ω-Hydroxyemodin Limits Staphylococcus aureus Quorum Sensing-Mediated Pathogenesis and Inflammation

Abstract

Antibiotic-resistant pathogens are a global health threat. Small molecules that inhibit bacterial virulence have been suggested as alternatives or adjuncts to conventional antibiotics, as they may limit pathogenesis and increase bacterial susceptibility to host killing. Staphylococcus aureus is a major cause of invasive skin and soft tissue infections (SSTIs) in both the hospital and community settings, and it is also becoming increasingly antibiotic resistant. Quorum sensing (QS) mediated by the accessory gene regulator (agr) controls virulence factor production essential for causing SSTIs. We recently identified ω-hydroxyemodin (OHM), a polyhydroxyanthraquinone isolated from solid-phase cultures of Penicillium restrictum, as a suppressor of QS and a compound sought for the further characterization of the mechanism of action. At concentrations that are nontoxic to eukaryotic cells and subinhibitory to bacterial growth, OHM prevented agr signaling by all four S. aureus agr alleles. OHM inhibited QS by direct binding to AgrA, the response regulator encoded by the agr operon, preventing the interaction of AgrA with the agr P2 promoter. Importantly, OHM was efficacious in a mouse model of S. aureus SSTI. Decreased dermonecrosis with OHM treatment was associated with enhanced bacterial clearance and reductions in inflammatory cytokine transcription and expression at the site of infection. Furthermore, OHM treatment enhanced the immune cell killing of S. aureus in vitro in an agr-dependent manner. These data suggest that bacterial disarmament through the suppression of S. aureus QS may bolster the host innate immune response and limit inflammation.

INTRODUCTION

Due to the widespread and seemingly inevitable development of bacterial resistance to antibiotics shortly after their introduction, there is a great need for alternatives or adjuncts to classical antimicrobials (1–3). Along with ongoing efforts to identify novel antibacterial targets, interventions that are not directly bactericidal may prove efficacious. These include approaches aimed at modifying or augmenting the host response, as well as approaches that inhibit bacterial virulence mechanisms and thus limit pathogenesis (1–3). Many pathogenic bacteria coordinate the expression of virulence factors important for invasive infection and pathogenesis through a density-dependent communication system called quorum sensing (QS) (4, 5). Therefore, approaches aimed at disrupting QS hold promise to limit pathogenesis in the host and/or serve as adjuncts to extend the utility of existing antibiotics (4, 6–10).

Skin and soft tissue infections (SSTIs) represent the majority of infections caused by Staphylococcus aureus (11–13), and many of the virulence factors contributing to SSTIs are globally regulated by the accessory gene regulator (agr) (Fig. 1A) (14–16). The agr system utilizes a small secreted autoinducing peptide (AIP) to activate a receptor histidine kinase, AgrC, in the bacterial cell membrane. AgrC phosphorylates the transcription factor AgrA, which in turn activates transcription at the P2 and P3 promoters of the operon. P3 activation drives the production of the effector of the operon, RNAIII, which regulates the expression of >200 virulence genes that contribute to invasive infection (14). S. aureus isolates have one of four agr alleles (agr-I to agr-IV), each encoding factors that secrete a unique AIP (AIP1 to AIP4, respectively) that is detected by a cognate AgrC histidine kinase; isolates from each allele can cause human disease (17, 18). Importantly, we and others have shown that the disruption of agr signaling by mutagenesis, monoclonal antibodies, or host factors limits S. aureus infection and reduces pathogenesis (14, 19–24), demonstrating that agr QS is a robust target for combating invasive S. aureus infection.

FIG 1.

(A) Schematic of the S. aureus accessory gene regulator quorum-sensing system and structure of ω-hydroxyemodin. 1, The agr P2 promoter drives the expression of the four genes of the operon agrBDCA. 2, AgrD is a propeptide that is cyclized to form autoinducing peptide (AIP) and secreted via AgrB. AIPs from the four agr alleles vary in length from seven to nine amino acids, but all contain a five-membered thiolactone ring. 3, Secreted AIP binds to its cognate receptor AgrC, activating its histidine kinase function leading to the phosphorylation of AgrA. 4, AgrA binds to the divergent promoters P2 and P3 as well as the promoters for the transcription of the phenol-soluble modulin (PSM) toxins. 5, P2 drives a positive-feedback loop resulting in the upregulation of the agr operon, whereas P3 drives the transcription of the effector molecule RNAIII. RNAIII leads to the upregulation of virulence factors that contribute to invasive infection. (B) Structure of ω-hydroxyemodin (OHM). Molecular mass, 286.24 Da.

