Bovine adenovirus (BAdV)-3 is a non-enveloped icosahedral DNA virus, which replicates in the nucleus of infected cells, and is being developed as a vector for vaccination for humans and animals. The genome of BAdV-3 is organized into early, intermediate and late genes and it has thirty three predicted open reading frames (Reddy et al., 1998). The late region of BAdV-3 is divided into seven families (L1-L7) (Reddy et al., 1998). One of the proteins expressed in the L-6 region encodes a protein called pVIII, which is a minor capsid protein connecting the core with the inner surface of the capsid. The objective of the current study was to characterize pVIII protein of BAdV-3 and to examine its role in the life cycle of BAdV-3. Anti-pVIII serum detected a protein of 24 kDa at 12-48 hr post infection and an additional protein of 8 kDa at 24-48 hr post infection. While a 24 kDa protein is detected in empty capsids, only the C-terminal cleaved protein of 8 kDa is detected in the mature virion suggesting that amino acids 147-216 of conserved C- terminus of BAdV-3 pVIII are incorporated in mature virions. The pVIII protein predominantly localizes to the nucleus of BAdV-3 infected cells utilizing the classical importin α /β dependent nuclear import pathway. Analysis of mutant pVIII demonstrated that amino acids 57-72 of the conserved N-terminus bind to importin α-3 with high affinity and are required for the nuclear localization. Detection of hexon associated with both, precursor (24 kDa) and cleaved (8 kDa) form of pVIII suggests that the C-terminus of pVIII interacts with Hexon. Based on yeast II hybrid screening assay, we identified the cellular protein DDX3 as an interacting protein partner of pVIII. Earlier, targeting of DDX3 by few viral proteins has defined its role in mRNA transport (Yedavalli et al., 2004) and induction of interferon production (Schroder et al., 2008; Wang et al., 2009). Here, we provide evidence regarding the involvement of DDX3 in cap dependent cellular mRNA translation and show that targeting of DDX3 by the adenovirus pVIII protein abolishes cap-dependent mRNA translation function of DDX3 in virus infected cells. Adenovirus late protein pVIII interacts with DDX3 in transfected and bovine adenovirus (BAdV-3) infected cells. pVIII inhibited capped mRNA translation in-vitro and in-vivo by limiting the amount of DDX3 and eIF3. Diminished amount of DDX3 and eIFs including eIF3, eIF4E and PABP were present in cap binding complex in BAdV-3 infected or pVIII transfected cells with no trace of pVIII in the cap binding complex. The total amount of eIFs appeared similar in uninfected or BAdV-3 infected cells. The co-immunoprecipitation experiments indicated the absence of direct interaction between pVIII and eIF3, eIF4E or PABP. These data indicate that interaction of pVIII with DDX3 depletes eIF3, eIF4E and PABP from the cap-binding complex. We conclude that DDX3 promotes cap-dependent cellular mRNA translation and BAdV-3 pVIII inhibits translation of capped cellular mRNA by excluding functional cap-binding complex from the capped cellular mRNA. BAdV-3 infection of DDX3 positive cells significantly inhibits cellular protein synthesis at late times post-infection. Interestingly, knockdown of DDX3 resulted in significant reduction in virus yield and expression of BAdV-3 late proteins at late times post-infection. Our results suggest that selective translation of BAdV-3 late mRNAs observed at late time post-infection of DDX3 positive cells is abrogated in DDX3 knock down cells. Moreover, the reduction in the extent of protein synthesis is evidenced by less functional 80S and polysomes in pVIII expressing plasmid transfected cells. Alternatively, DDX3 and pVIII binds to BAdV-3 tripartite leader (TPL) and the translation of mRNAs containing TPL at their 5’ ends is enhanced in the presence of pVIII and DDX3 proteins. From this observation, we concluded that pVIII and DDX-3 might promote the translation of late viral mRNAs by interacting with TPL.

