The Staphylococcus aureus non-coding RNA IsrR regulates TCA cycle activity and virulence

Abstract

Staphylococcus aureus has evolved mechanisms to cope with low iron (Fe) availability in host tissues. Staphylococcus aureus uses the ferric uptake transcriptional regulator (Fur) to sense titers of cytosolic Fe. Upon Fe depletion, apo-Fur relieves transcriptional repression of genes utilized for Fe uptake. We demonstrate that an S. aureus Δfur mutant has decreased expression of acnA, which codes for the Fe-dependent enzyme aconitase. This prevents the Δfur mutant from growing with amino acids as sole carbon and energy sources. We used a suppressor screen to exploit this phenotype and determined that a mutation that decreases the transcription of isrR, which produces a regulatory RNA, increased acnA expression, thereby enabling growth. Directed mutation of bases predicted to facilitate the interaction between the acnA transcript and IsrR, decreased the ability of IsrR to control acnA expression in vivo and IsrR bound to the acnA transcript in vitro. IsrR also bound transcripts coding the alternate tricarboxylic acid cycle proteins sdhC, mqo, citZ and citM. Whole-cell metal analyses suggest that IsrR promotes Fe uptake and increases intracellular Fe not ligated by macromolecules. Lastly, we determined that Fur and IsrR promote infection using murine skin and acute pneumonia models.

Graphical Abstract

Graphical Abstract.

Introduction

Staphylococcus aureus is an infectious agent and a primary cause of morbidity and mortality worldwide. It is commonly associated with community- and hospital-acquired infections, and it can cause a variety of ailments, from minor soft tissue infections to more severe diseases, including septicemia and endocarditis (1). A subset of S. aureus-associated infections is caused by strains resistant to commonly prescribed antibiotics, complicating treatment (2).

The acquisition of ionic iron (Fe) is essential for S. aureus. Staphylococcus aureus strains defective in the acquisition, uptake or proper usage of intracellular Fe have decreased virulence in models of infection (3–6). The importance of Fe is highlighted by the fact that ∼2% of the genomic protein-coding open reading frames (ORFs) code for Fe acquisition systems. Fe functions as a cofactor to many proteins, including iron–sulfur (Fe–S) cluster enzymes utilized for respiration and the tricarboxylic acid (TCA) cycle (7).

One strategy that mammals use to combat infections is to limit the availability of nutrients that are essential for bacterial growth. This process, known as nutritional immunity, limits bacterial access to trace metals, including Fe ions (8). Inside the host, Fe is abundant, but it is typically found associated with proteins in prosthetic groups, including heme or Fe–S clusters. Non-cofactor-associated Fe, in large part, is ligated by proteins, including transferrin, lactoferrin and calprotectin (4,9,10). Therefore, the ‘free’ or loosely ligated Fe concentrations within host tissues are low to restrict pathogen growth (11). Hosts further limit Fe availability during infection by reducing the absorption of dietary Fe in a process referred to as anemia of inflammation (12). Pathogen-containing macrophages export Fe ions and increase the expression of ferritin to decrease free Fe titers (13,14). Individuals with the inborn disease hemochromatosis have increased Fe titers and an increased prevalence of infection, highlighting the importance of Fe limitation to prevent infection (15).

Given the scarcity of Fe in host tissues, bacteria must alter their gene expression profiles to colonize tissue and promote infection. A key part of the response to low Fe is the upregulation of Fe uptake systems, including the expression of siderophores, which are high-affinity extracellular Fe ion chelating molecules that can compete with host proteins for Fe ions (16). Once bound to Fe, siderophores are transported back into the cell via specific surface receptors. Other strategies for Fe uptake include the use of cell-wall-associated transferrin-binding proteins, monovalent Fe uptake and acquisition of Fe from host hemoglobin (17).

In S. aureus, the ferric uptake transcriptional regulator (Fur) and partner Fpa sense and respond to cytosolic Fe titers (11,18). When associated with Fe, Fur typically acts as a transcriptional repressor (19,20). Growth in a low-Fe medium promotes Fur demetallation, altered affinity for DNA and expression of the Fur regulon. RegPrecise 3.0 predicts that there are 20 ORFs or operons in S. aureus that contain consensus Fur-binding sites in their operators (21). In large part, these are genes involved in Fe uptake or storage. An S. aureus Δfur mutant, or growth during Fe limitation, increases and decreases glycolysis and TCA cycle expression, respectively (19,22). The mechanism behind the Fur-dependent regulation of central metabolic pathways in S. aureus has remained elusive, but a recent study suggests that IsrR plays a role (23).

In Escherichia coli and Bacillus subtilis, Fur directly controls the transcription of ryhB and fsrA, respectively (24,25). Both loci produce small non-coding regulatory RNAs (sRNAs) that modulate gene expression by directly pairing with several messenger RNAs (mRNAs). In doing so, these sRNAs alter metabolism to spare Fe ions and redirect usage to prioritize essential processes. A recent publication demonstrated that the sRNA IsrR [originally named Tsr25 (26)] was required for growth during divalent metal starvation, suggesting that it is a functional analog of RyhB or FsrA (27). The expression of IsrR resulted in decreased expression of Fe-dependent enzymes formate dehydrogenase (fdhA) and glutamate synthase (gltB2). A direct interaction was noted between IsrR and the fdhA transcript in vitro. The isrR locus was initially identified in a global transcriptomic analysis as the most upregulated sRNA in human serum (26). Consistent with a role in pathogenesis, an ΔisrR mutant caused decreased mortality in a murine septicemia model (27).

Staphylococcus aureus relies on appropriate metabolic adaptation to host encountered stresses such as Fe limitation (19). Additionally, central metabolism and TCA cycle function impact virulence factor production and infection (28,29). This study was initiated to determine why an S. aureus Δfur mutant has decreased expression of the TCA cycle enzyme aconitase. We demonstrate that IsrR modulates the expression of acnA and directly interacts with the acnA mRNA transcript. IsrR also alters the expression of additional genes coding TCA cycle enzymes. We demonstrate that Fur and IsrR contribute to cellular Fe homeostasis and virulence using murine infection models of skin and pneumonia.

Materials and methods

Chemicals, bacterial strains and growth conditions

Unless specified, the S. aureus strains used in this study (Table 1) were isogenic and constructed in the community-associated S. aureus MRSA strain USA300_LAC that was cured of the native plasmid pUSA03 that confers erythromycin (Erm) resistance (30). All bacteria were grown at 37°C in tryptic soy broth (TSB; MP Biomedicals) or a chemically defined medium containing the 20 canonical amino acids with or without 10 mM glucose (3). Solid tryptic soy agar (TSA) and chemically defined media were generated by adding 1.5% (w/v) agar (VWR). Liquid cultures were shaken at 200 RPM. TSB treated with Chelex 100 resin (Bio-Rad Laboratories, Inc.) was prepared as previously described (31). Unless stated otherwise, cells were cultured in 10-ml-capacity culture tubes containing 2.0 ml of liquid medium. Liquid phenotypic analysis was conducted in 96-well microtiter plates containing 200 μl of media per well using a BioTek 808E visible absorption spectrophotometer with continuous shaking at high shake speed. The optical density of cultures was measured at 600 nm (A₆₀₀). For quantitative growth, strains were grown overnight and washed with phosphate buffered saline (PBS) before diluting to an optical density (A₆₀₀) of 0.05 in 200 μl of media. For spot plate growth analyses using solid media, strains were cultured for 18 h in TSB before harvesting by centrifugation. Cells were washed with PBS, standardized to an optical density of 2 (A₆₀₀), serially diluted in PBS and 5 μl aliquots were spotted upon solid media. Fe salts were added as ferrous sulfate.

Table 1.

Staphylococcus aureus USA300_LAC strains used in this study

Strain name	Genotype	Reference
JMB1100	USA300_LAC WT	A. Horswill
JMB10842	Δfur::tetM	(34)
JMB10495	Δfur::tetM isrR*	This study
JMB11112	Δfur::tetM SAUSA300_1309::Tn	This study
JMB11113	Δfur::tetM isrR* SAUSA300_1309::Tn	This study
JMB11292	ΔisrR	This study
JMB11293	Δfur::tetM ΔisrR	This study
JMB1886	geh::pLL39	(69)
JMB11803	acnA::Tn	(37)
JMB7868	ΔacnA::tetM	(28)
JMB14375	sdhA::Tn	This study
JMB11804	acnA::Tn Δfur::tetM	This study
JMB11805	acnA::Tn ΔisrR	This study
JMB11806	acnA::Tn Δfur::tetM ΔisrR	This study
JMB11448	SAUSA300_1456::Tn geh::pLL39	This study
JMB11449	fur E11Stop SAUSA300_1456::Tn geh::pLL39	This study
JMB11395	ΔisrR SAUSA300_1456::Tn geh::pLL39	This study
JMB11392	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39	This study
JMB11393	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39_isrR	This study
JMB13983	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39_isrR_C1	This study
JMB13984	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39_isrR_C2	This study
JMB13985	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39_isrR_C1_C2	This study
JMB14888	fur E11Stop ΔisrR SAUSA300_1456::Tn geh::pLL39_isrR_C2a	This study

Antibiotics were added at the following final concentrations: 100 μg ml^-1 ampicillin (Amp); 10 μg ml^-1 chloramphenicol (Cm) to select for plasmids and 3.3 μg ml^-1 Cm to maintain plasmids (called TSB-Cm); 5 μg ml^-1 Erm; 3 μg ml^-1 tetracycline (Tet); and 100 ng ml^-1 anhydrotetracycline. Protein concentrations were determined using Bradford reagent (Bio-Rad Laboratories, Inc., Hercules, CA). Unless stated otherwise, all chemicals were purchased from Sigma–Aldrich (St Louis, MO).

Plasmid and strain construction

The restriction minus strain S. aureus RN4220 was used for transformations (32) and transductions were done using bacteriophage 80α (33). 5α competent E. coli (New England Biolab; NEB) cultured in lysogeny broth was used for plasmid preparation.

