INTRODUÇÃO: O carcinoma adrenocortical é uma neoplasia rara que carreia um prognóstico reservado. Recentemente, uma série de estudos demonstrou o potencial do perfil de miRNAs na diferenciação entre adenomas e carcinomas adrenocorticais, estratificação de risco e prognóstico. Entretanto, pouco se sabe ainda sobre a regulação pós-transcricional de miRNAs. Nesse contexto, o LIN28 é uma proteína ligadora de RNAs altamente conservada que surgiu como um modulador do let-7, uma importante família de miRNAs amplamente conhecida por seus efeitos supressivos tumorais. Além do let-7, o LIN28 também mostrou regular e ser regulado pelo mir-9, mir-30 e mir-125. OBJETIVOS: Analisar a expressão gênica e proteica do LIN28 em uma grande coorte de tumores adrenocorticais (TACs) de adultos e pediátricos, além de investigar a variação no número de cópias dos genes LIN28A e LIN28B e a expressão dos miRNAs regulatórios do LIN28 (família let-7, mir-9, mir-30 e mir-125) em um subgrupo desta coorte. MÉTODOS: A expressão proteica do LIN28 foi avaliada em um total de 266 TACs de adultos (78 adenomas e 188 carcinomas) e 44 pediátricos (35 clinicamente benignos e 9 clinicamente malignos). A expressão dos genes LIN28A e LIN28B foi avaliada em um subgrupo de 86 TACs adultos e pediátricos e a análise da variação no número de cópias destes genes em 58 TACs. O estudo de expressão das famílias dos miRNAs let-7, mir-9, mir-30 e mir-125 foi realizado em 28 carcinomas adrenocorticais de adultos. RESULTADOS: Em adultos, o gene LIN28A mostrou-se hiperexpresso em carcinomas agressivos quando comparado a adenomas [7,0 (0 a 174,3) vs. 3,6 (0 a 18,3); p = 0,006, respectivamente] e observou-se uma tendência a maior expressão quando comparados a carcinomas não agressivos [7,0 (0 a 174,3) vs. 7,1 (0 a 17,1); p = 0,092]. A expressão do LIN28B foi negativa na grande maioria (92%) dos TACs de adultos. Curiosamente, uma imunorreatividade fraca para o LIN28 foi significativamente associada com diminuição da sobrevida livre de doença nessa população (p = 0,01), mas para sobrevida global apenas uma tendência foi observada (p = 0,117). Na análise multivariada, somente o índice Ki67 >= 10% (RR 5,7, 95% IC 3,0-10,8; p= 0,0001) e imunorreatividade fraca para o LIN28 (RR 2,3, 95% IC 1,2-4,4; p = 0,008) foram preditores independentes de recorrência em adultos. De forma interessante, a expressão do mir-9, um regulador negativo do LIN28A/B, foi significativamente maior em carcinomas agressivos quando comparados a não agressivos [2076 (36 a 9307) vs. 133,4 (2,4 a 5193); p = 0,011] e fortemente associada com a redução da sobrevida global (p = 0,01) e livre de doença (p = 0,01). Na população pediátrica, não se observou diferença significativa entre expressão da proteína LIN28, assim como dos genes LIN28A e mir-9, entre tumores clinicamente benignos e malignos. Nas crianças, a hiperexpressão do LIN28B foi significativamente associada com redução da sobrevida livre de doença (p = 0,026), mas não da sobrevida global (p = 0,406). A análise da variação do número de cópias mostrou que somente uma criança com tumor virilizante benigno apresentou amplificação do LIN28B e uma mulher com carcinoma adrenocortical metastático apresentou deleção do LIN28B. Não houve variação no número de cópias para o gene LIN28A. Um índice de Ki67 >= 20% nas crianças foi capaz de discriminar pacientes com pior prognóstico: houve uma associação significativa tanto com diminuição da sobrevida global (p = 0,015) como da sobrevida livre de doença (p = 0,001) em 36 TACs pediátricos com Weiss >- 3. CONCLUSÕES: A imunorreatividade fraca para o LIN28 foi associada à diminuição da sobrevida livre de doença em uma grande coorte de carcinomas adrenocorticais de adultos. O gene LIN28A teve expressão aumentada em carcinomas agressivos de adultos, sugerindo uma regulação pós-transcricional negativa da expressão proteica do LIN28. A hiperexpressão do mir-9, um regulador negativo do LIN28, mostrou-se um importante preditor de desfecho desfavorável nos adultos. Adicionalmente, a hiperexpressão do gene LIN28B mostrou-se um potencial marcador de mau prognóstico na população pediátrica. Um índice de Ki67 >= 10% em adultos e >= 20% em crianças foram associados a mau prognóstico / INTRODUCTION: Adrenocortical carcinoma is a rare neoplasm with overall poor prognosis. Recently, several studies demonstrated the potential of miRNA profiling in differentiating between adrenocortical adenomas and carcinomas, risk stratification and prognosis. Nevertheless, little is known about posttranscriptional regulation of miRNAs. LIN28 is a highly conserved RNA-binding protein that has emerged as a modulator of the processing of let-7, an important family of miRNAs widely known for its tumor-suppressive effects. Besides from let-7, LIN28 has also shown to regulate and be regulated by mir-9, mir-30 and mir-125. OBJECTIVES: To analyze LIN28 gene and protein expression in a large cohort of adult and pediatric adrenocotical tumors (ACTs), and investigate the copy number variation analysis for LIN28A and LIN28B genes and the expression of LIN28 regulatory microRNAs (let-7 family, mir-9, mir-30 e mir-125) in a subgroup of this cohort. METHODS: LIN28 protein expression was assessed in a total of 266 adult (78 adenomas and 188 carcinomas) and 44 pediatric ACTs (35 clinically benign and 9 clinically malignant). LIN28A and LIN28B gene expression was evaluated in a subgroup of 86 adult and pediatric ACTs and copy number variation analysis of these genes in 58 ACTs. The expression of let-7 family, mir-9, mir-30 and mir-125 was performed in 28 adult carcinomas. RESULTS: In adults, LIN28A gene was overexpressed in aggressive carcinomas when compared with adenomas [7.0 fold change (from 0 to 174.3) vs. 3.6 (from 0 to 18.3); p = 0.006, respectively] and a trend towards greaten expression when compared with non-aggressive carcinomas [7.0 (from 0 to 174.3) vs. 7.1 (from 0 to 17.1); p = 0.092]. LIN28B expression was undetectable in the great majority (92%) of adult ACTs. Surprisingly, weak LIN28 staining was significantly associated with