Evidence in different systems suggests that local translation is involved in synaptic plasticity in both neuron and muscle, but the mechanism by which this occurs is still poorly understood. The mRNA-binding protein Syncrip is conserved from fly to mammals and is thought to be involved in localized translation in both oocytes and neurons. Previous work has shown that Syncrip associates with mRNAs encoding key synaptic proteins at the Drosophila larval neuromuscular junction. Here I show that Syncrip is necessary for the structure and function of the neuromuscular junction. First, the loss of Syncrip leads to overgrowth of the neuromuscular junction. Second, Syncrip is required for proper expression of the Ca²⁺-sensor Synaptotagmin1 at the presynapse, and loss of Syncrip causes a decrease in vesicle release probability. Third, while it was not possible to measure mRNA distribution in neurons, Syncrip mutants, like other perturbations in synaptic plasticity, correlate with changes in mRNA localization in muscle. Fourth, the overexpression and loss of Syncrip function suggest that the nuclear and nucleolar trafficking of the eukaryotic translation initiation factor eIF4E may be important to regulating synaptic morphology. These data suggest that Syncrip is involved in mRNA localization and translation in synaptic plasticity.

Establishment of cell polarity requires the involvement of several posttranscriptional regulatory mechanisms, including mRNA localization and translational control. A family of highly conserved RNA binding proteins in vertebrates, VICKZ (V̲g1RBP/V̲era, I̲MP-1, 2, 3, C̲RD-BP, K̲OC, Z̲BP-1) proteins, has been shown to act in these two processes. Previous studies of the posttranscriptional mechanisms mediated by VICKZ family members have been largely limited by the lack of genetic approaches in certain vertebrate systems. Identification of Imp, the Drosophila member of the VICKZ family, opened the possibility to use genetic approaches to investigate the roles of a VICKZ family member in mRNA localization and translational control. In this dissertation, we show that Imp is associated with Squid and Hrp48, two heterogeneous proteins (hnRNP) that complex with one another to regulate localized expression of gurken (grk). In addition, Imp binds grk mRNA with high affinity in vitro and is concentrated at the site of grk localization in midstage oocytes. Mutation of the Imp gene does not substantially alter grk expression, but does partially suppress the grk mis-expression phenotype of fs(1)k10 mutants. In contrast, overexpression of Imp in germ line cells results in mislocalization of grk mRNA and protein. The opposing effects of reduced and elevated Imp activities on grk expression suggest that Imp acts in regulation of grk expression, but in a redundant way. To further explore the mechanisms by which localized expression of grk is regulated by Imp, a deficiency screen was conducted to search for dominant modifiers of the dorsalized phenotype resulting from Imp overexpression. Twelve genomic regions were identified to contain dominant modifiers of the Imp overexpression phenotype. Further characterization of mutants of genes within these genomic regions led to identification of five modifiers, including cyclin E (cycE), E2f transcriptional factor 1 (E2f1), lingerer (lig), snail (sna) and mushroom body expressed (mub). E2f1 encodes a transcriptional factor that is involved in regulating the G1 to S phase transition during mitosis. Mutation of E2f1 results in altered grk mRNA and protein distribution within oocyte, revealing a role for this gene in regulation of grk expression. / text

Cytoplasmic mRNA-containing granules in eukaryotic cells play key roles inthe storage, localisation and degradation of mRNA. In yeast, depletion of glucoseleads to the rapid inhibition of translation initiation and consequent appearance of Pbodiesand EGP-bodies. P-bodies contain factors of the mRNA decay pathway andtherefore, are likely to be sites in which mRNAs targeted for degradation arelocalised. In contrast, EGP-bodies lack decay components and contain onlytranslation initiation factors and RNA binding proteins. Thus EGP-bodies have beensuggested to be storage repositories for mRNAs that need to be rapidly translatedfollowing glucose readdition. In this study we utilised the m-TAG system to investigate the localisation ofendogenous MS2-tagged mRNAs with P-bodies and EGP-bodies. A triplefluorescent labelled system developed show that a class of unregulated mRNAslocalised into P-bodies following glucose starvation. It was also observed that thesespecific abundant classes of mRNAs can be found in aggregates prior to any cellularstress and upon glucose starvation these aggregates coalesce into larger granules thatcolocalise with P-body components. This coalescence of mRNA aggregatesfollowing glucose starvation does not rely upon the recruitment of mRNA decayfactors and appears to precede this event. Indeed mRNAs in mutants deficient in Pbodyformation still develop large aggregates following glucose stress. In unstressedcells it appears that the mRNA granules are implicated in high-level translation ofthese specific abundant mRNAs. Following the inhibition of translation initiation inresponse to stress, these granules nucleate P-body formation via aggregation and therecruitment of mRNA decay factors.