Recently, we reported on a synthetic small-molecule inhibitor of S. aureus QS called savirin, providing proof of principle that small-molecule-mediated inhibition of QS can be efficacious against S. aureus in vivo (25). Natural products also represent a wealth of bioactive compounds, as approximately 65% of the antibacterials introduced in the last 30 years are natural products or compounds designed based on a lead natural-product pharmacophore (26). Therefore, we extended our search for inhibitors of QS to natural products and in doing so identified a series of polyhydroxyanthraquinones isolated from cultures of the fungus Penicillium restrictum that inhibited S. aureus QS (27). Among these, ω-hydroxyemodin (OHM) (Fig. 1B) demonstrated the most potent activity for inhibiting agr-I signaling. Therefore, we sought to characterize OHM as an inhibitor of agr signaling and to evaluate its efficacy in vivo.

Approaches aimed at augmenting the host response and those aimed at inhibiting bacterial virulence mechanisms provide alternatives to conventional antibiotic therapy. However, these two approaches are not mutually exclusive, and because many S. aureus virulence factors antagonize the host innate immune response, we proposed that inhibiting bacterial virulence would itself augment host defense. Specifically, we postulated that small-molecule-mediated disruption of S. aureus QS-dependent virulence would not only limit pathogenesis but would also reduce inflammation and result in enhanced bacterial clearance. Here, we report that OHM inhibits S. aureus QS by isolates from all four agr alleles at concentrations that are noncytotoxic to S. aureus or eukaryotic cells. Mechanistically, OHM inhibits agr activation by binding directly to AgrA and blocking binding to agr promoter DNA. Importantly, as predicted, OHM limits tissue damage and inflammation and promotes bacterial clearance in a mouse model of S. aureus SSTI. In addition, OHM promotes the killing of agr⁺, but not agr-negative, S. aureus by both mouse macrophages and human polymorphonuclear leukocytes (PMNs), and it limits neutrophil lysis caused by agr-regulated S. aureus-secreted virulence factors. This is the first report of a polyhydroxyanthraquinone with in vivo efficacy against S. aureus QS-dependent virulence. In addition, these data demonstrate that antivirulence approaches can limit disease by disarming the bacteria while concurrently bolstering host innate defense.

MATERIALS AND METHODS

Ethics statements.

The 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 (97), the Animal Welfare Act, and U.S. federal law. The protocol was approved by the Institutional Animal Care and Use Committee of the University of New Mexico Health Sciences Center. The polymorphonuclear cells were isolated from whole blood samples from consenting healthy human volunteers, according to the protocol (no. 11-005) approved by the Human Research Protections Office of the University of New Mexico institutional review board.

Bacterial strains and growth conditions.

Methicillin-resistant S. aureus (MRSA) strain USA300 LAC (agr-I) was provided as a generous gift from Frank DeLeo (Rocky Mountain National Laboratories, National Institutes of Health, Hamilton, MT). S. aureus strains AH1677 (agr-I), AH430 (agr-II), AH1747 (agr-III), and AH1872 (agr-IV) expressing yellow fluorescent protein (YFP) under the control of the agr::P3 promoter were previously described (28). Staphylococcus epidermidis strain AH3408 (agr-I) expressing superfolder green fluorescent protein (sGFP) under the control of the agr::P3 promoter was also previously described (29). S. aureus strains AH3469 (AgrC wild type [WT]) and AH3470 (AgrC R238H) are described below. Unless otherwise noted, the bacteria were cultured at 37°C and 220 rpm, with at least a 5:1 air-to-culture ratio in Trypticase soy broth (TSB) (Becton, Dickinson and Company, Sparks, MD). Early exponential-phase bacteria were prepared as described previously (30). The frozen stocks were maintained at −80°C in TSB supplemented with 10% glycerol. The bacteria were enumerated by serial dilution and plating onto Trypticase soy agar containing 5% sheep blood (Becton, Dickinson and Company), followed by overnight incubation at 37°C. The limit of detection was 2 log₁₀ CFU.

agr::P3 promoter activation assays.