reduced disease-free survival in this population (p = 0.01), but for overall survival only a trend was detectable (p= 0.117). In the multivariate analysis, only Ki67 index >- 10% (HR 5.7, 95% CI 3.0-10.8; p = 0,0001) and weak LIN28 staining (HR 2.3, 95% CI 1.2-4.4; p = 0,008) were independent predictors of recurrence in adult patients. Interestingly, mir-9 expression, a negative LIN28A/B regulator, was significantly higher in aggressive than in non-aggressive ACCs [2076 (from 36 to 9307) vs. 133.4 (from 2.4 to 5193); p = 0.011] and was highly associated with reduced overall survival ( p= 0.01) and disease-free survival (p = 0.01). In the pediatric population, no significant difference was observed in the expression of LIN28 protein and LIN28A and mir-9 gene expression between clinically benign and clinically malignant tumors. Additionally. overexpression of LIN28B was significantly associated with reduced disease-free survival (p = 0.026), but not with overall survival (p = 0.406). Copy number variation analysis showed that only a child with a virilizing benign tumor had LIN28B amplification and a woman with a metastatic adrenocortical carcinoma had LIN28B deletion. No LIN28A copy number variation was detected. A Ki67 >= 20% in children was able to discriminate patient with worse prognosis: there was a significant associtation with reduced overall (p = 0,015) and disease-free survival (p = 0,001) in 36 pediatric ACTs with Weiss >- 3. CONCLUSIONS: Weak LIN28 staining was associated with reduced disease-free survival in a large cohort of adult adrenocortical carcinoma. LIN28A had higher expression in aggressive carcinomas in adults, suggesting there might be negative posttranscriptional regulation of LIN28 protein expression. Interestingly, overexpression of mir-9, a negative LIN28A regulator, predicted poor outcome in adult patients. In addition, LIN28B overexpression was an potential marker of poor prognosis in the pediatric population. A Ki67 index >- 10% in adults and >- 20% in children were associated with poor prognosis

Proteínas PUF regulam a estabilidade e a tradução através da ligação a seqüências específicas nas regiões 3\' não traduzidas (3\' UTR) dos mensageiros. A ligação é mediada por um domínio de ligação conservado constituído por 8 repetições de aproximadamente 36 aminoácidos cada. Experimentos realizados no sistema triplo-híbrido de levedura mostraram que os homólogos PUF de Arabidopsis APUM-1, APUM-2 e APUM-3 são capazes de ligar especificamente à seqüência chamada de Elemento de Resposta a NANOS (NRE) reconhecida pelo homólogo PUF de Drosophila. A utilização de bibliotecas de expressão de RNA em ensaios no sistema triplo-híbrido permitiu a identificação de seqüências de ligação consenso para as três proteínas APUM. Análises computacionais identificaram elementos de ligação a APUM em regiões 3\' UTR de importantes transcritos relacionados ao controle do meristema do caule e à manutenção das células totipotentes. Nós mostramos que os homólogos APUM-l, APUM-2 e APUM-3 reconhecem elementos de ligação a APUM nas regiões 3\' UTR dos transcritos WUSCHEL, CLAVATA-1, ZWILLE e FASCIATA-2. Ensaios de RT-PCR e Western blot semiquantitativos mostraram que a quantidade dos transcritos WUSHEL e CLAVATA-1 é alterada em plantas antisenso induzíveis para APUM-l, APUM-2 e APUM-3. A relevância biológica dessas interações foi observada através de ensaios de coimunoprecipitação, confirmando, portanto, o primeiro caso de regulação traducional descrito para os mensageiros WUSCHEL e CLAVATA-1. Análises computacionais adicionais para a identificação de outros homólogos PUF em Arabidopsis encontraram vinte e cinco proteínas possuindo repetições PUF. Entre elas, os homólogos APUM-4, APUM-S e APUM-6 apresentam alta similaridade com as proteínas APUM-l, APUM-2 e APUM-3, sendo capazes de ligar especificamente à seqüência NRE e aos elementos de ligação a APUM presentes nas regiões 3\' UTR dos transcritos WUSCHEL, CLAVATA-1, ZWILLE e FASCIATA-ts resultados indicam que vários homólogos PUF podem agir como reguladores traducionais em Arabidopsis através de um mecanismo molecular conservado entre as espécies, podendo abrir uma nova área de investigação da regulação de mRNA em plantas. / PUF proteins regulate stability and translation through sequence-specific binding to 3\' untranslated regions (UTR) oftarget mRNA transcripts. Binding is mediated by a conserved PUF domain which contains 8 repeats of approximately 36 amino acids each. Through three-hybrid assays, we have found that APUM-1, APUM-2 and APUM-3 Arabidopsis PUF homologs can bind specifically to the NANOS Response Element sequence (NRE) recognized by Drosophila PUF homologo Using Arabidopsis RNA libraries in three-hybrid screenings, we were able to identify APUM binding consensus sequences. A computational analysis allowed us to identify the APUM binding element within the 3\' UTR in many Arabidopsis transcripts, even in important mRNAs related to shoot and stem cell maintenance. We show that APUM-1, APUM-2 and APUM-3 are able to bind specifically to APUM binding elements in the 3\' UTR of WUSCHEL, CLAVATA-1, ZWILLE and FASCIATA-2 transcripts. RT-PCR and Western blot semiquantitatives assays showed altered WUSHEL and CLAVATA-1 amounts in APUM-1, APUM-2 and APUM-3 antisense plants. The biologic relevance of these interactions was observed with co-immunoprecipitation assay, which confirmed the first example of translational regulation to WUSCHEL and CLA VATA-1 transcripts. Computational analysis to identify others PUF homologs in Arabidopsis found twenty five proteins presenting PUF repeats. Among them, we found that APUM-4, APUM-S and APUM-6 homologs are very similar to APUM-1, APUM2 and APUM-3 and also are able to bind specifically to the NRE sequence and to APUM binding elements in the 3\' UTR of WUSCHEL, CLAVATA-1, ZWILLE and FASCIATA-2 transcripts. Our results indicate that the APUM proteins may act as regulators in Arabidopsis through an evolutionarily conserved mechanism, which may open up a new approach to investigate mRNA regulation in plants.