Synthetic DNA (Supplementary Table S2) was synthesized by Twist Biosciences (San Francisco, CA) and DNA primers (Supplementary Table S1) were synthesized by Integrated DNA Technologies (Coralville, IA). Plasmids are listed in Table 2. Quick Ligase, restriction enzymes, competent E. coli and HiFi DNA Assembly Kit were purchased from NEB. All bacterial strains were polymerase chain reaction (PCR)- or sequence-verified before use. Plasmid DNA and PCR products were sequenced by Azenta Life Sciences (South Plainfield, NJ).

Table 2.

Plasmids used in this study

Name	Function	Reference
pJB38	Mutant generation	(70)
pJB38_ΔisrR	ΔisrR mutant generation	This study
pEPSA5	Inducible expression complementation	(71)
pEPSA5_isrR	Inducible expression isrR	This study
pEPSA5_as_isrR	Inducible expression isrR antisense	This study
pEPSA5_trunk_isrR	Inducible expression truncated isrR	This study
pEPSA5_isrR*	Inducible expression isrR*	This study
pEPSA5_acnA	Non-native promoter expression	This study
pOS-1-plgt	Genetic complementation lgt promoter	(72)
pOS-1-plgt_fur	Genetic complementation lgt promoter	(5)
pOS_pisrR_gfp	isrR transcriptional reporter	This study
pOS_pisrR*_gfp	isrR* transcriptional reporter	This study
pLL39	Genetic complementation	(73)
pLL39_isrR	Native promoter complementation	This study
pLL39_isrR_C1	Native promoter complementation	This study
pLL39_isrR_C2	Native promoter complementation	This study
pLL39_isrR_C1_C2	Native promoter complementation	This study
pLL39_isrR_C2a	Native promoter complementation	This study
pOS_plgt_acnA_gfp	acnA translational reporter	This study
pOS_plgt_acnA1_gfp	Mutated acnA translational reporter	This study
pOS_plgt_acnA2_gfp	Mutated acnA translational reporter	This study

The pJB38_ΔisrR was constructed by combining two digested PCR amplicons corresponding to the DNA upstream and downstream of the isrR locus with digested pJB38. The upstream region of isrR was amplified using primer pair tsr25 AF and tsr25 AR. The downstream region was amplified using primer pair tsr25 BF and tsr25 BR. pJB38 was digested using SacI and KpnI. The upstream amplicon was digested with SacI and MulI. The downstream region was digested with MulI and KpnI.

The pEPSA5_isrR complementation vector was created using the tsr25 pEPSA 5 EcoRI and tsr25 pEPSA 3 SalI primer pair with JMB1100 as template DNA. The pEPSA5_as_isrR complementation vector was created using the tsr25 pMAL 5 SalI and tsr25 pMAL 3 EcoRI primer pair with JMB1100 as template DNA. The pEPSA5_isrR* vector was created using the tsr25* pEPSA 5 EcoRI and tsr25 pEPSA 3 SalI primer pair with JMB10495 cells as template DNA. The pEPSA5_trunk_isrR was created using the tsr25 trunk pEPSA 5 EcoRI and tsr25 pEPSA 3 SalI primer pair with JMB1100 cells as the DNA template. The pEPSA5_acnA expression vector was created using the pEPSA5_acnAEcoRI (acnA RBS) and pEPSA_acnA3SalI primer pair. All pEPSA5 vectors and inserts were digested with EcoRI and SalI.

The pLL39_isrR complementing plasmid, pLL39_isrR_C1, pLL39_isrR_C2 and pLL39_isrR_C1_C2 were created using the Tsr25 comp 5 SalI and Tsr25 comp 3 BamHI primer pair with JMB1100, synthetic DNA isrR_C1, synthetic DNA isrR_C2 and synthetic DNA isrR_C1_C2 used as template DNA, respectively. All pLL39 vectors and inserts were digested with SalI and BamHI. The pLL39_isrR_C2a plasmid was generated by site-directed mutagenesis using the pLL39_isrR as DNA template, the SDM_IsrR_a_FWD and SDM_IsrR_a_REV primer pair and the Q5 Site-Directed Mutagenesis Kit (NEB).

The pOS_pisrR_gfp and pOS_pisrR*_gfp transcriptional reporters were constructed using the pOStsr25_2forHindIII and pOStsr25_2revkpnI primer pair with JMB1100 and JMB10495 cells as template DNA, respectively. PCR products were digested and ligated into KpnI and HindIII digested pOS_saeP1_gfp.

The pOS_plgt_acnA_gfp, pOS_plgt_acnA1_gfp and pOS_plgt_acnA2_gfp translational reporters were generated using pOS_plgt digested with NdeI. The aconitase insert of pOS_plgt_acnA_gfp was created using the plgt_acnA and acnA_gfp Wt A rev primer pairs with JMB1100 as template DNA. The gfp insert of pOS_plgt_acnA_gfp was created using the acnA_gfp WT A for and gfp_plgt rev primer pair with the pOS_saeP1_gfp plasmid as template DNA. The aconitase insert of pOS_plgt_acnA1_gfp was created using the plgt_acnA and acnA_gfp MutC rev primer pair with JMB1100 as template DNA. The gfp insert of pOS_plgt_acnA1_gfp was created using the acnA_gfp MutC for and gfp_plgt rev primer pair with the pOS_saeP1_gfp plasmid as template DNA. The aconitase insert of pOS_plgt_acnA2_gfp was created using the plgt_acnA and acnA_a rev primer pair with JMB1100 as template DNA. The gfp insert of pOS_plgt_acnA2_gfp was created using the acnA_a for and gfp_plgt rev primer pair with the pOS_saeP1_gfp plasmid as template DNA. All digested pOS_plgt vectors, acnA and gfp inserts were ligated using the NEB Hifi DNA Assembly Master Mix.

RNA extraction, cDNA synthesis and qPCR

To analyze RNA abundances corresponding to isrR, S. aureus strains were cultured overnight and diluted into 2 ml TSB to an optical density (A₆₀₀) of 0.1 in 10-ml culture tubes. The cell cultures were incubated with shaking until an A₆₀₀ of 0.5 before 1 ml of cells was harvested by centrifugation, washed with PBS and resuspended in 500 μl RNAprotect (QIAGEN).

To analyze isrR transcript stability, the Δfur::tetM ΔisrR strain containing pEPSA5_isrR or pEPSA5_isrR* were cultured overnight in TSB with 10 μg ml^-1 Cm in 10-ml culture tubes at 37°C with shaking. Cultures were diluted into 5 ml of fresh TSB with 10 μg ml^-1 Cm and 2% xylose to an A₆₀₀ of 0.1 in 30-ml culture tubes. The cells were cultured with shaking until an A₆₀₀ of 0.5. Next, 1.5 ml of culture was added to 10-ml culture tubes containing rifampicin (100 μg ml^-1) and incubated with shaking at 37°C. Cells were harvested at indicated timepoints by centrifugation, washed with PBS and resuspended in 500 μl RNAprotect (QIAGEN). RNA extraction, complementary DNA (cDNA) synthesis and transcript quantification (QuantStudio 3; Bio-Rad Laboratories, Inc., Hercules, CA) were performed as previously described (31).

Northern blot analyses

RNA (3 μg/lane) was loaded onto a formaldehyde agarose gel and electrophoresed for 1 h 30 min at 120 V. RNAs were transferred onto a positively charged nylon membrane overnight by capillary transfer with 20× Saline-Sodium Citrate (SSC) buffer (1× SCC is 0.15 M NaCl plus 0.015 M sodium citrate) buffer and ultraviolet-cross-linked to the membrane. The presence of ribosomal RNA and ladder bands was visualized by staining the membrane with methylene blue (0.04% in 0.5 M acetate solution). To detect isrR, a radiolabeled probe was made as follows: PCR targeting isrR using primers isrR5north and isrR3north was performed and the PCR mixture was radiolabeled using the Roche Random Prime Labeling Kit. Approximately 1 μg of PCR product was used with [α-32P] ATP according to the manufacturer’s protocol. Probes were purified using Illustra MicroSpin G-25 columns (GE Healthcare). Membranes were prehybridized overnight at 45 °C in ULTRAhyb-Oligo buffer (Thermo Fisher Scientific) and then incubated with radiolabeled probe overnight at 45°C. After incubation, membranes were washed with 2×, 1× and 0.5× SSC buffer and visualized using a phosphor imager screen.

Suppressor screen, genome sequencing and SNP mapping

Ten independent cultures of the Δfur::tetM strain (JMB10842) were grown overnight in 2 ml of TSB in 10-ml-capacity culture tubes. One milliliter of cells was pelleted by centrifugation, resuspended in 1 ml PBS and then diluted 1:100 in PBS. One hundred microliters of the dilution was spread on chemically defined agar media containing amino acids without glucose. One colony from each plate (i.e. one per culture for ten total) was retained and struck for isolation. The increased growth phenotypes were verified by serially diluting cultures and spot plating them on solid defined amino acid media with and without glucose. We repeated the assay for a second time with five independent overnight cultures for a total of 15 suppressed strains.

For chromosomal DNA isolation, the Δfur::tetM suppressed strains and the Δfur::tetM parent strain were cultured overnight in 2 ml of TSB and genomic DNA was purified using the MaterPure Gram Positive Genomic DNA Purification Kit (LGC Biosearch Technologies). DNA was sequenced by SeqCenter (Pittsburgh, PA, USA) using Illumina technology.

