Comparative genomics reveals 104 candidate structured RNAs from bacteria, archaea, and their metagenomes

Abstract

Background: Structured noncoding RNAs perform many functions that are essential for protein synthesis, RNA processing, and gene regulation. Structured RNAs can be detected by comparative genomics, in which homologous sequences are identified and inspected for mutations that conserve RNA secondary structure.

Results: By applying a comparative genomics-based approach to genome and metagenome sequences from bacteria and archaea, we identified 104 candidate structured RNAs and inferred putative functions for many of these. Twelve candidate metabolite-binding RNAs were identified, three of which were validated, including one reported herein that binds the coenzyme S-adenosylmethionine. Newly identified cis-regulatory RNAs are implicated in photosynthesis or nitrogen regulation in cyanobacteria, purine and one-carbon metabolism, stomach infection by Helicobacter, and many other physiological processes. A candidate riboswitch termed crcB is represented in both bacteria and archaea. Another RNA motif may control gene expression from 3'-untranslated regions of mRNAs, which is unusual for bacteria. Many noncoding RNAs that likely act in trans are also revealed, and several of the noncoding RNA candidates are found mostly or exclusively in metagenome DNA sequences.

Conclusions: This work greatly expands the variety of highly structured noncoding RNAs known to exist in bacteria and archaea and provides a starting point for biochemical and genetic studies needed to validate their biologic functions. Given the sustained rate of RNA discovery over several similar projects, we expect that far more structured RNAs remain to be discovered from bacterial and archaeal organisms.

Figure 1

SAM/SAH riboswitches. (a) SAM/SAH motif consensus diagram. Possible additional base-pairing interactions are shown (Additional File 1). The legend applies to all other consensus diagrams in this report. (b) Sequence and proposed secondary structure of SK209-52 RNA. In-line probing annotations are derived from the data in c. Asterisks identify G residues added to improve in vitro transcription yield. (c) In-line probing gel with lanes loaded with 5' ³²P-labeled RNAs subjected to no reaction (NR), partial digestion with RNase T1 (T1), partial digest under alkaline pH (^-OH), in-line probing reaction without added compound (-), or in-line probing reactions with various concentrations of SAM. Selected bands in the RNase T1 partial digest lane (products of cleavage 3' of G residues) are numbered according to the nucleotide positions in b. Uncleaved precursor (Pre) and two internucleotide linkages whose cleavage rates are strongly affected by SAM (3' of nucleotides 42 and 45) are marked. The full gel image is provided in Additional File 1. (d) Plot of the normalized fraction of RNAs whose cleavage sites (linkage 23 not shown in c ) have undergone modulation versus the concentration of SAM present during the in-line probing reaction. The curve represents an ideal one-to-one binding interaction with a K_D of 8.6 μM.

Figure 2

Riboswitch candidates crcB, yjdF, wcaG, manA, pfl, epsC, and ykkC-III. Annotations are as described in Figure 1a. The transcription terminators that often overlap crcB or pfl RNAs are not depicted because they are not consistent in all representatives. They are annotated in Additional File 3. Question marks signify base-paired regions ("P4?" in yjdF, "P2?" in pfl, and "pseudoknot?" in manA) with weaker covariation or structural conservation. The pseudoknot in the epsC motif was predicted by others (Wade Winkler, personal communication, 2009). A portion of this figure was adapted from the supplementary data of a previous publication [21].

Figure 3

Riboswitch candidates glnA and Downstream-peptide. Annotations are as described in Figure 1a. Purple lines and numbers indicate conserved sequences or structures common to the two motifs.

Figure 4

Cyanobacterial motifs related to photosynthesis. Annotations are as described in Figure 1a.

Figure 5

Examples of other candidate RNAs. Annotations are as described in Figure 1a. The Bacteroidales-1 motif has more conserved nucleotides than depicted (Additional File 6). A portion of this figure was adapted from the supplementary data of a previous publication [21].