Les "protéine arginine méthyltransférases" (PRMT) sont impliquées dans de nombreux processus cellulaires : transcription, maturation et transport des ARN, traduction, transduction du signal, réplication et réparation de l'ADN, et apoptose. Différents travaux ont montré que des dérégulations de ces mécanismes impliquant les PRMT peuvent induire certains cancers, faisant de ces enzymes de nouvelles cibles potentielles en chimiothérapie. Il s’avère donc crucial de comprendre le mode d’action des PRMT à l’échelle atomique, à la fois au niveau fondamental et pour le développement de nouveaux médicaments. Les travaux décrits ici s’intéressent à la protéine PRMT4/CARM1 et s’appuient sur des études structurales par bio-cristallographie, pour comprendre les mécanismes de la réaction de méthylation catalysée par CARM1 et découvrir des inhibiteurs spécifiques, mais aussi sur des études en solution, pour caractériser l’interaction entre CARM1 et ses substrats. / Protein arginine methyltransferases (PRMTs) are involved in several cellular mechanisms: transcription, RNA maturation and transport, translation, signal transduction, DNA replication and repair, and apoptosis. Different studies showed that deregulation of those mechanisms involving PRMTs can induce some cancers, making these enzymes new potential targets for chemotherapy. It is therefore crucial to understand the mode of action of PRMTs at the atomic scale, both at the fundamental level and for the development of new drugs. The studies described here focus on PRMT4/CARM1 and rely on structural studies by bio-crystallography, in order to understand the methylation mechanisms catalyzed by CARM1 and to discover specific inhibitors, but also on in vitro studies, to characterize the interaction between CARM1 and its substrates.

Les protéines arginine méthyl transférases ("PRMTs") sont impliquées dans de nombreux processus cellulaires essentiels. La protéine CARMI ("Coactivator-associated arginine methyltransferase 1", appelée aussi "PRMT4") a été initialement identifiée par sa fonction co-activatrice de la transcription impliquantplusieurs récepteurs nucléaires des hormones. CARMI est une enzyme qui catalyse la réaction de méthylation sur les histones via un donneur de méthyl naturel, la S-adénosY-L -méthionine (SAM). De nombreux travaux ont montré que CARMI est surexprimée dans les cancers du sein et de la prostate. L' objectif de ce travail est la compréhension à l'échelle moléculaire du mode d'action de CARMI et l'étude du mécanisme de reconnaissances moléculaires et de transferts d' informations gouvernés par la protéine CARMI. La structure cristallographique obtenue de cette enzyme en présence de cofacteur, la S-AdénosyhHomocystéine ou la Sinefungine a eu un effet stabilisant. Ainsi, notre stratégie a été de créer des molécules hameçons basées sur le motif de la SAM capables d' ancrer un peptide mimant la séquence de l' histone H3, pour ensuite les tester en co-cristallisation avec CARMI. Ainsi, grâce à la diffraction aux rayons X, les interactions mises en jeu dans le complexe CARMlImolécules hameçons/peptide pourront être déterminées. Cette stratégie s'est effectuée en trois étapes : la première étape, décrite dans le chapitre 2, a consisté en la synthèse d'analogues de la SAM obtenus grâce à des modifications réalisées autour de l'atome de soufre. Ces composés nous ont permis d' explorer la « poche du sulfonium ». Puis la seconde étape, décrite dans le chapitre 3, a été la synthèse • d'analogues de bisubstrats nécessaires pour l'exploration de la « poche de l'arginine ». Dans une dernière étape, décrite dans les chapitres 4 et 5, nous avons abordé la synthèse d'adduits SAM-peptide pouf pouvoir étudier le « domaine de fixation du peptide ». Dans le quatrième chapitre, la méthode de choix est la création d'un lien covalent entre une molécule hameçon électrophile etun peptide par chimie de click in-situ : par réaction de cycloaddition de Huisgen; par réaction entre des molécules hameçons électrophiles capables de piéger un peptide cystéine ou un peptide arginine. Ces essais se sont révélés infructueux et une nouvelle stratégie a été employée en utilisant des molécules ancres. Dansle cinquième chapitre des molécules ancres ont donc été préformés pour ensuite être testés en cocristallisation dans CARMl. / Protein aginine methyltransferases (PRMTs) have been implicated in a variety of biological processes. Coactivator-associated arginine methyltransferase 1 (CARM1 , also known as PRMT4) was identified as an enhancer of the transcriptional activation by several nuclear hormone receptors CARM1 is an enzyme which methylates the