A identificação de mecanismos moleculares é importante para entender os eventos fisiológicos que ocorrem durante o processo de maturação sexual. O objetivo deste trabalho foi semiquantificar o RNAm para os receptores de gonadotrofinas em células da granulosa de folículos dominantes em crescimento de novilhas Nelore pré-púberes, buscando associação com a fertilidade precoce. As novilhas foram separadas em Precoce (P, n=6) e Não Precoce (NP, n=16), de acordo com o diagnóstico positivo de gestação aos 15 meses de idade. Realizou-se exame ultra-sonográfico dos ovários durante 17 dias para a identificação da onda folicular, o folículo dominante com 8-9 mm foi aspirado via trans-vaginal. O RNA total foi extraído das amostras e convertido a DNA complementar que teve uma região amplificada para a expressão dos genes para os receptores de gonadotrofinas. A semiquantificação foi feita por análise digital onde levou-se em consideração a intensidade média das bandas (unidade arbitrária - UA) em gel de poliacrilamida (8%) corado com uma solução de TBE-brometo de etídio (0,02%). Foi utilizado o RNAm para o gene do GAPDH como controle interno para corrigir a expressão dos receptores de gonadotrofinas em função da quantidade de RNA em cada amostra. Não houve diferença entre P e NP na quantidade de RNAm para o receptor de FSH (FSHr, 1,14±0,07 e 1,19±0,04 UA) e de LH (LHr 0,79±0,12 e 0,78±0,05 UA, respectivamente). Foi detectada a presença de splicing alternativo do RNAm para o receptor de LH (LHrs), pelo aparecimento de banda com peso molecular diferente, em todos os animais do experimento, constatando-se através do sequenciamento que o material apresentava a deleção do exon 10. As novilhas P apresentaram maior percentagem de LHrs em relação ao LHr (LHrs/LHr) (83,66±4,26%) quando comparadas a NP (70,05±3,39%). A quantidade de RNAm para receptores de gonadotrofinas parece não estar relacionada com o aparecimento de fertilidade precoce, mas a maior percentagem de deleção do exon 10 no RNAm para o LHr pode estar envolvida com a fertilidade precoce em novilhas Nelore. / Identification of molecular mechanisms is important to understand the physiological events occurring during the process of sexual maturation. The objective of this work was to semiquantify the mRNA for gonadotropins receptors from granulosa cells of developing dominant follicles in pre-pubertal Nellore heifers and associate the results to precocious fertility. The heifers were sorted in Precocious (P, n=6) and Non Precocious (NP, n=16), according to positive pregnancy at 15 months of age. Ovarian ultrasound examination was performed through 17 days for follicular wave identification, the growing dominant follicle was aspirated at 8-9 mm. Total mRNA was extracted from the granulosa cells samples, converted to complementary DNA and had a region amplified for gene expression for gonadotropins receptors. The semiquantification was performed by digital analysis of intensity average for the bands (arbitrary unit - AU) in poliacrilamida gel (8%) stained with etidio TBE-bromide solution (0,02%). mRNA for GAPDH gene was used as internal control to correct the expression of gonadotropins receptors to the amount of mRNA in each sample. There was no difference between P and NP in the amount of mRNA for FSH receptors (FSHr, 1,14±0,07 and 1,19±0,04 AU) and LH receptors (LHr 0,79±0,12 and 0,78±0,05 AU, respectively). Heifers presented alternative splicing of mRNA for LH receptors (LHrs) evidenced through sequencing, as the absence of exon 10. P heifers presented greater percentage of LHrs in relation to LHr (LHrs/LHr) (83,66±4,26%) when compared the NP (70,05±3,39%). The amount of mRNA for gonadotropins receptors was not related with precocious fertility, but the greater percentage of exon 10 deletion from mRNA for LHr was associate the precocious fertility in Nelore heifers.

