Genome sequencing has yielded unparalleled insights into fundamental biological processes and the genetics that guide them. In contrast to expectations that protein-coding genes would be the primary output of eukaryotic genomes, however, it is now clear that the vast majority of transcription is devoted to noncoding RNAs (ncRNAs). Although originally regarded as 'transcriptional noise', it is now clear that these transcripts are essential regulators of genetic activity. In this thesis I build upon the hypothesis that the genomes of eukaryotes encode a regulatory 'RNA machine' dominated by ncRNAs. In the Introduction (Chapter 1) I discuss how prior gene models may have inadvertently prevented a full understanding of ncRNAs, review the transcriptional landscape of eukaryotes, and examine the biogenesis and function of small regulatory RNAs. In support of a role for ncRNAs in complex metazoa, Chapter 2 presents data showing a positive correlation between the proportion of non-protein-coding DNA and biological complexity, suggesting that the evolutionary trajectory of intricate developmental phenotypes may have been facilitated by ncRNAs. In the following chapters two more 'gears' are added to the RNA machine. Chapter 3 details the discovery of snoRNA-derived RNAs - an evolutionarily ancient class of Argonaute-assocaited RNA whose biogenesis overlaps with microRNAs (miRNAs) and silencing RNAs (siRNAs). Likewise, Chapter 4 reports a new class of ~18 nt transcription initiation RNAs (tiRNAs) derived from regions proximal to transcription start sites. tiRNAs are enriched at GC-rich promoters and regions of active transcription, implicating them in transcriptional regulation. Chapter 5 presents evidence that tiRNAs are restricted to metazoa, and describes a model of RNA Polymerase II dependent tiRNA biogenesis. This thesis concludes with a general discussion of the implications of these findings, and the potential development of RNA therapeutics. Gathering evidence suggests that eukaryotic genomes are driven by a complex and interwoven network of RNA regulatory feedback loops. This thesis takes a small step towards developing a complete picture of this system.

Genome sequencing has yielded unparalleled insights into fundamental biological processes and the genetics that guide them. In contrast to expectations that protein-coding genes would be the primary output of eukaryotic genomes, however, it is now clear that the vast majority of transcription is devoted to noncoding RNAs (ncRNAs). Although originally regarded as 'transcriptional noise', it is now clear that these transcripts are essential regulators of genetic activity. In this thesis I build upon the hypothesis that the genomes of eukaryotes encode a regulatory 'RNA machine' dominated by ncRNAs. In the Introduction (Chapter 1) I discuss how prior gene models may have inadvertently prevented a full understanding of ncRNAs, review the transcriptional landscape of eukaryotes, and examine the biogenesis and function of small regulatory RNAs. In support of a role for ncRNAs in complex metazoa, Chapter 2 presents data showing a positive correlation between the proportion of non-protein-coding DNA and biological complexity, suggesting that the evolutionary trajectory of intricate developmental phenotypes may have been facilitated by ncRNAs. In the following chapters two more 'gears' are added to the RNA machine. Chapter 3 details the discovery of snoRNA-derived RNAs - an evolutionarily ancient class of Argonaute-assocaited RNA whose biogenesis overlaps with microRNAs (miRNAs) and silencing RNAs (siRNAs). Likewise, Chapter 4 reports a new class of ~18 nt transcription initiation RNAs (tiRNAs) derived from regions proximal to transcription start sites. tiRNAs are enriched at GC-rich promoters and regions of active transcription, implicating them in transcriptional regulation. Chapter 5 presents evidence that tiRNAs are restricted to metazoa, and describes a model of RNA Polymerase II dependent tiRNA biogenesis. This thesis concludes with a general discussion of the implications of these findings, and the potential development of RNA therapeutics. Gathering evidence suggests that eukaryotic genomes are driven by a complex and interwoven network of RNA regulatory feedback loops. This thesis takes a small step towards developing a complete picture of this system.

