Endogenous viral elements (EVEs) have been recently investigated as a source of transgenerational immune memory. These “viral fossils” are abundant in Aedes mosquitoes and partner with the host’s primary antiviral defense system, the RNA interference (RNAi) pathways. This partnership appears unique to mosquitoes, which encode an expansion of the Piwi endoribonucleases. To interrogate EVE-Piwi partnerships and their role in antiviral defense, I performed a comparative small RNA analysis of two naturally occurring EVE-virus pairs – one in the mosquito Aedes albopictus, and one in the midge Chaoborus americanus. Both express an EVE related to the nucleoprotein of their respective bunyavirus. My results show that Piwis generally do not have antiviral functions in Chaoborus, however EVEs are associated with Piwi recruitment to matched viral RNAs. These findings raise the possibility that RNAi-mediated EVE-virus interactions may be more common among insects than currently recognized.

La spermatogenèse est le processus par lequel les cellules germinales se différencient pour former les spermatozoides. Elle se déroule à l’intérieur des tubes séminifères. Pendant la période embryonnaire, les précurseurs des cellules germinales adultes constituent un pool de cellules appelées cellules germinales primordiales (Primordial Germ Cells, PGCs), qui vont migrer pour aller coloniser les gonades (Durcova-Hills and Capel, 2008; Surani et al., 2008). Au cours de leur migration, les PGCs vont subir une reprogrammation épigénétique de l’ensemble de leur génome, qui leur sera nécessaire pour passer de l’état somatique à l’état de totipotence (Ohinata et al., 2005). Durant cette reprogrammation, l’ADN est massivement démethylé, entrainant l’activation de plusieurs gènes qui sont importants pour le développement des PGCs, mais également l’activation des éléments transposables (ETs) (Hajkova et al., 2008; Sasaki and Matsui, 2008; Surani and Hajkova, 2010). Ces derniers constituent environ 50% du génome des mammifères. Ils sont subdivisés en deux classes et sont connus par leur capacité à être mobilisés dans le génome (Zamudio and Bourc'his, 2010). Cette mobilisation se fait au hasard et constitue ainsi un risque considérable de mutations, qui peuvent provoquer des tumeurs, des pathologies de développement et une infertilité (Zamudio and Bourc'his, 2010). Pour cela, leur expression doit être contrôlée pour maintenir l’intégrité du génome de la lignée germinale. Pour toutes ces raisons, les PGCs ainsi que les cellules germinales en cours de méiose ont développé des stratégies de défenses pour contrôler la mobilisation et l’expression des ETs basées entre autre sur la voie des piwi-interacting RNA (piRNA) (Chuma and Pillai, 2009; Pillai and Chuma, 2012b). Dans le laboratoire du Pr. Stéphane Viville, mes travaux de thèse se sont concentrés sur l’étude d’un gène nommé Tex19 pour Testis Expressed gene chez la souris. Nous avons démontré que ce gène est spécifique des mammifères et est dupliqué chez le rat et la souris en deux paralogues nommés Tex19.1 et Tex19.2. Deux domaines hautement conservés ont été identifiés par alignement multiple des protéines TEX19 et nommés MCP et VPTEL. Ces domaines ne présentent aucune homologie avec des domaines déjà caractérisés, prévenant ainsi toute prédiction de leurs fonctions (Kuntz et al., 2008). L’étude du profil d’expression de Tex19.1 et Tex19.2 a montré que ces deux gènes sont exprimés dans l’ectoderme et les PGCs. Ils sont aussi co-exprimés dans le testicule de l’âge embryonnaire à l’âge adulte. Néanmoins, seul Tex19.1 est exprimé dans les ovaires et le précurseur du placenta appelé cône ectoplacentaire (Celebi et al., 2012). Le knockout (KO) de Tex19.1 provoque une infertilité masculine chez la souris avec un arrêt de la spermatogenèse au stade pachytène, accompagnée d’une surexpression d’un rétrotransposon, MMERVK10C (Ollinger et al., 2008). Récemment, il a été démontré que Tex19.1 joue aussi un rôle dans le développement du placenta (Reichmann et al., 2013). Au cours de mes trois années de thèse, nous avons approfondie l’étude du KO de Tex19.1dans le testicule, les cellules embryonnaires souches (Embryonic Stem Cells, ESCs) et le placenta (Tarabay et al., 2013). Nous avons également étudié le phénotype observé suite au double KO de Tex19.1 et Tex19.2. [...] / We recently characterized two new mammalian specific genes, Tex19.1 and its paralog Tex19.2. Both genes are expressed in pachytene spermatocytes in adult testes. In addition, Tex19.1 is expressed in pluripotent cells (ES, EG, iPS and PGC cells), the inner cell mass of the blastocysts and the placenta. In order to decipher Tex19 functions, we generate three types of knockout (KO): i) KO of Tex19.1 ii) KO of Tex19.2 iii) double KO (DKO) of both genes. All Tex19.1-/- KO animals are growth-retarded and half of them die just after birth. This phenotype is probably linked to placenta defects. Surviving adults Tex19.1-/- KO males display a variable spermatogenesis phenotype, associated with an up-regulation of one endogenous retrovirus, MMERVK10C. Tex19.2 KO mice exhibit a subtle phenotype. Few seminiferous epitheliums are degenerated while the rest appear normal. DKO show a fully penetrant phenotype similar to the most severe Tex19.1-/- phenotype. DKO males exhibit small testes. Despite the presence of spermatogonia and spermatocytes, spermatogenesis is blocked at the pachytene stage. By RNA deep-sequencing on 10 days old DKO and WT testes, prior to histological phenotype, 114 genes are significantly up-regulated and 320 genes significantly down-regulated in the DKO compared to the WT. Gene ontology analyses show that among of these genes, two essential pathways are altered: meiosis and the piRNA pathway. Consistent with that, GST-pulldown and immunoprecipitation experiments demonstrate that MIWI, MILI, MAEL and MVH are partners of TEX19. Considering PIWI proteins function in the silencing of transposable elements through the piRNA pathway, we checked if TEX19 paralogs bind piRNA. By immunoprecipitation using WT and KO testes, we show that both TEX19.1 and TEX19.2 bind small RNA of 30 nucleotides through their VPTEL domain. This study highlights the pivot role of Tex19 paralogs in three essential functions of mammalian life cycle, i.e. pluripotency, placenta-supported in utero growth and fertility. The functional similarities of both paralogs, through the expression control of one endogenous retrovirus and the binding of piRNAs, lead us to propose that Tex19 paralogs are new members of the piRNA pathway.

