Une réponse rapide et adaptée est nécessaire à la survie des cellules soumises à un stress. La réponse cellulaire au stress (HSR pour Heat-Shock response) médié par le facteur de transcription HSF1 est induite par les contextes environnementaux (chaleur, hypoxie, …) et les processus biologiques normaux et pathologiques (vieillissement, inflammation, …) associés à une accumulation de protéines endommagées (Morimoto, 1998). Ces protéines endommagées forment des agrégats toxiques aux conséquences létales pour les cellules.Conservé chez tous les eucaryotes, HSF1 orchestre les actions nécessaires à la survie et à la croissance des cellules malgré le stress. Ses cibles les mieux connues sont les gènes codants pour les Heat Shock Protein (HSP) qui font office de chaperon moléculaire. Une caractéristique de la HSR chez l’Homme est l’accumulation massive du facteur HSF1 en foyers nucléaires nommés Nuclear Stress Bodies (nSBs). Curieusement, ces foyers ciblent l’hétérochromatine péricentrique composée de séquences répétées en tandem de type Satellite III (SATIII), particulièrement au niveau du locus 9q12. HSF1 induit une forte transcription en ARN SATIII Sens (Jolly et al., 2004). Le rôle des nSBs est une des problématiques majeures de notre équipe cependant jusqu’à présent aucune fonction n’a été confirmée pour ces structures.Les nSBs, spécifiques aux cellules humaines, n’ont été décrits que dans des cellules en culture. Mon projet de thèse a consisté dans un premier temps à montrer la présence des nSBs in vivo chez l’Homme. Cette étude, réalisée sur du tissu testiculaire nous a également permis d’identifier une nouvelle cible SATIII majeure pour HSF1, la région Yq12. Dans les testicules, les nSBs sont associés à des processus méiotiques et post-méiotiques, suggérant un rôle dans le remodelage de l’hétérochromatine. Dans un deuxième temps, nous avons cherché à mieux comprendre le rôle des nSBs lors de la HSR. Nous avons pu montrer que l’étape de transcription des SATIII induit une déstabilisation de l’hétérochromatine péricentrique caractérisée par une dissociation des facteurs HP1 (Heterochromatin Protein 1) alpha et beta et une perte de la marque répressive H3K9me3. Au cours de la période de récupération qui accompagne la reformation de l’hétérochromatine, une transcription séquentielle d’ARN SATIII Sens puis Anti-sens précède la restructuration des loci 9q12. Nous avons également pu montrer que la transcription des SATIII est associée à un blocage de la mitose. Nous montrons que dans les cellules stressées, une altération de ce point de contrôle par un Knock down des ARN sat III par des approches LNA conduisent à une l’instabilité génomique des cellules tumorales avec apparition de cellules polynucléées. / A rapid and well-adapted response is required for cell survival upon stress. The cellular stress response (HSR) is mediated by the transcription factor Heat Shock Factor 1 (HSF1) (Morimoto, 1998). It is activated by environmental stress (heat, hypoxia, ...) and by a series of patho-physiological contexts (aging, inflammation, ...) involving protein damages.The best-characterized targets of HSF1 are genes encoding for Heat Shock Protein (HSP) acting as molecular chaperone. A specific feature of the HSR in human cells is the presence of HSF1 nuclear foci named Nuclear Stress Bodies (NSBs). Surprisingly, nSBs target pericentric heterochromatin consisting in tandem repeats of type III Satellite (SATIII) sequences, primarily at the 9q12 locus. HSF1 triggers a strong transcriptional activation of this locus (Jolly et al., 2004). The role of nSBS is a major issue since no function related to these structures has been reported so far.So far, nSBs have been only identified in cells in culture. My thesis project has been to further explore whether these structures also existed in normal tissues. Indeed, we have been able to identify the presence of nSBs in testis where they were found to be associated to meiotic and post-meiotic stages, suggesting a role related to heterochromatin remodeling. Moreover, we have identified the Yq12 locus as a new target of nSBs in these tissues. Secondly, we have brought new evidence that sat III sequences triggers a transient dissociation of HP1 (heterochromatin Protein 1) α and β as well as a loss of the repressive epigenetic H3K9me3 histone mark at pericentric heterochromatin. Interestingly we have also found that, following stress, a sequential accumulation of SATIII RNA in a Sense and Antisense orientation occurs, suggesting that this specific pattern of expression plays an important role in heterochromatin reformation. Finally, we have found that the accumulation of SATIII RNA is associated with a slowdown of mitosis. Indeed we have found that in stressed cells, accumulation of sat III impcats the progression of mitosis and that a knock down of sat III RNA using LNA approaches releases this blockade, leading to genomic instability of tumor cells and to the appearance of poly nucleated cells.