arginines of histones via a natural methyl donor, the SAdenosyh-Methionine (SAM). Recent studies have shown that CARM-1 is over-expressed in breastumors and in hormone dependent prostate tumors. The goal of this work is to understand at the molecular level the mode of binding of substrate/product arginine-containing peptides, reflectingstates prior and subsequent to methylation and the detailed mechanism of action of this protein. Several crystal structures of the catalytic domain of CARM1 have shown that cofactor-binding, such as S-Adenosyl-L -Homocysteine or sinefungin, produces large conformational changes in the catalytic domain. These crystal structures clearly illustrate that SAM binding is a prerequisite for peptide binding and build up the productive peptide binding site. Our strategy was to design fishhook molecules derivatives of the SAM capable of anchoring a mimic peptide of the histone H3 in order to test the co-crystallization in CARM1. Consequently, thanks to X-Ray structure, interactions involvement in the complexe CARM1/fishhook molecule/peptide could be determined. This strategy was do ne in three steps: the first one, described in the chapter 2, consisted in synthesizing SAManalogues with several modifications around sulfur atom. These compounds permitted to explore the "sulfonium pocket". The second step, described in chapter 3, consisted in synthesizing analogues of bisubstrats to explore "arginine binding pocket". Finally, the last step, described in the chapter 4 and 5, consisted in synthesizing SAM-peptide adducts to study "peptide binding domain". ln the chapter 4, the chosen method is the creation of covalent link between this molecule and a peptide by click chemistry in-situ: by reaction of Huisgen's cycloaddition; by reaction between electrophilic fishhook molecules capable of capturing with a cystein or arginine peptides. Unfortunately, ail of these trials have been unsuccessful. Consequently SAM-peptide adducts were performed to be co-crystallized in CARM1. This part was described in the last chapter.

<div>Protein N-terminal methyltransferases (NTMTs) are a family of enzymes that methylate the α-N-terminus of a variety of protein substrates. Both NTMT1 and NTMT2 recognize a unique N-terminal X-P-K/R motif (X represents any amino acid other than D/E) to install 1-3 methyl group(s) on the substrates. NTMT1 plays important roles in mitosis regulation, chromatin interactions, and DNA damage repair. Another member NTMT2 shares ~50% sequence similarity and the same substrate recognition motif although NTMT2 was initially characterized as a mono-methyltransferase. To understand the molecular mechanism of NTMT2, we obtained the first co-crystal structure of NTMT2 in complex with its peptide substrate. After an extensive investigation of substrate recognition and methylated products of NTMT1/2, we found out that NTMT2 can fully methylate G/P-PKRIA peptides despite a predominant mono-methyltransferase. Moreover, we identified a gatekeeper N89 in NTMT2 that controls the substrate entry and the product specificity of NTMT2.</div><div>To elucidate the biological functions of NTMT1/2-catalyzed N-terminal methylation, we applied two different strategies to discover cell-potent inhibitors. Guided by the co-crystal structures of NTMT1 in complex with previously reported inhibitors, we designed and synthesized a series of new peptidomimetic inhibitors. By introducing more hydrophobic groups, the most cell-potent peptidomimetic inhibitor GD562 (IC50 = 0.93 ± 0.04 µM) exhibited over 2-fold increased inhibition on cellular N-terminal methylation levels with an IC50 value of ~50 µM compared to previously reported peptidomimetic inhibitor DC541. Meanwhile, we also discovered the first potent small molecule inhibitor Genz-682452 (IC50 = 0.5 ± 0.04 µM) after screening ~58,000 compounds. Subsequent structural modifications led to the discovery of GD433 (IC50 = 27 ± 0.5 nM) with a 20-fold increased potency compared to the initial hit Genz-682452. Inhibition mechanism indicated both inhibitors bind to peptide-binding pocket and co-crystal structures of both Genz-682452 and GD433 with NTMT1 confirmed their binding modes. Furthermore, GD433 shows over 7-fold selectivity over other major 40 protein methyltransferases and DNA methyltransferase and exhibits improved selectivity for NTMT1 over glucosylceramide synthase (GCS). GD433 significantly decreases the cellular N-terminal methylation level of NTMT1 substrates RCC1 and SET at 10 nM in both HEK293 and HCT116 cells, providing a valuable probe for cell-based studies in the future.<br></div><p><br></p>

