• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of narLink to Publisher's site
Nucleic Acids Res. Dec 2008; 36(21): 6781–6794.
Published online Oct 25, 2008. doi:  10.1093/nar/gkn742
PMCID: PMC2588501

The 5′ end of two redundant sRNAs is involved in the regulation of multiple targets, including their own regulator

Abstract

Small RNAs are widespread regulators of gene expression in numerous organisms. This study describes the mode of action of two redundant Escherichia coli sRNAs, OmrA and OmrB, that downregulate the expression of multiple targets, most of which encode outer membrane proteins. Our results show that both sRNAs directly interact with at least two of these target mRNAs, ompT and cirA, in the vicinity of the translation initiation region, consistent with control of these targets being dependent on both Hfq and RNase E. Interestingly, these interactions depend on short stretches of complementarity and involve the conserved 5′ end of OmrA/B. A mutation in this region abolishes control of all OmrA/B targets tested thus far, thereby highlighting the crucial role of the OmrA/B 5′ end. This allowed us, by looking for mRNA sequences complementary to the OmrA/B 5′ end, to identify ompR as an additional direct target of these two sRNAs. Since the OmpR transcriptional regulator activates expression of both omrA and omrB genes, this newly identified control should result in an autoregulatory loop limiting the amount of OmrA/B sRNAs.

INTRODUCTION

Bacterial small RNAs (sRNAs) play major roles in the physiology of the cell. Not fully recognized until the last decade, now about 100 of these molecules have been identified in the Escherichia coli genome (1), which represents more than 2% of the total number of genes. These sRNAs are usually between 50 and 400 nt in length and in most cases do not encode any protein. In general, their synthesis is tightly regulated and, as a result, they are expressed under highly specific conditions. Once synthesized, sRNAs either positively or negatively regulate the expression of one or several target genes. Several major stress regulons have been shown to include at least one sRNA involved in the cellular stress response (2).

Bacterial sRNAs do not all follow the same mode of action; the most extensively studied group is the class of Hfq-binding sRNAs, comprising at least one-third of the experimentally verified E. coli sRNAs (3). These sRNAs usually base-pair with their target-mRNAs and regulate the translation and/or stability of the mRNA at the post-transcriptional level. In most cases, this regulation is negative and sRNAs inhibit ribosome binding by base-pairing in the vicinity of the target ribosome-binding site. Nevertheless, several examples of positive control by sRNAs have been reported that are generally based on the same mechanism: the pairing of the activator sRNA to its target mRNA inhibits the formation of an inhibitory secondary structure and thereby derepresses translation of the target gene (4–6).

Hfq is an RNA chaperone that was first identified for its role in the replication of the RNA phage Qβ, but now appears to have multiple roles for cellular RNAs (7). It was shown to stabilize several Hfq-binding sRNAs in vivo (8) and to enhance base-pairing between sRNA and target-mRNA in vitro (8–11); besides affecting sRNA accumulation, Hfq is also likely required for sRNA action in vivo (12), at least in Gram-negative bacteria. In addition, Morita et al. (13) showed that Hfq copurified with Hfq-binding sRNAs, such as SgrS or RyhB, and with the RNase E endonuclease. Hfq could therefore recruit a ribonucleic complex, which would account for the observed RNase E-dependent destabilization of a target-mRNA after its pairing with a regulatory sRNA (as in reference 14 for instance).

As the role of many sRNAs has started to be investigated, it appears that at least eight of these molecules, and probably more, downregulate the synthesis of outer membrane proteins (OMP) in E. coli and other enterobacteria (15–19). Two of these, OmrA and OmrB, are sRNAs encoded by adjacent genes on the E. coli chromosome. They are almost identical at their 5′ and 3′ ends, but have a distinct central region. We have previously shown that both OmrA and OmrB are made in response to the activation of the EnvZ-OmpR two-component system and that their production results in a decrease in the level of several OMPs, such as the OmpT protease and the CirA, FecA and FepA receptors for iron–siderophore complexes (20).

This study focuses on the mechanism by which OmrA/B sRNAs downregulate their targets. Our results show that a conserved short region at the 5′ end of OmrA/B directly base-pairs with ompT and cirA mRNAs in the translation initiation region (TIR). Interestingly, this 5′ end of OmrA/B is also involved in the regulation of their other targets, including the newly identified ompR target. Since OmpR activates the transcription of omrA and omrB, these sRNAs should limit their own synthesis by directly downregulating their activator.

MATERIAL AND METHODS

Bacterial strains

Strains used in this study are listed in Table 1. Unless otherwise indicated, cells were grown aerobically in LB medium at 37°C. When necessary, antibiotics were used at the following concentrations: 50 μg/ml ampicillin on plates or 150 μg/ml in liquid cultures, 25 μg/ml kanamycin and 10 or 25 μg/ml chloramphenicol. Mutant alleles of the rne gene (14) as well as Δfur::cm (20), Δfur::kan (21), ΔomrAB::kan (20) and hfq::cm (22) were introduced by P1 transduction. For cloning procedures, PCR amplification was carried out using the Expand High Fidelity PCR system (Roche) and DH5α was used as the recipient strain.

Table 1.
Strains used in this study

Strains carrying ompT-lacZ or cirA-lacZ translational fusions were constructed as follows. DNA fragments corresponding to nts –220 to +30 of either ompT or cirA followed by a Flag tag (sequence 5′-GACTACAAGGACGACGATGACAAA-3′) were cloned in frame with lacZ between the EcoRI and BamHI sites (for ompT) or between the EcoRI and SmaI sites (for cirA) of the pRS414 plasmid (23) to give plasmids pRSompT10 or pRScirA10, respectively. These plasmids were then crossed with λRS468 bacteriophage and monolysogens were constructed as previously described (23). Mutations in these translational fusions were introduced in the pRSompT10 or pRScirA10 plasmid using the Quick-change II XL site-directed mutagenesis kit (Stratagene) and mutant monolysogens were then made as above.

Strains carrying variants of the ompT chromosomal copy (strains MG1281, MG1307 and MG1309) derive from a strain where a cat-sacB cassette was recombined in the ompT gene as previously described. An oligonucleotide carrying the Flag tag sequence or PCR products of a fragment of ompT DNA with the mut2 change either by itself or in combination with the Flag tag were then used to replace the cassette by selecting recombinants able to grow on sucrose plates. Candidates were checked for chloramphenicol sensitivity and the full ompT gene was sequenced. The mal::lacIq allele was then introduced in two steps. First, a crp::cat mutation was P1 transduced into these strains. Then, after P1 transduction of mal::lacIq, candidates were selected as clones that grew on minimal medium + 0.2% sorbitol, but not on minimal medium + 0.2% maltose.

Finally, the strain carrying a pBAD-ompR-lacZ fusion (or pBAD-ompRmut2-lacZ) was made by replacing the cat-sacB cassette of a pBAD-cat-sacB construct that was inserted upstream of lacZ gene in strain PM1205 (from P. Mandin, unpublished) by a DNA fragment carrying nts –123 to +30 of ompR (or ompRmut2).

Plasmids

Plasmids pBRplacOmrA, pBRplacOmrB, pBAD-OmrA and pBAD-OmrB have been described previously (20). mut2* and mut3* derivatives of pBRplacOmrA and pBRplacOmrB plasmids were made using the Quick-change II XL mutagenesis kit (Stratagene). After sequencing, the AatII-EcoRI fragment was subcloned into pBRplac to get rid of unwanted secondary mutations.