L'ARN génomique du Virus de l'Immunodéficience Humaine-1 (VIH-1) est multifonctionnel. Il constitue le génome encapsidé dans les virions et sert d'ARN messager pour la traduction des protéines virales Gag et Gag-Pol. La traduction de ces protéines dépend exclusivement de la machinerie traductionnelle cellulaire et est initiée par deux mécanismes différents : l'initiation canonique dépendante de la coiffe et l'initiation par entrée interne des ribosomes (IRES). Le VIH-1 présente deux IRES, l'un dans la région 5' non traduite (5'-UTR) qui est stimulé en phase G2/M du cycle cellulaire et l'autre dans la région codante de Gag. Ce dernier permet l'initiation de la traduction sur deux AUG en phase et conduit à la production de la protéine Gag pleine longueur mais également à la production d'une isoforme alternative de Gag, tronquée en région N-terminale. Le rôle de cette isoforme reste mal connu. Toutefois la mutation du second AUG chez VIH-1 et donc la suppression de la seconde isoforme de Gag provoque une diminution importante du taux de la réplication virale. La conservation structurelle et fonctionnelle de l'IRES Gag parmi les lentivirus suggère un rôle important de cette isoforme et de l'IRES gag dans le cycle viral. Nos travaux visent à comprendre à un niveau moléculaire les relations hôtes-pathogènes lors de la traduction des messagers viraux. Je me suis particulièrement intéressée aux rôles de la sous unité ribosomale 40S et de l'hélicase cellulaire DDX3 dans l'initiation de la traduction de la polyprotéine Gag du VIH-1. La première partie de ma thèse est consacrée à l'étude de l'interaction entre la sous unité ribosomale 40S et l'IRES gag du VIH-1. Par l'utilisation d'approches complémentaires, nous avons pu démontrer la présence de deux sites distincts de liaison au ribosome qui sont présents à proximité des deux codons d'initiation. Nous avons ensuite évalué à la fois in vitro et in cellulo (en collaboration avec l'équipe de T. Ohlmann, CIRI-ENS-Lyon) l'effet de la délétion de chacun des sites de liaison au 40S sur l'efficacité de traduction de la polyprotéine Gag. Nos résultats valident l'importance fonctionnelle des sites de liaison au ribosome pour une production optimale des deux isoformes de la polyprotéine Gag. La seconde partie de mon travail a consisté à définir le rôle de DDX3 dans l'initiation « coiffe-dépendante » de la traduction de la polyprotéine Gag. DDX3 est une hélicase à ARN à boîte DEAD impliquée dans de nombreux processus cellulaires tels que la régulation du cycle cellulaire et la réponse immunitaire innée mais également dans tous les aspects du métabolisme de l'ARN comme la transcription, l'épissage, l'export nucléaire ou encore la traduction. Plus récemment, il a été montré que DDX3 est nécessaire à la traduction de l'ARN génomique du VIH-1, cependant son rôle exact n'a pas encore été défini. Nous avons purifié une forme recombinante de la protéine en fusion avec la MBP (Maltose Binding Protein) et effectué des cinétiques enzymatiques afin de caractériser ses propriétés biochimiques. Contrairement à ce qui a été précédemment décrit, nos résultats montrent que DDX3 possède une activité ATPase strictement ARN-dépendante avec des constantes cinétiques similaires à celles de son homologue chez la levure, Ded1p. Nous avons également évalué l'activité hélicase de la protéine en présence de substrats de longueur et de nature variables (duplex ARN/ARN ou des hétéroduplex ADN/ARN). D'un point de vue fonctionnel, nous avons réalisé une première série d'expériences qui confirme la stimulation exercée par DDX3 sur la traduction de Gag in vitro. Ces résultats permettent d'envisager la caractérisation biochimique fine des interactions DDX3-ARN viral ainsi que de disséquer le rôle de DDX3 dans l'expression du génome viral. / The Human Immunodeficiency Virus (HIV) genomic RNA is multifunctional. It acts both as a genome that is packaged within virions and as messenger RNA translated to yield the Gag and Gag-Pol polyproteins. The translation of these proteins relies exclusively on the cellular translation machinery and is initiated through two mechanisms: the canonical cap-dependent initiation pathway and the use of internal ribosome entry sites (IRESes). HIV-1 has two IRESes, one located within the 5' UTR (5' UnTranslated Region) that is stimulated during the G2/M phase of the cell cycle, and the other embedded within the Gag polyprotein coding region. The later drives translation initiation from two AUG in frame and results in the production of the full-length Gag protein but also of an additional N-terminally truncated Gag isoform. Few things are known about this isoform, but the mutation of the second AUG causes a significant decrease in the rate of viral replication. The structural and functional conservation of Gag IRES among lentiviruses suggests an important role of this isoform and thus of the IRES in the viral cycle. Our work aims to understand at a molecular level the host-pathogen relationships in the translation of the viral messenger RNA. My work focused on the roles of the 40S ribosomal subunit and of the cellular helicase DDX3 in the translation initiation of Gag. During the first part of my Phd, I studied the interaction between the 40S ribosomal subunit and HIV-1Gag IRES. Following complementary approaches, we evidenced two distinct ribosome binding sites present close to the two the initiation sites of Gag. Then, we evaluated the effect of each 40S binding site deletion on Gag translation efficiency, both in vitro and in cellulo (in collaboration with the team of T. Ohlmann, CIRI-ENS-Lyon). Taken together, our results confirm the functional relevance of the two ribosomal binding sites to ensure optimal production of the two Gag isoforms. The second part of my Phd project aims to define the role of DDX3 in the translation