Exceptional structured noncoding RNAs revealed by bacterial metagenome analysis

Abstract

Estimates of the total number of bacterial species indicate that existing DNA sequence databases carry only a tiny fraction of the total amount of DNA sequence space represented by this division of life. Indeed, environmental DNA samples have been shown to encode many previously unknown classes of proteins and RNAs. Bioinformatics searches of genomic DNA from bacteria commonly identify new noncoding RNAs (ncRNAs) such as riboswitches. In rare instances, RNAs that exhibit more extensive sequence and structural conservation across a wide range of bacteria are encountered. Given that large structured RNAs are known to carry out complex biochemical functions such as protein synthesis and RNA processing reactions, identifying more RNAs of great size and intricate structure is likely to reveal additional biochemical functions that can be achieved by RNA. We applied an updated computational pipeline to discover ncRNAs that rival the known large ribozymes in size and structural complexity or that are among the most abundant RNAs in bacteria that encode them. These RNAs would have been difficult or impossible to detect without examining environmental DNA sequences, indicating that numerous RNAs with extraordinary size, structural complexity, or other exceptional characteristics remain to be discovered in unexplored sequence space.

Figure 1. Size and structural complexity of new-found RNAs compared to the ten largest known bacterial ncRNAs with complex structures

Structural complexity is represented by the number of multistem junctions plus pseudoknots (see full Methods for details). RNAs described in this report are in bold type. HEARO and Group I ribozyme symbols overlap. Narrowly distributed RNAs (present in only one bacterial class) are not included.

Figure 2. GOLLD RNAs

a, Simplified consensus sequence and secondary structure model for the most common architecture of GOLLD RNAs. Annotated 5′ and 3′ ends reflect L. brevis transcripts observed by RACE experiments (Supplementary Fig. 3). b, Phage induction and expression of GOLLD RNA. Experimental details are presented in the full Methods.

Figure 3. HEARO RNAs

a, Consensus sequence and secondary structure model for HEARO RNAs. Annotations are as described in the legend to Fig. 3a. b, Typical sequence signature of HEARO genomic integration (see also Supplementary Fig. 6). (Top) HEARO element and flanking sequence in Anabaena variabilis ATCC 29413, plasmid C (NC_007412.1). Green text designates DNA corresponding to the first five nucleotides of conserved HEARO RNA. Blue text designates DNA corresponding to the conserved RUGA motif at each integration site. (Bottom) Homologous genome sequence lacking the HEARO element from Nostoc sp. PCC 7120, plasmid pCC7120beta (NC_003240.1). Red nucleotides identify positions that vary between the two genomes.