Overnight cultures of S. aureus agr::P3 reporter strains were grown in TSB supplemented with 10 μg/ml chloramphenicol (Cam). The cultures were diluted 1:250 into fresh TSB with Cam, and 100-μl aliquots were transferred to 96-well microtiter plates (Costar 3603; Corning, Tewksbury, MA) prefilled with 100 μl of medium and a 2-fold serial dilution series (200 to 0.1 μM) of OHM. OHM was purified from solid-phase cultures of P. restrictum, as described previously (27), and was >95% pure, as measured by ultrahigh-performance liquid chromatography (UPLC) (see Fig. S1 in the supplemental material). After mixing in the microtiter plate, the effective dilution was 1:500, and the final OHM concentration ranged from 100 to 0.05 μM, with a final dimethyl sulfoxide (DMSO) concentration of 0.1% (vol/vol) in all wells. Four dilution series were prepared for each reporter; in addition, 4 mock DMSO dilution series were included for each reporter strain. The microtiter plates were incubated at 37°C with shaking (1,000 rpm) in a Stuart SI505 incubator (Bibby Scientific, Burlington, NJ) with a humidified chamber. Fluorescence (top reading of 493 nm excitation, 535 nm emission, and gain of 60) and optical density at 600 nm (OD₆₀₀) readings were recorded at 30-min increments using a Tecan Systems (San Jose, CA) Infinite M200 plate reader.

S. epidermidis AH3408 (agr-I::P3-sGFP) was cultured overnight in TSB supplemented with 10 μg/ml erythromycin (Erm). To collect exogenous S. epidermidis AIP1 peptide, the spent medium was centrifuged at 3,000 × g, passed through a 0.2-μm HT Tuffryn membrane (Pall, Port Washington, NY), and stored at −20°C until use. An overnight culture of AH3408 was diluted 1:200 into 500 μl of TSB (broth) or TSB with 10% spent medium containing 5 μg/ml OHM or DMSO (vehicle). The cultures were incubated for 24 h at 37°C, centrifuged, and resuspended in 10% formalin fixative for 1 min. The cultures were washed twice by centrifugation and resuspended in phosphate-buffered saline (PBS). The mean channel fluorescence (MCF) of sGFP was analyzed using an Accuri C6 flow cytometry system (BD Biosciences, San Jose, CA). The data were normalized to the broth cultures containing no exogenous AIP1.

Quantitative PCR.

For transcriptional quantification of mouse mRNA, 2.25-cm² sections of skin including and surrounding the abscess were excised, minced, and stored in RNAlater (Qiagen, Valencia, CA) at −20°C until use. mRNA was purified using the RNeasy kits (Qiagen), and cDNA was generated using a high-capacity RNA-to-cDNA kit (Applied Biosystems, Foster City, CA). Quantitative PCR was performed using an ABI 7900HT real-time PCR system with TaqMan Gene expression master mix, according to the manufacturer's directions (Applied Biosystems). Predesigned primer and probe sets (Integrated DNA Technologies, Coralville, IA) were used for the quantitation of mouse il-6, il-1β, tnfα, nlrp3, and hprt. The data are represented as the fold increase relative to hprt compared to that in uninfected tissue.

For the quantification of S. aureus gene transcription, 500-μl cultures at 2 × 10⁷ CFU/ml of LAC and/or LACΔagr were grown in TSB at 37°C, with aeration, for the indicated times with 50 nM exogenous AIP1 (Biopeptide Co., Inc., San Diego, CA) and treatments (vehicle versus OHM), as indicated in the appropriate figure legends. The bacteria were stored at −20°C in RNAprotect cell reagent, according to the manufacturer's recommendations (Qiagen), until the RNA was purified as previously described (25). cDNA generation and quantitative PCR (qPCR) were performed as described above for eukaryotic qPCR. The primer and probe sets for the quantification of the S. aureus genes are listed in Table S1 in the supplemental material.

Rabbit red blood cell lysis assay.

The assay was performed as previously described (31). Briefly, LAC was cultured in 5 ml of TSB for 8 h with the indicated treatments, centrifuged, and the supernatants were filtered through a 0.2-μm HT Tuffryn membrane (Pall). Serial 2-fold dilutions of the supernatant were incubated at 37°C for 1 h in a 4% solution of rabbit red blood cells (rRBCs). Lysis was assessed spectrophotometrically at OD₄₅₀. The data were analyzed by nonlinear regression fit to a four-parameter logistic curve and represented as the HA₅₀, which equals 1/the dilution required for 50% complete lysis.