Eukaryotic cells exhibit an extensive transcriptional diversity. Only about a quarter of the total RNA in the human cell can be accounted for by messenger RNA (mRNA), which convey genetic code for protein generation. The remaining part of the transcriptome consists of rather heterogenous molecules. While some classes are well defined and have been shown to carry out distinct functions, ranging from housekeeping to complex regulatory tasks, a big fraction of the transcriptional output is categorized solely based on the lack of protein-coding capacity and transcript length. Several studies have shown, that as a group, mRNA-like long non-coding RNAs (lncRNAs), are under stabilizing selection, however at much weaker levels than mRNAs. The conservation at the level of primary sequence is even lower, blurring the contrast between exonic and intronics parts, which impedes traditional methods of genome-wide homology search. As a consequence their evolutionary history is a fairly unexplored field and apart from a few experimentally studied cases, the vast majority of them is reported to be poorly conserved. However, the pervasive transcription and the highly spatio-temporal specific expression patterns of lncRNAs suggests their functional importance and makes their evolutionary age and conservation patterns a topic of interest. By employing diverse computational methods, recent studies shed light on the common conservation of lncRNA’s secondary and gene structures, highlighting the significance of structural features on functionality. Splice sites, in particular, are frequently retained over very large evolutionary time scales, as they maintain the intron-exon-structure of the transcript. Consequently, the conservation of splice sites can be utilized in a comparative genomics approach to establish homology and predict evolutionarily well-conserved transcripts, regardless of their coding capacity. Since splice site conservation cannot be directly inferred from experimental evidence, in the course of this thesis a computational pipeline was established to generate comparative maps of splice sites based on multiple sequence alignments together with transcriptomics data. Scoring schemes for splice site motifs are employed to assess the conservation of orthologs. This resource can then be used to systemically study the conservation patterns of RNAs and their gene structures. This thesis will demonstrate the versatility of this method by showcasing biological applications of three distinct studies. First, a comprehensive annotation of the human transcriptome, from RefSeq, ESTs and GENCODE, was used to trace the evolution of human lncRNAs. A large majority of human lncRNAs is found to be conserved across Eutheria, and many hundreds originated before the divergence of marsupials and placental mammals. However, they exhibit a rapid turnover of their transcript structures, indicating that they are actual ancient components of the vertebrate genome with outstanding evolutionary plasticity. Additionally, a public web server was setup, which allows the user to retrieve sets of orthologous splice sites from pre-computed comparative splice site maps and inspect visualizations of their conservation in the respective species. Second, a more specific data set of non-colinearly spliced latimerian RNAs is studied to fathom the origins of atypical transcripts. RNA-seq data from two coelacanth species are analyzed, yielding thousands of circular and trans-spliced products, with a surprising exclusivity of the majority of their splice junctions to atypically spliced forms, that is they are not used in linear isoforms. The conservation analysis with comparative splice site maps yielded high conservation levels for both cir- cularizing and trans-connecting splice sites. This fact in