In most eukaryotes, alternative splicing is an important regulatory mechanism of gene expression that results in a single gene coding for multiple protein isoforms, thus largely increases the diversity of the proteome. RNA-seq is widely used for genome-wide splicing isoform quantification, and several effective and powerful methods have been developed for splicing analysis with RNA-seq data. However, it remains problematic for genes with low coverages or large number of isoforms. These difficulties may in principle be ameliorated by exploiting correlations encoded in the structured data sources. This thesis contributes to developments of Bayesian methods for splicing analysis by leveraging additional information in multiple datasets with structured prior distributions. First, we developed DICEseq, the first isoform quantification method tailored to time-series RNA-seq experiments. DICEseq explicitly models the correlations between experiments at different time points to aid the quantification of isoforms across experiments. Numerical experiments on both simulated and real datasets show that DICEseq yields more accurate results than state-of-the-art methods, an advantage that can become considerable at low coverage levels. Furthermore, DICEseq permits to quantify the trade-off between temporal sampling of RNA and depth of sequencing, frequently an important choice when planning experiments. Second, we developed BRIE (Bayesian Regression for Isoform Estimation), a Bayesian hierarchical model which resolves the difficulties in splicing analysis in single-cell RNA-seq (scRNA-seq) data by learning an informative prior distribution from sequence features. This method combines the quantification and imputation for splicing analysis via a Bayesian way, which is particularly useful in scRNA-seq data due to its extreme low coverages and high technical noises. We validated BRIE on several scRNA-seq data sets, showing that BRIE yields reproducible estimates of exon inclusion ratios in single cells. Third, we provided an effective tool by using Bayes factor to sensitively detect differential splicing between different single cells. When applying BRIE to a few real datasets, we found interesting heterogeneity patterns in splicing events across cell population, for example alternative exons in DNMT3B. In summary, this thesis proposes structured Bayesian methods to integrate multiple datasets to improve splicing analysis and study its biological functions.

La Dystrophie Myotonique de type 1 (DM1) est cause par l’expansion anormale d’un triplet CTG dans la partie 3’UTR du gène DMPK, entrainant l’agrégation du transcrit mutant dans des inclusions ribonucléoprotéiques appelées foci. D’après plusieurs études structurales sur des courtes répétitions CUG, il a été proposé que les expansions CUG se replient en une structure en tige-boucle qui interfère avec l’activité de plusieurs facteurs lié au métabolisme de l’ARN et altère leur fonction cellulaire. Le facteur d’épissage muscleblind-like 1 (MBNL1) a été identifié par sa capacité à interagir avec les répétitions CUG. In vivo, ces répétitions entrainent la séquestration de cette protéine aboutissant en une déplétion nucléaire. Un autre facteur d’épissage, la CUG Binding protein (CUGBP1), est également impliqué dans la pathologie. Au lieu d’être séquestré par les répétitions, la stabilité protéique de CUGBP1 est augmentée dans les tissus DM1 entrainant un gain d’activité pour ce facteur. La séquestration de MBNL1 et la stabilisation de CUGBP1 résultent en la dérégulation de l’épissage alternatif de plusieurs transcrits musculaires et du cerveau et la réexpression d’isoformes protéiques fœtales dans les tissus adultes. Cependant, de récentes études suggèrent que d’autres facteurs ou voies de signalisation que celles faisant intervenir MBNL1 et CUGBP1 pourraient être impliquées dans la pathologie DM1.Le but de mon travail de thèse a été d’identifier de nouveaux facteurs ayant la capacité d’interagir avec les répétitions CUG. A l’aide d’une purification sur chromatographie d’affinité utilisant un ARN