TECHNICAL FIELD

The present invention generally relates to ribonucleic acid (RNA) molecules, and in particular to stable RNA molecules useful in gene expression control.

BACKGROUND

Small interfering RNAs (siRNAs) and microRNAs (miRNAs) are ~22 nucleotide long non-coding RNA molecules, discovered more than two decades ago. These small RNA (sRNA) molecules play an important role in gene regulation through association with the Argonaute (Ago) family of proteins and inducing post-transcriptional gene silencing of target messenger RNA (mRNA) in animals and plants. Thus, siRNAs and miRNAs are considered as novel therapeutic agents for the treatment of a wide range of disorders, including cancers and infections. Therapeutic siRNAs are designed to inhibit specific genes through perfect base pairing between the siRNA and the target mRNA, thereby causing a degradation of the target mRNA. Therapeutic miRNAs are used in the so-called replacement therapy, in which they mimic the function of endogenous miRNAs.

Therapeutic siRNAs and miRNAs are commonly introduced to the cell in the form of double stranded small RNAs, in which one strand, the guide stand, will bind to an Argonaute protein to form the RNA- induced silencing complex (RISC), which is the effector complex. The other strand, the passenger stand, is believed to be degraded. The strand selection step is not well understood and depends on many factors, including target availability and thermodynamic properties of the small RNA duplex. Although significant advances have been made to ensure the loading of only the guide strand, for example through various chemical modifications in the passenger strand, the off-target effects caused by passenger strand loading is still a major complication.

Accordingly, loading of single-stranded small RNAs onto RISC has been proposed to avoid the problems of double-stranded small RNAs mentioned above. However, in order to make such an approach feasible in vivo, the synthetic single-stranded small RNA needs to be chemically modified in order to avoid rapid degradation.

RNA activation (RNAa) is a sRNA guided and Ago-dependent gene regulation phenomenon, in which promoter-targeted short double-stranded RNAs induce target gene expression at the transcriptional level. Such promoter targeted small double-stranded RNAs have been termed antigene RNA (agRNAs) or small activating RNA (saRNA) in the art.

The problems described above for siRNA and miRNA molecules are also relevant for saRNA molecules.

Accordingly, there is still a need for efficient molecules that can be used in controlling gene expression, such as by RNA interference or RNA activation.

EP 2 562 257 discloses a method for stabilizing functional nucleic acids. A hairpin-shaped DNA is ligated to a double-stranded nucleic acid fragment or a single-stranded nucleic acid fragment having at least one stem portion and at least one loop portion. The hairpin-shaped DNA is said to improve the resistance of the single- or double-stranded nucleic acid fragment to degradation by nucleolytic enzymes.

SUMMARY

It is a general objective to provide a stable RNA molecule useful in gene expression control.

This and other objectives are met by embodiments as described herein.

The present invention is defined in the independent claims. Further embodiments of the invention are defined in the dependent claims.

An RNA molecule of the invention comprises a RNA fragment and a hairpin RNA fragment. The hairpin RNA fragment comprises a first stem region comprising the RNA sequence GCGUCG, a loop region comprising the RNA sequence CUGUCCA and a second stem region comprising the RNA sequence CGAGCGC. The RNA fragment is the functional domain of the RNA molecule and can, for instance, be used to silence or activate target genes.

The RNA molecule of the invention has several advantages over traditional siRNA, miRNA and saRNA molecules. Firstly, the RNA molecule is a single-stranded molecule thereby having reduced off-target effects since it does not contain any passenger strand as compared to traditional siRNA, miRNA and saRNA molecules comprising a guide strand and a passenger strand. No chemical modification to ribonucleotides is required to stabilize the RNA molecule and inhibit degradation thereof. Flence, standard synthesis method can be used to produce the RNA molecule. Double-stranded siRNA, miRNA and saRNA molecules are marred by unspecific toxicity, typically by activating components of the innate immune response. The RNA molecule of the invention has considerably lower toxicity.

BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments, together with further objects and advantages thereof, may best be understood by making reference to the following description taken together with the accompanying drawings, in which:

Fig. 1 . Production of RISC associated sRNAs from the HAdV-37 MLP transcriptional start site. A) Genomic organization of HAdV-37 showing the location of different viral transcription units. B) Distribution of viral small RNA reads (shown as percentage of total mapped viral reads) across the HAdV-37 genome both in the cytoplasmic fraction and F-Ago2-bound fraction. C) Close up view of the HAdV-37 MLP showing the mapped read coverage around the first leader (boxed) and the location of the MLP-TSS- sRNAs (SEQ ID NO: 38 and 39). D) Size distribution of the MLP-TSS-sRNA (-3 on left side and +1 on the right side) both in the cytoplasmic fraction and Ago2-associated fraction.

Fig. 2. Read count for the most abundant cellular microRNAs in in total cytoplasmic (Cyto) and Ago2- associated (F-Ago2-IP) in HAdV-37 infected cells. The abundance of the MLP-TSS-sRNA is shown in white. Fig. 3. The 5’ end of the HAdV-37 MLP-TSS-sRNAs are modified and produced during the late phase of infection. Northern blot analysis of cytoplasmic RNA extracted from HAdV-37 (5 FFU/cell) and mock (M) infected cells. The presence of a modified 5’ end was probed using TAP treatment. The same membrane was sequentially probed for the MLP-TSS-sRNA, 5’ mivaRNAII, Iet7b and tRNA lysine. Fig. 4. Northern blot analysis of cytoplasmic RNA extracted from HAdV-37 and mock (M) infected cells. The presence of a modified 5’ end was probed using tobacco acid pyrophosphatase (TAP) or RNA 5’ pyrophosphohydrolase (RppH) treatment. The same membrane was probed for (order from top to bottom): the MPL-TSS-sRNA, full length VA RNAII and tRNA lysine. Fig. 5. 293-Flag-Ago2 cells were infected with HAdV-37 (5 FFU/cell) or mock (M) infected. At different time points (6, 12, 18, 24 hpi) total cytoplasmic RNA was extracted and analyzed by Northern blotting using probes detecting the MPL-TSS-sRNA, 5’ mivaRNAII and tRNA lysine. Fig. 6. The HAdV-37 MLP-TSS-sRNAs are capped and enriched in Ago2 containing RISC. A) Northern blot analysis of total cytoplasmic (Cyto) and Flag-Ago2-associated (F-Ago2-IP) RNA. The membrane was probed sequentially for MLP-TSS-sRNA and 5’ mivaRNAII. The lower panel is a Western blot for Flag- Ago2 and Actin in the cytoplasmic and Ago2-bound fractions. B) Northern blot analysis demonstrating that the Ago2-associated MLP-TSS-sRNAs contain a 5’ cap structure. Ago2-associated RNAs were used for immunoprecipitation with an m7G cap antibody. The membrane was sequentially probed for MLP- TSS-sRNA, 5’ mivaRNAII and Iet7b.

Fig. 7. An siRNA knockdown approach of proteins involved in miRNA processing and RNA export was used to prove the pathway of MPL-TSS-sRNA production. A) Northern blot analysis to determine the effect of knockdown on association of the HAdV-37 MPL-TSS-sRNA and 5’ mivaRNAII with Ago2 complexes. B) Western blotting of the indicated proteins to measure the efficiency of siRNA knockdown. HEK-293 cells were transfected with the indicated plasmids for 24 hours, followed by a mock (M) infection or HAdV-37 (5 FFU/cell) infection for 34 hours. C) Northern blot analysis of total cytoplasmic (Cyto) and Flag-Ago2-associated (Flag-IP) RNA. The membrane was probed against the MPL-TSS-sRNA and Iet7b. D) Western blot analysis of Flag-epitope tagged proteins (Ago1 and Ago2) and GAPDH in the Cyto and Flag-IP fractions.

Fig. 8. The HAdV-37 MLP-TSS-sRNAs are products of premature termination of RNA polymerase II transcription. A) Schematic diagram of the different constructs used for the transfection assay. UnMLT contains the MLP fused to the first 200 nts downstream from the transcriptional start site, Trip contains the MLP fused to a cDNA copy of the spliced tripartite leader sequence, and RevTrip contains a cDNA copy of the tripartite leader where the second and third exons had been inserted in the reverse orientation. B) Northern blot analysis of total cytoplasmic RNA isolated 24 hours post transfection with the indicated plasmids. The same membrane was sequentially probed for the MLP-TSS-sRNA and tRNA lysine. C) Northern blot analysis of total (Input), RNAPII associated small RNA (RNAPII-IP) and Ago2-associated (Ago2-IP) RNA isolated from HAdV-37 infected 293-Flag-Ago2 cells at 30 hpi, The membrane was probed sequentially for the MLP-TSS-sRNA, 5’ mivaRNAII and tRNA lysine. D) Western blot analysis of RNAPII, Ago2 and Actin in the fractions displayed in C).

Fig. 9. A) ARPE19 cells were infected with HAdV-37 (20 FFU/cell) or mock (M) infected. At 24 hpi, total cytoplasmic (Cyto) and Ago-associated (Ago2-IP) RNA was probed sequentially for the MPL-TSS-sRNA and tRNA lysine by Northern blot analysis. B) Western blot for Ago2 and GAPDH in the Cyto and Ago2- IP fractions. C) Northern blot analysis of Flag-Ago2 associated RNA after transfection with either 5’ m7G- capped or monophosphate MPL-TSS-sRNA oligonucleotides for the indicated tie pointes. The membrane was probed for the MPL-TSS-sRNA. Fig. 10. The HAdV-37 MLP-TSS-sRNAs repress complementary targets. A) Schematic diagram of the different luciferase reporters used: pmirGlo lacks an MLP-TSS-sRNA binding site, pmirGlo(+) contains a complementary binding site, pmirGlo(Rev) has the binding site in reverse orientation and pmirGlo(mut) has multiple mutations in the binding site. B) and C) ARPE-19 cells were infected with HAdV-37 or mock infected. Following a one hour incubation cells were transfected with the indicated reporter plasmids. At 23 hours post transfection total cell lysate was prepared and luciferase activity measured. The result from three independent experiments is shown in the graph. Error bars indicate the standard deviation. D) ARPE-19 cells were co-transfected with the in wiro-transcri bed and capped MLP-TSS-sRNA or a control RNA (Ctrl) together with the indicated reporter plasmids. Luciferase expression was measured at 18 hrs post transfection. The mean from three independent experiments is shown in the graph. Error bars indicate the standard deviation (* p< 0.05, ** p<0.01 , *** p<0.001 ). E) RNA secondary structure prediction for the 31 nt MLP-TSS-sRNA, in the upper panel, and a shorter sRNA version that lacks the last 12 nt (MLP-TSS-sRNA(trunc)) in the lower panel. F) Northern blot analysis of total RNA isolated at the indicated time post transfection with 10 nM RNA oligonucleotides (either MLP-TSS-sRNA or MLP- TSSsRNA(trunc)) in 293-Flag-Ago2 cells. The same membrane was sequentially probed for the MLP- TSS-sRNA and tRNA lysine. On the right side Northern blot of 0.01 pmole of each RNA oligonucleotides probed for the MLP-TSS-sRNA. G) Hela cells were co-transfected with RNA oligonucleotides (either MLP-TSS-sRNA or MLP-TSSsRNA(trunc)) together with pmirGlo or pmirGlo(+) plasmids. Normalized Luciferase expression of pmirGlo(+) relative to pmirGlo was measured at the indicated time points, The results are presented as the mean from three independent experiments. Error bars indicate the standard deviation.