combination with their abundance strongly suggests that atypical RNAs are evolutionarily old and of functional importance. Lastly, comparative splice site maps are used to investigate the role of lncRNAs in the evolution of the Alzheimer’s disease (AD). The human specificity of AD clearly points out a phylogenetic aspect of the disease, which makes the evolutionary analysis a very promising field of research. Protein- coding and non-protein-coding regions, that have been identified to be differentially expressed in AD patients, are analyzed for conservation of their splice site and evolution of their exon-intron-structure. Both non-coding and protein-coding AD-associated genes are shown to have evolved more rapidly in their gene structure than the genome at large. This supports the view of AD as a consequence of the recent rapid adaptive evolution of the human brain. This phylogenetic trait might have far reaching consequences with respect to the appropriateness of animal models and the development of disease-modifying strategies. / Eukaryotische Zellen legen eine umfangreiche transkriptionelle Vielfalt an den Tag. Nur etwa ein Viertel der in der menschlichen Zelle enthaltenen RNA ist messenger RNA (mRNA), welche den genetischen Code für die Proteingenerierung übermittelt. Der verbleibende Anteil des Transkriptoms besteht aus eher heterogenen Molekülen. Während einigen wohldefinierten Klassen spezifische Funktionen zugeordnet werden können, welche von Zellhaushalt bis zu komplexen regulatorischen Aufgaben reichen, wird ein großer Teil der transkriptionellen Produktion ausschließlich auf Grundlage der fehlenden Kodierungskapazität und der Transkriptlänge kategorisiert. Einige Studien zeigten, dass mRNA-ähnliche lange nicht-kodierende RNA (lncRNA) als Gruppe unter stabilisierender Selektion stehen, wenn auch in einem weitaus geringeren Ausmaß als mRNAs. Die Konservierung auf Ebene der primären Sequenz ist sogar noch niedriger, wodurch der Kontrast zwischen exonischen und intronischen Elementen verschwimmt und Methoden der traditionellen Homologiesuche erschwert werden. Infolgedessen ist die evolutionäre Geschichte der lncRNAs ein recht unerforschtes Gebiet und abgesehen von ein paar vereinzelten Fallstudien wird die große Mehrheit als schwach konserviert vermeldet. Die tiefgreifende Transkription und die in Raum und Zeit hochspezifischen Expressionsmuster von lncRNA deuten jedoch auf deren funktionelle Bedeutung hin und machen ihr evolutionäres Alter und ihre Konservierungsmuster zu einem Thema von Interesse. Durch die Verwendung von computergestützten Methoden konnten jüngste Studien die verbreitete Konservierung von Sekundär- und Genstruktur von lncRNAs aufzeigen, was die Signifikanz von strukturellen Merkmalen in Bezug auf deren Funktionalität unterstreicht. Spleißstellen im besonderen werden oft über lange evolutionäre Zeitspannen erhalten, da sie die Intron-Exon-Struktur des Transkripts bewahren. Folglich, kann die Konservierung von Spleißstellen durch einen Ansatz der vergleichenden Genomik benutzt werden, um Homologie herzuleiten und evolutionär gut konservierte Transkripte unabhängig von deren Kodierungskapazität zu prognostizieren. Da es nicht möglich ist die Spleißstellenkonservierung direkt anhand von experimentellen Indikatoren abzulesen, wurde im Zuge dieser These eine computergestützte Methode entwickelt, welche, basierend auf multiplen Sequenzalignments und Transkriptomikdaten, “Vergleichskarten” von Spleißstellen erstellt. Ein Punktebewertungssystem für Spleißstellenmotive wird benutzt um die Konservierung der Orthologen zu beurteilen. Diese Resource kann anschließend verwendet werden um systematisch die Konservierungsmuster von RNAs und deren Genstrukturen zu untersuchen. Diese Arbeit wird die Vielseitigkeit dieser Methode demonstrieren, indem