An overview of gene regulation in bacteria by small RNAs derived from mRNA 3' ends

Abstract

Over the past two decades, small noncoding RNAs (sRNAs) that regulate mRNAs by short base pairing have gone from a curiosity to a major class of post-transcriptional regulators in bacteria. They are integral to many stress responses and regulatory circuits, affecting almost all aspects of bacterial life. Following pioneering sRNA searches in the early 2000s, the field quickly focused on conserved sRNA genes in the intergenic regions of bacterial chromosomes. Yet, it soon emerged that there might be another rich source of bacterial sRNAs-processed 3' end fragments of mRNAs. Several such 3' end-derived sRNAs have now been characterized, often revealing unexpected, conserved functions in diverse cellular processes. Here, we review our current knowledge of these 3' end-derived sRNAs-their biogenesis through ribonucleases, their molecular mechanisms, their interactions with RNA-binding proteins such as Hfq or ProQ and their functional scope, which ranges from acting as specialized regulators of single metabolic genes to constituting entire noncoding arms in global stress responses. Recent global RNA interactome studies suggest that the importance of functional 3' end-derived sRNAs has been vastly underestimated and that this type of cross-regulation between genes at the mRNA level is more pervasive in bacteria than currently appreciated.

Keywords: 3′ UTR; bacteria; post-transcriptional control; regulatory networks; sRNA.

Figure 1.

The many different sources of sRNAs as ‘parallel transcription output’. The canonical sRNA biogenesis pathways (left column) refer to sRNA production by transcription of stand-alone noncoding genes (right) located next to protein-coding genes (left). Alternative pathways (right column) produce sRNAs from mRNA loci by premature transcription termination in the 5′ region, by transcription starting inside the coding sequence (CDS) but using the same terminator or by mRNA processing. The focus of this review is on sRNAs that accumulate as 3′ end processing products of mRNAs and thus carry a monophosphate at their 5′ end.

Figure 2.

Biogenesis of 3′ UTR-derived sRNAs in Gram-negative and -positive bacteria. (A) In Gram-negative bacteria, the major endonuclease RNase E is the primary nuclease to produce 3′ UTR-derived sRNAs from their parental mRNAs, as shown here for the CpxQ sRNA and the cpxP mRNA. In this case, the Hfq protein is also required for biogenesis. (B) While RNase E is lacking in Gram-positive bacteria, RNase III was shown to free the RsaC sRNA from the mntABC operon mRNA by recognizing a double-stranded RNA structure.

Figure 3.

CxpQ as a well-understood example of 3′ UTR-derived sRNAs. (A) Nucleotide sequence alignment highlighting the strong conservation of the CpxQ sRNA in comparison to the 3′ region of the cpxP mRNA (STM: S. Typhimurium; KPN: Klebsiella pneumoniae; PAN: Pantoea spp.; Con: consensus sequence). (B) Inner membrane (IM) stress leads to the phosphorylation of the transcription factor CpxR. Phosphorylated CpxR activates transcription of cpxP mRNA that is translated, yielding CpxP protein, and processed by RNase E to yield CpxQ sRNA. While CpxP is involved in the degradation of misfolded inner membrane proteins (IMPs), CpxQ acts as a post-transcriptional regulator by repressing the translation of several IMPs. Thus, both the coding and noncoding parts of the cpxP mRNA act cooperatively to maintain IM homeostasis in the Cpx pathway.

Figure 4.

Enterobacterial 3′ UTR-derived sRNAs are involved in diverse pathways. (A) The DicF sRNA stems from inside the E. coli ydfABC-dicF-dicB-ydfD operon mRNA that is transcribed from a defective prophage region. Processing of the mRNA by both RNase E and RNase III yields DicF, which in turn is involved in the regulation of cell division and metabolism. Additionally, in EHEC, DicF is also important for the regulation of the pathogen's virulence by upregulating the transcriptional activator PchA. (B) RNase E-dependent processing at the 3′ end of the sdhCDAB-sucABCD mRNA generates the sRNA SdhX, which selectively acts on ackA of the bicistronic ackA-ptA operon to regulate acetate metabolism. Additionally, SdhX exhibits divergent targetomes in E. coli and Salmonella. (C)Premature transcriptional termination of the gltIJKL operon and subsequent RNase E-mediated cleavage frees the sRNA SroC from its parental operon. While SroC is involved in regulating motility by repressing translation of fliE, SroC has an expanded regulatory capacity through sponging the sRNAs GcvB and MgrR to further affect metabolism and LPS modification, respectively. (D) NarS is processed off the 3′ end of narK mRNA by RNase E. By selectively inhibiting translation of nirC as part of the nirBDC-cysG operon, this sRNA works synergistically with the NarK protein to fine-tune nitrite uptake.

