In evolution alternative genetic trajectories can potentially lead to similar phenotypic outcomes. However, certain trajectories are preferred over others. These preferences bias the genomes of living organisms and the underlying processes can be observed in ongoing evolution. We have studied a variety of biases that can be found in bacterial chromosomes and determined the selective causes and functional consequences for the cell. We have quantified codon usage bias in highly expressed genes and shown that it is selected to optimise translational speed. We further demonstrated that the resulting differences in decoding speed can be used to regulate gene expression, and that the use of ‘non-optimal’ codons can be detrimental to reading frame maintenance. Biased gene location on the chromosome favours recombination between genes within gene families and leads to co-evolution. We have shown that such recombinational events can protect these gene families from inactivation by mobile genetic elements, and that chromosome organization can be selectively maintained because inversions can lead to the formation of unstable hybrid operons. We have used the development of antibiotic resistance to study how different bacterial lifestyles influence evolutionary trajectories. For this we used two distinct pairs of antibiotics and disease-causing bacteria, namely (i) Mycobacterium tuberculosis that is treated with rifampicin and (ii) Escherichia coli that is treated with ciprofloxacin. We have shown that in the slow-growing Mycobacterium tuberculosis, resistance mutations are selected for high-level resistance. Fitness is initially less important, and over time fitness costs can be ameliorated by compensatory mutations. The need for rapid growth causes the selection of ciprofloxacin resistance in Escherichia coli not only to be selected on the basis of high-level resistance but also on high fitness. Compensatory evolution is therefore not required and is not observed. Taken together, our results show that the evolution of a phenotype is the product of multiple steps and that many factors influence which trajectory is the most likely to occur and be most beneficial. Over time, selection will favour this particular trajectory and lead to biased evolution, affecting genome sequence and organization.

La quantité de données générée dans le cadre d'étude à grande échelle du réseau d'interaction protéine-protéine dépasse notre capacité à les analyser et à comprendre leur sens; d'une part, par leur complexité et leur volume, et d'un autre part, par la qualité du jeu de donnée produit qui semble bondé de faux positifs et de faux négatifs. Cette dissertation décrit une nouvelle méthode de criblage des interactions physique entre protéines à haut débit chez Saccharomyces cerevisiae, la complémentation de fragments protéiques (PCA). Cette approche est accomplie dans des cellules intactes dans les conditions natives des protéines; sous leur promoteur endogène et dans le respect des contextes de modifications post-traductionnelles et de localisations subcellulaires. Une application biologique de cette méthode a permis de démontrer la capacité de ce système rapporteur à répondre aux questions d'adaptation cellulaire à des stress, comme la famine en nutriments et un traitement à une drogue. Dans le premier chapitre de cette dissertation, nous avons présenté un criblage des paires d'interactions entre les protéines résultant des quelques 6000 cadres de lecture de Saccharomyces cerevisiae. Nous avons identifié 2770 interactions entre 1124 protéines. Nous avons estimé la qualité de notre criblage en le comparant à d'autres banques d'interaction. Nous avons réalisé que la majorité de nos interactions sont nouvelles, alors que le chevauchement avec les données des autres méthodes est large. Nous avons pris cette opportunité pour caractériser les facteurs déterminants dans la détection d'une interaction par PCA. Nous avons remarqué que notre approche est sous une contrainte stérique provenant de la nécessité des fragments rapporteurs à pouvoir se rejoindre dans l'espace cellulaire afin de récupérer l'activité observable de la sonde d'interaction. L'intégration de nos résultats aux connaissances des dynamiques de régulations génétiques et des modifications protéiques nous dirigera vers une meilleure compréhension des processus cellulaires complexes orchestrés aux niveaux moléculaires et structuraux dans les cellules vivantes. Nous avons appliqué notre méthode aux réarrangements dynamiques opérant durant l'adaptation de la cellule à des stress, comme la famine en nutriments et le traitement à une drogue. Cette investigation fait le détail de notre second chapitre. Nous avons déterminé de cette manière que l'équilibre entre les formes phosphorylées et déphosphorylées de l'arginine méthyltransférase de Saccharomyces cerevisiae, Hmt1, régulait du même coup sont assemblage en hexamère et son activité enzymatique. L'activité d'Hmt1 a directement un impact dans la progression du cycle cellulaire durant un