces cerevisiae. Por fim, desenvolvemos um método para agrupamento de famílias gênicas com histórias evolutivas similares que não requer a reconstrução de árvores filogenéticas, aumentando a eficiência computacional. Aplicamos o método desenvolvido para reconstrução da filogenia de Aeromonas, o qual mostrou-se compatível com dados presentes na literatura. / Horizontal Gene Transfer (HGT) is one of main mechanisms of bacterial evolution, affecting virtually all gene families. In this document we identified and assessed putative horizontal transfers of genes from two functional classes from two distinct taxonomic levels. We characterized the distribution and evolution of 45 genes important to N2 fixation among 479 Proteobacteria genomes. We identified five potential distinct acquisitions of such genes by Proteobacteria lineages. The distinct origins are consistently identified in 36 out of the 45 assessed genes. We computed possible horizontal transfers of the 45 genes among the 479 Proteobacteria genomes, and we identified enrichments of HGT, likely related to phylogenetic and ecological signals. We developed a semi-automated pipeline to identify effectors of the Type III Secretion System within Aeromonas, which reported 21 putative effector families distributed among 105 genomes. Among the 21 likely effectors 17 have been described in Aeromonas for the first time, highlighting the sensibility of our pipeline. Our colaborators performed cytotoxicity tests for the 21 likely effector families identified by in silico analysis, and only four did not inhibited Saccharomyces cerevisiae growth. Lastly, we developed a method to cluster gene families according to shared evolutionary history, without the requirement of phylogenetic tree reconstruction, increasing computational efficiency. We applied this proposed method during Aeromonas phylogenetic reconstruction, and it showed up compatible with data available on the literature.

La pathogénicité de P. aeruginosa (P. a) repose sur de nombreux facteurs de virulence dont le système de sécrétion de type III (SST3). Ce complexe multiprotéique est constitué d'une aiguille se terminant par un translocon composé des protéines PopB et PopD. En s'insérant dans les membranes plasmiques, le translocon permet le passage des exotoxines dans le cytoplasme de la cellule cible. L'induction de la synthèse et de la sécrétion des exotoxines est dépendante d'un contact entre P. a et la cellule cible. Dans ce travail, nous avons examiné l'influence de facteurs cellulaires sur l'efficacité de translocation des toxines. L'utilisation d'un système rapporteur fluorescent CCF2/β-lactamase a permis de visualiser l'injection de toxine. En parallèle, l'association des protéines du translocon avec la membrane de la cellule hôte a été évaluée par immunodétection de PopB/D après fractionnement des membranes sur gradient de sucrose. Les cellules promyelocytaires HL-60 et promonocytaires U937 sont résistantes à l'injection de toxine, bien que PopB et PopD soient associées à la membrane. Après différenciation en neutrophiles, or monocytes/macrophages, ces cellules deviennent sensibles à l'injection sans que l'on détecte de variation notable de la quantité de protéines du translocon insérées dans la membrane. Le traitement des cellules HL-60 sensibles avec un agent déplétant le cholestérol, entraine une diminution de l'injection de toxine. De plus, la protéine PopB est retrouvée dans la fraction membranaire, obtenue par purification sur gradient de sucrose, contenant le marqueur des radeaux lipidiques flotilline. Par une approche pharmacologique, nous apportons la preuve que, en plus de la composition de la membrane, des voies de signalisation intracellulaires impliquées dans la polymérisation de l'actine sont essentielles pour la formation d'un pore fonctionnel. / The pathogenesis of Pseudomonas aeruginosa (P.a) implies multiple virulence factors among which the type III secretion system (T3SS). This multiprotein complex is composed of a needle through which four exotoxins are exported. The protein PopB and PopD form an oligomeric structure (translocon) at the end of the needle that inserts into the host cell membrane and translocates the exotoxins into the cytoplasm. Synthesis and toxin secretion is induced on contact with eukaryotic cell. In this work, we examined the influence of host cell elements on exotoxin translocation efficiency. The delivery of T3SS toxins was investigated using a CCF2/β-lactamase fluorescent reporter system In parallel, the association of translocon proteins with host plasma membranes was evaluated by immunodetection of PopB/D following sucrose gradient fractionation of membranes. Promyelocytic HL-60 cells and promonocytic U937 cells were found to be resistant to toxin injection even though PopB/D associated with host cell plasma membranes. Differentiation of these cells to neutrophil- or macrophage-like cells resulted in an injection-sensitive phenotype without any significant change in the level of membrane-inserted translocon proteins. Treatment of sensitive HL-60 cells with a cholesterol-depleting agent, resulted in a diminished injection of toxin. Moreover, the PopB translocator was found in the membrane fraction obtained from sucrose-gradient purifications and containing lipid-raft marker flotillin. Through a pharmacological approach, we brought evidence that, in addition to membrane composition, some general signalling pathways involved in actin polymerization may be critical for the formation of a functional pore.