Les "protéine arginine méthyltransférases" (PRMT) sont impliquées dans de nombreux processus cellulaires : transcription, maturation et transport des ARN, traduction, transduction du signal, réplication et réparation de l'ADN, et apoptose. Différents travaux ont montré que des dérégulations de ces mécanismes impliquant les PRMT peuvent induire certains cancers, faisant de ces enzymes de nouvelles cibles potentielles en chimiothérapie. Il s'avère donc crucial de comprendre le mode d'action des PRMT à l'échelle atomique, à la fois au niveau fondamental et pour le développement de nouveaux médicaments. Les travaux décrits ici s'intéressent à la protéine PRMT4/CARM1 et s'appuient sur des études structurales par bio-cristallographie, pour comprendre les mécanismes de la réaction de méthylation catalysée par CARM1 et découvrir des inhibiteurs spécifiques, mais aussi sur des études en solution, pour caractériser l'interaction entre CARM1 et ses substrats.

Because it can expand the range of a protein’s interactions or modulate its activity, post-translational methylation of arginine residues in proteins must be duly coordinated and ‘decoded’ to ensure appropriate cellular interpretation of this biological cue. This can be achieved through modulation of the enzymatic activity/specificity of the protein arginine methyltransferases (PRMTs) and proper recognition of the methylation ‘mark’ by a subset of proteins containing ‘methyl-sensing’ protein modules known as ‘Tudor’ domains. In order to gain a better understanding of these regulatory mechanisms, we undertook a detailed biochemical characterization of the predominant member of the PRMT family, PRMT1, and of the novel Tudor domain-containing protein 3 (TDRD3). First, we found that PRMT1 function can be modulated by 1) the expression of up to seven PRMT1 isoforms (v1-7), each with a unique N-terminal region that confers distinct substrate specificity, and by 2) differential subcellular localization, as revealed by the presence of a nuclear export sequence unique to PRMT1v2. Second, our findings suggest that TDRD3 is recruited to cytoplasmic stress granules (SGs) in response to environmental stress potentially by engaging in methyl-dependent protein-protein interactions with proteins involved in the control of gene expression. We also found that arginine methylation may serve as a general regulator of overall SG dynamics. Finally, we uncovered that alteration of PRMT1, TDRD3, and global arginine methylation levels in breast cancer cells may be closely associated with disease progression and poor prognosis. Therefore, further studies into the pathophysiological consequences ensuing from misregulation of arginine methylation will likely lead to the development of novel strategies for the prevention and treatment of breast cancer.

Because it can expand the range of a protein’s interactions or modulate its activity, post-translational methylation of arginine residues in proteins must be duly coordinated and ‘decoded’ to ensure appropriate cellular interpretation of this biological cue. This can be achieved through modulation of the enzymatic activity/specificity of the protein arginine methyltransferases (PRMTs) and proper recognition of the methylation ‘mark’ by a subset of proteins containing ‘methyl-sensing’ protein modules known as ‘Tudor’ domains. In order to gain a better understanding of these regulatory mechanisms, we undertook a detailed biochemical characterization of the predominant member of the PRMT family, PRMT1, and of the novel Tudor domain-containing protein 3 (TDRD3). First, we found that PRMT1 function can be modulated by 1) the expression of up to seven PRMT1 isoforms (v1-7), each with a unique N-terminal region that confers distinct substrate specificity, and by 2) differential subcellular localization, as revealed by the presence of a nuclear export sequence unique to PRMT1v2. Second, our findings suggest that TDRD3 is recruited to cytoplasmic stress granules (SGs) in response to environmental stress potentially by engaging in methyl-dependent protein-protein interactions with proteins involved in the control of gene expression. We also found that arginine methylation may serve as a general regulator of overall SG dynamics. Finally, we uncovered that alteration of PRMT1, TDRD3, and global arginine methylation levels in breast cancer cells may be closely associated with disease progression and poor prognosis. Therefore, further studies into the pathophysiological consequences ensuing from misregulation of arginine methylation will likely lead to the development of novel strategies for the prevention and treatment of breast cancer.

