La O-GlcNAcylation est une modification post-traductionnelle appartenant au groupe des glycosylations. C’est une modification dynamique, analogue à la phosphorylation, dont la réversibilité est contrôlée par un couple d’enzymes, la O-GlcNAc transférase (OGT) et la O-GlcNAcase (OGA). Bien que les fonctions de la O-GlcNAcylation aient été étudiées à différents niveaux de régulation cellulaire, aucune étude n’a porté jusqu’ici sur sa localisation potentielle au niveau des radeaux lipidiques (rafts). Or, il a été récemment démontré que l’élévation des niveaux de O-GlcNAcylation participait activement au phénomène de résistance à l’insuline et un nombre important de preuves a également été apporté sur le rôle critique des microdomaines lipidiques dans ce phénomène de résistance. On sait également depuis peu qu’après stimulation par l'insuline, l’OGT est recrutée à la membrane plasmique par l'intermédiaire de son domaine PPO (PIP-binding activity of OGT), et que cette relocalisation inactive la voie de signalisation PI3-kinase/Akt par antagonisme vis-à-vis de la phosphorylation. Ces trois observations suggèrent que l’OGT puisse constituer un relai essentiel permettant de transmettre l’information portée par des stimuli externes au travers de la membrane plasmique, et plus particulièrement des rafts, par le biais d’une régulation des voies de signalisation. Au cours de ma thèse, nous avons adapté une technique permettant de purifier les microdomaines lipidiques, en procédant par une extraction par le détergent CHAPSO suivi d'une étape d'ultracentrifugation. Nos expériences ont montré pour la première fois la localisation de l’OGT et de protéines O-GlcNAcylées, parmi lesquelles nous avons identifié la chaîne bêta du récepteur de l'insuline et Hsp70, au niveau des microdomaines lipidiques. Le recrutement de ces éléments est contrôlé par l’insuline: nous avons en effet observé une association progressive de la glycosyltransférase aux microdomaines lipidiques après stimulation par cette hormone. Nous avons démontré que, en réponse à l'insuline, l’OGT est recrutée au niveau des lipides rafts, ce qui suggère une régulation négative de la voie PI3-Kinase/Akt par O-GlcNAcylation, au niveau des microdomaines lipidiques. Cette étude est la première suggérant une relation directe entre la O-GlcNAcylation des protéines associées aux rafts et la régulation de la signalisation cellulaire par l'insuline. En marge de ses travaux, nous avons mis en évidence l'existence de la O-GlcNAcylation chez les deux parasites apicomplexes Toxoplasma gondii et Plasmodium falciparum. Une approche biochimique et bioinformatique a également permis de prouver l'existence propre de l'OGT chez le toxoplasme. / The attachment of O-linked beta-N-acetylglucosamine (O-GlcNAc) to proteins is a post-translational modification belonging to the group of glycosylations. This is a dynamic modification, analogous to phosphorylation, which reversibility is controlled by a couple of enzymes, namely the O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA). Although the functions of O-GlcNAcylation have been studied at different levels of cell regulation, until now none study refers its potential location at the level of lipids microdomains (rafts). Nevertheless, it has been recently shown an elevation of the O-GlcNAcylation levels are actively involved in the phenomenon of insulin resistance and a significant number of publications also reports the critical role of lipid microdomains in this phenomenon of resistance. Little while ago, we know that after stimulation with insulin, the OGT is recruited to the plasma membrane through its PPO domain (PIP-binding activity of OGT), and that this relocation inactivates the PI3-kinase/Akt signaling pathway by exerting an antagonistic effect on its phosphorylation. These three observations suggest that OGT may play an essential role in signal transmission through its interaction with the plasma membrane and more particularly with the lipid microdomains. During my thesis, we first adapted a technique to purify lipid microdomains, using a method combining an extraction with the detergent CHAPSO followed by an ultracentrifugation. Our experiments show for the first time that OGT and some O-GlcNAcylated proteins, among which we identified the insulin receptor beta-chain and Hsp70, are located in lipid rafts. The recruitment of these elements is controlled by insulin: we indeed observed a progressive association of the glycosyltransferase to the lipid microdomains after stimulation by this hormone. We demonstrated that in response to insuline, the OGT is recruited to lipid rafts, suggesting a down-regulation of the PI3-Kinase/Akt pathway by O-GlcNAcylation at the level of the lipid microdomains. This study is the very first suggesting a direct relationship between the O-GlcNAcylation of the lipid rafts-associated proteins and the regulation of cell signaling by insulin. In margin of these works, we have brought to the fore O-GlcNAcylation in the two apicomplexan parasites Toxoplasma gondii and Plasmodium falciparum. A biochemical and bioinformatic study led us to prove that toxoplasma possesses its own OGT.

