Une détérioration des ARNm, le vecteur de l’information génétique, ou une altération des mécanismes régulant leur synthèse, leur traduction, ou leur dégradation peut être responsable de l’apparition d’une maladie. Ainsi, il est important de comprendre les mécanismes affectant les ARNm et notamment leur dégradation. Cette dernière est initiée par la réaction de désadénylation caractérisée par le raccourcissement progressif de la queue poly(A) de l’ARNm. Un complexe responsable de la désadénylation est CCR4-NOT. Les protéines BTG/Tob lient la désadénylase CAF1 de CCR4-NOT via leur domaine APRO. Dans cette thèse, l’interaction de BTG2 avec PABPC1, la protéine liant la queue poly(A), a été étudiée. Cette association augmente l’activité désadénylase de CAF1 in vitro et in cellulo. Le rôle de cette liaison sur les propriétés antiprolifératives de BTG2 et l’impact de BTG2 sur la traduction ont aussi été analysés. Un modèle expliquant l’activation de la désadénylation par BTG2 est proposé. / Damaging mRNA, the molecules carrying the genetic information, or altering the mechanisms responsible for their synthesis, their translation or their degradation can be responsible for initiating diseases. It is thus important to understand the mechanisms impacting mRNA, especially their degradation. The latter is initiated by the deadenylation reaction characterized by the progressive shortening of the mRNA 3’ poly(A) tail. One of the complexes responsible for deadenylation is CCR4-NOT. BTG/Tob proteins are able to bind the CAF1 deadenylase of CCR4-NOT via their APRO domain. In this thesis, interaction of BTG2 with PABPC1, the factor binding poly(A) tail, was analyzed. This binding stimulates CAF1 deadenylase activity in vitro and in cellulo. The role of this interaction on the antiproliferative properties of BTG2 and the impact of BTG2 on translation were also investigated. A model explaining deadenylation activation by BTG2 is proposed.

O Vírus Respiratório Sincicial Humano, do inglês human Respiratory Syncytial Virus (hRSV), é uma das maiores causas de doenças respiratórias agudas, principalmente em crianças e bebês entre seis meses e dois anos de idade. Não há drogas eficazes ou vacina aprovada até o momento para esse vírus, apesar das décadas de intensa pesquisa e grande quantidade de dados sobre ele acumulados. O genoma do hRSV codifica onze proteínas e a compreensão das interações entre essas proteínas virais e as proteínas do hospedeiro é essencial para que possíveis alvos terapêuticos contra o hRSV sejam identificados. No laboratório, anteriormente, foi dado enfoque às interações entre as proteínas celulares e as proteínas virais de matriz (M), nucleoproteína (N) e fosfoproteína (P). Neste trabalho, analisamos as interações da proteína viral M2-1 (cofator essencial para a transcrição) através da mesma estratégia utilizada naqueles experimentos, de fusão a FLAG (gerando FLAG-M2-1) e imunoprecipitação com anticorpos contra esse peptídeo. As proteínas co-imunoprecipitadas, identificadas por espectrometria de massas, foram: poly(A)-binding protein cytoplasmic 1 (PABPC1), Y-box binding protein 3 (YBX3), e Nuclease-sensitive element-binding protein 1 (YBX1). M2-1 é capaz de integrar-se ao complexo chamado de semelhante a corpúsculos de inclusão (IB like, do inglês), formado por N e P, que é similar estruturalmente aos corpúsculos de inclusão encontrados em células infectadas (IBs). Essa propriedade foi usada para analisar que proteínas celulares seriam recrutadas para esse outro nível de organização dessas três proteínas virais, envolvidas na transcrição. O complexo FLAG-N/P/M2-1 co-imunoprecipitou as proteínas celulares: Hsp70, Hsp90 (Heat shock proteins 70 e 90), Npm (Nucleophosmin), que podemos agrupar como chaperonas; PABPC1, YBX1, YBX3, ligantes de RNA; e sub-unidade pICIn do metilossomo, associada a modificação pós-tradução. Detalhamos a análise para YBX3, obtendo evidências adicionais de sua interação com M2-1 em ensaios de complementação de proteína fragmentada (Split-NanoLuc), e de co-localização por imunofluorescência indireta. Finalmente, utilizamos a metodologia de expressão em bactérias para demonstrar a interação entre M2-1 e os domínios funcionais de