WO2010084371A1 - Novel circular interfering rna molecules - Google Patents

Abstract

A circular interfering RNA (ciRNA) molecule comprising the sense and the antisense strands of an interfering RNA molecule targeting a gene of interest, wherein said sense and antisense strands are closed at one end by a loop structure and at the other end by a splice junction sequence generated by splicing activity of permuted intron-exon sequences, and wherein the circular RNA molecule inhibits the expression of the gene of interest in cells expressing said gene of interest. Genetic constructs and transcription vectors for the production of said ciRNA. Method for producing large amounts of said ciRNA in yeast mitochondria in a form which is stable and easy to purify.

Description

NOVEL CIRCULAR INTERFERING RNA MOLECULES

The invention relates to novel circular interfering RNA molecules (ciRNA) which can be produced by in vitro or in vivo transcription. The invention relates to genetic constructs and transcription vectors useful for the production of said circular interfering RNAs. The invention relates also to a method for producing large amounts of said circular interfering RNAs in yeast in a form which is stable and easy to purify.

RNA interference (RNAi) is a post-transcriptional gene-silencing mechanism in which double-stranded RNA (dsRNA) triggers degradation of homologous messenger RNA in the cytoplasm (reviewed in Shuey et al, Drug Discovery Today, 2002, 7, 1040-1046 and Tuschl T. and Borkhardt A., MoI. Interv., 2002, 2, 158-167). In vivo, long dsRNAs are cleaved by the RNase III class endoribonuclease Dicer into 21-23 base duplexes having 2-base 3 '-overhangs and 5'- terminal phosphate groups. These species called "small interfering RNAs" (siRNAs) bind to a ribonuclease complex called RNA-induced silencing complex (RISC) that guides the small siRNA to its homologous mRNA target. Consequently, RISC cuts the mRNA approximately in the middle of the region paired with the antisens siRNA, after which the mRNA is further degraded. Target recognition is highly sequence specific since one or two base pair mismatches between the siRNA and the target gene will greatly reduce silencing effect. In addition, it was shown that 27mer RNA duplexes can be up to 100-fold more potent than traditional 21mer duplexes (Kim et al, Nat. Biotechnol., 2005, 23, 222-226). Furthermore, combination of asymmetric 2- base 3 '-overhang on the antisense strand with 3'-DNA residues on the blunt end (R 25D/27 duplex) result in a duplex form which directs dicing to predictably yield a single primary cleavage product (Rose et al, Nucleic Acids Res., 2005, 33, 4140- 4156).

RNAi is an important mechanism for the regulation of gene expression in a broad range of eukaryotic organisms, including both plants (Van Der Krol et al, Plant Cell., 1990, 2, 291-299) and animals (Fire et al, Nature, 1998, 391, 806-81 1).

The use of exogenous siRNA has also become a powerful tool in functional genomics to knock down single genes for detailed study or hundreds to thousands of genes in high-throughput functional genomic surveys. The potential of siRNAs for use as therapeutic agents to reduce activity of specific gene products is also receiving considerable attention. Today, around fifty siRNAs are in pre-clinical phase. Silencing can be induced in target cells by directly introducing synthetically produced siRNAs or short hairpin RNAs (shRNAs), or alternatively by transfecting cells with engineered plasmid or viral vectors expressing siRNAs or shRNAs under the transcriptional control of RNA polymerase II or III promoter (U6 and Hl). Short hairpin RNAs are siRNA derived molecules that contain the sense strand and the antisense strand of siRNA molecules and a short loop sequence between the sense and antisense strands. As opposed to siRNAs which are formed of two separate RNA molecules, shRNAs are formed of one RNA molecule that folds into a hairpin. shRNAs are converted into small interfering RNA (siRNA) by the cellular ribonuclease III, Dicer (Dykxhoorn et al, Nat. Rev. MoI. Cell. Biol., 2003, 4, 457-467; Hammond, S.M., FEBS Lett., 2005, 579, 5822-5829; Hannon, G.J. and Rossi, J.J., Nature, 2004, 431, 371-378; Rossi J.J., Hum. Gene Ther., 2008, 19, 313- 317).

Although plasmids or viral vectors are highly effective at expressing siRNAs or shRNAs, in vitro in cell culture as well as in vivo in laboratory animals, they cause safety and ethic problems in clinical applications. The vector DNA may be inserted in the chromosomal DNA and thus induce alteration of this DNA which may lead to cancer by activation of cellular oncogenes by insertional mutagenesis. Therefore, synthetic siRNAs or shRNAs are considered as the RNAi molecules of choice for clinical purposes. There are currently two methods for producing synthetic siRNAs and shRNAs: (1) chemical synthesis, and (2) in vitro transcription of linear "run-off templates or small single-stranded DNA circles by bacteriophage RNA polymerases (RNAP) from their cognate promoters (T7, T3 or SP6; Seyhan et al., oligonucleotides, 2006, 16, 353-363). Chemical synthesis is very costly. In vitro transcription by phage

RNA polymerase from their natural promoters results in 5'-termini triphosphate that can trigger an interferon response in vivo. In addition, due to the requirement of phage promoters for 5'-GpuPu sequences for transcription initiation, shRNA transcripts may have extra 5 '-nucleotides that can constrain the sequences that can be targeted. Also, the 3' ends may have an additional n+1 nucleotide not encoded by the template. In addition efficient transcription of siRNAs requires two separate dsDNA templates and four oligodeoxynucleotides must be synthesized for each siRNA duplex. Many of the problems encountered with in vitro transcription of linear "run off templates may be circumvented by in vitro transcription of multimeric shRNAs generated by rolling circle transcription of small circular single-stranded DNAs lacking promoter, primer or terminator sequences using T7 polymerase (Seyhan et al, oligonucleotides, 2006, 16, 353-363).

In addition, to the two preceding methods, the European Patent EP 1646724 describes a method for producing a heterologous RNA of interest including a siRNA, which uses yeast lacking mitochondrial DNA {rho ) whose mitochondria are transformed with a DNA encoding the heterologous RNA of interest (synthetic rho ) for producing the RNA of interest in their mitochondria. This RNA is readily isolated in a stable form and in large amounts, from the mitochondria of the synthetic rho^' strain, insofar as the only RNAs produced in the mitochondria of said synthetic rho^' strain are those which are encoded by the DNA used for the transformation.

However, the synthetic RNA molecules made of natural (unmodified) nucleotides which are produced by the preceding methods are not stable in biological fluids since their extremities are rapidly degraded by 3'-exonuclease, one of the major enzymes involved in the degradation of nucleic acids in vivo. Therefore, the production of stable synthetic siRNAs is required for therapeutic applications.

Chemically synthesized siRNA can be stabilized by the modification of the terminal ribose moieties with a 2'-deoxy, 2'-O-methyl or 2'-fluoro group or replacement of the terminal phosphodiester bonds with phosphorotioates. However, the chemically modified siRNAs are costly, they are often less active than their unmodified counterpart and they may be toxic to human.

Recently, a dumbbell-shaped circular RNA that contains a 15 to 23 bp stem sequence encoding the firefly luciferase gene (siRNA sequence) and two 9- mer loops was synthesized from two RNA strands that were closed at both ends with the loop sequences, using T4 RNA ligase (Abe et al, J. Am. Chem. Soc, 2007, 129, 15108-15109). The RNA dumbbell is resistant to degradation in serum, because of the shape of the molecule, an endless structure that cannot be degraded by 3'-exonuclease. Dumbbells with stem length longer than 23 bp are specifically recognized and cleaved efficiently by the human Dicer enzyme to generate siRNA that has a prolonged RNAi activity. However, this dumbbell-shaped circular RNA is costly since the two RNA strands are synthesized by chemical synthesis and subsequently closed at both ends using T4 RNA ligase. In addition, this RNA dumbbell may suffer from a lack of reproducibility since the ligation step may not always be completely efficient.

