Fluorescence-based investigations of RNA-small molecule interactions

Abstract

Interrogating non-coding RNA structures and functions with small molecules is an area of rapidly increasing interest among biochemists and chemical biologists. However, many biochemical approaches to monitoring RNA structures are time-consuming and low-throughput, and thereby are only of limited utility for RNA-small molecule studies. Fluorescence-based techniques are powerful tools for rapid investigation of RNA conformations, dynamics, and interactions with small molecules. Many fluorescence methods are amenable to high-throughput analysis, enabling library screening for small molecule binders. In this review, we summarize numerous fluorescence-based approaches for identifying and characterizing RNA-small molecule interactions. We describe in detail a high-information content dual-reporter FRET assay we developed to characterize small molecule-induced conformational and stability changes. Our assay is uniquely suited as a platform for both small molecule discovery and thorough characterization of RNA-small molecule binding mechanisms. Given the growing recognition of non-coding RNAs as attractive targets for therapeutic intervention, we anticipate our FRET assay and other fluorescence-based techniques will be indispensable for the development of potent and specific small molecule inhibitors targeting RNA.

Fig. 1

Fluorescent assays evaluating RNA-small molecule interactions. (A) Schematic of a small molecule microarray screen targeting RNA. A small molecule library is printed onto a microarray slide through a covalent linker. Fluorescently-labeled RNA is flowed over the microarray, unbound RNA is washed off, and bound RNA remains on the slide. Small molecules bound by the RNA are identified by recording fluorescence across the microarray. (B) Schematic of a fluorescence indicator displacement assay. A nonspecific fluorescent nucleic acid-binding indicator is added to RNA in solution. When bound, the indicator is highly fluorescent. If a small molecule binds to the RNA, it displaces one or more indicator molecules, resulting in a decrease in fluorescence. (C) Schematic of a fluorescence polarization assay to determine binding of a protein to RNA. RNA is fluorescently labeled, and fluorophores are excited with polarized light. Unbound RNA rotates quickly in solution, randomizing the orientation of the fluorophores, resulting in the emission of depolarized light. Upon binding to a protein (purple), the larger RNA-protein complex rotates more slowly on the timescale of the fluorophore lifetime. Thus, the fluorophore remains in a more well-defined orientation and emits polarized light. (D) Schematic of 2-AP fluorescence to study conformational changes. 2-AP exhibits high fluorescence (blue, glow) when exposed to solution. Conformational change induced by small molecule binding alters the 2-AP local environment leading to stacking interactions which quench fluorescence (blue, no glow). (E) Schematic of DSF. RNA-binding dyes bind preferentially to single-stranded regions of RNA. As temperature increases, the RNA starts to unfold, resulting in more single-stranded regions accessible to the dye. As temperature is further increased, the dye dissociates from the RNA and fluorescence signal is reduced. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2

A flowchart of the experimental protocol for our FRET/DS-FRET assay. Sections in the text corresponding to each step are indicated.

Fig. 3

Design and generation of bimolecular FRET constructs. (A) Secondary structures of wild type M1^TH and two bimolecular FRET constructs. For M1^ET, we remove the basal linker of the triplex for inclusion of Cy3 (green star) and Cy5 (red star) on the bottom of the triplex region. For M1^AB, we remove a portion of the apical duplex and incorporate Cy3 and Cy5 near the top of the triplex. (B) Representative size-exclusion chromatogram demonstrating purification of bimolecular, dually-labeled M1^ET. Absorbance was monitored at 260 nm (black trace), 550 nm (green trace), and 650 nm (red trace), corresponding to RNA, Cy3, and Cy5, respectively. The major peak featuring absorbance of all three signals, eluting at 9 min, contains the complexed M1^ET triple helix. The peak at 14 min, with signals for just RNA and Cy5, contains excess M1^T-Cy5. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

FRET assay evaluates small molecule-induced conformational changes. (A) Schematic of a small molecule disruption of RNA structure, promoting a less folded conformation recorded as a decreased FRET efficiency. (B) A representative dose-response curve for a compound that decreases FRET. (C) A representative dose-response curve for a compound that increases FRET.

Fig. 5

Differential scanning FRET evaluates changes to RNA thermal stability. (A) Schematic of differential scanning FRET experiment. The reaction begins at room temperature, and the RNA is fully folded, exhibiting high E_FRET (1). As temperature increases, the RNA begins to unfold, and E_FRET decreases (2). At high temperatures, the RNA is completely unfolded, and there is minimal FRET between the fluorophores (3). (B) A representative plot of E_FRET versus temperature with increasing concentrations of compound. As the compound concentration is increased from low (red) to high (green), the E_FRET curves shift to lower temperatures. (C) A plot of the derivative of the E_FRET curves in (B). (D) A plot of the T_m values determined from the peaks of the derivative curves in (C). In this example, the small molecule decreases the stability of the RNA. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6

Potential challenges when evaluating RNA-small molecule interactions with dual-reporter FRET/DS-FRET assay. (A) Representative plot of E_FRET and corresponding Cy3 and Cy5 fluorescence values as a function of increasing compound titration. Decreasing fluorescence values for both Cy3 and Cy5 indicate fluorophore quenching due to interactions with the small molecule. In such cases the E_FRET curve does not represent a genuine conformational change. (B-C) Representative Cy3 and Cy5 melt curves in the presence of (B) vehicle alone and (C) 50 μM of a compound that quenches Cy5 fluorescence. (D-F) Representative example of melt curves for which non-sigmoidal Cy3 and Cy5 curves complicate reliable determination of a T_m. (D) Cy3 and Cy5 melt curves. Note the presence of two inflection points in each curve. (E) The E_FRET calculated from the Cy3 and Cy5 curves in (D). (F) The negative derivative of the E_FRET curve in (E). The presence of a second peak in the derivative plot arises from the non-sigmoidal melt curves.