<p>Glycogen storage disease type Ia is an autosomal recessive disorder caused by a mutation in the glucose-6-phosphatase (G6Pase) catalytic subunit, encoded in humans by G6PC. G6Pase dephosphorylates glucose-6-phosphate (G6P) in the liver to generate glucose that can be shuttled to the bloodstream to maintain normoglycemia. Patients with GSD Ia typically present at 6 months of age with sever hypoglycemia, which is lethal if untreated. The current treatment is a strict dietary regimen in which children must be fed every 2 hours overnight or given nasogastric tube feeding, and adults must consume uncooked cornstarch around the clock to maintain normal blood sugar levels. This treatment maintains survival but fails to prevent other symptoms related to metabolism of the excess G6P, and patients develop hepatic adenomas that may become hepatocellular carcinoma later in life, in addition to progressive renal complications.</p><p>To overcome the problems persisting during dietary therapy, the Koeberl lab has sought to develop gene therapy approaches that use adeno-associated virus (AAV) vectors to replace the G6pase activity, restoring normoglycemia and normal metabolic processes. However, the vast majority of AAV-delivered genetic material exists as episomes that do not replicate as cells divide, so the effects of AAV gene therapy on GSD Ia mouse and dog models have proven temporary. We hypothesized that driving integration of therapeutic vector genomes into an affected individual's genome would improve beneficial effects' longevity.</p><p>We tested several approaches to accomplish this, and have found positive effects using a zinc finger nuclease (ZFN) that targets the mouse safe harbor ROSA26 locus to induce homologous recombination of the G6PC donor vector into the mouse genome. We were able to see an improvement in mouse survival to 8 months of age, an increase in G6Pase activity at 3 months of age, and a decrease in glycogen accumulation at 3 months of age, when the ZFN vector is administered alongside the G6PC vector, compared with mice that received the G6PC vector alone.</p><p>We have also taken an alternative approach to overcoming the long-term complications of the current dietary treatment, which would augment rather than replace the current treatment. We have examined several drugs known to induce autophagy in other disease models or cell culture systems, to determine if we could manipulate autophagic activity in G6PC knockdown hepatocytes or GSD Ia mice. We have found positive results using rapamycin, a well-studied MTOR inhibitor, in mice and cells, and have screened several other drugs as well, finding positive effects for bezafibrate, mifepristone, carbamazepin, and lithium chloride, in terms of lipid reduction (which accumulates as a symptom of GSD Ia) and/or LC3-II enhancement, which is reduced in GSD Ia due to downregulation of autophagy during G6P accumulation.</p> / Dissertation

A produção de proteínas terapêuticas recombinantes compreende moléculas complexas e de alto valor agregado, incluindo a enzima glucocerebrosidase (GCase). Sua deficiência resulta na Doença de Gaucher, passível de tratamento por meio da terapia de reposição enzimática. A forma ativa da GCase recombinante usada na terapia apresenta resíduos terminais de manose expostos no seu perfil de glicosilação. Perfil este que espera-se ser reproduzido por meio da construção de uma linhagem de Pichia pastoris com um padrão de glicosilação humanizado, por meio da deleção de dois genes envolvidos no sistema de glicosilação da levedura: alg3 e och1, responsáveis pela posterior hiper-manosilação característica desse organismo. Assim, a expressão da GCase será usada como modelo no desenvolvimento desta linhagem de Pichia pastoris que permita a expressão de glicoproteínas com um perfil humanizado específico de glicosilação. Além da produção da linhagem mutante pela técnica de CRISPR/Cas9, propomos a construção de duas linhagens controle: uma expressando a proteína GCase para análise do seu padrão selvagem de glicosilação em P. pastoris e outra expressando a proteína Cas9 de Streptoccocus pyogenes (SpCas9). A linhagem P. pastoris/GCase foi construída testando-se duas sequências sinal de secreção diferentes: fosfatase alcalina (PHO1) e albumina humana (Alb). Resultados de western blot mostraram a GCase no lisado celular e baixos níveis de proteína secretada no sobrenadante de cultura, sendo mais expresso na linhagem contendo a sequência PHO1. A linhagem P. pastoris/SpCas9 foi construída e a enzima SpCas9 foi detectada via western blot no lisado celular após indução com metanol. Para a produção da linhagem com padrão de glicosilação humanizado propôs-se a deleção dos genes alg3 e och1 e a inserção, pela