Les mucines bronchiques des patients atteints de mucoviscidose présentent des modifications de glycosylation qui se traduisent par la sur-expression de motifs sialyl-Lewisx (NeuAcα2-3Galβ1-4[Fucα1-3]GlcNAc-R, sLex) et 6-sulfo-sLex (6-sulfo-sLex), qui sont des récepteurs préférentiels de la bactérie Pseudomonas aeruginosa, responsable de l’infection chronique des patients atteints de mucoviscidose. Nous avons montré que le traitement au TNF d’explants bronchiques humains induisait une surexpression de motifs sLex et 6-sulfo-sLex, et que celle-ci était accompagnée d’une surexpression du gène de l’α2,3-sialyltransférase ST3GAL4. Nous avons pu mettre en évidence un effet similaire du TNF dans la lignée A549, et démontrer l’importance du gène ST3GAL4 dans la synthèse de sLex et 6-sulfo-sLex par interférence ARN. Nous avons ensuite identifié un transcrit majeur, BX répondant au TNF dans les explants bronchiques humains et dans les cellules A549, et cloné la région génomique possiblement impliquée dans le contrôle de l’expression de ce transcrit. Cette région présente une activité promotrice de base et répond au TNF. L’étude de différentes constructions de cette région et d’inhibiteurs pharmacologiques de voies de signalisation ont ensuite été utilisé dans le but de comprendre les mécanismes moléculaires et cellulaires conduisant à l’augmentation de l’expression de ST3GAL4 par le TNF. Enfin, nous avons étudié avec la lignée cellulaire NCI-H292 l’effet du TNF sur l’adhésion de différentes souches de Pseudomonas aeruginosa. Nos résultats préliminaires indiquent que le traitement par le TNF induit une surexpression de ST3GAL4, du sLex, et une augmentation de l’adhésion des souches PAO1 et PAK, dépendante de l’adhésine FliD, et de la présence d’acide sialique. / Bronchial mucins from patients suffering from cystic fibrosis (CF) exhibit glycosylation alterations, especially increased amounts of sialyl-Lewisx (NeuAcα2-3Galβ1-4[Fuc[α1-3]GlcNAc-R, sLex) and 6-sulfo-sialyl-Lewisx (6-sulfo-sLex) structures. These epitopes are preferential receptors for Pseudomonas aeruginosa, the bacteria responsible for the chronicity of airway infection of CF patients. Several studies have shown that inflammation may affect glycosylation and sulfation of glycoproteins, including mucins. We previously proposed that the inflammatory context in the respiratory tract of CF patients was responsible for the overexpression of sLex and 6 sulfo-sLex structures. Indeed, we and others have shown that pro-inflammatory cytokines can modulate glycosylation in the bronchial mucosa and in bronchial cell lines. In particular, TNF, IL-6 and IL-8 increase the expression and activity of several members of the sialyl-, fucosyl- and sulfotransferases families in human bronchial explants. During my thesis, I showed that TNF regulates the expression of a specific transcript (BX) of the ST3GAL4 gene that encodes the main α2,3-sialyltransferase (ST3Gal IV) involved in the biosynthesis of sLex and 6-sulfo-sLex epitopes in both human bronchial explants and A549 lung cancer cells. A 2 kb genomic region was cloned and we have shown by reporter gene experiments that this region could be responsible for both basal BX transcript expression and response to TNF treatment in A549 cells. Different genetic constructions of this sequence and the use of pharmacological inhibitors allowed us to study the molecular and cellular mechanisms involved in ST3GAL4 up-regulation induced by TNF treatment. In parallel, we have studied the effect of TNF-induced sLex and 6- sulfo -sLex overexpression on P. aeruginosa adhesion, using the NCI-H292 pulmonary carcinoma cell model. Our preliminary results indicate that TNF treatment induces an enhance expression of sLex derivatives that correlates with an enhance adhesion of both PAO1 and PAK strain. This increased adhesion is dependant of the presence of a functional FliD protein, the bacterial lectin recognizing sLex and 6- sulfo -sLex, and is abolished by sialidase treatment. Identifying mechanisms by which sLex and 6-sulfo-sLex structures are overexpressed on bronchial mucins from CF patients may allow the identification of new targets to fight bronchial inflammation and infection by P. aeruginosa.