RNA extraction and northern blot analysis

RNA was extracted from cells in exponential phase as previously described (20). Typically, cells were grown in LB + ampicillin to an OD600 of about 0.3, and synthesis of OmrA/B was induced using 100 μM IPTG (Figures 1A and B, B,3B,3B, B,6C6C and and7B)7B) or 0.2% arabinose (Figure 1C). RNA was harvested at the indicated time-points. For Figure 5, RNA was extracted from strain MG1189 transformed with different plasmids and grown in LB medium supplemented with ampicillin and 100 μM IPTG at the same time that samples were taken to assay the β-galactosidase activity, i.e. at an OD600 of about 0.4. After extraction, RNA was ethanol-precipitated, resuspended in DEPC-water and its concentration was estimated by measuring the OD at 260 nm.

Figure 1.
Degradation of ompT and cirA mRNA after OmrA/B induction is dependent on the RNAse E endonuclease, but not on RNAse III. RNA levels of OmrA/B as well as of ompT and cirA mRNA were followed by northern blot analysis at 0, 4 and 10 min after induction of ...
Figure 3.
OmrA and OmrB control ompT and cirA expression in an Hfq-dependent manner. (A) Specific β-galactosidase activity of the cirA-lacZ or ompT-lacZ fusion strains transformed with an OmrA- or OmrB-overexpressing plasmid or the corresponding control ...
Figure 5.
The 5′ end of OmrA/B interacts with the translation initiation region of cirA mRNA in vivo. Compensatory changes were introduced in both OmrA/B and/or the cirA-lacZ translational fusion (A). As in Figure 4, their effect on the control was assayed ...
Figure 6.
The 5' end of OmrA/B base-pairs with ompT mRNA in vivo. (A) Compensatory changes depicted in (A) were introduced in OmrA/B, ompT and ompT-Flag RNAs. Nucleotides in red are conserved between OmrA and OmrB and the ones in gray correspond to the ompT start ...
Figure 7.
The 5′ end of OmrA/B is involved in the regulation of multiple targets, including ompR. (A) Predicted base-pairing between ompR mRNA and OmrA/B. Nucleotides in red are conserved between OmrA and OmrB and the ones in gray correspond to the ompR ...

Northern analysis was then performed as follows. When the level of OmrA, OmrB or SsrA was analyzed, 3 μg total RNA were separated on a 5 or 10% acrylamide TBE-urea gel (Biorad) and then transferred onto a nylon membrane as previously described (20). To analyze the level of other RNAs, a constant amount of total RNA (between 6.5 and 15 μg) was separated onto a 1% agarose gel in MOPS buffer and transferred to a nylon membrane by capillary action. Detection was carried out using biotinylated probes and the Brightstar Biodetect kit (Ambion). SsrA and ompA RNAs were used as negative controls. Sequences of probes specific to OmrA, OmrB, SsrA, ompT, cirA, fecA, fepA or ompA were previously published (20). The sequence of the ompR-probe is 5′Biotin-cacgagtcagcaggcgatccatctgttctgcattagcgacgc3′. In Figures 5–7, wt and mutant OmrA or OmrB were detected with OmrAmut-probe (5′Biotin-caggttggtgcaagagacagggtacgaagagcgtaccg3′) or OmrBmut-probe (5′Biotin-cgcaggctggtgtaattcatgtgctcaacccgaagttga3′), respectively. When needed, membranes were stripped by boiling in SDS 0.5% for 10 min and could then be hybridized with another probe. Up to three different probes were used per membrane.

β-Galactosidase assays

Cells were grown in LB medium supplemented with ampicillin and 100 μM IPTG. Duplicate samples of 0.5 ml were removed in exponential phase (OD600 about 0.4), mixed with 0.5 ml of cold Z-buffer and their β-galactosidase activity was assayed as described (24). Results are the average of at least two independent experiments.

Western blot analysis

Cells were grown in LB medium containing ampicillin and IPTG at a final concentration of 100 μM. When the OD600 reached 0.4, proteins were precipitated with 5% TCA. The pellet was then washed with 80% acetone and resuspended in SDS-sample buffer (New England Biolabs), so that the concentration would correspond to 15 OD600/ml. For detection of the OmpT-Flag protein, 10 μl of these samples were then subject to electrophoresis on 10% Bis–Tris NuPAGE Gel (Invitrogen) in MOPS buffer, transferred to a nitrocellulose membrane and flagged proteins were detected using the anti-Flag M2 monoclonal antibody-alkaline phosphatase conjugate (Sigma) following manufacturer's instructions, combined with the Lumi-Phos chemiluminescent substrate (Pierce). For the detection of EFTu, 7.5 μl of samples diluted 100-fold were separated and transferred as above, and then probed using a 1:3000 dilution of anti-EFTu antiserum and the ECL detection kit (Amersham).

RESULTS

Degradation of ompT and cirA mRNAs in response to OmrA/B production requires the RNase E endonuclease

Since several targets of Hfq-binding sRNAs have been shown to be degraded by RNase E (14,25,26), this essential endonuclease was a good candidate for degrading OmrA/B targets. This was tested by looking at the effect of two different mutations of the enzyme on ompT and cirA mRNA levels after induction of OmrA or OmrB expression (Figure 1). The first mutant, rne-3071, is a thermosensitive allele carrying a single amino acid change in the N-terminal domain of the protein that affects the catalytic activity of the protein; the protein becomes inactive at elevated temperature and the strain is unable to grow (27). The second mutant, rne-131, carries a two nucleotide deletion resulting in a frameshift and the appearance of a premature stop codon (28). The resulting truncated protein lacks the domain involved in the interaction with other components of the degradosome; the C-terminal domain is not essential and the strain carrying this truncation can be grown at most temperatures (29).

Cells carrying either rne mutation and ectopically expressing OmrA or OmrB from an IPTG-inducible promoter were grown to exponential phase (OD600 > 0.3), transferred to nonpermissive temperature to inactivate the RNase E if needed (for the rne-3071 mutant, panel A) and the level of ompT and cirA mRNAs was followed by northern blot after induction of OmrA/B expression. In wild-type (wt) cells, at both 37°C and 43.5°C, degradation of ompT mRNA is visible as soon as 4 min after the induction of OmrA/B expression and most of the ompT message is gone by 10 min (Figure 1, top of panels A and B). Similar results were obtained with cirA mRNA; it was no longer detectable after 4 min of OmrA/B induction, while the control mRNA for ompA was unaffected. In the rne-3071 mutant, both ompT and cirA mRNAs were stabilized; no degradation was observed during the 10 min that followed induction (Figure 1A, bottom panel). RNase E endonucleolytic activity is therefore required for the degradation of at least these two target mRNAs of OmrA/B. Degradation of these targets was also decreased in the rne-131 mutant (Figure 1B), suggesting that the degradosome may play a role as well. In these experiments, accumulation of OmrA/B was unaffected by rne-3071 or rne-131 alleles and control of both ompT and cirA by these sRNAs was similar at 37°C and 43.5°C.

The requirement for endoribonuclease RNase III, that belongs to the same family as the eukaryotic Dicer endonuclease (30), was investigated as well, using a null allele of the rnc gene, encoding RNase III. Because we found that in the rnc strain used (strain MG1272), both OmrA and OmrB were strongly expressed from pBRplac derivatives, even without induction (data not shown), the OmrA/B RNAs were ectopically expressed from a PBAD promoter using arabinose in this set of experiments (Figure 1C). Under these conditions, the decrease in ompT mRNA is not as striking as with expression from a Plac promoter, although the decrease in cirA is complete (compare Figure 1B and C, top panels). The decrease in ompT and cirA mRNA levels after OmrA/B induction is at least as marked in both rnc+ and rnc cells (Figure 1C), showing that RNase III is not required in the degradation of these targets of OmrA/B.