initiation of Gag. DDX3 is a RNA DEAD-box helicase involved in many cellular processes such as cell cycle regulation and the innate immune response but also in all aspects of RNA metabolism such as transcription, splicing, mRNA nuclear export and translation. Recently DDX3 has been shown to favor HIV-1 Gag translation. To define its role, we first purified a recombinant form of the protein and performed kinetic experiments to analyze its biochemical properties. Contrary to what has been previously described, MBP-DDX3 displays a strictly RNA-dependent ATPase activity with kinetic constants similar to those displayed by its yeast counterpart Ded1p. We next evaluated MBP-DDX3 helicase activity towards RNA duplexes or RNA/DNA hybrids, with different length and single strand overhangs. Our preliminary results indicate that DDX3 alone is sufficient to enhance Gag translation in our in vitro system which paves the way to fine biochemistry experiments such as reconstruction of functional initiation complexes assembled onto Gag RNA and evaluation of its role on Gag RNA structure.

Swine influenza viruses (SIV) are a common and an important cause of respiratory disease in pigs. Pigs can serve as mixing vessels for the evolution of reassortment viruses containing both avian and human signatures, which have the potential to cause pandemics. NS1 protein of influenza A viruses is a major antagonist of host defence and it regulates multiple functions during infection by interacting with a variety of host proteins. Therefore, it is important to study swine viruses and NS1-interacting host factors in order to understand the mechanisms by which NS1 regulates virus replication and exerts its host defense functions. Influenza A viruses enter the host through the respiratory tract and infect epithelial cells in the respiratory tract, which form the primary sites of virus replication in the host. Thus, studying SIV infection in primary swine respiratory epithelial cells (SRECs) would resemble conditions similar to natural infection. The objectives of this study were to identify NS1-interacting host factors in the virus-infected SRECs and to understand the physiological role of at least one of the factors in influenza virus infection. The approaches to meet this objective were to generate a recombinant SIV carrying a Strep-tag in the NS1 protein, infect SRECs with the Strep-tag virus, purify NS1-interacting host protein complex from the infected cells by pull-down using strep-tactin resin and then study the physiological role of one of the NS1-interacting partners during influenza infection. Using a reverse-genetics strategy, a recombinant virus carrying the Strep-tag NS1 was successfully rescued and the SRECs were infected with this recombinant virus. The Strep-tag in the NS1 protein facilitated the isolation of an intact NS1-interacting protein complex and the proteins present in the complex were identified by liquid chromatography-tandem mass spectrometry. The identified proteins were grouped to enrich for different functions using bioinformatics. This gave an insight into the different functions that NS1 may regulate during infection and the potential host partners involved in these functions. Among the host proteins identified as potential interaction partners, RNA helicases were particularly of interest to study. Influenza being an RNA virus, RNA helicases could have important functions in the virus life cycle. Among the identified RNA helicases, DDX3 has been shown to regulate IFNβ induction and affect the life cycle of a number of viruses. However, its function in influenza A virus life cycle has not been studied. Hence, this study explored whether DDX3 has any role in the influenza A virus life cycle. Immunoprecipitation studies revealed viral proteins NP and NS1 as direct interaction partners with DDX3. DDX3 is a known component of stress granules (SGs) and influenza A virus lacking the NS1 gene is reported to induce SG formation. Therefore, the role of DDX3 in SG formation, induced by PR8 influenza A virus lacking NS1 (PR8 del NS1) was explored. The results from this study showed that DDX3 co-localized with NP in SGs indicating that DDX3 may interact with NP in the SGs. NS1 protein was found to inhibit virus-induced SGs and DDX3 downregulation impaired virus-induced SG formation. The contribution of the different domains of DDX3 to viral protein interaction and virus-induced SG formation was also studied. While DDX3 helicase domain did not interact with NS1 and NP, it was essential for DDX3 localization in virus induced SGs. Moreover, DDX3 downregulation resulted in the increased replication of PR8 del NS1virus, accompanied by an impairment of SG induction in infected cells. Since DDX3 is reported to regulate IFNβ induction, the role of DDX3 in influenza A virus induced IFNβ induction was also examined. Using small molecule inhibitors and siRNA-mediated gene knockdown, the RIG-I pathway was identified as the major contributor of influenza induced IFNβ induction in newborn porcine tracheal epithelial (NPTr) cells. DDX3 downregulation and overexpression also showed that DDX3 has an inhibitory effect on IFNβ expression induced by both influenza infection and low molecular weight (LMW) poly I:C treatment, which is also a RIG-I ligand. RNA competition assay to identify the mechanism of DDX3-mediated inhibition, showed that RIG-I binding affinity to its ligands LMW poly I:C and influenza viral RNA (vRNA) is much higher than that of DDX3. Furthermore, DDX3 downregulation enhanced titers of the PR8 del NS1 virus, while it did not affect the titers of the wild-type strains of PR8 and SIV/SK viruses. Overall, the results show that DDX3 has an antiviral role and the SG regulatory function of DDX3 has a profound effect on virus replication than the IFNβ regulatory function.