Therefore, there is a need for novel circular RNA molecules encoding interfering RNA molecules (circular RNAi or ciRNA) that can be produced by methods which are less expensive and more reproducible.

RNA splicing reactions can generate circular exon sequences when the 3' end of the exon is joined to a splice site at an upstream rather than a downstream position (donor (5') splice site is 3' of the acceptor (3') splice site ; figure 1). Circular RNAs generated by splicing have been demonstrated with in vitro manipulated Group I and Group II intron sequences.

Group I catalytic introns are self-splicing ribozymes. Their size is variable (from 68 over 300 nucleotides; most are over 400 nucleotides). They catalyze their own excision from mRNA, tRNA and rRNA precursors in bacteria, lower eukaryotes and higher plants. However, their occurrence in bacteria seems to be more sporadic than in lower eukaryotes, and they have become prevalent in higher plants. The genes that group I introns interrupt differ significantly: they interrupt rRNA, mRNA and tRNA genes in bacterial genomes, as well as in mitochondrial and chloroplast genomes of lower eukaryotes, but only invade rRNA genes in the nuclear genome of lower eukaryotes. In higher plants, these introns seem to be restricted to a few tRNA and mRNA genes of the chloroplast and mitochondria. In Group I introns, RNA forms the active site and directs the cleavage and ligation reaction at the 5' and 3' splice sites. Splicing is initiated with an external G nucleotide (cofactor) and processed by two sequential transesterification steps catalysed by RNA in vitro. The catalytic core of Group I introns consists of two structural domains, P4-P6 and P3-P9. A portion of P4-P6 has been proposed to interact with the substrate helix Pl which contains the 5' splice site, and P3-P9 contains the binding site for the guanosine nucleophile. When provided as separate molecules, the P4-P6 and P3-P9 domains of the Tetrahymena Group I intron can self-assemble into an active structure (Tanner et al, Science, 1997, 275, 847-849). Most Group I introns are able to splice themselves in the absence of proteins, i.e. the RNA itself is catalytic (ribozyme). However, not all group I introns are truly catalytic. Splicing of some group I introns in vivo is modulated by a number of proteins which play a role in folding and 3D structure (maturases) and are encoded by various genes independent from the introns or by the introns themselves. Group I catalytic intron and flanking 5' and 3' exons sequences are available in the sequence data base. For example, the cyanobacteria Anabaena sp. pre-tRNA-Leu gene sequence including the self-splicing Group I intron sequence (Xu et al, Science, 1990, 250, 1566-1570) corresponds to the accession numbers GenBank M38962 and M38961. The Tetrahymena ribosomal Group I intron sequence (399 bases in its circular form) is described for example in figure 1 of G. Dinter-Gottlieb, Proc. Natl. Acad. Sd., USA, 1986, 83, 6250-6254. Using Group I self-splicing permuted intron-exon (PIE) sequences in which the order of the splice sites has been reversed, it is possible to catalyze the circularization of a variety of RNA exon sequences in vitro, as well as in vivo in bacteria and in yeast (Puttaraju, M. and Been, M.D., Nucleic Acids Res., 1992, 20, 5357-5364 and GenBank X69005). This circularization is accomplished only by the self-splicing activities of Group I intron sequences and the mechanism is a host- independent process. The RNA molecule to be made circular is prepared by inserting a DNA sequence encoding the RNA of interest in a DNA construct comprising the following sequences from a Group I self-splicing intron (figure 1) : (1) a sequence encoding a 3' portion of the intron, (2) a sequence encoding the 3' splice site, (3) a sequence encoding the RNA of interest inserted at the fusion point of the 3 'exon fused end-to-end to the 5 'exon or of a fragment of the fused exons having few nucleotides (less than 50, preferably less than 20 nucleotides) of the 3' exon (Exon 2) sequence flanking the 3' splice site and few nucleotides of the 5' exon (Exon 1) sequence flanking the 5' splice and eventually a cloning site inserted at the junction of the fused exon sequences, (4) a sequence encoding the 5' splice site, and (5) a sequence encoding a 5' portion of the intron. The DNA construct is cloned in an expression plasmid under the transcriptional control of an appropriate promoter. Transcription from the promoter, results in the production of an autocatalytic messenger RNA (mRNA circular RNA precursor) which generates a circular form of the RNA of interest by self-splicing.

A permuted Anabaena-deήved Group I intron was used to generate circular hepatitis delta virus ribozyme and a circular form of Bacillus subtilis RNaseP in vitro and in E.coli (US Patent 5,712,128 by Been et al; Puttaraju et al, Nucleic

Acids Res., 1993, 21, 4253-4258; Puttaraju et al, Nucleic Acids Symposium Ser.,

1995, 33, 92-94; Puttaraju, M. and Been, M.D., The Journal of Biological Chemistry,

1996, 42, 26081-26087), as well as to generate a circular TAR RNA decoy in vitro (Bohjanen et al, Nucleic Acids Res., 1996, 24, 3733-3738).

A permuted bacteriophage T4-derived Group I intron was also used to generate circular messenger RNA sequences in vitro, in E.coli and in yeast (Ford E. and Ares M., Proc. Natl. Sci. USA, 1994, 91, 3117-3121; US Patent 5,773,244 by Ares M. and Ford, E.E.) as well as a circular form of streptavidin RNA aptamer in vitro (Umekage, S. and Kikuchi, Y., Nucleic Acids Symposium Series N°50, 2006, 323-324).

The circular RNA molecules generated by self-splicing of Group I permuted intron-exon (PIE) sequences comprise exon sequences flanking the splice junction; for example in the case of Anabaena-derived Group I intron the circular RNA comprises at least a 7 bp stem terminated by a 7 nucleotide loop (see for example, figure 1 of Puttaraju et al, Nucleic Acids Res., 1993, 21, 4253-4258; figure 1 of Bohjanen et al, Nucleic Acids Res., 1996, 24, 3733-3738 or figure 2 of Puttaraju, M. and Been, M.D., The Journal of Biological Chemistry, 1996, 42, 26081-26087).

To date, the circular RNA molecules which have been generated are enzymatic molecules (ribozyme, RNase P) or protein ligands (sptreptavidine, Tat). In these circular RNA molecules, the secondary structures from the Group I PIE splice- junction sequences were added at the end of a stem comprising the 5' and 3' ends of the RNA molecules. In this design, the structure and the function of the RNA molecule were not altered substantially since the secondary structures from the Group I PIE splice-junction sequences were added in a region distant from the active site (cleavage or binding site). However, these secondary structures from the Group I PIE splice- junction sequences have never been introduced in a RNAi molecule, as a strategy for making circular RNAi molecules. It is not known whether a stable circular RNAi molecule could be generated by self-splicing of Group I permuted intron-exon (PIE) sequences. It is not known also if such circular RNAi molecule comprising secondary structures from the Group I PIE splice-junction sequences would be cleaved efficiently by the Dicer enzyme to achieve RNAi effect in cells.

Furthermore, it is not known whether, the method described in the EP Patent 1646724 could be used to generate circular RNAi molecules by self- splicing of Group I PIE in the mitochondria of yeasts lacking mitochondrial DNA (rho°). First, the self-splicing activity of Group I PIE in the mitochondria of rho⁰ has never been demonstrated before. It is not known if the conditions which are essential for self-splicing (pH, GTP and RNA concentrations) are present in the mitochondria. Secondly, the stability of circular RNA in mitochondria lacking mitochondrial DNA is questionable since the transcription and the regulation of the pool of mRNA are not optimal in this type of mitochondria.