These results are consistent with findings for other Hfq-dependent sRNAs in E. coli, in which RNase E but not RNase III was found to be primarily responsible for mRNA degradation (reference 14 for instance).

Repression of ompT or cirA translational fusions by OmrA/B differs greatly in extent

For many Hfq-dependent sRNAs, pairing with negatively regulated mRNAs overlaps or is near the RBS and translation initiation codon. Initial pairing predictions for OmrA and B and these regions yielded a short possible pairing for OmrA/B and cirA, overlapping the RBS (Figure 2B). Nucleotides 2 to 24 of OmrA or OmrB, most of which are conserved between these two sRNAs (Figure 2A), are predicted to be involved in this interaction, in two stretches, one of 8 nt and the other of 12 nt. They would imperfectly base-pair with the nucleotides –35 to –10 of cirA mRNA, i.e. in the vicinity of the cirA TIR, which suggests that OmrA/B could inhibit ribosome binding to cirA message. For ompT, the longest stretch of predicted pairing is downstream from the start codon, at nt 12–20 of the translated region (Figure 2C); once again, this pairing is with the 5′ end of OmrA and OmrB.

Figure 2.
Predicted base-pairing between OmrA/B and ompT or cirA mRNA. (A) Sequences of E. coli OmrA and OmrB sRNAs. (B) Prediction of pairing with cirA mRNA. (C) Prediction of pairing with ompT mRNA. Nucleotides in red are conserved between OmrA and OmrB. Start ...

In order to examine whether these regions of the targets were necessary and sufficient for OmrA/B action, we constructed translational fusions between either ompT or cirA, and the lacZ reporter gene. Both CirA and OmpT are outer membrane proteins; in order to use β-galactosidase fusions to measure their expression, we designed fusions that should not be exported from the cytoplasm. More precisely, the DNA region from 220 nt upstream of the ATG start codon and 30 nt downstream (i.e. encoding 10 amino acids, which is only part of the signal sequence) was fused in frame with an eight-amino-acid FLAG tag and the lacZ gene. These constructs were then introduced in single copy at the lambda attachment site of a Δlac strain.

The effect of OmrA/B long-term expression on the expression of these fusions was then assayed by measuring the β-galactosidase activity associated with either fusion after transformation with a plasmid overexpressing either sRNA. Because cirA expression is Fur-repressed at the transcriptional level, activity of the cirA-lacZ fusion was measured in a Δfur::cm background.

As shown in Figure 3A and Table S1, OmrA and OmrB reduced the activity of cirA-lacZ 10.6- and 5.5-fold, respectively under these conditions. Both OmrA and OmrB bind tightly to Hfq (31,32), and therefore were expected to depend on Hfq for function. The requirement for Hfq for control by OmrA/B was assayed on cirA-lacZ expression. In an hfq mutant strain, control of cirA-lacZ by OmrA/B was fully abolished (Figure 3A). This result confirms that the cirA-lacZ fusion mimics the major expectations for OmrA and OmrB regulation of cirA, i.e. strong negative regulation and dependence upon Hfq. Therefore, it seems likely that the fusion contains the necessary elements for Omr regulation.

In contrast, the effect of OmrA/B on the expression of ompT-lacZ was extremely weak (Figure 3A and Table S1). This result suggested that elements within the ompT transcript that are important for the full regulation by OmrA and OmrB are not present in this fusion, and therefore the regulation of ompT may differ somewhat from that of cirA. Because of this modest effect of OmrA/B on the translational fusion, the requirement for Hfq was analyzed by a direct examination of the levels of ompT mRNA after OmrA/B induction in an hfq mutant (Figure 3B). In contrast to a wt strain, no degradation of ompT mRNA was visible in this mutant for as long as 10 min after induction, clearly showing that control of ompT by OmrA/B is Hfq-dependent as well.

In the absence of Hfq, both OmrA and OmrB fail to accumulate as much as in wt cells (Figure 3B), which explains, at least in part, the need for Hfq in the control of their targets. Lower levels of Hfq-dependent RNAs are commonly seen in hfq mutants, due to rapid turnover of these sRNAs (14). However, in vitro experiments clearly demonstrate direct roles for Hfq in pairing of sRNAs and their targets (33). It seems likely that Hfq is important in both capacities for OmrA and OmrB in vivo action.

Role of the 5′ leader of cirA mRNA for control by OmrA/B

Many targets of sRNAs have relatively long 5′ untranslated regions (UTRs) that participate in the regulation, for instance by adopting a structure that can promote the interaction with the regulatory sRNA (34). At first sight, cirA could fall into the category of these targets, since it carries a long UTR. Indeed, two promoters of cirA have been identified: p1, leading to transcription of a 173-nt untranslated leader, and p2, leading to transcription of a 160-nt untranslated leader (35). Both are present in the cirA-lacZ translational fusion used in this study. Two deletions were constructed to examine the role of this UTR in regulation by OmrA and OmrB. Deletion Δ1 extends from nts –149 to –54, deleting 96 nt; deletion Δ2 extends from nts –149 to –21, deleting 129 nt (Figure 4A). Activity of the cirAΔ1-lacZ fusion was similar to the full-length fusion, and was subject to OmrA and OmrB-dependent repression (Figure 4B). Although the degree of repression was less than that observed with the full-length leader, these results show that OmrA and OmrB do not require cirA sequences beyond those from –54 to +30 for regulation.

Figure 4.
Part of cirA leader is dispensable for post-transcriptional control by OmrA/B. Different mutations, depicted in (A), were introduced in the cirA-lacZ translational fusion. Their effect on the control by OmrA/B was then assayed by measuring the β-galactosidase ...

The Δ2 deletion increased the basal activity (in the absence of OmrA and OmrB) by 1.75-fold, and control by OmrA and OmrB was almost fully abolished (Figure 4B, Table S1), suggesting that at least some of the region between nts –54 to –21 of the cirA leader is crucial for OmrA/B action.

These results for the Δ1 and Δ2 fusions are consistent with the predicted pairing since Δ1 does not enter the predicted pairing region whereas Δ2 does (Figure 4A). Several mutations were then introduced into the cirA-lacZ fusion to further examine whether pairing with the TIR of cirA is required for OmrA/B action.

In the predicted pairing (Figure 4A), the longest stretch of continuous pairing, arbitrarily referred to as pairing region I, is an 8- or 9-nt region that extends from nts −35 to −28/−27 of cirA, pairing with a region that is not completely conserved between OmrA and OmrB. The second stretch of pairing involves 12 nt of OmrA and B, starting very close to the 5′ C of these sRNAs, and three regions of cirA, of 3 nt, 6 nt and 3 nt respectively, separated in each case by a single unpaired base (Figure 4A). These pairing regions were arbitrarily referred to as IIa, IIb and IIc, respectively. Changes in region I or IIa of cirA that are expected to disrupt pairing in these regions (mutI and mutIIa respectively, Figure 4A) did not disrupt regulation by OmrA/B (Figure 4B, Table S1). In addition, combining these mutations (mutI+IIa) gave regulation very similar to that seen in the wt case (Figure 4B). Therefore, pairing in regions I and IIa is not required for control by OmrA/B in our experimental conditions, which leaves only the pairing of 9 nt in regions IIb and IIc. Consistent with a crucial role for these regions, a mutation expected to disrupt the three last base-pairs of region IIb (mutIIb), almost fully abolished control by OmrA or OmrB (Figure 4B, Table S1).

