Les domaines SH3 (Src Homology domain) représentent l'un des modules protéiques le plus largement répandu dans la nature. Ils participent à des interactions intra- et intermoléculaires avec d’autre partenaires au travers de la formation et de la dissociation de complexes multi-protéiques. Le gène nef du Virus d'Immunodéficience Humain (VIH-1) code pour la protéine nef, importante pour la réplication du virus et le développement optimal du SIDA (Syndrome d’Immunodéficience Acquise) chez les personnes infectées. De précédentes études ont mis en évidence que la protéine nef utilise un mode « tertiaire » d’interaction pour mettre en place une affinité et une sélectivité élevées envers les domaines SH3 des kinases de la famille Src (SFKs). Savoir si cette stratégie de reconnaissance tertiaire des domaines SH3 peut être retrouvée dans des protéines cellulaires humaines est donc une question importante pour évaluer le degré de spécificité de la protéine nef comme cible anti-HIV. Nous avons identifié Alix (ALG-2 [apoptosis-linked gene 2]-interacting protein X) comme protéine originale interagissant avec le domaine SH3 de la kinase de cellules Hématopoïétique (Hck). Alix possède une sélectivité comparable à nef envers les domaines SH3 de SFKs. Nous avons combiné une analyse biophysique et structurale, alliant des méthodes telles que la microcalorimetrie à titration isotherme(‘ITC’), la Résonance Plasmonique de Surface (‘SPR’), des méthodes in vitro dites de ‘GST pulldown’, l'interférométrie (‘NPOI’), la Résonance Magnétique Nucléaire (‘NMR’ - HSQC) et la diffusion des rayons X aux petits angles (SAXS) pour explorer les caractéristiques définissant le mode d’interaction entre Alix et le domaine SH3 de la kinase Hck. Cette étude démontre que la protéine cellulaire Alix est unique, structurellement différente mais fonctionnellement semblable à nef. / Src homology (SH) 3 domains is one of the most wide-spreaded protein modules found in nature. They mediate both inter- and intra-molecular protein-protein interactions (PPIs) through the formation and dissociation of multi-protein complexes. These SH3-mediated interactions are responsible for signal transduction, cytoskeleton organization and other cellular processes. The nef gene of Human immunodeficiency virus (HIV-1) encodes the HIV-1 Nef protein, which is important for optimal virus replication and development of AIDS (acquired immunize deficiency syndrome) in HIV-1 infected persons. Previous studies show that the HIV-1 Nef protein uses a “tertiary” binding mode to achieve high affinity and selectivity toward SH3 domains of Src-family kinases (SFKs). Whether this strategy of ‘tertiary’ binding mode of SH3 domains can be found in human cellular proteins, besides HIV-1 Nef, is an important question in the specificity of the HIV-1 Nef protein as an anti-HIV target. We identified Alix (ALG-2 [apoptosis-linked gene 2]-interacting protein X) as a novel protein interacting with Hemopoietic cell kinase (Hck) SH3 domain. Alix has similar selectivity towards SH3 domains of SFKs as the HIV-1 Nef. We have combined biophysical and structural biology analysis, including ITC (isothermal titration calorimetry), SPR (surface Plasmon resonance), GST (glutathione S-transferase) pull-down, interferometry, HSQC (heteronuclear single quantum coherence) and SAXS (small-angle X-ray scattering) to explore the characteristics of Alix-SH3 recognition mode. This study shows that Alix as a unique cellular protein, which is structurally different but functionally similar in recognizing HIV-1 Nef. The structural information of the Alix-Hck association facilitates the understanding of how Hck and Alix assist viral budding and cell surface receptor regulation.