contenant 95 répétitions CUG comme appât, nous avons identifié l’ARN hélicase p68/DDX5. p68 fait partie de la famille des protéines DEAD-box, caractérisée par un core protéique conservé constitué de neufs domaines hautement conservés, dont le motif DEAD, à l’origine du nom de ces protéines. p68 est impliquée dans de nombreux aspects du métabolisme de l’ARN, dont la transcription, l’épissage, l’export, la traduction et la dégradation des ARN. Nous avons montré, que p68 colocalise avec les foci CUG dans un modèle cellulaire exprimant la partie 3’UTR du gène DMPK contenant de longues répétitions CTG. Nous avons identifié que p68 augmente l’interaction de MBNL1 sur les répétitions CUG et une structure secondaire particulière d’un élément régulateur de l’ARN pré-messager cardiac Troponin T (TNNT2), dont l’épissage est dérégulé dans la pathologie. L’insertion de mutations dans le core de l’hélicase de p68 abolit l’effet de p68 sur la fixation de MBNL1 ainsi que la colocalisation de p68 avec les expansions CUG in vivo, suggérant que le remodelage des structures secondaires ARN de manière ATP-dépendante par p68 facilite l’interaction de MBNL1. Nous trouvons également que la compétence de p68 pour réguler l’inclusion de l’exon alternatif 5 de TNNT2 dépend de l’intégrité des sites de fixation de MBNL1.Nous proposons que p68 agit comme un modificateur de l’activité de MBNL1 sur ces cibles d’épissage ainsi que sur les expansions CUG à l’origine de la pathologie. / Myotonic Dystrophy type I (DM1) is caused by an abnormal expansion of CTG triplets in the 3’ UTR of the DMPK gene, leading to the aggregation of the mutant transcript in nuclear RNA foci. Based on structural studies on short CUG repeats, it has been proposed that expanded CUG repeats fold into an imperfect hairpin structure that interferes with the activities of RNA binding proteins and alters their normal cellular function. The muscleblind-like 1 protein (MBNL1) was identified by its ability to bind to CUG repeats. It has been shown that the expanded mutant transcript promotes the sequestration of the MBNL1 splicing factor in nuclear RNA foci. CUGBP1 is another splicing factor that is involved in DM1. Instead of being sequestered by the repeats, the steady-state level of CUGBP1 is increased in DM1 tissues, leading to a gain of activity of the protein. The sequestration of MBNL1 and the up-regulation of CUGBP1 in DM1 results in the misregulation of alternative splicing of a subset of muscle and brain-specific transcripts, leading to the re-expression of fetal isoforms in adult tissues. However, several recent studies suggest that factors or signaling pathways other than MBNL1 and CUGBP1 could be involved in DM1 pathogenesis.The aim of this work was to isolate new factors that bind to CUG repeats. Using an affinity chromatography strategy with an RNA containing 95 pure CUG repeats, we identified the RNA helicase p68 (DDX5). p68 is a prototype of DEAD-box RNA helicase proteins. This family is characterized by a conserved core, consisting of nine conserved motifs including the DEAD signature, which gives rise to the name to these proteins. p68 is involved in many aspects of RNA metabolism including transcription, RNA processing, RNA export, translation and mRNA degradation. We showed that p68 colocalized with RNA foci in cells expressing the 3’UTR of the DMPK gene containing expanded CTG repeats. We found that p68 increased MBNL1 binding onto pathological repeats and the stem-loop structure regulatory element within the cardiac Troponin T (TNNT2) pre-mRNA, splicing of which is misregulated in DM1. Mutations in the helicase core of p68 prevented both the stimulatory effect of the protein on MBNL1 binding and the colocalization of p68 with CUG repeats, suggesting that remodeling of RNA secondary structure through a ATP-dependant manner by p68 facilitates MBNL1 binding. We also found that the competence of p68 for regulating TNNT2 exon 5 inclusion depended on the integrity of MBNL1 binding sites.We propose that p68 acts as a modifier of MBNL1 activity on splicing targets and pathogenic RNA.