The inventor has designed a circular siRNA (ciRNA) molecule that contains a siRNA sequence targeting the luciferase gene closed at one end by a small loop and at the other end by the stem and the loop derived from Anabaena Group I PIE splice-junction sequences (figure 2A). The ciRNA molecule was generated by in vitro or in vivo transcription (bacteria, yeast) from a transcription plasmid comprising a T7 promoter operatively linked to a DNA construct derived from the Anabaena Group I PIE constructs previously described by Puttaraju et al. (Puttaraju, M. and Been, M.D., Nucleic Acids Res., 1992, 20, 5357-5364; Puttaraju et al, Nucleic Acids Res., 1993, 21, 4253-4258; Puttaraju, M. and Been, M.D., The Journal of Biological Chemistry, 1996, 42, 26081-26087). The ciRNA molecule is stable (resistant to 3'- Exonuclease digestion; figure 6) and has an RNAi activity which is at least equivalent to that of the corresponding shRNA molecule (figure 9).

The invention relates to a circular RNA molecule (ciRNA) which comprises the sense and the antisense strands of an interfering RNA molecule (RNAi) targeting a gene of interest, wherein said sense and antisense strands are closed at one end by a loop structure (first loop) and at the other end by a splice junction sequence generated by splicing activity of permuted intron-exon sequences, and wherein the circular RNA molecule inhibits the expression of the gene of interest in cells expressing said gene of interest.

Definitions - "interfering RNA molecule or RNAi molecule" refers to a double- stranded RNA molecule of at least 19 bp, comprising complementary sense and antisense strands, wherein the antisense strand is complementary to the sequence of a target gene and the introduction of the RNAi molecule in cells expressing the target gene inhibits the expression of said target gene. The interfering RNA molecule may be long ( > 100 pb) or short (small; < 100 pb); small interfering RNA (siRNA) molecules consist preferably of 19 to 30 bp.

- "complementary" refers to the ability of a nucleic acid to form hydrogen bond(s) by either traditional Watson-Crick base-pairing or other non- traditional type base-pairing. In reference to the nucleic acid molecules of the present invention, the binding free energy for a nucleic acid molecule with its complementary sequence is sufficient to allow the relevant function of the nucleic acid to proceed, e.g., RNAi activity. Determination of binding free energies for nucleic acid molecules is well-known in the art (see, e.g., Turner et al, 1987, CSH Symp. Quant. Biol., 1987, LII, pp 123-133, Frier et al, P.N.A.S., 1986, 83, 9373-9377; Turner at al., J Am. Chem. Soc, 1987, 109, 3783-3785). A percent complementarity indicates the percentage of contiguous residues in a nucleic acid molecule that can form hydrogen bonds (e.g., Watson-Crick base-pairing) with a second nucleic acid sequence (e.g., 5, 6, 7, 8, 9, or 10 nucleotides out of a total of 10 nucleotides, in the first oligonucleotide being base- paired to a second nucleic acid sequence having 10 nucleotides represents 50 %, 60 %, 70 %, 80 %, 90 % and 100 % complementarity, respectively). "Perfectly complementary" means that all the contiguous residues of a nucleic acid sequence will hydrogen bond with the same number of contiguous residues in a second nucleic acid sequence.

- "few nucleotides" refers to less than 50 nucleotides, preferably less than 20 nucleotides, more preferably around 15 nucleotides.

- "target gene" refers to a gene whose expression is to be down- regulated. - "vector" refers to a nucleic acid molecule capable of transporting another nucleic acid to which it has been linked.

The inhibition of the target gene expression by the circular RNA molecule of the invention is assayed at the RNA or protein level, by methods well- known in the art, for example by real time quantitative RT-PCR, Northern-blot, FACS or Western-blot.

The portion of the ciRNA molecule comprising the sense strand of the RNAi molecule, the first loop and the antisense strand of the RNAi forms a shRNA. Therefore, this portion of the ciRNA molecule is also referred to as shRNA. According to a preferred embodiment of said ciRNA molecule, the sense and antisense strands of the RNAi molecule comprise at least 19 ribonucleotides, for example 21 to 27 (e.g. 21, 22, 23, 25, 26 or 27) ribonucleotides.

The RNAi molecule may have overhanging ribonucleotide(s) at one or both end(s), preferably, 1 to about 5 (e.g. about 1, 2, 3, 4, 5) overhanging ribonucleotides. The overhanging ribonucleotides which are advantageously at the 3 'end of the antisense strand, are preferably uridines.

The splice junction sequence may be generated by splicing activity of any permuted intron-exon sequence consisting of : (1) a portion of the 3' half of an intron sequence including the 3' splice site, (2) exon sequence (sequences flanking said intron or an exogenous sequence) and (3) a portion of the 5' half of said intron sequence including the 5' splice. According to a preferred embodiment of said ciRNA, the splice junction sequence is generated by splicing activity of permuted intron- sequences derived from an intron of a yeast mitochondrial gene. For example, from Saccharomyces cerevisiae mitochondrial 21 S rRNA gene, rl intron encoding a putative transposase (GenBank accession number Ml 1280) or S. cerevisiae COXl gene Group I al 4 intron. According to another preferred embodiment of said ciRNA, the splice junction sequence is generated by self-splicing activity of permuted intron-exon sequences. Preferably, the permuted intron-exon sequences are derived from a Group I catalytic intron, such as for example the Group I catalytic intron of the cyanobacteria Anabaena sp. pre-tRNA-Leu gene or the Tetrahymena ribosomal gene. The splice junction sequence may comprise or consist of a loop structure (second loop) which is similar or identical to that of the first loop. For example, the splice junction sequence may form a loop structure (Figure 2A) or it may comprise a stem-loop structure, derived for example from Anabaena sp. Group I PIE splice-junction sequence (figure 2B). In this case, the strands of the RNAi molecule are connected to the strands of the stem-loop structure, either directly or by 3 to 10 non- complementary nucleotides forming a bulge (figure 2B and SEQ ID NO: 2).

The loop(s) (first and/or second loop) which may be identical or different comprise at least 3 ribonucleotides, for example about 3, 4, 5, 6, 7, 8, 9, or 10 ribonucleotides. According to another preferred embodiment of said ciRNA, the loop(s) comprise at least 5 ribonucleotides. For example, the loop(s) comprise the sequence SEQ ID NO: 1.

The invention relates also to a genetic construct (DNA construct; figure 3) designed for the production of the ciRNA molecule of the invention, said genetic construct comprising: (1) a transcription initiation region (promoter), (2) a sequence encoding a 3' portion of a self-splicing Group I intron (Intron 2), (3) a sequence encoding the 3' splice site of said self-splicing Group I intron, (4) a sequence encoding the shRNA of the invention inserted at the junction (fusion point) of the 3'exon (Exon 2) of said self-splicing Group I intron fused end-to-end to the 5'exon (Exon 1) of said self-splicing Group I intron or of a fragment of the fused exons having few nucleotides of the 3' exon sequence flanking the 3' splice site and few nucleotides of the 5' exon sequence flanking the 5' splice and eventually a cloning site inserted at the junction of the fused exon sequences, (5) a sequence encoding the 5' splice of said self-splicing Group I intron (Intron 1), (6) a sequence encoding a 5' portion of said self-splicing Group I, and (7) a transcription termination region. The sequences (2) to (6) of the construct are operatively linked to the transcription initiation (1) and termination (7) regions. In vitro or in vivo transcription from the promoter produces an autocatalytic unspliced RNA (ciRNA RNA precursor or ciRNA precursor) which generates a circular form of the siRNA of interest by self-splicing. The genetic construct comprises advantageously restriction sites 5' and 3' of the promoter and transcription termination regions (figure 3). The DNA construct of the invention may be derived from the Group I PIE constructs previously described (Puttaraju, M. and Been, M.D., Nucleic Acids Res., 1992, 20, 5357-5364; Puttaraju et al, Nucleic Acids Res., 1993, 21, 4253-4258; Puttaraju, M. and Been, M.D., The Journal of Biological Chemistry, 1996, 42, 26081- 26087; US Patent 5,712,128 ; Puttaraju et al, Nucleic Acids Symposium Series, 1995, 33, 92-94; Ford E. and Ares M., Proc. Natl. Sci. USA, 1994, 91 , 31 17-3121 ; US Patent 5,773,244; Umekage, S. and Kikuchi, Y., Nucleic Acids Symposium Series N°50, 2006, 323-324; Bohjanen et al, Nucleic Acids Res., 1996, 24, 3733-3738). Alternatively, other Group I PIE constructs may be derived from other Group I autocatalytic introns, using the same strategy.