Whole genome sequencing data were analyzed using the CLC Genomics Workbench software package (QIAGEN). Reads were aligned to the S. aureus genome, using the USA300_FPR3757 genome sequence as a reference, as previously described (34). Quality-based variant detection was then performed to identify polymorphisms in each strain. A minimum threshold detection frequency of 80% was employed. The lists of polymorphisms generated for each suppressor mutant strain were cross-referenced against the parental strain (JMB10842). Common polymorphisms were eliminated (as were polymorphisms in homopolymeric nucleotide tracts) resulting in the identification of specific genetic variations between the suppressor strains and parental strain. The single nucleotide polymorphism (SNP) in the promoter of the isrR gene was confirmed in the suppressed strains using Sanger sequencing. These strains were referred to as Δfur isrR* and JMB10495 was used as a representative strain.

Transcriptional reporter assays

Overnight cultures of S. aureus strains containing a plasmid-based transcriptional reporter were cultured overnight in 2 ml TSB supplemented with 3.3 μg ml^-1 Cm in 10-ml culture tubes at 37°C with shaking. Cultures were then diluted to an optical density (A₆₀₀) of 0.05 in triplicate into 2 ml TSB supplemented with 3.3 μg ml^-1 Cm ± 250 μM 2,2′-dipyridyl (DIP) in 10-ml culture tubes and incubated at 37°C with shaking for 16 h. Optical density (A₆₀₀) and GFP fluorescence (excitation 485 nm, emission 520 nm) were measured in microtiter plates using a Varioskan Lux plate reader (Thermo Fisher Scientific).

EMSA assays

PCR fragments containing T7-acnA (from −41 to +541), T7-citM (from −46 to +725), T7-citZ (from −41 to +850), T7-lukH (from −100 to +807), T7-mqo (from −248 to +328) and T7-sdhC (full-length, from −260 to +615) were used as DNA template for in vitro transcription with T7 RNA polymerase. RNAs were finally purified and radiolabeled when required as previously described (35).

5′-radiolabeled IsrR (20 000 cpm/sample, concentration <1 pM) and aforementioned cold RNAs were separately denatured at 90°C in the buffer GR- (20 mM Tris–HCl pH 7.5, 60 mM KCl, 40 mM NH₄Cl, 3 mM DTT), cooled 1 min on ice and incubated at room temperature for 15 min in presence of 10 mM MgCl₂. Renatured RNAs were then mix and incubated at 37°C for 15 min. Finally, samples were loaded on a 6% polyacrylamide gel under non-denaturing conditions (300 V, 4°C). Results are representative of two independent experiments.

Enzyme assays

Aconitase assays

Aconitase (AcnA) assays were conducted as previously described (36). Briefly, strains were cultured overnight in 2 ml of TSB with or without 10 μg ml^-1 Cm before diluting them to an optical density (A₆₀₀) of 0.05 in 2 ml of TSB in 10-ml culture tubes. For strains containing pEPSA5, strains were cultured in TSB supplemented with 0.25% or 2% xylose for pEPSA_acnA and pEPSA_isrR containing strains respectively, and 3.3 μg ml^-1 Cm. Strains were cultured with shaking at 200 RPM at a 45 degree angle for 16 h. After incubation, 1 ml of cells was pelleted down and washed twice with PBS. After assaying AcnA activity as previously described (37), protein concentrations were determined using Bradford protein colorimetric assay modified for 96-well plate (Bio-Rad Protein Assay Dye Reagent Concentrate).

Succinate dehydrogenase assays

Succinate dehydrogenase (Sdh) assays were conducted as previously described (38). Cells were cultured and lysed as described for AcnA assays. The oxidation of succinate by succinate dehydrogenase was followed spectrophotometrically using the redox dye 2,6-dichlorophenol indophenol (DCPIP). The reaction mixture (1 ml) contained potassium phosphate buffer (0.1 M, pH 7.4), KCN (3 mM), DCPIP (25 μM), N-methylphenazonium methosulphate (2.2 mM), succinic acid (20 mM, pH 7) and 20 μl of cell lysate. The reduction of DCPIP (molar extinction coefficient 21 mM^-1 cm^-1) was followed at 600 nm for 2 min after addition of all reagents as described (39). The succinate-dependent slope (difference between the slope of absorbance/time of the sample with succinate minus the slope of the sample without succinate) was used to calculate specific Sdh activity. Protein concentrations were determined as described for the aconitase assay.

Malate quinone oxidoreductase assays

Cells were cultured and lysed as described for AcnA assays. Reaction mixture was the same as described for Sdh assays except for the use of sodium malate (20 mM) instead of succinic acid.

Streptonigrin sensitivity

Cells were cultured overnight and diluted to an optical density (A₆₀₀) of 0.05, after which 100 μl of the dilute culture was added to 4 ml of soft TSB agar (0.3% w/v) and overlaid on TSA media. After soft agar solidified, 2.5 μl of streptonigrin (1 mg ml^-1) was spotted. Zone of clearance was measured after 1-day incubation at 37°C.

CAS siderophore assay

Overnight cultures in TSB were diluted 100-fold into 1 ml of Chelex (Bio-Rad Laboratories, Inc.)-treated TSB with the addition of 25 μM zinc acetate, 25 μM MnCl₂, 1 mM MgCl₂ and 100 μM CaCl₂ in 10-ml glass culture tubes. The cultures were incubated at 37°C with shaking for 18 h. The chrome azurol S (CAS) siderophore assay was performed on the spent media using the modified microplate method as previously reported (40,41).

Whole cell metal quantification

Staphylococcus aureus strains were grown for 18 h overnight in TSB before diluting them to an optical density (A₆₀₀) of 0.05 into 7.5 ml of TSB or Chelex-treated (Bio-Rad Laboratories, Inc.) TSB in 30-ml capacity culture tubes as described previously (31). Cells were allowed to grow with shaking for 8 h. Pre-weighted metal-free 15-ml propylene tubes were used to pellet the cells in a tabletop centrifuge at 4°C (Eppendorf, Hauppauge, NY). Pellets were washed three times with 10 ml of ice-cold PBS. Samples were kept at −80°C or on dry ice until processing.

Cell pellets were acid digested with 2 ml of Optima grade nitric acid (Thermo Fisher Scientific, Waltham, MA) and 500 μl hydrogen peroxide (Sigma–Aldrich, St Louis, MO) for 24 h at 60°C. After digestion, 10 ml of UltraPure water (Invitrogen, Carlsbad, CA) was added to each sample. Elemental quantification on acid-digested liquid samples was performed using an Agilent 7700 inductively coupled plasma mass spectrometer (Agilent, Santa Clara, CA). The following settings were fixed for the analysis: Cell Entrance = −40 V, Cell Exit = −60 V, Plate Bias = −60 V, OctP Bias = −18 V and collision cell Helium Flow = 4.5 ml min⁻¹. Optimal voltages for Extract 2, Omega Bias, Omega Lens, OctP RF and Deflect were determined empirically before each sample set was analyzed. Element calibration curves were generated using ARISTAR ICP Standard Mix (VWR). Samples were introduced by a peristaltic pump with 0.5-mm-internal diameter tubing through a MicroMist borosilicate glass nebulizer (Agilent). Samples were initially up taken at 0.5 rps for 30 s followed by 30 s at 0.1 rps to stabilize the signal. Samples were analyzed in Spectrum mode at 0.1 rps collecting three points across each peak and performing three replicates of 100 sweeps for each element analyzed. Sampling probe and tubing were rinsed for 20 s at 0.5 rps with 2% nitric acid between each sample. Data were acquired and analyzed using the Agilent Mass Hunter Workstation Software version A.01.02.

Murine models of infection

Acute pneumonia model

Experiments were conducted as previously described (42). Briefly, mice were intranasally infected with 2–4 × 10⁷ colony forming units (CFU) in 50 μl of PBS under anesthesia (ketamine and xylazine). Bacterial loads were enumerated at 24 h post infection from bronchoalveolar lavage fluid (BALF) by washing the airway 3 times with 1 ml of PBS, and homogenized lung tissue. Bacterial counts were quantified by serial dilution using CHROMagar S. aureus plates (BD Biosciences).

Skin and soft tissue model

All studies were conducted in accordance with an approved protocol at The University of Kansas Medical Center. Female C57BL/6J mice (Jackson Laboratories) were used in a subcutaneous skin infection model as previously described (43). Briefly, 8-week-old mice were injected with ∼2.1e7 CFU of mid-exponential phase cells suspended in PBS. Mice were imaged daily, and lesion sizes determined using ImageJ. On the final day, mice were euthanized and the lesion plus ∼3 mm of neighboring tissue were homogenized in 1× HBSS 0.2% HAS 10 mM HEPES buffer using an MP Biomedicals Fastprep-24 homogenizer with Lysing Matrix H tubes following the manufacture’s protocol for skin. A sample was taken to determine bacterial titers by dilution plating. Next, the samples were clarified by centrifugation, treated with 1× protease inhibitor (Roche), and frozen at −80°C until processed for cytokines. Cytokines were measured using the BD cytometric bead array mouse flex set protocol on a BD Aria flow cytometry machine and analyzed in FCS express.

Ethics statement

Animal work in this study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the NIH (National Academies Press, 2011), the Animal Welfare Act and US federal law. Protocols were approved by the Institutional Animal Care and Use Committee of Rutgers New Jersey Medical School of Newark, New Jersey, USA, as well as The University of Kansas Medical Center, Kansas City, Kansas, USA.

Statistical analysis

For two group comparisons (controls versus treatment or between bacterial strains), Student’s t-tests were performed. Multiple group comparisons for animal data were performed using an ANOVA with a Kruskal–Wallis test or a Mann–Whitney non-parametric test for two group comparisons. All analyses were conducted with Sigmaplot 11, Microsoft Excel or Prism 9.

Results

Fe starvation decreases aconitase activity in a Fur-dependent manner

Previous work demonstrated that both the absence of Fur in S. aureus or Fe depletion resulted in a redirection of central metabolism where glycolysis is increased, and the TCA cycle is downregulated (24,25). The shift toward fermentative metabolism increases Fe availability, and since the TCA cycle contains Fe–S cluster requiring enzymes, its downregulation could be a way to decrease non-essential Fe usage in an Fe-sparing response (19,44).

