Searching genomes for ribozymes and riboswitches

Abstract

New regulatory RNAs with complex structures have recently been discovered, among them the first catalytic riboswitch, a gene-regulatory RNA sequence with catalytic activity. Here we discuss some of the experimental approaches and theoretical difficulties attached to the identification of new ribozymes in genomes.

Figure 1

Biochemical reactions naturally catalyzed by RNA. (a) Precursor tRNA hydrolysis by bacterial RNase P yields a phosphate-containing 5' end of the mature tRNA and a 3'-hydroxyl group at the 5' cleavage product. (b-d) Transesterification reactions catalyzed by (b) the small nucleolytic ribozymes, (c) group I introns, and (d) group II introns, in which different chemical groups serve as the attacking nucleophile. In the small nucleolytic ribozymes (b), a defined 2'-hydroxyl attacks the neighboring 3',5'-phosphodiester bond, resulting in a 2',3'-cyclic phosphate and a 5'-hydroxyl in the respective cleavage products. In the first step of group I intron splicing (c), the 3'-hydroxyl of the exogenous guanosine (G) cofactor attacks the 5'-exon-intron junction and sets the 5' exon free, which leads to the covalent attachment of the cofactor to the 5' end of the intron. In a second transesterification reaction, the 5' exon forms a conventional 3',5' bond with the 3' exon, releasing the linear intron with the additional guanosine [1]. In group II introns (d), the conserved branch-point adenosine (A) serves as the nucleophile, leading to the formation of a lariat intron. (e) Peptide-bond formation catalyzed by the ribosome.

Figure 2

The hammerhead ribozymes are based on a three-way junction and there are two main types. (a) Type I has the ends of the single-stranded RNA on stem I; (b) type III has the ends of the single-stranded RNA on stem III. For unknown reasons, potential type II ribozymes (ends of the single-stranded RNA on stem II) have never been observed. The three-dimensional architecture is maintained by coaxial stacking of stems II and III, which, through constraints in the conserved three-way junction residues [92], orients stem I so that loop-loop interactions between stems I and II form (Figure 3) [40,42]. The internal loop of stem II (IL2) is often replaced by a capping loop (CL2); similarly, CL1 in type III can be replaced by an internal loop (IL1) followed by another hairpin. Although only one structure has been fully characterized, sequence alignments show that the loop-loop interactions (mainly constituting non-Watson-Crick pairs) are very diverse.

Figure 3

Schematic diagrams of the interaction networks maintaining the three-dimensional architecture of two different ribozymes. (a) The HDV ribozyme [7,93]; (b) the active hammerhead ribozyme [42]. The HDV ribozyme has a convoluted pseudoknotted topology: the color lines indicate the path of the sugar-phosphate backbone. The nomenclature is as follows [75]. Each nucleotide has three edges with hydrogen bonding possibilities: the Watson-Crick edge (denoted by a circle), the Hoogsteen edge (denoted by a square) and the sugar edge (denoted by a triangle). A pairwise base-base interaction can be formed either with the attached sugar moieties on the same side of the line of approach (cis-configuration, the symbols are closed) or with the sugars on either sides of the line of approach (the trans-configuration, the symbols are open). To avoid ambiguities, when annotating tertiary contacts, the nucleotides that are involved have been boxed. When the base of a nucleotide is in the syn-conformation with respect to the sugar it is marked in bold. The rectangles indicate the position actually occupied in space by a nucleotide. In (b), the cleavage occurs 3' of the red C.

Figure 4

Different local topologies can give rise to similar tertiary contacts in group I introns. (a) The invariant core of a group I intron [36,94] is illustrated in schematic form with the paired segments indicated by P and the loop regions by L. The dashed lines indicate the contacts between the peripheral elements, which are indicated by the numbers in circles. (b) Three different group I introns illustrate distinct ways of achieving a similar tertiary contact (involving non-Watson-Crick A-minor base-base interactions between a GAAA tetraloop and two stacked pairs) connecting distant regions. In each case region 9 folds towards region 5 (as indicated by the shaded region) but, in the Twort ribozyme [95] this is via a three-way junction, in the Tetrahymena ribozyme [96], it is via a large bend (this is not the natural junction, however), and in the Azoarcus ribozyme [97], it is via a kink-turn. Each motif has a different sequence and set of structural constraints [77,92].

Figure 5

Identification of catalytic RNA from a genomic library. (a) Preparation of the genomic library. Genomic DNA is first partially digested and fragments of approximately 150 bp (blue) are gel-purified and incubated with Taq polymerase to give them 3' A overhangs. In the next step, ligation of covalently closed oligonucleotides (yellow and purple) to the library prevents the unwanted combination of DNA fragments. After removal of DNA hairpins, a T7 promoter (magenta) is then added by PCR, yielding an amplified linear library. (b) The in vitro selection scheme. The library is further amplified by PCR using a 5'-phosphorylated reverse primer and a biotinylated forward primer that allows the isolation of the phosphorylated strand using streptavidin beads. Single strands are individually circularized by ligation with a splint oligonucleotide and the second strand is added by incubation with Taq polymerase and deoxynucleoside triphosphates. The resulting nicked double-stranded library is suitable for rolling-circle transcription by T7 polymerase [98], yielding multimeric RNA species potentially encoding sites of self-cleavage (red triangles). The RNA is then incubated for self-cleavage, and active molecules (dimers) are size-selected. The scheme is completed by preparation of the next-generation DNA library using reverse transcription-PCR (RT-PCR). Modified from [29].