via de reparo por recombinação homóloga (HDR), de marcas de resistência aos antibióticos higromicina ou canamicina. Para tal, propusemos a construção de dois vetores finais de expressão do sistema CRISPR/Cas9 em P. pastoris, cada um contendo a enzima SpCas9 e os RNAs guia (gRNAs) para deleção do gene alg3 ou och1, e também a construção de dois fragmentos para HDR contendo o gene de resistência ao antibiótico flanqueado por regiões de 1Kb de homologia com a região de deleção do gene alg3 ou och1. A construção dos vetores e fragmentos para HDR foram inicialmente feitas por meio de técnicas de clonagem clássica. No entanto, apesar de inúmeras tentativas, resultados de PCR e sequenciamento mostraram o insucesso das construções. Partiu-se então para a técnica de Gibson Assembly®, através da qual os dois fragmentos para HDR foram construídos. Porém, os vetores de expressão contendo SpCas9 e os gRNAs ainda apresentam dificuldades na sua construção. Esforços ainda estão sendo feitos para a construção dos vetores e consequente tentativa de estabelecimento das linhagens mutantes. O sucesso no estabelecimento de um sistema de expressão de proteínas heterólogas com este padrão de glicosilação humano específico permitirá a obtenção e possível comercialização da GCase em sua forma terapêutica. Além disso, permitirá possíveis edições genômicas futuras para um padrão de maior complexidade de glicosilação humanizado, criando uma plataforma nacional para produção de outras glicoproteínas terapêuticas de interesse biotecnológico. / The production of therapeutic recombinant protein comprises complex and high valued molecules, including the glucocerebrosidase enzyme (GCase). Its deficiency results in Gaucher Disease, susceptible of treatment by enzymatic replacement therapy. The active form of recombinant GCase employed in therapy presents exposed terminal mannose residues in its glycosylation pattern. We hope to reproduce such pattern by constructing a Pichia pastoris strain with a specific human glycosylation pattern through the deletion of two genes involved in yeast glycosylation system, alg3 and och1, responsible for the final hyper-mannosylation characteristic of this organism. Therefore, the expression of GCase will be a case model for the development of the recombinant Pichia pastoris strain that could allow the expression of glycoproteins with a specific humanized glycosylation profile. Despite the establishment of the mutant strain using the CRISPR/Cas9 technique, we propose the construction of two control strains: one expressing the GCase protein for analysis of its wild type glycosylation pattern and another one expressing the Cas9 protein from Streptoccocus pyogenes (SpCas9). The P. pastoris/GCase strain was constructed testing two different secretion signal sequences: alkaline fosfatase (PHO1) and human albumin (Alb). Western blot results have shown GCase in cell lysate and in low expression levels in culture supernatant, being more expressed in the strain containing the PHO1 signal sequence. P. pastoris/SpCas9 strain was constructed and SpCas9 enzyme was detected via western blot in cell lysate after the induction with methanol. To produce the strain with the humanized glycosylation pattern, the deletion of alg3 and och1 genes was proposed along with the insertion, by homology directed repair pathway (HDR), of hygromycin and kanamycin antibiotics resistance marks. In order to do so, we have proposed the construction of two final expression vectors of the CRISPR/Cas9 system in P. pastoris, each one containing SpCas9 enzyme and the guide RNAs (gRNAs) for deletion of alg3 or och1, and also the construction of two fragments for HDR containing the antibiotics resistance gene flanked by 1Kb regions of homology with the deleted regions of alg3 or och1. Vectors and HDR fragments constructions were initially performed using classic cloning techniques. However, despite numerous tries, PCR and sequencing results have shown the failure of the constructions. Then, we moved on to the Gibson Assembly® technique, through which the two HDR fragments were built. Still, the expression vectors containing SpCas9 and the gRNAs presented difficulties in its assembly. Efforts continue to be made to successfully construct the remaining vectors and to establish the mutant lineage. Success in the establishment of a heterologous protein expression system with specific human glycosylation pattern will allow the obtainment and possible commercialization of the therapeutic form of GCase. Furthermore, it will also allow possible future genomic editing to a high complexity human glycosylation pattern, creating a national platform for the production of other therapeutic glycoproteins of biotechnological interest.