stress, stabilisant les transcrits de CLB2 et permettant la synthèse de Cln3p. Nous avons utilisé notre criblage afin de déterminer les régulateurs de la phosphorylation d'Hmt1 dans un contexte de traitement à la rapamycin, un inhibiteur de la kinase cible de la rapamycin (TOR). Nous avons identifié la sous-unité catalytique de la phosphatase PP2a, Pph22, activé par l'inhibition de la kinase TOR et la kinase Dbf2, activé durant l'entrée en mitose de la cellule, comme la phosphatase et la kinase responsable de la modification d'Hmt1 et de ses fonctions de régulations dans le cycle cellulaire. Cette approche peut être généralisée afin d'identifier et de lier mécanistiquement les gènes, incluant ceux n'ayant aucune fonction connue, à tout processus cellulaire, comme les mécanismes régulant l'ARNm. / The quantity of data generated within the framework of protein-protein interaction network large-scale studies exceeds our capacity to analyze them and to understand their meaning; on one hand, by their complexity and their number, and on the other hand, by the quality of the produced data, which are populated with spurious interactions. This dissertation describes new applications of a protein-fragments complementation assay (PCA) to screen for interactions among all proteins in the budding yeast Saccharomyces cerevisiae. This approach is carried out in intact cells, with proteins expressed in their native contexts and under their endogenous promoter, thus assuring correct post-translational modifications and subcellular localization. A further novel application of PCA is described for investigating proteome wide changes in response to cellular adaptation to stresses, such as nutrient starvations and drug treatments. Finally, as a result of the latter strategy applied to characterizing proteome-wide response to the immunosuppressant drug, rapamycin, I describe the discovery of an unforeseen mechanism of modulating cell cycle progression through control of cyclin mRNA stability. In the first chapter of this dissertation, I present a pairwise screen of interactions among proteins resulting from the ~6000 open reading frames in Saccharomyces cerevisiae. We identified 2770 interactions among 1124 proteins. We estimated the quality of our screen by comparing our results to curated gold standard data and coverage of known interactions to all previous studies. The majority of our interactions were novel, but overlap with data from previous studies was as high as 40%. PCA is based on refolding of the reporter protein from complementary N- and C- terminal fragments following interaction of the two proteins to which they are fused. Thus, reporter activity is sterrically limited to interactions in which the termini of the proteins to which the complementary reporter fragments are fused are sufficiently close in space. In the case of our reporter, this limit was 8 nm. Thus PCA is a molecular ruler, providing information on both direct protein-protein interactions and sterrically restricted distances between proteins in complexes. We benchmarked and demonstrated correct topological relationships for a number of known complexes, including the proteasome, RNA polymerase II and the nuclear pore complex. Thus our study provided, for the first time, a topological map of complex organization in a living cell. The integration of the results from such efforts with those of gene regulation dynamics and protein modifications will lead to a fuller understanding of how complex cellular processes are orchestrated at a molecular and structural level in the living cell. In chapter 2, I describe the results of an application of PCA to study the dynamic rearrangement of the proteome under a specific stress; treatment of cells with rapamycin. The results of these efforts were the identification of a novel mechanism of cell cycle control at the level of cyclin mRNA. Specifically, we discovered that the balance between the phosphorylated and dephosphorylated forms of the Saccharomyces cerevisiae arginine methyltransferase, Hmt1, regulates both its assembly into a hexamer and its enzymatic activity. The Hmt1 activity modulates cell cycle progression through stabilizing the B cyclin CLB2 mRNA. We then used PCA to identify the Hmt1 regulators under rapamycin treatment. We identified the catalytic subunit of the PP2a phosphatase, Pph22, activated by the inhibition of TOR, and the kinase Dbf2, activated during entry into mitosis, as the phosphatase and the kinase responsible for the modification of Hmt1 and for its regulatory functions in the cell cycle. I thus, in the end close the circle I began in this summary, going from large-scale discovery of protein-protein interactions, to mapping dynamics of proteome changes during an adaptation and finally to mechanistic insight into a primordial control mechanism in cellular dynamics. The strategies that we devised to discover this mechanism can be generalized to identify and mechanistically link genes together, including those of unknown function, to any cellular process.