Recentemente, tem sido demonstrado que a maioria dos RNAs transcritos em células humanas são RNAs não-codificadores de proteínas (ncRNAs) originados de íntrons ou regiões intergênicas. Em trabalhos anteriores realizados por nosso grupo, foram descritos longos ncRNAs transcritos de regiões intrônicas de genes codificadores e cuja expressão foi correlacionada ao grau de diferenciação de tumores de próstata, apontando para a relevância fisiológica desta classe de transcritos. Apesar de sua abundância, as propriedades, funções e regulação da grande maioria dos ncRNAs ainda não foram elucidadas. O objetivo do presente trabalho foi investigar a biossíntese de ncRNAs intrônicos em células humanas, primordialmente a contribuição da RNA Polimerase II (RNAP II), bem como aspectos de sua regulação. Primeiramente, o modelo de regulação da expressão gênica por hormônio andrógeno foi utilizado para avaliação da participação direta de um fator de transcrição de RNAP II, o Receptor de Andrógeno (AR), na modulação da transcrição de ncRNAs intrônicos. Utilizando-se a técnica de imunoprecipitação da cromatina, foi detectada a ligação do AR ao elemento de resposta a andrógeno (ARE) presente em um possível promotor de um transcrito intrônico antisenso (derivado do locus Myo5A), cuja expressão é aumentada em células da linhagem LNCaP tratadas com o hormônio. A ligação ao ARE foi induzida pelo tratamento, sugerindo que o efeito do andrógeno na expressão do ncRNA é mediado pelo AR. Em uma segunda abordagem, o efeito da inibição da transcrição por RNAP II com α-amanitina por 24 h em células LNCaP foi avaliado com o uso de microarranjos de oligonucleotídeos representando transcritos total ou parcialmente intrônicos, além de éxons de genes codificadores. A expressão de menos de 20 % dos transcritos intrônicos foi afetada, fração significativamente menor que a observada para os transcritos exônicos (40 %). Ainda que a maioria dos ncRNAs intrônicos diferencialmente expressos tenha sua abundância diminuída, interessantemente, 13 a 16 % foram aumentados, contrastando com aproximadamente 2 a 3 % de exônicos que aumentaram. Os resultados obtidos neste trabalho indicam que a RNAP II atua na transcrição de ncRNAs intrônicos, mas que uma fração considerável pode ser transcrita por outra RNA Polimerase. / It has been recently shown that the bulk of the transcription in human cells is comprised of non-protein-coding RNAs (or noncoding RNAs - ncRNAs) transcribed from introns and intergenic regions of the genome. Previous work from our group has demonstrated that expression of long intronic ncRNAs can be correlated to the degree of prostate tumor differentiation, underscoring the physiological relevance of these transcripts. However, the properties, functions, and regulation of this huge population of ncRNAs remain largely unknown. The present work aimed to investigate the biosynthesis of intronic ncRNAs and aspects of its regulation in human cells, focusing on the contribution of RNA Polymerase II (RNAP II). Initially, the model of regulation of gene expression by androgen hormone was used in order to evaluate the participation of the RNAP II transcription factor Androgen Receptor (AR) in the transcriptional regulation of intronic ncRNAs. Chromatin immunoprecipitation experiments revealed the binding of the AR in an androgen response element (ARE) present in a putative promoter driving the expression of an antisense intronic transcript in Myo5A locus in LNCaP cells. The interaction occurred in an androgen-inducible fashion, along with the up-regulation of the transcript, suggesting that hormone activation occurred in a direct manner mediated by the AR. In a different approach, the effect of RNAP II inhibition with α-amanitin for 24 h in LNCaP cells was analyzed using an oligoarray representing totally and partially intronic transcripts, as well as exons of proteincoding genes. The expression of less than 20 % of the intronic transcripts was affected by the treatment, contrasting to a significantly higher fraction observed for exonic messages (40 %). Moreover, most differentially expressed intronic transcripts were down-regulated, but strikingly 13 to 16 % were up-regulated in cells with blocked RNAP II, while this fraction for exonic transcripts was about 2 %. The results described here demonstrate that RNAP II in fact plays a role in intronic transcription in human cells, but also highlight that another transcriptional system may account for the biogenesis of a fraction of intronic ncRNAs.