O dermatófito Trichophyton rubrum é um fungo filamentoso, antropofílico, que invade tecidos queratinizados causando infecções superficiais e cutâneas. Algumas drogas são usadas para o tratamento das dermatofitoses, sendo o ácido undecanóico (AUN) uma delas. O AUN é o mais tóxico dos ácidos graxos saturados de cadeia média, utilizado como medicamento de uso tópico. O estudo de expressão gênica e mecanismos regulatórios são fundamentais para ampliar o conhecimento dos mecanismos moleculares envolvidos na resposta à exposição a estes agentes citotóxicos. Portanto, o objetivo deste trabalho foi caracterizar os mecanismos moleculares envolvidos no processo adaptativo da exposição ao ácido undecanóico, através da análise global do transcriptoma e mecanismos regulatórios como o processamento alternativo. Para tanto o micélio de T.rubrum foi submetido ao ácido undecanóico por 3 e 12 h de exposição, em triplicata biológica, e o RNA resultante foi submetido ao sequenciamento por RNAseq. O sequenciamento gerou aproximadamente 58 milhões de reads por biblioteca, as quais foram filtradas e alinhadas com o genoma de referência utilizando-se os softwares FASTQC e bowtie2, respectivamente. A análise de expressão gênica diferencial foi feita por meio do pacote do Bioconductor DESeq e, para esta análise, foi utilizada a amostra de 0h como referência. Foram identificados 492 genes diferencialmente expressos em resposta ao AUN, sendo 385 e 210 genes modulados em resposta a 3 e 12 horas de exposição, respectivamente. Estes genes estão relacionados a vários processos celulares envolvendo transporte transmembrana, degradação de xenobióticos, metabolismo de lipídeos e aminoácidos, secreção de enzimas proteolíticas e patogênese, sugerindo que o AUN ativa duas principais vias de sobrevivência em resposta a este agente estressor, a degradação e o efluxo da droga. Posteriormente, foram realizadas as análises de processamento alternativo quanto ao uso diferencial de exons por meio do algoritmo HTSeq e DEXSeq e retenção de introns utilizando-se algoritmos construídos na linguagem Perl. Os genes envolvidos em algum tipo de processamento alternativo estão relacionados com funções metabólicas variadas como tradução, transporte vesicular, metabolismo lipídico, biogênese ribossomal, sequência de ligação ao DNA, regulação da transcrição e processamento do pré-mRNA. Estes resultados contribuem para uma melhor compreensão dos mecanismos moleculares envolvidos na resposta de sobrevivência perante a exposição ao AUN. / The dermatophyte Trichophyton rubrum is a filamentous fungus, anthropophilic that invades keratinized tissue causing superficial infections on the skin. Some drugs are used for the treatment of dermatophytosis, being the undecanoic acid (AUN) one of them. The AUN the most toxic of the saturated medium chain fatty acids, is used as a topical medicine. The study of gene expression and the regulatory mechanisms are fundamental to understand the molecular mechanisms involved in response to exposure of these cytotoxic agents. So, the aim of this study was to characterize the molecular mechanisms involved in the adaptive process of the exposure to undecanoic acid, by the global analysis of the transcriptome and regulatory mechanisms such as alternative splicing. To this end, the mycelium of T. rubrum was exposed to undecanoic acid for 3 or 12 hours in biological triplicate, and the resulting RNA was sequenced by RNA-Seq. The sequencing generated approximately 58 million of reads per library, which were filtered and aligned with the reference genome using the softwares and Bowtie2 and FASTQC, respectively. The analysis of differential gene expression was performed through the Bioconductor package DESeq and, for this analysis, we used as reference sample the 0h time. We identified 492 differentially expressed genes in response to UDA, being 385 and 210 genes modulated in response to 3 and 12 hours of exposure, respectively. These genes are related to various cellular processes involving the transmembrane transport, xenobiotics degradation, lipids and amino acids metabolism, secretion of proteolytic enzymes and pathogenesis, suggesting the activation of two major survival pathways in response to this stressor, degradation and drug efflux, by UDA. Also, the alternative splicing analysis was performed through the differential use of exons using the algorithms HTSeq and DEXSeq and intron retention using algorithms built in Perl language. The genes involved in some kind of alternative splicing are associated with various metabolic functions such as translation, vesicular transport, lipid metabolism, ribosomal biogenesis, DNA binding sequence, transcription regulation and processing of pre-mRNA. These results contribute to increase the knowledge of the molecular mechanisms involved in the survival response upon exposure to the AUN.