L’organisation cellulaire repose sur une distribution organisée des macromolécules dans la cellule. Deux ARNm, cen et ik2, montrent une colocalisation parfaite aux centrosomes. Ces deux gènes font partie du même locus sur le chromosome 2L de Drosophila melanogaster et leur région 3’ non traduite (3’UTR) se chevauchent. Dans le mutant Cen, le transport de Ik2 est perturbé, mais dans le mutant Ik2, la localisation de cen n’est aucunement affectée. Ces résultats suggèrent que cen est le régulateur principal de la co-localisation de cen et ik2 aux centrosomes et que cette co-localisation se produit par un mécanisme impliquant la région complémentaire au niveau du 3’UTR des deux transcrits. La localisation de cen au niveau des centrosomes dans les cellules épithéliales de l’embryon est conservée dans différentes espèces de Drosophile : D. melanogater, D. simulans, D. virilis et D. mojavensis. Cependant, la localisation de ik2 n’est pas conservée dans D. virilis et D. mojavensis, deux espèces dont les gènes cen et ik2 sont dissociés dans le génome. Ces résultats suggèrent que la proximité de Cen et Ik2 dans le génome est importante afin d’avoir un événement de co-localisation de ces deux transcrits. J’ai généré différentes lignées de mouches transgéniques dans lesquelles un transgène contenant la séquence GFP fusionnée à différentes partie de Cen (partie codante, 3’UTR, Cod+3’UTR) qui sont sous le contrôle du promoteur UAS et qui sont gal4 inductibles. La région codante de l’ARNm cen était suffisante pour avoir un ciblage précis du transcrit aux centrosomes. / Messenger RNA (mRNA) localization plays a key role in establishing cellular architecture and function. The centrocortin (cen) and IkB Kinase-like 2 (ik2) mRNAs are co-localized to centrosomes in embryonic epithelial cells. Interestingly, both of these genes are organized in a head-to-head configuration in the genome, with their 3’ untranslated regions (3’UTRs) overlapping on opposite DNA strands. Here we show that gene positioning of cen and ik2 is important for the co-localization of these transcripts during Drosophila embryogenesis. The localization of cen is conserved within the Drosophila phylogeny and ik2 cannot localize when it is separated from the cen locus. Also, loss of function mutants of cen show a complete loss of ik2 localization, proposing that cen is the main driver of the co-localization. Structure-function analysis revealed that the coding region of cen is necessary for its centrosomal targeting, suggesting that a cis-regulatory motif that drives its localization is located in the coding region. This study reveals for the first time the importance of gene positioning for RNA localization. We suggest a model where cen mRNA is the main driver of centrosomal localization, which may occur through post-transcriptional interaction/annealing of these mRNAs via their 3’UTRs.