RNAs não codificantes (ncRNAs) têm ganho crescente prestígio nos últimos anos devido a recentes e contínuas descobertas revelando sua diversidade e importância. Porém, a identificação dessas moléculas ainda é um problema em aberto. Em particular, Plasmodium falciparum é um desafio para a pesquisa de ncRNAs, onde poucos foram identificados até o momento. P. falciparum é o parasita que causa uma malária humana letal. A descoberta de novos ncRNAs neste organismo pode auxiliar no desenvolvimento de novos tratamentos. Este trabalho faz um estudo sobre técnicas computacionais para a predição de ncRNAs e, utilizando como objeto de estudo P. falciparum, propõe uma metodologia de predição que seja aplicável inclusive a genomas com viés composicional. A ênfase deste estudo foi a predição de ncRNAs família-específicos, utilizando o componente RNA da telomerase como objeto de estudo. Este é um importante RNA que, devido à sua alta taxa de mutação, é de difícil identificação. Este RNA ainda não foi identificado em P. falciparum. No entanto, evidências biológicas indicam que este RNA é presente, funcional e deve ser essencial ao parasita, caracterizando-se como um alvo de drogas. Além disso, foi realizado um trabalho preliminar sobre a predição de ncRNAs em geral em P. falciparum utilizando uma abordagem comparativa. / Noncoding RNAs (ncRNAs) have been receiving increasing prestige in the last years due to recent and continuous discoveries revealing their diversity and importance. However, the identification of these molecules is still an open problem. In particular, Plasmodium falciparum is a challenge for the ncRNA research, in which few ncRNAs have been identified. P. falciparum is the parasite that causes a lethal human malaria. The discovery of new ncRNAs in this organism may help in the development of new treatments. This work does a research of computational techniques for the ncRNA prediction and, by using P. falciparum as target, proposes a prediction methodology which is also applicable to compositionally biased genomes. The emphasis of this study was the prediction of family-specific ncRNAs, by using the telomerase RNA component as target. This is an important RNA that has a high mutation rate, being difficult to predict. This RNA has not been identified in P. falciparum, yet. However, biological evidences indicate this RNA is present, functional and might be essential for the parasite, being a drug target. In addition, this work presents preliminary results about the prediction of general ncRNAs in P. falciparum by using a comparative approach.

O carcinoma de célula renal (CCR) subtipo célula clara é o câncer mais letal e prevalente do sistema urinário. A transformação maligna no CCR está possivelmente associada à mudanças no perfil de expressão de oncogenes e genes supressores de tumor, e acredita-se que estas alterações sejam críticas para o desenvolvimento do fenótipo maligno. Para identificar novos genes e vias moleculares associadas à transformação maligna no CCR célula clara, foram analisados perfis de expressão gênica de amostras pareadas de tumor e tecido não tumoral adjacente de 6 pacientes. Foi utilizada uma plataforma de microarrays de cDNA contendo 2.292 sondas mapeando éxons de genes codificadores e 822 sondas de RNAs não-codificadores mapeando em regiões intrônicas. A transcrição intrônica foi detectada em todos os tecidos normais e neoplásicos. Utilizando uma combinação de dois testes estatísticos e uma validação por leave-one-out, foi selecionado um subconjunto de 64 transcritos com expressão significativamente alterada em CCR célula clara em relação ao tecido não tumoral adjacente, estando a maior parte (86%) com expressão diminuída em CCR. Entre os transcritos com expressão diminuída, 49 mapearam em regiões não-traduzidas ou éxons de genes codificadores e 6 mapearam em regiões intrônicas de genes codificadores conhecidos. Os níveis de expressão diminuída de SIN3B, TRIP3, SYNJ2BP e NDE1 (p < 0,02), e de transcritos intrônicos derivados dos loci de SND1 e ACTN4 (p < 0,05), foram confirmados em CCR célula clara por Real-time RT-PCR. Um subconjunto de 25 transcritos se mostrou alterado em 6 amostras adicionais de CCR não célula clara, indicando alterações transcricionais comuns em CCR independentemente do subtipo histológico ou do estado de diferenciação do tumor. Além disso, foi analisado o perfil de metilação dos genes com expressão diminuída em tumor SIN3B, TRIP3, SYNJ2BP e GPX3. Nossos resultados indicam um novo conjunto de candidatos a gene supressor de tumor, que 8 podem desempenhar um papel importante na transformação maligna de células renais normais. / The clear cell subtype of renal cell carcinoma (RCC) is the most lethal and prevalent cancer of the urinary system. The carcinogenesis in RCC is thought to be associated with changes in the expression of several genes, and this alteration in gene expression is believed to be critical to the development of the malignant phenotype. To investigate new genes and molecular pathways associated with malignant transformation in clear cell RCC, gene expression profiles of matched samples of tumor and adjacent non-neoplastic tissue obtained from 6 patients were analysed. A custom-built cDNA microarray platform was used, comprising 2,292 probes that map to exons of genes and 822 probes for noncoding RNAs mapping to intronic regions. Intronic transcription was detected in all normal and neoplastic renal tissues. A subset of 64 transcripts with levels significantly deregulated in clear cell RCC relative to the matched non-tumor tissue, mostly (86%) downregulated in CCR, was