Glioblastoma multiforme (G BM), ou astrocitoma grau IV, é o tumor mais comum e letal do sistema nervoso central. Uma de suas características mais marcantes é seu alto potencial invasivo do tecido normal adjacente. Neste processo, o remodelamento da matriz extracelular, modulado por enzimas que degradam seus componentes e por inibidores destas enzimas, é crucial. Foi descrito que a expressão de MMP-2 e MMP-9, membros da família das metaloproteinases de matriz, aumentam conforme a progressão de astrocitomas. A variante canônica de RECK suprime a invasão tumoral e metástase através da inibição da atividade de, pelo menos, três MMPs: MMP-2, MMP-9 e MMP-14. Uma correlação positiva tem sido observada entre a abundância da expressão de RECK em amostras tumorais e um prognóstico mais favorável para pacientes com diversos tipos de tumores. Neste estudo, variantes de splicing do gene supressor de tumor RECK foram identificadas através da análise de Expressed sequenced Tags (ESTs), isoladas por RT-PCR, sequenciadas e clonadas. Três novas variantes de splicing do gene RECK foram identificadas e caracterizadas. O perfil de expressão dos transcritos de RECK foi determinado através de ensaios de RT-PCR quantitativo em um painel de tecidos normais e, também, durante a progressão de astrocitomas. Foram utilizadas, para esta análise, amostras macro dissecadas de tumores de pacientes com astrocitomas grau I (n=15), II (n=15), III (n=15) e GBMs (n=30). Os resultados mostram que maior expressão de RECK canônico, acompanhada de maior razão de expressão da variante canônica em relação às variantes de splicing alternativo, correlaciona positivamente com maior sobrevida global de pacientes com GBM, sugerindo seu papel como potenciais biomarcadores para o prognóstico destes pacientes. Análise funcional das isoform as de RECK em células U87 MG revelou que as células superexpressando as isoformas não apresentam inibição do processo de invasão celular, como observado para superexpressão da proteína canônica. Dentre as isoformas analisadas, destaca-se RECK-B, isoforma potencialmente ancorada à membrana plasmática por GPI, como a proteína canônica RECK, sugerindo uma possível colocalização destas variantes. Observa-se que células superexpressando RECK-B apresentam maior capacidade tumorigênica. Os resultados indicam que as variantes de RECK e o balanço entre a expressão destas variantes, apresentam um papel importante no comportamento e na agressividade de GBMs, tendo potencial valor na clínica. Além disso, para abrir perspectivas para o estudo das variantes de RECK, o balanço de expressão dos transcritos canônico e alternativos deste gene foi explorado durante os processos de diferenciação osteogênica e adipogênica. Os resultados indicam que a expressão da variante canônica é mais abundante em relação à expressão de suas isoformas em estágios tardios da adipogênese, sendo que o perfil inverso é observado em relação à isoforma B durante a osteogênese, sugerindo que o balanço entre os níveis de expressão das isformas de RECK possui um potencial papel biológico que deve ser explorado durante esses processos. Em conjunto, os resultados demonstram a existência de, pelo menos, três variantes de splicing do gene supressor de tumor RECK com envolvimento na tumorigênese e na diferenciação celular, abrindo novas perspectivas para o estudo e a aplicação do gene RECK na clínica. / Glioblastoma multiforme (GBM) or grade IV astrocytoma is the most common and lethal tumor of the central nervous system. One of the most striking features of GBMs is their invasive potential of the normal surrounding brain tissue. It has been described that MMP-2 and MMP-9 expression levels increase during astrocytoma progression. Canonical RECK suppresses tumor invasion and metastasis by negatively regulating at least three matrix metalloproteinases, namely: MMP-9, MMP-2 and MT1-MMP. A positive correlation has been observed between the abundance of RECK express ion in tumor samples and a more favorable prognosis for patients with several types of tumors. In this study, splice variants of the RECK tumor suppressor gene were identified by Expressed Sequence Tag (EST) analysis, isolated by RT-PCR, sequenced and cloned. Three novel alternatively spliced variants of the RECK tumor suppressor gene were identified and characterized. The RECK transcripts expression profiles were investigated using quantitative RT-PCR assays in a normal tissue RNA panel and, also, during astrocytoma progression in macrodissected tumor samples of patients with astrocytoma grades I (n=15), II (n=15), III (n=15) and IV/GBM (n=30). The results show that higher canonical RECK expression, accompanied by a higher ratio of canonical to alternative transcript expression, positively correlated with higher overall survival rate after chemotherapeutic treatment of GBM patients. Our findings suggest that these RECK transcript variants may potentially be used as biomarkers for prognosis of GBM patients. U87 MG cells overexpressing each RECK alternative variant were generated and found to lack the supressive role of cellular invasion processes found upon overexpressing the canonical protein. Among the characterized isoforms, RECK-B stands out, since this isoform is potentially anchored to the cell membrane by a GPI anchor, exactly as the canonical RECK and, also, since cells overexpressing RECK-B display greater tumorigenic capacity. The results indicate that RECK variants and the balance between the expressions of these variants, play an important role in the behavior and aggressiviness of GBMs, therefore have a potential translational application. In addition, in order to investigate new perspectives for the analysis of these isoforms, the expression balance of RECK transcripts was assessed during osteogenesis and adipogenesis, by qRT - PCR. The results show that the expression of the canonical RECK variant is more abundant that that of its alternative isoforms in later stages of adipogenic differentiation. The opposite profile is found regarding RECK-B during osteogenesis, suggesting that the balance between the expressions of these transcripts may have a potential role during these processes. Taken together, the results show the existence of, at least, three alternatively spliced variants of the RECK tumor suppressor gene, which are involved in tumogigenesis and cellular differentiation, o pening new perspectives for studies and clinical application of the RECK gene.