Metallothioneins (MT) are small metal-binding proteins. Metallothionein isoform MT-1 alters localization from cytoplasm to nucleus as cells enter S phase. The perinuclear localization of MT-1 mRNA is essential for this nuclear localization. The localization of the MT-1 mRNA requires a signal in the 3'untranslated region (3'UTR). This thesis describes studies of the localization signal in the MT-1 3'UTR and its binding to transacting cellular proteins.

Dystroglycan is a heterodimeric cell adhesion molecule consisting of alpha and beta subunits, which links the actin cytoskeleton to the extracellular matrix. Dystroglycan has an important structural role and is integral for maintaining tissue integrity, but new avenues have research have implicated it in other roles within the cell due to its ability to bind a number of interacting proteins including cytoskeletal components and proteins involved in signal transduction. To elucidate the functions of dystroglycan at the cellular level, fibroblasts were differentiated from dystroglycan null embryonic stem cells. Analysis of these cells by western blot, however, showed them to be expressing dystroglycan and this was confirmed by RT-PCR, genomic PCR and northern blot. Subsequently, Swiss 3T3 fibroblasts with greater than 60% reduction in dystroglycan expression were generated by stable retroviral infection of a shRNA construct. Dystroglycan deficient cells were smaller and were found to have a reduction in cell-substrate adhesions, when compared to a control cell line. Dystroglycan deficiency did not affect cell motility or polarity in fibroblasts, but did result in a reduction in the rate of cell proliferation, which was thought to be the consequence of an increase in apoptosis. Dystroglycan deficiency also inhibited the formation of filopodia and lamellipodia in response to dominant active Cdc42 or Rac1 respectively, suggesting that it an important mediator of Rho GTPase-mediated cytoskeletal rearrangements. This study has also identified a possible new function for dystroglycan during cytokinesis. Endogenous dystroglycan was localised to the cleavage furrow and midbody of dividing fibroblasts, where it was found to co-localise with the ERM family member, ezrin. Expression of mutated and truncated dystroglycan-GFP constructs in dividing HeLa cells revealed that the cytoplasmic domain is required for cleavage furrow localisation, but ezrin-binding is not. A model for dystroglycan function at the cleavage furrow is presented whereby dystroglycan recruits ezrin to the site of furrow ingression and together they serve to tether the actomyosin contractile ring to the plasma membrane during cytokinesis. These findings provide further insight into the multifaceted functions of dystroglycan within the cell and also raised interesting questions about possible new roles for dystroglycan.

La deltalactoferrine (ΔLf) est un facteur de transcription qui possède des propriétés anti-tumorales. Son activité et sa stabilité sont hautement contrôlées par des modifications post-traductionnelles comme la O-GlcNAcylation et la phosphorylation. Au cours de ma thèse, nous avons pu montrer que la ΔLf était modifiée par SUMO-1 et mettre en évidence 5 sites de SUMOylation. Nous avons produit un ensemble de mutants pour lesquels un seul site était préservé et un mutant M5S invalidé pour les cinq sites de SUMOylation. Nous avons pu montrer que cinq lysines étaient la cible de la machinerie de SUMOylation et que K13, K308 et K379 étaient les sites accepteurs majeurs. Nous avons ensuite étudié l’activité transcriptionnelle des différents mutants et montré que la SUMOylation réprime l’activité transcriptionnelle de la ΔLf. Dans un second temps nous avons étudié le dialogue entre différentes modifications post-traductionnelles. Nous avons pu démontrer qu’une compétition SUMO/ubiquitine existait et que sous la forme SUMOylée le mutant K379 avait une durée de ½ vie plus longue confirmant que la SUMOylation protégeait la ΔLf de la dégradation protéasomale. Nous avons également pu montrer que la ΔLf était acétylée principalement sur la lysine K13. En modulant le niveau de SUMOylation ou d’acétylation du mutant K13 nous avons pu montrer que la compétition SUMO/acétylation était impliquée dans la régulation de l’activité transcriptionnelle de la ΔLf. En effet, l’acétylation