Emerging structural themes in large RNA molecules

Abstract

Extensive networks of tertiary interactions give rise to unique, highly organized domain architectures that characterize the three-dimensional structure of large RNA molecules. Formed by stacked layers of a near-planar arrangement of contiguous coaxial helices, large RNA molecules are relatively flat in overall shape. The functional core of these molecules is stabilized by a diverse set of tertiary interaction motifs that often bring together distant regions of conserved nucleotides. Although homologous RNAs from different organisms can be structurally diverse, they adopt a structurally conserved functional core that includes preassembled active and/or substrate binding sites. These findings broaden our understanding of RNA folding and tertiary structure stabilization, illustrating how large, complex RNAs assemble into unique structures to perform recognition and catalysis.

Figure 1. Architecture of RNA molecules

A) Diagrams of various RNA molecules (> 100 nucleotides) illustrating some of their general properties. Two perpendicular views of each structure are shown. Large RNAs are formed by stacked stems that are stabilized by a variety of tertiary interactions. The Thermoanaerobacter tengcongensis glmS riboswitch ribozyme structure illustrates the packing of two long stems that form a single, extended layer. The Thermotoga maritima lysine riboswitch illustrates an elongated molecule formed by three packed stems, forming a two layer molecule. The two major stems interact with each other through a kissing loop (red/orange). The Bacillus subtilis M-box riboswitch also forms parallel stacked stems. More complex molecules have a more globular, but still quasi-planar structure. In some cases, the addition of protein increases their functionality. In the case of the phage Twort group I ribozyme, binding to the homodimeric N. crassa CYT-18 protein promotes the splicing of many introns. Note that the protein binds on the surface of the RNA. The Oceanobacillus iheyensis group II intron structure represents a remarkable example of an RNA molecule with a complex fold. This large RNA forms a tight, compact structure around the active site region (red). Similar to proteins, large RNA molecules can consist of functional domains. The structure of the RNase P holoenzyme is an excellent example of a ribozyme composed of two functional domains, a catalytic (blue) and a specificity (yellow) domain. In this case, the protein also binds on the surface and does not alter the RNA structure. B) Diagrams of the structure of the Thermus thermophilus 30S ribosome subunit with and without the protein subunits. The RNA moiety (left) is also quasi-planar, despite its large size. In the presence of the proteins (right) the overall shape remains, with the protein molecules mostly found on the surface. The structures (on a clockwise fashion starting from the top left corner) correspond to PDBs: 2Z75, 3DIL, 2QBZ, 2RKJ, 3IGI, and 3OK7 [11,13,14,16,17,43,48]. The structures on each panel are drawn at the same scale, the bar on the right corresponds to 100 Å.

Figure 2. Large ribozymes have defined core architectures that are similar across different organisms

A) Secondary structure schematic diagram emphasizing the shared functional core between the known ancestral (A-type) and bacillus (B-type) P RNA structures of RNase P [11,32]. The dashed box encompasses elements involved in core formation, whereas grey denotes regions that are variable between the structures. Regions that form junctions and that contain universal (cyan) or bacterially (purple) conserved nucleotides are also shown. All conserved nucleotides are labeled. Crystal structures of the T. maritima and Bacillus stearothermophilus P RNAs are shown for comparison [11,32] (PDB 3OK7 and 2A64). B) Schematic diagram of the shared functional core amongst three group I intron structures from different organisms. Colors and labels are as defined in A) with the green, blue, and purple labels in the variable grey helical region corresponding to the different intron structures from Tetrahymena thermophila, Azoarcus sp. BH72, and phage Twort, respectively [12,49,50] (PDB 1X8W, 1U6B, 1Y0Q) [12,49,50]. Despite the presence of different peripheral domains, all three group I structures exhibit the same core domain architecture. Grey dashed regions represent engineered RNA motifs added to promote crystal formation. The diagram in B) was adapted from [30].

Figure 3. Proposed active sites based on RNase P, group I, and group II intron structures

Whereas these catalytic RNAs require at least two metal ions, putatively magnesium (pink spheres), not all ribozymes contain metal cofactors within their active sites. In these large RNA molecules the overall active site structure does not change upon substrate binding and is largely preassembled, although correct substrate positioning likely enhances the rigidity of the active site and binding of catalytic metal ions. A) Structure of the T. maritima RNase P holoenzyme in complex with tRNA [11] (PDB 3OKB) shows components that comprise the active site scaffold. B) The Azoarcus group I intron (ribo-ΩG) with bound exons [42] (PDB 1ZZN). The group I active site (light blue) contains two A-rich regions and structurally conserved base triples that help to position an external nucleotide (ΩG, orange) and exons (red) during cleavage and ligation. The nucleophile and scissile bond for the ligation step are shown in yellow. C) Structure of the O. iheyensis group II intron shows three metal ions in the region where exon ligation occurs [33,46] (PDB 3EOG). Yellow dashed lines indicate metal-ligand interactions that are ≤ 3.0 Å for the RNase P and group II structures, and within ≤ 2.2 Å for the Azoarcus group I structure.

Figure 4. Recognition of molecular targets by large RNA molecules is based on shape complementarity and specific interactions

A) T. tengcongensis glmS ribozyme riboswitch [16] (PDB 2Z75) in complex with its ligand cofactor, glucosamine-6-phosphate (red), reveals the importance of specific stabilizing interactions that orient and position the substrate within the cofactor binding pocket. B) T. maritima RNase P holoenzyme recognition of tRNA^Phe (red) [11] (PDB 3OKB) is primarily achieved by precise positioning of two base-stacks (marked by an *) and an A-minor motif (•) in the specificity domain with the unique, tertiary fold of the tRNA TΨC- and D-loops (pink panel). The complex is further stabilized by base-pairing interactions between the RNase P catalytic domain and the tRNA 3′-CCA sequence (blue panel). C) Azoarcus group I intron [42] (PDB 1ZZN) with both exons (red) positioned by base-pairing and base-stacking interactions in a catalytic state prior to the exon ligation step (pink and blue panels). Three base-stacks (*, green panel) stabilize an important junction (dark purple) in close proximity to the active site. The 3′-terminal guanosine (ΩG,*) nucleophile is both positioned by and embedded within a conserved cluster of stacked base triples (orange panel). In all panels, purple and white spheres represent magnesium atoms and water molecules, respectively.