The term 'Retinitis pigmentosa' (RP) represents a group of inherited, late-onset diseases characterised by progressive retinal degeneration due to photoreceptor death. Mutations in the RP2 gene are found in 7-18% of patients with X-linked RP, one of the most severe forms. The RP2 gene product is a membrane-associated protein which encompasses two distinct domains. The N-terminal domain is well characterised as possessing GTPase-activating protein (GAP) activity towards the small GTPase ARL3 and thus regulate the transport of lipid-modified proteins within the photoreceptor cell. However, it is not known if the loss of this particular function of RP2 is the sole reason that causes the disease, while the role of the protein's C-terminus remains unknown. This thesis focuses on the characterisation of two novel protein-protein interactions of RP2 with the aim to investigate novel roles of the protein. Firstly, evidence is provided that a highly-conserved cluster of RP2 residues that span both the N- and C-terminus participate in direct interaction with Osteoclast-stimulating factor 1 (OSTF1). Two hypotheses are explored about the potential role of the complex in SRC-mediated RP2 phosphorylation and the regulation of cell motility. Secondly, the catalytic subunit of DNA-dependent protein kinase (DNA PK) is identified as a novel interaction partner of RP2 in cultured cells. The two proteins are shown to co-localise in the nuclear and membrane compartments of a retinal-derived cell line and might engage in a kinase-substrate relationship. So far, no evidence was found that RP2 participates in the canonical function of DNA PK which is the regulation of DNA double-stranded breaks. Finally, the CRISPR/Cas9 genome editing method was applied on zebrafish embryos to generate a novel vertebrate animal model for the loss of RP2 function. One out of three different zebrafish lines with rp2 mutations was shown by histology to have mild late-onset thinning of the photoreceptor outer segments. The present thesis reports previously unexplored aspects of RP2's function and will, therefore, contribute to understanding the molecular mechanisms that underlie RP. Moreover, this thesis will contribute to the discussion about the usefulness of zebrafish as an RP model.

L’édition du génome repose sur la création de cassures double brin à un endroit précis du génome à l’aide de nucléases artificielles (ZFN, TALEN, CRISPR/Cas9) et sur les différents systèmes de réparation que la cellule va mettre en place pour réparer ces dommages. Les deux systèmes de réparation principaux sont le NHEJ (Non Homologous End Joining) et la RH (Recombinaison Homologue). Le NHEJ consiste en une ligation directe des extrémités de la coupure pouvant induire de petites insertions ou délétions avant la ligation. Ces mutations, si elles sont introduites dans un exon, vont modifier le cadre de lecture et pouvoir inactiver le gène cible (Knock Out). La RH permet la réparation de la cassure en recopiant les informations présentes sur la chromatide soeur. Si un ADN exogène comportant des homologies avec la séquence à réparer est inséré avec les nucléases artificielles, la cellule peut le prendre comme matrice de réparation, il est ainsi possible d’insérer n’importe quelle mutation ou transgène de manière précise (Knock In). Ici, différentes stratégies ont été développées pour optimiser ces approches d’édition du génome. Le couplage du domaine Nter de la protéine CtIP à la nucléase Cas9 permet d’augmenter le taux d’insertion par homologie d’un transgène au site de coupure. Le couplage de l’exonucléase Trex2 à la nucléase Cas9 nickase permet quant à lui d’augmenter le taux de mutation après coupure. Ces nouvelles approches peuvent être largement utilisées et permettent de faciliter l’édition du génome. / Genome editing relies on the ability of artificial nucleases (TALEN or CRISPR/Cas9 system) to induce double strand break into a precise and unique sequence in a whole genome and on the different DNA repair system. The two major DNA repair systems are NHEJ (Non Homologous End Joining) and HR (Homologous Recombination). NHEJ consists on DNA end direct ligation. This system can lead to deletion or insertion at the cut site. These mutations, when induced in an exon, can induce reading frame change and gene inactivation (Knock out). HR consists on the use of sister chromatid to copy lost information in order to complete the double strand break. If an exogenous DNA with homologies with the targeted DNA is inserted with artificial nucleases, it can be used as a template and can permit to introduce any transgene at the cut site (Knock In). In this work, different strategies were used to optimize genome editing. By fusing Nter part of CtIP to Cas9, the KI rate of an exogenous DNA is increased and by fusing Trex2 exonuclease to Cas9, the mutation rate induced is also increased. These two approaches can be widely used to improve genome editing strategies.

