Surprisingly little is known about the organization and distribution of tRNAs and tRNA-related sequences on a genome-wide scale. While tRNA complements are usually reported in passing as part of genome annotation efforts, and peculiar features such as the tandem arrangements of tRNAs in Entamoeba histolytica have been described in some detail, comparative studies are rare. We therefore set out to systematically survey the genomic arrangement of tRNAs in a wide range of eukaryotes to identify common patterns and taxon-specific peculiarities. We found that tRNA complements evolve rapidly and that tRNA locations are subject to rapid turnover. At the phylum level, distributions of tRNA numbers are very broad, with standard deviations on the order of the mean. Even within fairly closely related species, we observe dramatic changes in local organization. Consistent with this variability, syntenic conservation of tRNAs is also poor in general, with turn-over rates comparable to those of unconstrained sequence elements. We conclude that the genomic organization of tRNAs shows complex, lineage-specific patterns characterized by extensive variability, and that this variability is in striking contrast to the extreme levels of sequence-conservation of the tRNA genes themselves. Our comprehensive analysis of eukaroyotic tRNA distributions provides a basis for further studies into the interplay between tRNA gene arrangements and genome organization in general. Secondly, we focused on the investigation of small non-coding RNAs (ncRNAs) from whole transcriptome data. Since ncRNAs constitute a significant part of the transcriptome, we explore this data to detect and classify patterns derived from transcriptome-associated loci. We selected three distinct ncRNA classes: microRNAs, snoRNAs and tRNAs, all of which undergo maturation processes that lead to the production of shorter RNAs. After mapping the sequences to the reference genome, specific patterns of short reads were observed. These read patterns appeared to reflect RNA processing and, if so, should specify the RNA transcripts from which they are derived. In order to investigate whether the short read patterns carry information on the particular ncRNA class from which they orginate, we performed a random forest classification on the three distinct ncRNA classes listed above. Then, after exploring the potential classification of general groups of ncRNAs, we focused on the identification of small RNA fragments derived from tRNAs. After mapping transcriptome sequence data to reference genomes, we searched for specific short read patterns reflecting tRNA processing. In this context, we devised a common tRNA coordinate system based on conservation and secondary structure information that allows vector representation of processing products and thus comparison of different tRNAs by anticodon and amino acid. We report patterns of tRNA processing that seem to be conserved across species. Though the mechanisms and functional implications underlying these patterns remain to be clarified, our analysis suggests that each type of tRNA exhibits a specific pattern and thus appears to undergo a characteristic maturation process.

Surprisingly little is known about the organization and distribution of tRNAs and tRNA-related sequences on a genome-wide scale. While tRNA complements are usually reported in passing as part of genome annotation efforts, and peculiar features such as the tandem arrangements of tRNAs in Entamoeba histolytica have been described in some detail, comparative studies are rare. We therefore set out to systematically survey the genomic arrangement of tRNAs in a wide range of eukaryotes to identify common patterns and taxon-specific peculiarities. We found that tRNA complements evolve rapidly and that tRNA locations are subject to rapid turnover. At the phylum level, distributions of tRNA numbers are very broad, with standard deviations on the order of the mean. Even within fairly closely related species, we observe dramatic changes in local organization. Consistent with this variability, syntenic conservation of tRNAs is also poor in general, with turn-over rates comparable to those of unconstrained sequence elements. We conclude that the genomic organization of tRNAs shows complex, lineage-specific patterns characterized by extensive variability, and that this variability is in striking contrast to the extreme levels of sequence-conservation of the tRNA genes themselves. Our comprehensive analysis of eukaroyotic tRNA distributions provides a basis for further studies into the interplay between tRNA gene arrangements and genome organization in general. Secondly, we focused on the investigation of small non-coding RNAs (ncRNAs) from whole transcriptome data. Since ncRNAs constitute a significant part of the transcriptome, we explore this data to detect and classify patterns derived from transcriptome-associated loci. We selected three distinct ncRNA classes: microRNAs, snoRNAs and tRNAs, all of which undergo maturation processes that lead to the production of shorter RNAs. After mapping the sequences to the reference genome, specific patterns of short reads were observed. These read patterns appeared to reflect RNA processing and, if so, should specify the RNA transcripts from which they are derived. In order to investigate whether the short read patterns carry information on the particular ncRNA class from which they orginate, we performed a random forest classification on the three distinct ncRNA classes listed above. Then, after exploring the potential classification of general groups of ncRNAs, we focused on the identification of small RNA fragments derived from tRNAs. After mapping transcriptome sequence data to reference genomes, we searched for specific short read patterns reflecting tRNA processing. In this context, we devised a common tRNA coordinate system based on conservation and secondary structure information that allows vector representation of processing products and thus comparison of different tRNAs by anticodon and amino acid. We report patterns of tRNA processing that seem to be conserved across species. Though the mechanisms and functional implications underlying these patterns remain to be clarified, our analysis suggests that each type of tRNA exhibits a specific pattern and thus appears to undergo a characteristic maturation process.