AgrC constitutive reporter assay.

The agrBDCA operon was amplified from strain LAC using the primers AgrB+RBS 5′–KpnI (GTTGGTACCCAGTGAGGAGAGTGGTGTAAAATTG) and AgrA 3′-SacI (GTTGAGCTCCTTATTATATTTTTTTAACGTTTCTCACCGATG) and ligated into pRMC2 (32). To make a variant with constitutive AgrC activity, we chose the AgrC R238H mutation, which was previously shown to have similar activity in the presence and absence of the AIP2 inhibitor and maximal activity in the absence of AIP1 (33). The AgrC R238H variant was generated by the QuikChange (Agilent Technologies) site-directed mutagenesis method, using the primers AgrC R238H fwd (CAACGAAATGCGCAAGTTCCATCATGATTATGTCAATATC) and AgrC R238H rev (GATATTGACATAATCATGATGGAACTTGCGCATTTCGTTG). To build a destination strain for assessing alpha-hemolysin production, we selected an agrC transposon mutant (NE873) from the Nebraska Transposon Mutant Library (34) and integrated the pLL29 plasmid at the phage 11 attachment site to confer tetracycline resistance (35). The above-described pRMC2 constructs were transformed into this strain to make reporters AH3469 (AgrC WT) and AH3470 (AgrC R238H). To test OHM, AH3469 and AH3470 were grown overnight with 10 μg/ml Cam and were diluted 1:500 into 5 ml of fresh medium with 10 μg/ml Cam and 0.025 μg/ml anhydrotetracycline. AIP2 control, OHM, or DMSO (vehicle) was added to each strain at the concentrations indicated. The cultures were grown at 37°C and shaken at 220 rpm for 6.5 h. The bacteria were pelleted by centrifugation, and the alpha-hemolysin-containing supernatants were passed through a 0.2-μm HT Tuffryn membrane (Pall). Rabbit red blood cell lysis assays were conducted as above but with an rRBC concentration of 1% and 25% supernatant (vol/vol) to yield complete lysis. The values are presented as the mean relative lysis compared to that with vehicle treatment.

Eukaryotic cytotoxicity.

A549, HEK293, or HepG2 cells were seeded in a 96-well tissue culture plate at 2.5 × 10⁴ cells per well and incubated at 37°C with 5% CO₂. XTT [2,3-Bis(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide] and phenazine methosulfate (PMS) were purchased from Sigma-Aldrich (St. Louis, MO). The XTT assay was previously described (36). After 24 h, the spent medium was removed, and fresh medium containing the indicated drug concentrations or vehicle was added to the cells and incubated for an additional 24 h. To avoid potential interference with the absorbance readings due to the red color of OHM, drug-containing medium was replaced with 100 μl of 0.3 mg/ml XTT with 0.015 mg/ml PMS in Hanks' balanced salt solution (HBSS) and incubated for 1 h. Cell viability was assessed by the metabolic reduction of tetrazolium measured at OD₄₉₀. The data are presented as the percentage of viable cells compared to that in the vehicle control.

EMSA and flow cytometry-based AgrA_C promoter binding assays.

Escherichia coli expressing the AgrA C-terminal DNA binding domain (AgrA_C) along with a 6-histidine tag was generously provided by Chuan He (University of Chicago, Chicago, IL) and purified as previously described (37). Electrophoretic mobility shift assays (EMSA) were performed as previously described (25), with purified AgrA_C and agr P2 promoter, and a 16-bp duplex DNA probe with a 3′ 6-fluorescein (P2-FAM) (Integrated DNA Technologies, Coralville, IA). The duplex DNA contained the high-affinity LytTR binding site located in both agr P2 and P3 promoters (38). Briefly, 2 μM AgrA_C was incubated for 10 min at room temperature (RT) with vehicle or the indicated concentrations of OHM in Tris-acetate-EDTA (TAE) buffer with 10 mM dithiothreitol (DTT). Next, 20 ng of P2-FAM DNA probe was added and incubated for an additional 10 min. The reaction mixtures were loaded onto a 10% PAGE gel and run at 50 V in the dark for 20 min. DNA migration was assessed by imaging on a FluorChem R system (ProteinSimple, Santa Clara, CA).