die biologische Anwendung in drei verschiedenen Studien präsentiert wird. Zuerst wird eine umfassende Annotation des menschlichen Transkriptoms, basierend auf RefSeq, EST und GENCODE, benutzt, um die Evolution von humanen lncRNAs nachzuvollziehen. Es konnte festgestellt werden, dass eine große Mehrheit der menschlichen lncRNAs innerhalb der Eutheria konserviert ist und mehrere hundert bereits vor der Auseinanderentwicklung von Beuteltieren und höheren Säugetieren entstanden. Dennoch zeigen sie eine rasante Veränderung in ihren Transkriptstrukturen, welche darauf hindeutet, dass sie tatsächlich alte Bestandteile von Vertebratengenomen mit bemerkenswerter evolutionärer Formbarkeit sind. Zusätzlich wurde ein öffentlicher Webserver aufgesetzt, der dem Nutzer ermöglicht Datensätze orthologer Spleißstellen aus vorgenerierten Vergleichskarten zu extrahieren und Visualisierungen der Konservierung in den jeweiligen Spezies zu betrachten. Als zweites wird ein spezifischerer Datensatz von nicht-linear gespleißten Latimeria-RNA untersucht um die Ursprünge untypischer Transkripte zu ergründen. Die Analyse der RNA-seq Daten zweier Exemplare des Quastenflossers ergab tausende zirkulärer und Transspleiß-Produkte, wobei die Mehrheit der Spleißverbindungen eine überraschende Exklusivität für untypisch gespleißte Formen aufzeigt, d.h. diese werden nicht für lineare Isoformen genutzt. Die Konservierungsanalyse mit Spleißstellen-Vergleichskarten ergibt hohe Konservierungsniveaus sowohl für zirkulärisierende als auch für trans-verbindende Spleißstellen. Diese Tatsache in Kombination mit ihrem häufigen Vorkommen, deutet stark darauf hin, dass untypische RNAs evolutionär alt und von funktioneller Bedeutung sind. Zuletzt werden Spleißstellen-Vergleichskarten benutzt um die Rolle von lncRNAs in der Evolution der Alzheimer-Krankheit (AK) zu untersuchen. Die Spezifität der AK auf den Menschen weist klar auf einen phylogenetischen Aspekt der Krankheit hin, was deren evolutionäre Analyse zu einem vielversprechenden Forschungsgebiet macht. Proteinkodierende und nicht-proteinkodierende Regionen, bei denen eine differentielle Expression in AK-Patienten erkannt wurde, werden auf die Konservierung ihrer Spleißstellen und Evolution ihrer Exon-Intron-Strukturen hin analysiert. Es kann nachgewiesen werden, dass sich die Genstruktur von sowohl nicht-kodierenden als auch von proteinkodierenden AK-assoziierten Genen schneller entwickelt als das Genom im Allgemeinen. Das unterstützt die Auffassung, dass AK die Folge einer kürzlichen rasanten adaptiven Evolution des menschlichen Gehirns ist. Diese phylogenetische Eigenschaft könnte weitreichende Konsequenzen in Bezug auf die Angemessenheit von Tiermodellen und die Entwicklung von krankheitsmodifizierenden Strategien haben.

Myc deregulation critically contributes to many cancer etiologies. Recent work suggests that Myc and its direct interactors can confer a distinct epigenetic state. Our goal is to better understand the Myc-conferred epigenetic status of cells. We have previously identified the long non-coding RNA (lncRNA), H19, as a target of Myc regulation and shown it to be important for transformation in lung and breast cells. These results prompted further analysis to identify similarly important Myc-regulated lncRNAs. Myc-regulated lncRNAs associated with the cell cycle and transformation have been identified by microarray analysis. A small number of candidate lncRNAs that were differentially expressed in both the cell cycle and transformation have been validated. Given the increasing importance of lncRNAs and epigenetics to cancer biology, the discovery of Myc-induced, growth associated lncRNAs could provide insight into the mechanisms behind Myc-related epigenetic signatures in both normal and disease states.

No description available.