Fig. 1 1 . MLP-TSS-sRNA inhibits HAdV-37-replication. A) Schematic drawing showing the position of the MLP-TSS-sRNA in relation to the E2B mRNAs. B)-D) Effect of synthetic MLP-TSS-sRNA overexpression on HAdV-37 growth. B) ARPE-19 cells were transfected with a scrambled sRNA (Scr) or the MLP-TSS- sRNA (15 nM) for four hours, after that cells were infected with HAdV-37 (10 FFU/cell) for 12 hours followed measuring relative expression of both Adpol and pTP mRNAs by RT-qPCR. The results are presented as the mean from three independent experiments. Error bars indicate the standard deviation (** p<0.01 , * p<0.05). C) The same transfection and infection protocol as in panel B was used but the infection extended to 24 hpi at which time point the viral DNA copy number was determined by qPCR. Results are presented as the mean from three independent experiments. Error bars indicate the standard deviation (* p<0.05). D) ARPE-19 cells were transfected with the scrambled sRNA (Scr) or the MLP-TSS- sRNA at two different concentrations (15 and 50 nM) for four hours, followed by infection with HAdV-37 (10 FFU/cell). At 24 hpi viral late proteins were detected by Western blot analysis using a polyclonal anticapsid antibody. Panels E-G, effect of MLP-TSS-sRNA sponge constructs on HAdV-37 growth. E) HeLa cells infected with HAdV-37, were transfected with pmirGlo(+), expressing a reporter mRNA containing an MLP-TSS-sRNA binding site, or control plasmid pmirGlo, lacking the binding site. Adpol and pTP mRNA expression was measured 24 hpi by RT-qPCR. The results are presented as the mean from three independent experiments. Error bars indicate the standard deviation (* p< 0.05, ** p<0.01 , *** p<0.001 ). F) Three independent HeLa stable cell lines constitutively expressing the pmirGlo(+) or pmirGlo reporter mRNAs were infected with HAdV-37 and viral late protein expression detected by Western blot analysis at 24 hpi using a polyclonal anti-capsid antibody. G) Same as in F) with the exception that production of new virus particles was titrated in crude cell lysates prepared from the cells at 24 hpi. The results are presented as the mean from two independent experiments. Error bars indicate the standard deviation (* p<0.05, ** p<0.01 ).

Fig 12. The 5’ end of the MLP-TSS-sRNA constitute the target recognition domain. A) Schematic diagram of the 3’ UTR of the different luciferase reporters used: pmirGlo (-) lacks an MLP-TSS-sRNA binding site, pmirGlo (WT) contains a complementary binding site and pmirGlo (5’ end Mut) which has mismatch mutations disrupting the base pairing between the first 10 nts of the MLP-TSS-sRNA and the target mRNA. B) ARPE-19 cells were infected with HAdV-37 or mock infected. Following a one-hour incubation cells were transfected with the indicated reporter plasmids. At 23 hours post transfection total cell lysate was prepared and luciferase activity measured. The result from three independent experiments is shown in the graph. Error bars indicate the standard deviation. C) ARPE-19 cells were co-transfected with the synthetic MLP-TSS-sRNA or a control RNA (Scr) together with the indicated reporter plasmids. Luciferase expression was measured at 18 hrs post transfection.

Fig 13. The MLP-TSS-sRNA functional domain can be substituted with a miRNA mimic. A) Schematic diagram of the chimeric miR26a-TSS and miR155-TSS where the first (5’) nts of the hsa-mir-26a-5p (10 nts) and hsa-mir-155-5p (12 nts) were fused to the MLP-TSS-sRNA structural domain. B) HeLa cell were transfected with the MLP-TSS-RNA (TSS-neg), the miR26a-TSS or the miR155-TSS at two different concentration (10 and 50 mM). At 36 hours post transfection, western blot analysis was performed on total protein lysates to detect expression of two known targets for hsa-mir-26a-5p (Rb1 and Enx1 ) and hsa-mir-155-5p (Bachl ), respectively. Actin was used as a loading control. C) The stability of the chimeric TSS-sRNAs was assayed by transfecting HeLa cells with either the MLP-TSS-sRNA or the chimeric miR26a-TSS or miR155-TSS synthetic RNAs. Cells were harvested at the indicated time points and total RNA prepared and analysed by Northern blot analysis. 0.1 pmole stands for the amount of the synthetic ssRNA loaded directly to the gel as a size marker.

Fig 14. Comparing the efficiency of the hsa-mir-26a-5p commercial mimic and the chimeric miR26a-TSS. A) HEK293 were transfected with either MLP-TSS-RNA (TSS-neg), miR26a-TSS, hsa-mir-26a-5p commercial mimic (miR26a-mimic) or an unrelated control mimic (mimic neg) for 36 hours followed by western blot analysis detecting Enx1 protein expression. Actin was used as a loading control. B) Quantification of the relative expression of the Enx1 protein in panel A. The ratio of Enx1 /Actin in TSS- neg transfected cells was set as one. The data represent the mean of three biological replicates. Error bars indicate the standard deviation. C) MicroRNA mimics are more toxic than their TSS counterparts. Equal amount of HeLa cells seeded in a 24-well plate were transfected with the indicated RNAs and the cell viability measured after a 24 hour incubation.

Fig 15. The MLP-TSS-sRNA functional domain can be substituted with a siRNA. A) Schematic diagram of the double stranded siRNA (SEQ ID: 40 and 41) targeting the Dicer mRNA and the Dicer guide strand (19 nts) fused to the MLP-TSS-sRNA stability domain (SEQ ID NO: 42). B) HEK293 were transfected with either MLP-TSS-RNA (TSS-neg), Dicer-TSS, siRNA Scr or a commercially available Dicer siRNA for 36 hours followed by western blot analysis detecting the Dicer protein. LaminB was used as a loading control. C) Quantification of the relative expression of the Dicer protein in panel B. The ration of Dicer/LaminB in TSS-neg transfected cells was set as one. The data represent the mean of three biological replicates Error bars indicate the standard deviation.

Fig. 16 is a schematic illustration of an RNA molecule according to an embodiment; and

Figs. 17A and 17B are schematic illustrations of a hairpin RNA fragment according to different embodiments.

DETAILED DESCRIPTION The present invention generally relates to RNA molecules, and in particular to stable RNA molecules useful in gene expression control.

RNA interference is the process in which RNA molecules inhibit gene expression or translation by neutralizing target messenger RNA (mRNA) molecules. RNA interference is based on the action of sRNA molecules denoted miRNA and siRNA.

The traditional RNA interference pathway for siRNA is mediated by the enzyme Dicer, which cleaves long double-stranded RNA (dsRNA) molecules into short double-stranded fragments of about 22 nt siRNAs comprising a passenger strand and a guide strand. The siRNA is believed to unwind allowing the guide strand to be incorporated into the RISC, whereas the passenger strand is believed to be degraded. The guide strand in the RISC then base pairs with a complementary sequence in an mRNA molecule and induces cleavage by Argonaute 2 (Ago2), which constitutes the catalytic component of the RISC. miRNAs are genomically encoded non-coding RNAs involved in regulating gene expression. Mature miRNAs are structurally similar to siRNAs produced from exogenous dsRNA. However, miRNAs undergo extensive post-transcriptional modification in order to reach maturity. A miRNA is expressed from a much longer RNA-coding gene as a primary transcript known as a pri-miRNA, which is processed, in the cell nucleus, to a ~70-nucleotide stem-loop structure called a pre-miRNA by the microprocessor complex. This microprocessor complex includes an RNase III enzyme called Drosha and a dsRNA-binding protein DGCR8. The dsRNA portion of this pre-miRNA is bound and cleaved by Dicer to produce the mature miRNA molecule that can be integrated into the RISC complex. Thus, miRNA and siRNA share the same downstream cellular machinery. RNA interference (RNAi) molecule as used herein include any RNA molecule capable of RNA interference, i.e., inhibiting gene expression or translation of target mRNA molecules. RNAi molecule thereby includes siRNA molecules, and in particular the guide strand of such a siRNA molecule; miRNA molecules, and in particular the guide strand of such a miRNA molecule; and miRNA mimics. Correspondingly, RNA activation is the process in which RNA molecule induce gene expression by targeting promoters of genes. RNA activation is based on the action of sRNA molecules denoted saRNA, and sometimes agRNA. The molecular mechanism of RNA activation is not fully understood. Similar to RNA interference, it has been shown that mammalian RNA activation requires members of the Ago clade of Argonaute proteins, particularly Ago2, but possesses kinetics distinct from RNA interference. In contrast to RNA interference, promoter-targeted saRNAs induce prolonged activation of gene expression associated with epigenetic changes. It is currently suggested that saRNAs are first loaded and processed by an Ago protein to form an Ago-RNA complex, which is then guided by the RNA to its promoter target. The target can be a noncoding transcript overlapping the promoter or the chromosomal DNA. The RNA-loaded Ago then recruits other proteins, such as RHA, also known as nuclear DNA helicase II, and CTR9 to form an RNA-induced transcriptional activation (RITA) complex. RITA can directly interact with RNAP II to stimulate transcription initiation and productive transcription elongation, which is related to increased ubiquitination of H2B.

RNA activation (RNAa) molecule as used herein include any RNA molecule capable of RNA activation, i.e., inducing prolonged activation of gene expression. RNAa molecule thereby includes saRNA molecules, and in particular the guide strand of such a saRNA molecule; and agRNA moleucles, and in particular the guide strand of such an ag RNA molecule.

RNAi molecules and RNAa molecules may collectively be denoted RNA regulator molecules or RNA regulators as the RNA molecules are capable of regulating or controlling gene expression.

The RNA molecule of the embodiments is sometimes denoted sRNA molecule to indicate that it is a small RNA molecule, i.e., <200 nt.

Therapeutic siRNA and miRNA (mimics) have been proposed and developed for various applications by inhibiting specific genes through perfect base pairing the between the siRNA or miRNA and the target mRNA, thereby causing degradation of the target mRNA. Correspondingly, saRNA has been proposed to induce activation of target genes.

A problem with the prior art siRNA, miRNA and saRNA molecules is that they are dsRNA molecules. There is evidence that the passenger strand of the siRNA, miRNA and saRNA molecules may negatively affect the loading of the guide strand to the RISC or the RITA. Double-stranded siRNA, miRNA and saRNA molecules furthermore show unspecific toxicity, typically by activating components of the innate immune response. Accordingly, single-stranded sRNAs have been proposed as RNA interference molecules in order to improve the loading onto the RISC avoiding off-target effects associated with dsRNA molecules. However, such single-stranded sRNAs are rapidly degraded in vivo, thereby requiring chemical modifications in order to prevent or at least reduce degradation of the single-stranded sRNAs. Such chemical modifications necessitate extra process steps and costs during the manufacture of the chemically modified single-stranded sRNAs.