Src kinase plays a crucial role in a multitude of fundamental cellular processes. Src is an essential component of signalling pathways controlling cellular proliferation, motility or differentiation, and is often found deregulated in tumours. Src activity is therefore maintained under stringent and complex regulation mediated by SH3 and SH2 domains and the phosphorylation state of tyrosines 416 and 527. Active Src adopts an open conformation whereas inactive state of the kinase is characterised by a compact structure stabilised by inhibitory intramolecular interactions. We identified phosphorylation of tyrosine 90 within binding surface of SH3 domain as a new regulatory switch controlling Src kinase activation. Using substitutions mimicking phosphorylation state of the residue we demonstrated that tyrosine 90 phosphorylation controls Src catalytic activity, conformation and interactions mediated by the SH3 domain, representing a positive regulatory mechanism leading to elevated activation of mitogenic pathways and increased invasive potential of cells. Based on correlation between compactness of Src structure and its catalytic activity, we constructed a FRET-based sensor of Src conformation enabling to measure the dynamics of Src activation in cells with spatio-temporal resolution. We found that...

Protein CAS is a major tyrosine-phosphorylated protein in cells transformed by v-crk and v-src oncogenes. It is a multidomain adaptor protein, which serves as a scaffold for assembly of signalling complexes which are important for migration and invasiveness of Src-transformed cells. A novel phosphorylation site in N-terminal SH3 domain was identified - tyrosine 12 located on binding surface of CAS SH3 domain. To study biological importance of tyrosine 12 phosphorylation, non-phosphorylable (Y12F) and phosphomimicking ( Y12E) mutant of CAS were prepared. We found that phosphomimicking mutation Y12E leads to decreased interaction of CAS SH domain with kinase FAK a phosphatase PTP-PEST and also reduce tyrosine phosphorylation of FAK. Using GFP-tagged CAS protein, we show that Y12E mutation caused delocalization of CAS from focal adhesion but has no effect on localization of CAS to podosome-type adhesion. Non-phosphorylable mutation Y12F cause hyperphosphorylation of CAS substrate domain and decrease turnover of focal adhesion and associated cell migration of mouse embryonal fibroblasts (MEFs) independent to integrin singalling. Analogically to migration, CAS Y12F decrease invasiveness of Src-transformed MEF. The results of this diploma thesis show that phosphorylation of Tyr12 in CAS SH3 domain is...

No description available.

L’identification de molécules organiques capables de moduler des interactions protéine-protéine (PPIs) est longtemps restée un domaine peu exploité par la recherche pharmaceutique privée comme académique. Cependant, le développement de méthodologies innovantes pour l’étude des PPIs et la validation récente de ce type d’inhibiteurs dans des essais précliniques, démontrent que les PPIs constituent une nouvelle source de cibles importantes. Les composés capables de moduler ces interactions représentent une nouvelle classe d’outils prometteurs, tant en recherche fondamentale qu’en thérapeutique. Elles peuvent aider à différencier les multiples fonctions portées par une même protéine, à replacer la protéine dans une cascade de réactions, ainsi qu’à disséquer et reconstituer des réseaux de signalisations protéiques. Ces molécules permettront également de faire émerger de nouvelles familles d’agents pharmacologiques actifs dans diverses pathologies.Mon travail de thèse s'est projeté dans l'avenir de la recherche biomédicale, en ciblant les interactions protéine-protéine. J’ai pu durant mon doctorat mettre en œuvre plusieurs méthodologies pour étudier et caractériser des interactions protéiques afin de développer des inhibiteurs de ces interactions. J’ai ainsi pu travailler sur l’optimisation d’un composé inhibiteur de l’interaction de la protéine virale Nef VIH-1 avec les domaines SH3 des Src kinases, le composé DLC27. J’ai également pu mettre en évidence la pertinence biologique de ce composé, qui cible un mode d’interaction unique, ou très rare, au niveau cellulaire en étudiant l’interaction avec les domaines SH3 de deux protéines, ALIX (ALG2-Interacting Protein X) et la sous-unité p85 de la PI3K (phosphatidylinositol 3-kinase).J’ai également pu caractériser la surface et le mode d’interaction de protéines virales impliquées dans le complexe de réplication du virus du SRAS (Syndrome Respiratoire Aigu Sévère). Cette étude tend à montrer que la protéine virale nsp10 agit comme