According to a preferred embodiment of said genetic construct, the self-splicing Group I intron is from the cyanobacteria Anabaena sp. pre-tRNA-Leu gene. Preferably, the portion of the genetic construct which starts immediately downstream of the promoter and terminates immediately upstream of the sequence encoding the shRNA of the invention comprises the sequence SEQ ID NO: 3 and the portion of the genetic construct which starts immediately downstream of the sequence encoding the shRNA of the invention and terminates immediately upstream of the transcription termination region comprises the sequence SEQ ID NO: 4. Therefore, the sequence encoding the shRNA of the invention is inserted between the sequences SEQ ID NO: 3 and SEQ ID NO: 4.

The transcription initiation region may be from RNA polymerase promoters from the host-cell that is used for producing the RNA, including for example prokaryotic RNA polymerase promoters and eukaryotic RNA polymerase I, II or III (pol I, II or III) promoters. Transcripts from pol II or pol III promoters are expressed at high levels in all cells. Transcription units derived from genes encoding U6 small nuclear transfer RNA and adenovirus VA RNA are useful in generating high concentrations of desired ciRNA in cells.

The promoter may be constitutive or inducible. Examples of inducible promoters are: eukaryotic metal lothionine promoter which is induced by increased levels of heavy metals, prokaryotic lacZ promoter which is induced in response to isopropyl-β-D-thiogalacto-pyranoside (IPTG) and eukaryotic heat shock promoter which is induced by increased temperature. Alternatively, the transcription initiation region may be from an exogenous RNA polymerase (T7, T3 or SP6) promoter, providing that said exogenous RNA polymerase enzyme is expressed in the host cell in which the expression vector is introduced. According to another preferred embodiment of said genetic construct, the ciRNA RNA precursor is under the transcriptional control of the bacteriophage T7 RNA polymerase promoter and terminator sequences. Examples of said sequence are SEQ ID NO: 5 and SEQ ID NO: 6, respectively for the T7 RNA polymerase promoter and the T7 terminator. According to another preferred embodiment of said genetic construct, the ciRNA RNA precursor is under the transcriptional control of transcription initiation and termination regions which are functional in yeast mitochondria.

According to another preferred embodiment of said genetic construct, it comprises the COX2 3'UTR sequence (SEQ ID NO: 7) which includes the dodecameric motif (aataatattctt ; SEQ ID NO : 8) that protects the mitochondrial RNA from the RNA degradosome. Preferably, the C0X2 3'UTR sequence is inserted 3' to the sequence encoding the shRNA and 5' to the transcription terminator.

A preferred genetic construct derived from Anabaena sp. Group I PIE sequences and in which the ciRNA RNA precursor is expressed under the transcriptional control of the T7 RNA polymerase promoter and terminator sequences comprises the sequence encoding the shRNA of the invention inserted between the sequences SEQ ID NO: 9 and SEQ ID NO: 10. This construct is flanked by EcoRl and Smal sites at both ends and comprises Pst\ and Pmel sites 3' of the T7 promoter as well as Xbal, Nrul, Spel and EcoRV sites 5'of the T7 terminator. In this construct the COX2 3'UTR sequence (SEQ ID NO: 7) flanked at both ends by Spel and EcoRV sites is inserted 5' to the T7 terminator.

A similar strategy may be used to generate genetic constructs from other permuted intron-exon sequences, as described above, For example, from PIE sequences derived from an intron of a yeast mitochondrial gene, such as Saccharomyces cerevisiae mitochondrial 21 S rRNA gene, rl intron encoding a putative transposase (GenBank accession number Ml 1280 and SEQ ID NO: 11) or S. cerevisiae COXl gene Group I al 4 intron.

The invention concerns also a vector comprising the genetic construct as defined above. A vector which can be used in the present invention includes, but is not limited to, a viral vector, a plasmid, a RNA vector or a linear or circular DNA or RNA molecule which may consists of a chromosomal, non chromosomal, semisynthetic or synthetic nucleic acids. Preferred vectors are those capable of autonomous replication (episomal vector) and/or expression of nucleic acids to which they are linked (transcription vectors). Vectors can also comprise selectable markers, for example: neomycin phosphotransferase, histidinol dehydrogenase, dihydro folate reductase, hygromycin phosphotransferase, herpes simplex virus thymidine kinase, adenosine deaminase, glutamine synthetase, and hypoxanthine-guanine phosphoribosyl transferase for eukaryotic cell culture; TRPl for S. cerevisiae; tetracycline, rifampicin or ampicillin resistance in E. coli. Large numbers of DNA vectors suitable for in vitro and in vivo transcription are known to those of skill in the art and commercially available.

Preferably, the transcription vector is a plasmid containing a replication origin which is active in the host cells, and a selection marker. Preferably, the host cells are bacteria or yeasts. Yeast expression vectors, comprise advantageously the Ori5 mitochondrial replication origin sequence (SEQ ID NO: 12) for maintenance of the plasmid vector in the yeast cell progeny and eventually a COX2 gene fragment (SEQ ID NO: 13) for the selection of mitochondrial transformants. The invention concerns also eukaryotic or prokaryotic cells which are modified by a vector as defined above.

The invention concerns also a method for producing the ciRNA of the invention derived from the method described in the EP Patent 1646724, which method is characterized in that it comprises at least the following steps: (1) transforming the mitochondria of yeast cells (in particular

S. cerevisiae cells) lacking mitochondrial DNA (rho° strain) with a mitochondrial transcription vector comprising the genetic construct as defined above under the control of a promoter and a transcription terminator that are functional in yeast mitochondria, and a mitochondrial transformation reporter gene or a fragment of said reporter gene; a mitochondrial transformant or a synthetic rho strain is thus obtained;

(2) identifying the yeast mitochondrial transformants that have incorporated the DNA of interest;

(3) culturing the yeast mitochondrial transformants selected in step (2), preferably in the exponential growth phase;

(4) isolating the mitochondria from the yeast mitochondrial transformants cultured according to step (3), and (5) extracting and purifying the ciRNA of interest from said mitochondria.

The steps (1) to (5) are performed as described in the EP Patent 1646724.

The ciRNA of the invention may be used as a therapeutic agent to reduce activity of specific gene products. The ciRNA of the invention may also be used in functional genomics to knock down single genes for detailed study or hundreds to thousands of genes in high-throughput functional genomic surveys.

The invention also concerns a ciRNA molecule or a vector as defined above, as a medicament. The invention concerns also a pharmaceutical composition comprising at least a ciRNA of the invention in an acceptable carrier, such as stabilizer, buffer and the like.

A pharmaceutical composition or formulation refers to a form suitable for administration, e.g., systemic or local administration, into a cell or subject, including for example a human. Suitable forms, in part, depend upon the use or the route of entry, for example oral, inhalation, or by injection. These compositions or formulations are prepared according to any method known in the art for the manufacture of pharmaceutical compositions.