La leucémie à tricholeucocytes (ou HCL pour Hairy Cell Leukemia) est un syndrôme lymphoprolifératif B rare du sujet âgé, caractérisé par une infiltration médullaire et splénique de cellules présentant des protrusions cytoplasmiques. Des thérapies de première ligne efficaces existent contre ce cancer et l’intérêt de la recherche biomédicale dans ce domaine réside désormais dans le développement de nouvelles molécules actives contre les cellules leucémiques réfractaires aux traitements de référence. Parmi les voies de signalisation dérégulées dans l’HCL, celle des Rho-GTPases influe sur les phénomènes de croissance cellulaire et d’organisation du cytosquelette d’actine, perturbés dans les tricholeucocytes. Les travaux précédemment menés au laboratoire ont montré la sous-expression dans l’HCL d’une Rho-GTPase atypique, RhoH, dont l’expression ectopique dans un modèle cellulaire d’HCL atténue la progression tumorale. Afin de déterminer les cibles moléculaires de RhoH dans cette leucémie, une étude transcriptomique a été réalisée et a montré la sous-expression du marqueur de surface CD38 lorsque RhoH est surexprimée.Plus qu’un marqueur de différenciation lymphocytaire, CD38 est une molécule à effets pléïotropiques (à la fois récepteur, enzyme et protéine d’adhérence cellulaire), importante dans le développement des lymphocytes B. CD38 a également été décrit comme un marqueur délétère dans la leucémie lymphoïde chronique et représente une cible thérapeutique dans le myélome multiple. Bien qu’exprimé par un tiers des patients porteurs de l’HCL, son rôle dans cette leucémie restait jusqu’alors inconnu.Les travaux présentés dans cette thèse décrivent, d’une part, l’étude de la régulation du gène CD38 par RhoH dans l’HCL, et d’autre part, l’impact de la protéine CD38 dans la progression de cette leucémie. Des données préliminaires sur l’activité de fragments de promoteur du gène CD38 semblent indiquer un rôle du facteur de transcription Smad1 dans la régulation de ce gène par RhoH. Grâce à une technique de genome editing, nous avons produit deux lignées cellulaires HCL knock out pour le gène CD38. Ces modèles nous ont permis de déterminer que CD38 promeut la survie ainsi que l’adhésion à l’endothélium des cellules HCL, et modifie également leurs propriétés migratoires in vitro. Nous avons également observé que CD38 favorisait la progression tumorale dans un modèle murin de xénogreffe de ces lignées cellulaires. Enfin, des données produites par nos collaborateurs ont montré que CD38 est un marqueur de mauvais pronostic pour la rechute des patients atteints d’HCL et qu’il constitue une cible thérapeutique potentielle pour les 30% de patients qui l’expriment.Mimer l’effet de RhoH dans l’HCL à des fins thérapeutiques s’avèrerait délicat. Le ciblage de CD38 semble donc une alternative de choix. En effet, la sous-expression de RhoH dans l’HCL favorise l’expression de cette protéine, dont l’effet est délétère pour les patients puisqu’elle participe à la progression de la leucémie. Les anticorps monoclonaux thérapeutiques dirigés contre CD38 étant déjà utilisés en clinique pour traiter d’autres leucémies, ce travail ouvre la voie à l’extension de leur utilisation dans le traitement de l’HCL réfractaire, pour les patients qui l’expriment. / Hairy Cell Leukemia (HCL) is a B-lymphoproliferative disorder of the elderly, which is characterized by medullar and splenic homing of “hairy” cells bearing cytoplasmic protrusions. Efficient first-line therapies against this cancer do exist and the real challenge in biomedical research is now to develop new molecules targeting leukemic cells which are resistant to these first-line treatments. Among some deregulated signaling pathways that have been described in HCL, Rho-GTPases are noteworthy, mediating proliferation and reorganization of actin cytoskeleton, being both disrupted in hairy cells. Former works from our laboratory have shown the underexpression in HCL cells of an atypical Rho-GTPase called RhoH, which reconstitution decreased malignant progression in both in vitro and in vivo models of HCL. In order to determine the molecular targets of RhoH, a microarray study was performed that showed underexpression of the cell surface marker CD38 while RhoH is overexpressed.Not only a differentiation marker of lymphocytes, CD38 is a pleiotropic molecule (being at the same time a receptor, an enzyme and an adhesion protein), which is important in B-cell development. It is also known as a bad prognosis marker in chronic lymphocytic leukemia and a therapeutic target in multiple myeloma. Its role in Hairy Cell Leukemia has not been studied yet, despite its expression in one third of HCL patients.The work presented in this thesis deals with, on the one hand, the study of the regulation of CD38 gene by RhoH in HCL, on the other hand, the impact of CD38 protein on HCL progression. Preliminary data seem to indicate a potential role of Smad1 transcription factor in this mechanism of regulation of CD38 by RhoH. Thanks to genome editing technology, we produced two HCL cell lines knock out for the CD38 gene. These models allowed us to prove that CD38 enhances hairy cells survival and adhesion to endothelium, and modulates their migratory features in vitro. We also showed that CD38 promotes disease progression in vivo in an HCL xenograft mouse model. Finally, data from our collaborators indicated that CD38 is a bad prognosis marker for HCL relapses and could be a potential therapeutic target.Mimicking RhoH effects for therapeutic purposes would be somewhat tricky. Targeting CD38 seems an interesting alternative, as RhoH underexpression favours CD38 expression, which is deleterious for patients by enhancing malignant progression. As therapeutic monoclonal antibodies targeting CD38 are already used in clinics to treat other leukemias, this work brings the evidence of their potential usefulness against refractory HCL cases expressing this marker.