Mass production of translationally optimized bacteriophages (hereafter referred to as phages) is the need of the hour in the application of phages to therapy. Understanding translational efficiency of phages is the major preliminary step for mass producing efficient phages. The objective of this thesis is to understand factors affecting translational efficiency of phages. In chapter two, we hypothesized that weak translation initiation efficiency is responsible for weak codon concordance of Escherichia coli lambdoid phages with that of their hosts. We measured the strength of translation initiation using two indices namely minimum folding energy (MFE) and proportion of Shine-Dalgarno sequence (PSD). Empirical results substantiate our hypothesis suggesting lack of strong selection for improving codon adaptation in these phages is due to their weak translation initiation. In chapter three, we measured codon usage concordance between GC-rich and GC-poor Aeromonas phages with their GC-rich host Aeromonas salmonicida. We found low codon usage concordance in the GC-poor Aeromonas phages. We were interested in testing for the role of tRNAs in the GC-poor phages. We observed that the GC-poor phages carry tRNAs for codons that are overused by the phages and underused by the host. These findings suggest that the GC-poor Aeromonas phages carry their own tRNAs for compensating for the compositional difference between their genomes and that of their host. Previously several studies have reported observed avoidance of stable secondary structures in start site of mRNA in a wide range of species. We probed the genomes of 422 phage species and measured their secondary structure stability using MFE. We observed strong patterns of secondary structure avoidance (less negative MFE values) in the translation initiation region (TIR) and translation termination region (TTR) of all analyzed phages. These findings imply selection is operating at these translationally important sites to control stable secondary structures in order to maintain efficient translation.

