Report

Controlling Gene Expression in Living Cells Through Small Molecule-RNA Interactions

See allHide authors and affiliations

Science  09 Oct 1998:
Vol. 282, Issue 5387, pp. 296-298
DOI: 10.1126/science.282.5387.296

Abstract

Short RNA aptamers that specifically bind to a wide variety of ligands in vitro can be isolated from randomized pools of RNA. Here it is shown that small molecule aptamers also bound their ligand in vivo, enabling development of a method for controlling gene expression in living cells. Insertion of a small molecule aptamer into the 5′ untranslated region of a messenger RNA allowed its translation to be repressible by ligand addition in vitro as well as in mammalian cells. The ability of small molecules to control expression of specific genes could facilitate studies in many areas of biology and medicine.

In vitro genetic selections (1) have been used to isolate nucleic acid sequences (aptamers) that bind small molecules with high affinity and specificity (2). The ability to control gene expression by using cell-permeable small molecules offers several advantages, and small molecule manipulation of gene expression at the levels of transcription (3) and signal transduction (4) has been reported.

We selected RNA aptamers that specifically bound to the related aminoglycoside antibiotics kanamycin A and tobramycin (Fig. 1A) (5). We analyzed the ability of these aptamers to function in vivo by expressing them inEscherichia coli and testing for a drug-resistant phenotype (6). In the absence of drug, bacterial strains expressing no aptamer (bl-RSETA), the kanamycin aptamer (bl-kan1), or the tobramycin aptamer (bl-tob1) grew similarly (Fig. 1B). In the presence of 10 μM kanamycin A, bl-kan1 grew to saturation, whereas growth of bl-RSETA and bl-tob1 was negligible (Fig. 1C). In the presence of 10 μM tobramycin, bl-tob1 grew to saturation, and bl-kan1 grew to a subsaturating concentration (Fig. 1C) (7). Increasing the number of aptamers in the expression vector from one to three enhanced growth in the presence of antibiotic (Fig. 1D). Thus, a specific drug-resistant phenotype was conferred by expression of an aminoglycoside aptamer (8), which demonstrates a small molecule–aptamer interaction in vivo.

Figure 1

Selective interaction between aminoglycosides and aminoglycoside aptamers in vivo. (A) Structures of aminoglycoside antibiotics and their aptamers. Consensus aptamers were identified after 10 to 12 rounds of selection. (B to D) Growth curves. Overnight cultures ofE. coli BL-21 transformed with plasmids expressing RSETA, tob1, tob3, kan1, or kan3 were diluted 1:100 into medium containing the indicated concentration of aminoglycoside antibiotic. Optical density (660 nm) was measured at fixed intervals over 8 hours of growth at 37°C. (B) Growth in the absence of drug. (C) Growth in the presence of 10 μM kanamycin A or tobramycin. (D) Growth in the presence of 20 μM kanamycin A or tobramycin.

We next asked whether small molecule aptamers could be used to regulate gene expression. Eukaryotic translation initiation typically involves 5′-to-3′ scanning from the 5′-m7G cap to the start codon (9). Binding of a protein between the cap and start codon can repress translation, presumably by blocking either scanning or the ribosome-mRNA interaction (10). We thus asked whether the presence of a small molecule–aptamer complex within the 5′ untranslated region (UTR) would repress translation.

We constructed an mRNA that contained three copies of the tob aptamer inserted in the 5′ UTR of RSETA (tob3-RSETA). In vitro translation (11) of the control RSETA mRNA was unaffected by all concentrations of tobramycin or kanamycin tested, whereas addition of tobramycin inhibited in vitro translation of the tob3-RSETA mRNA in a dose-dependent fashion (Fig. 2) (12). In vitro translation of the tob3-RSETA mRNA was not inhibited by comparable concentrations of kanamycin A, which is not recognized by the tob aptamer.

Figure 2

An aminoglycoside aptamer translation switch functions in vitro. RNA transcripts containing zero (RSETA) or three copies of the tob aptamer cloned into the 5′ UTR of RSETA (tob3-RSETA) were translated in a wheat germ extract in the presence of [35S]methionine and 0, 30, or 60 μM tobramycin or kanamycin A. Protein products were analyzed by SDS–polyacrylamide gel electrophoresis (PAGE) (arrows) and quantitated by densitometry. For each transcript, translation in the absence of drug was set at 100%.