In Chloroplasten durchlaufen primäre Transkripte eine großen Anzahl von bzw. Reifungsprozesse. Diese Ereignisse spielen eine wichtige Rolle bei der Regulation der Genexpression und sind im Wesentlichen durch Proteinfaktoren, insbesondere RNA-Bindeproteine, reguliert. Der plastidäre Spleißfaktor MatK zählt zu den prokaryotischen Gruppe-II-Intron. MatK aus Nicotiana tabacum interagiert mit seinem Heimatintron trnK und sechs weiteren Gruppe IIA Introns. In dieser Untersuchung, MatK-Bindestellen konnten unterschiedlichen Regionen der Gruppe-II-Introns zugewiesen werden mit RIP-seq in Nicotiana tabacum. Die vorliegenden Ergebnisse zeigen, dass MatK im Vergleich zu seinen bakteriellen Vorfahren an Vielseitigkeit in der RNA-Erkennung gewonnen hat. MatK zeigt somit beispielhaft, wie eine Maturase die Fähigkeit erworben haben könnte, in trans auf mehrere Introns zu wirken. Quantitative Untersuchung und mathematische Modellierung der Expression von MatK und dessen Zielen offenbart ein komplexes Muster möglicher regulatorischer Feedback-Mechanismen. In dieser Studie konnte ein möglicher Feedback- Mechanismus durch Analyse von polysomal gebundenen Transkripten ausgeschlossen werden. Stabile Bindung von Proteinen an spezifische RNA-Bindestellen und anschließender Abbau der ungeschützten RNA kann zu Akkumulation von kleinen RNAs (sRNAs) führen. Solche Footprints von RNA-Bindeproteinen wurden durch die Untersuchung von Datensätzen kleiner RNAs in Chlamydomonas reinhardtii identifiziert. Zwei der sRNAs entsprechen den 5'' Enden der reifen psbB und psbH mRNAs. Beide sRNAs sind abhängig von Mbb1, einem TPR (Tetratrico-peptide repeat) Protein. Die beiden sRNAs besitzen eine hohe Ähnlichkeit in ihrer Primärsequenz und fehlen in der mbb1 Mutante. Dies legt nahe, dass auch andere der hier identifizierten sRNAs an 5'' Enden plastidärer mRNAs Protein-Bindestellen repräsentieren, die für die korrekte RNA-Prozessierung und RNA-Stabilisierung in Chlamydomonas Chloroplasten erforderlich sind. / In chloroplasts, primary transcripts are subjected to a number of processing events. These events play important roles in the regulation of gene expression and are extensively controlled by protein factors, especially by RNA-binding proteins. Chloroplast splicing factor MatK is related to prokaryotic group II intron maturases. Nicotiana tabacum MatK interacts with its home intron trnK and six additional group IIA introns. In this study, binding sites of MatK were narrowed down to varying regions of its group II targets by RIP-seq in Nicotiana tabacum. The results obtained demonstrate that MatK has gained versatility in RNA recognition relative to its bacterial ancestors. MatK thus exemplifies how a maturase could have gained the ability to act in trans on multiple introns during the dispersion of the group II introns through the eukaryotic genome early in the eukaryote evolution. Quantitative investigation and mathematical modeling of the expression of MatK and its targets revealed a complex pattern of possible feedback regulatory interactions. In this study, one possible feedback regulation mechanism was ruled out by the analysis of polysome associated transcripts. Stable binding of proteins to specific RNA sites and subsequent degradation of the unprotected RNA regions can result in small RNA, footprint of the RNA binding protein. Such footprints were identified by examining small RNA datasets of Chlamydomonas reinhardtii. Two of the sRNAs correspond to the 5’ ends of mature psbB and psbH mRNAs. Both sRNAs are dependent on Mbb1, a nuclear-encoded TPR (Tetratrico-peptide repeat) protein. The two sRNAs have high similarity in primary sequence, and both are absent in the mbb1 mutant. This suggests that sRNAs at the 5’ ends of chloroplast mRNAs identified here generally represent the binding sites of proteins, which function in RNA processing and RNA stabilization in Chlamydomonas chloroplast.