The chicken BG system is a highly polymorphic and polygenic multigene family encoding type I transmembrane proteins, with butryophilins as homologues in mammals, some of which are crucial in T cell regulation. There are three genomic locations where BG genes are found: one singleton BG gene (BG0) on chromosome 2, another singleton gene (BG1) in BF-BL region (the so-called minimal essential chicken MHC) on chromosome 16, and many BG genes arranged tandemly in the BG region just outside the MHC. BG genes in BG region have copy number variation between different chicken haplotypes, so it has been unclear which BG genes are alleles, as very little sequence information has been available for haplotypes other than B12, the best characterized one. Also, the functions of chicken BG genes have been a mystery for half a century, although there is evidence for cytoskeletal regulation and for viral disease resistance. Therefore, the aim of the research was to develop new procedures and reagents to understand the BG system. A novel PCR protocol was established to overcome the difficulty of amplifying full length polymorphic BG transcripts, and then was applied to systematically examine the BG cDNA sequences from T cells and B cells of four different chicken haplotypes. In total 23 BG genes were found, most with alternative splicing isoforms; most strikingly, the transcripts potentially encoding soluble BG proteins were only seen in B cells, indicating functional differences of the same gene in T and B cells. By comparing the dominantly expressed BG genes as 'functional alleles' in these cells, only the cytoplasmic tail region is clearly seen to be under selection, based on the overwhelming preponderance of non-synonymous changes. With many other unexpected findings discovered in this project, a clearer picture of chicken BG genes is presented, and more questions were raised for future study. In order to explore BG functions and further characterize BG proteins, fourteen stable cell lines were developed expressing fusion proteins of the Ig-V domains of the 14 BG genes from the B12 haplotype chicken with the human IgG1 Fc fragment. These BG-Fc fusion proteins were used in sandwich enzyme-linked immunosorbent assays (ELISAs) to screen 290 BG monoclonal antibody (mAb) tissue culture supernatants, and these BG mAbs were further characterized for specificity by western blot using BG-Fc fusion proteins. These solid tools (BG-Fc fusion proteins and BG mAbs) provide the basis to further understand chicken BG functions and answer other interesting questions.