L'inflammation est considérée comme un promoteur de la cancérogenèse. La cyclooxygénase-2 (COX-2), la forme inductible de la famille des cyclooxygénases est un médiateur important de l'inflammation. Cette enzyme est constitutivement exprimée dans un grand nombre de cancers tels que les cancers du sein, du colon ou de la prostate. De nombreuses études mettent en évidence que COX-2 est surexprimée lors des étapes pré-néoplasiques. La COX-2 représente de ce fait une cible thérapeutique potentielle en chimioprévention et également pour le traitement des cancers. L'utilisation d'inhibiteurs synthétiques de COX-2 qui ciblent l'activité enzymatique est le seul traitement clinique actuellement disponible pour réduire l'activité de COX-2. Cependant, ces agents présentent des effets secondaires sévères, ce qui limite leur prise chronique chimiopréventives ou au cours des traitements anti-cancéreux. Une stratégie alternative pour cibler la fonction de COX-2 est d'inhiber son expression. Un grand nombre d'études montrent que certains produits naturels (la curcumine, le resveratrol ou l'apigénine par exemple) inhibent, préférentiellement l'expression de COX-2, sans être toxique. Notre projet analyse l'effet de l'acide cinnamique, un produit naturel extrait de la plante Cinnamonium cassia, sur l'expression de COX-2 au cours de la cancérogenèse dans le but d'évaluer son intérêt en chimioprévention. Nous avons utilisé comme modèle les cellules mammaires non carcinogènes, MCF10A stimulées avec un ester de phorbol, le 12-phorbol myristate 13-acetate (PMA), qui induit l'expression de COX-2. Nous avons observé une diminution de l'expression de COX-2 au niveau de l'ARNm et de la protéine après le traitement avec différentes concentrations d'acide cinnamique (1 et 10μM). L'analyse des mécanismes impliqués dans la diminution de l'expression de COX-2 a mis en évidence que l'acide cinnamique régule l'expression de COX-2 de façon post-transcriptionnelle en réduisant la stabilité de son ARNm. Cet effet est associé à une prévention de la diminution de l'expression de microARN (miR)-16 et une inhibition de l'expression de p38 induite par l'acide cinnamique en réponse au traitement avec le PMA. / Inflammation is considered a cancer-promoting factor. Cyclooxygenase-2 (COX-2), the inducible form of the family of cyclooxygenases is an important mediator of inflammation, which has been found constitutively expressed in many forms of cancer including breast, colon or prostate. A number of studies show that COX-2 is stably expressed since the early pre-neoplastic stages. This encourages us to consider COX-2 as a potential target in chemoprevention as well as in the treatment of cancer. Synthetic inhibitors of COX-2, which target enzymatic activity, are the only clinical strategy to counteract COX-2. However, these compounds present severe side effects, a fact that limits their prolonged intake, like requested in chemoprevention or during anti-cancer treatment.An alternative strategy to target COX-2 is at the level of its expression. A number of studies show that several natural compounds including curcumin, resveratrol or apigenin preferentially target COX-2 expression without showing toxicity.Our study analyses the effect of cinnamic acid, a natural compound derived from Cinnamonium cassia on COX-2 expression during carcinogenesis, with the final aim to evaluate its potential in chemoprevention/therapy.For our chemopreventive purposes, we used the non-carcinogenic breast cell line MCF10A, stimulated by the phorbol ester 12-phorbol myristate 13-acetate (PMA), which typically induces COX-2. We show a reduction of induced COX-2 expression after treatment with different concentrations of cinnamic acid (1 and 10μM). The analysis of the mechanisms involved in COX-2 protein expression decrease shows that cinnamic acid regulated COX-2 expression at the post-transcriptional level by reducing COX-2 mRNA stability. This effect is associated with the ability of cinnamic acid to prevent downregulation of miR-16 expression and p38 activation in response to PMA treatment.