Mutational analyses of cancer genomes have identified novel candidate cancer genes with hitherto unknown function in cancer. To enable phenotyping of mutations in such genes, we have developed a scalable technology for gene knock-in and knock-out in human somatic cells based on recombination-mediated construct generation and a computational tool to design gene targeting constructs. Using this technology, we have generated somatic cell knock-outs of the putative cancer genes ZBED6 and DIP2C in human colorectal cancer cells. In ZBED6-/- cells complete loss of functional ZBED6 was validated and loss of ZBED6 induced the expression of IGF2. Whole transcriptome and ChIP-seq analyses revealed relative enrichment of ZBED6 binding sites at upregulated genes as compared to downregulated genes. The functional annotation of differentially expressed genes revealed enrichment of genes related to cell cycle and cell proliferation and the transcriptional modulator ZBED6 affected the cell growth and cell cycle of human colorectal cancer cells. In DIP2C-/-cells, transcriptome sequencing revealed 780 differentially expressed genes as compared to their parental cells including the tumour suppressor gene CDKN2A. The DIP2C regulated genes belonged to several cancer related processes such as angiogenesis, cell structure and motility. The DIP2C-/-cells were enlarged and grew slower than their parental cells. To be able to directly compare the phenotypes of mutant KRAS and BRAF in colorectal cancers, we have introduced a KRASG13D allele in RKO BRAFV600E/-/-/ cells. The expression of the mutant KRAS allele was confirmed and anchorage independent growth was restored in KRASG13D cells. The differentially expressed genes both in BRAF and KRAS mutant cells included ERBB, TGFB and histone modification pathways. Together, the isogenic model systems presented here can provide insights to known and novel cancer pathways and can be used for drug discovery.

Recent advances in genome engineering technologies based on the CRISPR-associated RNA-guided endonuclease Cas9 are enabling the systematic interrogation of genome function. Analogous to the search function in modern word processors, Cas9 can be guided to specific locations within complex genomes by a short RNA search string. Using this system, DNA sequences within the endogenous genome and their functional outputs are now easily edited or modulated in virtually any organism of choice. Cas9-mediated genetic perturbation is simple and scalable, empowering researchers to elucidate the functional organization of the genome at the systems level and establish causal linkages between genetic variations and biological phenotypes. To facilitate successful and specific Cas9 targeting, we first optimize the guide RNAs (sgRNA) to significantly enhance gene editing efficiency and consistency. We also systematically characterize Cas9 targeting specificity in human cells to inform the selection of target sites and avoid off-target mutagenesis. We find that SpCas9 tolerates mismatches between guide RNA and target DNA at different positions in a sequence-dependent manner, sensitive to the number, position and distribution of mismatches. We also show that Cas9-mediated cleavage is unaffected by DNA methylation and that the dosage of Cas9 and sgRNA can be titrated to minimize off-target modification. Additionally, we provide a web-based software tool to guide the selection and validation of target sequences as well as off-target analyses. We next demonstrate that Cas9 nickase mutants can be used with paired guide RNAs to introduce targeted double-strand breaks. Because individual nicks in the genome are repaired with high fidelity, simultaneous nicking via appropriately offset guide RNAs can reduce off-target activity by over 1,500-fold in human cells. In collaboration with researchers at the University of Tokyo, we further identified a PAM-interacting domain of the Cas9 nuclease that dictates Cas9 target recognition specificity. Finally, we present protocols that provide experimentally derived guidelines for the selection of target sites, evaluation of cleavage efficiency and analysis of off-target activity. Beginning with target design, gene modifications can be achieved within as little as 1-2 weeks. Taken together, this work enables a variety of genome engineering applications from basic biology to biotechnology and medicine.