une plateforme de reconnaissance pour ses partenaires, les protéines virales nsp14 et nsp16. Ces interactions permettent l’activation ou l’augmentation des activités respectives de nsp16 et nsp14 et jouent un rôle au niveau de la réplication virale. Suite à l’identification d’un ‘point chaud’ d’interaction, le résidu Tyr96 à la surface de nsp10, nous avons mis en évidence la première famille de molécules inhibitrices du complexe nsp10-nsp14 en couplant des méthodes informatiques (in silico) à des criblages expérimentaux. Ces molécules pourraient être utilisées comme antiviraux ou servir d’outils pour la recherche, en permettant par exemple de mieux comprendre et d’élucider les mécanismes moléculaires impliqués dans la réplication du virus du SRAS et des coronavirus en général. / Protein-protein interactions (PPIs) participate in and regulate almost all essential cellular functions. As a consequence, they are frequently involved in various pathologies (going from cancer development to viral replication and host cell infection) but their study remains a challenge.Thus understanding those interactions as well as finding small drug candidates able to modulate them, a field of research not currently fully developed, appear as the future of the healthcare industry.In this context, I chose to learn different techniques to study PPIs that are usually employed in academic (IMR laboratory, CNRS, France) or corporate environments (Genentech, USA). Moreover, I also worked on the development of small organic inhibitors of PPIs coupling in silico methodologies (chemo-informatics, Drug Design) to biological and structural validations.During my PhD, I could manage and work on different projects involving the study of PPIs involved in cancer signaling pathways as well as the development of potent antiviral drugs targeting the HIV and SARS viruses.My organizational, personal and scientific skills as well as the practical experience I developed on various techniques (from cell biology to biophysics, structural biochemistry and Drug Design), make me feel confident on the management of PPIs drug discovery projects.I am thus able to efficiently work on, and manage, the study of protein-protein interactions in various pathologies as well as the development of potent PPIs inhibitors, that will be a major breakthrough for Biotech/Pharma companies in the coming years.

L’usage excessif d’antibiotiques a provoqué l’émergence de résistance, constituant un problème sanitaire mondial. L’antibiotique triméthoprime (TMP) inhibe l’enzyme dihydrofolate réductase (FolA) des bactéries, interrompant la production d’un précurseur essentiel dans la synthèse des purines et empêchant ainsi la croissance bactérienne. Cependant, certaines bactéries produisent une seconde dihydrofolate réductase : une DfrB, appartenant à une famille d’enzymes hautement résistantes au TMP. Actuellement, dix membres de la famille DfrB ont été identifiés, qui partagent une identité de séquence élevée (74 – 98 %). Les enzymes DfrB sont constituées de domaines identiques de 78 acides aminés, de type ‘SH3-like’, qui s’homotétramérisent afin de former l’enzyme active. Les DfrB ne partagent aucune homologie de séquence ou de structure avec les FolA et aucun antibiotique n’a encore été développé pour contourner la résistance au TMP causée par les DfrB. Afin de mieux comprendre le domaine SH3-like, des homologues (DfrB-H) partageant 10 à 80 % d’identité avec la DfrB1 ont été identifiés et caractérisés. Ils possèdent une activité dihydrofolate réductase (Dfr) et confèrent de la résistance au TMP. De plus, afin de vérifier si les gènes dfrB se retrouvent dans divers environnements, une recherche dans une base de données métagénomiques a été entreprise, permettant de caractériser 10 nouvelles séquences homologues aux DfrB connues. En 2012, le groupe Pelletier a rapporté le premier inhibiteur spécifique d’une DfrB, et plusieurs autres depuis. Seule la DfrB1 a été caractérisée concernant son profil d’inhibition ainsi que sa thermostabilité inhabituelle. Ici, une méthode semi-automatisée sera développée pour caractériser les profils d’inhibition, de thermostabilité, de résistance au TMP et d’activité enzymatique de toutes les DfrB et des homologues identifiés, afin de les comparer à ceux de la DfrB1. Pour atteindre ces objectifs, des nouvelles méthodes à haut débit de détermination d’activité ainsi que des tests de concentration minimale inhibitrice (CMI) furent développés. Ces méthodes ont permis de déterminer que les profils de thermostabilité et d’inhibition de plusieurs DfrB et DfrB-H sont comparables aux profils de la DfrB1. De plus, le criblage de