régule positivement l’activité transactivatrice de la ΔLf alors que la SUMOylation l’inhibe. En résumé, nous avons pu montrer que la ΔLf était multi-SUMOylée et que la SUMOylation était un nouveau mécanisme de régulation à la fois de l’activité transcriptionnelle et de la stabilité de la ΔLf. En parallèle, nous avons montré que la ΔLf transactivait le gène Bax déclenchant l’apoptose et que la balance O-GlcNAc/Phosphate modulait la transactivation de Bax. / Deltalactoferrin (ΔLf) is a transcription factor which possesses antitumoral activities. Posttranslational modifications such as O-GlcNAcylation and phosphorylation, efficiently modulate its transcription factor activity and stability. During my PhD thesis we first showed that ΔLf is modified by SUMO-1 and mapped the five SUMO sites. We produced a series of mutants for which only one site was preserved and a null-mutant in which all five SUMO sites were invalidated. We showed that all lysine residues were SUMO acceptors and that K13, K308 and K379 were the main SUMO sites. We next studied the impact of SUMOylation on ΔLf activity and showed that SUMOylation negatively regulated the transactivation function of ΔLf. During the second part of my PhD, we investigated the crosstalk between different posttranslational modifications. We showed that K379 which is either ubiquitinated or SUMOylated, is a pivotal site for the control of ΔLf stability. We also showed that SUMOylation competes with ubiquitination and protects ΔLf from proteosomal degradation by positively regulating its stability. We demonstrated that K13 is the main acetylation site and that favoring acetylation at K13 reduced SUMOylation and increased ΔLf transcriptional activity. Collectively, our results indicate that multi-SUMOylation occurs on ΔLf to repress its transcriptional activity. Reciprocal occupancy of K13 by either SUMO-1 or an acetyl group may contribute to the establishment of finely regulated mechanisms to control ΔLf transcriptional activity. Moreover, competition between SUMOylation and ubiquitination at K379 coordinately regulates the stability of ΔLf toward proteolysis. Therefore SUMOylation of ΔLf is a novel mechanism controlling both its activity and stability. In parallel, we demonstrated that ΔLf transactivates the Bax promoter leading to the triggering of apoptosis and that the O-GlcNAc/Phosphate interplay controls Bax transactivation. Lf) is a transcription factor which possesses antitumoral activities. Posttranslational modifications such as O-GlcNAcylation and phosphorylation, efficiently modulate its transcription factor activity and stability. During my PhD thesis we first showed that ΔLf is modified by SUMO-1 and mapped the five SUMO sites. We produced a series of mutants for which only one site was preserved and a null-mutant in which all five SUMO sites were invalidated. We showed that all lysine residues were SUMO acceptors and that K13, K308 and K379 were the main SUMO sites. We next studied the impact of SUMOylation on ΔLf activity and showed that SUMOylation negatively regulated the transactivation function of ΔLf. During the second part of my PhD, we investigated the crosstalk between different posttranslational modifications. We showed that K379 which is either ubiquitinated or SUMOylated, is a pivotal site for the control of ΔLf stability. We also showed that SUMOylation competes with ubiquitination and protects ΔLf from proteosomal degradation by positively regulating its stability. We demonstrated that K13 is the main acetylation site and that favoring acetylation at K13 reduced SUMOylation and increased ΔLf transcriptional activity. Collectively, our results indicate that multi-SUMOylation occurs on ΔLf to repress its transcriptional activity. Reciprocal occupancy of K13 by either SUMO-1 or an acetyl group may contribute to the establishment of finely regulated mechanisms to control ΔLf transcriptional activity. Moreover, competition between SUMOylation and ubiquitination at K379 coordinately regulates the stability of ΔLf toward proteolysis. Therefore SUMOylation of ΔLf is a novel mechanism controlling both its activity and stability. In parallel, we demonstrated that ΔLf transactivates the Bax promoter leading to the triggering of apoptosis and that the O-GlcNAc/Phosphate interplay controls Bax transactivation.