miRNAs are small regulatory RNAs, which function as post-transcriptional mRNA regulators. They direct ribonucleoprotein complexes to cognate mRNA to repress them by translational inhibition and degradation. miRNAs regulate thousands of mRNAs in mammals and have been recognized as regulatory factors in most cellular and developmental processes. Dysregulation of the miRNA pathway can lead to severe defects and diseases. Interestingly, a unique situation exists in mouse oocytes, where all the miRNA pathway components are present, yet the pathway is dispensable and nonfunctional, the molecular foundation of this phenomenon and its significance still remain unclear. In spite of the pronounced effects of the miRNA pathway in gene regulation in somatic cells, study strategies of the pathway bare limitations. Current methods for studying the activity of the miRNA pathway employ corelative studies (such as NGS) or reporter assays, which have relatively low throughput and are prone to artifacts. Here, I present design and development of a new strategy for directly monitor global miRNA pathway activity and integrity in near physiological conditions in living cells, which could also be employed in vivo for studies of mouse oocytes. The strategy is based on fluorescently tagged endogenous proteins of the...

A produção de proteínas terapêuticas recombinantes compreende moléculas complexas e de alto valor agregado, incluindo a enzima glucocerebrosidase (GCase). Sua deficiência resulta na Doença de Gaucher, passível de tratamento por meio da terapia de reposição enzimática. A forma ativa da GCase recombinante usada na terapia apresenta resíduos terminais de manose expostos no seu perfil de glicosilação. Perfil este que espera-se ser reproduzido por meio da construção de uma linhagem de Pichia pastoris com um padrão de glicosilação humanizado, por meio da deleção de dois genes envolvidos no sistema de glicosilação da levedura: alg3 e och1, responsáveis pela posterior hiper-manosilação característica desse organismo. Assim, a expressão da GCase será usada como modelo no desenvolvimento desta linhagem de Pichia pastoris que permita a expressão de glicoproteínas com um perfil humanizado específico de glicosilação. Além da produção da linhagem mutante pela técnica de CRISPR/Cas9, propomos a construção de duas linhagens controle: uma expressando a proteína GCase para análise do seu padrão selvagem de glicosilação em P. pastoris e outra expressando a proteína Cas9 de Streptoccocus pyogenes (SpCas9). A linhagem P. pastoris/GCase foi construída testando-se duas sequências sinal de secreção diferentes: fosfatase alcalina (PHO1) e albumina humana (Alb). Resultados de western blot mostraram a GCase no lisado celular e baixos níveis de proteína secretada no sobrenadante de cultura, sendo mais expresso na linhagem contendo a sequência PHO1. A linhagem P. pastoris/SpCas9 foi construída e a enzima SpCas9 foi detectada via western blot no lisado celular após indução com metanol. Para a produção da linhagem com padrão de glicosilação humanizado propôs-se a deleção dos genes alg3 e och1 e a inserção, pela via de reparo por recombinação homóloga (HDR), de marcas de resistência aos antibióticos higromicina ou canamicina. Para tal, propusemos a construção de dois vetores finais de expressão do sistema CRISPR/Cas9 em P. pastoris, cada um contendo a enzima SpCas9 e os RNAs guia (gRNAs) para deleção do gene alg3 ou och1, e também a construção de dois fragmentos para HDR contendo o gene de resistência ao antibiótico flanqueado por regiões de 1Kb de homologia com a região de deleção do gene alg3 ou och1. A construção dos vetores e fragmentos para HDR foram inicialmente feitas por meio de técnicas de clonagem clássica. No entanto, apesar de inúmeras tentativas, resultados de PCR e sequenciamento mostraram o insucesso das construções. Partiu-se então para a técnica de Gibson Assembly®, através da qual os dois fragmentos para HDR foram construídos. Porém, os vetores de expressão contendo SpCas9 e os gRNAs ainda apresentam dificuldades na sua construção. Esforços ainda estão sendo feitos para a construção dos vetores e consequente tentativa de estabelecimento das linhagens mutantes. O sucesso no estabelecimento de um sistema de expressão de proteínas heterólogas com este padrão de glicosilação humano específico permitirá a obtenção e possível comercialização da GCase em sua forma terapêutica. Além disso, permitirá possíveis edições genômicas futuras para um padrão de maior complexidade de glicosilação humanizado, criando uma plataforma nacional para produção de outras glicoproteínas terapêuticas de interesse biotecnológico. / The