Plastiden enthalten ihr eigenes Genom, das u.a. für Untereinheiten des photosynthetischen Apparates kodiert. Die Expression dieses Apparates wird hauptsächlich posttranskriptionell reguliert. Dafür notwendige Faktoren sind vor allem RNA-Bindeproteine, welche fast ausschließlich kernkodiert und posttranslational in die Plastiden importiert werden. Dazu gehören auch die äußerst abundanten chloroplastidären Ribonukleoproteine (cpRNPs). Die bisher näher untersuchten Mitglieder der cpRNP-Familie aus Arabidopsis thaliana sind an der Prozessierung und Stabilisierung von plastidären Transkripten beteiligt und phylogenetisch eng miteinander verwandt. In dieser Studie wurden zwei noch unbekannte Mitglieder der cpRNP Familie, CP33A und CP33B, näher untersucht. CP33A ist ein essentielles Protein der Chloroplastenbiogenese. Mutanten von CP33A keimen nur in der Gegenwart einer externen Kohlenstoffquelle. Die Blätter sind albinotisch, in ihrer Struktur anomal und das gesamte Wachstum ist stark eingeschränkt. Untersuchungen der RNA-Interaktionspartner von CP33A durch RIP-Chip-Analysen (RNA-Immunopräzipitation und Chip-Hybridisierung) zeigen, dass CP33A mit allen mRNAs assoziiert. Des Weiteren führt der Verlust von CP33A zu einer starken Reduktion vieler Transkripte, vor allem RNAs, die durch die plastidär kodierte RNA Polymerase (PEP) transkribiert werden und unprozessierte Vorläufer-Transkripte. CP33B interagiert ebenfalls mit multiplen plastidären RNAs. Dabei zeigt CP33B eine Präferenz für psbA. Feinkartierung der CP33B-Bindung innerhalb des psbA Leserahmens verdeutlichten, dass CP33B vor allem mit dem 3´Ende des Transkriptes interagiert. Phänotypische und genetische Untersuchungen der cp33b-Nullmutante ließen keinen vom Wildtyp abweichenden Phänotyp identifizieren und zeigten dass CP33B keinen essentiellen Einfluss auf die Proteinakkumulation photosynthetischer Untereinheiten, die Expression plastidärer Transkripte, das Spleißen und die Edierung seiner Ziel-RNAs hat. / Plastids harbour their own genome, which encodes for essential subunits of the photosynthetic apparatus. The expression of these subunits is mainly regulated on the posttranscriptional level. The important factors for posttranscriptional processing are RNA-binding proteins (RBPs), which are almost exclusively nuclear-encoded and imported posttranslational into the plastids. Among them are the chloroplast ribonucleoproteins (cpRNPs). The cpRNPs are a family of highly abundant RNA-binding proteins found in the chloroplast of land plants. Members of the Arabidopsis thaliana cpRNP family, that have been investigated in more detail, are involved in processing and stabilization of plastid transcripts and are phylogenetically closely related. In this study two unknown members of the cpRNPs, CP33A and CP33B, which cluster outside of this phylogenetic group, are investigated. CP33A is essential for chloroplast biogenesis. Null alleles of CP33A only germinate in the presence of an external carbon source. cp33a seedlings are albino, show strong growth inhibition and an abnormal leaf structure. Investigating RNA-ligands of CP33A using RIP-Chip (coimmunoprecipitation coupled to microarray analysis) shows an association with all chloroplast mRNAs. The loss of CP33A leads to a reduction of almost all transcripts, predominantly affecting RNAs transcribed by the plastid-encoded RNA polymerase (PEP) and unspliced and unprocessed precursor mRNAs. CP33B also interacts with multiple plastid RNAs. The main target is the mRNA of psbA. More than 90% of the stromal psbA mRNA is associated with CP33B. Fine mapping efforts suggest that CP33B preferentially interacts with the 3’-end of the psbA reading frame. Phenotypic and genetic analyses of cp33b-null mutants do not show any differences compared to wild-type plants. CP33B has no essential impact on: Protein accumulation of photosynthetic subunits, expression of plastid transcripts, RNA-splicing or RNA-editing of its target RNAs.