PABPC1, porém esses ensaios não foram conclusivos. / Human Respiratory Syncytial Virus (hRSV) is one of the leading causes of acute respiratory diseases, especially in children and infants between six months and two years of age. There is no effective drug or vaccine approved so far for this virus, despite decades of intensive research and large amount of data on it. The genome of hRSV encodes 11 proteins and the understanding of the interactions between these viral proteins and host proteins is essential to identify possible therapeutic targets against hRSV. In the lab, previously, was given focus to the interactions between cellular proteins and viral proteins matrix (M), nucleoprotein (N) and phosphoprotein (P). In this paper, we analyze the viral M2-1 (cofactor essential for transcription) protein interactions through the same strategy used in those experiments: fusion with FLAG (generating FLAG-M2-1) and immunoprecipitation with antibodies against this peptide. The co-immunoprecipitated proteins, identified by mass spectrometry, were: Poly (A)-binding protein cytoplasmic 1 (PABPC1), Y-box binding protein 3 (YBX3), and Nuclease-sensitive element-binding protein 1 (YBX1). M2-1 is able to integrate the complex called similar to inclusion bodies (IB like), formed by N and P, which is similar structurally to the inclusion bodies found in infected cells (IBs). This property has been used to analyze which cellular proteins would be recruited for this new level of organization of these three viral proteins involved in transcription. The cellular proteins co-immunoprecipitated with the complex FLAG-N/P/M2-1, were: Hsp70, Hsp90 (Heat shock proteins 70 and 90), Npm (Nucleophosmin), that we can group as chaperones; PABPC1, YBX1, YBX3, RNA ligands; and the methylosome sub-unit pICIn, post-translational modification-associated. We detailed the analysis for YBX3, obtaining additional evidence of its interaction with M2-1 in fragmented protein complementation tests (Split-NanoLuc), and co-localization by indirect immunofluorescence. Finally, we used the methodology of expression in bacteria to demonstrate the interaction between M2-1 and functional domains of PABPC1, but these tests were not conclusive.

Neurodevelopmental disorders (NDD) refer to a group of conditions resulting from disturbances of the developing brain with a typical onset in childhood before puberty. Genetic causes make up a large part of developmental delays, which is why genetic examinations play a decisive role in the clarification of the causes of NDD. Due to the development of Next Generation Sequencing (NGS) and the increased use of genome-wide analyses in recent years, it has become clear that a large proportion of cases are due to rare, monogenic alterations in each case. Meanwhile, 1534 genes have been currently associated with NDD. Nevertheless, half of the evaluated cases remain without a valid diagnosis. However, this is a prerequisite for personalized support and the estimation of the development prognosis, as well as the differentiated assessment of the risk of recurrence for family members. To decipher the genetics of NDD, I systematically analyzed the exome sequences of 104 individuals with NDD and their relatives (see Figure 2). In 10 of the 104 cases, I was able to find variants in already known genes that partially explain the phenotype. I intensively evaluated all the cases for new candidate genes and identified 89 candidate genes in 58 individuals (see Supplementary, Table S1). In the remaining 46 individuals, no candidate gene could be identified. I then scored the candidate genes to prioritize them regarding the probability of being true NDD genes. From the detailed analysis of a relatively small cohort of NDD individuals (n=104) and the resulting 89 candidate genes, a total of 9 research collaborations have emerged (see Table 1). Of the candidate genes with further research in AutoCasC, six are in my top 20 candidate genes, which is a good indication of the efficiency of this systematic approach on deciphering the genetics of NDD. Studies on the candidate genes SKOR2, HCN2, SP9, CCDC66, and TANC1 are currently being worked on by collaborators worldwide and we could add our