sRNA-controlled iron sparing response in Staphylococci

Abstract

Staphylococcus aureus, a human opportunist pathogen, adjusts its metabolism to cope with iron deprivation within the host. We investigated the potential role of small non-coding RNAs (sRNAs) in dictating this process. A single sRNA, named here IsrR, emerged from a competition assay with tagged-mutant libraries as being required during iron starvation. IsrR is iron-repressed and predicted to target mRNAs expressing iron-containing enzymes. Among them, we demonstrated that IsrR down-regulates the translation of mRNAs of enzymes that catalyze anaerobic nitrate respiration. The IsrR sequence reveals three single-stranded C-rich regions (CRRs). Mutational and structural analysis indicated a differential contribution of these CRRs according to targets. We also report that IsrR is required for full lethality of S. aureus in a mouse septicemia model, underscoring its role as a major contributor to the iron-sparing response for bacterial survival during infection. IsrR is conserved among staphylococci, but it is not ortholog to the proteobacterial sRNA RyhB, nor to other characterized sRNAs down-regulating mRNAs of iron-containing enzymes. Remarkably, these distinct sRNAs regulate common targets, illustrating that RNA-based regulation provides optimal evolutionary solutions to improve bacterial fitness when iron is scarce.

Figure 1.

S. aureus IsrR sRNA is required for optimal growth when iron is scarce. (A) Experimental protocol scheme to select mutants with altered fitness in media containing iron chelators. (B) Evolution of mutant proportions in libraries (for composition, see Supplementary Table S4) grown in the presence of EDDHA 0.7 mM or DIP 1.25 mM normalized to the same libraries grown in the absence of iron chelators. Error bars indicate the standard deviation from three independent libraries. Upper parts; data with the complete 48 mutant libraries. Tag135 corresponding to the isrR mutant is highlighted by a red box. For Tag/mutant correspondence, see Supplementary Table S1. Note that Tag145 and Tag149 are in a single strain, which corresponds to a double mutant (ΔsprX2 ΔsprX1). Lower parts, selected data (enlargement): ΔisrR and three other tagged regions are shown. Loci 1, 2 and 3 correspond to tag insertions in non-transcribed regions expected not to affect bacterial growth. Histogram color code corresponds to sampling time color code from (A). (C) ΔisrR growth defect in iron-depleted media is complemented by an isrR ectopic chromosomal copy. For strain constructions see Supplementary Table S1 and Supplementary Figure S2A. Plating efficiency of indicated strains on BHI medium without (upper panels) or with (lower panels) DIP 1.25mM. Three independent biological clones are shown for each strain. For results with EDDHA, see Supplementary Figure S2C. (D) Multicopy isrR is toxic in iron-depleted media. Experiments were performed as for Figure 1C. pCont, pCN38; pIsrR, pCN38-IsrR. For strain and plasmid constructions see Supplementary Tables S1 and S2.

Figure 2.

isrR is regulated by Fur. (A) isrR genetic locus. The transcribed region is in red. For isrR 3′-5′RACE mapping results, see Supplementary Figure S3A. Predicted Fur boxes are bold and underlined. The staphylococcal Fur box consensus is taken from the RegPrecise database (30). (B) Northern blot experiment with HG003 and its fur derivative sampled at the indicated OD₆₀₀ and probed for IsrR and tmRNA (control) (n = 2). (C) isrR Fur box sequence (WT) and its mutant derivatives (P_isrR1, P_isrR2 and P_isrR1&2). Mutations altering the Fur boxes are shown in black boxes. (D) Fluorescence of transcriptional fusions placing gfp under the control of the isrR promoter and its mutant derivatives (as indicated) measured by flow cytometry, (n = 5). The first strain indicated ‘isrR’ is a control strain in which the isrR gene is not fused to gfp.

Figure 3.

isrR is conserved and Fur-regulated within the Staphylococcus genus. (A) Sequence conservation. The alignment was obtained using LocARNA (31) with isrR sequences from indicated strains generated by GLASSgo (28) as inputs. 80% conserved nucleotides within the tested sequences are colored (green, A; blue, U; red, G; ocher, C). Sequences shown are limited to the three first stem-loops (H1 to H3). There is poor nucleotide sequence conservation for H3 and the transcriptional terminator. Three C-rich regions are indicated (CRR1 to CRR3). (B) Northern blot probed for IsrR in S. aureus HG003, S. epidermidis ATCC 12228, S. haemolyticus JCSC1435, and S. ludgunensis N920143. Bacteria were grown in BHI or BHI supplemented with DIP 1.25 mM and were withdrawn at OD₆₀₀ 1 (n = 2).

Figure 4.

IsrR secondary structure model obtained using IPANEMAP. IsrR was probed with 1M7 at 37°C in 40 mM HEPES pH 7.5, 150 mM KCl with/without 5 mM MgCl₂. White: low reactivity to 1M7, yellow: moderate reactivity, red: high reactivity. Nucleotides in grey denote undetermined reactivity. Specific regions of the IsrR structure are indicated: Three stem–loop structures (H1 to H3), a rho-independent transcription terminator (T), and three C-rich regions (CRR1 to CRR3).