production of therapeutic recombinant protein comprises complex and high valued molecules, including the glucocerebrosidase enzyme (GCase). Its deficiency results in Gaucher Disease, susceptible of treatment by enzymatic replacement therapy. The active form of recombinant GCase employed in therapy presents exposed terminal mannose residues in its glycosylation pattern. We hope to reproduce such pattern by constructing a Pichia pastoris strain with a specific human glycosylation pattern through the deletion of two genes involved in yeast glycosylation system, alg3 and och1, responsible for the final hyper-mannosylation characteristic of this organism. Therefore, the expression of GCase will be a case model for the development of the recombinant Pichia pastoris strain that could allow the expression of glycoproteins with a specific humanized glycosylation profile. Despite the establishment of the mutant strain using the CRISPR/Cas9 technique, we propose the construction of two control strains: one expressing the GCase protein for analysis of its wild type glycosylation pattern and another one expressing the Cas9 protein from Streptoccocus pyogenes (SpCas9). The P. pastoris/GCase strain was constructed testing two different secretion signal sequences: alkaline fosfatase (PHO1) and human albumin (Alb). Western blot results have shown GCase in cell lysate and in low expression levels in culture supernatant, being more expressed in the strain containing the PHO1 signal sequence. P. pastoris/SpCas9 strain was constructed and SpCas9 enzyme was detected via western blot in cell lysate after the induction with methanol. To produce the strain with the humanized glycosylation pattern, the deletion of alg3 and och1 genes was proposed along with the insertion, by homology directed repair pathway (HDR), of hygromycin and kanamycin antibiotics resistance marks. In order to do so, we have proposed the construction of two final expression vectors of the CRISPR/Cas9 system in P. pastoris, each one containing SpCas9 enzyme and the guide RNAs (gRNAs) for deletion of alg3 or och1, and also the construction of two fragments for HDR containing the antibiotics resistance gene flanked by 1Kb regions of homology with the deleted regions of alg3 or och1. Vectors and HDR fragments constructions were initially performed using classic cloning techniques. However, despite numerous tries, PCR and sequencing results have shown the failure of the constructions. Then, we moved on to the Gibson Assembly® technique, through which the two HDR fragments were built. Still, the expression vectors containing SpCas9 and the gRNAs presented difficulties in its assembly. Efforts continue to be made to successfully construct the remaining vectors and to establish the mutant lineage. Success in the establishment of a heterologous protein expression system with specific human glycosylation pattern will allow the obtainment and possible commercialization of the therapeutic form of GCase. Furthermore, it will also allow possible future genomic editing to a high complexity human glycosylation pattern, creating a national platform for the production of other therapeutic glycoproteins of biotechnological interest.

Xanthomonas oryzae pv. oryzae (Xoo), l'agent causal de bactériose vasculaire du riz (BLB), injecte des protéines de liaison à l'ADN, appelées Transcription Activator-Like Effectors (TALEs) dans les cellules hôtes afin de moduler l'expression de gènes cibles. Certains TALEs agissent comme des facteurs de virulence majeurs, indispensables à la mise en place du BLB et ciblent des gènes de sensibilité du riz. Les TALEs majeurs de Xoo ciblent universellement les gènes de sensibilité de la famille SWEET. Il existe dans la nature un polymorphisme des séquences ADN des gènes SWEET reconnues par les TALEs qui confère une résistance à la maladie. L’utilisation de la technologie TALEN a permis d'introduire artificiellement ce type de mutation dans le promoteur du gène de sensibilité SWEET14 le rendant insensible aux TALEs et conférant une résistance à certaines souches de Xoo asiatiques. La caractérisation des populations de Xoo africaines montrent qu'elles sont distinctes de celles d’Asie. L’objectif du projet de thèse était de créer à l'aide des technologies d'édition des génomes des sources de résistances efficaces contre un large panel de Xoo maliennes.Dans une première partie, l'édition des boites ADN de SWEET14 ciblées par des TALEs majeurs de souches africaines a effectivement permis d'obtenir des résistances contre les souches utilisant le TALE TalF (initialement appelé Tal5) mais pas contre celles utilisant TalC. La caractérisation