clinical and genetic data (see Table 1). Further, I was more substantially involved in the identification of the candidate gene ATP2B1, subsequently studied functionally, and published in the American Journal of Human Genetics (IF 11,0) with me as coauthor as a novel NDD gene, and the ongoing research on a genotype-phenotype correlation with functional lines of evidence of NDD-individuals with variants in DOCK4. Furthermore, I have led the efforts for the genes RIPPLY2 and PABPC1. I was able to describe three individuals from two families with compound-heterozygous variants in RIPPLY2 in two sisters and a homozygous nonsense variant in an 8-year-old boy. All individuals had multiple vertebral body malformations in the cervical and thoracic region, small or absent rib involvement, myelopathies, and common clinical features of spondylocostal dysostosis 6 (SCDO6) including scoliosis, mild facial asymmetry, spinal spasticity, and hemivertebrae. At this time, RIPPLY2 was only associated as a candidate gene with SCDO6 and had only been described in a small cohort of seven individuals from five families in two publications. I could confirm that bi-allelic variants in RIPPLY2 cause congenital cervical spine malformation in spondylocostal dysostosis type 6 and broaden the phenotype by adding myelopathy with or without spinal canal stenosis and spinal spasticity to the symptom spectrum as a first author in a publication published in Clinical Genetics (IF 4,4). In the study on PABPC1, I describe four individuals with an overlapping phenotype of developmental delay, expressive speech delay, autistic features, and heterozygous de novo variants that cluster in the PABP domain of PABPC1. Further symptoms are seizures and behavioral disorders. Molecular modeling predicted that the variants are pathogenic and would lead to decreased binding affinity to mRNA metabolism-related proteins such as PAIP2. Co-immunoprecipitation confirmed this as it demonstrated a significant weakening of the interaction between mutant PABPC1 and PAIP2. Electroporation of mouse embryo brains showed that Pabpc1 knockdown decreases the proliferation of neural progenitor cells. The wild type Pabpc1 could rescue this disturbance, while three of the four variants did not. Together with partners from the Central South University in Changsha, China, I was able to propose that pathogenic missense variants in the PABP-domain of PABPC1 lead to a novel form of developmental disorder and published my work in Genetics in Medicine (IF 8,9). Through this, I demonstrated that systematic trio exome analysis and identification and characterization of candidate genes followed by prioritizing the genes based on systematic scoring and by building international cooperation to gather further individuals, describe the phenotypes, and prove that the pathogenicity of the variants is an excellent way to decipher the genetics of NDD. With this approach, I was able to describe PABPC1 as a novel NDD-gene and confirm the association between RIPPLY2 and SCDO6. Moreover, I contributed as a co-author to the publication of ATP2B1 as a novel NDD-gene and to the ongoing research on SKOR2, SP9, HCN2, CCDC66, TANC1, and DOCK4. More might follow in the future. The continuation of this research in genetic diagnostics is important for creating personalized support and prevention programs for individuals with neurodevelopmental delays, to be able to estimate the developmental prognosis, and to be able to assess the recurrence risk of other family members in a more differentiated way.:1 Introduction 1.1 Genetics of neurodevelopmental disorders 1.2 Identification of neurodevelopmental delay genes 1.3 Assessment of candidate genes 1.4 Rationale 1.5 Results 1.5.1 Identification of neurodevelopmental delay genes 1.5.2 Scoring of the identified candidate genes 1.5.3 Candidate genes under further research 2 Publications 2.1 Congenital cervical spine malformation due to bi-allelic RIPPLY2 variants in spondylocostal dysostosis type 6 2.2 De novo variants in the PABP-domain of PABPC1 lead to developmental delay 3 Summary 4 References 5 Internet resources 6 Supplementary 7 Presentation of personal scientific contribution 8 Declaration of Authorship