O gene do Retardo Mental do X Frágil (FMR1) possui 17 éxons e seu transcrito primário pode sofrer splicing alternativo, havendo, entre outros eventos, possibilidade de exclusão ou inclusão do éxon 12. O produto da expressão do FMR1, a proteína do retardo mental do X frágil (FMRP), possui papéis importantes no sistema nervoso central, atuando como repressora da tradução de RNAm em espinhas dendríticas e controlando a síntese de proteínas envolvidas na função sináptica. Entre dois domínios centrais do tipo KH presentes na FMRP, o segundo (KH-2) é responsável pela interação da proteína aos polissomos. O domínio KH-2 é codificado pelos éxons 9 a 13 do FMR1 e possui a alça variável mais longa já observada entre proteínas humanas, que é codificada pelos éxons 11 e 12. A inclusão do éxon 12 no RNAm do FMR1 causa uma extensão em fase dessa alça variável do KH-2 da FMRP. Estas isoformas apresentam expressão significativa em neurônios cortico-cerebrais e cerebelares do rato, no primeiro mês pós-natal. Este trabalho baseia-se em resultados prévios do grupo de pesquisa, em que se identificaram sequências curtas no íntron 12 do FMR1, com potencial para agir como acentuadores de splicing. Baseando-nos na hipótese de que essas sequências constituem elementos transcritos que se ligam a fatores proteicos do núcleo celular, potencialmente reguladores do splicing do pré-RNAm do FMR1, realizamos ensaios de precipitação por afinidade com extratos nucleares de córtex cerebral de rato e transcritos do loco, biotinilados. Análises por espectrometria de massas revelaram enriquecimento de proteínas nucleares, contendo domínios de ligação a RNA, principalmente aquelas relacionadas à regulação e processamento de pré-RNAm, sobretudo o splicing / Fragile X Mental Retardation 1 gene (FMR1) comprises 17 exons. Its primary transcript is subject to alternative splicing, allowing for the possibility of exon 12 inclusion or skipping, among other events. The product of FMR1 gene expression, fragile X mental retardation protein (FMRP), has important roles in the central nervous system, acting as a translational repressor in dendritic spines, thus controlling the synthesis of proteins involved in synaptic function. FMRP has two central KH domains. One of them (KH-2) is responsible for its interaction with polysomes. The KH-2 domain is encoded by FMR1 exons 9 to 13. It contains the longest variable loop ever observed among human KH-containing proteins, which is encoded by FMR1 exons 11 and 12. Exon-12 inclusion in FMR1 mRNA causes an in-frame extension of FMRP KH-2 domain variable loop. These isoforms appear significantly expressed in cortico-cerebral and cerebellar neurons of the rat in the first month after birth. We have previously identified short sequences within FMR1 intron 12 that may potentially act as splicing enhancers. Our study is based on the hypothesis that those sequences when transcribed should bind to nuclear protein factors that may function as FMR1 exon 12 pre-mRNA splicing regulators. To initiate an experimental approach to test that hypothesis, we conducted affinity precipitation assays with rat cerebral cortex nuclear extracts and biotinylated transcripts. Mass spectrometry analyses disclosed proteins that have been described to be enriched in the cell nucleus, contain RNA-binding domains, and be functionally related to pre-mRNA processing, notably splicing

Viele biologische Prozesse im Verhalten von ganzen Organismen, aber auch in den Prozessen und der biochemischen Zusammensetzung von Zellen zeigen einen zirkadianen Rhythmus, also einen Rhythmus mit einer Periode von etwa 24 Stunden. Diese 24-Stunden-Rhythmen sind in der Genexpression auf allen Ebenen zu finden: von der Tran- skriptionsinitiation bis zur Proteindegradation. Auf Transkriptebene, zirkadiane mRNA-Produktion und mRNA-Abundanz ist umfassend gemessen. Auf der anderen Seite, zirkadiane posttranskriptionelle Regulation ist weit weniger verstanden. In dieser Arbeit untersuche ich, wie bisher ungemessene, posttranskriptionelle Prozesse die rhythmischen Eigenschaften der Genexpression beeinflussen. Dazu beschreibe ich die Lebensstadien eines Bio-Moleküls mit einem Modell-Motiv, einer einfachen Differentialgleichung mit zeitabhängigen, rhythmischen Raten. Als erstes diskutiere ich die Einschränkungen von Phase und Amplitude zirkadianer Transkripte, die nur von konstanter PTR beeinflusst werden. Bei