La terminaison de la traduction se produit lorsqu’un codon stop entre au site A du ribosome où il est reconnu par le facteur de terminaison eRF1 accompagné du facteur eRF3. Cette étape de la traduction est encore mal comprise chez les eucaryotes. Au cours de ma thèse je me suis intéressée à l’étude de la fidélité de la terminaison de la traduction afin de mieux comprendre et caractériser les mécanismes moléculaires mis en jeu lors du décodage du codon stop.L’un de mes projets consistait à mieux caractériser une région du domaine N-terminal d’eRF1, la cavité P1, identifiée comme étant impliquée dans l’efficacité de terminaison. Grâce à une quantification de l’efficacité de translecture de mutants de la cavité P1, j’ai pu mettre en évidence le rôle de résidus clés comme les serines 33 et 70, impliquées dans le décodage spécifique du codon UGA probablement via une interaction directe entre les deux résidus, ou encore l’arginine 65 et la lysine 109, essentielles pour une terminaison efficace sur les trois codons stop. L’analyse par RMN de ces mutants a également permis de montrer que ces résidus étaient importants pour la conformation correcte de la cavité et potentiellement impliqués dans une interaction directe avec l’ARNm. La combinaison des données génétiques et structurales nous a permis de proposer un modèle d’interaction entre l’ARNm et le facteur de terminaison eRF1 dans lequel le codon stop serait reconnu en partie par l’intermédiaire de la cavité P1. Dans la cellule, la terminaison est toujours en compétition avec la translecture, qui correspond à l’incorporation d’un ARNt proche-cognat au niveau du codon stop. Afin d’identifier les acides aminés incorporés par translecture au niveau du codon stop, j’ai mis au point un système basé sur l’expression et la purification de protéines issues de la translecture qui sont ensuite analysées par spectrométrie de masse. J’ai pu mettre en évidence que la glutamine, la tyrosine et la lysine s’incorporent au niveau des codons UAA et UAG, alors que le tryptophane, la cystéine et l’arginine sont retrouvés au niveau du codon UGA. J’ai également pu montrer que le contexte en 5’ n’influençait pas l’incorporation des acides aminés au codon stop mais qu’en revanche, la présence de la paromomycine avait un impact sur la sélection des ARNt suppresseurs naturels. Ce projet permet d’apporter de nouvelles informations sur les règles de décodage grâce à l’analyse des appariements entre codons stop et anticodons des ARNt naturels suppresseurs. Il permet également d’envisager des perspectives thérapeutiques dans le cadre des maladies liées à la présence d’un codon stop prématuré et pour lesquelles le traitement repose sur l’utilisation de la translecture afin de ré-exprimer des protéines entières. / Translation termination occurs when a stop codon enters the A site of the ribosome where it is recognized by eRF1 (eukaryotic release factor 1), associated with eRF3. This step of translation is not yet understood in eukaryotes. During my PhD, I was interested in studying translation termination accuracy to better understand and characterize the molecular mechanisms involved in stop codon decoding.One of my project consisted in characterizing a region in eRF1 N-terminal domain, pocket P1, identified to be involved in termination efficiency. Through a quantification of readthrough efficiency of pocket P1 mutants, I have highlighted the role of key residues, like serine 33 and serine 70, implicated in specific recognition of UGA stop codon, probably through a direct interaction between the two amino acids, and also arginine 65 and lysine 109, essential for efficient termination on the three stop codons. The analysis of the mutants by NMR revealed that these residues are also important for proper conformation of the cavity and potentially involved in a direct interaction with mRNA. The combination of our genetic data and structural analysis allowed us to propose a model of interaction between termination factor eRF1 and the mRNA, in which the stop codon would be recognized partially through pocket P1.In cells, termination always competes with readthrough which corresponds to the incorporation of near-cognate tRNAs at the stop codon. To identify the amino acids inserted by readthrough at the stop codon, I have developed a reporter system based on the expression and purification of readthrough proteins that are analyzed by mass spectrometry. I found that glutamine, tyrosine and lysine are inserted at UAA and UAG stop codons, whereas tryptophan, cysteine and arginine are inserted at UGA stop codon. I also showed that the 5’ nucleotide context does not influence the incorporation of amino acids at the stop codons by readthrough, but that, in contrast, the presence of paromomycin impacted the selection of natural suppressors tRNAs incorporated by readthrough. This project gives us new insights into the decoding rules by analyzing the base pairings between stop codon and near-cognates anticodons. It also allows us to consider therapeutic prospects for the treatment of premature stop codon diseases which uses readthrough as a tool to re-express full-length proteins from mRNAs that are interrupted by the presence of a premature stop codon.

