Defining a role for Hfq in Gram-positive bacteria: evidence for Hfq-dependent antisense regulation in Listeria monocytogenes

Abstract

Small trans-encoded RNAs (sRNAs) modulate the translation and decay of mRNAs in bacteria. In Gram-negative species, antisense regulation by trans-encoded sRNAs relies on the Sm-like protein Hfq. In contrast to this, Hfq is dispensable for sRNA-mediated riboregulation in the Gram-positive species studied thus far. Here, we provide evidence for Hfq-dependent translational repression in the Gram-positive human pathogen Listeria monocytogenes, which is known to encode at least 50 sRNAs. We show that the Hfq-binding sRNA LhrA controls the translation and degradation of its target mRNA by an antisense mechanism, and that Hfq facilitates the binding of LhrA to its target. The work presented here provides the first experimental evidence for Hfq-dependent riboregulation in a Gram-positive bacterium. Our findings indicate that modulation of translation by trans-encoded sRNAs may occur by both Hfq-dependent and -independent mechanisms, thus adding another layer of complexity to sRNA-mediated riboregulation in Gram-positive species.

Figure 1.

Genetic organization and putative secondary structure of LhrA and its target lmo0850. (A) Chromosomal location of lhrA. The upstream region of lhrA was found to overlap with another gene, lmo2257 (dotted lines) encoding an unknown hypothetical protein. As discussed previously, the annotation of this open reading frame is highly questionable (2). (B) Chromosomal location of lmo0850, encoding a small hypothetical protein of unknown function. (C) MFOLD predicted secondary structure of LhrA RNA. The proposed interaction site with lmo0850 is indicated with asterisks. The two putative start codons of lmo0850 are boxed. The G → C substitutions introduced to abolish translation initiation in lmo0850-Mut1-lacZ and lmo0850-Mut2-lacZ as well as the U → A substitutions introduced to create lmo0850-STOP-lacZ are indicated in bold. The nucleotide substitutions introduced in lmo0850-Mut3-lacZ and lhrA-Mut3* are also highlighted in bold.

Figure 2.

LhrA inhibits lmo0850 translation in an Hfq-dependent manner. The left part shows a schematic overview of the seven lacZ fusions (A–G) tested by β-galactosidase assay in EGD wild-type, Δhfq and ΔlhrA mutant strains. Two constructs (pCK-lmo0850-lacZ and pCK-lmo0850-Mut2-lacZ) were tested in EGDlhrA-Mut3* and EGDΔhfq lhrA-Mut3* as well. The right part shows the corresponding specific β-galactosidase activity in cells harvested in the exponential growth phase. The presented activities are the averages of three independent experiments each conducted in duplicate.

Figure 3.

In vitro toeprint assay of lmo0850 RNA in the absence or presence of LhrA. In vitro transcribed lmo0850 RNA was incubated with 30S ribosomes in the absence or presence of excess levels of LhrA or LhrA-Mut3* RNA. Lanes 1 and 2: control reactions containing lmo0850 RNA only, or lmo0850 RNA together with 30S ribosomes. Lane 3: in the presence of 30S ribosomes and fMet tRNA, two distinct toeprint signals are observed precisely 13 nucleotides downstream from each of the two predicted start codons (AUG and UUG, respectively). Lanes 4 and 5: the addition of LhrA prevents formation of both toeprint signals. Lanes 6 and 7: the addition of LhrA-Mut3* does not prevent the formation of toeprint signals demonstrating that base pairing between LhrA and lmo0850 RNA is essential for regulation.

Figure 4.

LhrA downregulates lmo0850 transcript levels in an Hfq-dependent manner. (A) Northern blot showing the levels of lmo0850 mRNA, LhrA and 5S rRNA in EGD wild-type, Δhfq and ΔlhrA mutant strains at various time points during growth in BHI medium. (B) Northern blot showing the levels of lmo0850 mRNA, LhrA/LhrA-Mut3* and 5S rRNA in EGD wild-type (to the right) as compared to an lhrA-Mut3* strain (to the left) and the Δhfq lhrA-Mut3* strain (in the middle) at two different time points during growth in BHI medium. (C) Northern blot showing the effect of ectopic expression of LhrA from a high copy number plasmid. EGDΔlhrA containing an empty vector (pAT18) or the LhrA-expression vector (pAT18-lhrA) was grown in BHI medium. At the indicated time points, cells were harvested and total RNA was prepared, and the levels of lmo0850, LhrA and 5S rRNA was determined by northern blotting.

Figure 5.

Hfq stimulates LhrA-lmo0850 duplex formation. (A) In vitro binding assays of LhrA and lmo0850 RNA in the absence ( − ) or presence ( + ) of Hfq. In the left part of the panel, an end-labelled wild-type lmo0850 RNA fragment was used. In the right part of the panel, a lmo0850-Mut3 RNA fragment was used. Where indicated, in vitro transcribed LhrA RNA or LhrA-Mut3* RNA was added at 10 (+) or 100 (++) fold excess of lmo0850. (B) Time course experiment with end-labelled lmo0850 RNA fragment carried out in the absence (lanes1–5) or presence (lanes 6–10) of Hfq. The 5′-end-labelled lmo0850 RNA fragment was mixed with 100-fold excess LhrA RNA and then incubated at 37°C for 0, 1, 2, 5 or 10 min, chilled for 30 s on ice, and then loaded onto a gel. The experiment was repeated twice with similar results. (C) Quantification of the time course experiment in (B).

Figure 6.

Hfq_LMO is able to restore several key defects associated with Hfq in E. coli. (A) Growth curves of E. coli wild-type (SØ928) carrying the empty vector pNDM-220, and E. coli hfq⁻¹ carrying pNDM-220, pNDM-hfq_ECO or pNDM-hfq_LMO. Cells were cultivated in LB medium in the presence of 1 mM IPTG. The data shown are the result of three independent experiments each conducted in duplicate. (B) Resistance to oxidative stress. Overnight cultures were spread on agar plates and tested for their tolerance towards hydrogen peroxide by disk diffusion assay. Here, the averages of three independent experiments each conducted in triplicate are shown. The presence of three asterisks above a bar indicate a significant difference as compared to the hfq⁻ strain with P < 0.001. (C) Western blot analysis of the level of σ^S and GroEL (control). Cells were grown in LB medium containing 1 mM IPTG and cells were harvested at various time points during growth. E, exponential; T, transition phase; S, stationary phase. (D) Northern blot showing Hfq-dependent stabilization of RyhB and degradation of sodB mRNA. Cells were grown in LB medium containing 1 mM IPTG. At OD₆₀₀ = 0.4, the cultures were split and 2,2′-Dipyridyl (DIP) was added to one of the cultures. After 10 min, cells were harvested for RNA extractions, and RhyB RNA, sodB mRNA and 5S rRNA levels were analysed by northern blotting.