Sex biased expression characterizes ~1,000 genes in mammalian liver, and impart sex differences in metabolism and disease susceptibility. The sex-dependent temporal patterns of pituitary growth hormone (GH) secretion, pulsatile in males and more continuous in females, are known to sex-differentially activate transcriptional regulators (TFs), leading to widespread sex-differences in the mouse liver transcriptome. This thesis elucidates sex-biased gene expression patterns in the following studies. Gene structures, expression patterns and species conservation are characterized for ~15,000 liver-expressed intergenic long noncoding RNAs (lncRNAs), many of which are novel. Analysis of intergenic lncRNA promoters revealed unexpected high conservation and significant enrichment of TF binding compared to protein-coding promoters. A subset of intergenic lncRNAs showed strong sex-specific and GH-dependent gene expression, and whose transcription was tightly correlated with the surrounding chromatin environment and TF binding patterns. The pervasive role of genetic factors to regulate sex-biased genes was revealed by analyzing livers with matched genotype and gene expression data from Diversity Outbred (DO) mice, an outbred population with high natural allelic variance derived from eight inbred strains. Significant associations between genetic variants and gene expression (eQTLs) were identified, including many eQTLs with a strong sex-biased association. Remarkably, a large fraction of these sex-biased eQTLs were linked to either gain or loss of sex-specific gene expression in the DO founder strain predicted to be regulated by the eQTL. Thus, genetic factors are a major contributor to the variability of sex-biased genes, which has important consequences related to the individual variability of liver phenotypes with known sex-differences. Natural genetic perturbations in DO mice were leveraged to identify candidate lncRNAs that may regulate hypophysectomy (hypox) responsiveness. Co-regulated protein-coding gene clusters were discovered based on gene expression correlations across DO mouse livers, many of which are enriched for distinct hypox response classes. LncRNAs whose expression showed unexpected significant negative correlation with protein-coding gene clusters enriched for genes of the opposite-sex bias and inverse hypox class were hypothesized to play negative regulatory role. In sum, these studies expand the characterization of the sex-biased hepatic transcriptome and reveal contributions of genetic factors to the regulation of sex bias in mammalian liver. / 2020-11-27T00:00:00Z

Myc deregulation critically contributes to many cancer etiologies. Recent work suggests that Myc and its direct interactors can confer a distinct epigenetic state. Our goal is to better understand the Myc-conferred epigenetic status of cells. We have previously identified the long non-coding RNA (lncRNA), H19, as a target of Myc regulation and shown it to be important for transformation in lung and breast cells. These results prompted further analysis to identify similarly important Myc-regulated lncRNAs. Myc-regulated lncRNAs associated with the cell cycle and transformation have been identified by microarray analysis. A small number of candidate lncRNAs that were differentially expressed in both the cell cycle and transformation have been validated. Given the increasing importance of lncRNAs and epigenetics to cancer biology, the discovery of Myc-induced, growth associated lncRNAs could provide insight into the mechanisms behind Myc-related epigenetic signatures in both normal and disease states.

This project concerns an exploration of the presence and interactions of long non-coding RNA transcripts in an experimental atherosclerosis mouse model with relevance for human abdominal aortic aneurysm development. 187 long noncoding RNAs, two of them entirely novel, were found to be differentially expressed between angiotensin II treated (developing abdominal aortic aneurysms) and non-treated apolipoprotein E deficient mice (not developing aneurysms) harvested after the same period of time. These transcripts were also studied with regards to co-expression network connections. Eleven previously annotated and two novel long non-coding RNAs were present in two significantly disease correlated co-expression groups that were further profiled with respect to network properties, Gene Ontology terms and MetaCore© connections.