We tested the hypothesis that the Fe–S cluster requiring enzyme aconitase (AcnA), which converts citrate to isocitrate in the TCA cycle, would have decreased activity upon Fe limitation. We quantified the activity of AcnA in the parent strain, USA300_LAC, and the isogenic Δfur::tetM (Δfur) mutant. The strain lacking Fur had nearly undetectable AcnA activity, and the phenotype could be genetically complemented (Figure 1A). The divalent metal chelator DIP has a high affinity for Fe(II) and co-culture with DIP results in Fur derepression (44,45). When the wild-type (WT) strain was co-cultured with DIP, the activity of AcnA was significantly decreased (Figure 1B).

Figure 1.

An S. aureus Δfur mutant has decreased aconitase expression. (A) Aconitase activity was quantified in cell-free lysates harvested from the WT (JMB1100) and Δfur (JMB10842) strains carrying pOS-1-plgt or pOS-1-plgt_fur after culture in TSB-Cm medium. (B) Aconitase activity was quantified in cell-free lysates from the WT after culture in TSB ± 250 μM DIP. (C) Aconitase activity was quantified in cell-free lysates harvested from the acnA::Tn (JMB11803) and Δfur acnA::Tn (JMB11804) strains with pEPSA5_acnA after culture in TSB-Cm medium supplemented with 0.25% xylose. (D) Culture optical densities (A₆₀₀) were monitored for the WT, Δfur and acnA::Tn strains in a liquid-defined medium containing amino acids as the sole carbon and energy sources. (E) Culture optical densities (A₆₀₀) of the WT, Δfur and acnA::Tn strains were monitored in a liquid-defined medium containing amino acids supplemented with 10 mM glucose. The data shown represent the average of biological triplicates with standard deviations shown. Error bars are shown for all data but in some cases (panels D and E) are smaller than the symbols used. Student’s two-tailed t-tests were performed on the data and * represents a P-value of <0.05.

We examined whether decoupling acnA transcriptional activity from its native promoter would increase AcnA activity in the fur mutant. To this end, we placed acnA and its ribosomal binding site (RBS) under the transcriptional control of a xylose inducible promoter (xylRO) using the pEPSA5 plasmid. We examined AcnA activity in the extracts of the acnA::Tn and acnA::Tn Δfur strains containing pEPSA5_acnA after culture with xylose. The acnA::Tn Δfur strain had significantly decreased AcnA activity compared with the acnA::Tn strain (Figure 1C). These data demonstrate that acnA expression is altered in a strain lacking fur and this regulation is also independent of its native promoter.

Staphylococcus aureus requires a functional TCA cycle to use amino acids as carbon and energy sources (36,46). We compared the growth of the WT, Δfur and acnA::Tn strains cultured in a chemically defined medium containing amino acids with and without glucose. The Δfur and acnA::Tn strains had severe growth defects in the medium containing amino acids as the sole carbon and energy source, whereas the WT was capable of growth (Figure 1D). All three strains were capable of growth when the medium was supplemented with glucose (Figure 1E). We also found that the Δfur and acnA::Tn strains did not grow on a solid chemically defined medium containing amino acids for carbon and energy. Again, this phenotype was reversed by supplementing the medium with glucose (Supplementary Figure S1).

A null mutation in isrR promotes the growth of a Δfur strain on amino acids

We conducted suppressor analysis to provide insight into the defective growth of the Δfur strain on an amino acid medium. We individually plated fifteen cultures of the Δfur strain on a chemically defined solid medium containing only amino acids as carbon and energy sources. We isolated one colony from each plate, verified the phenotype of improved growth with amino acids and then determined the locations of the single nucleotide polymorphisms by whole genome sequencing. All 15 strains contained a C→T mutation located 604 base pairs upstream of the translation start site of arlR (SAUSA300_1308). This chromosomal location corresponds to tsr25 (recently renamed isrR), which codes for a sRNA (26,27). The mutations correspond to a G→A change within the isrR sequence at the +1 from the transcriptional start site as determined by RNA sequencing (47) and the +3 location as determined by 5′ Rapid Amplification of cDNA Ends (RACE) analysis (27). Henceforth, we refer to this allele as isrR*. The isrR* mutation permitted the growth of the Δfur mutant on both solid and liquid chemically defined amino acid media (Figure 2A, and Supplementary Figure S2). The Δfur isrR* strain also had increased AcnA activity compared with the Δfur mutant (Figure 2B).

Figure 2.

A null mutation in isrR suppresses the amino acid growth defect and increases aconitase (AcnA) activity of a Δfur mutant. (A) Culture optical densities (A₆₀₀) of the WT (JMB1100), Δfur (JMB10842), Δfur isrR* (JMB10495) and acnA::Tn (JMB11803) strains were monitored when cultured on defined medium containing amino acids as carbon and energy sources. (B) Aconitase activity was quantified in cell-free lysates harvested from the WT, Δfur and Δfur isrR* strains after culture in TSB medium. (C) Culture optical densities (A₆₀₀) of the WT, acnA::Tn, Δfur, Δfur ΔisrR (JMB11293) and Δfur ΔisrR acnA::Tn (JMB11806) strains were monitored in liquid-defined medium containing amino acids as carbon and energy sources. (D) Aconitase activity was quantified in cell-free lysates harvested from the WT with pLL39 (JMB1886), SAUSA300_1456::Tn (1456::Tn) pLL39 (JMB11448), fur* 1456::Tn pLL39 (JMB11449), ΔisrR 1456::Tn pLL39 (JMB11395), fur* ΔisrR 1456::Tn pLL39 (JMB11392) and fur* ΔisrR 1456::Tn pLL39_isrR (JMB11393) strains after culture in TSB medium. (E) Aconitase activity was quantified in cell-free lysates harvested from the acnA::Tn, Δfur acnA::Tn (JMB11804) and Δfur ΔisrR acnA::Tn strains containing pEPSA5_acnA after culture in TSB-Cm medium supplemented with 0.25% xylose. The data shown represent the average of biological triplicates with standard deviations shown. Error bars are shown for all data but in some cases (panels D and E) are smaller than the symbols used. Student’s two-tailed t-tests were performed on the data and * represents a P-value of <0.05.

Three findings suggested that the isrR* allele decreased IsrR function and thereby corrected the phenotypes of the Δfur strain. First, we linked a transposon (Tn; SAUSA300_1309::Tn) to the isrR* mutation and then used this strain as a DNA donor for transduction into the Δfur strain. We isolated two classes of transductants. One class corrected the growth of the Δfur mutant on amino acid media, and the second class did not. Sanger sequencing determined that the strains that grew on amino acid medium contained the isrR* allele, and all strains that did not grow contained the WT isrR allele. Second, we constructed ΔisrR and Δfur ΔisrR strains. Deletion of isrR in the Δfur strain permitted growth on solid and liquid defined amino acid medium (Figure 2C, and Supplementary Figure S3). Deletion of isrR in the Δfur strain also increased AcnA activity (Figure 2D). Introduction of the acnA::Tn mutation into the Δfur ΔisrR strain prevented growth in amino acid medium consistent with the hypothesis that the increased growth imparted by the null isrR mutations requires functional AcnA (Figure 2C, and Supplementary Figure S3). Third, we genetically complemented the ΔisrR strains. The pLL39 episome codes for Tet resistance so we could not use the Δfur strain. Instead, we used previously described strains containing a null fur_E11stop allele (fur*) that is genetically linked to a Tn in gene SAUSA300_1456 (18). As expected, the strain containing fur* had greatly reduced AcnA activity. The presence of the ΔisrR mutation in the fur* strain increased AcnA activity and growth on chemically defined amino acid medium, and the phenotypes could be genetically complemented (Figure 2D, and Supplementary Figure S4).

We also examined the effect of the ΔisrR mutation on AcnA activity when acnA was expressed from a non-native promoter. We quantified AcnA activity in cell lysates generated from the acnA::Tn, acnA::Tn Δfur and acnA::Tn ΔisrR Δfur strains containing pEPSA5_acnA after culture in TSB containing xylose (Figure 2E). Again, the activity of AcnA was decreased in the strain lacking Fur, and the introduction of the ΔisrR mutation increased AcnA activity, but not to the level of that in the parent strain.

The isrR* allele decreases isrR transcription

The isrR promoter contains two near consensus Fur box sequences (27). The isrR* mutation resulted in a base change in the Fur box proximal to the isrR transcription start site (Figure 3A). We conducted northern blot analyses to (i) determine if isrR transcript abundances responded to divalent metal starvation, and (ii) verify that it was regulated by Fur in USA300_LAC. The isrR transcript increased in abundance in the WT as we increased the concentration of DIP in the growth medium but was not detected in the ΔisrR strain (Figure 3B, and Supplementary Figure S5). Moreover, the isrR transcript accumulated in the Δfur strain when compared with the unchallenged WT strain, but there was no additional accumulation in the Δfur strain upon co-culture with DIP (Figure 3B). The isrR transcript was not detectable in the ΔisrR Δfur strain. These data are consistent with previous results demonstrating that isrR transcription is regulated by Fur (27) and suggest that Fur is the dominant divalent metal-dependent transcriptional regulator controlling isrR transcription.

Figure 3.