- / -

L’épithélium digestif est recouvert par une épaisse couche de mucus majoritairement composé de mucines qui joue un rôle essentiel dans le maintien de l’homéostasie avec la microflore résidente et dans la protection de l’épithélium digestif contre la colonisation par des organismes pathogènes. Au cours de ma thèse, je me suis intéressée à l’étude des interactions hôtes / bactéries dans le tractus gastro-intestinal. En utilisant différents modèles in vivo ou in vitro, nous avons plus particulièrement étudié l’influence de bactéries commensales et pathogènes sur la glycosylation des mucines du tractus digestif. Nous avons tout d’abord caractérisé le profil de glycosylation des mucines gastriques humaines d’individus sains, de patients asymptomatiques infectés par Helicobacter pylori et de patients atteints de métaplasie intestinale incomplète. Afin de mieux décrypter le dialogue qui s’établit entre les mucines et les bactéries, nous avons également caractérisé les modifications de glycosylation des mucines intestinales à partir de modèles animaux et cellulaires d’infections bactériennes. Nous avons étudié l’effet de deux bactéries commensales : Bacteroides thetaiotaomicron et Faecalibacterium prausnitzii et d’une bactérie pathogène : Shigella flexneri. Les mucines apparaissent comme un élément incontournable dans l’interaction hôte/bactéries. Nos travaux indiquent que la barrière mucosale, même si elle exerce un rôle de défense indéniable pour l’épithélium digestif, n’est pas infaillible vis-à-vis des pathogènes. Ils soulignent l’importance du dialogue tripartite entre les mucines, le microbiote et le système immunitaire dans la lutte contre les infections. / The gut epithelium is covered by a thick layer of mucus mainly composed of mucins which plays an essential role in maintaining homeostasis with the resident microflora but it also protects the digestive epithelium against colonization by pathogens. During my PhD work, I focused my interest in the study of host / bacteria interactions in the gastrointestinal tract. Using different in vivo or in vitro models, we specifically investigated the influence of commensal and pathogenic bacteria on the glycosylation of mucins of the digestive tract. We first characterized the glycosylation profile of human gastric mucins from healthy, from asymptomatic patients infected with Helicobacter pylori and from patients with incomplete intestinal metaplasia. To better understand the cross-talk between mucins and bacteria, we also studied changes in the glycosylation of intestinal mucins from animal and cellular models of bacterial infections. We analyzed the effect of two commensal bacteria: Bacteroides thetaiotaomicron and Faecalibacterium prausnitzii, and one pathogenic bacteria: Shigella flexneri. Mucins appear as a key element in the host / bacteria interactions. Our work indicates that the mucosal barrier, even though it plays a major role in the defense of gastrointestinal epithelium, is not infallible regarding pathogens. Our work emphasizes the importance of the tripartite dialogue between mucins, the microbiota and the immune system in the fight against infections.

CD23 is a Type II glycoprotein that exists in both membrane-bound (mbCD23) and soluble (sCD23) forms and functions as the low affinity receptor for IgE. CD23 interacts with a range of proteins to fulfil its function in vivo. Through interactions with its ligands IgE and CD21, CD23 regulates the levels of IgE in serum. Soluble CD23 also interacts with members of the integrin family of membrane receptors. Integrins are heterodimeric transmembrane receptors that participate in bidirectional signalling across membranes and have a critical role in cellular adhesion reactions. CD23 interacts with four integrins, αVβ3 and αVβ5 from the αV integrin family, and αMβ2 and αXβ2, from the β2 family of integrins. Using peptide array technology, we identified set of overlapping peptides derived from the soluble CD23 sequence that interact with integrins expressed on the surface of monocytic cells and with purified αVβ3 and αVβ5 integrins in in vitro Biacore assays. These peptides all contained a common basic tripeptide motif, termed the Arg-Lys-Cys (RKC) motif. Integrins traditionally recognise and bind the Arg-Gly-Asp (RGD) tripeptide in their ligands in a cation-dependent manner. However, the in vitro interaction between RKC-containing peptides and purified integrins was determined to be cation-independent, salt-sensitive and independent of RGD binding. The interaction was blocked by full length CD23. Substitution of the residues in the RKC motif reduced or abolished binding of the peptides to integrins expressed on cells and in vitro, as measured by Biacore analysis, and abolished the competition with CD23. Taken together, these data suggest that the RKC motif is the site in CD23 that is recognised and bound by αVβ3 and αVβ5 integrins. The RKC motif can be considered a novel recognition motif for integrins, as it is cation-independent, and its binding is not blocked by the presence of RGD-containing integrin ligands. Therefore it is likely that the RKC motif interacts with integrins at a site other than that used for RGD-binding, similar to the interactions that have been described for the binding of the HIV Tat protein. The interaction between sCD23 and its