The present invention is related to an RNA molecule that is stabilized not by chemical modification of the nucleotides but by the inclusion of a stabilizing domain in the form of a hairpin RNA fragment. This hairpin RNA fragment stabilizes a RNA fragment and thereby prevents or at least reduces or inhibits degradation of the RNA fragment. The resulting RNA molecule with the RNA fragment and the stabilizing hairpin RNA fragment can still be loaded into the RISC and the RITA and thereby inhibit gene expression or translation of target mRNA molecules or induce activation of target genes. Hence, the RNA molecule with the hairpin RNA fragment can be used in gene control applications, such as RNA interference and RNA activation applications.

An aspect of the invention therefore relates to an RNA molecule 1 , see Figs. 16, 17A and 17B. The RNA molecule 1 comprises a RNA fragment 2 and a hairpin RNA fragment 3. The hairpin RNA fragment 3 comprises a first stem region 10, a loop region 20 and a second stem region 30. The first stem region 10 comprises the RNA sequence GCGUCG, the loop region 20 comprises the RNA sequence CUGUCCA and the second stem region 30 comprises the RNA sequence CGAGCGC.

The RNA molecule 1 of the invention is a single-stranded RNA molecule comprising two main fragments; the RNA fragment 2 and the hairpin RNA fragment 3. The RNA fragment 2 could be regarded as the functional domain of the RNA molecule 1 , such as in terms of having a nucleotide sequence that is complementary to a target nucleotide sequence in an mRNA molecule or a target nucleotide sequence in a promoter. Hence, the RNA fragment 2 is able to base pair with the target nucleotide sequence. The hairpin RNA fragment 3 is then regarded as a stabilizing domain of the RNA molecule 1 since it prevents or at least significantly reduces degradation of the functional domain, i.e., the RNA fragment 2. Accordingly, the stability of the RNA molecule 1 with the RNA fragment 2 and the hairpin RNA fragment 3 is significantly higher as compared to the stability of the RNA fragment 2 alone.

Complementary as used herein with regard to the RNA fragment 2 and the target nucleotide sequence ranges from fully complementary, i.e., all nucleotides in the RNA fragment 2 will base pair with a respective nucleotide in the target nucleotide sequence, to allowing one or a few mismatches. Hence, the RNA fragment 2 does not need to be 100 % complementary to the target nucleotide sequence. Thus, as long as the RNA fragment 2 is capable of binding to the target nucleotide sequence allowing the RISC to degrade the mRNA molecule comprising the target mRNA sequence or the RITA to activate the gene controlled by the target promoter sequence there is sufficient complementarity between the RNA fragment 2 and the target nucleotide sequence.

The first stem region 10 is capable of base pairing with the second stem region 30 as shown in Figs. 16, 17A and 17B. Hence, the two RNA sequences of the two stem regions 10, 30 are at least partly complementary to each other.

The two stem regions 10, 30 are complementary to each other with an exception of a G in the second stem region 30: GCG-UCG

I I I I I I CGCGAGC

In an embodiment, the hairpin RNA fragment 3 comprises, from a 5’ end 1 1 to a 3’ end 32, the first stem region 10, the loop region 20 and the second stem region 30.

In another embodiment, the hairpin RNA fragment 3 comprises, from a 5’ end 1 1 to a 3’ end 32, the second stem region 30, the loop region 20 and the first stem region 10. In this embodiment, the base pairing between the stem regions 10, 30 as according to below:

GCG-UCG In both above-described embodiments, the base pairing between the two stem regions 10, 30 forms a stem of the hairpin with the loop region 20 forming a loop in between the stem regions 10, 30. In an embodiment, the RNA fragment 2 is connected to the first stem region 10 or to the second stem 30. Thus, the 3’ end 42 of the RNA fragment 2 is connected to the 5’ end 1 1 of the hairpin fragment 3 or the 5’ end 41 of the RNA fragment 2 is connected to the 3’ end 32 of the hairpin fragment 3. In other words, the hairpin domain 3 could be provided upstream or downstream, preferably downstream, of the RNA fragment 2 with regard to 5’ to 3’ direction of the RNA molecule 1 . Hence, in an embodiment, the RNA molecule 1 comprises, from a 5’ end 41 to a 3’ end 32, the RNA fragment 2 and the hairpin RNA fragment 3.

In the embodiments as shown in Figs. 16, 17A and 17B, the 3’ end 42 of the RNA fragment 2 is connected to the 5’ end 1 1 of the hairpin fragment 2 and of the first stem region 10. This is a currently preferred embodiment.

As mentioned in the foregoing, the first stem region 10 comprises the RNA sequence GCGUCG. In an embodiment, the first stem region 10 consists of the RNA sequence GCGUCG.

As mentioned in the foregoing, the loop region 20 comprises the RNA sequence CUGUCCA. In an embodiment, the loop region 20 consists of the RNA sequence CUGUCCA.

The second stem region 30 comprises the RNA sequence CGAGCGC. In an embodiment, the second stem region 30 consists of a RNA sequence selected from the group consisting of CGAGCGC, CGAGCGCA, CGAGCGCU, CGAGCGCG and CGAGCGCC, more preferably from the group consisting of CGAGCGC and CGAGCGCC.

In an embodiment, the hairpin RNA fragment 3 comprises the RNA sequence GCGUCGCUGUCCACGAGCGC (SEQ ID NO: 1).

In a particular embodiment, the hairpin RNA fragment 3 consists of a RNA sequence selected from the group consisting of GCGUCGCUGUCCACGAGCGC (SEQ ID NO: 1 ),

GCGUCGCUGUCCACGAGCGCC (SEQ ID NO: 2), GCGUCGCUGUCCACGAGCGCG (SEQ ID NO: 35), GCGUCGCUGUCCACGAGCGCU (SEQ ID NO: 36) and GCGUCGCUGUCCACGAGCGCA (SEQ ID NO: 37), preferably from the group consisting of GCGUCGCUGUCCACGAGCGC (SEQ ID NO: 1 ) and GCGUCGCUGUCCACGAGCGCC (SEQ ID NO: 2). In an embodiment, the hairpin RNA fragment 3 has a formula selected from the group consisting of formula (I), see Fig. 17A, and formula (II), see Fig. 17B:

The RNA fragment 2 of the RNA molecule 1 could be any RNA sequence, preferably a RNA sequence that could bind to a target nucleotide sequence, such as bind to a target mRNA sequence to achieve a RNA interference or a target promoter sequence to achieve a RNA activation. Flence, the RNA fragment 2 could be selected from any siRNA, miRNA, miRNA mimic, saRNA, in particular the guide strand thereof, or a portion of such a siRNA, miRNA, miRNA, saRNA, in particular the guide strand thereof.

In an embodiment, the RNA fragment 2 consists of from 5 ribonucleotides up to 50 ribonucleotides, preferably from 5 ribonucleotides up to 25 ribonucleotides. The RNA fragment 2 preferably consists of from 10 ribonucleotides up to 20 ribonucleotides, and more preferably from 10 ribonucleotides up to 15 ribonucleotides.

Accordingly, the RNA molecule 1 is preferably a so-called small RNA molecule.

Experimental data as presented herein shows that the 5’ end 41 of the RNA fragment 2 could be unphosphorylated, phosphorylated or comprise a 7-methylguanosine (m7g) cap and still be loaded into the RISC. In an embodiment, the RNA molecule 1 is a chimeric RNA molecule. Flence, in such an embodiment, the RNA fragment 2 is different from the functional domain of the adenovirus major late protein (MLP) transcriptional stat site (TSS) small RNA (MLP-TSS-sRNA). Thus, in this embodiment the RNA fragment 2 is different from ACUCUCUUCC (SEQ ID NO: 3) and CUCACUCUCUUCC (SEQ ID NO: 4). Chimeric RNA molecule as used herein indicates that the RNA molecule is a hybrid molecule comprising a functional domain from one source and a stabilizing domain from another source, preferably the MLP- TSS-sRNA. Another aspect of the invention relates to a nucleotide sequence encoding an RNA molecule 1 according to any of the embodiments.

In an embodiment, the nucleotide sequence comprises the nucleotide sequence GCGTCGCTGTCCACGAGCGC (SEQ ID NO: 5), and preferably the nucleotide sequence GCGT CGCTGT CCACGAGCGCC (SEQ ID NO: 6).

A further aspect of the invention relates to an expression vector comprising a promoter operatively connected to a nucleotide sequence according to any of the above mentioned embodiments. The expression vector, sometimes referred to an expression construct, can be any vector capable of being transcribed in a host cell to produce the RNA molecule 1 according to any of the embodiments.

The promoter is operatively connected to the nucleotide sequence to enable transcription of the nucleotide sequence in the host cells. The expression vector may contain, in addition to the promoter, other regulator sequences, such as enhancers, to obtain an efficient transcription of the nucleotide sequence.

The promoter can be selected among constitutive promoters and regulated promoters, such as inducible promoters. The actual promoter of the nucleotide sequence is preferably selected based on the host cell, in which the expression vector is to be introduced in order to transcribe the nucleotide sequence into the RNA molecule 1 .

Non-limiting, but illustrative, examples of promoters that can be used in the expression vector include lac promoter, T7 promoter, TAC promoter, AOX1 promoter, LAC4 promoter, vir promoter, CaMV 35S promoter, Ubi promoter, Act-1 promoter, Adh-1 promoter, CMV promoter, SV40 promoter, and EF-1 promoter. In an embodiment, the expression vector is selected among a plasmid vector and a viral vector. Nonlimiting, but illustrative, examples of viral vectors include retroviruses, lentivirus, adenoviruses, adeno- associated viruses and herpes simplex viruses. The expression vector may include nucleotide sequences encoding other molecules in addition to the RNA molecule 1 . For instance, the expression vector could comprise a nucleotide sequence encoding a ribozyme. Ribozymes are RNA molecules that are capable of catalyzing specific biochemical reactions, similar to the action of protein enzymes. The most common activities of natural or in vitro-ev olved ribozymes are the cleavage or ligation of RNA and DNA and peptide bond formation.

In an embodiment, the nucleotide sequence encoding the ribozyme is preferably arranged downstream of the nucleotide sequence encoding the RNA molecule 1 with regard to the promoter in the expression vector. In another embodiment, the nucleotide sequence encoding the ribozyme is arranged upstream of the nucleotide sequence encoding the RNA molecule 1 .

In a further embodiment, a first nucleotide sequence encoding a first ribozyme is arranged upstream of the nucleotide sequence encoding the RNA molecule 1 and a second nucleotide sequence encoding a second ribozyme is arranged downstream of the nucleotide sequence encoding the RNA molecule 1 .

The ribozyme is then preferably selected to enable cleavage of the resulting RNA transcript obtained from the expression vector in a host cell to release the RNA molecule 1 from the resulting RNA transcript. Hence, in an embodiment, the ribozyme is preferably configured to cleave the resulting RNA transcript transcribed from the expression vector at the 3’ end 32 of the RNA molecule 1 to thereby release the RNA molecule 1 .

