Chez les cellules eucaryotes, la transcription est réalisée par les ARN polymérases I, II et III. L’ARN polymérase III (Pol III) transcrit des petits ARNs non codants tels que les ARNt, l’ARN U6, l’ARNr 5S et certains microARN. Il a été montré précédemment que l’augmentation de l’activité transcriptionnelle de la Pol III était associée à la transformation tumorale. Au sein du laboratoire, nous avons décrit une nouvelle sous–unité de la Pol III, RPC32α, qui met en évidence l’existence de deux isoformes de la Pol III humaine : la Pol IIIα et la Pol IIIβ. La sous-unité RPC32β, présente dans la Pol III, est exprimée de façon ubiquitaire et parait comme essentielle pour la croissance cellulaire. La sous-unité RPC32α n’est pas essentielle pour la survie cellulaire et son expression est limitée aux cellules souches non différenciées et aux cellules tumorales. De plus, l’expression ectopique de RPC32α induit la transformation tumorale des cellules fibroblastes IMR90 et change totalement l’expression de nombreux transcrits Pol III mais aussi d’autres ARNm impliqués dans la tumorogenèse (Haurie et al., 2010). Les travaux décrits dans ce manuscrit ont eu pour but de mieux caractériser et comprendre les deux isoformes de la Pol III humaine. Nous avons purifié les Pol IIIα et Pol IIIβ afin d’identifier tous les composants protéiques des deux isoformes du complexe. Au cours de cette étude nous avons observé que des modifications post-traductionnelles de RPC32α semblaient jouer un rôle déterminant dans la capacité oncogénique de la Pol IIIα. Pour comprendre par quel mécanisme moléculaire les Pol IIIα et Pol III pouvaient influencer l’expression de transcrits Pol II, notamment lors de la transformation cellulaire induite par l’expression de RPC32α, nous avons étudié cette régulation lors de la surexpression des deux sous unités paralogues. Ainsi, nous avons mené des analyses ciblées révélant la régulation de certains gènes impliqués dans le développement, la différenciation et la tumorogenèse. Nous avons également essayé de décrire les changements d'expression des gènes au niveau globale en utilisant la technologie des puces à ADN. Cette approche innovante et puissante nous a permis d’obtenir une vision d’ensemble des transcrits Pol II différentiellement régulés lors de la surexpression de RPC32α et RPC32β Nous avons pu apprécier l’impact de la Pol III pour le développement embryonnaire, la différenciation cellulaire, la survie de la cellule, la prolifération tumorale ou encore à la réponse immunitaire innée. Les résultats de cette étude qui demandent à être confirmés ouvrent des voies de recherches particulièrement intéressantes qui mériteront d’être approfondies dans le futur / Transcription in eukaryotic nuclei is carried out by DNA-dependent RNA polymerases I, II, and III. Human RNA polymerase III (Pol III) transcribes small untranslated RNAs that include tRNAs, 5S RNA, U6 RNA, and some microRNAs. Increased Pol III transcription has been reported to accompany or cause cell transformation. In the laboratory, we described a Pol III subunit (RPC32β) that led to the demonstration of two human Pol III isoforms (Pol IIIα and Pol IIIβ). RPC32β-containing Pol IIIβ is ubiquitously expressed and essential for growth of human cells. RPC32α-containing Pol IIIα is dispensable for cell survival, with expression being restricted to undifferentiated ES cells and to tumor cells. In this regard, and most importantly, ectopic expression of RPC32α in fibroblast IMR90 enhances cell transformation and dramatically changes the expression of several tumor-related mRNAs and that of a subset of Pol III RNAs. These results identify a human Pol III isoform and isoform-specific functions in the regulation of cell growth, the differentiation and transformation. (Haurie et al., 2010).The work described in this manuscript enables identification and understanding the function of the two human Pol III isoforms. Pol IIIα and Pol IIIβ are purified in order to describe all the protein components of the two Pol III complex. During this study, we observed that post-translated modifications of RPC32α seem to have a crucial function in oncogenic capacity of Pol IIIα. To understand how Pol IIIα and Pol IIIβ can affect Pol II RNA expression, in particular during cell transformation induced by RPC32α, we studied this regulation during the overexpression of the two paralogue subunits. We performed focused analysis, this study revealed the regulation of certain genes involved in the development, the differentiation and the tumorigenesis. We also tried to describe global gene expression modification using microarray technology. This new and powerful approach enables to obtain a global view on mRNA regulation by overexpression of RPC32α and RPC32β. We observed the effect of Pol III on embryo development, cell differentiation, cell survival, tumor proliferation and on innate immune response. The results of this study need further confirmation pave the way for interesting projects which are worth going into detail for the future.