dizaines de composés potentiellement inhibiteurs a été effectué afin de poursuivre la recherche d’inhibiteurs spécifiques aux DfrB. En outre, nous signalons 10 nouvelles séquences homologues de DfrB qui confèrent une résistance élevée au TMP et possèdent une activité Dfr. La caractérisation de tous les membres DfrB et les homologues nous permettra d’acquérir une meilleure connaissance de leur mécanisme de résistance, de leur prévalence dans divers environnements et de soutenir notre développement de nouveaux inhibiteurs des DfrB. / The intensive usage of antibiotics has provoked the emergence of antibiotic resistance, causing a worldwide health issue. The antibiotic trimethoprim (TMP) targets the microbial dihydrofolate reductase enzyme (FolA), abrogating the production of an essential precursor in the synthesis of purines and thus preventing bacterial proliferation. However, some bacteria produce an additional dihydrofolate reductase: the highly TMP-resistant DfrB. Currently, ten DfrB family members have been identified, that share high sequence identity (74 – 98 %). DfrB enzymes consist of identical, 78 amino acid-long SH3-like domains, that homotetramerize to form the active enzyme. DfrB share no sequence or structural homology with FolA and no antibiotic has yet been developed to circumvent the TMP resistance caused by DfrB. In order to gain insight into the SH3-like domain of DfrB, homologues (DfrB-H) sharing 10 to 80 % identity with DfrB1 were identified and characterized, which displayed dihydrofolate reductase (Dfr) activity and conferred high TMP resistance. Also, to investigate if dfrB genes are identified in various environments, a metagenomic database search was undertaken to characterize ten new DfrB1 homologue sequences. In 2012, the Pelletier group reported the first specific inhibitor of a DfrB, and several others since. Only DfrB1 has been characterized regarding its inhibition profile as well as its unusual thermostability. Here, semi-automated methods will be developed to compare the inhibition, thermostability, TMP-resistance and enzymatic activity profiles of all DfrB and DfrB homologues to those of DfrB1. To address this objective, new high-throughput activity assays as well as Minimal Inhibitory Concentration (MIC) assays were developed. Using those methods, we determined that thermostability and inhibition profiles of several DfrB and DfrB-H were comparable to those of DfrB1. Also, a screen of several dozen potential inhibitory compounds was performed, to attempt to identify further specific DfrB inhibitors. In addition, we report 10 new DfrB homologues that confer high TMP resistance and possess Dfr activity. The characterization of all DfrB members and DfrB homologues will allow us to acquire greater knowledge on their antimicrobial resistance mechanism, their prevalence in different environments and support our development of new DfrB-specific inhibitors.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

G3BP1 and G3BP2 are members of a highly conserved family of multi-functional RNA binding proteins, which appear to co-ordinate signal transduction and post-transcriptional gene regulation. Both proteins are over-expressed in cancer, and G3BP1 promotes cell proliferation and survival. Aberrant expression of various RNA binding proteins is common in cancer, and several of these proteins influence tumorigenesis. Therefore, detailed examination of RNA binding proteins, such as G3BPs, may provide insights into the post-transcriptional mechanisms underlying tumorigenesis. Tumours arise as a consequence of genetic mutation or alteration, which often result from stress-induced DNA damage. Cancer progression is facilitated by various epigenetic stress adaptation mechanisms. Stressful stimuli induce transitory translational shut-off, mediated by phosphorylation of eukaryotic initiation factor alpha;(eIF2alpha;). This phosphorylation event leads to formation of discrete cytoplasmic foci known as stress granules (SGs), which are translationally-silent sites of mRNA sorting. It was initially thought that an RNA-binding protein, T-cell internal antigen 1 (TIA-1), was instrumental in both the formation and functioning of SGs, because over-expression of TIA-1 induces spontaneous SGs and concomitantly causes a decrease in reporter gene expression. It is now clear that SG content can change depending on the type of stress, and that various proteins, including G3BP1, can induce spontaneous SGs. In vitro evidence previously implicated both G3BP1 and G3BP2 as endoribonucleases, so it was suggested that G3BPs act to target mRNA for decay at the SG. This project sought to further