In another embodiment, the ribozyme is preferably configured to cleave the resulting RNA transcript transcribed from the expression vector at the 5’ end 41 of the RNA molecule 1 .

It is further possible to have a ribozyme capable of cleaving the resulting RNA transcript at both the 5’ end 41 and the 3’ end 32 of the RNA molecule 1 . Alternatively, the expression vector could include a first nucleotide sequence encoding a first ribozyme capable of cleaving the resulting RNA transcript at the 5’ end 41 of the RNA molecule 1 and a second nucleotide sequence encoding a second ribozyme capable of cleaving the resulting RNA transcript at the 3’ end 32 of the RNA molecule 1 .

Yet another aspect of the invention relates to a cell comprising a nucleotide sequence according to any of the embodiments and/or an expression vector according to any of the embodiments. The cell is, in an embodiment, selected from the group consisting of a bacterial cell; a yeast cell; a plant cell; a mammalian cell, such as a human cell, a mouse cell, a rat cell, a dog cell, a cat cell, a cow cell, a horse cell, a sheep cell, a goat cell, etc.

The nucleotide sequence and/or the expression vector could be present in the cell as extrachromosomal nucleotide sequence and/or expression vector. Alternatively, the cell could be stably transfected by the nucleotide sequence and/or expression vector so that the nucleotide sequence and/or expression vector is integrated into the genome of the cell.

The RNA molecule 1 of the present embodiments can be used in RNA interference applications that traditionally employ siRNAs, miRNAs and/or miRNA mimics. Hence, the RNA molecule 1 could be used to treat or inhibit a disease or disorder characterized by deleterious expression of a gene and where silencing of the expression of that gene would treat or at least inhibit the disease or disorder.

RNA interference has been used in various therapeutic applications including viral infections, cancer, and neurological diseases.

Antiviral treatment is one of the earliest proposed RNA interference-based medical applications, and two different types have been developed. The first type is to target viral RNAs. Many studies have shown that targeting viral RNAs can suppress the replication of numerous viruses, including HIV, HPV, hepatitis A hepatitis B, Influenza virus, and Measles virus. The other strategy is to block the initial viral entries by targeting the host cell genes. For example, suppression of chemokine receptors (CXCR4 and CCR5) on host cells can prevent HIV viral entry. While traditional chemotherapy can effectively kill cancer cells, lack of specificity for discriminating normal cells and cancer cells in these treatments usually cause severe side effects. Numerous studies have demonstrated that RNA interference can provide a more specific approach to inhibit tumor growth by targeting cancer-related genes, e.g., oncogenes. It has also been proposed that RNA interference can enhance the sensitivity of cancer cells to chemotherapeutic agents, providing a combinatorial therapeutic approach with chemotherapy. Another potential RNA interference-based treatment is to inhibit cell invasion and migration.

Compared with chemotherapy or other anti-cancer drugs, there are a lot of advantages of RNA interference drug. RNA interference acts on the post-translational stage of gene expression, so it does not modify or change DNA in a deleterious effect. RNA interference can also be used to produce a specific response in a certain type of way, such as by downgrading suppression of gene expression. In a single cancer cell, RNA interference can cause dramatic suppression of gene expression with just several copies. This happens by silencing cancer-promoting genes with RNA interference, as well as targeting an mRNA sequence.

RNA interference strategies also show potential for treating neurodegenerative diseases. Studies in cells and in mouse have shown that specifically targeting Amyloid beta-producing genes, e.g., BACE1 and APP, by RNA interference can significantly reduce the amount of Amyloid beta peptide, which is correlated with the cause of Alzheimer's disease. In addition, this silencing-based approaches also provide promising results in treatment of Parkinson's disease and polyglutamine disease.

Clinical Phase I and II studies of RNA interference therapies have demonstrated potent and durable gene knockdown in the liver, with some signs of clinical improvement and without unacceptable toxicity. Two Phase III studies are in progress to treat familial neurodegenerative and cardiac syndromes caused by mutations in transthyretin (TTR). Numerous publications have shown that in vivo delivery systems are very promising and are diverse in characteristics, allowing numerous applications. The nanoparticle delivery system shows promise. Examples of diseases and disorders targeted by RNA interference include cancer, targeting the genes KSP, VEGF, PKN3, EphA2, PLK1 , RRM2, KRAS; Ebola virus infection, targeting the genes VP24, VP35, Zaire Ebola L-polymerase; respiratory syncytial virus infections, targeting the gene RSV nucleocapsid; hypercholesterolemia, targeting the genes ApB, PCSK9; transthyretin-mediated amyloidosis, targeting the gene TTR; pachyonychia congenital, targeting the gene K6a (N171 K mutation); macular degeneration, targeting the genes VEGFR1 , VEGF; choroidal neovascularization, targeting the gene VEGFR1 ; optic atrophy and non-arteritic anterior ischemic optic neuropathy, targeting the gene CASP2; kidney injury and acute renal failure targeting the gene p53; ocular pain and dry-eye syndrome, targeting the gene TRPV1 ; ocular hypertension and open-angle glaucoma, targeting the gene ADRB2, familial adenomatous polyposis, targeting the gene CTNNB1 , and cicatrix scar prevention, targeting the gene CTGF.

The hairpin RNA fragment as disclosed herein can connected to any known siRNA, miRNA and miRNA mimic sequence, and in particular the guide strand thereof, as RNA fragment in order to stabilize the RNA fragment and inhibit degradation thereof. A resulting RNA molecule can then be used in any of the above mentioned RNA interference applications.

RNA activation has been studied in vivo, including its potential therapeutic applications in treating cancer and non-cancerous diseases, such as erectile dysfunction.

The hairpin RNA fragment as disclosed herein can connected to any known saRNA sequence, and in particular the guide strand thereof, as RNA fragment in order to stabilize the RNA fragment and inhibit degradation thereof. A resulting RNA molecule can then be used in any of the above mentioned RNA activation applications.

Thus, a further aspect of the invention includes an RNA molecule 1 according to any of the embodiments, a nucleotide sequence according to any of the embodiments, and/or an expression vector according to any of the embodiments for use as a medicament.

The invention also relates to a pharmaceutical composition comprising an RNA molecule 1 according to any of the embodiments, a nucleotide sequence according to any of the embodiments and/or an expression vector according to any of the embodiments. The pharmaceutical composition comprises the RNA molecule 1 , the nucleotide sequence and/or the expression vector in a pharmaceutically effective amount. Pharmaceutically effective amount as used herein refers to an amount that is necessary for an active ingredient of a pharmaceutical composition, i.e., functional RNA fragment 2 of the RNA molecule 1 , to exert its gene silencing or activating functions that imposes no or substantially no side effects harmful to organisms to which the pharmaceutical composition is administered. The specific dose varies depending on the type of functional RNA fragment 2 used, the target molecule, the dosage form to be employed, information regarding the subject, and the route of administration. When a pharmaceutical composition is administered to a human, a pharmaceutically effective amount and a preferable route of administration are generally determined based on the data obtained as a result of cell culture assays and animal experimentation. The final dose is determined and adjusted by a doctor in accordance with the individual subject. In such a case, examples of information regarding the subject that is to be taken into consideration include the extent or severity of a disease, general physical conditions, age, body weight, sexuality, eating habits, drug sensitivity, and resistance to treatment and the like.

In an embodiment, the pharmaceutical composition further comprises a pharmaceutically acceptable carrier. Such a pharmaceutically acceptable carrier could be a solvent and/or additive that is generally used in the pharmaceutical field.

Illustrative, but non-limiting, examples of solvents include water, saline, buffer, glucose solution and pharmaceutically acceptable organic solvents, e.g., ethanol, propylene glycol, ethoxylated isostearyl alcohol, polyoxylated isostearyl alcohol, and polyoxyethylene sorbitan fatty acid esters. Such solvents are preferably sterilized. It is preferable that the solvents is adjusted to be isotonic with blood, as the saline, as necessary.

Non-limiting, but illustrative, examples of additives include excipients, adsorption inhibitors, binders, disintegrators, fillers, emulsifiers, flow modifiers, and lubricants.

Non-limiting, but illustrative, examples of excipients include saccharides, such as monosaccharides, disaccharides, cyclodextrins, and polysaccharides, such as glucose, sucrose, lactose, raffinose, mannitol, sorbitol, inositol, dextrin, maltodextrin, starch, and cellulose; metal salts, such as sodium phosphate or calcium phosphate, calcium sulfate, and magnesium sulfate; citric acids; tartaric acids; glycine; low-, middle-, or high-molecular weight polyethylene glycol (PEG); Pluronic, and combinations of any thereof. Non-limiting, but illustrative, examples of adsorption inhibitors include Tween 80, Tween 20, gelatin, and/or human serum albumin.

Non-limiting, but illustrative, examples of binders include starch pastes using maize, wheat, rice, or potato starch, gelatin, Tragacanth, methyl cellulose, hydroxypropyl methylcellulose, carboxymethylcellulose sodium, and/or polyvinyl pyrrolidone.

Non-limiting, but illustrative, examples of disintegrators include the aforementioned starch, carboxymethyl starch, crosslinked polyvinyl pyrrolidone, agar, alginic acid, sodium alginate, and salts of any thereof.

Non-limiting, but illustrative, examples of fillers include the aforementioned sugar and/or calcium phosphate, e.g., tricalcium phosphate and calcium hydrogen phosphate. Non-limiting, but illustrative, examples of emulsifiers include sorbitan fatty acid ester, glycerine fatty acid ester, sucrose fatty acid ester, and propylene glycol fatty acid ester.

Non-limiting, but illustrative, examples of flow modifiers and lubricants include silicate, talc, stearate, and polyethylene glycol.

The pharmaceutical composition is preferably administered in a dosage unit form. The pharmaceutical composition can be administered through an oral route, directly into the tissue, e.g., subcutaneous, intramuscular, or intravenous administration, or outside the tissue, e.g., percutaneous, instillation, nasal, or transrectal administration. The pharmaceutical composition of the present invention is preferably administered in a dosage form adequate for the method of administration. When the pharmaceutical composition is directly administered into tissue, for example, injection through the blood stream is preferable. Thus, the dosage form may be an injection solution.

When the pharmaceutical composition is administered in the form of an injection, the site of injection is not particularly limited, provided that the RNA molecule 1 of the present invention can exert its functions on the target nucleotide sequence and achieve an objective of the administration of the pharmaceutical composition. Examples include intravenous, intraarterial, intrahepatic, intramuscular, intraarticular, intramedullary, intraspinal, intraventricular, percutaneous, subcutaneous, intracutaneous, intraperitoneal, intranasal, intestinal, and sublingual sites. An injection into a blood vessel, such as intraveneous or intraarterial injection, is preferable since this allows the pharmaceutical composition of the present invention to be immediately distributed throughout the body via the blood stream and invasiveness is relatively low. Alternatively, the pharmaceutical composition may be injected directly into a site at which pharmacological effects of the pharmaceutical composition of the present invention are needed, thus allowing a large amount of the pharmaceutical composition to directly act on the target site.

