Epissage d'ARN et d'ADN : comment la Paramécie annote son génome – ParaDice

Most eukaryotic genes are interrupted by non-coding introns that must be accurately removed from pre-messenger RNAs to produce translatable mRNAs. Splicing is guided locally by short conserved sequences, but genes typically contain many potential splice sites, and the mechanisms specifying the correct sites remain poorly understood. In most organisms, introns cannot be efficiently predicted solely on the basis of sequence motifs. Paramecium tetraurelia is no exception: its genome contains approximately 90,000 very short introns defined by very weak cis-acting signals. A recent statistical analysis of these introns has shown that 'readthrough' introns (introns that would not interrupt the reading frame if retained in messenger RNAs) are strongly counter-selected during evolution. The bias, which turned out to be universally observed in intron-rich genomes, would suggest that splicing is, in fact, relatively inefficient and that cells rely on Nonsense-Mediated mRNA Decay (NMD) to select translatable mRNAs. However, the bias would also be consistent with the controversial idea that translatability can influence splice site choice. In addition to RNA splicing, ciliates also undergo massive DNA splicing during the development of the somatic nucleus from the germline. In P. tetraurelia, this process includes the precise excision of approximately 60,000 Internal Eliminated Sequences (IESs) that are dispersed throughout the genome and very frequently interrupt the open reading frames of protein-coding genes. Like introns, IESs are single-copy sequences defined by very weak cis-acting signals that are not sufficient to specify the excision pattern across the genome. Recent advances have shown that a specialized RNA interference pathway, producing massive amounts of small RNAs (scnRNAs) from the germline during meiosis, is involved in their recognition. It has been suggested that these scnRNAs mediate a natural genomic subtraction that allows rearrangement patterns to be simply copied from the previously rearranged maternal somatic genome at each sexual generation. One of the main objectives of our project is to unravel the precise mechanisms by which additional sources of information are used to achieve specific recognition of large numbers of different sequences with only weak cis-acting signals. In the case of introns, we will test whether the observed translatability bias simply serves to correct the output of inefficient splicing by NMD, or whether the detection of in-frame stop codons (by NMD or some other mechanism) can also influence splice site choice and/or splicing efficiency. In the case of IESs, the genome scanning model remains to be experimentally validated by examining the fate of scnRNA subsets matching somatic and germline-specific sequences during sexual processes. Elucidation of this mechanism would be of general interest because of the intriguing similarities between ciliate scnRNAs and the recently discovered piRNAs of metazoans. Our project encompasses sequencing of the MIC genome, since only a very small fraction of IESs and other germline-specific sequences is currently known. Large-scale bioinformatic analyses will identify constraints on the evolution of IES sequences, providing useful insights into the mechanisms of their recognition and excision. The proposed experiments will allow us to sequence most IESs and to evaluate the complexity of repeated sequences eliminated by the imprecise mechanism. The first description of genome-wide rearrangements in a ciliate should be of fundamental interest to understand how these organisms have coped with transposable elements. Since our ultimate goal is to understand the genome rearrangements that occur during development, the present project also addresses the molecular mechanisms of IES excision and imprecise elimination of repeated sequences. We will investigate the role of a domesticated piggyBac transposase, which appears to be a promising endonuclease candidate fo