The isrR* mutation suppresses the phenotypes of the Δfur mutant by decreasing isrR transcription. (A) Top panel: Predicted consensus Fur box sequence as determined by RegPrecise. Bottom panel: Consensus Staphylococcal Fur box, isrR distal Fur box, isrR proximal Fur box and proximal Fur box in the isrR* strains with change highlighted in yellow and underlined. The transcriptional start site, as determined by 5′ RACE analysis, is highlighted in green (27). (B) Northern blot analysis of IsrR transcripts using total RNA isolated from WT (JMB1100), ΔisrR (JMB11292), Δfur (JMB10842), Δfur ΔisrR (JMB11293) and Δfur isrR* (JMB10495) strains after culture in TSB media containing 0, 120 or 500 μM DIP. (C) isrR transcript abundance in the WT, Δfur and Δfur isrR* strains after culture in liquid TSB media supplemented with or without 120 μM DIP. Transcript abundance was determined by quantitative PCR. (D) Quantification of transcripts corresponding to isrR after the Δfur ΔisrR strain carrying either pEPSA5_isrR or pEPSA5_isrR* were cultured in TSB medium containing 2% xylose and subsequently rifampicin was added (t = 0) to inhibit transcription. Transcript abundances were normalized to t = 0. (E) Relative fluorescence of the WT strain containing the pOS_pisrR_gfp or pOS_pisrR*_gfp transcriptional reporter after culture in TSB-Cm with or without 250 μM DIP. (F) Aconitase activity in cell-free lysates from the WT strain carrying pEPSA5, pEPSA5_isrR, pEPSA5_as_isrR, pEPSA5_isrR* and pEPSA5_trunk_isrR after culture in TSB-Cm with or without 2% xylose. Panel (B) contains representative northern blot. The data displayed in panels (C)–(F) represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data and * represents a P-value of <0.05.

To evaluate how the isrR* mutation was affecting the expression of isrR, we used northern blot analysis and quantitative real-time PCR (qPCR) to examine the abundances of the transcripts corresponding to isrR in the WT, Δfur and Δfur isrR* strains after culture in the presence and absence of DIP. We could detect the isrR transcript in the Δfur isrR* strain, and the basal expression in unchallenged cells appeared to be increased compared with the WT (Figure 3B). However, isrR did not appear to be significantly induced in the Δfur isrR* strain by co-culture with DIP.

We quantified these findings using quantitative polymerase chain reaction (qPCR). The abundance of the isrR transcript was increased in the WT upon co-culture with DIP (Figure 3C). The isrR transcript was elevated in the Δfur strain when compared with the WT, and it did not further increase in abundance upon co-culture with DIP. The abundance of isrR transcripts in the Δfur isrR* strain was slightly higher than the WT when unchallenged; however, the isrR transcript abundance did not significantly increase in the Δfur isrR* strain upon co-culture with DIP. These data suggest that the isrR* mutation decreases the transcriptional activity of the isrR locus under Fe starvation or that it promotes decreased isrR transcript stability.

We tested the hypothesis that the isrR* mutation was suppressing the defects of the Δfur strain by decreasing IsrR stability. We cultured the Δfur ΔisrR strain containing pEPSA5_isrR or pEPSA5_isrR* in the presence of xylose to induce isrR transcription. We then added rifampicin to inhibit RNA synthesis, isolated samples at different timepoints and quantified isrR transcripts. The transcripts corresponding to isrR and isrR* decayed at similar rates. In fact, the isrR* transcript had a higher relative abundance than the isrR transcript 30 min after transcription was halted (Figure 3D).

We next tested the hypothesis that the isrR* mutation was suppressing the defects of the Δfur strain due to decreased isrR transcription. We generated transcriptional reporter constructs where the isrR or isrR* promoter drove transcription of gfp. The WT strain containing either construct was cultured in TSB with and without DIP, and gfp fluorescence was quantified. Co-culture with DIP significantly increased gfp expression in both strains containing a transcriptional reporter (Figure 3E). However, the strain containing the isrR*_gfp had considerably decreased isrR transcriptional activity upon Fe-depletion compared with the strain expressing isrR_gfp.

We individually placed the expression of isrR, isrR*, antisense isrR (as_isrR) and a truncated isrR that begins two nucleotides downstream from the nucleotide mutated in isrR* (trunk_isrR), under the transcriptional control of xylRO promoter using pEPSA5. Induced expression of isrR, isrR* or trunk_isrR in the WT strain significantly decreased AcnA activity and decreased growth on solid defined media compared with the WT strain containing the empty vector (Figure 3F, and Supplementary Figure S6). The AcnA activity or the growth of WT expressing the as_isrR was not significantly different to that of the WT containing the empty vector. These results suggest that (i) the expression of isrR under metal-replete conditions decreases acnA expression, and (ii) the isrR* mutation does not affect the ability of the isrR* transcript to decrease acnA expression. These findings are consistent with the hypothesis that the isrR* mutation results in decreased transcription of the isrR locus and, thereby, suppresses the phenotypes of the Δfur strain.

The presence of IsrR results in decreased acnA expression

Our results are consistent with a model wherein the absence of Fur or upon Fe ion limitation, isrR is expressed and mediates the repression of acnA expression. We tested the hypothesis that IsrR alters acnA expression post-transcriptionally. We used data from a published RNA-sequencing dataset to observe the length of the acnA 5′ untranslated region (UTR) and estimate that it extends −41 bases from the translational start site under the growth conditions examined (Supplementary Figure S7) (26). We used the program IntaRNA [Freiburg RNA tools (48,49)] to help predict potential interactions between IsrR and the acnA transcript. One potential interaction site was identified that overlayed the 5′ UTR and the first two codons of the coding sequence with a base-pair minimal annealing energy of −14.2 kcal mol^-1. This interaction included a 100% overlap with the Shine Dalgarno sequence (AGGGGG) (Figure 4A).

Figure 4.

IsrR directly influences acnA translation. (A) IntaRNA predicted interaction between IsrR and acnA mRNA transcript. Predicted interaction includes the acnA Shine Dalgarno sequence (RBS) and the second cytosine-rich region (CRR_2) of IsrR. The acnA mRNA AUG start codon is underlined and bolded, and the AcnA coding sequence is in blue. The base indicating the predicted transcriptional start site from Supplementary Figure S7 is in green. (B) Relative fluorescence of the WT (JMB1100) and ΔisrR (JMB11292) strains containing the pOS_plgt_acnA_gfp translational reporter after culture in TSB-Cm media with or without 120 μM DIP. (C) Relative fluorescence of the WT, ΔisrR, Δfur (JMB10842), Δfur ΔisrR (JMB11293) and Δfur::tetM isrR* (JMB10495) strains containing the pOS_plgt_acnA_gfp translational reporter after culture in TSB-Cm medium. (D) Electrophoretic mobility shift assay (EMSA) using 20 000 cpm of radiolabeled IsrR and 0–500 μM of the acnA transcript. (E) EMSA using 20 000 cpm of radiolabeled IsrR and 0–500 μM of the lukH transcript. The data in panels (B) and (C) represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data and * represents a P-value of <0.05. Pictures of representative EMSA assays (n = 2) are displayed in panels (D) and (E).

To evaluate the effect of IsrR on acnA translation, we created a translational reporter where the constitutive lgt promoter drove transcription of a chimeric acnA_gfp generated by fusing the acnA 5′UTR and first two acnA codons in frame with the coding sequence of gfp (pOS_plgt_acnA_gfp). We used this construct to quantify gfp expression in the WT and ΔisrR strains after culture with and without DIP. The expression of gfp was decreased in the WT strain upon co-culture with DIP, but not in the ΔisrR strain (Figure 4B).

We next used the translational reporter to compare gfp expression in the WT, Δfur, Δfur ΔisrR and Δfur isrR* strains. When compared with the WT strain, the expression of gfp was decreased in the Δfur strain (Figure 4C). However, the presence of either the ΔisrR or isrR* alleles reversed the Δfur phenotype. These data are consistent with the hypothesis that IsrR modulates the expression of the acnA_gfp allele in vivo.

We tested the hypothesis that the isrR and acnA transcripts interact in vitro. We performed an EMSA using labeled IsrR and the 5′ of acnA transcript (from −41 to +541). RNA–RNA gel shifts demonstrated that titrating acnA transcript into samples containing IsrR decreased the rate at which IsrR migrated through the polyacrylamide matrix, suggesting a direct interaction between IsrR and the acnA transcript (Figure 4D). Using these data, we estimated that the K_d of IsrR for the acnA transcript is likely <50 nM. IntaRNA does not predict a strong interaction between IsrR and the lukH transcript, which codes for a leukotoxin. IsrR failed to interact with the lukH transcript using the same amounts of IsrR and lukH transcript as used for the acnA transcript EMSA (Figure 4E).

IsrR may interact with the acnA transcript RBS

To further analyze the predicted IsrR–acnA mRNA interaction, we modified the acnA translational reporter to contain base change substitutions in and around the acnA Shine Dalgarno sequence, which are predicted to interact with IsrR, but still allow RBS function (pOS_plgt_acnA1_gfp) (Figure 5A). We hypothesized that these nucleotides were involved in IsrR–acnA mRNA interaction, and base substitutions would affect IsrR binding, thus decoupling acnA_gfp expression from IsrR control. We quantified gfp expression from the acnA_gfp and the mutated acnA1_gfp translational reporters in the WT, Δfur, ΔisrR and Δfur ΔisrR strains. The acnA_gfp behaved as previously demonstrated, and the Δfur mutant had decreased gfp expression, which was restored upon deletion of isrR (Figure 5B). The fluorescence from acnA1_gfp showed no significant differences between the strains examined, suggesting that the nucleotide substitutions in the acnA RBS decreased IsrR-mediated expression control.

Figure 5.