du répertoire de TALEs des souches maliennes par des approches fonctionnelles (sensibilité des lignées éditées, expression de SWEET14 et autres gènes cibles de TALEs) et in silico (séquençage du génome de 8 souches) à révélé une diversité fonctionnelle de ces répertoires et la présence simultanée quasi systématique de versions actives et redondantes de TalF et TalC. La caractérisation de la sensibilité de variétés de riz locales aux souches de Xoo maliennes a montré que ces dernières possèdent un large spectre de virulence et qu'à une exception près, toutes les variétés testées sont sensibles au BLB. L'édition en multiplex par la technique CRISPR/Cas9 des boites TalF et TalC a aboli l'induction de SWEET14 en réponse à une souche malienne. Cependant, les lignées correspondantes sont restées sensibles à cette souche. Dans la dernière partie, pour expliquer ce résultat, nous avons postulé l'existence d'au moins un gène de sensibilité, cible de TalC et redondant avec SWEET14. Une approche bioinformatique a permis d'identifier un locus dont plusieurs caractéristiques en faisaient un candidat intéressant. Ce locus, nommé ATAC (pour Alternative TalC Target) est composé de deux gènes, ATAC1 et ATAC2 induits de façon bidirectionnelle par TalC. Nous avons montré qu'ATAC2, qui code pour un facteur de transcription bHLH atypique potentiellement impliqué dans l’élongation cellulaire et l'immunité chez le riz, se comporte comme un locus de sensibilité lorsqu'il est induit par des TALE artificiels dans un système gain de fonction. Nous avons édité simultanément le promoteur SWEET14 et le locus ATAC. Ces éditions devraient empêcher la reconnaissance du promoteur SWEET14 et du locus ATAC par TalC et TalF afin de conférer une résistance large au BLB au Mali. / Xanthomonas oryzae pv. oryzae (Xoo), the causal agent of bacterial leaf blight of rice (BLB), injects DNA binding proteins called Transcription Activator-Like Effectors (TALEs) into host cells to manipulate plant genes expression. Some TALEs behave as major virulence factors essential for BLB to occur by binding directly to target DNA boxes of rice susceptibility genes and inducing their expression. Xoo major TALEs universally target susceptibility genes of the SWEET family. In nature, polymorphism in the DNA sequence of SWEET genes recognized by TALEs confers resistance to BLB. Using the TALEN technology, this type of mutations has been artificially introduced in the promoter of the SWEET14 susceptibility gene to make it TALE-unresponsive and confer resistance to some Asian Xoo. The characterizations of Malian Xoo populations show that they are distinct from the Asian ones. The PhD project aimed to create broad tailored BLB resistance against Malian Xoo using genome editing technologies.First, editing SWEET14 DNA boxes targeted by major TALEs of African strains indeed yielded resistance against strains relying on TalF (initially named Tal5) but not against those relying on TalC. The characterization of Malian strains TALE repertoires using functional (edited lines susceptibility assays, SWEET14 and other TALE target expression studies) and in silico (genome sequencing of 8 strains) approaches uncovered functional diversity in these repertoires and, the almost systematic, simultaneous presence of active and redundant versions of TalF and TalC. In susceptibility assays of local rice varieties, Malian Xoo strains exhibited a broad virulence spectrum and, with one exception, all tested varieties were susceptible to BLB. Multiplex editing of TalF and TalC target boxes with the CRISPR/Cas9 technology abolished SWEET14 induction in response to a Malian strain. However the corresponding rice lines remained susceptible to this strain. Finally, to explain these results, we postulated the existence of, at least, a TalC target susceptibility gene redundant with SWEET14. Bioinformatics analysis identified a rice locus with several features electing it as a high priority candidate. This locus named ATAC (Alternative TalC Target) is composed of two genes, ATAC1 and ATAC2, bidirectionally upregulated by TalC. We further showed that ATAC2 which is predicted to code for an atypical bHLH transcription factor potentially involved in rice cell elongation and immunity, behaves as a susceptibility gene upon artificial TALEs-mediated induction in a gain of function assay. We used the CRISPR/Cas9 system to simultaneously edit the SWEET14 promoter and the ATAC locus. These mutations should prevent the recognition of the SWEET14 promoter and the ATAC locus by TalC and TalF, compromise their transcriptional induction and ultimately provide broad BLB resistance in Mali.