For the flow-based AgrA_C promoter binding assays, AgrA_C was biotinylated (AgrA_C-BTN) using a Thermo Scientific EZ-Link sulfo-NHS-LC-biotin kit (Thermo Scientific, Rockford, IL), according to the manufacturer's directions. AgrA_C-BTN was immobilized on 1-μm-diameter Dynabeads MyOne streptavidin T1 (Life Technologies, Grand Island, NY) (AgrA_C-SA), and the beads were suspended in PBS. DNA probe (P2-FAM) was added at a final concentration of 1.6 μM, along with equimolar competing unlabeled P2, vehicle control, or OHM at the indicated concentrations. OHM-mediated inhibition of AgrA_C-SA binding to DNA probe was measured as decreased mean channel fluorescence (MCF) compared to that of the vehicle control using an Accuri C6 flow cytometry system (BD Biosciences).

In silico docking on AgrA_C.

In silico docking calculations were performed using the Macintosh binary executable of AutoDock Vina (39). OHM was docked onto the B subunit of the AgrA_C crystal structure (RSCB Protein Data Bank [www.pdb.org], PDB ID 4G4K) (40, 41) stripped of heteroatoms. The search box was restricted to the C-terminal region of AgrA_C, as described for 9H-xanthene-9-carboxylic acid (40). Based on initial observations suggesting that OHM bound to the pocket between the side chains of His200, Agr218, Tyr229, and Val232, additional calculations were run in which the size of search box was varied and the side-chain torsion angles for different combinations of residues in the region were allowed to be flexible. The reported docking solution was obtained by allowing flexibility in the side chain torsion angles for His200, Agr218, Tyr229, and Val232 and by using a search box that was large enough to include both the pocket bounded by the side chains of His200, Agr218, Tyr229, and Val232 and the groove between Val232 and Lys236. Molecular modeling images were prepared using PDB ID 3BS1 and PyMOL (PyMOL molecular graphics system, version 1.5.0.4; Schrodinger, LLC).

Surface plasmon resonance analysis.

To overcome the potential interference for the oxidative inactivation of AgrA_C during surface plasmon resonance (SPR) analysis, the oxidation-resistant C199S mutation was introduced into the AgrA_C expression construct, as previously described (37), using the QuikChange II XL kit (Agilent Technologies). His-tagged AgrA_C-C199S was purified as described previously (25) but without the addition of Tris(2-carboxyethyl)phosphine (TCEP) or DTT during purification.

SPR binding and kinetics analyses were performed on a Biacore X100 instrument (GE Healthcare, Pittsburgh, PA) and evaluated with Biacore X100 evaluation software (version 1.0). His-tagged AgrA_C-C199S was immobilized at 10 μg/ml in PBS on a nitrilotriacetic acid (NTA) biosensor with the NTA reagent kit (GE Healthcare). For binding studies, OHM (analyte) was dissolved in running buffer (PBS, 5% DMSO [pH 9]) and applied at a flow rate of 30 μl/min with a 180-s contact time and 300-s dissociation time. The data were fit to a 1:1 binding model after the subtraction of blank injections and the removal of injection spikes from the sensorgrams. NTA biosensor chips were regenerated with the following sequence: two 60-s washes with 350 mM EDTA, a 60-s wash with PBS, and a 60-s wash with 500 mM imidazole, followed by a final 60-s wash with PBS. The analyses were performed at 25°C.

Mouse model of skin and soft tissue infection.

The mouse model of skin and soft tissue infection was previously described and was implemented with minor modifications (42). Early exponential-phase S. aureus strain LAC was diluted into USP-grade saline (Braun, Irvine, CA) to deliver 5 × 10⁷ to 7 × 10⁷ CFU per mouse. Aliquots of OHM were diluted in 0.5% hydroxypropylmethylcellulose (HPMC) (pH 11) to deliver 0.2 mg/kg per mouse (∼5 μg). The mice were anesthetized with 3% isoflurane at 3 liters/min. The bacteria and OHM were mixed 1:1 immediately before subcutaneous injection into the right flank, at a total volume of 50 μl. The mice were weighed prior to infection and every day postinfection. Additionally, the injection sites were photographed daily to determine the abscess and ulcer areas with ImageJ analysis (43). On day three or day seven postinfection, each mouse was euthanized by CO₂ asphyxiation, and a 2.25-cm² section of skin surrounding the abscess was excised. The tissue was mechanically homogenized and serially plated on sheep blood agar to determine the bacterial burden. The tissue homogenates were stored at −80°C until they were rapidly defrosted at 37°C for cytokine analysis. On day 3 postinfection, the homogenate was centrifuged at 12,500 × g for 10 min and the clarified supernatant analyzed with a custom-designed multiplex assay (Merck KGaA, Darmstadt, Germany) using a BioPlex 200 with BioPlex Manager software (Bio-Rad, Hercules, CA). The abscess tissues collected for hematoxylin and eosin (H&E) staining were fixed overnight in 10% formalin and embedded in paraffin. Five-micrometer sections were then stained with H&E and imaged using an Olympus IX70 microscope (Olympus, Center Valley, PA).