RNA sequencing (RNA-seq) technologies facilitate the characterisation of genes and transcripts in different cell types as well as their expression analysis across various conditions. Due to its ability to provide in-depth insights into transcription and post-transcription mechanisms, RNA-seq has been extensively used in functional genetics and transcriptomics, system biology and developmental biology in animals, plants, diseases, etc. The aim of this project is to use mathematical and computational models to integrate big genomic and transcriptomic data from high-throughput technologies in plant biology and develop new methods to identify which genes or transcripts have significant expression variation across experimental conditions of interest, then to interpret the regulatory causalities of these expression changes by distinguishing the effects from the transcription and alternative splicing. We performed a high resolution ultra-deep RNA-seq time-course experiment to study Arabidopsis in response to cold treatment where plants were grown at 20^oC and then the temperature was reduced to 4^oC. We have developed a high quality <i>Arabidopsis thaliana</i> Reference Transcript Dataset (AtRTD2) transcriptome for accurate transcript and gene quantification. This high quality time-series dataset was used as the benchmark for novel method development and downstream expression analysis. The main outcomes of this project include three parts. i) A pipeline for differential expression (DE) and differential alternative splicing (DAS) analysis at both gene and transcript levels. Firstly, we implemented data pre-processing to reduce the noise/low expression, batch effects and technical biases of read counts. Then we used the limma-voom pipeline to compare the expression at corresponding time-points of 4^oC to the time-points of 20^oC. We identified 8,949 genes with altered expression of which 2,442 showed significant DAS and 1,647 were only regulated by AS. Compared with current publications, 3,039 of these genes were novel cold-responsive genes. In addition, we identified 4,008 differential transcript usage (DTU) transcripts of which the expression changes were significantly different to their cognate DAS genes. ii) A TSIS R package for time-series transcript isoform switch (IS) analysis was developed. IS refers to the time-points when a pair of transcript isoforms from the same gene reverse their relative expression abundances. By using a five metric scheme to evaluate robustly the qualities of each switch point, we identified 892 significant ISs between the high abundance transcripts in the DAS genes and about 57% of these switches occurred very rapidly between 0-6h following transfer to 4^oC. iii) A RLowPC R package for co-expression network construction was generated. The RLowPC method uses a two-step approach to select the high-confidence edges first by reducing the search space by only picking the top ranked genes from an initial partial correlation analysis, and then computes the partial correlations in the confined search space by only removing the linear dependencies from the shared neighbours, largely ignoring the genes showing lower association. In future work, we will construct dynamic transcriptional and AS regulatory networks to interpret the causalities of DE and DAS. We will study the coupling and de-coupling of expression rhythmicity to the Arabidopsis circadian clock in response to cold. We will develop new methods to improve the statistical power of expression comparative analysis, such as by taking into account the missing values of expression and by distinguishing the technical and biological variabilities.

To increase diversity of enzymes and proteins, cells mix and match exonic and intronic regions retained in mature mRNAs by alternative splicing. An estimated 94% of all multi-exon genes express one or more alternatively spliced transcripts generating proteins with similar or modified functions. Cyclooxygenase is a signaling enzyme that catalyzes the rate-limiting step in the synthesis of diverse bioactive lipids termed prostaglandins. Prostaglandins are involved in myriad physiological and pathopysiological processes including vasoregulation, stomach mucosal maintenance, parturition, pain, fever, inflammation, neoplasia and angiogenesis and are inhibited by aspirin-like drugs known as NSAIDs. In 2002 an alternatively spliced, intron-1 retaining variant of COX-1 was cloned from canine brain tissue. This new variant, termed COX-3 or COX-1b, is an enzymatically active prostaglandin synthase expressed at relatively high levels in a tissue and cell type dependant manner in all species examined. In humans and most rodent species intron-1 is 94 and 98 nucleotides long respectively. Retention of the intron in these species introduces a frameshift and is predicted to result in translation of a very small 8-16kD protein with little similarity to either 72kD COX-1 or COX-2, calling into question the role of this variant. In this dissertation, I present my results from cloning and ectopically expressing a complete and accurate COX-3 cDNA from both rat and human. I confirmed that COX-3 mRNA encodes multiple large molecular weight cyclooxygenase-like proteins in the same reading frame as COX-1. Translation of these proteins relies on several recoding mechanisms including cap-independent translation initiation, alternative start site selection, and ribosomal frameshifting. Using siRNA and Western blotting I have identified some of these proteins in tissues and cells. Two COX-3 encoded proteins are active prostaglandin synthase enzymes with activities similar to COX-1 and represent novel targets of NSAIDs. Other COX-3 proteins have unknown function, but their size and cellular location suggest potential roles as diverse as cytosolic enzymes and nuclear factors.