vielen gemessenen Transkripten sind diese Einschränkungen verletzt. In diesen Fällen muss es eine rhythmische PTR geben. Ich untersuche, welche rhythmische PTR diese Fälle erklären können und führe einen statistischen Test ein, der auf unbeobachtete, rhythmische PTR testet. Durch die Analyse zweier Datensätze von Mausleber und -niere finde ich, dass 18% aller zirkadianen Gene in Niere und 34% in Leber rhythmisch posttranskriptionell reguliert sind. Im zweiten Teil analysiere ich weitere Aspekte von PTR in einem Hypothesen-getriebenen Ansatz. Ich zeige, dass Spleißen mit einem Rhythmus von 24 Stunden 12 Stunden-Rhythmen in der Abundanz von mRNA erzeugen kann. Als nächstes schlage ich ein Modell vor, das rhythmische Degradation von Mitgliedern der zirkadianen Uhr beschreibt. Schließlich erweitere ich das Modell-Grundmotiv zu einer partiellen Differentialgleichung (PDG), die das “Altern” von Molekülen beschreibt. / In almost all organisms on Earth, many behavioral, physiological, and biochemical activities oscillate with a circadian rhythm, a rhythm with a period of about 24 hours. In gene expression, the 24-hour-rhythm can be found on all stages: from transcription initiation to protein degradation. On the transcript level, circadian mRNA production and mRNA abundance are comprehensively charted through numerous genome-wide high throughput studies. Circadian post-transcriptional regulation, however, is less well understood. In this thesis, I will investigate how unobserved post-transcriptional processes influence rhythmic properties of gene expression. To this end, I quantify the life-stages of biomolecules using one modeling motif, a simple ordinary differential equation describing production and degradation with time-dependent rhythmic rates. This basic modeling motif is systematically varied to examine and discuss various influences of post-transcriptional regulation (PTR) on circadian mRNA expression. I first discuss the restrictions of rhythmic phase and amplitude of circadian transcripts influenced by non-rhythmic PTR. For many genes these restrictions are violated and we have to assume the existence of a rhythmic PTR. I discuss which rhythmic PTR can explain these findings and further introduce a statistical test to quantify the extent of unobserved rhythmic PTR. Analyzing two data sets on mouse liver and kidney, I find that 18% of circadian genes in kidney and 34% in liver are under rhythmic post-transcriptional control. In a second part, I analyze more specific aspects of PTR in a hypothesis-driven approach. Firstly, I find that splicing with a rhythm of 24 hours is able to generate 12-hour rhythms in abundance of mature mRNA. Secondly, I propose and analyze a model to investigate rhythmic degradation of core clock genes. And finally, I extend the core modeling motif to a partial differential equation (PDE) model that accounts for the “aging” process of molecules.

Les protéines de liaison à l’ARN PUMILIO 1 et 2 (PUM1/2) exercent un rôle central dans le maintien des cellules souches chez les Invertébrés en se fixant, en association avec des partenaires protéiques, sur la région 3’ UTR de certains ARNm, régulant ainsi leur devenir. A ce jour, le rôle de PUM1/2 dans les cellules souches/progénitrices hématopoïétiques (CSPHs) a été peu étudié. La perte de la coordination entre auto-renouvellement et différenciation des CSPHs peut aboutir à des hémopathies chez l'Homme, d’où la nécessité de comprendre les mécanismes sous-jacents. Notre équipe a mis en évidence, par une approche de shARN, que l’invalidation des protéines PUM1/2 dans les CSHs humaines et murines conduit à une réduction de leur expansion, associée à une apoptose accrue et un arrêt du cycle cellulaire en phase G0/G1, et aussi à une perte du potentiel clonogénique in vitro et du potentiel de reconstitution in vivo. L’objectif de notre travail a consisté à : a/ évaluer les effets de la surexpression de PUM1/2 dans les CSPHs, b/ déterminer l’implication de PUM1/2 dans les processus leucémiques, c/ étudier les mécanismes moléculaires responsables de l’activité de PUM1/2 en identifiant les cibles et les partenaires protéiques par une approche de protéomique globale. Nos résultats suggèrent qu’une surexpression modérée de PUM1 (2/3 fois) dans les cellules CD34+ limite la perte du potentiel clonogénique alors qu’une expression plus élevée (5/10 fois et plus) est toxique. L’analyse