O gene supressor tumoral RASSF1A tem sido associado com redução da proliferação celular em diversos tumores. Sua expressão é regulada por eventos epigenéticos que envolvem o complexo repressivo polycomb (PRC2), no entanto os mecanismos moleculares da modulação do recrutamento deste modificador epigenético para este locus ainda são desconhecidos. Neste trabalho identificamos e caracterizamos ANRASSF1, um RNA não codificador longo (lncRNA) intrônico unspliced, que é transcrito na fita oposta do gene RASSF1A, em várias linhagem celulares e tecidos, e se liga a PRC2. ANRASSF1 é transcrito pela RNAPII, possui cap-5´ e cauda poli-A, além de localizar-se no núcleo e possuir uma meia-vida em média quatro vezes menor comparada com outros lncRNAs ligados à PRC2. A super-expressão ectópica de ANRASSF1 reduziu os níveis de RASSF1A e aumentou a taxa de proliferação em células HeLa, enquanto seu silenciamento provocou efeito oposto. Essas mudanças nos níveis de ANRASSF1 não afetaram a abundância da isoforma RASSF1C em nenhuma das condições. A super-expressão de ANRASSF1 provocou um grande aumento tanto da ocupação de PRC2 como da marca de histona repressiva H3K27me3 especificamente na região promotora RASSF1A. Nenhum efeito da super-expressão de ANRASSF1 foi detectado na ocupação de PRC2 e na histona H3K27me3 nas regiões promotoras de RASSF1C e de outros quatro genes vizinhos, incluindo dois genes supressores tumorais bem caracterizados. Além disso, foi demonstrado que ANRASSF1 forma um híbrido de RNA/DNA e recruta SUZ12, um componente do PRC2, para o promotor de RASSF1A. Notavelmente, foi detectado pelo ensaio de RNase-ChIP que a degradação de ANRASSF1 diminui a ocupação de PRC2 neste promotor. Esses resultados demonstram um novo mecanismo de repressão epigenética do supressor tumoral RASSF1A, envolvendo um lncRNA unspliced antissenso, onde ANRASSF1 reprime seletivamente a expressão da isoforma de RASSF1 que sobrepõe o transcrito antissenso de modo local e específico. Considerando uma perspectiva mais ampla, nossos resultados sugerem que outros lncRNAs intrônicos unspliced não caracterizados no genoma humano podem contribuir para uma modulação epigenética local e específica de cada região em que os lncRNAs são transcritos. / Tumor-suppressor RASSF1A gene down-regulation has been implicated in increasing cell proliferation in several tumors. Its expression is regulated by epigenetic events involving polycomb repressive complex 2 (PRC2), however the molecular mechanisms modulating recruitment of this epigenetic modifier to the locus remain largely unknown. Here, we identify and characterize ANRASSF1, an endogenous unspliced long noncoding RNA (lncRNA) that is transcribed from the opposite strand of RASSF1 gene in several cell lines and tissues, and binds to PRC2. ANRASSF1 is transcribed by RNA Polymerase II, 5\'-capped, polyadenylated, displays nuclear localization, and has on average a four-fold shorter half-life compared to other lncRNAs that bind PRC2. ANRASSF1 ectopic overexpression decreases RASSF1A abundance and increases the proliferation rate of HeLa cells, whereas its silencing causes opposite effects. These changes in NRASSF1 levels do not affect RASSF1C isoform abundance. ANRASSF1 overexpression causes a marked increase both in PRC2 occupancy and in histone H3K27me3 repressive mark specifically at the RASSF1A promoter region. No effect of ANRASSF1 overexpression is detected on PRC2 occupancy and on histone H3K27me3 at the promoter regions of RASSF1C and of four other neighbor genes, including two well-characterized tumor suppressor genes. Additionally, we demonstrate that ANRASSF1 forms an RNA/DNA hybrid, and recruits SUZ12, a PRC2 component, to the RASSF1A promoter. Notably, depletion of ANRASSF1 disrupts SUZ12 occupancy on RASSF1A promoter as measured by RNAse-ChIP assay. Together, these results show a new mechanism of epigenetic repression of RASSF1A tumor suppressor gene involving an antisense unspliced lncRNA, in which ANRASSF1 selectively represses expression of the RASSF1 isoform overlapping the antisense transcript in a location-specific manner. In a broader perspective, our findings suggest that other non-characterized unspliced intronic lncRNAs transcribed in the human genome may contribute to a location-specific epigenetic modulation of genes.