A further aspect of the invention relates to a method of treating or inhibiting a disease or disorder in a subject. The disease or disorder is characterized by deleterious expression of a gene in the subject. The method comprises administering a pharmaceutical composition according to any of the embodiments to the subject. The RNA fragment 2 of the RNA molecule 1 is complementary to a target mRNA sequence in an mRNA molecule transcribed from the gene.

A related aspect of the invention defines a method of treating or inhibiting a disease or disorder in a subject. The method comprises administering a pharmaceutical composition according to any of the embodiments to the subject. The RNA fragment 2 of the RNA molecule 1 is complementary to a target promoter sequence controlling transcription of a gene encoding a molecule useful in treating or inhibiting the disease or disorder. Yet another aspect of the invention relates to a method of controlling expression of a gene. The method comprises contacting an mRNA molecule encoded by the gene with an RNA molecule 1 according to any of the embodiments. The RNA fragment 2 of the RNA molecule 1 is complementary to a target mRNA sequence in the mRNA molecule. A related aspect of the invention defines a method of controlling expression of a gene. The method comprises contacting a promoter of the gene with a RNA molecule 1 according to any of the embodiments. The RNA fragment 2 of the RNA molecule 1 is complementary to a target nucleotide sequence in the promoter. In an embodiment, contacting the mRNA molecule or promoter comprises introducing the RNA molecule 1 according to any of the embodiments, a nucleotide sequence according to any of the embodiments and/or an expression vector according to any of the embodiments into a cell capable of transcribing the gene into the mRNA molecule. The nucleotide sequence according to any of the embodiments can be transcribed in the cell and/or the promoter of the expression vector according to any of the embodiments is active or inducible active in the cell.

A further aspect of the invention relates to a method for enhancing resistance of an RNA fragment 2 to degradation by a nucleolytic enzyme. The method comprising providing, at a 5’ end 41 or a 3’ end 42 of the RNA fragment 2, a hairpin RNA fragment 3. The hairpin RNA fragment 3 comprises a first stem region 10 comprising the RNA sequence GCGUCG, a loop region 20 comprising the RNA sequence CUGUCCA and a second stem region 30 comprising the RNA sequence CGAGCGC. The hairpin RNA fragment 3 can be ligated to the 5’ end 41 or the 3’ end of the RNA fragment 2 in an embodiment. In another embodiment, a synthesis method can be used to produce an RNA molecule 1 comprising the RNA fragment 2 and the hairpin RNA fragment 3. In a further embodiment, a synthesis method is used to produce a nucleotide sequence encoding the RNA molecule 1 comprising the RNA fragment 2 and the hairpin fragment 3.

EXAMPLES

EXAMPLE 1

This example shows that the adenovirus major late promoter (MLP) produces a 31 -nucleotide transcriptional start site (TSS) small RNA (MLP-TSS-sRNA) that retains the 7-methylguanosine (m7G) cap and is incorporated onto Ago2-containing RNA-induced silencing complexes (RISC) in human adenovirus-37 infected cells. RNA polymerase II CLIP (UV-cross linking immunoprecipitation) experiments suggest that the MLP-TSS-sRNA is produced by promoter proximal stalling/termination of RNA polymerase II transcription at the site of the small RNA 3’ end. The MLP-TSS-sRNA is highly stable in cells and functionally active, downregulating complementary targets in a sequence and dose dependent manner. The MLP-TSS-sRNA is transcribed from the opposite strand to the adenoviral DNA polymerase and pre-terminal protein mRNAs, two essential viral replication proteins. The MLP-TSS-sRNA is shown to act in trans to reduce DNA polymerase and pre-terminal protein mRNA expression. As a consequence, the MLP-TSS-sRNA has an inhibitory effect on the efficiency of viral DNA replication. The results suggest that this novel sRNA may serve a regulatory function controlling viral genome replication during a lytic and/or persistent adenovirus infection in its natural host. MicroRNAs (miRNAs) are abundant 22 nucleotide (nt) regulatory RNAs, typically derived from endogenous transcripts, with short hairpins, that modulates essential cellular processes at the posttranscriptional level in most eukaryotes. Canonical miRNAs are produced from long primary miRNA transcripts by two RNaselll like cleavage events. First, the primary miRNA is processed by the Drosha/DGCR8 microprocessor complex in the nucleus into a 60-80 nt precursor miRNA hairpin that is exported to the cytoplasm. Here the Dicer enzyme cleaves the loop of the hairpin to produce a ~22 base pair mature miRNA duplex. Based on the relative thermodynamic stability of the duplex ends one or both of the strands are loaded onto one of four Argonaute (Ago) proteins to form the RNA-induced silencing complex (RISC). The RISC complex function through miRNA-directed base pairing to target mRNA transcripts for posttranscriptional gene silencing.

Human adenoviruses (HAdVs) encode for one or two virus-associated RNAs (VA RNAs), designated as VA RNAI and VA RNAII. The VA RNAs are approximately 160 nt long non-coding RNAs, transcribed by RNA polymerase III that fold into highly structured RNAs that show features similar to pre-miRNAs. The terminal stem of the VA RNAs from several HAdVs is processed by Dicer into small viral miRNAs, so- called mivaRNAs, which associate with Ago2 containing RISC. The HAdV-5 mivaRNAs functions as miRNAs that regulates host cell gene expression by targeting complementary sequences in the 3’-UTR of an mRNA. However, the significance of these interactions for a lytic infection, at least in tissue culture cells, has been questioned. Although most of the mivaRNA research has been done in the HAdV-5 model system, the mivaRNAs expressed in HAdV-1 1 - and HAdV-37-infected cells are more abundant and more efficiently associated with Ago2-containing RISC.

In this example, a novel mechanism to produce an Ago2 associated small RNA is demonstrated. It is shown that the promoter proximal region of the adenovirus major late promoter (MLP) produces a transcriptional start site (TSS) small RNAs, denoted MLP-TSS-sRNAs herein, which associates with Ago2 in HAd-infected cells. The MLP produces a major 31 nt MLP-TSS-sRNA with a 5’ end that coincides with the MLP +1 start site. The MLP-TSS-sRNAs were detected in all HAds analyzed (HAdV-4, HAdV-5, HAdV-1 1 and HAdV-37) at varying levels. The HAdV-37 MLP-TSS-sRNAs contains a 7-methylguanosine (m7G)-cap and are enriched in Ago2 containing RISC. The MLP-TSS-sRNAs were capable of inhibiting reporter gene expression and had a capacity to suppress viral DNA replication during a lytic virus infection. Further, it is shown that the HAdV-37 MLP-TSS-sRNAs, in contrast to the mivaRNAs, are produced by a distinct pathway not requiring the Exportin5 nuclear export receptor or the Dicer enzyme. The data suggests that the MLP-TSS-sRNAs is generated by a novel mechanism where a short prematurely terminated RNA polymerase II (RNAPII) transcript becomes Ago2 associated.

RESULTS

Production of RISC associated sRNAs from the HAdV-37 MLP transcriptional start site

In a recent study the pool of viral sRNAs from several adenovirus infections were sequenced (Kamel et al. 2014), primarily to map the structure and abundance of the VA RNA derived mivaRNAs expressed in the different serotypes. As expected, the vast majority of sRNAs mapped to the VA RNA region (Fig. 1 B). However, at a closer examination of the data it was noted that a small fraction of the sRNAs enriched in RISC mapped to the MLP TSS (Fig. 1 C) with a major 31 nt species with a 5’ end at the classical MLP +1 cap site and a minor 34 nt species (20-fold lower) with a 5’ end located at position -3 relative to the MLP cap site. Since both MLP-TSS RNAs were enriched in immuno-purified Flag-Ago2 containing RISC, they are referred to as (+1 and -3 start site RNAs) as MLP-TSS-sRNAs. The same MLP-TSS-sRNAs were detected in all serotypes analyzed (HAdV-4, HAdV-5, HAdV-1 1 and HAdV-37) but were most highly expressed in HAdV-37-infected cells (Table 1 ). Therefore all subsequent experiments were done using this virus as the model virus.

Table 1 - Summary of viral sRNA accumulation from the MLP-TSS region in different adenovirus serotypes

The HAdV-37 +1 and -3 MLP-TSS-sRNAs share a major 3’ end, located 9 nts upstream of the first leader 5’ splice site (Fig. 1 C). The length distribution of the +1 TSS sRNAs was predominantly of 31 nts with a trailing fraction generating a second peak at 35 nts in length and minor peaks of 24 and 32 (Fig. 1 D). The -3 MLP-TSS-sRNAs had a major peak at 34 nts but a larger heterogeneity of both shorter and longer types (Fig. 1 D).

The data shows that the HAdV-37 MLP-TSS-sRNAs are the most abundant RISC-associated adenoviral sRNAs outside the VA region (Fig 1 B), accumulating to levels comparable to many of the most abundant cellular miRNAs expressed in HAdV-37-infected cells (Fig. 2).

The 5’ end of the HAdV-37 MLP-TSS-sRNAs are modified

To characterize the MLP-TSS-sRNAs carbodiimide-mediated cross-linking of RNA to a nylon membrane was used (Pall and Hamilton 2008). This method requires a free 5’ phosphate group for chemical binding to the membrane. This feature of the method was used to investigate whether the 5’ end of the MLT- TSS-sRNAs was modified. Since the 5’ end of the large majority of these RNAs coincide with the well- characterized MLP cap site tobacco acid pyrophosphatase (TAP) or RNA 5' Pyrophosphohydrolase (RppH) treatment was used to determine whether they contain a 5’ end modification that hinders cross- linking of the RNAs to the membrane. The TAP and RppH enzymes remove the m7G-cap from mRNAs leaving a 5’ monophosphate group. As shown in Fig. 3, TAP treatment of total cytoplasmic RNA from HAdV-37-infected cells increased the cross-linking efficiency of MLP-TSS-sRNA by 13 to 15 fold (lanes 3 and 4). In contrast, crosslinking of HAdV-37 VA RNAII and tRNA-lysine, which are RNA polymerase III products, or let-7b miRNA were not enhanced by TAP treatment (lanes 3 and 4). Similar results were obtained using RppH (Fig. 4). Based on these findings the RNA was TAP treated before performing Northern blot analysis in all subsequent experiments. Interestingly, a series of longer transcripts were also enhanced after TAP or RppH treatment (Fig. 3, lane 4, Fig. 4). These RNAs most likely correspond to promoter proximal pre-terminated MLP transcripts (see below). As expected from the expression profile of the MLP and the VA RNAs the MLP-TSS-sRNAs starts to accumulate towards the late phase of infection (12 to 18 hours post infection (hpi)) whereas the full length VA RNAII was detectable already at 6 hpi and the processed imivaRNAII at 12 hpi (Fig. 5).