de l’expression de PUM1/2 par RT-qPCR dans les échantillons de Leucémies Aigue Myeloïdes (LAM) (GOELAMSthèque) montre une augmentation significative dans les échantillons les plus immatures (LAM0-2) comparés aux contrôles sains. La perte de PUM1/2 par shARN dans les cellules primaires de leucémies ainsi que dans des lignées issues de différents processus leucémiques réduit fortement leur survie. La recherche des partenaires associés à PUM par spectrométrie de masse a permis de découvrir Argonaute2 et MOV10 (tous les 2 impliqués dans la machinerie des miRNA), ainsi que des protéines de liaison aux ARNs, ELAV1 déjà connue pour son implication dans le maintien des CSH murines et IMP3, impliqué dans de nombreux cancers et dans la régulation du cycle cellulaire. L’invalidation de IMP3 ou ELAV1 dans les CSPHs conduisent, in vitro, aux mêmes effets observés avec la perte du PUM 1/2, une diminution de l’expansion avec une augmentation de l’apoptose, et la perte du potentiel clonogénique. Enfin, nous avons identifié FoxP1 (Forkhead box P1) comme nouvelle cible directe de PUM1/2, dont le rôle est encore très peu décrit dans l’hématopoïèse. L’étude fonctionnelle de FoxP1 sur les CSPHs par shARN mime les effets observés avec les facteurs PUM1/2. De plus, la surexpression de FoxP1 restaure partiellement les activités antiprolifératives et pro-apoptotiques générées par les shPUM1/2. Enfin, le profil d’expression de FoxP1 dans les LAM corrèle avec le profil d’expression de PUM1/2. Nos résultats confirment le rôle majeur joué par les protéines PUM1/2 en partie via la régulation positive de FoxP1 qui contribue au maintien les CSPHs normales et pathologiques. / Pumilio 1 and 2 (PUM1/2) RNA-binding proteins exert a central role in stem cell maintenance among Invertebrates by binding the 3'UTR of mRNA targets in association with protein partners, thus regulating mRNA stability/translation. Nothing is known regarding normal and pathologic hematopoietic stem and progenitor cells (HSPCs). Loss of coordination between self-renewal and differentiation of HSPCs can lead to leukemia in humans, hence the need to understand the mechanisms. Our team has highlighted the fundamental role played by the post-transcriptional regulators Pumilio (PUM) 1/2 on normal HSPC properties. By a shRNA approach, PUM 1/2 knockdown in human and murine HSPCs leads to: a/ a reduced expansion associated with an increased apoptosis and a cell cycle arrest in G0/G1 phase, b/ the loss of their clonogenic capacity and their in vivo reconstitution potential. The objective of our work is to: a/ evaluate the effects of PUM 1/2 overexpression in HSPC, b/ determine PUM1/2 involvement in leukemic processes; c/ investigate the molecular mechanisms responsible of PUM activity in HSPC by identifying protein targets and partners. Our results showed that a moderate overexpression of PUM1 (2 to 3 fold) in normal CD34+ HSPCs limits the loss of their clonogenic potential, while a higher expression (5 to 10 fold or more) is toxic. The expression analysis of PUM1/2 transcripts in Acute Myeloid Leukemia (AML) (GOELAMSthèque) showed a significant increase in the most immature samples (AML0-2) as compared to healthy controls. PUM1/2 knockdown by shRNA in AML cells significantly reduced their survival. The same effect was observed in cell lines from several leukemic processes. We identified various PUM-associated partners by mass spectrometry, Argonaute2 and MOV10 (involved in the miRNA machinery), and the RNA-binding proteins IMP3 (involved in several cancer and in cell cycle regulation) and HuR/ELAV1 (already known to be involved in murine HSPCs maintenance). IMP3 or ELAV1 knockdown in HSPCs in vitro lead to the same effect of a PUM1/2 invalidation, a decreased expansion with an increased apoptosis and the loss of clonogenic potential. Finally, we identify the forkhead box P1 (FOXP1) transcription factor as a new direct target up-regulated by PUM1 and PUM2. Functional study of FoxP1 knockdown by shRNA in HSPCs mimic PUM1/2 activities. Moreover, FOXP1 overexpression partially rescued shPUM antiproliferative and pro-apoptotic effects. Also, the PUM1/2 and FOXP1 expression levels in leukemic primary cells were measured by RT-qPCR and revealed a positive correlation. Our results reveal that PUM1/2 are direct positive regulators of FOXP1 which contributes to the maintenance of normal and leukemic HSPCs.