OmrA and OmrB base-pair with cirA mRNA in vivo

To directly test whether the predicted cirA-OmrA/B pairing in regions IIb and IIc was critical for regulation, effects of mutations and compensatory mutations in the sRNAs and the cirA-lacZ fusion were tested in vivo. Mutations in cirA-lacZ that changed blocks of 3–5 nt were made as shown in Figure 5; the compensating mutations, indicated with a asterisk (*), were constructed in both OmrA and OmrB and gave wt levels of RNA (Figure 5C).

In the cirA-lacZ mutant referred to as mut2, the nine predicted base-pairs in regions IIb and IIc are disrupted by a 3-nt change (Figure 5A). Probably because the first G of the cirA Shine–Dalgarno (SD) sequence is changed into a C with this mutation, the activity of the cirAmut2-lacZ fusion was extremely low (<2 Miller units when transformed with the empty vector, data not shown). This rendered it virtually impossible to measure the effect of OmrA/B on the expression of this fusion. Therefore, we took advantage of an up-mutant that we constructed while in the process of characterizing the cirA untranslated region. This mutant, referred to as ‘up’, consists in a change of 4 out of the 6 nt between the SD sequence and the AUG start codon (Figure 5A). The activity of the resulting cirAup-lacZ fusion is more than 8-fold higher than the activity of the wt fusion (Figure 5B, compare scale on two panels). In addition, consistent with the fact that the ‘up’ change affects nucleotides that are not predicted to base-pair with the sRNAs, OmrA and OmrB were found to efficiently repress the activity of the cirAup-lacZ fusion (Figure 5B, Table S1). Although the molecular mechanism underlying the increase in expression with the ‘up’ mutation is still unknown, we combined this change with other changes in the pairing region in order to study the effects of compensatory mutations in OmrA/B.

Two sets of mutations in the 5′ end of OmrA or OmrB were tested. Mutating positions 3–6 of the 5′ end of OmrA/B to create mut2* (Figure 5A) drastically reduced the ability of the Omr mutant RNAs to regulate the cirAup-lacZ fusion (Figure 5B, right panel), consistent with required pairing of the 5′ end of OmrA and OmrB. Mutating the 5 nt between positions 6 and 10 (Figure 5A, mut3*) also greatly reduced regulation of the ‘up’ fusion. Both sets of mutations also reduced regulation of the wt fusion (Figure 5B, left panel), consistent with the ‘up’ mutation not changing the basic-pairing properties of the cirA-lacZ fusion. Complementary mutations in the pairing regions of cirA were tested, in the context of the ‘up’ version of the cirA-lacZ fusion. While the up/2 double mutant had relatively low expression in the absence of the sRNAs, there was very little effect of expression of wt OmrA or OmrB (1.5- and 1.3-fold decrease, respectively). In contrast, OmrAmut2* and OmrBmut2* repressed the expression of the cirAup/2-lacZ fusion significantly (repression factors are 3.1 and 2.7 for OmrAmut2* and OmrBmut2*, respectively), supporting the predicted base pairing (Figure 5B, Table S1).

This was further confirmed by using the mut3/mut3* set of mutants. As previously, mut3 was combined with the up change in cirA-lacZ and the effect of OmrA/B was analyzed on the resulting cirAup/3-lacZ fusion as well as on the cirAup-lacZ. Wt OmrA and OmrB repressed the expression of the up fusion much better than the up/3 fusion; repression of 27.9-fold for OmrA regulation of the ‘up’ fusion was reduced to 2.3-fold for the up/3 fusion; for OmrB, repression of the up fusion was 11.5-fold and was reduced to 1.4 for the up/3 fusion (Figure 5B, right panel). When the mut3* compensating changes were present in the Omr sRNAs, regulation of the up/3 fusion was restored: OmrA/Bmut3* repressed the up fusion poorly (about 1.5 fold) but the up/3 fusion well (about 6-fold) (Figure 5B, Table S1).

All together, data presented in Figures 4 and and55 clearly show that the 5′ terminal region of OmrA/B interacts in vivo with the TIR of cirA mRNA and that a relatively short pairing region of no more than 9 nt is sufficient for the regulation, at least when OmrA/B are overexpressed. It is worth noting that none of these mutations fully abolished regulation under these growth conditions, suggesting that the remaining even shorter pairing can provide some modest degree of regulation.

OmrA and OmrB base-pair with ompT mRNA in vivo

As for cirA, the importance of RNase E and Hfq in the control of ompT expression (see above) suggested a base-pairing with OmrA/B and such a base-pairing can be predicted in the vicinity of the ompT TIR (Figure 6A). Interestingly, the region of OmrA/B that base-pairs with cirA (i.e. nts 2 to 10 of the sRNAs) is also predicted to be involved in the interaction with ompT, since nts 1 to 33 of OmrA or 1 to 32 of OmrB are expected to form an imperfect duplex with nts –15/–14 to +20 of ompT mRNA. Surprisingly, most of the region of ompT mRNA involved in this base-pairing is part of the coding sequence rather than of the 5′ UTR.

In order to determine whether this interaction occurs in vivo, the effect on the control of ompT expression of mutations in the longest stretch of predicted base pairs (the 9 bp between nts 1 and 9 of OmrA/B and nts 12 to 20 of ompT mRNA) was analyzed. Mutants in OmrA/B used in this set of experiments are OmrA/Bmut2*, also used previously, while the mutant in ompT mRNA, referred to as ompTmut2, consists of the compensatory change (Figure 6A).

Results obtained using these compensatory mutations and the previously described ompT-lacZ translational fusion as a reporter for ompT expression support the pairing between the 5′ end of OmrA/B and the early coding region of ompT mRNA (Table S1). However, given the very modest effect seen with the ompT-lacZ reporter, a number of other approaches were used to examine the importance of the predicted base-pairing. In both of the approaches discussed below, a full-length ompT gene is present and therefore, if sequences downstream of the ompT-lacZ fusion joint are critical for regulation, these should be present.

An ompT-Flag allele was used to further confirm this direct interaction in vivo. In this OmpT-Flag construct, an eight-amino-acid Flag tag (sequence DYKDDDDK) was introduced between amino acids S110 and N111 of OmpT (the numbering refers to the full-length protein, i.e. including the signal sequence). Based on the structure of OmpT (36), the Flag tag is expected to be in the second extracellular loop. As shown by western analysis using an anti-Flag antibody, the amount of OmpT-Flag clearly decreases in the presence of the plasmids overproducing OmrA or OmrB compared to the empty vector, but the mut2* forms of OmrA/B are less effective (Figure 6B, lanes 1–5). In this regard, this OmpT-Flag is a better reporter than the OmpT-β-galactosidase fusion for the action of OmrA/B, consistent with the idea that regions outside of those present in the fusion may contribute to regulation. Consistent with the predicted pairing, OmrAmut2* and OmrBmut2*, but not the wt OmrA and OmrB, decreased the cellular level of the OmpTmut2-Flag protein (Figure 6B, lanes 6–10) made from an ompTmut2-Flag allele where the compensatory mutation was introduced. This clearly shows that OmrA/B base-pair with ompT mRNA in vivo in the beginning of the coding region. Therefore, the first 10 nt of both OmrA and OmrB mediate their interaction with at least two direct targets, ompT and cirA.