Interactions between the acnA Shine Dalgarno sequence nucleotides and a cytosine-rich region (CRR) of IsrR may influence IsrR-mediated acnA translational repression. (A) Partial sequences of the acnA_gfp and acnA1_gfp translational reporters. Black indicates acnA 5′ UTR, the acnA Shine Dalgarno sequence is underlined, blue indicates the first two codons of acnA, green indicates the start of the gfp sequence and an asterisk above the nucleotide denotes that they are predicted to interact with IsrR. Red nucleotides in the lower sequence indicate substitutions in the acnA1_gfp translational reporter. (B) Relative fluorescence of the WT (JMB1100), Δfur (JMB10842), ΔisrR (JMB11292) and Δfur ΔisrR (JMB11293) strains containing the acnA_gfp or acnA1_gfp translational reporters cultured in TSB-Cm medium. (C) Portions of IsrR and IsrR variant sequences (isrR_C1, isrR_C2 and isrR_C1_C2) that are predicted to interact with acnA mRNA. Blue indicates nucleotides involved in the predicted interaction, including cytosine-rich region one (CRR_1; dark blue) and CCR_2 (light blue). Underlined nucleotides indicate the IsrR CRRs. Red indicates the nucleotide substitutions in the isrR variants. (D) Relative fluorescence of the 1456::Tn pLL39 (JMB11448) (parent), fur* ΔisrR 1456::Tn pLL39 (JMB11392), fur* ΔisrR 1456::Tn pLL39_isrR (JMB11393), fur* ΔisrR 1456::Tn pLL39_isrR_C1 (JMB13983), fur* ΔisrR 1456::Tn pLL39_isrR_C2 (JMB13984) and fur* ΔisrR 1456::Tn pLL39_isrR_C1_C2 (JMB14314) containing the pOS_plgt_acnA_gfp translational reporter. (E) Aconitase activity in cell-free lysates harvested from the strains in panel (D) after culture in TSB medium. The data in panels (B), (D) and (E) represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data, and * represents a P-value of <0.05.

Coronel-Tellez et al. determined that IsrR has CRRs, which may be used to bind mRNA targets (27). IntaRNA predicted individual interactions between the IsrR CRR_1 (Supplementary Figure S8) and CRR_2 (Figure 4A) with the acnA transcript. We set out to determine if CRR_1 or CRR_2 is primarily responsible for driving the interactions between IsrR and the acnA 5′ UTR. We designed three different IsrR variants: one with mutations near CRR_1 (isrR_C1), one with mutations near the CRR_2 (isrR_C2) and one that combined both the CRR_1 and CRR_2 mutations (isrR_C1_C2) (Figure 5C). The mutated isrR variants and the WT isrR were individually cloned into the pLL39 episome and integrated onto the chromosome of the ΔisrR fur* strain (null fur allele). We then used the acnA_gfp translational reporter to quantify the effect of the isrR alleles on gfp expression. The ΔisrR fur* strains carrying isrR or isrR_C1 decreased gfp expression, suggesting that both alleles function to modulate the expression of acnA_gfp (Figure 5D). The expression of acnA_gfp in the strains carrying the isrR_C2 or isrR_C1_C2 alleles phenocopied the strain carrying the empty vector, suggesting that these isrR variants lost the ability to control acnA_gfp expression. We found that isrR transcript levels accumulated to similar levels in each strain, demonstrating that the difference in acnA translational activity did not result from differential IsrR accumulation (Supplementary Figure S9).

We next monitored AcnA activity in cell lysates of the WT and ΔisrR fur* strains carrying the empty vector or the different isrR alleles. The ΔisrR fur* strains carrying isrR or isrR_C1 decreased AcnA activity, suggesting that both alleles function to modulate acnA expression (Figure 5E). The activity of AcnA in the ΔisrR fur* strains carrying the isrR_C2 or isrR_C1_C2 alleles behaved like the ΔisrR fur* strain carrying the empty vector suggesting that these isrR variants lost the ability to control acnA expression. These findings are consistent with the hypothesis that the CRR_2 region of IsrR is utilized to interact with the acnA transcript and instigate translational repression.

We sought further confirmation that IsrR may interact with acnA transcript RBS using CRR_2. We constructed an IsrR variant that contained a single base change in the CRR_2 (isrR_C2a) (Figure 6A). Expression of isrR_C2a in the ΔisrR fur* strain resulted in increased aconitase activity compared with the strain expressing isrR (Figure 6B). The ΔisrR fur* strain containing isrR_C2a also had increased gfp expression from the acnA_gfp transcriptional reporter compared with the strain expressing isrR, suggesting that the one base pair change in CRR_2 was sufficient to lessen IsrR regulatory control over acnA_gfp expression (Figure 6C). We next constructed an acnA translational reporter that contained a compensatory mutation intended to base pair with the mutation introduced in isrR_C2a (pOS_plgt_acnA2_gfp) (Figure 6D). The ΔisrR fur* strain containing isrR decreased expression of acnA_gfp compared with the parent strain, but not expression of acnA2_gfp, suggesting that the one base pair change decreased the ability of IsrR to regulate acnA2_gfp expression (Figure 6E). Importantly, the ΔisrR fur* strain containing the isrR_C2a allele decreased acnA2_gfp expression, demonstrating that introducing a compensatory mutation into isrR regained some control over acnA2_gfp expression. The isrR and isrRC2a transcripts accumulated to similar levels in the ΔisrR fur* strain (Supplementary Figure S9).

Figure 6.

Compensatory interactions between IsrR CRR_2 and the acnA RBS reestablish IsrR-mediated control of aconitase expression. (A) IntaRNA predicted interaction between the pOS_plgt_acnA2_gfp and IsrR_C2a sequences. The mutated bases for both acnA2_gfp and IsrR_C2a are in red, blue indicates the first two codons of acnA, the acnA mRNA AUG start codon is underlined and green indicates the first codon of gfp. (B) Aconitase activity in cell-free lysates harvested from fur* ΔisrR 1456::Tn pLL39 (JMB11392), fur* ΔisrR 1456::Tn pLL39_isrR (JMB11393) and fur* ΔisrR 1456::Tn pLL39_isrR_C2a (14 888) after culture in TSB medium. (C) Relative fluorescence of the strains in panel (B) containing the pOS_plgt_acnA_gfp translational reporter after culture in TSB-Cm medium. (D) Partial sequences of the acnA_gfp and acnA2_gfp translational reporters. Black indicates acnA 5′ UTR, the acnA Shine Dalgarno sequence is underlined, blue indicates the first two codons of acnA, green indicates the start of the gfp sequence and an asterisk above the nucleotide denotes that they are predicted to interact with IsrR. The red nucleotide in the lower sequence indicates the substitution in the acnA2_gfp translational reporter. (E) Relative fluorescence of the strains in panel (B) containing the pOS_plgt_acnA_gfp or pOS_plgt_acnA2_gfp translational reporters after culture in TSB-Cm medium. The data displayed in panels (B), (C) and (E) represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data, and * represents a P-value of <0.05.

IsrR represses the expression of additional genes coding TCA cycle enzymes

We sought to determine whether IsrR also controls the expression of genes coding for additional TCA cycle enzymes. We used IntaRNA and identified predicted IsrR interaction sites in the 5′UTR of sdhC (succinate dehydrogenase), citM (citrate synthase) and citZ (citrate importer) (Supplementary Figures S10–S12). We also identified a site in mqo (malate quinone oxidoreductase) in the 5′ region of the coding sequence (Supplementary Figure S13). We performed RNA–RNA gel shifts to examine if IsrR could interact with the sdh, mqo, citM or citZ transcripts in vitro (Figure 7A–C). Gel shifts show IsrR directly interacts with all four transcripts with an estimated K_d that is <50 nM.

Figure 7.

IsrR interacts with TCA cycle mRNAs and affects expression. (A) EMSA using 20 000 cpm of radiolabeled IsrR and 0–500 nM of the sdhC transcript. (B) EMSA using 20 000 cpm of radiolabeled IsrR and 0–500 nM of the mqo transcript. (C) EMSA using 20 000 cpm of radiolabeled IsrR and 0–500 nM of the citZ or citM transcripts. (D) Activity of succinate dehydrogenase (Sdh) was quantified in cell-free lysates generated from the WT (JMB1100), Δfur (JMB10842), ΔisrR (JMB11292), Δfur ΔisrR (JMB11293) and sdhA::Tn (JMB14375) strains after culture in TSB medium. (E) Activity of malate quinone oxidoreductase (Mqo) in cell-free lysates generated from the WT, Δfur, ΔisrR and Δfur ΔisrR strains. Pictures of representative EMSA assays (n = 2) are displayed in panels (A)–(C). The data displayed in panels (D) and (E) represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data, and * represents a P-value of <0.05.

We next tested the hypothesis that IsrR would negatively regulate the expression of sdh and mqo. We cultured the WT, Δfur, ΔisrR and Δfur ΔisrR strains in TSB and assessed the activities of succinate dehydrogenase (Sdh) and malate quinone oxidoreductase (Mqo) in cell-free lysates (Figure 7D and E). The Sdh and Mqo activities were decreased in the Δfur mutant, and activity was partially restored upon the deletion of isrR. These results suggest that IsrR mediates the repression of both Fe–S cluster and non-Fe using TCA cycle enzymes in response to Fe limitation.

IsrR impacts Fe ion homeostasis

IsrR was demonstrated to downregulate the expression of non-essential Fe utilizing genes including gltA, fdh, miaB and anaerobic nitrate respiration (nasD and narG) (27,50). This work also showed that IsrR is required for growth upon Fe depletion, leading to a model wherein IsrR contributes to Fe sparing upon Fur demetallation. TCA cycle enzymes are a cellular Fe sink, and herein, we demonstrate that IsrR represses the expression of TCA cycle genes in the absence of Fur, which contributes to the proposed Fe-sparing model.

We tested the hypothesis that isrR expression promotes Fe uptake and/or an increase in Fe ions not ligated by macromolecules (also called ‘free Fe’). The antibiotic streptonigrin, when combined with intracellular Fe(II) and a reducing agent, promotes killing by causing double-stranded DNA breaks (3,7). Therefore, increased killing by streptonigrin is correlated with an increased pool of Fe that is not chelated by macromolecules. We assayed streptonigrin sensitivity by spotting streptonigrin on top agar overlays containing the WT, Δfur, ΔisrR or Δfur ΔisrR strains. The Δfur strain displayed increased streptonigrin sensitivity compared with the WT (Figure 8A). This phenotype was expected since a Δfur mutant has derepressed Fe uptake (18). The Δfur ΔisrR had decreased streptonigrin sensitivity when compared with Δfur, suggesting a role for IsrR in increasing streptonigrin sensitivity on a Δfur strain. We confirmed a role for IsrR in increasing streptonigrin sensitivity by inducing the expression of isrR or an antisense isrR (as_isrR) in the WT strain using the pEPSA5 vector. Induced expression of isrR, but not as_isrR, resulted in increased streptonigrin sensitivity (Figure 8B). These data are consistent with the hypothesis that upon Fe limitation and Fur derepression, isrR expression increases the pool of free intracellular Fe.