Figure 5.

Stability of fdhA and gltB2 mRNAs is not significantly affected by IsrR. HG003 (left panel) and its isogenic ΔisrR derivative (right panel) were grown in rich medium with or without the addition of DIP (as indicated). At t₀, rifampicin (Rif) was added to the growth medium. Cultures were sampled at t₀, 1, 3, 5, 10 and 20 min after addition of rifampicin, total RNA was extracted, and the amounts of fdhA mRNA, gltB2 mRNA, tmRNA (loading control), and IsrR were assessed by Northern blot. Histograms show the quantification of fdhA and gltB2 mRNAs normalized to t₀ from two rifampicin assays as shown in the upper panel. Vertical axis, arbitrary units. Error bars indicate the standard deviation from two independent experiments (n = 2).

Figure 6.

Secondary structure models for fdhA and gltB2 5′UTR alone or in interaction with IsrR. Top panels: Secondary structure model obtained with IPANEMAP for fdhA (A) and gltB2 (B) 5′UTRs using 1M7 reactivity as constraints. Nucleotides are coloured according to their reactivity in the absence of IsrR with the indicated colour code. ND, not determined. Bottom panels: model for interaction between IsrR and fdhA (A) and gltB2 (B) mRNAs based on changes in reactivity in the presence of IsrR. mRNA Shine-Dalgarno sequence is shown in a blue rectangle and the start codon in a green rectangle; IsrR C-rich regions are shown in red rectangles.

Figure 7.

Translational down-regulation by IsrR and the CRR contribution. Leader fusions between the first codons of fdhA or gltB2 and GFP were constructed (Supplementary Figure S9 and Table S2). The cloned fragments include the interaction regions with IsrR as described (Figure 6). HG003 ΔisrR derivatives with either a control plasmid (no IsrR; pRMC2ΔR), or plasmids expressing IsrR (pRMC2ΔR-isrR), IsrRΔCRR1 (pRMC2ΔR-isrRΔCRR1), IsrRΔCRR2 (pRMC2ΔR-isrRΔCRR2), IsrRΔCRR3 (pRMC2ΔR-isrRΔCRR3) were transformed with each engineered reporter gene fusion. Translational activity from the reporters in the presence of the different isrR derivatives were evaluated by fluorescence scanning of streaked clones on plates (n = 3). Fully active IsrR derivatives are shown in red. Translational activity of the reporter genes with the different isrR derivatives was also determined in liquid culture. Fluorescence of the strains was measured in 6 h cultures using a microtiter plate reader. Results are normalized to 1 for each fusion with the control plasmid. Error bars indicate the standard deviation from three independent experiments (n = 3). Statistical analyses were performed using a t-test with Welch's correction: ^**** represents P-value < 0.0001, *** represents P-value of 0.0003, ** represents P-value of 0.0029, * represents P-value between 0.0111–0.0184.

Figure 8.

IsrR controls nitrite production. IsrR prevents nitrate conversion to nitrite. Strains were grown in anaerobic conditions for two hours and nitrate (20 mM) was added to cultures. Growth media: left panel, BHI; right panel, TSB. TSB medium was used here since we observed that Δfur mutants grow poorly in BHI under anaerobic conditions. Samples for nitrite measurement were withdrawn at times 0 and 150 (left panel) or 240 min (right panel) upon nitrate addition. p, pRMC2ΔR; p-IsrR, pRMC2ΔR-IsrR. Histograms represent the relative nitrite concentration (Griess assay, OD₅₄₀) normalized to the bacterial mass (OD₆₀₀). Results are normalized to 1 for ΔisrR p (left panel) and WT (right panel) samples prior to nitrate addition. Error bars indicate the standard deviation from three independent experiments (n = 3). Statistical analyses were performed using a t-test with Welch's correction: ** represents P-value between 0.001–0.004; ns, non-significant.

Figure 9.

IsrR is required for S. aureus full virulence. Kaplan–Meier survival probability plots in a septicemia model of mice infected with either HG003 (WT, black), HG003 ΔisrR (red) and HG003 ΔisrR locus2::isrR⁺ (ΔisrR complemented, blue). Survival was monitored for 8 days post-infection. Results shown are from 10 mice per group; the experiment was performed twice, and data combined. The Mantel-Cox test was used to determine P values. NS, non-significant.

Figure 10.

IsrR control of dissimilatory nitrate reduction. isrR is Fur-regulated. Its expression is induced in iron-free growth conditions. IsrR base pairs to SDs of mRNAs encoding components of formate dehydrogenase, nitrate reductase, nitrite reductase and glutamate synthase, thus preventing translation of the encoded iron-sulfur containing enzymes and nitrate dissimilatory reduction.