Alpha-hemolysin quantification.

For the detection of alpha-hemolysin (Hla) in abscess tissue homogenates by Western blot assay, clarified homogenates were rapidly thawed and an aliquot electrophoresed on a 16% Tris-glycine SDS-PAGE gel (Life Technologies, Grand Island, NY) before transfer to a polyvinylidene fluoride membrane. The membranes were blocked for 1 h at RT, using TBST (20 mM Tris [pH 7.5], 150 mM NaCl, 0.1% Tween 20) with 5% nonfat dry milk. Hla was detected using anti-Hla antibody (ab15948; Abcam, Cambridge, MA) at a 1:1,000 dilution and alkaline phosphatase-conjugated secondary antibody. The immunoreactive bands were developed with nitroblue tetrazolium (NBT)/5-bromo-4-chloro-3-indolyl-phosphate (BCIP) (Thermo Scientific) and intensity measured using a FluorChem R system and the AlphaView software (ProteinSimple). Relative intensity was calculated as the measured intensity divided by the total protein concentration based on absorbance at 280 nm.

Mouse macrophage killing of S. aureus.

Murine macrophage cells (RAW 264.7) were maintained at 37°C in 5% CO₂ in high-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, 10 mM HEPES, with 100 U/ml penicillin and 100 μg/ml streptomycin. Twenty-four hours prior to the experiments, the RAW cells were washed with PBS and the medium replaced with DMEM, as described above but with 2% FBS without antibiotics. Early exponential-phase LAC or LACΔagr was cultured in 3 ml of TSB at 2 × 10⁷ CFU/ml at 37°C, with aeration for 5 h, with 50 nM exogenous AIP1 (Biopeptide Co.) and 5 μg/ml OHM or DMSO (vehicle). The bacteria were centrifuged, washed in PBS, sonicated, and suspended at 1 × 10⁸ to 2 × 10⁸ in DMEM but with 1% FBS without antibiotics. The bacteria were opsonized overnight at 4°C with rabbit anti-S. aureus IgG at 100 μg/ml (catalog no. YVS6881; Accurate Chemical & Scientific Co., Westbury, NY). The RAW cells were washed with PBS and suspended at 2 × 10⁷ cells/ml in DMEM with 1% FBS without antibiotics and combined with opsonized bacteria at a multiplicity of infection (MOI) of 1:1. The cells were centrifuged at 500 × g for 3 min to initiate contact and incubated at 37°C in 5% CO₂ for 1 h to allow phagocytosis. Lysostaphin (catalog no. L-0761; Sigma-Aldrich) was added at 2 μg/ml for 15 min to kill extracellular bacteria and then removed by centrifugation and replacement with fresh medium. Half of the samples were immediately processed for CFU determination, and the other half were incubated for an additional 4 h before CFU enumeration. Intracellular bacteria were enumerated by preliminary dilution into PBS–0.1% Triton X-100, followed by sonication and plating onto blood agar.

Human PMN assays.

PMNs were purified from normal healthy venous blood, as described by Nauseef (44). The purified PMNs were suspended in HBSS without divalent cations at ≤3 × 10⁷ cells/ml and kept on ice until use.