Les protéines synthétisées et sécrétées par les bactéries jouent des rôles importants pour leur survie. Les bactéries à Gram négatif ont développé des voies de sécrétion en tant qu'armes principales pour transporter des facteurs de virulence dans l'environnement extracellulaire ou dans des cellules hôte. L'un de ces systèmes, le T9SS a été principalement étudié chez l'agent pathogène oral Porphyromonas gingivalis et chez la bactérie mobile Flavobacterium johnsoniae. Un autre complexe, le T2SS est le principal déterminant de la virulence de la bactérie Pseudomonas aeruginosa, un agent pathogène de la fibrose kystique. Dans le cadre de ma thèse, j'ai résolu la structure atomique de plusieurs composants centraux du T9SS et du T2SS. Concernant le projet T9SS, j'ai essayé de cristalliser le domaine cytoplasmique de GldL de F. johnsoniae. La co-cristallisation de GldL avec des Nbs a été réalisée sans succès. Néanmoins, les structures cristallines de deux nanobody contre GldL ont été résolues par remplacement moléculaire. De plus, j'ai également travaillé sur la protéine PG1058 de P. gingivalis. J'ai résolu sa structure par diffraction anomale à la longueur d’onde du selenium. Concernant le projet T2SS, je me suis concentré sur la partie N-terminale de XcpQ, une sous-unité de la sécrétine. J'ai résolu la structure cristalline de XcpQN012 seul et en complexe avec le nanobody vhh04 à une résolution de 2,98 Å et de 2,9 Å, respectivement. Enfin, j'ai participé à la détermination structurale de TssK, un composant de plaque de base du système de T6SS et déterminer la structure cristalline d'un nanobody contre le domaine périplasmique de PorM. / Proteins synthesized and secreted by bacteria serve many important roles in their survival. In particular, Gram-negative bacteria have evolved secretion pathways as the main weapons for transporting virulence factors into target cells or into the extracellular environment. One of these systems, the type IX secretion system (T9SS) or the Por secretion system, has been studied mainly in the oral pathogen Porphyromonas gingivalis and the gliding bacterium Flavobacterium johnsoniae. Another complex, the type II secretion system (T2SS) is the main determinant of the virulence of Pseudomonas aeruginosa, a cystic fibrosis pathogen. In my PhD thesis, I solved the atomic structure of several core components of both T9SS and T2SS.For the T9SS project, I tried to crystallize the cytoplasmic domain of GldL from F. johnsoniae. The co-crystallization of GldL with Nbs was unsuccessfull. The crystal structures of two nanobodies against GldL were solved by molecular replacement. I also worked on the PG1058 protein of P. gingivalis. I obtained crystals of the selenomethionine-derivatized PG1058 OmpA_C-like domain that diffracted up to 1.55 Å, and solved its structure by single-wavelength anomalous diffraction. For the T2SS project, I focused on the N-terminal part of XcpQ, a subunit of the secretin. I solved the crystal structure of XcpQN012 alone and in complex with nanobody vhh04 at a resolution of 2.98 Å and 2.9 Å, respectively. In addition, I also took part in the structural determination of the base plate component TssK of the T6SS and determined the crystal structure of one nanobody (vhh19) against the periplasmic domain of PorM.