investigate this proposal, and in this way gain insight into the specific function of G3BPs in post-transcriptional regulation during tumorigenesis. Characterisation of G3BP1 and G3BP2 expression and localisation patterns in human cells and cancer was necessary before functional analyses in human cell systems could be undertaken. Both proteins were found to be over-expressed in breast cancer, irrespective of cancer stage or grade. G3BP1 and G3BP2 were also expressed in all human cell lines tested, despite previously observed tissue-specific expression. These results support the notion that G3BP expression is switched on in parallel with cell proliferation, and as such, may influence tumorigenesis. The results of further analyses suggested that the diverse functions attributed to G3BP1 and G3BP2 may be facilitated by isoform-specific expression, various post-translational modifications and sub-cellular localisation. Despite the absence of a canonical endoribonuclease domain, it was previously reported that site-specific phosphorylation of G3BP1 enables the protein to degrade a synthetic c-myc RNA substrate in vitro. This finding implicated G3BP in the specific regulation of a proto-oncogene. Tailored reporter assays were thus designed in order to address the in vivo consequences of G3BP's putative endoribonuclease activity. Contrary to expectations, all G3BP family members increased or maintained the expression of a range of reporters, at both the mRNA and protein level, irrespective of the presence of any particular cis-acting element, coding sequence or promoter. These results support the emerging notion that G3BPs positively affect the expression of at least some of their target mRNAs, and may also indirectly promote transcription. In contrast to the theory that G3BPs degrade proto-oncogenic mRNA/s, these findings are consistent with a role for G3BP in promoting cell proliferation and survival. Further analyses showed that G3BP1 and G3BP2 simultaneously increased reporter gene expression and induced SG formation. These findings highlighted the fact that SGs are dynamic sorting stations for mRNAs, and not merely sites of stalled translation. This result also supports the notion that a variety of proteins may be recruited to the SG to facilitate a multitude of mRNA fates. Although the precise role of the SG in stress adapation is not known, it is clear that an appropriate integrated stress response (ISR) is required for cells to survive in sub-optimal conditions. It was found that specific G3BP1 knockdown inhibited SG formation and cell survival, and this appeared to occur downstream of eIF2alpha; phosphorylation. The phosphorylation of eIFalpha; is the only factor known to be necessary for SG formation and cell survival. This data is the first to implicate SG formation itself, downstream of eIF2alpha; phosphorylation, in the survival phase of the ISR. The results also suggest that G3BP1 plays a pivotal role in the post-transcriptional mechanisms underlying stress adaptation. To facilitate future analysis of G3BP roles in the regulation of specific transcripts and in SG biology, a pilot study to identify G3BP RNA ligands was undertaken. Immunoprecipitation of epitope-tagged G3BP1 from stable cell lines facilitated purification and isolation of RNA in association with G3BP1. Specific RNA transcripts were subsequently detected and identified by microarray. Many genes were enriched in the G3BP1 immunoprecipitate. Transcript enrichment in the control immunoprecipitate was comparatively weak and seemingly random, suggesting that several replicates will enable generation of a reliable target list. This work forms a promising basis for further investigations into G3BP functionality, and also provides a platform for broader and more large-scale analyses of the mechanisms of post-transcriptional gene regulation. The work presented in this thesis addressed the potential post-transcriptional mechanisms by which the G3BP family of proteins mediate cell proliferation and survival. Both G3BP1 and G3BP2 were shown to be over-expressed in tumours and each appeared to promote reporter gene expression. G3BP1 was also found to play a pivotal role in stress adaptation. A technique to identify novel RNA ligands was assessed, and it was found that G3BP1 may interact with various mRNA transcripts. It is hypothesised that the G3BP family of proteins, and in particular G3BP1, function to determine the fate of specific RNAs in response to cellular stress and other stimuli. In this way, G3BP proteins may facilitate appropriate responses to extra-cellular stimuli which allow for cell proliferation and survival.