Finally, the ability of OmrA/B or OmrA/Bmut2* to induce the degradation of ompT or ompTmut2 mRNAs, as well as their flagged derivatives, was investigated. As previously shown, wt OmrA and OmrB induce degradation of ompT mRNA as soon as 10 min after induction. Surprisingly however, the level of ompT-Flag or ompTmut2-Flag mRNAs were not affected by the induction of wt or mut2* OmrA/B, respectively (Figure 6C), even though the synthesis of the corresponding proteins are controlled by these sRNAs (Figure 6B). This strongly suggests that mRNA degradation is not required for translational control of ompT by Omr sRNAs. Finally, ompTmut2 message was not degraded after a 10-min induction of mut2* forms of OmrA/B (Figure 6C) and similar results were obtained after a 30-min induction of the sRNAs (data not shown). A 4-nt change in the ompT mRNA is thus sufficient to change its sensitivity to Omr-promoted degradation.

The 5′ end of OmrA/B is also involved in the regulation of other targets of these sRNAs

Results presented so far show the crucial role of the conserved 5′ end of OmrA/B in the base-pairing with ompT and cirA targets. Consistent with these results, mutants in the 5′ end of these sRNAs, OmrAmut2* or OmrBmut2*, failed to induce the decay of ompT and cirA mRNAs seen with the wt sRNAs (Figures 6C and and7B).7B). We have previously shown that OmrA and OmrB have other targets, including two additional genes for outer membrane proteins, fecA and fepA. The effect of the mut2* mutation in OmrA/B on the mRNA level of these other targets was investigated as well. Figure 7B clearly shows that, as for ompT and cirA, mRNA levels of fecA and fepA are decreased after induction of the wt forms of OmrA/B (see also reference 20), but not after induction of their mut2* counterparts. Therefore, the 5′ end of OmrA/B plays a major role in the regulation of at least four targets. In the case of ompT and cirA, this is due to a direct base-pairing between the target-mRNA and OmrA/B 5′ end. The same may be true for fecA and fepA as well, but we cannot exclude indirect regulation of these targets by OmrA/B, in which case the mut2* mutation could affect the level of an intermediate regulator instead.

Both OmrA and OmrB directly regulate the expression of their transcriptional activator, ompR

Given these results, we used the program TargetRNA (37) to search for potential pairing between the first 10 nt of OmrA/B and the TIR of E. coli mRNAs, with the idea that additional potential targets might be found. The search was done using the following criteria: focus around nts –30 to +20 of mRNA relative to the start codon, a hybridization seed of 6 bp allowing G:U base-pairs, and a P-value threshold of 0.3. With these fairly permissive criteria, ompT was ranked 5 (P-value = 0.0338288) and cirA was ranked 48 (P-value = 0.137068), whereas fecA and fepA were not present among the first 100 predicted targets. Interestingly, the first predicted target is yjcH (P-value = 0.00800168), and mRNA levels for both genes of the yjcH-yjcG operon were strongly decreased after induction of OmrA or OmrB in microarrays experiments (20). Similarly, ranked at the eighth position with a P-value of 0.0338288 is ompR, encoding the transcriptional regulator of several genes including omrA and omrB. ompR is the first gene of the ompR-envZ bicistronic operon and both ompR and envZ mRNA levels were found to be reproducibly decreased by 2-fold after OmrA or OmrB induction in E. coli (20). The same was true for ompR in Salmonella (Papenfort,K. and Vogel,J., personal communication). Altogether, these results suggest that OmrA/B could base-pair with the leader of yjcH or ompR and thereby regulate the expression of the corresponding operons. Besides ompT, yjcH and ompR, the other genes among the 10 first predicted targets were either poorly expressed or their expression was unchanged in the presence or absence of OmrA or B as assayed by microarray experiments (20).

The possibility that ompR could be a direct target of OmrA/B was further reinforced when we investigated the effect of ectopically overproducing wt and mut2* OmrA/B on ompR mRNA levels by northern blot. When we used a 42-nt probe complementary to nts 98 to 137 of ompR mRNA, two distinct mRNA species were detected (Figure 7B). The larger band migrates between 2.0 and 2.5 kb (data not shown), which is consistent with the size of the ompR-envZ bicistronic mRNA. The smaller band migrates above 1.0 kb and could correspond to an alternative transcription product, a processing product or could be due to a nonspecific hybridization, but this was not investigated further. As shown in Figure 7B, induction of wt OmrA or OmrB, but not of OmrAmut2* or OmrBmut2*, decreased the level of ompR-envZ mRNA, while the amount of the smaller transcript was not affected. This result shows that OmrA/B negatively controls the expression of the ompR-envZ operon, and that, as for other targets of these sRNAs, the 5′ end of OmrA/B is involved in this control.

Given these results, it seems likely that the predicted pairing mentioned above between the ompR 5′ UTR and OmrA/B (shown in Figure 7A) exists in vivo. In order to test this, a translational ompR-lacZ fusion was constructed. Transcription of this fusion is driven by the PBAD promoter and the transcription start site is expected to be the same as that for the ompR promoter p1 (38). The 5′ UTR should therefore be 123-nt long. lacZ is fused in frame after the 30th nucleotide of the ompR coding region (i.e. after the 10th amino acid). The activity of the resulting fusion was decreased by 3- or 2.5-fold, respectively when OmrA or OmrB was ectopically overexpressed (Figure 7C, Table S1), showing that the region between nts –123 and +30 of ompR mRNA is sufficient for some control by OmrA/B. Mutants in the 5′ end of OmrA/B such as OmrA/Bmut2* were unable to control the expression of this fusion. However, OmrA/Bmut2* downregulated by 4.2-fold the expression of an ompRmut2-lacZ fusion bearing the compensatory changes; wt OmrA/B had no effect on this fusion (Figure 7C, Table S1). These data show that OmrA/B post-transcriptionally regulate the expression of ompR by base-pairing with its 5′ UTR in vivo.

DISCUSSION

OmrA and OmrB, two adjacent and partially homologous Hfq-binding sRNAs, are positively regulated by the EnvZ/OmpR two-component system and negatively regulate a set of outer membrane proteins. Here, we have examined the nature of the interaction of OmrA and OmrB with their targets. As for other Hfq-binding RNAs, we find that the action of OmrA/B depends on Hfq and on RNase E. We also find that the conserved 5′ end of OmrA and OmrB is required for interaction with and/or regulation of all of the tested targets. Interestingly, a similar pattern of interaction has recently been observed for RybB sRNA (39; Mika,F. and Vogel,J., in preparation). Finally, we find that one direct target of OmrA and OmrB is the mRNA encoding the transcriptional regulators of these sRNAs, EnvZ and OmpR, thus providing a feedback loop in which activation of transcription will downregulate expression of the activators.

Common pairing interactions with OmrA and OmrB for multiple targets

We found that interaction between the 5′ end of OmrA/B and nts –10 to –19 of cirA mRNA is sufficient for the control of cirA. This pairing region in cirA is conserved in other enterobacteria where OmrA/B are found, such as Salmonella, Yersinia and Klebsiella species (Figure S1), suggesting that cirA expression is subject to post-transcriptional control by OmrA/B in these species as well. Pairing with cirA seems to follow the general scheme of negatively acting sRNAs: it involves the SD sequence and most likely regulates cirA expression by competing with ribosome binding. Consistent with that, the activity of a translational cirA-lacZ fusion carrying the region of pairing was strongly decreased upon overexpression of OmrA/B.