The mitochondrial genome of the nematode Romanomermis culicivorax encodes for miniaturized hairpin-like tRNA molecules that lack D- as well as T-arms, strongly deviating from the consensus cloverleaf. The single tRNA nucleotidyltransferase of this organism is fully active on armless tRNAs, while the human counterpart is not able to add a complete CCA-end. Transplanting single regions of the Romanomermis enzyme into the human counterpart, we identified a beta-turn element of the catalytic core that—when inserted into the human enzyme—confers full CCA-adding activity on armless tRNAs. This region, originally identified to position the 30 -end of the tRNA primer in the catalytic core, dramatically increases the enzyme’s substrate affinity. While conventional tRNA substrates bind to the enzyme by interactions with the T-arm, this is not possible in the case of armless tRNAs, and the strong contribution of the beta-turn compensates for an otherwise too weak interaction required for the addition of a complete CCA-terminus. This compensation demonstrates the remarkable evolutionary plasticity of the catalytic core elements of this enzyme to adapt to unconventional tRNA substrates.

Les bactéries acidophiles du genre Acidithiobacillus joue un rôle important dans les activités industrielles de récupération des métaux au sein des sites miniers. Dans cette thèse, la séquence du génome de la bactérie psychro-tolerante Acidithiobacillus ferrivorans CF27 a été re-séquencée. L’analyse comparative du génome de CF27 et des autres bactéries du genre Acidithiobacillus a permis de montrer: (i) une synthénie conservée entre 2 clusters de tRNAs trouvés dans les génomes de At. ferrivorans CF27 et At. ferrooxidans ATCC 23270, et qui ont contribué à la redondance génique des tRNAs chez ces 2 organismes. Notre analyse in silico à grande échelle de ces clusters de tRNAs au sein des génomes procaryotes a montré que les clusters de tRNAs sont présents dans très peu de phyla bactériens; (ii) la présence d’une importante proportion de gènes spécifiques chez CF27 et SS3, ce qui indique la très grande variabilité du contenu génique dans les génomes d’Acidithiobacillus et ainsi la nature unique de chaque groupe d’espèces. L’expression de ces gènes spécifiques a été confirmée chez CF27 cultivés en présence de Fer et soufre; et (iii) une composition taxonomique chimérique des génomes de la classe des Acidithiobacillia, confirmant ainsi que ce groupe appartient à une classe taxonomique particulière. Ces résultats apporte de nouvelles connaissances sur l’adaptation de CF27 à son environnement, ainsi que la nature chimérique des génomes de la classe taxonomique Acidithiobacillia. J’ai participé au projet ‘Thioredoxine réductase (TR)’ dont l’objectif est de définir la fonction biochimique, la structure moléculaire, ainsi que l’histoire évolutive de TRi, une réductase atypique. / The acidophilic Acidithiobacillus bacteria play an important role in industrial biomining operations for metal recovery. In this thesis, the genome sequence of a psychrotolerant Acidithiobacillus ferrivorans CF27 were first refined. The comparative genome analysis between CF27 and the closely related Acidithiobacillus genomes revealed: (i) a syntenic conservation of two tRNA array units which are only present in At. ferrivorans CF27 and At. ferrooxidans ATCC 23270 genomes and mainly contribute to the tRNA gene redundancy in both organisms. Moreover, our large-scale genome survey of the tRNA array units in prokaryotic organisms showed that tRNA arrays appear in few phyla; (ii) a high proportion of species-specific genes in CF27 and SS3 strains indicated the high variability of gene content in Acidithiobacillus genomes and therefore the unique nature of each group of species. Given that mRNA expression of some CF27 specific genes were confirmed in Fe(II)-grown cells and sulfur attached cells in CF27, these results highlighted the functional importance of specific genes for CF27 lifestyle; and (iii) the mosaic taxonomic composition of members of the Acidithiobacillia class, and thus confirmed that this group belongs to a particular taxonomic class, distinct to other proteobacterial groups. Taken together, our results provide insights into At. ferrivorans lifestyle as well as the chimeric genome nature of the Acidithiobacillus organisms. In addition, I also participated to the ‘Thioredoxin reductase’ project which aims to define the biochemical function, molecular structure and evolution of TRi, an atypical thioredoxin reductase.