Figure 5.

3′ UTR-derived sRNAs in V. cholerae. (A) RNase E-mediated processing yields the sRNA FarS from its parental mRNA fabB. Through inhibition of fadE translation, FarS is involved in the regulation of the fatty acid metabolism of the pathogen. (B) The 3′ end of the oppABCDF operon contains the sRNA OppZ, which is generated through RNase E processing. OppZ acts self-regulating by initiating Rho-dependent transcription termination downstream of oppA. This leads to the fine-tuning of the protein stoichiometry of the OppABCDF system.

Figure 6.

Further examples of 3′ UTR-derived sRNA in diverse organisms. (A) The sRNA RaiZ derived from the 3′ end of the raiA mRNA acts in ProQ-dependent manner by suppressing the globally acting DNA-binding protein hupA affecting metabolism and virulence of Salmonella and E. coli. (B) Processing of the sodF mRNA by an unknown ribonuclease gives rise to the s-SodF sRNA. The sRNA allows S. coelicolor to switch the superoxide dismutase (SOD) in response to oxidative stress under Ni-limited conditions. (C) In Staphylococcus, the sRNA RsaC is processed via RNase III from the mntABC operon at low Mn²⁺ levels. Under these conditions, RsaC represses the Mn-dependent SOD sodA allowing an efficient response against oxidative stress. (D) The sRNA RsaG is generated via 5′ → 3′ degradation by RNase J1/J2 of the full-length uhpT mRNA. Binding of RsaG to its target can lead to either stabilization or degradation of the bound mRNAs. In case of the former, RsaG binds the rex mRNA, which is a redox regulator. In case of the latter, binding of RsaG accelerates degradation of the lactate dehydrogenase (ldh1) mRNA and thus impacts metabolism. (E) The sRNA SorX is part of the 3′ end of RSP_0847 in R. sphaeroides and generated through an unknown RNase combined with RNase E. The released sRNA inhibits spermidine uptake through translational inhibition of the polyamine uptake transporter potA and thus supports fighting off oxidative stress. (F) The sRNA PcrX represents the 3′ end of the photosynthesis complex pufQBALMX operon and is generated via RNase E. PcrX acts self-regulating by destabilizing its parental mRNA through a yet unknown mechanism.

Figure 7.

Common regulatory networks of 3′ UTR-derived sRNAs. (A) In V. cholerae, the transcription factor FadR activates the expression of FarS and its parental mRNA fabB, of which the latter is involved in fatty acid synthesis. Simultaneously, FadR and FarS repress fadE expression that is part of the opposing β-oxidation pathway for fatty acids. Thus, this mixed feed-forward loop enables an efficient fatty acid metabolism. (B)The autoregulatory sRNA OppZ ensures a proper balance of the different proteins of the Opp peptide uptake system in V. cholerae. Through causing premature transcription termination of its parental mRNA, for all but the peptide-binding protein OppA, OppZ limits expression of oppBCDF as well as itself. (C) Mixed feed-forward loops can also be involved in the regulation of stress. As exemplified by the CpxR-dependent protein CpxP and sRNA CpxQ of the Enterobacteriales. Through translational inhibtion of diverse IMPs, CpxQ reduces the CpxR-mediated transcription of CpxP as well as the sRNA itself. Both components of the Cpx pathway ensure a reduction of misfolded IMPs and thus alleviate IM stress.

Figure 8.

Global methods to uncover 3′ UTR-derived sRNAs and their targets. (A) A schematic overview for the workflow of methods relying on proximity ligation to identify sRNAs and their targets such as RIL-seq or CLASH. Split read mapping of the ligated RNA products allows the identification of the interacting transcripts and can thus globally uncover interactions between 3′ UTRs of genes and potential targets. (B) Grad-seq can enable the identification of novel 3′ UTR-derived sRNAs in a global manner. Fractionation of a cell lysate and subsequent sequencing can uncover differential migration patterns of a parental mRNA and its 3′ UTR and thus indicate an additional function of the latter as an sRNA, as shown for the example of SroC (data acquired from Smirnov et al. 2016).

Figure 9.

3′ UTR-derived sRNAs are a conserved feature of their parental mRNA. (A) Sequence alignment for sucC and the sRNA SdhX (ECO: E. coli; KPN: Klebsiella pneumoniae; ROR: Raoultella ornithinolytica; Con: consensus sequence).(B) Sequence alignment for oppF and the sRNA OppZ (VCH: V. cholerae; VFU: V. furnissii; VAN: V. anguillarum; Con: consensus sequence). The RNase E cleavage sites are indicated.