Genome rearrangements can make and break small RNA genes

Abstract

Small RNAs (sRNAs) are short, transcribed regulatory elements that are typically encoded in the intergenic regions (IGRs) of bacterial genomes. Several sRNAs, first recognized in Escherichia coli, are conserved among enteric bacteria, but because of the regulatory roles of sRNAs, differences in sRNA repertoires might be responsible for features that differentiate closely related species. We scanned the E. coli MG1655 and Salmonella enterica Typhimurium genomes for nonsyntenic IGRs as a potential source of uncharacterized, species-specific sRNAs and found that genome rearrangements have reconfigured several IGRs causing the disruption and formation of sRNAs. Within an IGR that is present in E. coli but was disrupted in Salmonella by a translocation event is an sRNA that is associated with the FNR/CRP global regulators and influences E. coli biofilm formation. A Salmonella-specific sRNA evolved de novo through point mutations that generated a σ(70) promoter sequence in an IGR that arose through genome rearrangement events. The differences in the sRNA pools among bacterial species have previously been ascribed to duplication, deletion, or horizontal acquisition. Here, we show that genomic rearrangements also contribute to this process by either disrupting sRNA-containing IGRs or creating IGRs in which novel sRNAs may evolve.

Keywords: E. coli; Salmonella; gene origination; intergenic regions; sRNA.

Fig. 1.—

Expression profiles within nonsyntenic IGRs. Putative sRNAs were detected by RNA-seq analysis of transcript levels within IGRs in E. coli (A, B) and Salmonella (C, D). For uniformity, the number of sequencing reads mapped to IGRs is limited to 2,000 (dashed line). Arrows showing the orientation of ORFs and putative sRNAs are not drawn to scale.

Fig. 2.—

Salmonella IGR formed through an HGT-mediated genome rearrangement. Most homologs of STM14_1870 (stbE, blue arrow) and STM14_1871 (stbD, green arrow) are situated on bacterial plasmids, and STM14_1872 (purple arrow) is a prophage gene. Both the STM14_1869–STM14_1870 IGR and the sRNA (SesR2) are present only in Salmonella. In Shigella, an insertion sequence (IS1294) flanks the stbE gene.

Fig. 3.—

Evolution of a new sRNA promoter. Sequences immediately upstream of STM14_1869 (4471230–4471245) and its ortholog yjgH in E. coli (1636369–1636384) are aligned. Numbers of RNA-seq reads mapping to this region are shown (black, Salmonella; blue, E. coli). The new Salmonella σ⁷⁰ promoter and sRNA (SesR2) transcription start site are boxed. Asterisks indicate point mutations that differentiate the Salmonella sequence from the corresponding region in E. coli.

Fig. 4.—

Distribution of the uspF–ompN IGR among enteric species. (A) Alignment of genomic regions containing the uspF–ompN IGR in E. coli and three other enteric species. Note that in both Citrobacter koseri and Klebsiella pneumoniae, small ORFs (gray arrows situated between uspF [purple] and ompN [blue]) have been predicted to occur in this IGR. (B) Phylogenetic tree (modified from Petty et al. 2010) showing the presence or absence of the uspF–ompN IGR among species.

Fig. 5.—

Loss of uspF–ompN IGR through genome rearrangement. The uspF–ompN IGR of E. coli was fragmented in Salmonella due to the translocation of ompN to a site adjacent to STM14_1775. The predicted FNR- and CRP-binding sites (yellow and blue, respectively; overlapping region in green) upstream of the sRNA (EcsR1) transcription start site (sRNA TSS) are shown. A predicted Rho-independent terminator (stem-loop structure) situated 3′ of the sRNA is also depicted.

Fig. 6.—

Biofilm formation is influenced by EcsR1. Escherichia coli biofilms stained with crystal violet were measured (A₆₀₀) after 48-h growth at 28 °C and normalized to OD₆₀₀ value. A wild-type strain, an EcsR1-deleted strain (ΔEcsR1), a ΔEcsR1 strain containing pBAD with cloned EcsR1 (ΔEcsR1-pBAD-EcsR1), and a ΔEcsR1 strain containing empty pBAD (ΔEcsR1-empty pBAD) were tested. Asterisks indicate a statistically significant difference between wild-type and ΔEcsR1 strains (P < 0.0001).