Les protéines arginine méthyl transférases ("PRMTs") sont impliquées dans de nombreux processus cellulaires essentiels. La protéine CARMI ("Coactivator-associated arginine methyltransferase 1", appelée aussi "PRMT4") a été initialement identifiée par sa fonction co-activatrice de la transcription impliquantplusieurs récepteurs nucléaires des hormones. CARMI est une enzyme qui catalyse la réaction de méthylation sur les histones via un donneur de méthyl naturel, la S-adénosY-L -méthionine (SAM). De nombreux travaux ont montré que CARMI est surexprimée dans les cancers du sein et de la prostate. L' objectif de ce travail est la compréhension à l'échelle moléculaire du mode d'action de CARMI et l'étude du mécanisme de reconnaissances moléculaires et de transferts d' informations gouvernés par la protéine CARMI. La structure cristallographique obtenue de cette enzyme en présence de cofacteur, la S-AdénosyhHomocystéine ou la Sinefungine a eu un effet stabilisant. Ainsi, notre stratégie a été de créer des molécules hameçons basées sur le motif de la SAM capables d' ancrer un peptide mimant la séquence de l' histone H3, pour ensuite les tester en co-cristallisation avec CARMI. Ainsi, grâce à la diffraction aux rayons X, les interactions mises en jeu dans le complexe CARMlImolécules hameçons/peptide pourront être déterminées. Cette stratégie s'est effectuée en trois étapes : la première étape, décrite dans le chapitre 2, a consisté en la synthèse d'analogues de la SAM obtenus grâce à des modifications réalisées autour de l'atome de soufre. Ces composés nous ont permis d' explorer la " poche du sulfonium ". Puis la seconde étape, décrite dans le chapitre 3, a été la synthèse * d'analogues de bisubstrats nécessaires pour l'exploration de la " poche de l'arginine ". Dans une dernière étape, décrite dans les chapitres 4 et 5, nous avons abordé la synthèse d'adduits SAM-peptide pouf pouvoir étudier le " domaine de fixation du peptide ". Dans le quatrième chapitre, la méthode de choix est la création d'un lien covalent entre une molécule hameçon électrophile etun peptide par chimie de click in-situ : par réaction de cycloaddition de Huisgen; par réaction entre des molécules hameçons électrophiles capables de piéger un peptide cystéine ou un peptide arginine. Ces essais se sont révélés infructueux et une nouvelle stratégie a été employée en utilisant des molécules ancres. Dansle cinquième chapitre des molécules ancres ont donc été préformés pour ensuite être testés en cocristallisation dans CARMl.

Because it can expand the range of a protein’s interactions or modulate its activity, post-translational methylation of arginine residues in proteins must be duly coordinated and ‘decoded’ to ensure appropriate cellular interpretation of this biological cue. This can be achieved through modulation of the enzymatic activity/specificity of the protein arginine methyltransferases (PRMTs) and proper recognition of the methylation ‘mark’ by a subset of proteins containing ‘methyl-sensing’ protein modules known as ‘Tudor’ domains. In order to gain a better understanding of these regulatory mechanisms, we undertook a detailed biochemical characterization of the predominant member of the PRMT family, PRMT1, and of the novel Tudor domain-containing protein 3 (TDRD3). First, we found that PRMT1 function can be modulated by 1) the expression of up to seven PRMT1 isoforms (v1-7), each with a unique N-terminal region that confers distinct substrate specificity, and by 2) differential subcellular localization, as revealed by the presence of a nuclear export sequence unique to PRMT1v2. Second, our findings suggest that TDRD3 is recruited to cytoplasmic stress granules (SGs) in response to environmental stress potentially by engaging in methyl-dependent protein-protein interactions with proteins involved in the control of gene expression. We also found that arginine methylation may serve as a general regulator of overall SG dynamics. Finally, we uncovered that alteration of PRMT1, TDRD3, and global arginine methylation levels in breast cancer cells may be closely associated with disease progression and poor prognosis. Therefore, further studies into the pathophysiological consequences ensuing from misregulation of arginine methylation will likely lead to the development of novel strategies for the prevention and treatment of breast cancer.

