A short oligonucleotide sequence as in a single-stranded antisense oligo nucleotides (AON) or in double-stranded small interfering RNAs (siRNA) can modulate the gene expression by targeting against the cellular mRNA, which can be potentially exploited for therapeutic purposes in the treatment of different diseases. In order to improve the efficacy of oligonucleotide-based drugs, the problem of target affinity, nuclease stability and delivery needs to be addressed. Chemical modifications of oligonucleotides have been proved to be an effective strategy to counter some of these problems. In this thesis, chemical synthesis of conformationally constrained nucleosides such as 7′-Me-carba-LNA-A, -G, -MeC and -T as well as 6′, 7′-substituted α-L-carba-LNA-T (Papers I-III) was achieved through a key free-radical cyclization. 1D and 2D NMR techniques were employed to prove the formation of bicyclic ring system by free-radical ring closure as well as to identify the specific constrained conformations in sugar moieties. These sugar-locked nucleosides were transformed to the corresponding phosphoramidites and incorporated into antisense oligonucleotides in different sequences, to evaluate their physicochemical and biochemical properties for potential antisense-based therapeutic application. AONs modified with 7′-Me-carba-LNA analogues exhibited higher RNA affinities (plus 1-4°C/modification) (Papers I & III), but AONs containing α-L-carba-LNA analogues showed decreased affinities (minus 2-3°C/ modification) (Paper II) towards complementary RNA compared to the native counterpart.  It has been demonstrated in Papers I-III that 7′-methyl substitution in α-L-carba-LNA caused the Tm drop due to a steric clash of the R-configured methyl group in the major groove of the duplex, whereas 7′-methyl group of carba-LNA locating in the minor groove of the duplex exerted no obviously negative effect on Tms, regardless of its orientation. Moreover, AONs containing 7′-Me-carba-LNA and α-L-carba-LNA derivatives were found to be nucleolytically more stable than native AONs, LNA modified AONs as well as α-L-LNA modified ones (Papers I-III). We also found in Paper II & III that the orientations of OH group in C6′ of α-L-carba-LNAs and methyl group in C7′ of 7′-Me-carba-LNAs can significantly influence the nuclease stabilities of modified AONs. It was proved that the methyl substitution in cLNAs which points towards the vicinal 3′-phosphate were more resistant to nuclease degradation than that caused by the methyl group pointing away from 3′-phosphate. Additionally, AONs modified with 7′-Me-carba-LNAs and α-L-carba-LNAs were found to elicit the RNase H mediated RNA degradation with comparable or higher rates (from 2-fold to 8-fold higher dependent upon the modification sites) as compared to the native counterpart. We also found that the cleavage patterns and rates by E. coli RNase H1 were highly dependent upon the modification sites in the AON sequences, regardless of the structural features of modifications (Papers II & III). Furthermore, we have shown that the modulations of Tms of AON/RNA duplexes are directly correlated with the aqueous solvation (Paper III).

Dans les dernières années, une explosion de la recherche sur les ARN a eu lieue à cause de nombreuses découvertes démontrant l’importance de l’ARN dans plusieurs processus biologiques. Ainsi, de grandes quantités d’ARN sont devenues indispensables au bon déroulement de plusieurs études, notamment pour la biologie structurale et la caractérisation fonctionnelle. Cependant, il existe encore peu de méthodes de purification simples, efficaces, fiables et produisant un ARN sous forme native. Dans les dernières années, le laboratoire Legault a mis au point une méthode de purification par affinité utilisant une étiquette ARiBo pour la purification d’ARN transcrits in vitro par la polymérase à ARN du phage T7. Cette méthode de purification d’ARN a été spécifiquement développée pour maximiser la pureté et le rendement. De plus, elle est très rapide et fonctionne avec plusieurs types d’ARN. Cependant, comme plusieurs autres méthodes de purification, cette méthode produit des ARN avec des extrémités 5′ hétérogènes. Dans ce mémoire, des solutions sont proposées pour remédier au problème d’hétérogénéité en 5ʹ′ des ARN transcrits avec la polymérase à ARN du phage T7 et purifiés par la méthode ARiBo. La première solution consiste à choisir la séquence en 5′ parmi celles des 32 séquences testées qui ne présentent pas d’hétérogénéité en 5ʹ′. La seconde solution est d’utiliser une étiquette clivable en 5ʹ′ de l’ARN d’intérêt, tel que le ribozyme hammerhead, déjà utilisée pour ce genre d’application, ou le système CRISPR/Cse3 que nous proposons dans l’article présenté dans ce mémoire. De plus, nous avons adapté la méthode ARiBo pour rendre possible la purification d’un long ARN de 614 nt, le polycistron miR-106b-25. Nous avons également démontré la possibilité d’utiliser la méthode ARiBo pour l’isolation de protéines qui se lient à un ARN donné, le précurseur de miRNA pre-miR-153-2. En conclusion, ce mémoire démontre la possibilité d’adapter la méthode ARiBo à plusieurs applications. / In recent years, the field of RNA research has exploded due to several discoveries demonstrating the importance of RNA in many biological processes. Along with the increased interest in this field, large amounts of RNA have become essential to the success of several studies, in particular for structural biology and functional characterization. However, there are still very few native purification methods that are simple, efficient and reliable. In the past few years, the Legault laboratory has established an affinity purification method using an ARiBo tag to purify RNAs produced by in vitro transcription with the T7 RNA polymerase. This RNA purification method was specifically developed to maximise purity and yield. In addition, this method is fast and works with several types of RNAs. However, like several other purification methods, this method produces RNAs with 5' heterogeneity. This Master’s thesis propose solutions to overcome the problem of 5' heterogeneity for RNAs transcribed with the T7 RNA polymerase and purified with the ARiBo method. The first solution proposed is to choose a 5' sequence among those of the 32 sequences tested that do not present 5'- heterogeneity. The other possibility is the use of a cleavable tag at the 5'-end of the RNA of interest, such as the hammerhead ribozyme, already used for this purpose or the CRISPR/Cse3 system, which is presented here. Furthermore, we have adapted the ARiBo method to purify an RNA of 614 nt, the miRNAs cluster miR- 106b-25. We also demonstrate the possibility to use the ARiBo method to isolate proteins that bind a given RNA, the miRNA precursor pre-miR-153-2. In conclusion, this Master’s thesis demonstrates the possibility of adapting the ARiBo method for several applications.