Les ARN polymérases sont des acteurs indispensables de la transcription. Chez les eucaryotes il existe trois ARN polymérases (I, II et III). La ARN polymérase III (Pol III) possède 17 sous-unités, dont une qui existe sous deux formes: RPC32α et RPC32β. Seulement une des deux formes peut être intégrée dans la Pol III, créant ainsi deux polymérases différentes Pol IIIα et Pol IIIβ. Alors que RPC32β est présent dans les cellules somatiques, RPC32α est exprimé surtout dans des cellules souches et des cellules tumorales. Aujourd’hui rien n’est connu sur leurs rôles respectifs. Le cancer du sein est un problème majeur de santé publique car c’est le cancer féminin le plus fréquent. Plusieurs types de cancer du sein sont identifiés selon la présence ou absence de certains récepteurs hormonaux. Des cancers qui testent négative pour le récepteur d’oestrogène et de progestérone et qui ne surexpriment pas le récepteur pour les facteurs de croissance épidermiques humains 2 (HER2) sont appelés triple-négative. Ils ont un pronostique peu favorable, due à l’agressivité de ce type de cancer et un manque de thérapie cibles. Pour étudier le rôle de RPC32α il fallait identifier un model tumorale. En collaboration avec Jean-Paul Feugeas (INSERM UMR 1098) une étude transcriptomique a été fait sur 2627 échantillons cliniques de tissus de sein. L’étude montre que RPC32α est surexprimé dans les cancers triple-négative, alors que son homologue RPC32β est surexprimé dans les tissues normaux. Une analyse sur six lignées de cancer du sein et une ligné non-tumorale ont pu confirmer les résultats de l’analyse transcriptomique. Le modèle de cancer du sein a donc été validé. Une caractérisation des différentes lignées de cancer du sein a démontré que d’autres sousunités de la Pol III n’étaient pas surexprimées dans les cancers triple-négative. La surexpression de RPC32α n’était donc pas une conséquence d’une hyperactivité de la Pol III. Une analyse des transcrits synthétisé par la Pol III a montré que en générale les transcrits de la Pol III étaient plus fortement exprimé dans les cancers triple-négative que dans d’autres cancers. Afin d’étudier l’implication de RPC32α dans les phénomènes de tumorisation, plusieurs lignées cellulaires dépourvues de RPC32alpha ont été créé utilisant la technique CRISPRCAS9. L’absence de RPC32α n’a pas induit une augmentation de transcription ni de l’ARN de 4 RPC32α, ni de celle de RPC32β. Il n’existe donc pas de boucle de rétroaction pour RPC32α et les deux homologues ne sont pas co-régulés. Plusieurs, mais pas tous les transcrits synthétisé par la Pol III ont une expression fortement baissé dans les lignées mutants. Le fait que pas tous les transcrits ne soit affectés par la perte de RPC32α, indique qu’il existe une spécificité de transcription pour Pol IIIα et Pol IIIβ. Les cellules des linges mutants ne présentaient pas de phénotype différent des cellules mères et la croissance était la même dans toutes les lignées. Par contre les tests de croissance en agar-mou ont révélé que les lignées mutants formaient 85% de moins de colonies, indiquant que RPC32α est nécessaire pour la croissance tumorigénique in vitro. Pour tester l’effet de la perte de RPC32α sur la croissance tumorigénique in vivo, des cellules mutants et des cellules mères ont été injecté dans des souris. Les souris greffées avec des cellules mutantes montrent un départ de tumorisation retardé. Au bout de six semaines elles avaient de tumeurs deux fois plus petit que les souris avec des cellules mères. Après ablation de la tumeur primaire, les souris ont été surveillées pour l’apparition de métastases. Quatre semaines plus tard les souris greffées avec des cellules mutantes avaient 100 fois moins de métastases que les souris contrôles. Ces résultats montrent que RPC32α est nécessaire pour la tumorisation in vitro et in vivo. La protéine semble surtout jouer un rôle dans la formation des métastases, qui sont un des problèmes majeurs dans le traitement des cancers. / The RNA polymerases are key players of transcription. Eukaryotes have three RNA polymerases (I, II and III). The RNA polymerase III (Pol III) has 17 subunits, one of which exists in two alternative forms: RPC32α and RPC32β. Only one of the two forms can be integrated into the enzymes, thus generating either Pol IIIα or Pol IIIβ. While RPC32β is found in all somatic cells, RPC32α is expressed in stem cells and tumor cells. To date nothing is known of their respective roles. Breast cancer is one of the major public health problems, as it is the most common cancer in women. Several types of breast cancers are distinguished, according to the presence or absence of hormonal receptors. Cancers that test negative for estrogen receptors, progesterone receptors and that do not overexpress the human epidermal growth factor receptor 2, are called triple-negative breast cancers. They tend to have a poor prognosis, due to