The TRIB1 locus (8q24.13) is a novel locus associated with plasma TGs and CAD risk. Trib1 is a regulator of MAPK activity, and has been shown to regulate hepatic lipogenesis and VLDL production in mice. However, the functional relationship between common SNPs at the TRIB1 locus and plasma lipid traits is unknown; TRIB1 has not been identified as an eQTL. This cluster of SNPs falls within an intergenic region 25kb to 50kb downstream of the TRIB1 coding region. By phylogenetic footprinting analysis and DNA genotyping, we identified an evolutionarily conserved region (CNS1) within the risk locus that harbours two common SNPs in tight LD with GWAS risk SNPs and significantly associated with CAD. We investigated the regulatory function of CNS1 by luciferase reporter assays in HepG2 cells and demonstrate that this region has promoter activity. In addition, the rs2001844 risk allele significantly reduces luciferase activity, suggesting that altered expression of the EST-based gene may be associated with plasma TGs. We identified an EST within the risk locus directly downstream of CNS1. We performed 5'/3' RACE using HepG2 RNA, identified multiple variants of this EST-based gene, and confirmed its transcription start site within CNS1. We hypothesize that this EST is a long noncoding RNA due to low abundance, poor conservation, and absence of significant ORF. Over-expression of a short variant implicates its function in the regulation of target gene transcription, although the mechanism of action remains unknown. We conclude that the risk locus at 8q24.13 harbours a novel EST-based gene that may explain the relationship between GWAS SNPs at this locus and plasma lipid traits.

The TRIB1 locus (8q24.13) is a novel locus associated with plasma TGs and CAD risk. Trib1 is a regulator of MAPK activity, and has been shown to regulate hepatic lipogenesis and VLDL production in mice. However, the functional relationship between common SNPs at the TRIB1 locus and plasma lipid traits is unknown; TRIB1 has not been identified as an eQTL. This cluster of SNPs falls within an intergenic region 25kb to 50kb downstream of the TRIB1 coding region. By phylogenetic footprinting analysis and DNA genotyping, we identified an evolutionarily conserved region (CNS1) within the risk locus that harbours two common SNPs in tight LD with GWAS risk SNPs and significantly associated with CAD. We investigated the regulatory function of CNS1 by luciferase reporter assays in HepG2 cells and demonstrate that this region has promoter activity. In addition, the rs2001844 risk allele significantly reduces luciferase activity, suggesting that altered expression of the EST-based gene may be associated with plasma TGs. We identified an EST within the risk locus directly downstream of CNS1. We performed 5'/3' RACE using HepG2 RNA, identified multiple variants of this EST-based gene, and confirmed its transcription start site within CNS1. We hypothesize that this EST is a long noncoding RNA due to low abundance, poor conservation, and absence of significant ORF. Over-expression of a short variant implicates its function in the regulation of target gene transcription, although the mechanism of action remains unknown. We conclude that the risk locus at 8q24.13 harbours a novel EST-based gene that may explain the relationship between GWAS SNPs at this locus and plasma lipid traits.