The 5′ end of E. coli OmrA/B also base-paired with nts 12 to 20 of ompT ORF. Even though this pairing occurs downstream of the RBS, it is probably close enough to inhibit ribosome binding. Indeed, the interaction between Salmonella RybB sRNA and the early coding sequence of ompN (downstream of the AUG) is sufficient to inhibit translation initiation of this target (39). The ompT gene is part of the cryptic lambdoid DLP12 prophage and no close homolog of this gene was found in Klebsiella pneumoniae or Salmonella genomes (40). Nevertheless, members of the family of outer membrane proteases defined by OmpT, the Omptin family, are encoded by Yersinia (pla gene) or Salmonella (pgtE gene), but the OmrA/B-ompT pairing is not conserved with these genes (Figure S1). Therefore, the synthesis of Yersinia Pla and Salmonella PgtE may not be directly controlled by OmrA/B, even though we cannot exclude an unidentified alternative pairing in these cases.

Pairing with the same 5′ region of OmrA and OmrB is also required for regulation of envZ/ompR; in this case, pairing is again upstream of the AUG and partially overlapping the RBS. As for cirA, this pairing is conserved in other enterobacteria (Figure S1). Finally, the same 4-nt mutation (mut2*) that disrupts pairing with ompR, ompT, and cirA also abolishes regulation of fepA and fecA, suggesting that all OmrA and OmrB targets depend on this 5′ region.

Pairing of the same region of an sRNA to multiple targets may be a general rule; for instance, a conserved region of GcvB sRNA is involved in its interaction with as many as seven different targets (41). In the case of OmrA/B, the participation of the 5′ end of the sRNA may provide increased flexibility or allow easier accessibility to mRNAs. Pairing to multiple targets through the 5′ end of a sRNA has also been observed with Salmonella RybB sRNA, that interacts with several OMP-encoding mRNAs (Mika,F. et al., submitted). Interestingly, MicC and MicF, that respectively inhibit synthesis of OmpC and OmpF porins, also paired with the corresponding target-mRNAs through their 5′ end (42,43). This is reminiscent of the 5′ end seed used for pairing by eukaryotic miRNAs (44).

The conservation of this 5′ end region between OmrA and OmrB is consistent with the fact that both sRNAs can regulate all targets identified so far. Such redundancy in regulation of shared targets is not unusual for duplicated sRNAs (45,46), but does raise the question of why these small RNAs are duplicated in some species, but not others such as Yersinia pestis. It is also worth noting that, in addition to conservation of the 5′ end (which extends well beyond the pairing region, to nt 21), the last 35 nt of OmrA/B are conserved as well. Although some of these 3′ end nts are part of the transcriptional terminator, this conservation pattern may suggest that additional targets of OmrA/B remain to be identified that pair with the 3′ conserved regions, or that this region plays other roles in sRNA function.

Other components of cirA regulation

cirA mRNA has a rather long leader [173- or 160-nt long if transcription is driven by p1 or p2, respectively (31)]. Even though pairing of OmrA/B with nts –10 to –19 of cirA leader is sufficient for the control in our experimental conditions, we found that deleting nts –20 to –149 (i.e. outside of the sufficient pairing region) almost fully abolished this regulation (Figure 4, deletion Δ2). Some of the leader may therefore ensure an optimal conformation for the pairing region to interact with the sRNAs.

On the other hand, it is clear that most of this leader is not necessary for control by OmrA/B, given that a deletion of nts –54 to –149 of the leader still gives good regulation (Figure 4, deletion Δ1). It is therefore conceivable that, in addition to the effects of OmrA/B, other post-transcriptional controls regulate the expression of cirA via this leader region.

Yet another level of control is suggested by the surprising effect of the ‘up’ mutation (Figure 5): changing 4 out of the 6 nt between the SD sequence and the start codon of cirA increased the activity of the cirA-lacZ fusion by more than 8-fold. Preliminary data indicate that changing either the G or A just upstream of the AUG into an U is sufficient to increase the expression of cirA-lacZ by about 3-fold (our unpublished data). It seems therefore that the increase in expression could be due to the disruption of the GGAA sequence between nts –3 and +1, which resembles an SD sequence and may be used to initiate translation from an alternative start codon. Interestingly, this GGAA sequence is conserved in Salmonella and Klebsiella, and, in Yersinia, there is an AGAG sequence instead, that could also be considered as a potential SD sequence (Figure S1). Whether this model is true and whether it is a process regulating, rather than simply decreasing, cirA expression remains to be investigated.

Complexity of ompT regulation and mRNA turnover

ompT regulation by OmrA and OmrB is easily demonstrated at the mRNA level for the wt ompT message; mRNA disappears rapidly upon induction of the sRNA (Figures 1, ,33 and and6).6). However, in experiments done with an ompT-lacZ fusion carrying the 5′ UTR and the first 30 nt of ompT, regulation by OmrA/B overproduction is poor (Figure 3A). Since this suggested the possibility that elements beyond nt +30 were required, control was analyzed using the chromosomal copy of ompT, either in its wt or Flag-tagged form. In contrast to the ompT-lacZ fusion, ompT mRNA downstream of nt +30 is present in these constructs, but may be modified by the insertion of the 24-nt Flag sequence. The accumulation of the Flagged proteins is clearly controlled by OmrA/B, both in the wt situation as well as when compensatory changes were introduced in ompT-flag mRNA and in OmrA/B (Figure 6B). However, compensatory changes did not restore the ability of OmrA/B to induce the degradation of ompTmut2 mRNA or ompTmut2-Flag mRNA (Figure 6C). In general, degradation of the ompT mRNA in response to the Omr RNAs was perturbed by every change we made. For instance, the ompT-Flag mRNA (not carrying a mutation) is poorly sensitive to degradation by wt sRNAs (Figure 6C).

The mut2 mutation, in which the fifth and sixth codons of ompT are changed from CUUCUG (Leu-Leu) to CUGUUC (Leu-Phe), also decreases the level of OmpT-Flag protein (Figure 6B, compare the two vector lanes). This is independent of OmrA/B, since the experiment was done in a ΔomrAB strain. This sequence is part of the signal sequence and it is possible that the mutation affects the efficiency of export of the resulting protein to the outer membrane. Thus, it seems likely that mRNA structure, stability, and translation efficiency are perturbed by the addition of a Flag tag and by the mut2 change, hinting at further complexity in the regulation of ompT.

It was shown previously that the control of the IICBGlc inner membrane protein synthesis by SgrS sRNA required the membrane localization of the target-mRNA, presumably because this decreased ribosome loading on the mRNA, allowing the sRNA to successfully compete with the ribosome (47). Even though the localization process is somewhat different for outer membrane proteins such as OmpT, one can hypothesize that, in a similar way, the translation of ompT may be affected by sequences within the signal sequence in ways that perturb OmrA/B action. This could explain, at least in part, why OmrA/B only weakly decreases the expression of the ompT-lacZ fusion.