The HAdV-37 MLP-TSS-sRNAs are capped and enriched in Ago2 containing RISC

To determine whether the Ago2-associated MLP-TSS-sRNAs contain an m7G-cap Flag-Ago2 complexes from cytoplasmic extracts prepared from HAdV-37-infected 293-FlagAgo2 cells at 24 hpi were immuno- purified. The processed 5’ mivaRNAII was enriched in the Ago2-RISC (Fig. 6A, lane 4). In agreement with the sRNA sequencing data the MLP-TSS-sRNA was also found in the Ago2-RISC (Fig. 6A, lane 4). As a control of specificity, it was shown that the actin protein did not contaminate the Ago2-RISC fraction (Fig. 6A, lanes 3 and 4). The Ago2-RISC RNAs were purified by phenol extraction and subjected to a second round of immunoprecipitation using an anti-m7G-cap monoclonal antibody. As shown in Fig. 6B, the MLP-TSS-sRNAs were recovered in the m7G immunoprecipitate (lane 4), whereas the 5’ mivaRNAII and the Iet7b miRNA, as expected, were not detectable in the m7G immunoprecipitate (lanes 3 and 4). These results suggest that the MLP-TSS-sRNA retain the m7G-cap added during the transcription initiation process. Further, the TAP treatment of the cytoplasmic RNA fraction (Fig. 3) strongly argues that at least 95% of the MLP-TSS-sRNAs are capped.

The HAdV-37 MLP-TSS-sRNAs are produced by a pathway distinct from the mivaRNAs

Previous studies have shown that VA RNA export requires Exportin5 (Yi et al. 2003; Lu and Cullen 2004) and that the mivaRNAs are produced by Dicer cleavage (Kamel et al. 2014). Considering the novel features of the MLP-TSS-sRNA, it was of interest to study whether the mivaRNAs and the MLP-TSS- sRNAs require the same cellular machinery for export and processing. To test this an siRNA approach was used to knockdown candidate proteins. As shown in Fig. 7A, Dicer knockdown reduced 5’ mivaRNA processing whereas MLP-TSS-sRNA production was unaffected (lanes 2, 3). Also knockdown of the Dis3 RNAse in the nuclear exosome complex did not impair the maturation of the MLP-TSS-sRNA or the 5’ mivRNAII (lane 7). Knockdown of Exportin5 reduced mivaRNA production (lane 6). In contrast, targeting Exportin5 did not impair export of the MLP-TSS-sRNA. To investigate whether MLP-TSS-sRNA production is dependent on the Ago2 slicer activity, like mir-451 (Cheloufi et al. 2010; Cifuentes et al. 2010), HEK293 cells were transfected with plasmids expressing Flag-epitope-tagged wild type Ago2 or the catalytically dead Ago2 (D669A) mutant proteins. Cells were infected with HAdV-37 and the sRNA content in the cytoplasmic Ago2 fractions analyzed at 24 hpi. The result shows that the Ago2 catalytic activity is not required for MLP-TSS-sRNA production (Fig. 7C, lanes 9 and 10). The analysis further demonstrated that the MLP-TSS-sRNA also associates with Ago1 , which lacks slicer activity.

The MLP-TSS-sRNA is an RNA polymerase II transcription termination product

In order to gain further insights about the biogenesis of the MLP-TSS-sRNA, the sequence of the potential precursor transcript(s) that could be the origin of MLP-TSS-sRNA was investigated. For this experiment 293-FlagAgo2 cells were transfected with three pGL4 reporter plasmids (Fig. 8A) for 24 hours followed by total cytoplasmic RNA extraction and Northern blot analysis. Reporter plasmid UnMLT corresponds to the unspliced precursor RNA, and contains the MLP fused to the first 200 nts downstream of the transcriptional start site. Plasmid Trip contained the MLP fused to a cDNA copy of the spliced tripartite leader sequence, whereas reported plasmid RevT rip contained a cDNA copy of the tripartite leader where the second and third exons had been inserted in the reverse orientation. As shown in Fig. 8B, all three reporter plasmids accumulated a small RNA around 31 nts long. This result indicates the MLP-TSS- sRNAs can be produced from both the unspliced MLP precursor RNA and the spliced tripartite leader.

Most interestingly, the MLP-TSS-sRNA was also produced in cells transfected with the plasmid containing a reversed copy of the tripartite leader exons 2 and 3. This plasmid contains the MLP-TSS- sRNA sequence plus only two extra nucleotides from the first leader exon (Fig. 1 C). These results argue that the MLP-TSS-sRNA is not processed from a precursor RNA requiring downstream sequences beyond two nucleotides from the MLP-TSS-sRNA 3’ end. Alternatively the MLP-TSS-sRNA could be generated as a direct product of RNAPII stalling/termination downstream of the MLP initiation site. To test this hypothesis it was determined whether an increased accumulation of RNAPII was detected at the site of the MLP-TSS-sRNA 3’ end. For this experiment 293Ago2 cells were infected with HAdV-37 or transfected with plasmid UnMLT. Following 24 hours incubation cells were UV-irradiated to halt further transcription, and RNAPII complexes isolated by immunoprecipitation. The supernatant after RNAPII-IP was used to immunopurify Ago2 complexes and the sRNA content in both fractions analyzed by Northern blot. As shown in Fig. 8C, a distinct band (lanes 4-6) of the size of the Ago2-associated MLP-TSS-sRNA (lane 8) appeared in the RNAPII precipitate particularly in HAdV-37-infected cells (lane 5). The same band was also recovered from plasmid-transfected cells, although with a considerably lower efficiency (lane 6). The more pronounced site for RNAPII stalling in the transfected cells was a site located approximately 60 nts downstream of the MLP TSS. This sRNA can be seen in most experiments (lane 5 and Figs. 2-5 and 7-9) and is also Ago2-RISC associated (lanes 8, 9). The specificity of the immunoprecipitations was confirmed by probing the same membrane for VA RNAII and tRNA lysine, which are RNA polymerase III products. The specificity of the immunoprecipitations was also confirmed by Western blotting of the indicated proteins (Fig. 8D).

These results are compatible with the hypothesis that the MLP-TSS-sRNA is an RNAPII transcript produced by RNA polymerase II stalling/termination shortly after initiation at the MLP.

The HAdV-37 MLP-TSS-sRNAs repress complementary targets

Importantly, the MLP-TSS-sRNA associates with the endogenous Ago2 protein in HAdV-37-infected ARPE-19 (Fig. 9A). The specificity of this interaction was confirmed by the lack of tRNA lysine and the GAPDH protein in the Ago2 immunoprecipitate (Fig. 9B).

To determine whether the MLP-TSS-sRNAs loaded onto RISC are functional, the capacity of the complex to regulate expression of a firefly luciferase reporter construct was tested with a complementary MLP- TSS-sRNA binding site in the 3’ UTR (pmirGlo(+), Fig. 10A). For this experiment ARPE-19 cells were first infected with two concentrations of HAdV-37 followed by transfection of reporter plasmids pmirGlo, lacking an MLP-TSS-sRNA binding site, or pmirGlo(+) at 1 hpi. Cells were harvested at 24 hpi and the effect of the HAdV-37 infection on luciferase expression measured. As shown in Fig. 10B, the relative luciferase expression was progressively reduced in pmirGlo(+)-infected cells whereas luciferase expression in cells transfected with the pmirGlo reporter construct was unaffected. To further investigate the specificity of this interaction, the MLP-TSS-sRNA binding site was cloned in the reverse orientation, generating plasmid pmirGlo(Rev) (Fig. 10A). As shown in Fig. 10C, an HAdV-37 infection significantly suppressed gene expression from pmirGlo(+) without having a negative effect on luciferase expression from reporter plasmids pmirGlo or pmirGlo(Rev). Taken together, these results suggest that during a HAdV-37 infection RISC programmed with the MLP-TSS-sRNAs are functionally active capable of suppressing reporter gene expression in a sequence specific and dose dependent manner.

To confirm this observation a reductionist approach was used to minimize the potentially complex effects caused by the expression of the whole repertoire of viral proteins. For this experiment the m7G-capped MLP-TSS-sRNA was produced by in vitro transcription and co-transfected with luciferase reporter plasmids pmirGlo, pmirGlo(+) or pmirGlo(mut), where the MLP-TSS-sRNA binding site was mutated. As shown in Fig. 10D, the MLP-TSS-sRNA was capable of down-regulating luciferase expression from the reporter containing a complementary binding site. Moreover mutations in this binding site caused a complete loss of the inhibitory effects on luciferase expression. This experiment is significant because it suggests that the MLP-TSS-sRNA is capable of regulating gene expression in a sequence specific manner also in the absence of other viral components.

The stability of the MLP-TSS-sRNA is dependent on a 3’ hairpin structure

Interestingly, the m7G-cap did not contribute significantly to the stability of the MLP-TSS-sRNA. Thus, the MLP-TSS-sRNA with a 5’ phosphate was as stable as the capped counterpart in transfected 293- Flag-Ago2 cells (Fig. 9C). On the other hand, secondary structure prediction identified a potential short hairpin at the 3’ end of the MLP-TSS-sRNA (Fig. 10E). To investigate whether this hairpin was important as a stabilizing element this secondary structure was disrupted by deleting the last 12 nucleotides from the MLP-TSS-sRNA, generating MLP-TSS-sRNA(trunc). The full-length and truncated MLP-TSS-sRNA synthetic RNAs where transfected into 293Ago2 cells and small RNA stability measured in total RNA at 12, 24, 48, 72 hour post transfection by Northern blot. As shown in Fig. 10F, the full-length MLP-TSS- sRNA was highly stable during the course of the experiment (lanes 3-6) whereas the MLP- TSSsRNA(trunc) was rapidly degraded and barely detectable at any of the time points assayed (lanes 8- 1 1 ). Importantly, the input of synthetic sRNA was the same for the full-length and truncated MLP-TSS- sRNA (lanes 1 and 2). The low stability of MLP-TSS-sRNA(trunc) suggests that it should be inefficient as a suppressor of target mRNAs. As shown in Fig. 10G, MLP-TSS-sRNA efficiently repressed luciferase expression during the three day experiment whereas MLP-TSS-sRNA(trunc) caused moderate repression at 24 hours post transfection, a repression that was lost at the later time points.

These data suggest that the hairpin structure at the 3’ end of the MLP-TSS-sRNA is critical for stability and consequentially also important for the sRNA function.

The MLP-TSS sRNA suppresses HAdV-37 DNA replication

The adenovirus E2B region encodes for two viral proteins that are essential for viral DNA replication: the preterminal protein (pTP), which is the primer protein needed to initiate viral DNA replication and the Adpol, which is the DNA polymerase copying the viral genome. Interestingly, the MLP-TSS-sRNA originates from the opposite strand to the adenovirus E2B region (Figs. 1A and 1 1 A). Thus, the MLP- TSS-sRNA is perfectly complementary to the pTP and Adpol mRNAs and might therefore be functioning as a regulator repressing virus growth by targeting the expression of the E2B mRNAs. To determine whether the expression level of pTP and Adpol are key factors controlling subsequent steps in virus replication, like DNA replication and late protein synthesis, ARPE-19 cells were transfected with a synthetic MLP-TSS-sRNA or a scrambled sRNA for four hours, followed by infection with HAdV-37. The transfection approach would mimic an overexpression of the MLP-TSS-sRNA. As shown in Fig. 1 1 B, transfection of the MLT-TSS-sRNA reduced expression of both Adpol and pTP mRNAs by more than 50%. As a consequence it also caused a decrease in the efficiency of viral DNA replication (Fig. 1 1 C), which was accompanied by a reduction in viral capsid proteins synthesis (Fig. 1 1 D).