In one embodiment, the invention features a composition wherein the ciRNA molecule or vector is associated to a compound that allows the delivery of the ciRNA/vector into target cells. The compound may be a membrane peptide, transporter, lipid, hydrophobic moiety, cationic polymer, PEL Preferably, the ciRNA and the compound are formulated in microspheres, nanoparticules or liposomes. Furthermore, the ciRNA molecule or vector may be associated with a compound that allows a specific targeting of the target cell, such as a ligand of a cell-surface antigen or receptor, for example a peptide or an antibody specific for said antigen/receptor . A pharmaceutically effective dose is that dose required to prevent, inhibit the occurrence or treat (alleviate a symptom to some extent, preferably all the symptoms) of a disease or state. The pharmaceutically effective dose of the ciRNA depends upon the type of disease, the composition used, the route of administration, the type of mammal being treated, the physical characteristics of the specific mammal under consideration, concurrent medication, and other factors, that those skilled in the medical arts will recognize. Generally, an amount between 0.1 mg/kg and 100 mg/kg body weight/day of active ingredients is administered.

The ciRNA of the invention may be administered by a single or multiple route(s) chosen from: local (intratumoral, for example intracerebral (intrathecal, intraventricular)), parenteral (percutaneous, subcutaneous, intravenous, intramuscular, intraperitoneal), oral, sub-lingual, or inhalation.

The practice of the present invention will employ, unless otherwise indicated, conventional techniques of cell biology, cell culture, molecular biology, transgenic biology, microbiology, recombinant DNA, and immunology, which are within the skill of the art. Such techniques are explained fully in the literature. See, for example, Current Protocols in Molecular Biology (Frederick M. AUSUBEL, 2000, Wiley and son Inc, Library of Congress, USA); Molecular Cloning: A Laboratory Manual, Third Edition, (Sambrook et al. , 2001 , Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press); Oligonucleotide Synthesis (M. J. Gait ed., 1984); Mullis et al. U.S. Pat. No. 4,683,195; Nucleic Acid Hybridization (B. D. Harries & S. J. Higgins eds. 1984); Transcription And Translation (B. D. Hames & S. J. Higgins eds. 1984); Culture Of Animal Cells (R. I. Freshney, Alan R. Liss, Inc., 1987); Immobilized Cells And Enzymes (IRL Press, 1986); B. Perbal, A Practical Guide To Molecular Cloning (1984); the series, Methods In ENZYMOLOG Y (J. Abelson and M. Simon, eds.-in-chief, Academic Press, Inc., New York), specifically, VoIs.154 and 155 (Wu et al. eds.) and Vol. 185, "Gene Expression Technology" (D. Goeddel, ed.); Gene Transfer Vectors For Mammalian Cells (J. H. Miller and M. P. Calos eds., 1987, Cold Spring Harbor Laboratory); Immunochemical Methods In Cell And Molecular Biology (Mayer and Walker, eds., Academic Press, London, 1987); Handbook Of Experimental Immunology, Volumes I-IV (D. M. Weir and C. C. Blackwell, eds., 1986); and Manipulating the Mouse Embryo, (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y., 1986).

In addition to the preceding features, the invention further comprises other features which will emerge from the description which follows, which refers to examples illustrating the ciRNA molecules of the invention and their use according to the invention, as well as to the appended drawings in which: - figure 1 illustrates the production of circular RNA molecules using permuted intron-exon sequences. FP: fusion point. 5'ss: 5' splice site. 3'ss: 3'splice site. SJ: splice junction.

- figure 2 represents the structure of circular RNA molecules according to the invention. - figure 3 represents the structure of a genetic construct designed for the production of the circular molecule according to the invention. Intron 2: intron 3' half. Exon 2: 3'exon sequence flanking the 3' splice site. Intron 1 : intron 5' half. Exon 1 : 5'exon sequence flanking the 5' splice site.

- figure 4 represents the sequence of a genetic construct designed for the production of the circular molecule according to the invention.

- figure 5 represents the pVciLuc (A) and pVmutciLuc (B) plasmid map.

- figure 6 illustrates the resistance of the circular RNA ciLUC to exonuclease R treatment. The linear precursor of ciLUC produced by in vitro transcription was circularized by incubation with GTP in Hepes buffer for 18 h at 32 °C. An aliquot of the reaction mixture was incubated with exonuclease R (IUI per microgram of RNA) for Ih at 37 °C and analyzed by electrophoresis on denaturing 1 % agarose gel (ciLuc (+)). An aliquot of the reaction mixture, not treated with exonuclease R (ciLuc (-)) and a linear RNA (LIN) were used as controls. - figure 7 represents the pucMod (A) and pPT24 (B) vector map.

- figure 8 represents the T7 vector map. - figure 9 illustrates the RNA interference activity of the ciLUC RNA. Huh-7 cells (10⁵ cells) were transfected with 900 ng of RNA. The expression of the two luciferase genes was assayed at 48 h. The values correspond to the percentage of luminescence inhibition by the RNA (ratio of RLUs in the cells transfected with RNA versus control cells (no RNA) x 100). The RNA interference activity of the circular siRNA, ciLuc, was compared to that of the corresponding shRNA (siLUC) and a non-relevant shRNA (siNCE). The plotted data are the means ± standard deviation of 4 i experiments.

Example 1: Design of a ciRNA targeting the luciferase gene (ciLUC) shLUC is a known shRNA targeting the luciferase gene

(5'CUUACGCUGAGUACUUCGAUUGAACGAAGAAUCGAAGUACUCAGCGU AAGUUUUU3'; SEQ ID NO: 14); it contains a 21mer RNA duplex targeting the luciferase gene, a 8-nucleotide loop (sequence underlined) and a 5-nucleotide 3'- overhang on the antisense strand. shLUC was used to construct a circular interfering RNA molecule named ciLUC:

5'AAAAUCGCUAGGCGCUUACGCUGAGUACUUCGAUUGAA CGAAGAAUCGAAGUACUCAGCGUAAGUUUUUCUGGGCUAGCGACUUS'; SEQ ID NO: 15) using the Group I PIE construct derived from the Anabaena pre- tRNA^Leu gene described previously by Puttaraju et al, Nucleic Acids Res., 1993, 21, 4253-4258. The 5' and 3' ends of the ciLUC sequence correspond to the 3' and 5' splice sites respectively. The structure of ciLUC is presented in figure 2B; it contains the shLUC sequences mentioned above and additional sequences (underlined) consisting of the tRNA^Leu anticodon stem-loop of Anabaena group I PIE (in bold) and sequences from a HDV ribozyme (AGGCG and CUGGGCU). The tRNA^Leu anticodon stem-loop of Anabaena group I PIE contains a 7-nucleotide loop (AAAA TUUC) corresponding to the splice junction (arrow) and a 5 base-pair stem (UCGCU and AGCGA).

A non-circular RNA (mutant ci-LUC) which differs from ciLUC by the introduction of two mutations (AA to UG and UU to AC; marked with asterisks), respectively in the 3' and 5' splice sites, was also produced: 5'

U*G*AAUCGCUAGGCGCUUACGCUGAGUACUUCGAUUGAACGAAGAAUC GAAGUACUCAGCGUAAGUUUUUCUGGGCUAGCGACA*C*3'; SEQ ID NO:

16).

Example 2: Construction of ciLUC transcription vectors

A genetic construct was designed for the transcription of ciLUC in vitro and in vivo in bacteria and yeast (figures 3 and 4). In this construct, the ciRNA precursor is under the transcriptional control of the T7 RNA polymerase promoter and T7 terminator sequences (SEQ ID NO: 5 and SEQ ID NO: 6, respectively).