Chez les eucaryotes, plusieurs protéines liant l'ARN ou RBPs agissent sur l'ARNm à différents niveaux, de l'épissage à la traduction. Récemment, un grand nombre d’ARN non-codant des protéines (npcRNAs) ont été identifiés chez les eucaryotes et ont été montré comme interagissant avec une variété de ribonucléoprotéines (RNP) pour contrôler l'expression des gènes au niveau post-transcriptionnel. Nous avons identifié une Nuclear-Speckle RBP (ou NSR) qui interagit avec le npcRNA, ENOD40, un lncARN qui s'accumule au cours de la formation des racines latérales et des nodules chez les légumineuses. Durant cette thèse nous avons analysé le rôle des NSR d’Arabidopsis thaliana ainsi que leur lien avec les npcARN.Deux gènes AtNSRs homologues existent chez Arabidopsis nommés NSRa et NSRb, ces gènes codent des protéines localisées dans des speckles nucléaires avec certaines protéines apparentées à l’épissage. Fait intéressant, les fusions AtNSR-GFP sont relocalisées dans des granules cytoplasmiques dans certaines cellules des racines différenciées ainsi que lors d’une co-expression éctopique de ENOD40. Le gène AtNSRb est régulé par l'auxine alors AtNSRa est constitutif. Les simples mutants Atnsr ne montrent pas de phénotype, mais la croissance des racines des doubles mutants est partiellement insensible à l'auxine, ce qui suggère une fonction redondante de ces protéines dans les racines. La localisation observée pour ces protéines nous a mené à explorer un rôle des NSRs dans l’épissage, nous avons donc analysé le profil d'épissage de 288 gènes en réponse à l'auxine chez Arabidopsis et comparé ces profils entre le WT et les mutants nsra/nsrb. Tout d’abord nous avons remarqué que l’épissage général ne variait pas, en revanche, l’analyse de 288 gènes alternativement épissés montre que le profil d'épissage de 77 gènes semble être modifié durant la réponse à l'auxine et 51 gènes nécessitent les protéines AtNSR pour ce changement. Afin de vérifier l’interaction des NSRs avec les cibles d’AS et avec les npcARN nous avons co-immunoprécipité les NSRs et nous avons identifié au moins 5 cible d’AS et 2 npcARN. L’expression de l’ARN ENOD40 ainsi que du partenaire npcARN module L’AS chez Arabidopsis. Dans un deuxième chapitre, nous avons exploré le rôle des NSRs dans le PTGS déclenché par un transgène contenant un intron ce qui nous a permis de lier l’épissage alternatif et le silencing. Nous proposons donc que les NSRs pourraient lier l’épissage alternatif et l’action des ARN non codants, notamment lors de la croissance de la racine. / In eukaryotes, several RNA binding proteins (RBPs) act on mRNA at various levels from splicing to translation. Recently a large number of non-protein coding RNAs (npcRNAs) have been identified in eukaryotes and shown to integrate into a variety of ribonucleoproteins (RNP) to control posttranscriptional gene expression. Our laboratory has identified a plant Nuclear-Speckle RBP (or NSR) that interacts with an npcRNA, ENOD40 that accumulates during lateral root and nodule formation in legumes. NSR is relocalised into a cytoplasmic RNP in the ENOD40-expressing cells. During this PhD, we have analysed the role of NSRs in Arabidopsis thaliana and its link with npcRNAs. Two AtNSR homologs from Arabidopsis thaliana, named AtNSRa and AtNSRb, code for proteins also localised in nuclear speckles together with certain splicing-related proteins. Interestingly, AtNSR-GFP fusions are relocalised into cytoplasmic granules in certain differentiated root cells and by ectopic expression of the ENOD40 RNA. The AtNSRb gene is regulated by auxin whereas AtNSRa is constitutive. Root growth and lateral root formation of double nsra/nsrb mutants is partially insensitive to auxin. The localisation of these proteins prompted us to explore roles in splicing. No defects in general splicing were observed however analysis of 288 alternatively spliced genes in WT and nsra/nsrb roots in response to auxin revealed 77 changes in splicing profiles in response to auxin from which 51 required AtNSRs. In order to validate the interaction of NSRs with alternatively spliced mRNAs and npcRNAs, we have co-immunoprecipitated NSRs and identified at least 5 interacting alternatively spliced mRNAs and 2 npcRNAs. Expression of the ENOD40 RNA or one interacting ncRNA modulate alternatively splicing in Arabidopsis. In a second chapter, we explored the role of NSRs in the modulation of PTGS triggered by intron-containing transgenes allowing us to link alternatively splicing and silencing. We propose that NSRs may link alternative splicing and the action of non-coding RNA, notably during root growth and development.