La traduction locale permet un contrôle spatial de l'expression des gènes. Dans ce travail, j'ai participé à deux cribles de localisation d'ARNm concernant plus de 1000 transcrits. Le premier était un crible double ARNm/protéine qui utilisait une approche de BAComics pour co-détecter les ARNm et la protéine pour laquelle ils codent. Le second a été réalisé à l'aide d'une nouvelle approche smFISH à haut-débit et a analysé tous les ARNm codant pour des protéines centrosomales et des régulateurs mitotiques. Le premier crible a révélé des cas de traduction locale dans divers compartiments subcellulaires, et notamment au niveau des protrusions cytoplasmiques, des centrosomes, de l’appareil de Golgi, des endosomes et des pores nucléaires, ce qui n'avait jamais été décrit auparavant. De manière remarquable, la traduction du peptide naissant était nécessaire pour le transport de nombreux transcrits localisés. De plus, j'ai montré que plusieurs ARNm (tels que ASPM et DYNC1H1) sont traduits dans des structures dédiées appelées usines de traduction.Le deuxième crible a révélé 8 transcrits localisés et traduits au niveau des centrosomes. J'ai montré que la localisation de ces 8 transcrits est régulée par le cycle cellulaire et qu'elle nécessite également la traduction du polypeptide naissant. En utilisant le gène ASPM comme modèle, j'ai visualisé des ARNm et des polysomes uniques avec les systèmes MS2 et SunTag, respectivement. Cela a révélé un transport dirigé des polysomes ASPM vers les centrosomes au début de la mitose, lorsque cet ARNm commence à être localisé. Ces données fournissent des preuves fortes d'un mécanisme de ciblage co-traductionnel dépendant de moteurs moléculaires ainsi que de la protéine naissante. Cela va à l'encontre du dogme actuel selon lequel le transport d'ARNm est un processus basé sur l'ARN et agissant sur des molécules réprimées pour la traduction. En revanche, cela suggère que des mécanismes tels que celui utilisé par le SRP sont plus répandus qu'on ne le pensait auparavant. / Local translation allows a spatial control of gene expression. Here, I participated in two mRNA localization screens imaging more than 1000 transcripts in total: (i) the first was a dual mRNA/protein screen that used a BAComics approach to co-detect mRNAs and the protein they encode; (ii) the second was done using a new high-throughput smFISH approach to screen all genes that encode centrosomal proteins and mitotic regulators. The first screen revealed cases of local translation at various subcellular compartments including cytoplasmic protrusions, centrosomes, Golgi, endosomes and the nuclear pore, which was never described before. Remarkably, translation of the nascent peptide was required for the transport of many localized transcripts. In addition, I showed that several mRNAs (such as ASPM and DYNC1H1) are translated in dedicated structures called translation factories.The second screen revealed 8 transcripts that are localized and translated at the centrosome. I showed that the localization of these 8 transcripts is regulated by the cell cycle, and that it also requires translation of the nascent polypeptide. Using the endogenous ASPM gene as a model, I imaged single mRNAs and polysomes with the MS2 and SunTag systems, respectively. This revealed a directed transport of ASPM polysomes towards centrosomes at the onset of mitosis, when this mRNA starts localizing. These data provide definitive evidence for a co-translational targeting mechanism dependent on motors as well as the nascent protein. This argues against the current dogma that mRNA transport is an RNA-based process acting on translationally repressed molecules. Instead, it suggests that SRP-like mechanisms are more widespread than previously thought.

Le transport subcellulaire et la traduction localisée des molécules d'ARNm semble être un processus très répandu et important pour contrôler la distribution asymétrique des protéines dans les cellules. L’ARNm, Bsg25D, connu pour se localiser aux centrosomes et aux microtubules astraux dans les embryons de drosophile au cours des premiers événements d'embryogenèse, a été sélectionné pour déterminer le rôle et l'importance du ciblage de l'ARNm à l'appareil mitotique lors de la division cellulaire. La localisation de Bsg25D aux centrosomes dans les embryons de drosophile est conservée entre espèces telles que D. melanogaster, D. simulans et D. yakuba. Bsg25D encode une protéine qui est étroitement liée à la Ninein (Nin) et à la Ninein-like protein (Nlp), deux protéines associées aux centrosomes présentes dans les cellules mammifères. L’analyse structure-fonction démontre que la région codante et la région 3’UTR de Bsg25D sont nécessaires pour son ciblage. Ceci suggère qu’un élément de régulation en cis, qui favorise sa localisation se situe dans la région codante + 3’UTR. / The subcellular transport and localized translation of mRNA molecules is emerging as a highly prevalent and important process for controlling asymmetric protein distribution in cells. A candidate mRNA, Bsg25D, known to localize to centrosomes and astral microtubules in Drosophila embryos during early events of embryogenesis, was selected to determine the role and importance of mRNA targeting to the mitotic apparatus during cell division. The localization of Bsg25D to centrosomes in Drosophila embryos is conserved between species such as D. melanogaster, D. simulans and D. yakuba. Bsg25D encodes a protein closely related to centrosome-associated proteins Ninein (Nin) and Ninein-like protein (Nlp) in mammalian cells. Structure function analysis revealed that the coding and 3’UTR of Bsg25D are necessary for its targeting pattern, suggesting that a cis-regulatory motif that drives its localization, is in the coding + 3’ UTR region.

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal.