integrin receptors is important in the regulation of cytokine production by monocytic cells. Most monocytic cells will express a combination of the different CD23-binding integrins simultaneously and, therefore, the cytokine output of that cell will be the net result of the interaction of CD23 with a combination of integrins. We used monoclonal antibodies to investigate the role of individual integrins in cytokine production. Antibodies were selected to allow comparision of the cytokine response between, a) integrin families, b) integrin subunits, and c) integrin epitopes. The cytokine profile induced by integrin ligation did not differ between the αV and β2 integrin families, although the concentration of cytokines produced varied depending on the heterodimer targeted. However, within a particular family, cytokine production induced by integrin ligation was specific and relied on the recognition of a precise epitope on the integrin. Cytokine production by CD23-binding integrins appears to require the ligation of the α subunit of the integrin heterodimer. We identified an antibody directed against the αVβ3 integrin that induced high levels of cytokine production. Cytokine production following ligation of the integrin with this antibody was dependent on activation of the ERK/MAPK pathway in cells. This production of cytokines and phosphorylation of ERK was enhanced by the addition of macrophage colony stimulating factor (M-CSF) and partially inhibited by an anti-TLR-2 antibody. Chronic stimulation (<3 days) of THP-1 cells with the anti-αVβ3 antibody in the presence of M-CSF led to morphological changes in the cell line associated with the development of a more macrophage-like phenotype and the continued production of cytokines. Analysis of the changes in cell surface marker expression and cytokine profiles suggested that the THP-1 cells had undergone an M2b programme of macrophage activation in response to αVβ3 ligation. Data presented herein reinforce the importance of the role of integrins in the control of adhesion-independent signalling pathways in suspension cells.

Les Oligosaccharides soluble (OS) sont essentiellement générés durant le processus de dégradation des N-glycoprotéines nouvellement synthétisées et mal conformées (ERAD) et durant la voie « turn-over » des glycoprotéines matures. Dans le but de déterminer si la modification post-traductionnelle O-GlcNAc est effectivement impliquée dans le processus de dégradation des N-glycoprotéines, nous avons analysé par spectrométrie de masse les OS provenant des cellules CHO après traitement par l’inhibiteur PUGNAc. Le PUGNAc ou « O-(2-acetamido-2-deoxy-D-glucopyranosylidene) amino-N-phenylcarbanate » est un inhibiteur puissant de l’O-GlcNAcase (OGA) qui catalyse l’hydrolyse du résidu O-GlcNAc des résidus sérine et thréonine des protéines O-GlcNAcylées.L’analyse par spectrométrie de masse révèle l’apparition d’une population d’OS de structures anormaux dans les cellules CHO suite au traitement par le PUGNAc. Cette population a été identifiée comme ayant des structures possédant des résidus GlcNAc au niveau de leur extrémité non-réductrice issues d’une dégradation lysosomale incomplète des glycoprotéines. Contrairement au PUGNAc, le NButGT, un autre inhibiteur de l’OGA, n’aboutit pas à l’apparition de cette population. Ainsi, Nous avons démontré que ces structures s’accumulent exclusivement dans la fraction membranaire conséquence de l’inhibition des β-hexosaminidases lysosomaux par le PUGNAc. Notre étude avait permis, d’un part, de mettre en évidence la capacité du PUGNAc de mimer une maladie de surcharge lysosomale et, d’autre part, de montrer un autre aspect des effets indésirables induits par le PUGNAc et qui nécessite d’être pris en considération lors de l’utilisation de cet inhibiteur. / Free oligosaccharides (fOS) are generated as a result of glycoproteins catabolism that occurs in two principal distinct pathways: the endoplasmic reticulum-associated degradation (ERAD) of misfolded newly synthesized N-glycoproteins and the mature N-glycoproteins turnover pathway. We analyzed fOS by Mass spectrometry in PUGNAc CHO treated cells in order to investigate whether O-GlcNAc modified proteins were involved in N-glycoprotein degradation process. The O-(2-acetamido-2-deoxy-D-glucopyranosylidene) amino-N-phenylcarbanate (PUGNAc) is a potent inhibitor of the O-GlcNAcase (OGA) catalyzing the cleavage of β-O-linked 2-acetamido-2-deoxy-beta-D-glucopyranoside (O-GlcNAc) from serine and threonine residues of post-translationally modified proteins. Mass spectrometry (MS) analysis revealed the appearance of an unusual population of fOS after PUGNAc treatment. The structures representing this population have been identified as containing non-reducing end GlcNAc residues resulting from incomplete lysosomal fOS degradation. Only observed after PUGNAc treatment, the NButGt, another OGA inhibitor, did not lead to the appearance of the population. These structures have clearly been shown to accumulate in membrane fractions as the consequence of lysosomal β-hexosaminidases inhibition by PUGNAc. As Lysosomal Storage Disorders (LSD) are characterized by the accumulation of fOS in various tissues, our study evokes that PUGNAc mimics a LSD and shows another off target effects that needs to be taken into account in the use of this drug.