Figure 8.

IsrR impacts Fe homeostasis. (A) The WT (JMB1100), Δfur (JMB10842), ΔisrR (JMB11292) and Δfur ΔisrR (JMB11293) strains were plated as top agar overlays on solid TSA, followed by spotting 2.5 μg streptonigrin. (B) The WT strain containing pEPSA5_isrR or pEPSA_as_isrR were plated as top agar overlays on solid TSA-Cm with or without 2% xylose, followed by spotting 2.5 μg streptonigrin. For panels (A) and (B), the zones of clearance resulting from streptonigrin growth inhibition was quantified. (C–E) The ratio of ⁵⁶Fe and ²⁴Mg abundances were quantified in whole cells using inductively coupled mass spectrometry (ICP-MS) after culture in Chelex-treated TSB (panel C), TSB (panel D) or TSB supplemented with 50 μM Fe (panel E). The ratio of ⁵⁶Fe/²⁴Mg is displayed for WT, Δfur, ΔisrR and Δfur ΔisrR strains. (F) Siderophore production from spent culture supernatants from the WT, Δfur, ΔisrR and Δfur ΔisrR strains was quantified. The data displayed represent the average of biological triplicates with standard deviations shown. Student’s two-tailed t-tests were performed on the data, and * represents a P-value of <0.05.

We next evaluated the role of IsrR in Fe ion uptake. We quantified total ⁵⁶Fe pools using ICP-MS after growth in (i) TSB, (ii) TSB with 50 μM Fe(II) and (iii) TSB treated with Chelex to decrease the titers of divalent metals (Figure 8C–E). The Δfur, ΔisrR and Δfur ΔisrR strains had increased titers of ⁵⁶Fe compared with the WT in all three media. However, the ΔisrR and Δfur ΔisrR strains had decreased ⁵⁶Fe levels compared with Δfur in Chelex-treated TSB and TSB supplemented with Fe. The intermediate Fe levels of the Δfur ΔisrR suggest that IsrR contributes to the increased Fe levels in strains lacking Fur.

Lastly, we examined whether IsrR has a role in siderophore production and/or uptake. We quantified total siderophore production in the WT, Δfur, ΔisrR and Δfur ΔisrR strains after culture in Chelex-treated TSB. The Δfur strain had increased siderophore production compared with the WT (Figure 8F). The Δfur ΔisrR double mutant strain produced fewer siderophores than the Δfur, but more than the ΔisrR and WT strains. Taken together, these findings verify a role for IsrR in the Fe ion homeostasis and suggest that it has a role in siderophore production.

IsrR and Fur contribute to S. aureus pathogenesis

The connection between S. aureus metabolism and virulence potential is well described. Pathogenesis requires nutritional adaptation to the niche and altered virulence factor production is often linked to nutrient availability changes (51,52). Staphylococcus aureus is one of the most common causes of bacterial pneumonia and the most common pathogen isolated in skin and soft tissue infections (53,54). Host-induced hypoferremia is a commonality in both bacterial lung and skin and soft tissue infections (55,56).

We first used a model of acute pneumonia to define the respective roles of IsrR and Fur in pathogenesis. In both BALF and lung tissue, we observed the importance of IsrR in pathogenesis (Figure 9A and B). While we did not observe a significant impact of Fur in this model, the inactivation of isrR led to a decrease in bacterial survival. In BALF, the inactivation of isrR led to a nearly 40-fold reduction in bacteria; this held true in the lung with a 25-fold decrease.

Figure 9.

IsrR and Fur are important for tissue damage and colonization during infection. (A and B) The WT (JMB1100), Δfur (JMB10842), ΔisrR (JMB11292) and Δfur ΔisrR (JMB11293) strains were tested in a model of acute pneumonia infection. Data represent bacterial counts 24 h after intranasal infection in BALF (panel A) and lung tissue (panel B). (C and D) The WT, Δfur, ΔisrR and Δfur ΔisrR strains were injected subcutaneously into C57BL/6J mice and total lesion size (panel C) and necrosis size (panel D) were monitored over time. (E) Local bacterial titers were quantified four days post-infection. For panels C and D, the data are representative experiment with n = 10, and error bars represent the SEM. Each dot is an individual animal for panels A, B and E, and the bar or line represents the mean. Error bars represent the SEM and may be smaller than symbols. *, ** and *** indicates P< 0.05, P< 0.01, and P< 0.001, respectively, by Mann-Whitney test.

We next sought to determine how the absence of IsrR or Fur would impact skin infection. To this end, we performed a murine model of skin infection and monitored several outcomes. We observed no difference in lesion size between the ΔisrR mutant and WT strain (Figure 9C). By contrast, the lesion sizes significantly decreased for the Δfur and Δfur ΔisrR strains compared with the WT-infected mice. While lesion size reports the overall surface lesion, necrosis size quantifies fully necrotic tissue that forms during infection and is observed as a scab-like structure. The same trends were observed for necrosis between the strains (Figure 9D).

To determine if reduced lesion or necrosis size in the absence of Fur was due to changes in bacterial colonization, bacterial titers at the site of infection were determined at 4 days post-infection. We observed an approximately 3-fold decrease in bacterial titers in all mutant strain-infected mice compared with WT-infected mice at this timepoint (Figure 9E). Since the ΔisrR mutant showed reduced bacterial numbers but not decreased lesion formation, we interpret this to mean that the smaller lesions observed in the Δfur and Δfur ΔisrR strains are not likely due to changes in bacterial titers at the site of infection.

Tissue damage likely results from a combination of bacterial factors and the host’s immune response. Considering the reduced tissue damage in the absence of Fur, we measured cytokine and chemokine levels at the site of infection to provide insight into the immunological changes occurring during infection with fur mutants. We used a panel of proinflammatory cytokines that are important for immune cell recruitment and/or activation. We did not observe a change in KC, MCP-1, GM-CSF or MIP1α between the strains (Supplementary Figure S14). Thus, tissue damage differences did not correlate with any changes in these chemokines and cytokines. IL-1β, G-CSF and IL-6 were decreased in all mutant bacterial infections. In contrast, we observed increased IL-1α but decreased TNFα and G-CSF in a Fur-dependent manner, and these changes correlate with decreased lesion size when Fur was absent. We did not perform statistical analysis on CCL5 (RANTES), but while it was readily detectable in WT-infected mice, most mutant mice had levels below the limit of quantification. The finding that some cytokine levels, but not others, differed when Fur was absent suggests specific immunological changes are occurring during infection in a Δfur mutant. What those changes are in immune cell recruitment or activation and the mechanism by which this occurs will require additional investigation.

Discussion

This study was initiated to understand why a strain lacking Fur has decreased TCA cycle function. Work by others led to the hypothesis that growth in Fe-limiting conditions decreases the levels of Fe-bound Fur, resulting in altered affinity for DNA and derepression of the Fur regulon (11,19). Bioinformatic work led us to predict that the genes that Fur directly regulates are utilized for Fe uptake and a gene for Fe storage (dps) (18). However, we discovered that a Δfur mutant had greatly reduced AcnA activity. We also noted that neither Δfur or acnA::Tn mutants could grow using amino acids as carbon or energy sources, providing us with a phenotype that we could exploit.

We used an unbiased suppressor screen approach to identify IsrR as the Fur-regulated negative regulator of aconitase expression. To our knowledge, this is the first case of a suppressor screen identifying an sRNA through the suppressive effects of a null mutation permitting growth on an otherwise non-permissive growth condition. This finding showcases the power of suppressor screen genetics and highlights the broad impacts that IsrR has on S. aureus physiology.

The suppressor mutations in isrR (called isrR*) allowed for increased growth of the Δfur mutant on defined amino acid media. Interestingly, all fifteen of the isrR* mutant strains, isolated during two separate screening events, contained the same mutation in the operator of isrR, suggesting that there is something unique about this allele. In alternate organisms, small proteins, such as Hfq in E. coli, promote interaction between a sRNA and a target mRNA (57). The finding that suppressor mutations only mapped to the promoter of isrR and not to alternate loci, including hfq, suggests that S. aureus does not require a single protein to facilitate interaction between IsrR and target RNAs. The recent study by Coronel-Tellez et al. found that a strain with a deletion mutation in the gene predicted to code the Hfq homolog did not alter IsrR-dependent regulation of expression (27). Alternatively, there could be more than one chaperone that aids IsrR regulation that share functional overlap.

The isrR operator has two Fur-boxes and the isrR* mutation is in the Fur-box that is proximate to the transcription start site (27). The mutation resulted in significantly reduced isrR transcription upon Fe depletion or in a Δfur mutant. It also resulted in an inability for isrR transcription to be induced in a fur mutant during low Fe. The isrR* mutation did not alter isrR transcript degradation, and strains expressing isrR* or isrR using a non-native promoter resulted in phenotypic similarities, suggesting that isrR* is functional in vivo, but the mutation decreases transcription.

The deletion of isrR increased the AcnA activity of a Δfur mutant, and over-expression of isrR in the WT decreased AcnA activity. We identified a predicted IsrR binding site in the acnA transcript and demonstrated that IsrR interacted with the acnA transcript in vitro. The predicted interaction site overlayed the acnA Shine Dalgarno sequence. Introducing mutations into the Shine Dalgarno sequence of acnA that preserve the RBS but impact predicted base-pairing with IsrR decreased IsrR-mediated control of acnA expression. Additionally, mutations impacting the CRR_2 of IsrR, predicted to interact with the acnA transcript, also decreased IsrR control over acnA expression. Lastly, we were able to change one base pair in the acnA RBS, resulting in decreased IsrR control over acnA expression. The introduction of a compensatory mutation into the CRR_2 sequence of IsrR resulted in a partially corrected ability of IsrR to control acnA expression. While not definitive, these data are consistent with a model wherein IsrR binds the acnA transcript and mediates translational repression through occlusion of the RBS. This model is supported by previous findings where IsrR mediates translational repression of fdhA and gltB2, which have similar predicted IsrR pairing sites as the acnA transcript and don’t appear to trigger mRNA degradation (27).