La localisation des ARNm au niveau des microtubules et des centrosomes laisse voir le centrosome et le fuseau mitotique comme des complexes ribonucléoprotéiques. Cependant, le mécanisme de localisation des ARNm à ces différentes structures ainsi que leurs fonctions dans la régulation de la mitose restent encore incompris. L’objectif était ici de caractériser des protéines de liaison à l’ARN (RNA Binding Proteins, RBPs) fonctionnellement impliquées dans la localisation des ARNm mitotiques chez la Drosophile et d’évaluer la conservation de la fonction de ces RBPs dans les cellules humaines. La déplétion de RBPs par RNAi générée dans des Drosophiles mutantes résulte en des phénotypes distincts de localisation anormale de l’ARNm centrosomique cen et en des défauts mitotiques différents selon le RBP ciblé, suggérant des fonctions différentes de ces RBPs. De plus, dans les jeunes embryons, les RBPs Bru-2 et Mask semblent être fonctionnellement importants pour la mitose via la régulation de l’ARNm cen, donnant un aperçu de la possible fonction mitotique de RBPs dans la régulation d’un ARN centrosomique. De plus, il a été observé dans un criblage d’immunofluorescence dans des cellules HeLa en métaphase que HNRNPUL1 colocalise au fuseau et aux centrosomes. HNRNPUL1 pourrait être impliqué dans la régulation de l’ARNm CDR2 (orthologue de cen) puisque la déplétion de l’orthologue de HNRNPUL1 dans la Drosophile, CG30122, résulte en une localisation anormale de l’ARNm centrosomique cen. / The localization of mRNA to microtubules and centrosomes has led to the suggestion that the centrosome and mitotic spindle are in fact ribonucleoprotein complexes. However, the mechanism of mRNA localization to those structures and its functional contribution in mitosis regulation remain poorly characterized. The objectives here were to identify RNA Binding Proteins (RBPs) involved in mitotic mRNA localization in Drosophila and to assess the conservation of the function of these RBPs in human cells. RNAi-mediated RBP depletion in Drosophila mutants leads to distinct phenotypes of abnormal localization of the centrosomal cen mRNA, and to different mitotic defects depending on the targeted RBP, suggesting different functions for these RBPs. Moreover, in young embryos, Bru-2 and Mask RBPs seem to be functionally important for mitosis through cen mRNA regulation, giving insight into a possible RBP mitotic function in regulating a centrosomal mRNA. In addition, data from an immunofluorescence screen on HeLa cells at metaphase suggests that HNRNPUL1 colocalizes to the spindle and centrosomes. HNRNPUL1 may be involved in the regulation of CDR2 mRNA (cen ortholog) because depletion of the HNRNPUL1 ortholog in flies, CG30122, disrupted cen mRNA localization.

Thèse numérisée par la Division de la gestion de documents et des archives de l'Université de Montréal

The failure to balance protein synthesis, folding, and degradation in the endoplasmic reticulum (ER) leads to the accumulation of unfolded proteins, leading to ER stress. Cells respond to this stress by activating a response signaling pathway known as the Unfolded Protein Response (UPR). One of the branches of the UPR induces the phosphorylation of eIF2α (Eukaryotic Initiation Factor 2) to attenuate global protein synthesis, allowing for a chance to clear misfolded and unfolded proteins. This phosphorylation of eIF2α is opposed by a phosphatase, containing a catalytic subunit, Protein Phosphatase 1, and a scaffolding protein, either GADD34 or CReP. Inhibition of eIF2α phosphatases has shown to promote survival in cell types by prolonging the effects of the UPR. This research focuses on understanding the gene expression patterns and localization of UPR specific genes with the presence of constant ER stress. Zebrafish are an ideal model for this research because they are a good mimic of what happens in humans and provide the ability to study gene expression and localization patterns at different stages during ER stress and its recovery. The eIF2α phosphatases were shown to have a protective effect on apoptosis when overexpressed in acute ER stress but were shown to have a protective effect on apoptosis when knocked out in chronic ER stress. We sought to determine the flow of gene expression of these phosphatases as well as other UPR specific genes, such as BiP and CHOP, to determine the contradictory effects of acute versus chronic ER-stress induced apoptosis. We studied the changes in gene expression for these genes in zebrafish embryos by isolating RNA and performing RT-qPCR after the induction of ER stress with pharmacological drugs across multiple time points. There was increased gene upregulation and mRNA localization to the fin epidermis and eye of GADD34, CReP, and BiP in acute ER stress from 2 hours to 6 hours, and these genes steadily declined in chronic ER stress from 24 hours to 48 hours. CHOP is a late-phase pro-apoptotic protein whose gene expression was upregulated in chronic ER stress from 12 hours to 48 hours. This data was also supported by mRNA localization studies performed by conducting whole mount in-situ hybridization on zebrafish embryos treated with ER stress inducers for 4 hours and 24 hours. Our results indicate that all UPR genes examined are affected by ER stress and their expression patterns are dependent on the time length of ER stress induction, allowing us to get a more in-depth working model of this branch of the UPR signaling pathway in zebrafish.