This construct contains sequences from the Anabaena Group I PIE plasmid pRIOO (Puttaraju, M. and M.D. Been, Nucleic Acids Res., 1992, 20, 5357- 5364; see in particular figure 1) and from the derived plasmids, pR120, containing a shorter fused-exon sequence and pRCl, containing an HDV ribozyme into the exon (Puttaraju et al, Nucleic Acids Res., 1993, 21, 4253-4258; see in particular figure Ic and 2). These sequences (figure 4) consist of two fragments: a first fragment including (A) plasmid sequences, (B) a 3' portion of the intron, (C) the 3' splice site, the first nucleotides of the 3' exon and HDV ribozyme sequences (AAAATCGCTAGGCG) and a second fragment including: (D) HDV ribozyme sequences, the last nucleotides of the 5' exon and the 5¹ splice site (CTGGGCT AGCGACTT), (E) a 3' portion of the intron and (F) other plasmid sequences.

The shLUC coding sequence is inserted between the first and the second fragments.

The construct contains also the COX2 3'UTR sequence (SEQ ID

NO: 7) including the dodecameric motif aataatattctt (SEQ ID NO: 8) that protects the mitochondrial RNA from the RNA degradosome. The COX2 3'UTR sequence is inserted 3' to the sequence encoding the ciRNA precursor and 5' to the T7 transcription terminator.

In addition, the construct contains restriction sites. It is flanked by EcoRl and Smal sites at both ends and comprises Pstl and Pmel sites 3' to the T7 promoter as well as Xbal, Nrul, Spel and EcoRV sites 5 'to the T7 terminator. In this construct the COX2 3'UTR sequence (SEQ ID NO: 7) is flanked at both ends by Spel and EcoRV sites as well as by Xbal and Nrul sites, at its 3' end only. The sequence of this construct corresponds to SEQ ID NO: 17. Two control constructs were also generated: a first control (SEQ ID NO: 18) in which the sequence encoding ciLUC was replaced with the sequence encoding mutant ciLUC, and a second control encoding shLUC only (deletion of the intron and exon sequences from Anabaena Group I PIE). The genetic constructs encoding the ciLUC or shLUC RNAs were generated by synthesis of two complementary oligonucleotides corresponding to each strand of the DNA construct and annealing of the oligonucleotides. Each construct was then inserted into the EcoRV site of the pUC57 cloning vector (# SDl 176; GENSCRIPT) to give the recombinant plasmids pVciLuc (figure 5A) and pVshLuc, using standard recombinant DNA techniques. The plasmid pVmutciLuc (figure 5B) containing the genetic construct encoding mutant-ciLUC was derived from pVciLuc by site-directed mutagenesis of the ciLuc insert (SEQ ID NO: 17) at positions 204-205 (AA to TG mutation) and 286-287 (TT to AC mutation). The constructs were also cloned into appropriate vectors for in vivo transcription in yeast or bacteria using T7 RNA polymerase.

Example 3: Production of ciLUC by in vitro transcription and circularization of the transcript

1) Material and methods a) Production of linear unspliced ciRNA-precursor by in vitro transcription The in vitro transcription plasmids containing the genetic constructs encoding ciLUC, mutant-ciLUC or shLUC (example 2) were digested with appropriate restriction enzyme(s) to create linear templates for in vitro transcription using T7 RNA polymerase. The plasmids pVciLuc and pVmutciLucwere digested with EcoKV (25 UI) for 3h at 37 °C in the buffer recommended by the manufacturer and purified by ethanol precipitation. In vitro transcription reactions (40 μL) containing 2 μg of linear plasmid DNA were incubated for 3 h at 37 °C using the

MEGAscript® kit, according to the manufacturer's instructions (AMBION). Dnase (2 μL) was then added to the transcription reactions which were then incubated for 15 min at 37 °C. Lithium chloride 7.5 M, 50 mM EDTA (100 μL) was then added to the transcription reactions and the mixtures were incubated for Ih at 80 °C and then centrifuged for 45 min at 13200 rpm and at 4 °C. The RNA pellets were then washed in ethanol (70 %) and resuspended in sterile water (100 μL). b) Circularization of the RNA produced by in vitro transcription

The RNA produced by in vitro transcription was incubated in Hepes buffer (40 raM Hepes, pH 7.5; 200 mM NaCl; 20 mM MgCl₂) for 5 min at 50 °C.

GTP (200 μM) was then added to the RNA sample and the mixture was incubated for 18 h at 32 °C. The reactions were kept on ice for subsequent analysis on the same day or stored at - 80 ⁰C. c) Endonuclease/Exonuclease treatments

The RNA was digested with exonucleases having a 3'-5'(RnaseR, RnaseT, PNPase), or 5 '-3' (Exoribonuclease I and II) activity, endonucleases specific for single-strand (Rnase A or RNAse Tl) or double-strand (Rnase V) RNA, according to the manufacturer's instructions. Exonuclease treatment was performed in the presence of 1 UI of Exonuclease R (TEBU-BIO) per microgram of RNA, for Ih at 37 °C. d) Gel electrophoresis RNA samples were mixed with denaturation loading dye solution (95

% formamide, 0.025 % SDS, 0.025 % Bromophenol Blue, 0.05 % Ethidium bromide; # LDS-DT-I, MITOPROD) and 0.5 μL ethidium bromide, incubated 10 min at 65- 70°C and separated by electrophoresis in native or denaturing 1 % agarose gels, at a voltage inferior or equal to 150 V. Denaturing agarose gel (1 g agarose, 10 mL 1OX MOPS buffer (0.4 MOPS, 0.1 M sodium acetate, 10 mM EDTA), 18 mL formaldehyde, 72 mL water, 1 μL ethidium bromide (10 mg/mL)) were used with running buffer containing IX MOPS buffer (0.04 MOPS, 0.01 M sodium acetate, 1 mM EDTA). Native agarose gel (1 g agarose, 10 mL 1OX TAE buffer, 90 mL water, 1 μL ethidium bromide (10 mg/mL) were used with IX TAE buffer. A RNA ladder (# LAD-DT-25, MITOPROD) was used as a molecular weight marker. Bands on gel were visualized using a UV (302 nm) transilluminator. e) Northern Blot Analysis

The RNA samples were separated by gel electrophoresis as described above, transferred to a PVDF membrane and detected by standard hybridization technique with a ³²P labeled DNA probe specific for the intron or exon portion of the Anabaena Group I PIE sequences, or for the shLUC, prepared using the

Rediprime TM II Random primed Labelling System kit (# RPN 1633, AMERSHAM), following the manufacturer's instructions. Bands on the membrane were visualized by autoradiography. f) Anion exchange HPLC

RNA sample, eventually diluted in loading buffer A (30 mM lithium perchlorate, 20 mM sodium acetate, pH 6.5), was denatured at 65-70°C for 10 min. Then, it was loaded at a low flow-rate on a anion exchange analytic column (DNApac200, DIONEX) equilibrated with the loading buffer A. The chromatography was performed at a low flow-rate (lmL/min) with a gradient of elution buffer B (30 to 300 mM lithium perchlorate). The column was maintained at the desired temperature (usually 65-70°C); the buffers were at room temperature. Alternatively, the column was at room temperature and the buffers were at 65-70°C. All chromatographic runs were carried out on an AKTA Purifier 10 (GE Healthcare). Data analysis and reporting were performed on computers using the Unicorn 5.0 control system software. The different peaks were identified and the area of each peak was measured to evaluate the efficiency of production of the circular RNA. Purified circular RNA was recovered from the corresponding peak. g) RNA quantification

A sample containing RNA was diluted in a neutral buffer (usually water) and the amount of RNA present in the sample was determined by measuring the absorbance (OD) of the diluted sample at 260 nm using a spectrophotometer. The level of contaminants was evaluated by measuring the absorbance at 230 nm and 280 nm and calculating the ratios OD₂₆₀/OD₂₈o and OD₂₆₀/OD₂₃₀. Salts (EDTA, acetate), solvents (trizol) and protein (peptid bond) absorb at 230 nm; a good sample has a OD₂₆₀/OD₂₃o ratio > 1.8. Polysaccharides, glycogen, fats and lipids absorb at 280 nm; a good sample has a OD₂₆₀/OD₂₈o ratio > 1.6 and < 2.1. If the ratio is below, the RNA is not completely solubilized or the sample contains many proteins. If the ratio is greater than 2.1 , the RNA is degraded, h) RNA stability assay