京都大学 / 新制・課程博士 / 博士(生命科学) / 甲第24119号 / 生博第481号 / 新制||生||64(附属図書館) / 京都大学大学院生命科学研究科統合生命科学専攻 / (主査)教授 野田 岳志, 教授 鈴木 淳, 教授 高田 穣 / 学位規則第4条第1項該当 / Doctor of Philosophy in Life Sciences / Kyoto University / DFAM

L’ARN hélicase à boite DEAD de la levure S.cerevisiae Ded1 est une protéine essentielle dont la fonction a été conservée au cours de l’évolution. Ses homologues fonctionnels sont impliqués dans le développement et le cycle cellulaire. Ded1 a longtemps été associée à l’étape de scanning de la région 5’UTR des ARNm au niveau de l’initiation de la traduction. Nous avons utilisé différentes approches comme les co-immunoprécipitations, des analyses de spectrométrie de masse, des tests de complémentation génétique, de séparation des complexes sur gradients de saccharose, des expériences de localisation in situ et d’enzymologie pour montrer que Ded1 interagissait physiquement avec des complexes cytoplasmique et nucléaire de liaison à la coiffe des ARNm. Nous avons également montré que Ded1 peut passer du noyau vers le cytoplasme par différentes voies d’export nucléaire. De façon intéressante, ses partenaires protéines sont capables de stimuler son activité ATPase. De plus, nous avons montré qu’il existait un lien génétique entre Ded1 et ses partenaires. Nous avons également montré que Ded1 colocalise partiellement avec ses partenaires dans des gradients de saccharose, suggérant que Ded1 pourrait être associée à certains mRNPs. Nos résultats encore préliminaires indiquent que Ded1 pourrait s’associer à d’autre ARNs coiffés. Ainsi, Ded1 pourrait remodeler les complexes associés à différentes étapes de la vie des ARN coiffés. / The budding yeast DEAD-box RNA helicase Ded1 is an essential yeast protein that is closely related to a subfamily of DEAD-box proteins that are involved in developmental and cell-cycle regulation. Ded1 is generally considered to be a translation-initiation factor that helps the 40S ribosome scan the mRNA from the 5' 7-methylguanosine cap to the AUG start codon. We have used IgG pulldown experiments, mass spectroscopy analyses, genetic experiments, saccharose gradients, in situ localizations, and enzymatic assays to show that Ded1 is a cap-associated factor that actively shuttles between the cytoplasm and the nucleus. We show that Ded1 physically interacts with various cap-associated factors and that its enzymatic activity is stimulated by these factors. By using various mutated proteins, we show that Ded1 is genetically linked to these factors. Ded1 comigrates with these factors on saccharose gradients, but the peak of Ded1 sediments slightly heavier than for the other factors, which suggests that Ded1 is predominately associated with a subset of the mRNPs. Finally, purification of the protein complexes associated with Ded1 and subsequent analysis by nanoLC-MS/MS indicates that Ded1 is associated with both nuclear and cytoplasmic mRNPs. Preliminary experiements showed that Ded1 can associate with other capped RNA. We conclude that Ded1 may function as a remodeling factor that is needed to form the different complexes associated with the different processing steps of the capped RNA.