Direct structural analysis of modified RNA by fluorescent in-line probing

Abstract

Chemical probing is a common method for the structural characterization of RNA. Typically, RNA is radioactively end-labelled, subjected to probing conditions, and the cleavage fragment pattern is analysed by gel electrophoresis. In recent years, many chemical modifications, like fluorophores, were introduced into RNA, but methods are lacking that detect the influence of the modification on the RNA structure with single-nucleotide resolution. Here, we first demonstrate that a 5'-terminal (32)P label can be replaced by a dye label for in-line probing of riboswitch RNAs. Next, we show that small, highly structured FRET-labelled Diels-Alderase ribozymes can be directly probed, using the internal or terminal FRET dyes as reporters. The probing patterns indeed reveal whether or not the attachment of the dyes influences the structure. The existence of two dye labels in typical FRET constructs is found to be beneficial, as 'duplexing' allows observation of the complete RNA on a single gel. Structural information can be derived from the probing gels by deconvolution of the superimposed band patterns. Finally, we use fluorescent in-line probing to experimentally validate the structural consequences of photocaging, unambiguously demonstrating the intentional destruction of selected elements of secondary or tertiary structure.

Figure 1.

A fluorescent label can replace a ³²P in riboswitch binding studies. (a) In-line probing analysis of the 5′-Atto 633 labelled lysC lysine riboswitch aptamer from B. subtilis. Reactions were performed in the absence (lane B) or in the presence of l-lysine (1 nM to 1 mM). NR, T1 and OH represent unreacted RNA, a G-specific sequencing ladder and a partial alkaline digest, respectively. Selected bands in the T1 lane are annotated with the nucleotide position. Regions 1–3 represent sites where the extent of spontaneous cleavage is modulated by the lysine concentration and were assigned either directly or by analogy to published probing patterns (32,34). (b) In-line probing analysis of the 5′-³²P-labelled lysC lysine riboswitch aptamer. (c) Secondary structure representing the lysine riboswitch aptamer. Important tertiary interactions and structural elements are marked. Nucleotides depicted in blue show spontaneous cleavage under in-line probing conditions. Nucleotides depicted in red show lysine-induced changes of their spontaneous cleavage. (d) Comparison of the intensity of the cleavage bands in all three regions of lysine-induced cleavage modulation. For both labels, the intensity of the cleavage band at 1 mM l-lysine is divided by the cleavage intensity in the absence of ligand (after background correction). (e) Plot depicting the normalized fraction of RNA cleaved (after correction for loading differences using the band of strong, lysine-independent cleavage annotated with ‘correction’) in region 1 for both labels (³²P in black and Atto 633 in red) versus the concentration of l-lysine.

Figure 2.

In-line probing analysis with two terminal labels. (a) Secondary structure representing the DAse ribozyme. Important structural features are indicated. Dashed lines represent a nested pseudoknot. U50 is put in brackets as it is present only in constructs A and C. The table shows the position of the dye labels and photocleavable NPE groups used in this study. (b) Schematic representation of an RNA labelled with fluorescent dyes on both termini. Stars represent the dyes and white gaps cleavage sites resulting from a chemical probing experiment. The three panels labelled with Cy3, Cy5 and + represent gel scans of the dyes and a superimposed image of both scans, respectively. Yellow bands represent superposition in both scans. (c) In-line probing analysis of the 5′-³²P-labelled wild-type DAse ribozyme. Reactions were performed at three different Mg²⁺-concentrations (0, 20, 80 mM). (d) In-line probing analysis of a ^Cy35′-^Cy5U50-labelled DAse construct A. Both dyes were scanned independently from the same gel (images separated by a double line). Selected cleavage bands are annotated on the left (for the Cy3 scan) and on the right (for the Cy5 scan) of the gel.

Figure 3.

In-line probing analysis with one terminal and one internal label. (a) In-line probing analysis of a ^Cy35’-^Cy5U33-labelled DAse construct B. The right image is a subtraction of the Cy3 scan from the Cy5 scan after normalization. Selected cleavage bands are annotated on the left (for the Cy3 scan) and on the right (for the subtraction image) of the gel. The position of the internal label is indicated by the dashed lines. (b) Schematic representation of the probing experiment. The panel labelled with Cy5–Cy3 represents a subtraction image with all values <0 set to zero. The yellow box indicates which cleavage products superimpose in both scans (indicated as yellow bands). The dashed box shows the cleavage products visible in the subtraction image.

Figure 4.

In-line probing analysis with two internal labels. (a) Schematic representation of the experiment. The panels labelled with Cy5–Cy3 and Cy3–Cy5 are subtraction images with all values <0 set to zero. Yellow bands in the superimposed image indicate cleavage products present in both scans. The dashed box shows the part of the sequence that can be analysed unambiguously. The dashed line shows the border above which only cleavage products carrying both labels are present. (b) In-line probing analysis of four DAse ribozyme constructs (A–D). The images show a section of the unmodified Cy3 scan from the gels visible in Figure 2d, 3a and Supplementary Figure S3. Below each image, the relative catalytic activity, with respect to the wild-type ribozyme, is given.

Figure 5.

In-line probing analysis of five DAse ribozyme constructs (B–H). All constructs are labelled with two fluorescent dyes (^Cy35′–^Cy5U33). Constructs E–H are modified with photocleavable NPE protecting groups in different positions. The image shows the Cy3 scan, except for the rightmost panel, which is the subtraction image (Cy5–Cy3) of construct H. Black connecting lines illustrate the influence of an attached NPE group on the electrophoretic mobility.