The polyglutamine diseases are a clinically heterogeneous group of inherited neurodegenerative disorders caused by expansion of polyglutamine-encoding (CAG)n trinucleotide repeats within the disease genes. It is increasingly clear that the amino acid sequences flanking the polyglutamine expansion in each disease protein, i.e. the specific protein context, contribute to selective neuronal toxicity by influencing the behavior of the disease protein within selectively vulnerable neuronal populations. In the studies described here, I explore the role that protein context plays in the polyglutamine disease, Spinocerebellar ataxia type 3 (SCA3). Toward this end, I utilize biochemical, cell-based, and animal models to gain a broader understanding of the SCA3 disease protein, ataxin-3, and generate tools for further exploration of the molecular properties of ataxin-3 that modulate its toxicity during disease. In Chapter 1, I provide an overview of the recognized polyglutamine diseases, emphasizing the elements of protein context that are distinct among the polyglutamine disease proteins and may contribute to the neuropathological and clinical heterogeneity within this family of diseases. Alternative splicing of the polyglutamine disease gene products adds an additional level of complexity to the tissue-specific protein context of expanded polyglutamine, yet this phenomenon has been underinvestigated. In Chapter 2, I examine the significance of ataxin-3 splice variation. Several minor 5' variants and both known 3' splice variants of ataxin-3, a deubiquitinating enzyme, are expressed at the mRNA level in brain. At the protein level, however, the C-terminal splice isoform with three ubiquitin interacting motifs (3UIM ataxin-3) is the predominant isoform in brain, independent of age or (CAG)n expansion. Although both C-terminal ataxin-3 splice isoforms display similar in vitro deubiquitinating activity, 2UIM ataxin-3 is more prone to aggregate and is more rapidly degraded by the proteasome. These observations demonstrate how alternative splicing of sequences distinct from the polyglutamine-encoding (CAG)n repeat can alter disease-related components of protein context. Knock-in models of polyglutamine diseases utilize pathogenic (CAG)n expansions within the endogenous genomic, transcript, and protein context to recreate key features of individual polyglutamine diseases. In chapter 3, I describe the creation of the first knock-in mouse model of SCA3. Hemizygous knock-in mice transmit the knock-in allele in Mendelian ratios and broadly express both the expanded Atxn3(Q3KQ82) protein and the wildtype murine Atxn3(Q6) protein. In this chapter, I also compare the gene targeting efficiencies and rates of chromosomal instability of a novel C57BL/6J ES cell line (UMB6JD7) and two well established ES cell lines (W4 and Bruce4.G9). Of these, Bruce4.G9 ES cells proved superior based on lower rates of aneuploidy and the production of germline transmitting chimeras. Finally, in Chapter 4 I discuss questions and concepts raised during the course of these studies, and suggest avenues of future research aimed at broadening our understanding of ataxin-3 physiology and of protein context-dependent elements in polyglutamine disease pathogenesis.

<p>Our genome operates by sending instructions, conveyed by mRNA, for the manufacture of proteins from chromosomal DNA in the nucleus of the cell to the protein synthesizing machinery in the cytoplasm. Alternative splicing is a natural process in which a single gene can encode multiple related proteins. During RNA splicing, introns are selectively removed resulting in alternatively spliced gene products. Alternatively spliced protein products can have very different biological effects, such that one protein isoform is disease-related while another isoform is desirable. Splice switching opens the door to new drug targets, and antisense oligonucleotides (asONs), designed to switch splicing, are effective drug candidates. Cellular uptake of oligonucleotides(ONs) is poor, therefore utilization of cell-penetrating peptides (CPPs), well recognized for intracellular cargo delivery, is a promising approach to overcome this essential issue. Most CPPs are internalized by endocytosis, although the mechanisms involved remain controversial.</p><p>Here, evaluation of CPP-mediated ON delivery over cellular membranes has been performed. A protocol that allows for convenient assessment of CPP-mediated cellular uptake and characterization of corresponding internalization routes is established. The protocol is based on both fluorometric uptake measurements and a functional splice-switching assay, which in itself is based on biological activity of conveyed ONs. Additionally, splice switching ONs (SSOs) have been optimized for high efficiency and specificity. Data suggest that SSO activity is improved for chimeric phosphorothioate SSOs containing locked nucleic acid (LNA) monomers. It is striking that the LNA monomers in such chimeric constructs give rise to low mismatch discrimination of target pre-mRNA, which highlight the necessity to optimize sequences to minimize risk for off-target effects.</p><p>The results are important for up-coming work aimed at developing compounds consisting of peptides and novel synthetic nucleic acids, making these entities winning allies in the competition to develop therapeutics regulating protein expression patterns.</p>