PMN phagosomal killing of S. aureus was conducted as previously described, with the following alterations (45). Prior to opsonization, early exponential-phase LAC or LACΔagr was cultured in 3 ml of TSB at 2 × 10⁷ CFU/ml at 37°C, with aeration for 5 h, with 50 nM exogenous AIP-I (Biopeptide Co.) and 5 μg/ml OHM or DMSO (vehicle). The bacteria were centrifuged, washed in PBS, and opsonized at 5 × 10⁶ CFU/ml in HBSS, with divalent cations supplemented with 20 mM HEPES, 1% human serum albumin (HSA), and 10% pooled human serum. Following a 20-min incubation with tumbling at 37°C, the bacteria were pelleted, washed in PBS, and resuspended in HBSS with divalent cations supplemented with 20 mM HEPES. The PMNs and opsonized bacteria were combined at an MOI of 1:1 and incubated for 10 min at 37°C. The extracellular bacteria were removed by centrifugation at 500 × g for 5 min, followed by the resuspension of infected PMNs in HBSS with divalent cations supplemented with 20 mM HEPES and 1% HSA. The infected PMNs were incubated at 37°C for 120 min, and the aliquots were removed at 0, 30, 60, and 120 min. The aliquots were diluted into PBS–0.1% Triton X-100 to lyse the cells and then serially diluted and plated on blood agar for CFU enumeration.

The lysis of PMNs by the S. aureus supernatant was conducted as previously described, with minor modifications (25). Briefly, LAC was cultured in 3 ml of TSB for 5 h with 5 μg/ml OHM or vehicle, centrifuged, and the supernatants were filtered through a 0.2-μm HT Tuffryn membrane (Pall). The supernatants were stored at −80°C and thawed on ice prior to use. The PMNs were washed with PBS and resuspended in RPMI supplemented with 10 mM HEPES and 1% HSA. PMNs at a density of 3 × 10⁶ cells/ml in 100 μl were added to 100 μl of RPMI, RPMI with 10% TSB (vol/vol), or RPMI with 10% S. aureus supernatant prepared as described above. The PMNs were incubated at 37°C and 5% CO₂ for 2 h. Following incubation, the supernatants were collected by centrifugation at 3,000 × g for 5 min and assessed for lactate dehydrogenase (LDH) release, according to the manufacturer's specifications (CytoTox 96 nonradioactive cytotoxicity assay; Promega Co., Madison, WI). Triton X-100 was added at a final concentration of 0.1% (vol/vol) as a 100% lysis control, while cell-free RPMI with 5% TSB served as a blank. The data are normalized to 100% lysis control.

Statistical analysis.

Statistical evaluations were performed using GraphPad Prism version 5.04. The in vitro data were analyzed by the two-tailed Student's t test, and the in vivo data were analyzed by the Mann-Whitney U test for nonparametrics. The results were considered significantly different at a P value of <0.05.

RESULTS

ω-Hydroxyemodin is a universal inhibitor of S. aureus quorum sensing.

Our original report focused on OHM inhibition of QS using an S. aureus strain derived from the USA300 agr-I isolate LAC (27). However, isolates from all four agr alleles contribute to disease in humans. Therefore, to be of maximum utility, a QS inhibitor must antagonize QS by all four agr alleles. To address this, we assessed the ability of OHM to inhibit quorum sensing by isolates of all four agr types using reporter strains expressing yellow fluorescent protein (YFP) under the control of the agr::P3 promoter. OHM inhibited QS by all four agr types at concentrations that do not impact bacterial growth (Fig. 2A). As expected based on the inhibition of agr::P3 promoter activation, OHM decreased the transcription of the agr effector RNAIII and agr-regulated virulence factors, including phenol-soluble modulin alpha (psmα) and alpha-hemolysin (hla) (see Fig. S2A in the supplemental material). OHM also inhibited the production of Hla, as demonstrated by the red blood cell lysis assay (see Fig. S2B in the supplemental material). Importantly, at concentrations required for agr inhibition, OHM was nontoxic to human alveolar (A549), kidney (HEK293), and hepatocyte cell lines (Fig. 2B). Therefore, these data demonstrate that at concentrations that are noncytotoxic to eukaryotic cells, OHM is a universal inhibitor of S. aureus QS.

FIG 2.

ω-Hydroxyemodin inhibits S. aureus quorum sensing by all four agr alleles. (A) Effect of OHM on agr::P3 promoter activation (open symbols) and cell growth (closed symbols), measured by flow cytometry and OD₆₀₀, respectively, for agr-I (red circles), agr-II (blue squares), agr-III (green triangles), and agr-IV isolates (black diamonds). (B) Percent cell viability of A549 (blue circles), HEK293 (red squares), and HepG2 (green triangles) cells measured by XTT assay after 24 h of incubation with the indicated concentrations of OHM. The dashed vertical lines indicate the concentration used for the in vitro assays. Data are shown as means ± SEM. The experiments were performed in triplicate or quadruplicate.