The stability of the RNAs (ciLUC, mutant-ciLUC, shLUC) produced by in vitro transcription was assayed by 3'exonuclease digestion as described above. It was also assayed by incubating the ciLUC and shLUC RNAs at 22 °C and 37 °C for different time periods and compairing the ciLUC and shLUC chromatographic profiles by anion echange HPLC, as described above. 2) Results

The RNA products obtained by in vitro transcription from the vectors encoding the ciLUC, mutant-ciLUC or shLUC RNAs, and circularization or not, of the linear transcripts in the presence of GTP, were treated with various exonucleases and endonucleases, and analysed by anion exchange HPLC, native or denaturing gel electrophoresis, and eventually by Northern-blot analysis, to identify the different RNA molecules produced (linear precursor (unspliced), intermediates and products (circular (spliced), 5' and 3' linear fragments; see figure Ic of Puttaraju M. and M.D. Been, Nucleic Acids Res., 1992, 20, 5357-5364 or figure 1 of the present application). The proportion of circular RNA (ciLUC) was measured to determine the efficiency of production of this circular RNA by in vitro transcription of a linear unspliced precursor and circularization by self-splicing in the presence of GTP. The proportion of circular RNA (ciLUC) may be compared with that of shRNA (shLUC) produced in the same conditions. Circular RNA (at least 80 % pure) was purified by HPLC. The stability of the circular RNA (ciLUC) was tested and compared with that of mutant- ciLUC produced in the same conditions.

The results presented in figure 6 show that the circular siRNA ciLUC can be produced by in vitro transcription of a linear precursor containing Anabaena Group I PIE sequences and circularization of the exon sequences containing shLUC, by self-splicing in the presence of GTP. The RNA resulting from the incubation of the in vitro transcript in the presence of GTP is circular as demonstrated by its resistance to exonuclease digestion. Example 4: Production of ciLUC by in vivo transcription in bacteria 1) Material and methods a) Bacteria transformation

The transcription vectors comprising the genetic constructs encoding ciLUC, mutant-ciLUC or shLUC (example 2) were transformed into an E.coli strain expressing the T7 RNA polymerase from a lac promoter (T7 express competent (OZYME), BL21(DE3)(NOVAGEN or AGILENT TECHNOLOGIES), using the heat shock method. BL21(DE3), F ompT gal dcm Ion hsdS_B(r_B ^" m_B ^") λ(DE3 [lad lacUV5- T7 gene 1 indl sam7 nin5]), is an E. coli B strain with DE3, a λ prophage carrying the T7 RNA polymerase gene and lacl^q. Transformed plasmids containing T7 promoter driven expression are repressed until IPTG induction of T7 RNA polymerase from a lac promoter. b) Transcription induction and total RNA extraction

Transformed bacteria were cultivated in fresh medium complemented with the appropriate antibiotic to reach 2 OD/ml. IPTG (2mM) was added to the cultures which were then incubated for 3 hours at 37 °C under agitation. After OD measurement, the culture was centrifugated and the RNA was extracted from the pellet using trizol® (INVITROGEN) or trireagent® (EUROMEDEX). The cell pellet was resuspended in Trireagent (1 ml solution for each 20 OD; sample volume should not exceed 10 % of the volume of Trireagent). Chloroform (0.2 mL per ImL Trireagent) was then added and the mixture was vortexed for 15 sec (2 times), stored few minutes (2 to 15 min) at room temperature and centrifugated at 8500g to 1200Og for 15-20 min at 4 °C. The aqueous phase was transferred to a fresh tube and isopropanol was added (0.5 mL for each 20 OD). The mixture was stored 5 to 10 min at room temperature and centrifugated for 15 to 20 min at 4 to 25 °C at 9000g to 1200Og. The supernatant was removed and the RNA pellet was washed with 75 % ethanol (0.5 to ImL for each 20 OD) and centrifugated at 750Og for 5 min at 4 to 25°C. The ethanol was removed and the RNA pellet was briefly air-dried for few min. The RNA was dissolved in water or in buffer A (for chromatography) by passing the solution a few times through a pipette tip and incubating if necessary for 10-15 min at 55-60⁰C, to improve the solubilization. 2) Results The RNA products obtained by in vivo transcription from bacteria expressing the T7 RNA polymerase under the control of a lac promoter, transformed with vectors encoding the ciLUC, mutant-ciLUC or shLUC RNAs (example 2), induced or not by IPTG, were treated with various exonucleases and endonucleases, and analysed by anion exchange HPLC, native or denaturing gel electrophoresis, and eventually by Northern-blot analysis, as described in example 3, to identify the different RNA molecules produced (linear precursor (unspliced), intermediates and products (circular (spliced), 5' and 3' linear fragments; see figure Ic of Puttaraju M. and M.D. Been, Nucleic Acids Res., 1992, 20, 5357-5364 or figure 1 of the present application). The proportion of circular RNA (ciLUC) was measured to determine the efficiency of production of this circular RNA by in vivo transcription and in vivo circularization by self-splicing. The proportion of circular RNA (ciLUC) may be compared with that of mutant-ciLUC produced in the same conditions. Circular RNA (at least 80 % pure) was purified by HPLC. The stability of the circular RNA (ciLUC) was tested and compared with that of mutant-ciLUC produced in the same conditions. Example 5: Production of ciLUC by in vivo transcription in yeast mitochondria 1) Material and methods The production of heterologous RNA by transcription in yeast mitochondria is described in the EP Patent 1646724.

The genetic constructs encoding ciLUC, mutant-ciLUC or shLUC described in example 2 were cloned individually into the EcoRl site of pucMod or pPT24 (figure 7) mitochondrial transcription vectors. The pucMod vector is a pUC57 derived plasmid containing the Ori5 sequence which allows the maintenance of the plasmid in the mitochondria of the yeast progeny and a COX2 gene fragment (SEQ ID NO: 13) for the selection of the mitochondrial transformants. The plasmid pPT24 is described in Thorness, P.E. and T.D. Fox, Genetics, 1993, 134, 21-28.

The W303-1B strain (Matα, ade2, trpl, his3, Ieu2, ura3), called W303- IB (ATCC No. 201238)/ A/50 was transformed with a vector (figure 8) containing the a T7 RNA polymerase gene operatively linked to a mitochondrial targeting sequence (MTS), COX4 leader sequence or ATP9 MTS, under the control of the galactose inducible promoter, GALlO, and an auxotrophic marker (LEU2 or ADEI).

