Evolutionary rates vary among rRNA structural elements

Abstract

Understanding patterns of rRNA evolution is critical for a number of fields, including structure prediction and phylogeny. The standard model of RNA evolution is that compensatory mutations in stems make up the bulk of the changes between homologous sequences, while unpaired regions are relatively homogeneous. We show that considerable heterogeneity exists in the relative rates of evolution of different secondary structure categories (stems, loops, bulges, etc.) within the rRNA, and that in eukaryotes, loops actually evolve much faster than stems. Both rates of evolution and abundance of different structural categories vary with distance from functionally important parts of the ribosome such as the tRNA path and the peptidyl transferase center. For example, fast-evolving residues are mainly found at the surface; stems are enriched at the subunit interface, and junctions near the peptidyl transferase center. However, different secondary structure categories evolve at different rates even when these effects are accounted for. The results demonstrate that relative rates and patterns of evolution are lineage specific, suggesting that phylogenetically and structurally specific models will improve evolutionary and structural predictions.

Figure 1.

In bacteria, stems dominate in high-rate categories, unpaired regions in low-rate categories. For each structural category, we calculated the percent of positions (y-axis) in each rate category (x-axis). For example, of all the positions in stems, 8% is in rate category 1, 22% is in rate category 2, etc. The graph contains four different series, one for each structural category. Within one series the values add up to 100%. The data is calculated from the RDB variability categories superimposed on the RDB secondary structure model of E. coli.

Figure 2.

Pairwise sequence comparisons for bacteria and archaea. Stems evolve fastest, both in the complete set of bacteria and in individual lineages. However, the different classes of unpaired regions always evolve at significantly different rates. The scatterplots show the fraction divergence per structural category (y-axis) versus the fraction divergence overall (x-axis); see Methods for definition. (A) SSU bacteria. (B) SSU archaea. (C) LSU bacteria. (D) Bacteria without Actinobacteria. (E) Proteobacteria. (F) Firmicutes.

Figure 3.

Comparison between rates inferred using CRW and RDB rate categories showing that loops dominate high-rate classes in eukaryotes. The graphs show the fraction of positions (y-axis) per rate class (x-axis) for each structural element. Both graphs show data for S. cerevisiae. Left: CRW conservation values, CRW structural model, excluding the controversial region, including unclassified positions as highest rate class. Right: RDB variability scores, RDB structural model, including the controversial region and unclassified positions. These two figures present essentially the same data. The difference is caused by the different rate categories used by the two data sources. Rate categories 1 and 5 in CRW correspond roughly to 1/2 and 5/6/7 in RDB, respectively.

Figure 4.

Pairwise sequence comparisons for eukaryotes. Hairpins (blue) evolve fastest, both overall and in each lineage individually. The scatterplots show the fraction divergence per structural category (y-axis) versus the fraction divergence overall (x-axis); see Methods for definition. (A) SSU eukaryotes. (B) Viridiplantae and Metazoa. (C) Alveolata. (D) Fungi. (E) Stramenopiles. (F) LSU eukaryotes.

Figure 5.

Distribution of rate categories as a function of distance from the PTC. Each bar graph shows the fraction of atoms in each rate category (y-axis) versus the distance from the PTC [x-axis; last bin contains all atoms >100 Å (or 140 Å on the SSU graph) away from the PTC]. The fractions within a distance bin (vertical column) add up to 1.0. (A) LSU all structural elements. (B) SSU all structural elements. (C) LSU stem. (D) LSU loop. (E) LSU bulge. (F) LSU junction/other.

Figure 6.

Distribution of structural elements and rate categories in the ribosome. Each panel shows the T. thermophilus ribosome, with residues highlighted according to proximity to specific structural features and colored either by rate or by structural category. Panel (A) shows the large subunit with residues within 20 Å of the PTC (near the tRNA ends) and residues >80 Å away from the PTC (outer shell) highlighted. The residues are colored by rate category (fast evolving sites in orange/red, slow evolving sites in cyan/blue, sites changing at an intermediate rate in gray). The small subunit, 5S rRNA, and proteins are hidden. The PyMol script that generates these figures is available as Supplementary Data to allow interactive exploration of these features. Panel (B) shows all residues in rate category 1 colored blue, all residues in rate category 7 and 8 colored in red colors. Panel (C) shows all residues within 15 Å from PTC (right), within 15 Å from the tRNA path in the SSU (left) and within 10 Å on both sides of the subunit interface (middle), colored by structural element. Stem in yellow, loop in blue, bulge in red and junction in green.

Figure 7.

Distribution of structural elements in the ribosome as a function of distance from specific structural features. Each bar graph shows the fraction of atoms in each structural element (y-axis) versus the distance from a particular feature (x-axis; last bin contains all atoms more than that distance away). The fractions within a distance bin (vertical column) add up to 1.0. (A) SSU, tRNA path. (B) LSU, PTC. (C) SSU, subunit interface. (D) LSU, subunit interface.