Improved silencing properties using small internally segmented interfering RNAs

Abstract

RNA interference is mediated by small interfering RNAs (siRNAs) that upon incorporation into the RNA-induced silencing complex (RISC) can target complementary mRNA for degradation. Standard siRNA design usually feature a 19-27 base pair contiguous double-stranded region that is believed to be important for RISC incorporation. Here, we describe a novel siRNA design composed of an intact antisense strand complemented with two shorter 10-12 nt sense strands. This three-stranded construct, termed small internally segmented interfering RNA (sisiRNA), is highly functional demonstrating that an intact sense strand is not a prerequisite for RNA interference. Moreover, when using the sisiRNA design only the antisense strand is functional in activated RISC thereby completely eliminating unintended mRNA targeting by the sense strand. Interestingly, the sisiRNA design supports the function of chemically modified antisense strands, which are non-functional within the context of standard siRNA designs. This suggests that the sisiRNA design has a clear potential of improving the pharmacokinetic properties of siRNA in vivo.

Figure 1.

Testing the knock down of EGFP by sisiRNA and related constructs. The EGFP expression was assessed both at the RNA and protein level in H1299 cells. (A) Fluorescence microscopy analysis of EGFP expression in H1299 cells. Cells were treated with 50 nM of the indicated combinations of RNA/LNA oligoes and analyzed 48 hours after for EGFP expression. (B) Northern blot showing EGFP mRNA expression 48 and 120 hours after initial transfection with 50 nM of the indicated combinations of oligonucleotides. (C) Western blot showing the expression of EGFP protein in cells at 48, 120 and 180 hours after transfection with 50 nM of the indicated RNA constructs. The filter was reprobed with an antibody specific to hnRNPC1 protein as a loading control. (D) Flow cytometry analysis showing the mean green fluorescence of 50.000 cells 48 hours posttransfection (based on three experiments).

Figure 2.

Knockdown of GAPDH mRNA by sisiRNA. (A) Northen blot analysis of GAPDH mRNA expression in H1299 cells 48 h after transfection of either siGAPDH, siGAPDH (LNA), sisiGAPDH 1, siGAPDH 2 or siEGFP as indicated. Upper panel: GAPDH mRNA as evaluated by northern blotting using a GAPDH probe. Lower panel: ethidium bromide staining of 18S rRNA (loading control). Experiments were performed in triplicates. (B) Quantification of GAPDH mRNA levels in (A). All values are normalized to siEGFP transfected cells.

Figure 3.

Serum stability of the sisiRNA design. (A) LNA-modified sisiRNA and LNA-modified siRNA have increased serum stability compared to unmodified siRNA. The siRNA variants were incubated in 80% FCS and aliquots taken at indicated time points. Serum-stability was evaluated by PAGE followed by SYBR Gold® staining. Whereas siRNA is degraded within 1–1½ h of serum incubation, significant amount of LNA-modified sisiRNA and LNA-modified siRNA are still present after 13 h of incubation. A size marker is indicated to the left. (B) LNA-base pairing is essential for the integrity of sisiRNA molecules upon incubation in 10% FCS. The indicated sisiRNA molecules carrying different or no LNA modifications were incubated in the presence (+) or absence (−) of 10% FCS for 24 h and duplex stability was evaluated by PAGE followed by SYBR Gold® staining. sisiRNA constructs with LNA in both strands exhibited full stability whereas sisiRNA containing only RNA were completely degraded upon serum incubation. The position of the LNA modifications are indicated schematically (vertical lines). (C) LNA-modified sisiRNA and LNA-modified siRNA do not activate the interferon system as evaluated by induction of ISG56 in T98G cells. The glioblastoma cell line T98G was transfected with 80 nM of the siRNA variants or 0.8 μg poly(I:C) as indicated and ISG56 mRNA levels evaluated by qPCR analysis 48 h posttransfection. Only transfection of poly(I:C) (pos. control) lead to high levels of ISG56 induction, whereas ISG56 levels for all siRNA-variants were indistinguishable from untreated cells or cells treated with transfection reagent alone (Mirus Trans-IT TKO®).

Figure 4.

Optimization the sisiRNA design. The knock down efficiencies between different sisiRNA and siRNA designs were compared by targeting EGFP mRNA. (A) Analyzing the effect of different gap positions in the sense strand. The numbers correspond to the size of the 5′ and 3′ fragment of the sense strand, respectively. (B) Analyzing the effect of modifications at the 3′ terminal nucleotide on the sense strand. The mean fluorescence of approximately 50 000 cells was measured by flow cytometry. The siRNA mismatch represents a siRNA that contains four mismatches to the EGFP target (Table 1).

Figure 5.

The sisiRNA design increases the specificity of gene silencing. The knockdown activity of the two strands was assessed by measuring luciferase expression from reporter constructs containing either the target sequence for the sense or antisense strand of the EGFP specific duplex (light and dark gray bars, respectively). The reporter constructs are drawn above (not to scale) and the siRNA constructs are indicated schematically below. The values are averaged over three completely independent experiments. The luciferase values of each experiment are normalized to make the sums of the luciferase activities in each of the experiments equal. For each reporter construct, the firefly luciferase (Fluc)/renilla luciferase (Rluc) ratio was normalized to mismatch controls.

Figure 6.

The sisiRNA design supports the silencing effects of chemically modified antisense strands. (A) Molecular structure of N2′-adamantylmethylcarbonyl 2′-amino-LNA-Thymine and N2′-pyren-1-ylmethyl 2′-amino-LNA-thymine. (B) Analyzing the effect of the sisiRNA design on the silencing efficiency of heavily modified antisense strands. The mean fluorescence of approximately 50 000 cells was measured by flow cytometry. The siRNA mismatch represents a siRNA that contains four mismatches to the EGFP target (Table 1). (C) The knockdown activity of the two siRNA duplex strands was assessed by measuring luciferase expression from reporter constructs containing either the target sequence for the sense or antisense strand of the EGFP specific duplex (light and dark gray bars, respectively). The reporter constructs are drawn above (not to scale). The experiment was performed in triplicate and for each reporter construct the firefly luciferase (Fluc)/renilla luciferase (Rluc) ratio was normalized to mismatch controls.