La O-GlcNAcylation est une glycosylation dynamique et réversible sous le contrôle de la O-GlcNAc Transférase (OGT) qui transfère un résidu de GlcNAc sur les Ser/Thr de protéines intracellulaires, et de la O-GlcNAcase (OGA). Plusieurs travaux dont ceux de notre équipe ont montré l'importance de la dynamique de O-GlcNAcylation pour la progression normale du cycle cellulaire, et plus particulièrement de la mitose. L’objectif de mes travaux de thèse était de comprendre comment la balance O-GlcNAc participe au contrôle des étapes précoces du cycle cellulaire. J’ai d’abord montré dans différentes lignées cellulaires que l’entrée en phase S s’accompagne d’une baisse marquée du niveau de O-GlcNAc, corrélée à une augmentation de l’expression et de l’activité de l’OGA endogène. Par protéomique, 58 protéines cytosoliques et nucléaires différentiellement O-GlcNAcylées à la transition G1/S ont ensuite été identifiées dans les cellules MCF7 synchronisées. Ces protéines interviennent dans des processus cellulaires essentiels à la phase G1 dont la régulation de la transcription, de la traduction et de la mise en conformation des protéines, et de la réplication de l’ADN. Par immunoprécipitation, les variations O-GlcNAc dépendantes du cycle cellulaire ont été confirmées sur les protéines cytosoliques CK8, hnRNP K et Caprine 1, et sur les protéines nucléaires du complexe de pré-réplication, MCM-3, -4, -6, et -7. Ces travaux montrent donc que la transition G1/S est étroitement liée à la dynamique de O-GlcNAcylation et soulignent un rôle potentiel de cette glycosylation dans le contrôle de l’initiation de la réplication de l’ADN et par là même, dans le maintien de l'intégrité génomique. / O-GlcNAcylation is a highly dynamic and reversible glycosylation which is governed by O-GlcNAc Transferase (OGT) that transfers the N-acetylglucosamine (GlcNAc) residue onto Ser/Thr of intracellular proteins, and O-GlcNAcase (OGA). Over the last decade, we and others have shown that dynamics of O-GlcNAcylation was important in regulating the cell cycle progression, and more particularly the mitosis events. The aim of my work was to explore how O-GlcNAc balance is implicated in the control of cellular proliferation by focusing on the early steps in the cell cycle. We highlighted in several cell lines that S-phase entry is associated with a marked decrease in the overall level of O-GlcNAcylated proteins, concordant with an increase in both the expression and activity of endogenous OGA. Then, using a proteomic approach we identified 58 cytoplasmic and nuclear proteins differentially O-GlcNAcylated between G0, G1 and S phases. These proteins are involved in key cellular functions that are essential for G1 and S progression, such as protein folding and translation, transcription or DNA replication. By immunoprecipitation, we further confirmed the cell cycle-dependent O-GlcNAc variations of CK8, hnRNP K, Caprin-1, and MCM -3, -4, -6, and -7 proteins which are part of the pre-replicative complex. To conclude, this study shows that there is a close link between the dynamics of O-GlcNAc and G1/S transition and provides a descriptive overview of differentially O-GlcNAcylated proteins at the G1/S transition, highlighting a potential role of O-GlcNAcylation in the initiation of DNA synthesis and therefore, in the maintenance of genome integrity.