Because it can expand the range of a protein’s interactions or modulate its activity, post-translational methylation of arginine residues in proteins must be duly coordinated and ‘decoded’ to ensure appropriate cellular interpretation of this biological cue. This can be achieved through modulation of the enzymatic activity/specificity of the protein arginine methyltransferases (PRMTs) and proper recognition of the methylation ‘mark’ by a subset of proteins containing ‘methyl-sensing’ protein modules known as ‘Tudor’ domains. In order to gain a better understanding of these regulatory mechanisms, we undertook a detailed biochemical characterization of the predominant member of the PRMT family, PRMT1, and of the novel Tudor domain-containing protein 3 (TDRD3). First, we found that PRMT1 function can be modulated by 1) the expression of up to seven PRMT1 isoforms (v1-7), each with a unique N-terminal region that confers distinct substrate specificity, and by 2) differential subcellular localization, as revealed by the presence of a nuclear export sequence unique to PRMT1v2. Second, our findings suggest that TDRD3 is recruited to cytoplasmic stress granules (SGs) in response to environmental stress potentially by engaging in methyl-dependent protein-protein interactions with proteins involved in the control of gene expression. We also found that arginine methylation may serve as a general regulator of overall SG dynamics. Finally, we uncovered that alteration of PRMT1, TDRD3, and global arginine methylation levels in breast cancer cells may be closely associated with disease progression and poor prognosis. Therefore, further studies into the pathophysiological consequences ensuing from misregulation of arginine methylation will likely lead to the development of novel strategies for the prevention and treatment of breast cancer.

DNA-Methylierung spielt eine wichtige Rolle bei der Kontrolle der Chromatinorganisation und Genregulation in höheren Eukaryoten und muss zusammen mit der genetischen Information in jedem Zellzyklus dupliziert werden. Bei Mammalia wird DNA durch die DNA-Methyltransferase 1 (DNMT1) methyliert, die dabei mit nuklearen Replikationsstellen (RF) assoziiert und so die Erhaltung des Methylierungsmusters mit der Duplikation der DNA verbindet. In dieser Arbeit wurden die Funktion der regulatorischen Sequenzen in der N-terminalen Domäne von DNMT1 bei der Kontrolle ihrer subnuklearen Lokalisierung während des Zellzyklus und die evolutionäre Konservierung dieser Sequenzen, sowie die Mechanismen die eine Assoziation von Proteinen mit RF vermitteln, untersucht. Es konnte gezeigt werden, dass DNMT1 eine dynamische Verteilung im Kern aufweist, die durch regulatorische Sequenzen zellzyklusabhängig gesteuert wird. Um die subnukleare Verteilung von DNMT1 während des Zellzyklus zu untersuchen, wurden RFP-Ligase Fusionsproteine hergestellt, die als Marker für die Identifikation von Zellzyklusstadien in lebenden Zellen dienen. Verschiedene, mit GFP fusionierte DNMT1 Mutanten wurden zusammen mit RFP-Ligase exprimiert und über einen ganzen Zellzyklus hinweg mit 4-dimensionaler Lebendzellmikroskopie verfolgt. Die PBD (PCNA-Bindungsdomäne) bewirkt die Lokalisierung von DNMT1 an RF während der S-Phase, und die TS (targeting sequence) vermittelt die Retention von DNMT1 an spät replizierendem Heterochromatin von der späten S- bis zur frühen G1-Phase. Im Gegensatz dazu scheint die PBHD (Polybromohomologiedomäne) für die Freisetzung von DNMT1 von perizentrischen Regionen während der G1-Phase notwendig zu sein. Eine Überexpression der TS zu Störung dieser Assoziation, senkt die Überlebensrate der Zellen und fördert die Bildung von Mikronuklei sowie die Verschmelzung von zentromerem Heterochromatin. Diese Ergebnisse zeigen eine neue Funktion für die TS bei der Assoziation von DNMT1 mit perizentrischem Heterochromatin von der später S- über die G2-Phase bis hin zur Mitose, die eine wichtige Voraussetzung für die Erhaltung der DNA-Methylierung und Heterochromatinstruktur und -funktion ist. Datenbankanalysen zeigten, dass es sich bei der TS um eine einzigartige Domäne innerhalb der DNMT1 Proteinfamilie handelt. Innerhalb der DNMT1 Familie besitzen nur die DNMT1 Proteine