The introduction of the ΔisrR mutation to the Δfur strain resulted in only partial recovery of AcnA activity. The acnA promoter contains a Fur box terminating at position −179. We are currently trying to determine if Fur directly acts as an activator of acnA transcription. While this manuscript was in revision, a manuscript was published that also demonstrated that IsrR controls acnA expression in S. aureus (58). The authors also demonstrated that AcnA, when not ligating an Fe–S cluster, can bind to mRNA transcripts, directly decreasing the expression of acnA (citB), citZ and citC, which code for TCA cycle enzymes. Further, they demonstrated that IsrR decreases the expression of ccpE, which codes for a transcriptional activator of acnA expression. Fur or IsrR may be required for the expression of the Fe–S cluster synthesis and Fe–S protein maturation systems, resulting in decreased holo-AcnA in the ΔisrR Δfur strain (59). Although we currently do not know why the AcnA activity in the ΔisrR Δfur is lower than that of the WT, it is apparent that acnA expression is complex and is regulated by several factors.

We demonstrated that IsrR binds to additional mRNA transcripts that code for enzymes of the TCA cycle, including sdh, mqo, citM and citZ. Activity assays were used to verify that Mqo and Sdh expression was decreased in a Δfur mutant, and this regulation was, in part, relieved by the deletion of isrR. As we witnessed with AcnA activity, the activities of Mqo and Sdh are lower in the ΔisrR Δfur strain when compared with the WT, suggesting that another factor other than IsrR is regulating expression in the absence of Fur.

The results herein, in combination with previous findings, demonstrate that IsrR represses the expression of mRNA coding the Fe-requiring proteins AcnA, Sdh, GltB2, FdhA and MiaB (27,50). Coronel-Tellez et al. found that an ΔisrR mutant had decreased growth when challenged with DIP on solid media supporting the model wherein IsrR functions to spare Fe in S. aureus (27). To further support this model, we examined intracellular Fe ion pools. As previously witnessed, a Δfur mutant was more susceptible to killing by streptonigrin, suggesting an increased free Fe pool (7). The deletion of isrR lessened this sensitivity of the Δfur mutant but did not return to the levels seen in WT. These data are consistent with a model wherein isrR expression promotes an increased Fe ion pool that is not ligated by macromolecules; however, it is currently unknown if this is the result of decreased expression of Fe-requiring proteins or increased Fe uptake or both phenomena. The expression of isrR in the Δfur mutant also increased total Fe pools when S. aureus was cultured in a rich complex medium under Fe deplete or replete conditions. Production of S. aureus siderophore staphyloferrin B (Sbn) requires citrate synthesized from o-phospho-L-serine and glutamate (60). It is tempting to speculate that IsrR decreases TCA cycle function to decrease glutamate anabolism and promote citrate accumulation that can be used for Sbn synthesis. Consistent with this speculation, the introduction of a ΔisrR mutation into the Δfur mutant decreased siderophore production. It is currently unknown why the ΔisrR strain had increased free Fe and cell-associated Fe compared with the WT strain. The simplest explanation is that the growth media used is unable to fulfill the Fe demands of S. aureus, resulting in partial Fur-dependent derepression of isrR transcription in the WT. In the WT, IsrR functions to decrease the expression of Fe-requiring enzymes, resulting in an overall decreased Fe load, which does not happen in the ΔisrR strain. It is also currently unknown why the Fe-related phenotypes associated with the ΔisrR mutation are dominant over the phenotypes associated with the Δfur mutation. The dominant nature of the ΔisrR mutation could be explained by IsrR positively influencing the expression of one or more Fe uptake systems. The absence of IsrR results in decreased Fe uptake in the ΔisrR Δfur double mutant strain compared with the Δfur strain. Further studies are necessary to parse out the role of IsrR on the regulation of Fe uptake.

So far, IsrR has been shown to regulate ten mRNAs that encompass nitrogen homeostasis (gltB2, narG and nasD), fermentation (fdhA), tRNA modification (miaB) and TCA cycle (acnA, sdhC, mqo, citM and citZ) (27,50). IsrR likely functions to repress the expression of these enzymes to spare Fe for use by alternate proteins essential for fitness; however, it is unclear what these proteins are. Fe–S cluster synthesis is essential in S. aureus, but the essential Fe–S protein(s) remain elusive (7).

The link among Fe, Fur and metabolism has been observed across bacteria, including S. aureus. This relationship was first explained by the discovery of the Fur-regulated sRNA RyhB in E. coli (25,61). RyhB expression results in an Fe-sparing response where non-essential genes coding for Fe-using proteins involved in processes like the TCA cycle (acnB, sdh and fumA), Fe storage (ftnA and bfr), oxidative stress (sodB), respiration (nuo) and Fe–S cluster assembly (isc) are downregulated in response to Fe limitation (25). RyhB homologs have been found in other enterobacteria such as Salmonella, Shigella and Yersinia, and functional analogs have been described in Pseudomonas (PrrF1 and PrrF2) and B. subtilis (FsrA) (24,62–65). As Coronel-Tellez et al. indicated, it is remarkable that IsrR, although not similar in sequence to these alternate sRNA, functions in a homologous manner to decrease the expression of Fe requiring enzymes and processes (27). Moreover, the expression of these alternate sRNAs is controlled by Fur.

The virulence defects of the isrR mutants may suggest a proper Fe sensing and response is required for full host colonization. The importance of Fe homeostasis is highlighted by results showing the colonization defects of strains lacking Fur or IsrR in lung and skin murine models of infection. IsrR was also shown to be required for full lethality of S. aureus in a mouse septicemia model (27). A Δfur mutant in the Newman genetic background had increased exoprotein protein, leukotoxin production, hemolysis production and increased killing of HL-60 cells. However, the Δfur mutant also had decreased survival in a neutrophil-killing assay and in a model of murine pneumonia. Interestingly, the depletion of neutrophils nullified the Δfur mutant’s defective lung colonization, suggesting that the Δfur mutant’s increased susceptibility to neutrophil killing contributed to its inability to colonize lung tissue (5). This contrasts with our results that did not observe a decrease in pathogenesis with the fur mutant. This might be due to the different strains of S. aureus used (USA300_LAC and Newman) and the earlier study relying on a higher infection dose. In the skin and soft tissue model of infection, the ΔisrR mutant had decreased CFU counts, comparable to levels of the Δfur mutants, but there was no significant decrease in the lesion or necrosis sizes compared with WT. The difference in lesion and necrosis sizes between the ΔisrR and the Δfur mutants could be due to a difference in virulence factor regulation. A difference in virulence factor expression could also impact the immune response, affecting the degree of inflammation. Alternatively, the lower CFU counts of the ΔisrR strain might explain the slightly smaller lesion and necrosis sizes compared with WT.

The effect of IsrR on virulence could be indirect and a consequence of metabolite imbalance. Altering metabolic status can impact metabolite pools, which can be sensed by transcriptional regulators that control virulence factor production (51,52). For example, TCA cycle repression could alter GTP and branched-chain amino acid pools, which are sensed by CodY, resulting in altered virulence factor production (66,67). Likewise, IsrR-mediated TCA cycle repression could increase pyruvate titers. Increased pyruvate upregulates virulence factor production in S. aureus through complex regulatory networks involving the Agr, Arl and Sae regulatory systems (68). Further study into the regulatory effects of IsrR should reveal how IsrR contributes to S. aureus virulence and will shed light on the cellular response of S. aureus to host encountered Fe limitation.

Thanks to high throughput sequencing technologies, hundreds of sRNA candidates have been identified in S. aureus. Nevertheless, only 50 sRNAs fit the requirements to be bona fide trans-acting sRNAs, and around fifteen have been associated with their mRNA targets and biological functions (35). The studies in this manuscript have furthered our understanding of one of these bona fide sRNA. The results presented demonstrate that under low Fe growth conditions, Fur derepresses isrR transcription. This results in the expression of isrR, which modulates the expression of genes that function in the TCA cycle (Figure 10). We also demonstrate that both IsrR and Fur contribute to cellular Fe homeostasis and virulence. Future studies will investigate the mechanisms by which IsrR and Fur alter pathogenesis.

Figure 10.

Model for IsrR-dependent regulation of TCA cycle expression. Growth in Fe ion replete conditions results in metalation of Fur and transcriptional repression of isrR, which produces an sRNA. Upon Fe limitation, Fur is demetallated, and isrR is expressed. IsrR forms complexes with the acnA, sdh, citM and citZ mRNA transcripts resulting in decreased expression and decreased carbon flux through the TCA cycle. Decreased expression of TCA cycle enzymes results in an inability to grow using amino acids for carbon and energy. This figure was created with Biorender.com.

Data availability

The data underlying this article are available in this article and in its online supplementary material. Strains and plasmids will be made available upon request.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

National Institute of Allergy and Infectious Diseases (NIAID) [1R01Al139100-01 to J.M.B.]; National Science Foundation (NSF) [1750624 to J.M.B.]; United States Department of Agriculture [NE-1028 to J.M.B.]; Agence Nationale de la Recherche [ANR-20-CE12-0021 to D.L.]; National Institutes of Health (NIH) [1R21AI156251 to Bose laboratory, R21AI153646 to Parker lab, AI172352-01A1 and ES007028 to M.J.M.]; New Jersey Commission on Cancer Research [COCR22RBG005 to Parker lab]. Funding for open access charge: National Science Foundation (NSF) [1750624].

Conflict of interest statement. None declared.