Proteínas PUF regulam a estabilidade e a tradução através da ligação a seqüências específicas nas regiões 3\' não traduzidas (3\' UTR) dos mensageiros. A ligação é mediada por um domínio de ligação conservado constituído por 8 repetições de aproximadamente 36 aminoácidos cada. Experimentos realizados no sistema triplo-híbrido de levedura mostraram que os homólogos PUF de Arabidopsis APUM-1, APUM-2 e APUM-3 são capazes de ligar especificamente à seqüência chamada de Elemento de Resposta a NANOS (NRE) reconhecida pelo homólogo PUF de Drosophila. A utilização de bibliotecas de expressão de RNA em ensaios no sistema triplo-híbrido permitiu a identificação de seqüências de ligação consenso para as três proteínas APUM. Análises computacionais identificaram elementos de ligação a APUM em regiões 3\' UTR de importantes transcritos relacionados ao controle do meristema do caule e à manutenção das células totipotentes. Nós mostramos que os homólogos APUM-l, APUM-2 e APUM-3 reconhecem elementos de ligação a APUM nas regiões 3\' UTR dos transcritos WUSCHEL, CLAVATA-1, ZWILLE e FASCIATA-2. Ensaios de RT-PCR e Western blot semiquantitativos mostraram que a quantidade dos transcritos WUSHEL e CLAVATA-1 é alterada em plantas antisenso induzíveis para APUM-l, APUM-2 e APUM-3. A relevância biológica dessas interações foi observada através de ensaios de coimunoprecipitação, confirmando, portanto, o primeiro caso de regulação traducional descrito para os mensageiros WUSCHEL e CLAVATA-1. Análises computacionais adicionais para a identificação de outros homólogos PUF em Arabidopsis encontraram vinte e cinco proteínas possuindo repetições PUF. Entre elas, os homólogos APUM-4, APUM-S e APUM-6 apresentam alta similaridade com as proteínas APUM-l, APUM-2 e APUM-3, sendo capazes de ligar especificamente à seqüência NRE e aos elementos de ligação a APUM presentes nas regiões 3\' UTR dos transcritos WUSCHEL, CLAVATA-1, ZWILLE e FASCIATA-ts resultados indicam que vários homólogos PUF podem agir como reguladores traducionais em Arabidopsis através de um mecanismo molecular conservado entre as espécies, podendo abrir uma nova área de investigação da regulação de mRNA em plantas. / PUF proteins regulate stability and translation through sequence-specific binding to 3\' untranslated regions (UTR) oftarget mRNA transcripts. Binding is mediated by a conserved PUF domain which contains 8 repeats of approximately 36 amino acids each. Through three-hybrid assays, we have found that APUM-1, APUM-2 and APUM-3 Arabidopsis PUF homologs can bind specifically to the NANOS Response Element sequence (NRE) recognized by Drosophila PUF homologo Using Arabidopsis RNA libraries in three-hybrid screenings, we were able to identify APUM binding consensus sequences. A computational analysis allowed us to identify the APUM binding element within the 3\' UTR in many Arabidopsis transcripts, even in important mRNAs related to shoot and stem cell maintenance. We show that APUM-1, APUM-2 and APUM-3 are able to bind specifically to APUM binding elements in the 3\' UTR of WUSCHEL, CLAVATA-1, ZWILLE and FASCIATA-2 transcripts. RT-PCR and Western blot semiquantitatives assays showed altered WUSHEL and CLAVATA-1 amounts in APUM-1, APUM-2 and APUM-3 antisense plants. The biologic relevance of these interactions was observed with co-immunoprecipitation assay, which confirmed the first example of translational regulation to WUSCHEL and CLA VATA-1 transcripts. Computational analysis to identify others PUF homologs in Arabidopsis found twenty five proteins presenting PUF repeats. Among them, we found that APUM-4, APUM-S and APUM-6 homologs are very similar to APUM-1, APUM2 and APUM-3 and also are able to bind specifically to the NRE sequence and to APUM binding elements in the 3\' UTR of WUSCHEL, CLAVATA-1, ZWILLE and FASCIATA-2 transcripts. Our results indicate that the APUM proteins may act as regulators in Arabidopsis through an evolutionarily conserved mechanism, which may open up a new approach to investigate mRNA regulation in plants.