Les ARNm codent pour les protéines, tandis qu'un grand nombre d'ARNs nommés longues ARNs non codants (ARNlnc) ne sont pas traduites en protéines. Les deux types d’ARNs existent en isoforms qui se distinguent à cause de l’épissage alternatif. Certains des ARNlnc jouent des rôles importants dans la croissance et différentiation cellulaire. Cependant, leurs fonctions dans la cytotoxicité de la chimiothérapie anti-cancéreuse médicamenteuse utilisant le 5-fluorouracile (5-FU) sont encore inconnues. Pendant mes travaux j'ai trouvé que le traitement par le 5-FU cause l’accumulation des ARNlnc. Ce phénomène est parfois, sous forme d’ARN double brin (ARNds) formé par une paire de transcrits chevauchant, corrélé négativement avec le niveau de la protéine codée par l'ARNm. Cette inhibition potentielle de la traduction des régulateurs du cycle cellulaire clés et les gènes essentiels en formant des l'ARNds peut éventuellement empêcher la progression du cycle cellulaire. Nos analyses prometteuses devraient inspirer des études approfondies des ARNlnc dans la cytotoxicité du 5-FU chez la levure et l’homme afin d’'améliorer la chimiothérapie. J'ai trouvé que la surexpression de RRP6, peut conduire à une résistance accrue au traitement par le 5-FU. Je démontre ensuite que l’ARNlnc MUT1312 forme des ARNds avec RRP6 qui sont négativement corrélés avec le niveau de la protéine Rrp6. Par ailleurs, la surexpression de MUT1312 pendant la mitose et associé avec une diminution d’Rrp6. Ainsi, mon étude suggère que MUT1312 soit impliqué dans la régulation de Rrp6 pendant la differentiation cellulaire. Mes recherches de MUT477/SWI4 indiquent la function importante de la méiose induite à long ARN non codantes en tant que forme d'ARN double brin potentiellement réguler la traduction. J'ai trouvé que SUT200 pourrait inhiber la transcription de CDC6 durant la méiose par read-through. Un cas comparable est MUT1465 et CLN2. J’ai fait un criblage in silico pour trouver des facteurs de transcription qui activent des MUTs durant la méiose. J’ai trouvé que la plupart des MUTs sont induites par Ndt80. MUT1465 est parmi eux : il pourrait être induite par Ndt80 ce qui inhiberait l’expression de CLN2 après l’initiation de la méiose. J’ai trouvé que la répression de certains MUTs par le complexe Ume6/Rpd3 en mitose est différemment régulée entre JHY222 et SK1. MUT100 qui ne possède pas l'élément USR1 fixé par Ume6, et qui est donc une cible indirecte, est déréprimé dans JHY22 ume6 mais pas dans SK1 ume6. Pour la régulation de l'étude de isoforme méiose, Nous avons trouvé que le complexe histone déacétylase Rpd3/Sin3/Ume6, empêche également l'induction de l'isoforme longue de BOI1 dans la mitose par liaison directe de liaison Ume6 à sa cible de URS1. Orc1 est importante pour la réplication de l'ADN. J’ai démontré que mORC1 est une cible directe de l'activateur Ndt80 et que son motif de fixation (MSE) est nécessaire pour l'induction de l’isoforme mORC1 et du gene méiotique SMA2 transcrit de façon divergente. J’ai trouvé qu'une souche incapable d’induire mORC1, contient des niveaux anormalement élevés d’Orc1 pendant la gamétogenèse, ce qui corréle mORC1 avec la baisse de la protéine Orc1. En conclusion, mes études au cours du doctorat révèlent des nouvelles cibles et ainsi offrent des nouvelles perspectives de l’amélioration de la chimiothérapie par le 5-FU. Les mécanismes incluent la formation d'un ARN double brin avec son ARNm anti-sens pour potentiellement inhiber la traduction de l'ARNm, et inhibition en aval de l'ARNm par transcription read-through d’une ARNlnc. Mon travail a également révélé un mécanisme de régulation des ARNlnc et les isoforms d’ARN pendent la croissance et la différentiation cellulaire. / RNAs are molecules with important functions in diverse cellular processes. mRNAs encode