A screen of chemical modifications identifies position-specific modification by UNA to most potently reduce siRNA off-target effects

Abstract

Small interfering RNAs (siRNAs) are now established as the preferred tool to inhibit gene function in mammalian cells yet trigger unintended gene silencing due to their inherent miRNA-like behavior. Such off-target effects are primarily mediated by the sequence-specific interaction between the siRNA seed regions (position 2-8 of either siRNA strand counting from the 5'-end) and complementary sequences in the 3'UTR of (off-) targets. It was previously shown that chemical modification of siRNAs can reduce off-targeting but only very few modifications have been tested leaving more to be identified. Here we developed a luciferase reporter-based assay suitable to monitor siRNA off-targeting in a high throughput manner using stable cell lines. We investigated the impact of chemically modifying single nucleotide positions within the siRNA seed on siRNA function and off-targeting using 10 different types of chemical modifications, three different target sequences and three siRNA concentrations. We found several differently modified siRNAs to exercise reduced off-targeting yet incorporation of the strongly destabilizing unlocked nucleic acid (UNA) modification into position 7 of the siRNA most potently reduced off-targeting for all tested sequences. Notably, such position-specific destabilization of siRNA-target interactions did not significantly reduce siRNA potency and is therefore well suited for future siRNA designs especially for applications in vivo where siRNA concentrations, expectedly, will be low.

Figure 1.

Development of siRNA and miRNA sensor cell lines to monitor siRNA off-target effects. (A) Schematic view of the luciferase based sensors used to monitor siRNA and off-target effects in stable cell lines. The siRNA sensor habors a single target site in the Renilla 3′ UTR perfectly complementary to the siRNA AS thereby leading to Renilla luciferease mRNA cleavage by Ago2. The miRNA seed sensor instead contain four copies of a target sequence matching only position 1–8 of the siRNA (seed match) whereas the miRNA full sensor additionally matches the bases at position 13–19. Both miRNA sensors monitor Ago non-cleavage mediated KD of Renilla luciferase levels indicative of siRNA off-targeting. Firefly luciferase is used as endogenous non-regulated control to normalize for differences in cell numbers. Drawings are not to scale, siRNAs are shown in gray. P, Promoters; renilla luc., Renilla luciferase; firefly luc, Firefly luciferase; pA, polyA site. (B) Validation of siRNA sensor and miRNA sensor behaviors. Unmodified siRNA, siEGFP1, was transfected into the siRNA sensor1 and miRNA full sensor1 and the relative Renilla luciferase protein and mRNA levels were determined 48 h after transfection by dual luciferase assays and qPCR, respectively. Relative Renilla luciferase levels were normalized to cells transfected with the unrelated siRNA, siBCR-ABL.

Figure 2.

Chemical modification of AS seed regions can reduce off-targeting. (A) Schematic overview of the chemical modifications investigated in this study. (B) Initial screen of siRNA potency (using the siRNA sensor) and off-targeting (miRNA seed sensor) of all chemically modified versions of the unmodified AS, W053 (see Table 1 for overview). Results from all ASs are shown in gray squares whereas selected oligos are shown in color-coding as indicated. ASs exhibiting reduced off-targeting relative to their siRNA potency are framed by the green triangle, whereas the majority of ASs exhibit a linear correlation between siRNA and off-target potency (marked by red square).

Figure 3.

UNA-modification of the AS seed most dramatically reduces siRNA off-targeting. Evaluation of siRNA potency and off-targeting potential of selected ASs at 0.1, 1 and 10 nM concentrations using the siRNA sensor1 (light gray lines) and miRNA seed/full sensor1′s (dark gray lines), respectively.

Figure 4.

UNA-modification of AS position 7 dramatically reduces siRNA off-targeting while preserving siRNA potency. Evaluation of siRNA potency and off-targeting potential of selected ASs at 0.1, 1 and 10 nM concentrations using the siRNA sensor2 (light gray lines) and miRNA seed/full sensor2′s (dark gray lines), respectively.

Figure 5.

UNA-modification destabilizes (off-) target interactions but still supports the siRNA function of ASs with weak seed region. (A) UNA-modification of the AS seed region strongly destabilizes initial (off-) target interactions. UNA-modified ASs, OMe modified (Dharm1) and unmodified AS (W053) were annealed to a 13-mer RNA with sequence complementarity to position 1–9 of the ASs and analyzed by native gel electrophoresis. Sizes of annealed duplexes and single stranded ASs are indicated to the right. (B) Evaluation of siRNA potency and off-targeting potential of selected ASs at 0.1–10 nM concentrations using the siRNA sensor3 (light gray lines) and miRNA seed/full sensor3′s (dark gray lines), respectively. The seed sequence of the siEGFP3 contains forms only two GC pairs with its target. ASs modified with UNA at position 7 further exhibits almost wild-type siRNA potency.

Figure 6.

UNA-modification of siRNA AS and SS reduces off-targeting of well-characterized endogenous off-targets while preserving on-target activity. (A) Evaluation of on-target (GRK4 mRNA) and off-target activity (hif-1α mRNA) for a siRNA directed against GRK4. (B) Evaluation of on-target (ICAM1 mRNA) and off-target activity (TNFR1 mRNA) for a siRNA directed against ICAM1. H1299 were transfected with the siRNAs at the indicated concentrations and mRNA levels were evaluated after 24 h by qPCR. The presented values are normalized to values from cells transfected with the unrelated siEGFP1.