der Metazoen die PBD. Das lässt vermuten, dass die Verknüpfung von Beibehaltung der DNA Methylierung mit der DNA Replikation nur in Metazoen auftritt, während in Pflanzen und Pilzen alternative Mechanismen zur Aufrechterhaltung des Methylierungsmusters, wahrscheinlich vermittelt durch die TS, zur Anwendung kommen. Die evolutionäre Konservierung von Mechanismen, zur Assoziation von Proteine mit RF in Säugerzellen, wurde durch die Analyse der Säugerproteine PCNA, DNA Ligase I und DNMT1 in Drosophila-zellen direkt getestet. Von allen untersuchten Proteinen assoziiert nur PCNA mit RF, während die anderen nur eine diffuse Verteilung innerhalb des Kerns zeigten, obwohl sie eine funktionale PBD enthalten. Überraschenderweise assoziierte auch die Drosophila DNA Ligase I in Säugerzellen nicht aber in Drosophila-zellen mit RF. Diese Ergebnisse weisen auf Unterschiede in der Dynamik und dem Aufbau der Replikationsmaschinerie in diesen entfernt verwandten Organismen hin, was mit der Vergrösserung und höheren Komplexität des Säugergenoms korreliert. / DNA methylation constitutes an essential epigenetic mark controlling chromatin organization and gene regulation in higher eucaryotes, which has to be duplicated together with the genetic information at every cell division cycle. In mammals duplication of DNA methylation is mediated by DNA methyltransferase-1 (DNMT1). It associates with sites of nuclear DNA replication, called replication foci (RF), and thereby couples maintenance of DNA methylation to DNA duplication. In this work, we have analyzed the role of regulatory sequences in the N-terminal domain of DNMT1 in controlling its subnuclear localization throughout the cell cycle, and the evolutionary conservation of these sequences and of the mechanisms that mediate association of proteins with RF. We provide evidence that DNMT1 shows dynamic subnuclear distribution that is controlled by the regulatory sequences depending on the cell cycle stage. To determine the subnuclear distribution of DNMT1 throughout the cell cycle, an RFP-Ligase fusion protein was developed as a marker that allows identification of the cell cycle stage in live cells. Various DNMT1 mutants fused to GFP were coexpressed with RFP-Ligase and imaged by 4-dimensional live cell microscopy during an entire cell cycle. The PBD (PCNA binding domain) drives the localization of DNMT1 at RF throughout S phase and the TS (targeting sequence) mediates retention of DNMT1 only at the late replicating pericentric heterochromatin from late-S phase until early-G1. In contrast, the PBHD (polybromo homology domain) seems to be required for unloading DNMT1 from the pericentric regions in G1. Overexpression of the TS to interfere with this association lowers cell viability and induces the formation of micronuclei and coalescence of centromeric heterochromatin. These results bring forth a novel function of the TS in mediating association of DNMT1 with pericentric heterochromatin from late-S phase through G2 until mitosis, which is important for maintenance of DNA methylation, and heterochromatin structure and function. Database searches indicate that the TS is a domain unique to the DNMT1 family of proteins. Amongst the DNMT1 family, only the metazoan DNMT1 proteins have the PBD. This suggests that coupling of maintenance of DNA methylation with DNA replication occurs only in metazoans, while plants and fungi have alternative mechanisms that maintain DNA methylation patterns, probably mediated by the TS. The evolutionary conservation of the mechanisms by which proteins associate with RF in mammalian cells was directly tested by analyzing the ability of mammalian replication proteins PCNA and DNA Ligase I as well as DNMT1 to associate with RF in Drosophila cells. Of all the proteins tested, only PCNA associated with RF while the others showed diffused nuclear distribution although they contain a functional PBD. Surprisingly, Drosophila DNA Ligase I associates with RF in mammalian but not in Drosophila cells. These results suggest differences in the dynamics and organization of the replication machinery in these distantly related organisms, which correlates with the increased size and complexity of mammalian genomes.