Dear Editor,

RNAi has become a mainstream molecular tool for assessing the functions of genes in mammalian cells 1. Large-scale RNA interference-based analyses are often complicated by false positive and negative hits due to off-target effects 2 and interferon response 3, which can be attributed at least in part to the use of high concentrations of siRNA. Lowering the amounts of siRNAs and shRNAs can effectively and expediently mitigate the off-target effect and interferon response 4. However, in RNAi experiments, lowering the concentration of siRNA is often accompanied by a lower knockdown efficiency. One of the key factors affecting RNAi efficiency is the stability of double-stranded siRNA. We reasoned that measures that could stabilize double-stranded RNA may lead to increased RNAi efficiency. Given that RNAi requires a number of cellular proteins, it should be possible, at least in theory, to regulate the efficiency of RNAi using small organic molecules. To stabilize ds-RNA used in RNAi, we therefore envisioned the following methods: (1) to design more stable siRNA (screening for certain sequences, modification of RNA); (2) to increase the activity of ds-RNA protective proteins (such as ds-RNA-binding proteins, or simply binding domains); (3) to inhibit the activity of ds-RNA-dissolving proteins (such as RNA helicases); (4) to stabilize ds-RNA and/or its protein complex with small organic compounds. Our longstanding interest in drug-nucleic acids interaction 5 led us to search for potential small molecular regulators of RNAi. We hypothesized that inhibitors of RNA helicases may increase the stability of double-stranded siRNA, so as to enhance RNAi efficiency. Since a large family of fluoroquinolone antibiotics target bacterial DNA gyrase complexed with the targeted DNA possibly in A-form (similar to RNA) 6 and since they also exhibit antiviral activity through interference with Tat-TAR interaction 7, we decided to screen a library of commercially available fluoroquinolone antibiotics, with the hope that some of the analogs may cross-inhibit relevant human RNA helicases. Herein, we report that enoxacin, one of the fluoroquinolone antibiotics known to inhibit bacterial gyrase and topoisomerase IV with minimal effects on their mammalian counterparts, can increase RNAi efficiency. We have found that enoxacin can reduce the concentrations of siRNA by 2∼5-fold for the same RNAi knockdown efficiency.

A dual-luciferase reporter assay system was used to screen small organic compounds that were capable of enhancing RNAi efficiency. The siRNA used in our screen is siFL867-885, which can effectively suppress the firefly luciferase reporter activity at 10 nM siRNA duplexes 8. This system has been reported to be a robust siRNA screening system 8. As a starting point, we diluted siFL867-885 to a concentration of 8.4 × 10−10 M, and at this concentration the RNAi knockdown was partial. A small library of fifteen widely used fluoroquinolone antibiotics (Figure 1B) was screened using the dual-luciferase reporter system. By statistical analysis of experimental data (Figure 1A), seven compounds were found to increase RNAi efficiency (Figure 1B, the compounds shown with asterisk (*)). Among these 7 fluoroquinolones exhibiting statistically significant RNAi-enhancing activities, enoxacin and norfloxacin were more active than others at the same concentration. We then chose enoxacin and norfloxacin for further characterization. It was found that the RNAi-enhancing activity of enoxacin was dose-dependent (see Supplementary information, Figure S1). The EC50 of enoxacin as an RNAi enhancer is about 30 μM. This concentration is far below the IC50 of enoxacin for inhibition of topoisomerase II in human macrophages (IC50 = 1 485 μM) (see Supplementary information, Figure S2) 9. At 120 μM, enoxacin started to cause non-specific reduction of the firefly luciferase activity (see Supplementary information, Figure S3). The RNAi efficiency of different concentrations of siRNA (siFL867-885) in the presence or absence of enoxacin was also examined. At 12 h post-transfection, HEK-293 cells were treated with or without 50 μM enoxacin. An optimum RNAi-enhancing effect by enoxacin was observed when the siRNA concentration was at 8.4 × 10−10 M. Generally, enoxacin can reduce the amount of siRNA by 2∼5-fold to achieve the same RNAi knockdown efficiency (Figure 1C).

Figure 1
figure 1

(A) Screening of 15 fluoroquinolones for RNAi enhancing activity in cultured HEK-293 cells. Statistically significant difference of RNAi efficiency was observed in the presence of 7 compounds as shown with asterisk. The Y-axis denotes the chemiluminescence data of firefly luciferase activities. The final concentrations of siRNA duplexes (siFL867-885) and compounds are 8.4 × 10−10 M and 50 μM, respectively. The plotted data represent the meanSE. The asterisk (*) denotes statistically significant difference (P < 0.05). (B) Chemical structures of 15 fluoroquinolones. The compounds showing statistically significant RNAi enhancing effect are labeled with asterisk (P < 0.05). (C) RNAi efficiency of different concentrations of siRNA (siFL867-885) in the presence or absence of enoxacin. At 12 h post-transfection, HEK-293 cells were treated without or with enoxacin at the final concentration of 50 μM. The Y-axis denotes the RNAi knockdown percentage of firefly luciferase activity after correction by renilla luciferase activity (using normalized expression ratios between the firefly luciferase avtivity and the renilla luciferase activity in the dual-luciferase reporter assay system) 6. The RNAi-enhancing activity by enoxacin reaches the maximum at the siRNA concentration of 8.4 × 10−10 M. All experiments with siRNAs were performed in triplicates and repeated at least twice. Each value represents the mean±SE.

Full size image

In summary, we have demonstrated that certain fluoroquinolone antibiotics such as enoxacin, in addition to their powerful clinic use for the treatment of infections in humans and animals 10, can be used to increase RNAi efficiency. Enoxacin can significantly reduce the amount of siRNA (by 2∼5-fold) to achieve the same RNAi efficacy. The precise mechanism of this RNAi enhancement remains unclear at present. While our manuscript was in preparation, a similar finding with more detailed analysis was reported by Jin and colleagues 11, who proposed that enoxacin acts by potentially increasing RISC loading efficiency through a mechanism depending on the protein factor TRBP. Nevertheless, one cannot rule out the possibility that the effect of enoxacin on RNAi is due to the cross-interaction with human RNA helicases and the stabilization of RNAi molecules, especially in view of the finding that human RNA helicase A (RHA) is an active RISC component and functions in RISC as an siRNA loading factor 12. Further mechanistic study (such as QSAR, photolabelling and affinity purification) is being actively pursued in our lab to elucidate the molecular basis of the observation reported here. Other ds-RNA stabilizing agents are still waiting to be explored for the regulation of RNAi. We believe that the finding of small molecular RNAi enhancers could potentially be used as a tool in genome-scale loss-of-function screening with RNAi, especially in cases that are highly dependent on the use of relatively low concentrations of siRNA and shRNA. It may also be a useful molecular probe to shed new light on the RNAi process. (Experimental materials and methods are depicted in the Supplementary information, Data S1)

( Supplementary Information is linked to the online version of the paper on the Cell Research website.)