Les organismes bactériens sont en contact direct avec leur environnement et doivent donc constamment s’acclimater aux variations de celui-ci. Pour cela, plusieurs leviers de régulations peuvent être actionnés. Récemment, la régulation post-transcriptionnelle par les ARN régulateurs a été proposée comme un mécanisme de régulation rapide et peu coûteux pour la cellule. Chez le phytopathogène Dickeya dadantii, la régulation de la virulence a quasi exclusivement été étudiée au niveau transcriptionnel et l’implication des ARN régulateurs dans la virulence reste très peu connue. Pour cela, nous avons tout d’abord étudié le rôle des chaperons à ARN dans la pathogénie de D. dadantii et mis en évidence leur implication dans de nombreux facteurs de virulence comme la production d’enzyme de dégradation de la paroi végétale. Puis, nous avons développé une nouvelle méthode d’identification d’ARN à partir de données RNA-seq. Cette méthode a été développée pour tirer profit des séquençages réalisés en paired-end, permettant de séquencer les deux extrémités d’un transcrit. Son évaluation dans sa capacité à détecter de manière précise des ARN connus a montré une performance supérieure aux méthodes de détection existantes. Enfin, cette nouvelle méthode a été appliquée sur des données de séquençage de petits transcrits. Cette analyse nous a permis d’identifier plus d’un millier d’ARN régulateurs potentiels, dont plusieurs pourraient être impliqués dans la régulation de la virulence. Ces travaux ont donc permis de mettre en lumière l’existence d’une régulation post-transcriptionnelle chez D. dadantii et de proposer des pistes concernant les acteurs et mécanismes concernés / Bacterial organisms are directly exposed to environmental conditions and have to respond to environmental stress. To do so, several regulation network are known. Recently, post transcriptional regulation with small RNAs was suggested to be a fast and cheap in energy regulation mechanism. In the phytopathogen Dickeya dadantii, investigations on pathogenic process mostly focused on its control by transcriptional regulators. Knowledge of post-transcriptional regulation of the virulence factors is still in its infancy.To this end, we first studied the impact of RNA chaperones in the virulence of D. dadantii and showed that they were involved in the regulation of several virulence factors, like production of cell wall degrading enzyme. Then, we developed a new method to detect sRNAs from paired-end bacterial RNA-seq data. This method take paired end sequencing into account, which allow the sequencing of the both ends of each fragment. A comparative assessment showed that this method outperforms all the existing methods in terms of sRNA detection and boundary precision. Finally, this method was applied to sequencing data. With this analysis, more than one thousand sRNAs has been detected, with the identification of several candidates potentially involved in virulence.Thereby, this work highlight the existence of post-transcriptionnal regulation in D. dadantii and suggest candidates and mechanisms involved in this regulation

Les rhabdomyosarcomes embryonnaires et aléolaires (RMS) appartiennent aux tumeurs des tissus mous les plus fréquentes chez les enfants dont elles représentent 2/3 des cas. Plusieurs données suggèrent que la dérégulation des cellules progénitrices du muscle squelettique pourrait jouer un rôle dans l'émergence des cellules de RMS qui ont aussi bien perdu le contrôle de la régulation de la prolifération cellulaire que la capacité à se différencier.Néanmoins les mécanismes de développement des RMS restent à caractériser. La famille des IMPs et notamment IMP-2, protéines liant les ARN, sont à la fois fortement exprimées dans le muscle en régénération in vivo mais aussi dans les cellules de RMS.Au cours de ma thèse, j’ai pu mettre en évidence le rôle d’IMP-2 dans la motilité des cellules de RMS et dans les cellules musculaires ainsi que dans le contrôle de l’intégrité du cytosquelette de microtubules (MTs) et dans le remodelage des adhésions focales. En effet, IMP-2 est impliqué à la fois dans la régulation de l’expression de MuRF-3, une protéine lié àla stabilisation des MTs et de Pinch-2, un important médiateur de l’adhésion cellulaire. / The RNA-binding proteins IMPs (IGF-II mRNA binding protein) first discovered in rhabdomyosarcoma cells (RMS) are expressed during embryonic development but their expression is decreased in adult tissues.We showed that IMPs and particularly IMP-2 are strongly expressed in mouse myoblatsts, during early regeneration of skeletal muscle in vivo and in and RMS. IMP-2 loss of function experiments using siRNA have shown that IMP-2 is necessary for microtubules stability(MTs), cell motility and invasion of myoblasts and RMS.Expression of IMP-2 specifically increases MTs stability by an enrichment of detyrosinated tubulin Glu-tubulin. Detyrosination is indispensable for myogenic differentiation and plays substantial role in tumor growth. Additionaly, MTs stabilization play an important role in focal adhesion remodeling, in cytoskeleton integrity, cell adhesion and cell motility.To get new insight into molecular mechanism underlying the function of IMP-2 in MTs stability and cell motility, full ranscriptome analysis was performed between IMP-2 knockdown (KD) myoblasts and control myoblatsts. We have further shown that IMP-2 controls the mRNA levels of many important mediators of cell adhesion such as PINCH-2, as well as multiple cytoskeleton remodeling, such as MuRF-3.We have identified a number of functionally relevant protein partners of IMP-2.Moreover subsequent RNAi screens have revealed the importance of IMP-2 regulated transcripts involved in cell motility and cell adhesion In conclusion, we show that IMP-2 dependent regulation of mRNA such as MuRF3 and PINCH2 largely contributes to the motility –deficient in IMP-2 KD cells. Moreover these results indicate clearly, that further analysis of IMP2 protein partners and RNA targets regulated by IMP-2 will help to characterized the function of IMP-2 and to propose a model of IMP-2 transcriptional regulation of gene expression in myoblasts and RMS cells.