the aggressive nature of the cancer and the lack of targeted therapies. To study the role of RPC32α, a tumor model needed to be identified. In collaboration with Jean-Paul Feugeas (INSERM UMR 1098) a transcriptomic study was performed on 2627 clinical breast tissue samples. The study showed that RPC32α was overexpressed in triplenegative breast cancer, whereas RPC32β was overexpressed in normal tissue. A study on six breast cancer cell lines and one non-tumorigenic line confirmed the results of the transcriptomic study. The breast cancer model was thus validated. A characterization of different breast cancer cell lines showed that other Pol III subunits were not overexpressed in triple-negative breast cancer. The overexpression of RPC32α was therefore not a mere consequence of a Pol III hyperactivity. An analysis of the transcripts synthesized by Pol III showed that overall the Pol III transcript levels were elevated in triplenegative breast cancer compared to other breast cancer subtypes. In order to study the role of RPC32α in tumorigenesis, several RPC32α knock-out cell lines were created using CRISPR-Cas9. The loss of RPC32α did not induce an increase in transcription of the RNAs of RPC32α or RPC32β. This shows that no feed-back loop exists for RPC32α and that the two homologues are not co-regulated. Various Pol III transcripts showed decreased expression levels in the knock-out cell lines. Yet not all transcripts were reduced in the absence of RPC32α. This indicates that some sort of transcription specificity must exist for Pol IIIα and Pol IIIβ. The knock-out cell lines did not show any alterations in their phenotype or growth rates. However, in soft agar assays the knock-out cell lines produced 85% less colonies than the mother cell line. This proves that RPC32α is necessary for tumorigenic growth in vitro. To find out if RPC32α was also necessary for tumorigenic growth in vivo, knock-out and wild type cells were injected into mice. The mice grafted with knock-out cells showed a slowed onset of tumor growth. After six weeks, the mice injected with knock-out cells had tumors half the size of the mice injected with wild type cells. The primary tumor was ablated and mice were tracked for metastasis. Four weeks later, mice injected with RPC32α knock-out cells had 100 times less metastasis than the control group. These results show that RPC32α is necessary for tumorigenic growth in vitro and in vivo. The protein seems also to be implicated in the formation of metastasis, which are one of the greatest problems in cancer treatment today.

Orientador: Celso Eduardo Benedetti / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Biologia / Made available in DSpace on 2018-08-21T05:36:49Z (GMT). No. of bitstreams: 1 Soprano_AdrianaSantos_D.pdf: 24018757 bytes, checksum: 29724fbb81a588e3bb656bf4c2df3390 (MD5) Previous issue date: 2012 / Resumo: O cancro cítrico, causado pela bactéria Xanthomonas citri (X. citri), afeta a maioria das espécies de Citrus, ocorre praticamente em todos continentes e se destaca como uma séria ameaça à citricultura brasileira. O mecanismo molecular pelo qual X. citri causa cancro não é inteiramente conhecido, entretanto, sabe-se que a bactéria utiliza o sistema secretório tipo III para injetar proteínas de patogenicidade, entre elas, PthAs da família AvrBs3/PthA, também conhecidas como efetores TAL (transcriptional activator-like). Os efetores TAL atuam como fatores de transcrição transativando genes específicos da planta que vão beneficiar a bactéria ou desencadear respostas de defesa. Com o objetivo de entender os mecanismos moleculares pelos quais os efetores TAL atuam, a técnica de duplo híbrido foi usada para identificar proteínas de laranja doce (Citrus sinensis) que interagem com PthA4, um dos efetores TAL de X. citri necessário para o desenvolvimento do cancro cítrico. A maioria das proteínas de laranja identificadas como alvos de PthA4 apresenta domínios de ligação à DNA ou RNA e está envolvida no controle da transcrição, estabilização de mRNAs e tradução. Várias dessas proteínas interagem entre si, sugerindo a presença de um complexo multiproteico como alvo de efetores TAL. Entre as proteínas envolvidas no controle da transcrição, destacamos a CsMAF1, uma proteína homóloga à MAF1 humana que atua como regulador negativo da RNA Polimerase III. Os resultados obtidos nesse trabalho revelam que CsMAF1 complementa o fenótipo do mutante maf1 de levedura, reprimindo a expressão de tRNAHis e que a expressão de PthA4 na cepa complementada restaura a síntese desse tRNA. Portanto, os dados mostram que CsMAF1 atua como um repressor da RNA Pol III em levedura e que PthA4 altera o estado repressor de CsMAF1 sobre a RNA Pol III. De forma surpreendente, verificamos que plantas de citros com níveis reduzidos de CsMAF1 apresentaram aumento significativo no número e intensidade de lesões hiperplásticas ou eruptivas quando