Le glioblastome (GBM) est la tumeur primitive du cerveau la plus fréquente et la plus agressive chez l’adulte. Le mauvais pronostic de cette pathologie peut être expliqué par la présence de cellules résistantes aux traitements à l’origine des rechutes appelées cellules souches de glioblastome (CSG). Caractérisées par une plasticité cellulaire, elles sont capables de s’adapter aux environnements défavorables à leur survie. Ainsi, malgré des traitements multimodaux agressifs, les bénéfices de la prise en charge du GBM restent très modestes ; il est donc nécessaire de mieux caractériser ces cellules afin de développer de nouvelles thérapies ciblant les CSG. Le rôle des mécanismes post-transcriptionnels de régulation de l’expression génique dans le maintien des cellules souches cancéreuses a été démontré dans différentes tumeurs. Cependant, peu de protéines de liaison à l'ARN (RBP), régulateurs clés de ces événements, ont été identifiés dans le contexte des CSG. Mon projet de thèse s’inscrit dans le cadre de l’étude de RBP régulant les propriétés d’auto-renouvèlement et de survie des CSG afin d’approfondir nos connaissances sur les mécanismes moléculaires qui contribuent à la formation ou récurrence des GBM. Dans un premier temps, j’ai cherché à définir un protocole permettant d’enrichir la population de CSG maintenues en conditions de culture non-adhérentes qui favorisent la formation de neurosphères et le maintien d’une hiérarchie cellulaire. Confrontée à une forte hétérogénéité de signatures moléculaires, j’ai choisi de prendre cette caractéristique en compte dans un second temps en menant une étude comparative du transcriptome de 5 modèles de cellules souches de glioblastome provenant de patients différents. Cette analyse inclut également des cellules différenciées in vitro à partir des CSG ainsi que des cellules souches neurales humaines. Grâce à une approche de séquençage de leur transcriptome, mes travaux de thèse ont permis d’identifier une famille de régulateurs post-transcriptionnels enrichis dans les CSG, les protéines nELAVL. Ces résultats ont pu être confirmés par l’analyse de leur expression protéique dans des modèles in vitro et dans des coupes de tumeurs provenant de différents patients atteints de GBM. Une co-expression des protéines nELAVL avec OLIG2 ou SOX2 a été observée confirmant ainsi leur association avec l’état souche. ELAVL4 correspond au membre de la famille nELAVL le plus réprimé lors de la différenciation des CSG. Des outils de modulation de son expression ont été développé en vue d’évaluer son rôle dans les CSG par des approches de perte d’expression ou de gain de fonction / Glioblastoma (GBM) is the most common and aggressive primary adult brain tumor. GBM dismal prognosis can be explained by the presence of treatment resistant cells responsible for tumor relapse known as glioblastoma stem cells (GSC). Characterized by cellular plasticity, they are able to adapt to hostile environments. Thus, despite aggressive multimodal treatments, GBM curative therapies have provided only a modest benefit; it is therefore necessary to better characterize these cells in order to develop new GSC targeted therapies. The role of posttranscriptional mechanisms of gene expression regulation in the maintenance of cancer stem cells has been demonstrated in different tumors. However, few RNA-binding proteins (RBPs), key regulators of these events, have been identified in GSC. My thesis project consists in identifying RBP that are critical for self-renewal and increased survival properties of GSC. The aim of this study is to deepen our knowledge of the underlying molecular mechanisms of GBM development or recurrence. At first, I established a protocol to enrich for GSC using nonadherent culture conditions that promote the formation of neurospheres comprising a cellular hierarchy. I chose to take into account the facing problem of GSC molecular heterogeneity in a second step and carried out a comparative study of the transcriptome of 5 glioblastoma stem cell cultures from different patients. This analysis also includes in vitro differentiated cells originating from GSC as well as human neural stem cells. Using a transcriptome sequencing approach, my thesis work has led to the identification of a family of post-transcriptional regulators enriched in GSC, nELAVL proteins. These results were confirmed by protein expression analysis in in vitro neurospheres and within tumor sections from different GBM patients. Co-expression of nELAVL proteins with OLIG2 or SOX2 was observed thus confirming their association with stemness. ELAVL4 repesents the most differentially expressed member of nELAVL family upon GSC differentiation. Knock-down and gain-offunction tools targeting ELAVL4 have been developed to further assess its roles in GSC maintenance

<p><i>Trypanosoma cruzi</i>, the causative agent of Chagas´ disease, is an evolutionarily ancient species with distinct biological and immunological characteristics. A fundamental understanding of the basic biology of the parasite is necessary in order to develop reliable therapeutic and prophylactic agents against <i>T. cruzi</i>. We have, as a part of the <i>T. cruzi</i> genome project launched by the WHO, generated ESTs corresponding to about one third of the functional genes in the parasite. Only about 1/3 of the unique ESTs could be assigned a function upon sequence comparison to all publicly available data. Comparative analysis of the ESTs to functional genes in <i>S.</i> <i>cerevisiae</i> and <i>C. elegans</i> as well as to sequence data from all other kinetoplastids provided primary insights into the evolutionary divergence of <i>T. cruzi.</i> </p><p>A novel dispersed gene family (<i>DGC3</i>) was identified and shown to be present specifically on chromosome 3 and its homologue. Sequence analysis of ten isolated <i>DGC3</i> genes revealed a high sequence similarity of almost 98% among copies. The <i>DGC3</i> genes were transcribed, <i>trans</i>-spliced with the spliced leader and polyadenylated, but did not seem to have any protein-coding property. These data preliminary suggest that it encodes a novel family of functional RNA. </p><p>In the <i>T. cruzi</i> CL Brener strain, the two alleles of a single copy gene encoding the trypanothione synthetase (TcTRS) enzyme appeared to be highly polymorphic. The divergence of the deduced protein sequence was 4%, almost ten-fold higher than another protein, trypanothione reductase, involved in the same pathway. The observed allelic divergence might influence the TcTRS activity thereby having implications for drug design. Moreover, the <i>TcTRS</i> gene was found to be flanked by a number of genes involved in diverse functions and located to a pair of homologous chromosomes with a size difference of about 2 Mbp. </p><p>A gene potentially encoding the polypyrimidine-binding protein (TcPTB) was identified and characterised regarding its organisation and function. The deduced amino acid sequence was shown to comprise four RRM domains generally present in other PTBs. Interestingly, the <i>TcPTB</i> gene appeared to be expressed in a stage-specific manner implicating different functions during parasite development.</p>

Trypanosoma cruzi, the causative agent of Chagas´ disease, is an evolutionarily ancient species with distinct biological and immunological characteristics. A fundamental understanding of the basic biology of the parasite is necessary in order to develop reliable therapeutic and prophylactic agents against T. cruzi. We have, as a part of the T. cruzi genome project launched by the WHO, generated ESTs corresponding to about one third of the functional genes in the parasite. Only about 1/3 of the unique ESTs could be assigned a function upon sequence comparison to all publicly available data. Comparative analysis of the ESTs to functional genes in S. cerevisiae and C. elegans as well as to sequence data from all other kinetoplastids provided primary insights into the evolutionary divergence of T. cruzi. A novel dispersed gene family (DGC3) was identified and shown to be present specifically on chromosome 3 and its homologue. Sequence analysis of ten isolated DGC3 genes revealed a high sequence similarity of almost 98% among copies. The DGC3 genes were transcribed, trans-spliced with the spliced leader and polyadenylated, but did not seem to have any protein-coding property. These data preliminary suggest that it encodes a novel family of functional RNA. In the T. cruzi CL Brener strain, the two alleles of a single copy gene encoding the trypanothione synthetase (TcTRS) enzyme appeared to be highly polymorphic. The divergence of the deduced protein sequence was 4%, almost ten-fold higher than another protein, trypanothione reductase, involved in the same pathway. The observed allelic divergence might influence the TcTRS activity thereby having implications for drug design. Moreover, the TcTRS gene was found to be flanked by a number of genes involved in diverse functions and located to a pair of homologous chromosomes with a size difference of about 2 Mbp. A gene potentially encoding the polypyrimidine-binding protein (TcPTB) was identified and characterised regarding its organisation and function. The deduced amino acid sequence was shown to comprise four RRM domains generally present in other PTBs. Interestingly, the TcPTB gene appeared to be expressed in a stage-specific manner implicating different functions during parasite development.

