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.

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 OmrA/B. This was done in RNase E (A and B) or RNase III (C) mutant cells as well as in their wild-type counterparts. Blots were also probed for SsrA or ompA mRNA as loading controls. This experiment was carried out using strains MG1273 (rne⁺), MG1274 (rne-3071), MG1277 (rne-131), MG1100 (rnc⁺) and MG1272 (rnc⁻).

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 codons and putative Shine–Dalgarno sequences are in gray.

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 vector. Strains used in this experiment are MG1193, MG1195 and MG1188. (B) Northern blot analysis of ompT mRNA levels after OmrA/B induction in an hfq⁺ (strain MG1099) or hfq⁻ (strain MG1279) background. SsrA and ompA RNAs were used as loading controls. When probed together, the relative levels of Omr sRNAs in hfq⁺ and hfq⁻ cells were similar to the ones shown in (B) (data not shown).

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 activity of the resulting strains transformed with pBRplac vector or its derivatives expressing OmrA or OmrB. Nucleotides in red are conserved between OmrA and OmrB, and nucleotides in gray correspond to the start codon and the Shine–Dalgarno sequence of cirA. (B) The wt strain is MG1189, and strains MG1206, MG1205, MG1253, MG1255, MG1261 and MG1256 carry derivatives of the cirA-lacZ fusion with mutations Δ1, Δ2, mutI, mutIIa, mutI+IIa and mutIIb, respectively. Nucleotides shown as X are not conserved between OmrA and OmrB (Figure 2).

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 by measuring the β-galactosidase activity of the resulting strains transformed with pBRplac or its derivatives (B). Mutant and wild-type forms of OmrA or OmrB accumulate at similar levels in the wild-type strain as shown by northern blot analysis (C). Strains used in this set of experiments are MG1189 (wt), MG1254 (up), MG1340 (up/2) and MG1341 (up/3).

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 codon and Shine–Dalgarno sequence. (B) Control was assayed by analyzing the accumulation of OmpT-Flag (strain MG1307) and OmpTmut2-Flag (strain MG1309) proteins when wild-type or mut2* OmrA/B are overexpressed (C) Levels of ompT (strain MG1099), ompTmut2 (strain MG1281) mRNAs as well as of their Flagged derivatives (strains MG1307 and MG1309, respectively) were monitored by northern blot before and after a 10-min induction of the different forms of OmrA/B sRNAs. SsrA and ompA RNAs were used as loading controls.

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 start codon and Shine–Dalgarno sequence. (B) Northern blot analysis of target mRNA before and after induction of wt or mut2* OmrA/B. SsrA and ompA RNAs were used as loading controls. OmrA, OmrB, SsrA, ompT, ompR and ompA RNAs were analyzed in the MG1099 strain, whereas the fur⁻ strain MG1100 was used for cirA, fecA and fepA mRNAs. Induction of OmrA/B as well as the levels of SsrA and ompA mRNA were as expected in MG1100 (data not shown). (C) Effect of compensatory changes in ompR-lacZ mRNA and/or OmrA/B sRNAs on the activity of a pBAD-ompR-lacZ translational fusion. Fusion strains are MG1398 (wt) and MG1403 (mut2).