Autoregulation by OmrA and OmrB

The expression of several sRNAs is subject to feedback control. For instance, RybB sRNA, whose transcription is regulated by the alternative sigma factor sigmaE, indirectly downregulates the expression of sigmaE by decreasing the synthesis of several OMPs (48,49). Similarly, production of the Fur-regulated RyhB sRNA represses the synthesis of iron-binding proteins, thereby increasing the cellular level of free iron, which will in turn repress the expression of the Fur regulon, which includes RyhB itself (50). Another example comes from Vibrio cholerae, where the four redundant Qrr sRNAs inhibit expression of the HapR regulator, which indirectly represses synthesis of the Qrr (51). In our case, OmrA/B sRNAs negatively autoregulate their transcription by directly repressing the ompR-envZ operon that encodes their transcriptional activators. This feedback loop should limit the production of OmrA/B after induction of the EnvZ-OmpR two-component system. However, it was found that there was almost no change in the transcription of two members of the EnvZ-OmpR regulon, ompC and ompF, over a wide range of EnvZ and OmpR levels (52), and the physiological relevance of OmrA/B autoregulation remains therefore to be investigated.

Length of pairing regions

Our results show that relatively short regions of complementarity between OmrA/B and their targets are sufficient for translational control by these sRNAs in our experimental conditions. It is especially striking for cirA, where mutagenesis of the leader indicates that two stretches of 6 and 3 bp with the sRNA are sufficient for control. sRNA–mRNA interactions relying on very short regions of complementarity are not limited to the Omr sRNAs, since Aiba and co-workers (33) showed that, when mutagenizing the duplex between SgrS sRNA and ptsG mRNA, only 6 bp affected control when point mutants were introduced. One drawback of both our and Aiba's group results is that the effect of mutations in either the sRNA or the target-mRNA was analyzed under conditions where the sRNA was overproduced, which could compensate for weaker interactions. Therefore, we cannot exclude that additional base-pairs contribute to the interaction with the targets when sRNAs are expressed at lower levels from the chromosome. In the future, it might be worthwhile to undertake the same kind of study after inducing the sRNA synthesis under conditions that would be closer to the physiological levels. It would also be interesting to determine whether having the pairing region at the very 5′ end of OmrA/B allows control through shorter base-pairing interactions.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

The Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. Funding for open access charge: Intramural Research program of the NIH.

Conflict of interest statement. None declared.

Supplementary Material

[Supplementary Data]

ACKNOWLEDGEMENTS

We are grateful to M. Hemm and M. Kawano for help with detection of Flag-tagged proteins, to M. Maurizi for the gift of anti-EFTu antibody and to J. Vogel for sharing unpublished results. We thank N. Majdalani and other members of the lab for their help and P. Mandin, N. De Lay, G. Storz and J. Vogel for comments on the manuscript.

