Large Noncoding RNAs in Bacteria

Abstract

Bacterial noncoding RNA (ncRNA) classes longer than 200 nucleotides are rare but are responsible for performing some of the most fundamental tasks in living cells. RNAs such as 16S and 23S rRNA, group I and group II introns, RNase P ribozymes, transfer-messenger RNAs, and coenzyme B₁₂ riboswitches are diverse in structure and accomplish biochemical functions that rival the activities of proteins. Over the last decade, a number of new classes of large ncRNAs have been uncovered in bacteria. A total of 21 classes with no established functions have been identified through the use of bioinformatics search strategies. Based on precedents for bacterial large ncRNAs performing sophisticated functions, it seems likely that some of these structured ncRNAs also will prove to carry out complex functions. Thus, determining their roles will provide a better understanding of fundamental biological processes. A few studies have produced data that provide clues to the purposes of some of these recently found classes, but the true functions of most classes remain mysterious.

FIGURE 1

Size and structural complexity of large and highly structured ncRNAs in bacteria. Structural complexity is represented by the number of multistem junctions and pseudoknots present in the predicted secondary-structure models, as described previously (18). Overlapping points representing different ncRNAs are depicted with split circles. Narrowly distributed ncRNAs and ncRNAs with <2 multistem junctions and pseudoknots were omitted. For example, noncoding RNAs such as large sRNAs and clustered regularly interspaced short palindromic repeat (CRISPR) RNAs are commonly >200 nucleotides long but have repetitive and simple hairpin secondary structures that are bound by proteins. Although 23S rRNA forms the active site for the peptidyltransferase reaction catalyzed by ribosomes, 16S rRNA functions in complex with the catalytic RNA component and is classified accordingly.

FIGURE 2

Consensus sequence and secondary-structure model for OLE RNAs. This model is based on the alignment of 657 unique representatives from genomic sequences from RefSeq version 63 and metagenomic sequences as described in reference . R and Y represent purine and pyrimidine nucleotides, respectively.

FIGURE 3

Consensus sequence and secondary-structure model for GOLLD RNAs. This model is based on the alignment of sequences identified in reference . Notable predicted substructures include 2 E-loops, 3 GNRA tetraloops, and 5 pseudoknots. Of the 20 hairpin loops, 5 form pseudoknots or represent GNRA tetraloops. A total of 12 of the remaining 15 hairpin loops carry highly conserved nucleotides, suggesting that they might be involved in forming RNA tertiary contacts that are important for the function of GOLLD RNA. Other annotations are as described for Fig. 2.