The biolistic method was used to transform the recombinant mitochondrial transcription vectors into the mitochondria of rho⁰ (lacking mitochondrial DNA) derivatives of W303- IB strain nuclear transformants (expressing the T7 RNA polymerase). The mitochondrial transformants were isolated by crossing with a tester strain (rho⁺, C OX2^') and isolation of cells capable of growing on a non- fermentable medium. Mitochondrial transformants were cultivated in YPGA medium supplemented with glucose and galactose to induce T7 RNA polymerase expression to reach 5 to 10 OD/ml. The culture was centrifugated and the mitochondria were isolated as described in the EP Patent 164 and the mitochondrial RNA was extracted using the RN AXEL® and RNABIND® reagents according to the manufacturer's instructions (EUROMEDEX). 2) Results

The RNA products obtained by transcription in the mitochondria of yeast expressing the T7 RNA polymerase under the control of a galactose inducible promoter, transformed with mitochondrial transcription vectors encoding the ciLUC, mutant-ciLUC or shLUC RNAs (example 2), induced or not by galactose, were treated with various exonucleases and endonucleases, and analysed by anion exchange HPLC, native or denaturing gel electrophoresis, and eventually by Northern-blot analysis, as described in example 3, to identify the different RNA molecules produced (linear precursor (unspliced), intermediates and products (circular (spliced), 5' and 3' linear fragments; see figure Ic of Puttaraju M. and M.D. Been, Nucleic Acids Res., 1992, 20, 5357-5364 or figure 1 of the present application). The proportion of circular RNA (ciLUC) was measured to determine the efficiency of production of this circular RNA by in vivo transcription and in vivo circularization by self-splicing in yeast mitochondria. The proportion of circular RNA (ciLUC) may be compared with that of mutant-ciLUC produced in the same conditions. Circular RNA (at least 80 % pure) was purified by HPLC.The stability of the circular RNA (ciLUC) was tested and compared with that of mutant-ciLUC produced in the same conditions. Example 6: RNA interference activity of ciLUC 1) Material and methods

Huh-7 (human hepatocarcinoma) derived cells expressing constitutively both the Firefly and Renilla luciferase genes were grown in 24 well plates (10⁵ cells/well) in DMEM medium supplemented with 20 % fetal calf serum (complete medium), for 16 h at 37 °C with 5 % CO₂. The transfection was performed by incubating 0.9 μg of RNA and 2.1 μL of DMRIE-C liposome (INVITROGEN) in 500 μL OptiMEM® (GIBCO), for 30 min at room temperature. The 500 μL transfection mixture was added to the cells rinsed with PBS and the cells were incubated for 4 h at 37 °C with 5 % CO₂. 500 μL of complete medium was added to the cells which were incubated for 48 h at 37 °C with 5 % CO₂. The cells were then rinsed three times with PBS lysed in 70 μL of lysis buffer (PROMEGA) and centrifuged 2 min at 13 000 g. The supernatant was recovered and the luciferase assay was performed on a 20 μL aliquot of the supernatant mixed with 50 μL of substrate using the Dual-Luciferase Assay System kit according to the manufacturer's instructions (PROMEGA). The light emitted by the luciferase-catalyzed chemoluminescent reaction (number of relative luciferase units or RLUs) was measured in the cells transfected with RNA and in the control cells (no RNA), using a luminometer. The silencing of the luciferase gene was calculated from the ratio of RLUs in the transfected cells versus control cells. 2) Results

The RNA interference activity of the ciLUC RNA (ciLuc) was compared with that of shLUC RNA (siLuc) and a non-relevant shRNA (siNCE) produced in vitro as described in example 3. The results presented in figure 9 show that the ciLUC RNA has an RNAi activity which is equivalent to that of the shLUC

RNA.

Claims

1°) A circular RNA molecule which comprises the sense and the antisense strands of an interfering RNA (RNAi) molecule targeting a gene of interest, wherein said sense and antisense strands are closed at one end by a loop structure (first loop) and at the other end by a splice junction sequence generated by splicing activity of permuted intron-exon sequences, and wherein the circular RNA molecule inhibits the expression of the gene of interest in cells expressing said gene of interest.

2°) The circular RNA molecule according to claim 1, wherein the sense and antisense strands of the RNAi molecule comprise 21 to 27 ribonucleotides. 3°) The circular RNA molecule according to claim 1 or claim 2, wherein the permuted intron-exon sequences are from the intron of a yeast mitochondrial gene.

4°) The circular RNA molecule according to claim 1 or claim 2, wherein the permuted intron-exon sequences are from a self-splicing Group I intron. 5°) The circular RNA molecule according to claim 4, wherein the self-splicing Group I intron is from the Cyanobacterium Anabaena sp. pre-tRNA-Leu gene or the Tetrahymena ribosomal gene .

6°) The circular RNA molecule according to anyone of claims 1 to 5, wherein the splice junction sequence comprises a loop structure (second loop). 7°) The circular RNA molecule according to claim 1 or claim 6, wherein the loop structure(s) comprise(s) at least 5 ribonucleotides.

8°) The circular RNA molecule according to claim 7, wherein the loop structure(s) comprise(s) the sequence SEQ ID NO: 1.

9°) The circular RNA molecule according to anyone of claims 1 to 8, wherein the splice junction sequence comprises a stem-loop structure in which the strands of the stem are connected to the strands of the RNAi molecule, directly or by 3 to 10 non-complementary ribonucleotides.

10°) The circular RNA molecule according to claim 9, wherein the splice junction sequence comprises the sequence SEQ ID NO: 2. 1 1°) A genetic construct designed for the production of the circular

RNA molecule according to anyone of claims 4 to 10, said genetic construct comprising: (1) a transcription initiation region, (2) a sequence encoding a 3' portion of a self-splicing Group I intron, (3) a sequence encoding the 3' splice site of said self- splicing Group I intron, (4) a sequence encoding the sense strand, the loop structure and the antisense strand of the siRNA molecule as defined in claim 1 or claim 2, inserted at the fusion point of the 3'exon of said self-splicing Group I intron fused end-to-end to the 5'exon of said self-splicing Group I intron or at the fusion point of a fragment of the fused exons having few nucleotides of the 3' exon sequence flanking the 3' splice site and few nucleotides of the 5' exon sequence flanking the 5' splice, (5) a sequence encoding the 5' splice of said self-splicing Group I intron, (6) a sequence encoding a 5' portion of said self-splicing Group I, and (7) a transcription termination region, wherein the sequences (2) to (6) of the genetic construct are operatively linked to the transcription initiation (1) and termination (7) regions, and the RNA molecule that is expressed from the promoter generates a circular form of the siRNA by self-splicing.

12°) The genetic construct according to claim 11, wherein the transcription initiation and termination regions are functional in yeast mitochondria.

13°) The genetic construct according to claim 11 or claim 12, wherein the transcription initiation and termination regions are from bacteriophage T7 RNA polymerase.

14°) The genetic construct according to anyone of claims 11 to 13, which comprises the COX2 3'UTR sequence SEQ ID NO: 7.

15°) The genetic construct according to anyone of claims 1 1 to 14, wherein the sequence encoding the sense strand, the first loop and the antisense strand of the siRNA molecule is inserted between the sequences SEQ ID NO: 9 and SEQ ID NO: 10. 16°) A transcription vector comprising the genetic construct according to anyone of claims 11 to 15.

17°) The transcription vector according to claim 16, which is a bacteria or yeast plasmid containing an origin of replication and a selection marker.

18°) The transcription vector according to claim 17, which is a yeast plasmid comprising the Ori5 mitochondrial replication origin sequence SEQ ID NO: 12 and the COX2 gene sequence fragment SEQ ID NO: 13. 19°) A eukaryotic or prokaryotic cell which is modified by a transcription vector according to anyone of claims 16 to 18.

20°) A method for producing the ciRNA according to anyone of claims 1 to 10, comprising at least the following steps: (1) transforming the mitochondria of yeast cells lacking mitochondrial

DNA (rho° strain) with a mitochondrial transcription vector comprising a genetic construct according to anyone of claims 12 to 15, and a mitochondrial transformation reporter gene or a fragment of said reporter gene;

(2) identifying the yeast mitochondrial transformants that have incorporated the DNA of interest;

(3) culturing the yeast mitochondrial transformants selected in step (2), preferably in the exponential growth phase;

(4) isolating the mitochondria from the yeast mitochondrial transformants cultured according to step (3), and (5) extracting and purifying the ciRNA of interest from said mitochondria.

21°) Use of a circular RNA molecule according to anyone of claims 1 to 10 for functional genomics studies.

22°) A circular RNA molecule according to anyone of claims 1 to 10 as a medicament.