REFERENCES

1. Vogel J, Sharma CM. How to find small non-coding RNAs in bacteria. Biol. Chem. 2005;386:1219–1238. [PubMed]
2. Gottesman S, McCullen CA, Guillier M, Vanderpool CK, Majdalani N, Benhammou J, Thompson KM, FItzGerald PC, Sowa NA, FitzGerald DJ. Small RNA regulators and the bacterial response to stress. Cold Spring Harb. Symp. Quant. Biol. 2007;71:1–11. [PMC free article] [PubMed]
3. Storz G, Gottesman S. In: The RNA World. 3rd. Gesteland RF, Cech TR, Atkins JF, editors. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 2006. pp. 567–594.
4. Prevost K, Salvail H, Desnoyers G, Jacques JF, Phaneuf E, Massé E. The small RNA RyhB activates the translation of shiA mRNA encoding a permease of shikimate, a compound involved in siderophore synthesis. Mol. Microbiol. 2007;64:1260–1273. [PubMed]
5. Majdalani N, Cunning C, Sledjeski D, Elliott T, Gottesman S. DsrA RNA regulates translation of RpoS message by an anti-antisense mechanism, independent of its action as an antisilencer of transcription. Proc. Natl Acad. Sci. USA. 1998;95:12462–12467. [PMC free article] [PubMed]
6. Urban JH, Vogel J. Two seemingly homologous noncoding RNAs act hierarchically to activate glmS mRNA translation. PLoS Biol. 2008;6:e64. [PMC free article] [PubMed]
7. Brennan RG, Link TM. Hfq structure, function and ligand binding. Curr. Opin. Microbiol. 2007;10:125–133. [PubMed]
8. Moller T, Franch T, Hojrup P, Keene DR, Bachinger HP, Brennan RG, Valentin-Hansen P. Hfq: a bacterial Sm-like protein that mediates RNA-RNA interaction. Mol. Cell. 2002;9:23–30. [PubMed]
9. Kawamoto H, Koide Y, Morita T, Aiba H. Base-pairing requirement for RNA silencing by a bacterial small RNA and acceleration of duplex formation by Hfq. Mol. Microbiol. 2006;61:1013–1022. [PubMed]
10. Zhang A, Wassarman KM, Ortega J, Steven AC, Storz G. The Sm-like Hfq protein increases OxyS RNA interaction with target mRNAs. Mol. Cell. 2002;9:11–22. [PubMed]
11. Rasmussen AA, Ericksen M, Gilany K, Udesen C, Franch T, Peterson C, Valentin-Hansen P. Regulation of ompA mRNA stability: the role of a small regulatory RNA in growth phase-dependent control. Mol. Microbiol. 2005;58:1421–1429. [PubMed]
12. Urban JH, Vogel J. Translational control and target recognition by Escherichia coli small RNAs in vivo. Nucleic Acids Res. 2007;35:1018–1037. [PMC free article] [PubMed]
13. Morita T, Maki K, Aiba H. RNase E-based ribonucleoprotein complexes:mechanical basis of mRNA stabilization mediated by bacterial noncoding RNAs. Genes Dev. 2005;19:2176–2186. [PMC free article] [PubMed]
14. Massé E, Escorcia FE, Gottesman S. Coupled degradation of a small regulatory RNA and its mRNA targets in Escherichia coli. Genes Dev. 2003;17:2374–2383. [PMC free article] [PubMed]
15. Guillier M, Gottesman S, Storz G. Modulating the outer membrane with small RNAs. Genes Dev. 2006;20:2338–2348. [PubMed]
16. Vogel J, Papenfort K. Small non-coding RNAs and the bacterial outer membrane. Curr. Opin. Microbiol. 2006;9:605–611. [PubMed]
17. Valentin-Hansen P, Johansen J, Rasmussen AA. Small RNAs controlling outer membrane porins. Curr. Opin. Microbiol. 2007;10:152–155. [PubMed]
18. Papenfort K, Pfeiffer V, Lucchini S, Sonawane A, Hinton JC, Vogel J. Systematic deletion of Salmonella small RNA genes identifies CyaR, a conserved CRP-dependent riboregulator of OmpX synthesis. Mol. Microbiol. 2008;68:890–906. [PubMed]
19. De Lay N, Gottesman S. The Crp-activated sRNA CyaR (RyeE) links nutritional status to group behavior. J. Bacteriol. 2008 submitted. [PMC free article] [PubMed]
20. Guillier M, Gottesman S. Remodelling of the Escherichia coli outer membrane by two small regulatory RNAs. Mol. Microbiol. 2006;59:231–247. [PubMed]
21. Massé E, Gottesman S. A small RNA regulates the expression of genes involved in iron metabolism in Escherichia coli. Proc. Natl Acad. Sci. USA. 2002;99:4620–4625. [PMC free article] [PubMed]
22. Tsui H-CT, Feng G, Winkler M. Negative regulation of mutS and mutH repair gene expression by the Hfq and RpoS global regulators of Escherichia coli K-12. J. Bacteriol. 1997;179:7476–7487. [PMC free article] [PubMed]
23. Simons RW, Houman F, Kleckner N. Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene. 1987;53:85–96. [PubMed]
24. Miller JH. A Short Course in Bacterial Genetics, A Laboratory Handbook and Manual. Cold Spring Harbor, NY, pp. 72–74: Cold Spring Harbor Laboratory Press; 1992.
25. Kimata K, Tanaka Y, Inada T, Aiba H. Expression of the glucose transporter gene, ptsG, is regulated at the mRNA degradation step in response to glycolytic flux in Escherichia coli. EMBO J. 2001;20:3587–3595. [PMC free article] [PubMed]
26. Vanderpool CK, Gottesman S. Involvement of a novel transcriptional activator and small RNA in post-transcriptional regulation of the glucose phosphoenolpyruvate phosphotransferase system. Mol. Microbiol. 2004;54:1076–1089. [PubMed]
27. McDowall KJ, Hernandez RG, Lin-Chao S, Cohen SN. The ams-1 and rne-3071 temperature-sensitive mutations in the ams gene are in close proximity to each other and cause substitutions within a domain that resembles a product of the Escherichia coli mre locus. J. Bacteriol. 1993;175:4245–4249. [PMC free article] [PubMed]
28. Kido M, Yamanaka K, Mitani T, Niki H, Ogura T, Hiraga S. RNase E polypeptides lacking a carboxyl-terminal half suppress a mukB mutation in Escherichia coli. J. Bacteriol. 1996;178:3917–3925. [PMC free article] [PubMed]
29. Lopez PJ, Marchand I, Joyce SA, Dreyfus M. The C-terminal half of RNase E, which organizes the Escherichia coli degradosome, participates in mRNA degradation but not rRNA processing in vivo. Mol. Microbiol. 1999;33:188–199. [PubMed]
30. Nicholson AW. The ribonuclease III superfamily: forms and functions in RAN maturation, decay and gene silencing. In: Hannon GJ, editor. RNAi: A Guide to Gene Silencing. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 2003. pp. 149–174.
31. Wassarman KM, Repoila F, Rosenow C, Storz G, Gottesman S. Identification of novel small RNAs using comparative genomics and microarrays. Genes Dev. 2001;15:1637–1651. [PMC free article] [PubMed]
32. Zhang A, Wassarman KM, Rosenow C, Tjaden BC, Storz G, Gottesman S. Global analysis of small RNA and mRNA targets of Hfq. Mol. Microbiol. 2003;50:1111–1124. [PubMed]
33. Kawamoto H, Koide Y, Morita T, Aiba H. Base-pairing requirement for RNA silencing by a bacterial small RNA and acceleration of duplex formation by Hfq. Mol. Microbiol. 2006;61:1013–1022. [PubMed]
34. Geissmann TA, Touati D. Hfq, a new chaperoning role: binding to messenger RNA determines access for small RNA regulator. EMBO J. 2004;23:396–405. [PMC free article] [PubMed]
35. Griggs DW, Tharp BB, Konisky J. Cloning and promoter identification of the iron-regulated cir gene of Escherichia coli. J. Bacteriol. 1987;169:5343–5352. [PMC free article] [PubMed]
36. Vandeputte-Rutten L, Kramer RA, Kroon J, Dekker N, Egmond MR, Gros P. Crystal structure of the outer membrane protease OmpT from Escherichia coli suggests a novel catalytic site. EMBO J. 2001;20:5033–5039. [PMC free article] [PubMed]
37. Tjaden B, Goodwin SS, Opkyke JA, Guillier M, Fu DX, Gottesman S, Storz G. Target prediction for small, noncoding RNAs in bacteria. Nucleic Acids Res. 2006;34:2791–2802. [PMC free article] [PubMed]
38. Tsui P, Huang L, Freundlich M. Integration host factor binds specifically to multiple sites in the ompB promoter of Escherichia coli and inhibits transcription. J. Bacteriol. 1991;173:5800–5807. [PMC free article] [PubMed]
39. Bouvier M, Sharma CM, Mika F, Nierhaus KH, Vogel J. Small RNA binding the 5′ mRNA coding region inhibits translational inhibition. Mol. Cell. 2008 in press. [PubMed]
40. McClelland M, Florea L, Sanderson K, Clifton SW, Parkhill J, Churcher C, Dougan G, Wilson RK, Miller W. Comparison of the Escherichia coli K-12 genome with sampled genomes of a Klebsiella pneumoniae and three Salmonella enterica serovars, Typhimurium, Typhi and Paratyphi. Nucleic Acids Res. 2000;28:4974–4986. [PMC free article] [PubMed]
41. Sharma CM, Darfeuille F, Plantinga TH, Vogel J. A small RNA regulates multiple ABC transporter mRNAs by targeting C/A-rich elements inside and upstream of ribosome-binding sites. Genes Dev. 2007;21:2804–2817. [PMC free article] [PubMed]
42. Chen S, Zhang A, Blyn LB, Storz G. MicC, a second small-RNA regulator of Omp protein expression in Escherichia coli. J. Bacteriol. 2004;186:6689–6697. [PMC free article] [PubMed]
43. Schmidt M, Zheng P, Delihas N. Secondary structures of Escherichia coli antisense micF RNA, the 5′-end of the target ompF mRNA, and the RNA/RNA duplex. Biochemistry. 1995;34:3621–3631. [PubMed]
44. Liu J. Control of protein synthesis and mRNA degradation by microRNAs. Curr. Opin. Cell Biol. 2008;20:214–221. [PubMed]
45. Wilderman PJ, Sowa NA, FitzGerald DJ, FitzGerald PC, Gottesman S, Ochsner UA, Vasil ML. Identification of tandem duplicate regulatory small RNAs in Pseudomonas aeruginosa involved in iron homeostasis. Proc. Natl Acad. Sci. USA. 2004;101:9792–9797. [PMC free article] [PubMed]
46. Lenz DH, Mok KC, Lilley BN, Kulkarni RV, Wingreen NS, Bassler BL. The small RNA chaperone Hfq and multiple small RNAs control quorum sensing in Vibrio harveyi and Vibrio cholerae. Cell. 2004;118:69–82. [PubMed]
47. Kawamoto H, Morita T, Shimizu A, Inada T, Aiba H. Implications of membrane localization of target mRNA in the action of a small RNA: mechanism of post-transcriptional regulation of glucose transporter in Escherichia coli. Genes Dev. 2005;19:328–338. [PMC free article] [PubMed]
48. Thompson KM, Rhodius VA, Gottesman S. σE regulates and is regulated by a small RNA in Escherichia coli. J. Bacteriol. 2007;189:4243–4256. [PMC free article] [PubMed]
49. Papenfort K, Pfeiffer V, Mika F, Luchhini S, Hinton JCD, Vogel J. σE-dependent small RNAs of Salmonella respond to membrane stress by accelerating global omp mRNA decay. Mol. Microbiol. 2006;62:1674–1688. [PMC free article] [PubMed]
50. Massé E, Vanderpool CK, Gottesman S. Effect of RyhB small RNA on global iron use in Escherichia coli. J. Bacteriol. 2005;187:6962–6971. [PMC free article] [PubMed]
51. Svenningsen SL, Waters CM, Bassler BL. A negative feedback loop involving small RNAs accelerates Vibrio cholerae's transition out of quorum-sensing mode. Genes Dev. 2008;22:226–238. [PMC free article] [PubMed]
52. Batchelor E, Goulian M. Robustness and the cycle of phosphorylation and dephosphorylation in a two-component regulatory system. Proc. Natl Acad. Sci. USA. 2003;100:691–696. [PMC free article] [PubMed]

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...