Major fraction of eukaryotic genomes is composed of transposons. Mobilization of these transposons leads to mutations and genomic instability. In animals, these selfish genetic elements are regulated by a class of small RNAs called PIWI interacting RNAs (piRNAs). Thus host piRNA pathway acts as a defense against pathogenic transposons. Many piRNA pathway genes are rapidly evolving indicating that they are involved in a host-pathogen arms race. In my thesis, I investigated the nature of this arms race by checking functional consequences of the sequence diversity in piRNA pathway genes. In order to study the functional consequences of the divergence in piRNA pathway genes, we swapped piRNA pathway genes between two sibling Drosophila species, Drosophila melanogaster and Drosophila simulans. We focused on RDC complex, composed of Rhino, Deadlock and Cutoff, which specifies piRNA clusters and regulates transcription from clusters. None of the D. simulans RDC complex proteins function in D. melanogaster. Rhino and Deadlock interact and colocalize in D. simulans and D. melanogaster, but D. simulans Rhino does not bind D. melanogaster Deadlock, due to substitutions in the rapidly evolving Shadow domain. Cutoff from D. simulans stably binds and traps D. melanogaster Deadlock. Adaptive evolution has thus generated cross-species incompatibilities in the piRNA pathway which may contribute in reproductive isolation.

The Piwi interacting RNA pathway (piRNA) transcriptionally and post-transcriptionally silences transposons in the germline to maintain host genome integrity and faithful transmission of the genetic materials. In Drosophilaovaries, maternally loaded piRNAs kick-start piRNA biogenesis and convert precursor transcripts into piRNAs to replenish the piRNA pool during oogenesis. piRNA clusters are the genomic source of piRNA precursors, which are determined by the HP1 homolog Rhino and accessary factors. Rhino specifically binds to piRNA cluster chromatin. I was intrigued by how Rhino localizes to piRNA clusters to specify piRNA precursors. TREX is a conserved mRNA biogenesis complex composed of UAP56 and the THO complex. Identification of UAP56 as a cluster transcript-processing factor established the link between piRNA biogenesis and the general mRNA processing machinery. In my thesis, I investigated the functions of UAP56 and THO in piRNA cluster transcript processing. I characterized an RNP specific to cluster transcripts, defined by binding with both factors, which is distinct from RNP of bulk mRNA transcripts, and found that assembly of these RNPs depends on Rhino. These findings imply that piRNA precursors are specified co-transcriptionally. Additionally, I found that TREX mutants lead to a loss of Rhino binding specificity. I propose that Rhino and TREX co-transcriptionally scan for cluster and transposon sequences to establish loci that produce piRNA precursors. Surprisingly, I also discovered a piRNA-independent function for TREX in transposon silencing. I showed that TREX mutants lead to transcriptionally activation of a number of transposon families without affecting their piRNA biogenesis and piRNA mediated repressive histone modifications. I propose that TREX could mediate a conserved transposon silencing mechanism.