infiltradas com X. citri, indicando que CsMAF1 desempenha um papel crítico no desenvolvimento dos sintomas do cancro cítrico. O aumento das lesões do cancro nas plantas silenciadas para CsMAF1 se correlaciona com um aumento expressivo de tRNAs, incluindo o tRNAHis, confirmando assim o papel repressor de CsMAF1 sobre a RNA Pol III em citros. Além disso, mostramos nesse trabalho que CsMAF1 é uma fosfoproteína que se encontra na forma dimérica em solução, uma característica singular ainda não descrita para membros dessa família de proteínas. Verificamos que CsMAF1 é fosforilada in vitro pelas quinases PKA e PKC e que apresenta sítios adicionais de fosforilação conservados para a quinase TOR, incluindo o resíduo Thr62. Curiosamente, tais sítios se localizam na interface de dimerização de CsMAF1, sugerindo que a fosforilação desses sítios deve regular a função da proteína e/ou seu estado multimérico. De fato, verificamos que a substituição do resíduo de treonina Thr 62 para ácido aspártico (Asp 62) diminui a proporção dímero:monômero de CsMAF1, indicando que a fosforilação de resíduos na interface do dímero desestabiliza o dímero, e que esse pode ser um mecanismo regulatório novo para essa classe de proteína. Desse modo, esses achados abrem novas perspectivas para o entendimento não só dos mecanismos moleculares envolvidos na regulação da RNA Pol III pela CsMAF1, como também do papel de PthA4 na interação com CsMAF1 e sua modulação da transcrição / Abstract: Citrus canker, caused by Xanthomonas citri (X. citri), is a disease that affects most of the Citrus species, occurs in almost all continents and stands as a threat to the Brazilian citrus industry. The molecular mechanism by which X. citri causes canker is poorly understood, however the bacterium injects pathogenicity proteins via the type III secretion system (T3S) including proteins of AvrBs3/PthA family, also known as transcriptional activator-like (TAL) effectors. TAL effectors have been extensively studied and are known to act as transcription factors that transactivate specific plant genes which either benefit the bacteria or trigger defense responses. To gain insights into the molecular mode of action of TAL effectors, a twohybrid screening was performed to identify sweet orange (Citrus sinensis) proteins that interact with PthA4, one of the X. citri TAL effectors required for citrus canker development. Among the proteins identified as PthA4 interactors, most are DNA and/or RNA-binding factors involved in chromatin remodeling and repair, transcriptional control and mRNA stabilization/modification. Several of these proteins interact with each other, suggesting the presence of a multiprotein complex as a target of TAL effectors. Among the proteins involved in transcription control, we selected for further studies the CsMAF1, a homolog of the human MAF1 that acts as a negative regulator of RNA polymerase III. The results presented here reveal that CsMAF1 complements the yeast maf1 mutant phenotype by repressing the tRNAHis transcription, and that PthA4 expression in the complemented strain restores the tRNAHis synthesis. Thus, the data show that CsMAF1 acts as a RNA Pol III repressor in yeast and that PthA4 somehow suppresses the repressor activity of CsMAF1 upon on the RNA Pol III. Surprisingly, we found that citrus plants with reduced levels of CsMAF1 showed a significant increase in the number, morphology and size of eruptive or hyperplastic lesions when infiltrated with X. citri, indicating the CsMAF1 plays a critical role in canker development. Increased canker lesions in CsMAF1 silenced plants correlated with a significant increase of tRNAs expression, including tRNAHis, thus confirming the repressor role of CsMAF1 upon the citrus RNA Pol III. Furthermore, we showed in this work that CsMAF1 is a phosphorylated and a dimer in solution, a feature that so far has not been reported for any member of this protein family. We found that CsMAF1 is phosphorylated in vitro by PKA and PKC, and has additional phosphorylation sites for the TOR kinase, including the Thr 62 residue. Interestingly, these phosphorylation sites are located at the dimerization interface of CsMAF1, suggesting that phosphorylation of such sites might regulate the function of the protein and / or its multimeric state. Indeed, mutation of threonine residue Thr62 to aspartic acid (Asp62) decreases the dimer:monomer CsMAF1 ratio, indicating that phosphorylation of the residues at the interface of the dimer destabilizes the dimer, and this may be a novel regulatory mechanism for this class of protein. Thus, these findings open new perspectives for the understanding of the molecular mechanisms involved in RNA Pol III regulation by CsMAF1, as well as for the role of PthA4 in the modulation of RNA Pol III transcription mediated by CsMAF1 / Doutorado / Genetica Vegetal e Melhoramento / Doutor em Genetica e Biologia Molecular