Although these results suggest that the synthetic MLP-TSS-sRNA can target the Adpol and pTP mRNAs, it was investigated whether the MLP-TSS-sRNA produced from the viral genome would exhibit a similar inhibitory effect. For this experiment the same plasmid transfection and virus infection approach as described in Fig. 10 were used. HeLa cells were infected with HAdV-37 followed by transfection with pmirGlo(+) or pmirGlo lacking a binding site (Fig. 10A). The rational in this experiment was that the pmirGlo(+) would function as a MLP-TSS-sRNA sponge that would sequester the virus produced MLP- TSS-sRNA and thereby alleviate the inhibitory effect of the sRNA on Adpol and pTP mRNA expression As shown Fig. 1 1 E, transfection with pmirGlo(+) caused, indeed, a significant upregulation of both Adpol, and pTP mRNA accumulation. To further confirm this result stable HeLa cell lines constitutively expressing the pmirGlo(+) or pmirGlo luciferase reporter mRNAs were generated. Three separate clones of respective plasmid transfected cells were infected with HAdV-37 and late protein expression visualized by Western blot at 24 hpi. As shown in Fig. 1 1 F, Hela(+) cells expressing the reporter mRNA with the MLP-TSS-sRNA sponge enhanced viral capsid protein expression considerably (lanes 5-7) compared to the HeLa(-) cells expressing the reporter mRNA with no binding site (lanes 2-4). The upregulation of late protein synthesis was also accompanied by an increase of late viral mRNA expression, as illustrated by an enhanced accumulation of tripartite leader containing mRNAs (Fig. 1 1 E). The enhanced late protein expression in the HeLa(+) cell lines (Fig. 1 1 F) was also manifested in a 3-4 fold increase in new virus progeny formation (Fig. 11 G). Taken together, these results suggest that the MLP-TSS-sRNA, indeed, serve a regulatory function during virus multiplication.

It is shown herein, using a TAP-treatment small RNA sequencing approach, that in HAdV-37-infected cells large amounts of a novel Ago2-associated sRNA (MLP-TSS-sRNA) is produced. This sRNA shows several unique features not previously shown for Ago2-associed RNAs. For example, the MLP-TSS- sRNAs are unusually long 31 or 34 nucleotides in length. The major 31 nt species has a 5’ end that coincides with the characterized MLP start site. Also, the MLP-TSS-sRNAs contain an m7G-cap at the 5’ end (Figs. 3, 4 and 6B). Based on available data (Figs. 3 and 4) it is estimated that approximately 95% of the MLP-TSS-sRNA produced during a HAdV-37 infected carry the m7G-cap. The MLP-TSS-sRNAs represent the first Ago2-associated sRNA that does not require the miRNA machinery for processing and maturation (Fig. 7). Interestingly, the MLP-TSS-sRNA also associates efficiently with Agd in HAdV-37- infected cells (Fig. 7C). The results indicate that this sRNA is produced by RNAPII stalling within the MLP first leader exon at a site coinciding with the MLP-TSS-sRNA 3’ end (Fig. 8C). The data suggest a model where the MLP-TSS-sRNAs are produced through repeated cycles of RNAPII initiation of transcription, stalling and premature termination. The data suggests that a major reason that the MLP-TSS-sRNA accumulates to high levels during an infection may be that it contains a hairpin-like structure at the 3’ end of the sRNA that protects it from degradation (Figs. 10E to 10G). Small RNAs can be degraded by multiple pathways. The sRNA sequencing data shows that the third largest population of MLP-TSS-sRNA is 24 nts long and accounts for around 8% of the total MLP-TSS-sRNA population in RISC (Fig. 1 D). This pool of small RNAs may represent processed products similar to miR-451 , which becomes trimmed by Poly(A)-specific ribonuclease PARN on the RISC complex. The crystal structure of an Ago2-miRNA complex shows that the 5’ end of the miRNA is anchored at the Ago2 MID domain, with extensive interactions between the 5’ residues including the 5’ monophosphate. Further, several reports has demonstrated that a 5’ phosphate is required for double-stranded siRNA loading onto RISC. These examples raise the question how the capped MLP-TSS-sRNAs associates with Ago2. It has, in fact, been proposed that the bulkiness of the m7G cap structure might impede on capped sRNA association with the Ago2 MID domain (Haussecker et al. 2008; Xie et al. 2013). The data suggests that the MLP-TSS-sRNA is produced as a stable single-stranded RNAPII stalled/preterminated transcript. It seems that single-stranded siRNA loading is more permissive compared to duplex loading, allowing that the siRNA 5’ end is modified without any negative effects on RISC activity. The failure to previously observe single-stranded siRNA loading in vivo appears to be due to a low stability of single-stranded siRNAs.

The VA RNAs are produced in massive amounts during virus infection and interfere with cellular miRNA production by suppressing pre-miRNA export and Dicer cleavage. Since the MLP-TSS-sRNA use an alternative non-canonical route for production the virus will maintain control over the miRNA pathway while being capable of producing other functional sRNAs. Importantly, RISC programmed with the MLP-TSS-sRNAs is functional with a capacity to suppress reporter gene expression in a sequence specific and dose dependent manner (Figs. 10B to 10D). The MLP-TSS-sRNA originates from the antisense strand of the E2B mRNAs with a perfect complementary binding site within the Adpol gene and the 3’ UTR of the pTP mRNA (Figs. 1 A and 1 1 A). Both Adpol and pTP are essential DNA replication proteins. Further, the E2B mRNAs are among the scarcest mRNAs expressed during an adenovirus infection, potentially through the MLP-TSS-sRNA targeting the pTP and Adpol mRNAs. Indeed, controlling pTP and Adpol mRNA accumulation (Fig. 1 1 B) by transfecting a synthetic MLT-TSS-sRNA had a significant inhibitory effect on adenovirus genome replication (Fig. 1 1 C) and late protein synthesis (Fig. 1 1 D). However, it was central to determine whether the MLP-TSS-sRNAs produced from the virus act similarly as the synthetic mimic. By using an MLP-TSS-sRNA sponge assay, either in a transient transfection assay (Fig. 1 1 E) or in HeLa stable cell lines (Fig. 1 1 F), it was demonstrated that sequestering the MLP-TSS-sRNA drastically increased E2B mRNA expression (Fig. 1 1 E), late viral protein synthesis (Fig. 1 1 F) and new virus progeny formation (Fig. 1 1 G).These results suggest that the MLP-TSS-sRNA has a function as a trans- acting viral regulator and not simply produced as transcriptional noise.

It is well established that the activity of the MLP increases as a result of the increase in viral DNA copy number following the start of viral DNA replication. In a hypothetical model one would predict that synthesis of Adpol and pTP stimulates viral DNA replication. As a consequence the activity of MLP increases, which results in an increase in MLP-TSS-sRNA production, which in turn reduces Adpol and pTP gene expression. Such a feedback mechanism may be important for the virus to control virus replication, particularly during the establishment and/or maintenance of persistent/latent infections where the capacity of the virus to multiple needs to be kept at a reduced level. MATERIALS AND METHODS

Cell culture and virus infection

HEK-293, 293-Flag-Ago2 and HeLa cells were grown in Dulbecco's modified Eagle’s medium (DMEM, Invitrogen) and A549 and ARPE-19 cells in DMEM/F12 medium (Invitrogen), supplemented with 10% fetal calf serum (FCS, Invitrogen), 1 % penicillin/streptomycin (PEST) at 37°C in 7% CO2 Virus titers were measured as fluorescence forming units (FFU) (Philipson 1961 ) using a pan-hexon antibody (MAB8052, Millipore, 1 :500 dilution). Infection of adherent cells was done as previously described (Kamel et al. 2014). HeLa stable cell lines constitutively expressing firefly reporter construct pmirGlo or pmirGlo(+), were made as previously described (Xu et al. 2007). For titration purposes crude cell lysate was prepared through 5 cycles of freezing (dry ice-ethanol) and thawing (37°C water bath).

Transient transfection and siRNA knockdown

Transfection of siRNA and plasmid DNA was done using the Jetprime (Polypustransfection) transfection reagent according to the manufacturer’s instructions. The sequence of the siRNAs used in this study is listed in Table 2. A scrambled ON-TARGET plus Non-targeting siRNA pool (GE Dharmacon) was used as a negative control. Table 2 - Northern blot probes, primers, siRNA and antibodies siRNA and RNA oligonucleotides

Primers

Antibodies

RNA immunoprecipitation

S15 cytoplasmic extract preparation and Flag/HA tagged Ago2 protein immunoprecipitation was as previously described (Xu and Akusjarvi 201 1 ). The endogenous Ago2 was immunoprecipitated using an anti-Ago2 antibody (Abeam, ab571 13). Cell lysates were incubated with the antibody (2 pg Ab/ml lysate) on a rotating wheel overnight at 4°C. Antibody-Ago2 complexes were purified on Dynabeads® Protein G magnetic beads (Thermo Fisher Scientific). The washing and RNA extraction was performed as described (Xu and Akusjarvi 201 1 ). For RNAPII immunoprecipitation, 293-Flag-Ago2 cells were irradiated twice for 400 mJ/cm ² at 254 nm followed by protein extraction with RIPA buffer (150 mM NaCI, 50 mM HEPES [pH 7.4], 0.5% Sodium Deoxycholate, and 0.1 % SDS, supplemented with 2U DNAsel/ml and protease inhibitors) and sonication 10 cycles (30 sec on and 30 sec off). RNAPII complexes were immunoprecipitated using the 8WG16 antibody (ab817), which recognizes the RNAPII CTD repeat, using the same method as described above with the modification that the beads were washed four times in lysis buffer, followed by proteinase K digestion and phenol extraction and ethanol precipitation. Treatment of RNA with Cap-removing enzymes was preformed using Tobacco acid Pyrophosphatase TAP (Epicenter) or RNA 5’ Pyrophosphohydrolase RppH (New England Biolabs Inc.) according to the manufacturer’s instructions.

Capped RNA immunoprecipitation

RNA prepared from the RISC immunoprecipitation was incubated in cap binding buffer (100 mM Flepes- KOH [pH 7.0], 5 mM MgCL, 5 mM KCI, 300 mM NaCI) together with 5 pg Anti-m3G-cap, m7G-cap Antibody (clone FLO, Millipore) on a rotating wheel overnight at 4°C. The antibody-capped RNA complexes were purified on Dynabeads® Protein G magnetic beads (Thermo Fisher Scientific) and washed three time in cap binding buffer, followed by RNA extraction as described below.