proteins, while a large number of RNAs called long noncoding RNAs (lncRNAs) are not translated into proteins. Both types of RNAs exist in various isoforms due to alternative splicing.Some of lncRNA play important roles in cell growth and differentiation. However, their functions in the cytotoxicity of the drug anticancer chemotherapy using 5-fluorouracil (5-FU) are still unknown. During my research I found that treatment with 5-FU causes accumulation of lncRNA. Acuumulated antisense lncRNA form double stranded RNA with the mRNAs , negatively correlated with the level of the protein encoded by the mRNA. This potential inhibition of translation of key cell cycle regulators and essential genes by forming dsRNA may possibly prevent the progression of the cell cycle. My results suggest that lncRNA are likely to play an important role in the cytotoxicity of 5-FU. Our promising testing should inspire in-depth studies of lncRNA in the cytotoxicity of 5-FU in yeast and humans to improve chemotherapy.Rrp6 is a 3'-5 'exoribonuclease, which plays an important role in the regulation and modification of rRNA, mRNA and lncRNA. I found that overexpression of RRP6, the homologue of the yeast EXOSC10 gene in mammals, can lead to increased resistance to treatment with 5-FU. I found that the lncRNA MUT1312 form dsRNA with RRP6 that are negatively correlated with the level of Rrp6 protein. Furthermore, overexpression of MUT1312 during mitosis and associated with a decrease of Rrp6. Thus, my study suggests that MUT1312 may involved in the regulation of Rrp6 during cell differentiation. I further explored the function of the double-stranded RNA in meiosis. My research about SWI4/MUT477 indicates the important function of meiosis induced long noncoding RNA as a form of double-stranded RNA potentially regulate translation. Another aspect of the function of lncRNA is to regulate the transcription of downstream mRNA. I found SUT200 could inhibit transcription of CDC6 during meiosis by read-through. A similar case is CLN2/MUT1465. I did an in silico screening to find transcription factors that activate MUTs during meiosis. I found that most MUTs are induced by Ndt80. MUT1465 is among them: it could be induced by Ndt80 which inhibit the expression of CLN2 after initiation of meiosis. I found that repression of certain MUTs by the Ume6 / Rpd3 complex in mitosis is regulated differently between JHY222 and SK1. MUT100 which does not have the Ume6 binding site URS1 element, and is therefore an indirect target is derepressed in JHY22 ume6 but not in SK1 ume6. For the study about regulation of meiosis isoform, we have found that the histone deacetylase complex Rpd3 / Sin3 / Ume6 prevents the induction of long isoform BOI1 in mitosis by direct binding Ume6 binding to its target URS1.Orc1 is important for DNA replication. I have demonstrated that mORC1 is a direct target of the Ndt80 activator and its binding motif (MSE) is required for induction of isoform mORC1 and meiotic gene SMA2 divergently transcribed. I found that a strain incapable of inducing mORC1 contains abnormally high levels of Orc1 during gametogenesis, which correlates with mORC1 declining Orc1 protein. Since eukaryotic genes often encode multiple transcripts with 5'-UTR of variable length, the findings are likely relevant to gene expression during development and disease in higher eukaryotes. In conclusion, my studies during PhD reveal new targets and thus offer new prospects for improving chemotherapy with 5-FU. Mechanisms include (1) the formation of a double strand with its antisense mRNAs to potentially inhibit translation of mRNA, and (2) downstream inhibition of mRNA transcription read-through of a lncRNA. My work also revealed a lncRNA regulatory mechanism and RNA isoforms dangling growth and cell differentiation.