The TRIB1 locus (8q24.13) is a novel locus associated with plasma TGs and CAD risk. Trib1 is a regulator of MAPK activity, and has been shown to regulate hepatic lipogenesis and VLDL production in mice. However, the functional relationship between common SNPs at the TRIB1 locus and plasma lipid traits is unknown; TRIB1 has not been identified as an eQTL. This cluster of SNPs falls within an intergenic region 25kb to 50kb downstream of the TRIB1 coding region. By phylogenetic footprinting analysis and DNA genotyping, we identified an evolutionarily conserved region (CNS1) within the risk locus that harbours two common SNPs in tight LD with GWAS risk SNPs and significantly associated with CAD. We investigated the regulatory function of CNS1 by luciferase reporter assays in HepG2 cells and demonstrate that this region has promoter activity. In addition, the rs2001844 risk allele significantly reduces luciferase activity, suggesting that altered expression of the EST-based gene may be associated with plasma TGs. We identified an EST within the risk locus directly downstream of CNS1. We performed 5'/3' RACE using HepG2 RNA, identified multiple variants of this EST-based gene, and confirmed its transcription start site within CNS1. We hypothesize that this EST is a long noncoding RNA due to low abundance, poor conservation, and absence of significant ORF. Over-expression of a short variant implicates its function in the regulation of target gene transcription, although the mechanism of action remains unknown. We conclude that the risk locus at 8q24.13 harbours a novel EST-based gene that may explain the relationship between GWAS SNPs at this locus and plasma lipid traits.

Les rétrotransposons sont des éléments génétiques mobiles qui constituent presque la moitié de notre génome. Seule la sous-famille L1HS appartenant à la classe des Long Interspersed Element-1(LINE-1 ou L1) a gardé une capacité de mobilité autonome chez l’Homme. Leur mobilisation dans la lignée germinale, mais Aussi dans certains tissus somatiques, contribue à la diversité du génome humain ainsi qu’à certaines maladies comme le cancer. Ainsi, de nouvelles copies de L1 peuvent directement s'intégrer dans des séquences codantes ou régulatrices, et altérer leur fonction. De plus, les séquences L1 contiennent elles-mêmes plusieurs éléments cis-régulateurs et leur insertion à proximité ou dans un gène peut produire des altérations génétiques plus subtiles. Afin d'explorer l'ensemble de ces altérations à l'échelle du génome, nous avons développé un logiciel dédié à l’analyse des données de séquençage d'ARN qui permet d'identifier des transcrits chimériques ou antisens impliquant les L1 et d'annoter ces isoformes en fonction des différents événements d’épissage alternatif subits. Au cours de ce travail, il est apparu que la compréhension du lien entre polymorphisme des insertions et phénotype nécessite une vue complète des différentes copies L1HS présentes chez un individu donné. Afin de disposer d'un catalogue aussi complet que possible de ces polymorphismes identifiés dans des échantillons humains sains ou pathologiques et publiés dans des journaux scientifiques, nous avons développé euL1db, la base de données des insertions de rétrotransposon L1HS chez l’Homme. En conclusion, ce travail aidera à comprendre l’impact des L1 sur l’expression des gènes, à l'échelle du génome. / Retrotransposons are mobile genetics elements, which form almost half of our genome. Only the L1HS subfamily of the Long Interspersed Element-1 class (LINE-1 or L1) has retained the ability to jump autonomously in humans. Their mobilization in the germline – but also in some somatic tissues – contributes to human genetic diversity and to diseases, such as cancer. L1 reactivation can be directly mutagenic by disrupting genes or regulatory sequences. In addition, L1 sequences themselves contain many regulatory cis-elements. Thus, L1 insertions near a gene or within intronic sequences can also produce more subtle genic alterations. To explore L1-mediated genic alterations in a genome-wide manner, we have developed a dedicated RNA-seq analysis software able to identify L1 chimeric or antisense transcripts and to annotate these novel isoforms with their associated alternative splicing events. During the course of this work, it appeared that understanding the link between L1HS insertion polymorphisms and phenotype or disease requires a comprehensive view of the different L1HS copies present in a given individual or sample. To provide a comprehensive summary of L1HS insertion polymorphisms identified in healthy or pathological human samples and published in peer-reviewed journals, we developed euL1db, the European database of L1HS retrotransposon insertions in humans. This work will help understanding the overall impact of L1 insertions on gene expression, at a genome-wide scale.