Artificial control of gene expression in mammalian cells by modulating RNA interference through aptamer-small molecule interaction

Abstract

Recent studies have uncovered extensive presence and functions of small noncoding RNAs in gene regulation in eukaryotes. In particular, RNA interference (RNAi) has been the subject of significant investigations for its unique role in post-transcriptional gene regulation and utility as a tool for artificial gene knockdown. Here, we describe a novel strategy for post-transcriptional gene regulation in mammalian cells in which RNAi is specifically modulated through RNA aptamer-small molecule interaction. Incorporation of an RNA aptamer for theophylline in the loop region of a short hairpin RNA (shRNA) designed to silence fluorescent reporter genes led to dose-dependent inhibition of RNAi by theophylline. shRNA cleavage experiments using recombinant Dicer demonstrated that theophylline inhibited cleavage of an aptamer-fused shRNA by Dicer in vitro. Inhibition of siRNA production by theophylline was also observed in vivo. The results presented here provide the first evidence of specific RNA-small molecule interaction affecting RNAi, and a novel strategy to regulate mammalian gene expression by small molecules without engineered proteins.

FIGURE 1.

Design of aptamer-fused shRNAs and effect of theophylline on RNAi induced by the shRNAs. (A) Secondary structures of putative transcripts from pE19, pE19T, pE20T, and pE21T. Boxed nucleotides indicate the anti-sense strand targeting EGFP, gray nucleotides indicate the loop sequence from the pSilencer vector (Ambion) (Brummelkamp et al. 2002), circled nucleotides indicate bases that interact with theophylline (Zimmermann et al. 1997), and arrowheads indicate putative Dicer cleavage sites. RNA polymerase III transcripts are known to have heterogenous 3′ terminus with one to four U residues (Bogenhagen and Brown 1981; Miyagishi and Taira 2002). For simplicity, structures with two U residues are depicted. It should be noted that the pE19 transcript (E19 shRNA) was predicted to have a 21-bp stem region consisting of a 19-bp stem encoding an EGFP sequence and an additional 2-bp stem from the loop sequence recommended by the manufacturer. (B) Dose-dependent inhibition of RNAi with pE19T by theophylline. HEK293 cells were cotransfected with pEGFP-N1 and pDsRed1-N1, and pE19 (without aptamer) or pE19T (with aptamer) in a 96-well microplate. Fluorescence measurements were performed 48 h after transfection, and the ratios of EGFP and DsRed fluorescence intensity were calculated. The ratios were then normalized to those from cells transfected with an shRNA expression vector with a scrambled sequence that has no significant homology with the human genome. The data are averages of triplicate transfections and error bars represent standard deviations. (C) Effect of shRNA stem length on theophylline-dependent RNAi inhibition. HEK293 cells were cotransfected with pEGFP-N1, pDsRed1-N1, and either pE19, pE19T, pE20T, or pE21T in a 96-well microplate.

FIGURE 2.

Design of an aptamer-fused shRNA targeting the DsRed1 gene and inhibition of RNAi by theophylline. (A) Secondary structures of putative transcripts from pD19 and pD19T. Boxed nucleotides indicate the anti-sense strand targeting DsRed, gray nucleotides indicate the loop sequence from the pSilencer vector, circled nucleotides indicate bases that interact with theophylline, and arrowheads indicate putative Dicer cleavage sites. Three U residues resulting in 2-nt 3′ overhang are depicted here for simplicity. Note that the putative transcript from pD19 has a 21-bp stem as well as that from pE19 (see the legend of Fig. 1A). (B) Dose-dependent inhibition of RNAi against the DsRed1 gene by theophylline. RNAi was induced for the DsRed1 gene with pD19 (without aptamer) or pD19T (with aptamer) in the presence or the absence of theophylline. The data shown are averages of normalized DsRed/EGFP values obtained from triplicate transfections with error bars representing standard deviations. Note that 10 mM theophylline pD19 data do not have an error bar because two of the transfection data did not reach the transfection efficiency standard as described in Materials and Methods (see the “Fluorescence measurements” section).

FIGURE 3.

Inhibition of Dicer-mediated shRNA cleavage by theophylline. In vitro transcribed E19, E19T, and E20T shRNAs were incubated with recombinant Dicer in the presence or the absence of theophylline. Reaction products were separated on a 15% denaturing polyacrylamide gel, and detected by Northern blotting using a 5′-biotinylated DNA oligonucleotide probe encoding an EGFP sense sequence. The asterisk indicates the position of partially cleaved shRNAs (putative). The sizes of shRNA precursors (49 nt for E19, 65 nt for E19T, 67 nt for E20T) and their diced products (siRNA, 21 nt) were verified using dsRNA Ladder (New England Biolabs) and SYBR Gold Nucleic Acid Gel Stain (Molecular Probes) (data not shown).

FIGURE 4.

Inhibition of shRNA cleavage by theophylline in vivo. HEK293 cells were cotransfected with pEGFP-N1 and pDsRed1-N1, and either pE19, pE19T, pE20T, or pE21T in 6-well plates. Small RNA species were isolated from the transfected cells after measuring EGFP/DsRed fluorescence. (A) Normalized fluorescence data of the transfected cells used to isolate small RNA species (cf. Fig. 1C). (B) Northern blotting of small RNAs from transfected cells. Expression of U6 snRNA was used as a loading control. In vitro-diced products of E19 and E19T in the presence of theophylline (cf. Fig. 3) were used as size markers. The asterisk indicates the position of partially cleaved shRNAs (putative), and the arrow indicates the position of unidentified cleavage products specifically observed in vivo (see text for details).