Recently, long non-coding RNAs (lncRNAs) emerged as important regulators of many cellular functions. Many nuclear lncRNAs regulate the expression of geomically proximal or overlapping protein coding genes. Less clear is whether intergenic lncRNAs can regulate transcription by modulating chromatin at genomically distant loci in an RNA-dependent manner. This thesis investigated molecular functions of Dali, an intergenic central nervous system expressed lncRNA conserved in therian mammals. Dali is transcribed from a locus 50 kb downstream of the Pou3f3 transcription factor gene and performs both genomically local and distal RNA-dependent roles. Its depletion disrupts the differentiation of neuroblastoma cells. Locally, Dali regulates transcription of the Pou3f3 locus. Distally, it preferentially binds near to and regulates active promoters across the genome, including by physically associating with the POU3F3 transcription factor. Dali also interacts with the DNMT1 DNA methyltransferase in mouse and human and regulates CpG island-associated promoters by modulating their DNA methylation levels in trans. This work is the first to demonstrate that a lncRNA can regulate the DNA methylation of CpG island-associated promoters in trans and one of the first large scale studies to identify direct transcriptional targets of a lncRNA genome-wide. It also provides a more detailed molecular dissection of the extended Pou3f3 locus and a framework for the prioritisation and comprehensive functional characterisation of nuclear lncRNAs.

Allergies and asthma result from immune responses to common environmental antigens. Despite the fact that both diseases have been described as having a strong genetic background, influences by environmental factors are not negligible. In fact, direct exposure to environmental pollutants (e.g tobacco smoke or chemicals) has already been associated with the development of such diseases and pieces of evidence pile up that early life exposure to these substances is critical. The group of chemicals called 'phthalates' are classified among the incriminated substances. Long non-coding RNAs (lncRNAs) are part of the cellular elements that can be influenced by such exposure. They have been extensively studied over the last decade, highlighting potential roles in both normal and pathological cell functions, and have been associated with various diseases including some immune-related ones such as allergic reactions. Nevertheless, the role they play in environmental health still requires to be clarified. Maternal exposure to benzyl butyl phthalate (BzBP), a chemical from the phthalate family present in daily-life products (e.g. vinyl flooring, adhesives), has been associated with allergies and asthma development in cohort studies. The underlying cellular mechanisms remain nonetheless mostly unknown. Because the immune cells T helper (TH) and B cells are key actors in such pathologies, we hypothesized that maternal exposure to BzBP may have an impact on asthma and allergy development by affecting these cells. The present study therefore focused on the effect of maternal exposure to BzBP on TH and B cells, with a particular interest in BzBP-induced changes in lncRNAs expression, motivated by their cellular functions and their putative responsiveness to chemicals. Using a murine model and RNA-sequencing, we observed that perinatal (during pregnancy and breastfeeding) and prenatal (pregnancy only) exposure to BzBP led to differential expression in TH cells, including for lncRNAs. With the aim of identifying putative roles of BzBP-regulated lncRNAs, weighted correlation network analysis was performed. The results suggest that some of the non-coding transcripts regulated by prenatal exposure to BzBP may be involved in TH cell-related functions such as cell activation, TH2-cytokines expression and regulation of the inflammatory response. Differential expression of lncRNAs due to prenatal exposure to BzBP was also observed in mouse B cells. Recently published RNA-sequencing data for B cell development and activation (Brazao et al., 2016) were integrated with our own data for correlation network analysis. The results revealed that some of the BzBP-regulated lncRNAs were correlated with genes involved in B cell-related functions. In particular, we hypothesized that two of these lncRNAs may play a role in germinal center B cell functions. Several non-coding transcripts that are regulated by prenatal exposure to BzBP and that may be involved in TH cell-related processes and in germinal center B cell functions were therefore identified. Due to the association of the concerned processes and of this B cell sub-population with the development of allergies and asthma, we suspect these genes to also be actors in such phenomenons after prenatal exposure to BzBP. We propose models of how these genes are involved in immune-related functions and how their regulation by BzBP may be critical. Taken together, these data thus provide a whole new layer of information of the toxicological effect of prenatal exposure to BzBP and of how this exposure might be associated with allergies and asthma development by providing an insight of underlying cellular mechanisms. Moreover, our study also contributes to filling the gaps in the knowledge concerning the roles of lncRNAs in environmental health by showing that BzBP influences non-coding transcripts that may be involved in pathologies.