Protein-protein interactions are of vital importance to the cell as they mediate the assembly of protein complexes that carry out diverse biological functions. Many proteins involved in cellular signaling are built by the combinatorial use of peptide recognition modules (PRMs), which are small protein domains that bind to their cognate ligands by recognizing short linear peptide motifs. Thousands of PRMs are found in nature, requiring improved methods to better elucidate their molecular determinants of binding and to allow accurate mapping of their interaction networks. In this thesis, I describe the development and application of phage-displayed peptide libraries to map the binding specificities of two common PRMs. First, I generated specificity profiles for 82 C. elegans and human PDZ domains that could be organized into a specificity map. The map revealed that PDZ domains have far greater substrate sequence specificity than previously believed, providing significant insights into the relationships between PDZ structure and specificity, and allowing specificity prediction for uncharacterized domains. My results were used to predict both endogenous and pathogenic PDZ interactions. This analysis revealed that viruses have evolved ligands that specifically mimic PDZ domains to subvert host cell immunity. Second, I analyzed the binding specificity for the SH3 domain family in S. cerevisae. I found that, like PDZ domains, SH3 domains have binding specificities that are more detailed than the conventional classification system. The phage-derived specificity profiles were combined with data from oriented peptide and yeast two-hybrid screening to generate a highly accurate SH3 domain interaction network. Given the prominent role of SH3 domains in endocytosis, the SH3 domain interaction data was used to predict the dynamic localization of several uncharacterized endocytosis proteins, which was subsequently confirmed by cell-based assays. The application of the techniques described here to other PRM families will significantly improve protein interaction maps for signaling pathways, which will illuminate our understanding of the cell circuitry, allow the use of PRMs as general affinity reagent and detection tools, and guide the development of small molecule inhibitors that mimic their peptide ligands for therapeutic intervention.

Protein-protein interactions are of vital importance to the cell as they mediate the assembly of protein complexes that carry out diverse biological functions. Many proteins involved in cellular signaling are built by the combinatorial use of peptide recognition modules (PRMs), which are small protein domains that bind to their cognate ligands by recognizing short linear peptide motifs. Thousands of PRMs are found in nature, requiring improved methods to better elucidate their molecular determinants of binding and to allow accurate mapping of their interaction networks. In this thesis, I describe the development and application of phage-displayed peptide libraries to map the binding specificities of two common PRMs. First, I generated specificity profiles for 82 C. elegans and human PDZ domains that could be organized into a specificity map. The map revealed that PDZ domains have far greater substrate sequence specificity than previously believed, providing significant insights into the relationships between PDZ structure and specificity, and allowing specificity prediction for uncharacterized domains. My results were used to predict both endogenous and pathogenic PDZ interactions. This analysis revealed that viruses have evolved ligands that specifically mimic PDZ domains to subvert host cell immunity. Second, I analyzed the binding specificity for the SH3 domain family in S. cerevisae. I found that, like PDZ domains, SH3 domains have binding specificities that are more detailed than the conventional classification system. The phage-derived specificity profiles were combined with data from oriented peptide and yeast two-hybrid screening to generate a highly accurate SH3 domain interaction network. Given the prominent role of SH3 domains in endocytosis, the SH3 domain interaction data was used to predict the dynamic localization of several uncharacterized endocytosis proteins, which was subsequently confirmed by cell-based assays. The application of the techniques described here to other PRM families will significantly improve protein interaction maps for signaling pathways, which will illuminate our understanding of the cell circuitry, allow the use of PRMs as general affinity reagent and detection tools, and guide the development of small molecule inhibitors that mimic their peptide ligands for therapeutic intervention.