We next attempted to reconfigure the system for regulating gene expression in vivo. Because aminoglycosides are relatively impermeable to the plasma membrane, can be cytotoxic, and have a general inhibitory effect on translation at high concentrations (13), we used a cell-permeable small molecule as the translation regulator.

Hoechst dye 33258 (H33258) and the closely related drug H33342 (Fig. 3A) are relatively nontoxic and cell-permeable (14). We isolated RNA aptamers that bound specifically to H33258 (15), two of which—H10 and H19—are shown (Fig. 3B). Both H10 and H19 bound to an H33258 affinity column and required a relatively high concentration (25 mM) of free H33258 for elution (Fig. 3C) (16).

Figure 3

Controlling translation in vitro and in vivo with cell-permeable small molecules. (A) Structure of Hoechst dyes H33258 and H33342. (B) Sequences and predicted secondary structures of two H33258 aptamers, H10 and H19, based on the computer modeling program Mulfold. A Hoechst dye aptamer consensus sequence (UUAN4–5UCU) was identified after 10 rounds of selection. The fixed-primer binding regions are shown in normal print, selected bases are boldface, and the consensus sequence is indicated with outline print. (C) Interaction of H10 and H19 aptamers with H33258. [32P]UTP-labeled aptamer (200,000 cpm) was loaded onto a 0.25-ml H33258 Sepharose column. Each column was then washed sequentially with 6 ml of binding buffer, 1 ml of binding buffer containing 5 mM H33258, and 1 ml of binding buffer containing 25 mM H33258. Percentages of total bound RNA eluted in each step are indicated. (D) In vitro translation. RNA transcripts containing zero (RSETA) or two copies of an H33258 aptamer cloned into the Bsa I site of pRSETA (H2-RSETA) were translated in a wheat germ extract in the presence of [35S]methionine and 0, 40, or 80 μM H33258. Protein products were analyzed by SDS-PAGE and quantitated by densitometry. For each transcript, translation in the absence of drug was set at 100%. (E) In vivo expression. H33258 aptamers H10 and H19 were cloned in tandem into the 5′ UTR (Sfi I–Avr II sites) of a β-galactosidase reporter gene (SVβgal; Promega) to generate SVH2βgal. CHO cells were cotransfected with 1 μg of SVβgal or SVH2βgal and 1 μg of a luciferase expression vector (pGL3). Cells were grown in the presence of 0, 5, or 10 mM H33342. Twenty-four hours after transfection, cell extracts were prepared and β-galactosidase and luciferase activities were determined.

To demonstrate that the H33258 aptamer could be used to regulate translation, we inserted one copy of H10 and H19 in tandem into the 5′ UTR of RSETA. Addition of H33258 inhibited in vitro translation of H2-RSETA but not the control RSETA in a dose-dependent fashion (Fig. 3D).

To test whether this small molecule–aptamer interaction could be used to control gene expression in vivo, we inserted one copy of H10 and H19 into the 5′ UTR of a mammalian β-galactosidase expression plasmid, SVβGal (Promega), generating the construct SVH2βgal. Chinese hamster ovary (CHO) cells were cotransfected with SVH2βGal or, as a control, the parental vector SVβGal and a luciferase reporter gene to provide an internal control (17). After transfection, cells were grown for 24 hours in the presence of 0, 5, or 10 μM H33342 and analyzed for β-galactosidase and luciferase activities (18).

In the absence of drug, two H33258 aptamers in the 5′ UTR had no effect on gene expression (compare SVβgal and SVH2βgal) (Fig. 3E), consistent with the in vitro translation data of Fig. 3D. Expression of the luciferase reporter was also not inhibited by H33342 (19). However, H33342 reduced β-galactosidase activity from SVH2βGal by greater than 90% in a dose-dependent fashion (Fig. 3E).

We have described how a small molecule and its RNA aptamer can be used to design a translation switch for controlling gene expression in living cells. The results also establish the possibility of using small molecules to regulate expression of endogenous genes.

  • * Present address: MBI Fermentas Inc., 7 Innovation Drive, Flamborough, ON, Canada L9H 7H9.

  • To whom correspondence should be addressed. E-mail: michael.green{at}ummed.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article