Mycobacterium tuberculosis has one ribosomal RNA operon. The survival of this bacillus thus depends on a transcription mechanism that can effectively couple gene expression to changes in the environment. σ factors are transcription proteins that bind to the RNA polymerase (RNAP) and dictate gene expression. Extra Cytoplasmic Function σ factors (ECF) are a subset of σ factors that coordinate environment-induced changes in transcription. The environment specific binding of ECF σ factors to the RNAP presents an effective mechanism for the bacillus to modulate gene expression. ECF σ factors, in turn, are regulated by their interaction with an anti-σ factor. The active σ factor is released from this complex upon specific cellular or environmental stimuli. The aim of this study was to understand the structural and mechanistic aspects of σ factor activation. Towards this goal, two ECF σ factors, σC and σL, were examined. Structural and biophysical studies on M. tuberculosis σC provided a novel insight into ECF σ factor regulation. Inter-domain interactions in σC were sufficient to occlude the DNA recognition regions even in the absence of an interacting protein. The structure of M. tuberculosis σL in complex with the anti-σ factor RslA provides a structural basis to rationalize the release of active σL under oxidative stress. The other chapters of this thesis include a description of the structure and biochemical features of a hypothetical protein Rv2704 that is co-transcribed with the primary σ factor σA. In an effort to understand the collaboration-competition-redundancy model of prokaryotic σ factors, we performed a computational analysis of this system compiling experimental data from the E. coli and B. subtilis model systems. These results are also presented in this thesis. Put together, the structural and biochemical characteristics of the σ factors presented in this thesis suggest substantial variations in the regulatory mechanisms of the M. tuberculosis σ factors when compared to the canonical E. coli or B. subtilis model systems. This thesis is organized as follows: Chapter 1: The introductory chapter of this thesis is organized to frame the pertinent mechanistic issues involved in the σ factor-regulatory protein interactions in the context of the underlying biology of M. tuberculosis. The first part of this chapter provides an overview of σ factors and a summary of the classification of these proteins and their roles in different prokaryotes. The latter part of this chapter is a summary of the pathogen M. tuberculosis in terms of its genetic composition, gene expression as well as aspects of virulence and pathogenecity. Chapter 2: This chapter describes the characterization of the ECF σ factor, σC. Here we report the structure of an ECF σ factor σC from M. tuberculosis. σC is essential for the lethality of M. tuberculosis in a mouse model of infection. Our studies suggest that M. tuberculosis σC differs from the canonical ECF σ factors as it has an N-terminal domain comprising of 126 amino acids that precedes the σC2 and σC4 domains. In an effort to understand the regulatory mechanism of this protein, the crystal structures of the σC2 and C4 domains of σC were determined. These promoter recognition domains are structurally similar to the corresponding domains of E. coli σA despite the low sequence similarity. Fluorescence experiments using the intrinsic tryptophan residues of σC2 as well as surface plasmon resonance measurements reveal that the σC2 and σC4 domains interact with each other. Mutational analysis suggests that the Pribnow box-binding region of σC2 is involved in this inter-domain interaction. Interactions between the promoter recognition domains in M. tuberculosis σC are thus likely to regulate the activity of this protein even in the absence of an anti-σ factor. Chapter 3 provides an account of the regulatory features of the ECF σ factor, σL. ECF σ factors are often regulated by their interactions with an anti-σ factor that can sense diverse environmental stimuli. Transcriptional responses to changes in the oxidation state are particularly important for M. tuberculosis as it adapts to the environment of the host alveoli and macrophages. Here we demonstrate that the protein RslA binds Zinc and can sequester σL in a reducing environment. Our data suggests that the cytosolic domain at the N-terminus of RslA alone is involved in binding σL. Under oxidizing conditions, the σL/RslA complex undergoes substantial conformational rearrangements that coincide with the release of the Zinc cofactor. In the absence of Zinc, the affinity of RslA for σL reduces by ca 8 fold compared to the holo form. The CXXC motif of RslA acts as a redox sensor. In response to oxidative stimuli, the proximal cysteines in this motif can form a disulfide bond with the release of the bound Zn2+ ion. This observation could be rationalized based on the crystal structure of the σL4/RslA complex. Put together, RslA is a distinct variant of the Zinc binding anti-σ factor (ZAS) family. The structural and biophysical parameters that control σL/RslA interactions demonstrate how variations in the rate of Zinc release and associated conformational changes in RslA could regulate the release of free σL in a measured response to oxidative stress. Chapter 4 is based on the biochemical and structural characterization of a hypothetical protein Rv2704. The gene for M. tuberculosis Rv2704 is located in the same operon as the principal σ factor σA. The biochemical and structural features of Rv2704 were thus examined to identify its role, if any, in the regulation of σA. This protein is a trimer in solution and adopts a chorismate mutase-like fold. The crystal structure reveals that Rv2704 is a member of the functionally diverse YjgF family of proteins. The important structural differences between Rv2704 and other YjgF proteins lie in the arrangement of secondary structural elements and the putative functional clefts between the subunit interface. Although Rv2704 does not interact with σA in vitro, the structural similarities to the YjgF family suggests that this protein could interact with a variety of metabolites, potentially influencing its function. Chapter 5 of this thesis is based on a computational analysis of σ factors. Four conformational segments of σ factors, referred to as σ1, σ2, σ3 and σ4 interact with specific regions of promoter DNA. ECF σ factors are a subset of σ factors that coordinate environment-induced transcription. ECF σ factors are minimalist σ factors with two DNA binding domains viz., σ2 and σ4 that recognize the –10 and –35 promoter elements and are unable to interact with either upstream-activating regions or the extended –10 element of the promoter. There are several ECF σ factors in a typical bacterium often characterized by substantial overlap in function. Here we present an analysis of B. subtilis ECF σ factors and their cognate promoters to understand functional overlap and redundancy in this class of proteins. As expected, conserved bases in the –10 element appear more critical for promoter selectivity than the –35 element. However, we note distinct conformational features in the –35 promoter interaction with the helix-turn-helix (HTH) motif when compared to a data-set of known HTH-DNA complexes. Furthermore, we note differences in –35 element interaction between σ factors that act alone and those that overlap in function. The σ factor promoter interactions were then examined vis-à-vis the estimated cellular concentration of these proteins and their affinity to bind the core RNAP. Put together, this analysis suggests that while the cellular protein concentration dictates the choice of an ECF σ factor to form a complex with the RNAP, conformational features of the –35 element serve to select potential collaborative members, a subset of which eventually initiate transcription. Collaborative arrangements and functional redundancy in ECF σ factors are thus possible within the limits placed by these two parameters. Chapter 6 is a summary of the work reported in this thesis and the conclusions that can be drawn based on these studies. The appendix section of this thesis comprises of technical details that were not included in the main text of this thesis. Appendix I describes the initial characterization of the M. tuberculosis σD/anti-σD complex. Appendix II provides the experimental protocols as well as some of the supplementary data to the work reported in Chapters 2-5 of this thesis.