O iodo em excesso promove o efeito Wolff-Chaikoff. Oligominerais já foram descritos como potenciais reguladores da expressão de proteínas. Tornou-se interessante avaliar se o iodo interferiria com a expressão do mRNA da NIS, em curtos períodos de tempo. Foram realizados, em ratos e células (de 30 min24h), estudos de expressão, comprimento de cauda poli-A e recrutamento para polissomos, do mRNA de NIS. Observou-se, in vivo e in vitro, que o excesso de iodo promoveu diminuição da expressão e do comprimento da cauda poli-A do mRNA de NIS, em todos os períodos estudados, além de promover menor recrutamento deste mRNA para os polissomos. A diminuição da cauda poli-A do mRNA de NIS pode ter aumentado sua instabilidade/degradação e também ter sido responsável por uma menor eficiência de tradução deste transcrito. Conclui-se que: (a) o iodo regula pós-transcricionalmente a expressão gênica da NIS, sendo fundamental nos processos que norteiam o efeito Wolff-Chaikoff e (b) oligoelementos têm relevância na regulação da expressão de proteínas relacionadas ao seu transporte. / Iodide in excess exerts the Wolff-Chaikoff effect. It is described that some minerals can regulate the expression of proteins. This study aimed to investigate if the iodide could modify the expression of NIS mRNA, in short periods of time. Rats and cells, divided in time-groups of 30 min up to 24h, were used in studies of expression, poly-A tale length and polysomal profile of NIS mRNA. Both in vivo and in vitro studies showed that the iodide treatment promoted a reduction in the expression and the poly-A length of NIS mRNA, in all time-groups, and decreased its recruitment to the polysomes. It is possible that the reduction of NIS mRNA poly-A tale length has increased the instability/degradation of this transcript, and impaired the translation efficiency of it. Concluding: a) the iodine exerts a post-transcriptional regulation of NIS mRNA expression, being essencial in the processes that guide the Wollf-Chaikoff effect; b) the oligoelements have an extremely important role in the expression regulation of proteins related to their transport.

Cells sense the properties of the surrounding environment and convert this information into changes in gene expression. Bacteria are, in contrast to many multi-cellular eukaryotes, remarkable in their ability to cope with rapid environmental changes and to endure harsh and extreme milieus. Previously, control of gene expression was thought to be carried out exclusively by proteins. However, it is now clear that small regulatory RNAs (sRNA) also carry out gene regulatory functions. Bacteria such as E. coli harbor a large class of sRNAs that bind to mRNAs to alter translation and/or mRNA stability. By identifying mRNAs that are targeted by sRNAs, my studies have broadened the understanding of the mechanisms that underlie sRNA-dependent gene regulation, and have shed light on the impact that this type of regulation has on bacterial physiology. Control of gene expression often relies on the interplay of many regulators. This interplay is exemplified by our discovery of mutual regulation between the sRNA MicF and the globally acting transcription factor Lrp. Through double negative feedback, these two regulators respond to nutrient availability in the environment which results in reprogramming of downstream gene expression. We have also shown that both the transcription factor CsgD, and the anti-sigma factor FlgM, are repressed by the two sRNAs OmrA and OmrB, suggesting that these sRNAs are important players in the complex regulation that allow bacteria to switch between motility and sessility. Bacterial populations of genetically identical individuals show phenotypic variations when switching to the sessile state due to bistability in gene expression. While bistability has previously been demonstrated to arise from stochastic fluctuations in transcription, our results suggest that bistability possibly may arise from sRNA-dependent regulatory events also on the post-transcriptional level.