In vitro transcription and m7G capping of the MLP-TSS-sRNA

The template for the in vitro transcription was a double-stranded DNA oligonucleotide in which the forward strand contained a T7-promoter sequence. In v/fro-transcription was done using the TranscriptAid T7 High Yield transcription kit (ThermoFisher) according to the manufacturer’s instructions. The RNA was purified by one round of phenol extraction and ethanol precipitation. Capping of the in vitro transcribed RNA was done using the vaccine virus capping system (New England Biolabs Inc.) as described by the manufacturer. The final RNA used for transfection was purified by one round of phenol extraction and ethanol precipitation. The RNA was dissolved in sterile H2O and stored at -80°C.

Luciferase assay

Luciferase assay was performed with the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions. Both firefly luciferase and renilla luciferase activity was measured on Infinite M200 luminometer (Tecan). Results shown are normalized ratio of firefly to renilla and are presented as means from at least three biological replicates. Statistical analysis was performed on Prism6 (GraphPad Software, USA) using a two-tailed unpaired t-test. The value of p < 0.05 was considered statistically significant.

Bioinformatics analysis

The cDNA library construction and small RNA sequencing has previously been described (Kamel et al. 2014). Sequencing reads were mapped to the human reference genome (GRCH37) and HAdV-37 genome (accession number DQ900900) using STAR aligner (Dobin et al. 2013). Visualization of mapped reads shown in Fig. 1 C was done using Interactive Genome Viewer (IGV) (Robinson et al. 201 1 ). To characterize viral sRNA, reads were mapped the corresponding adenovirus genome using BLASTN (no mismatch allowed). Read start site and viral sRNA read length was visualized using customized Perl scripts.

Total protein extraction

Cells were collected by low speed centrifugation and washed once with 1 xPBS. The pellet was suspended in 180 mI of RIPA buffer (150 mM NaCI, 50 mM HEPES [pH 7.4], 0.5% Sodium Deoxycholate, and 0.1 % SDS) supplemented with 1 U/ml of Benzonase (EMD Millipore) and incubated for 1 h at 4°C. Cells were further disrupted by addition of 10% SDS and 1 M DTT to final concentrations of 1 % and 100 mM, respectively, followed by a boiling step of 5 min (Holden and Horton 2009).

Total DNA/RNA extraction and quantitative real time PCR

Genomic viral DNA was extracted using TRI Reagent (Sigma). To estimate viral DNA copy number, 50 ng DNA of each sample was amplified using VAN specific primers and copy number was calculated from a standard curve created from serial dilutions of pud 9-VAII plasmid, which contain single copy of HAdV- 37 VAN gene, using the absolute quantification method (Whelan et al. 2003). RNA was extracted using TRI Reagent (Sigma) followed by cDNA synthesis using Superscript III kit (Thermo Fisher Scientific). Relative expression oiAdpol and pTP were normalized to HPRT1 and quantified using the MCt method (Schmittgen and Livak 2008). PCR was carried on Applied Biosystems 7900 system (Life Technologies) using HOT FIREPol® EvaGreen® qPCR Supermix (Solis BioDyne). Statistical analysis was performed on Prism6 (GraphPad Software, USA) using a two-tailed unpaired t-test. The value of p < 0.05 was considered statistically significant. The primers used in the qPCR are listed in Table 2.

Northern and Western blot analysis

Northern and Western blots were performed as described in detail in (Kamel et al. 2014). A list of oligonucleotide probes, details of the antibodies, dilutions and the vendors used are shown in Table 2.

EXAMPLE 2

This example demonstrates that the MLP-TSS-sRNA can be transformed into a general RNA regulatory method. MLP-TSS-sRNA can be subdivided into two domains; a 5’ functional domain and a 3’ stability domain that can be modified and fused to generate, for instance, chimeric siRNAs/miRNAs.

RESULTS

To test the proposed domain structure, it was analyzed whether the 3’ structural domain contributed to function. It is well established for classical siRNAs and miRNAs that the 5’ end of the guide strand is crucial for target recognition. To establish whether this is the case with MLP-TSS-sRNA, the role of the 5’ and 3’ domains in targeting complementary RNA sequences was investigated. As shown in Fig. 12, the MLP-TSS- sRNA, either produced during a HAdV-37 infection (A) or transfected as a synthetic RNA (B), was capable of repressing target mRNA sequences with full complementary (WT) whereas a target mRNA with no base pair complementarity was unaffected. Further mutating the complementarity of the ten 5’ nucleotides essentially abolished this repressive function of the MLP-TSS-sRNA. These experiments suggest that the MLP-TSS-sRNA, similar to a miRNA and a siRNA, depend on the 5’ end for target recognition. The data further support the model that the 3’ end of the MLP-TSS-sRNA is required for stability whereas the 5’ end is responsible for target recognition.

In a second line of experiments it was tested whether the sequence of the functional domain could be exchanged to unrelated siRNA and miRNA sequences.

For this experiment, chimeric MLP-TSS-sRNAs were generated where the adenovirus specific functional domain was replaced with the targeting sequence from the cellular miRNAs miR26a and miR155. As is shown in Fig. 13, the generated chimeric miRNAs contained the targeting sequence of two cellular miRNAs fused to the MLP-TSS stability domain. As we previously reported for the MLP-TSS-sRNA in Example 1 , the miR26a-TSS and miR155-TSS showed the same extreme stability as the MLP-TSS-sRNA with essentially no degradation at 96 hrs post transfection (Fig. 13C).

The effect of the chimeric miR26a-TSS and miR155-TSS on the validated targets Enx1 and Rb1 (miR26a) and Bachl (miR155) were tested. As shown Fig. 13B, the chimeric mir26a-TSS and mir155-TSS were indeed effectively suppressing their validated target genes. To compare silencing efficiency of the chimeric TSS sRNAs with a commercially available miRNA mimic, the knockdown levels of the Enx1 protein induced by transfecting mir26a-TSS or a double-stranded mir26a mimic were evaluated. As shown in Figure in Fig. 14A and quantitated in Fig. 14B, the mir26a-TSS was slightly less effective compared to the commercially available mir26a mimic. Since siRNAs are 22 nucleotides long and functions through perfect complementary of the small RNA to its target mRNA, it was tested whether the length of the functional domain could be varied. Note that in mir26a- TSS and mir155-TSS a targeting sequence of 10 and 11 nucleotides, respectively was appended to the MLP-TSS stability domain (Fig. 13A). For this experiment a 22 nt single stranded anti-sense siRNA directed against the Dicer mRNA was fused to the MLP-TSS stability domain, generating Dicer-TSS (Fig. 15A). This siRNA sequence has been extensively used in the literature. As shown in Fig. 15B and quantitated in Fig. 15C, Western blot analysis showed that the Dicer-TSS sRNA repressed Dicer protein expression to the same extent as a conventional double-stranded siRNA.

Double-stranded siRNAs shows a certain degree of unspecific toxicity, possible by activating components of the innate immune response, which has several arms that reacts to double-stranded RNA. To test the toxicity of the single-stranded TSS variants, HeLa cells were transfected with mir26a-TSS and mir155-TSS and the equivalent miRNA mimics, and the cells were counted 36 hours post transfection. As shown in Fig. 14C, the TSS miRNA variants showed a considerably lower toxicity (presented by higher cell count) compared to the double-stranded miRNA mimics.

MATERIALS AND METHODS

Luciferase assay Luciferase assay was performed with the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s instructions. Both firefly luciferase and renilla luciferase activity was measured on an Infinite M200 luminometer (Tecan). Results shown are normalized ratio of firefly to renilla and are presented as means from at least three biological replicates. Statistical analysis was performed on Prism6 (GraphPad Software, USA) using a two-tailed unpaired t-test. The value of p < 0.05 was considered statistically significant.

Transient transfection and siRNA knockdown

Transfection of siRNA and plasmid DNA was done using the Jetprime (Polypustransfection) transfection reagent according to the manufacturer’s instructions. The sequence of the siRNAs used in this study is listed in Table 3.

Table 3 - Northern blot probes, primers, siRNA and antibodies siRNA and RNA oligonucleotides

Antibodies

Northern and Western blot analysis

Northern and Western blot analysis were performed as described in detail in (Kamel et al. 2014). A list of oligonucleotide probes, details of the antibodies, dilutions and the vendors used are shown in Table 3.

The embodiments described above are to be understood as a few illustrative examples of the present invention. It will be understood by those skilled in the art that various modifications, combinations and changes may be made to the embodiments without departing from the scope of the present invention. In particular, different part solutions in the different embodiments can be combined in other configurations, where technically possible. The scope of the present invention is, however, defined by the appended claims.

REFERENCES

Cheloufi S, Dos Santos CO, Chong MM, Hannon GJ. 2010. A dicer-independent miRNA biogenesis pathway that requires Ago catalysis. Nature 465: 584-589.

Cifuentes D, Xue H, Taylor DW, Patnode H, Mishima Y, Cheloufi S, Ma E, Mane S, Hannon GJ, Lawson ND et al. 2010. A novel miRNA processing pathway independent of Dicer requires Argonaute2 catalytic activity. Science 328: 1694-1698. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. 2013. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29: 15-21 . Haussecker D, Cao D, Huang Y, Parameswaran P, Fire AZ, Kay MA. 2008. Capped small RNAs and MOV10 in human hepatitis delta virus replication. Nature structural & molecular biology 15: 714-721 .

Holden P, Horton WA. 2009. Crude subcellular fractionation of cultured mammalian cell lines. BMC Res Notes 2: 243.

Kamel W, Segerman B, Punga T, Akusjarvi G. 2014. Small RNA sequence analysis of adenovirus VA RNA-derived miRNAs reveals an unexpected serotype-specific difference in structure and abundance. PLoS One 9: e105746. Lu S, Cullen BR. 2004. Adenovirus VA1 noncoding RNA can inhibit small interfering RNA and MicroRNA biogenesis. J Virol 78: 12868-12876.

Pall GS, Hamilton AJ. 2008. Improved northern blot method for enhanced detection of small RNA. Nat Protoc 3: 1077-1084.

Philipson L. 1961 . Adenovirus assay by the fluorescent cell-counting procedure. Virology 15: 263-268.

Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. 201 1. Integrative genomics viewer. Nat Biotechnol 29: 24-26.

Schmittgen TD, Livak KJ. 2008. Analyzing real-time PCR data by the comparative C(T) method. Nat Protoc 3: 1101 -1 108.

Whelan JA, Russell NB, Whelan MA. 2003. A method for the absolute quantification of cDNA using real- time PCR. J Immunol Methods 278: 261 -269.

Xie M, Li M, Vilborg A, Lee N, Shu MD, Yartseva V, Sestan N, Steitz JA. 2013. Mammalian 5'-capped microRNA precursors that generate a single microRNA. Cell 155: 1568-1580. Xu N, Segerman B, Zhou X, Akusjarvi G. 2007. Adenovirus virus-associated RNAII-derived small RNAs are efficiently incorporated into the ma-induced silencing complex and associate with polyribosomes. J Virol 81 : 10540-10549.

Xu N, Akusjarvi G. 201 1 . Characterization of RISC-associated adenoviral small RNAs. Methods Mol Biol 721 : 183-198.

Yi R, Qin Y, Macara IG, Cullen BR. 2003. Exportin-5 mediates the nuclear export of premicroRNAs and short hairpin RNAs. Genes Dev 17: 301 1 -3016.