ω-Hydroxyemodin antagonizes AgrA function.

The ability of OHM to antagonize QS by all S. aureus agr alleles pointed to a target that is well conserved in the system. Therefore, we first focused on AgrC, the receptor histidine kinase activated by AIP binding. To determine whether OHM disrupted AgrC activation, we tested OHM for the inhibition of agr-mediated Hla expression determined by the lysis of rabbit RBCs, using an agr-I isolate expressing constitutively active AgrC (R238H [33]). Whereas the addition of inhibitory AIP (AIP2) reduced Hla expression by S. aureus expressing wild-type (WT) but not constitutively active AgrC, OHM inhibited Hla expression by both isolates (Fig. 3A). These results support a mechanism of action whereby OHM intracellularly inhibits agr signaling, downstream of AgrC activation.

FIG 3.

ω-Hydroxyemodin inhibits AgrA binding to promoter DNA. (A) Effect of OHM on agr::P3 promoter activation assessed by rabbit red blood cell lysis for AgrC-WT isolate AH3469 and AgrC-R238H isolate AH3470. The data are the mean relative lysis ± standard error of the mean (SEM) compared to the vehicle control. The experiments were performed in triplicate. ns, not significant. (B) Space-fill model of AgrA_C showing potential binding site for OHM. Inset, ball-and-stick representation of OHM binding site. (C) Flow cytometric bead-based assay to determine the effect of OHM on the binding of P2-FAM to biotinylated AgrA_C immobilized on streptavidin beads (SA beads). Unlabeled P2 binding to immobilized AgrA_C in competition with P2-FAM was included as a specificity control. The data are the mean ± SEM (n = 3). (D) Surface plasmon resonance analysis of OHM binding to 6-His-tagged AgrA_C-C199S immobilized on a nitrilotriacetic acid biosensor. The 50% inhibitory concentration (IC₅₀) was calculated as described in Materials and Methods. K_d, dissociation constant. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P ≤ 0.0001 by Student's t test.

The response regulator AgrA functions downstream of AgrC, and we and others have shown that small molecules that target AgrA disrupt QS (25, 40). Therefore, to further address the mechanism of action of OHM, we evaluated potential OHM binding sites on the crystal structure of the C-terminal AgrA DNA binding domain (AgrA_C) (39, 40). The most favorable binding site for OHM was near the AgrA_C-DNA interface (Fig. 3B). Docking studies positioned OHM in a pocket between the side chains of H200 and Y229, with Y229 recently identified as a major contributor to maximal AgrA activity (46), and three residues, R218, S231, and V232, which make direct interactions with bound DNA in the AgrA-DNA crystal structure (38). Given this, together with observations that OHM is within hydrogen bonding distance of R218 and that naturally occurring mutations at R218 result in agr-negative phenotypes (47), we predicted that OHM would inhibit AgrA binding to promoter DNA. To test this, we expressed AgrA_C and measured binding to fluorescently labeled duplex agr promoter DNA encompassing the high-affinity binding site located in the agr P2 and P3 promoters (P2-FAM). As expected, OHM demonstrated a dose-dependent inhibition of AgrA_C binding to agr promoter DNA by an electrophoretic mobility shift assay (EMSA) (see Fig. S3A in the supplemental material). In addition, we developed a bead-based assay to measure transcription factor binding to target DNA using flow cytometry. Biotinylated AgrA_C was immobilized on streptavidin beads (SA beads), and the binding to promoter DNA was measured by flow cytometry. As expected, OHM again demonstrated a dose-dependent inhibition of AgrA_C binding to agr promoter DNA (Fig. 3C). Furthermore, OHM bound directly to immobilized AgrA_C, as shown by surface plasmon resonance (SPR) analysis (Fig. 3D). Finally, because the amino acid sequence in the predicted OHM binding site of S. aureus AgrA is highly conserved with that of S. epidermidis AgrA, we postulated that OHM would also inhibit agr signaling by S. epidermidis. As expected, OHM significantly inhibited agr activation by agr-I S. epidermidis (see Fig. S3B in the supplemental material). Together, these data strongly suggest that OHM inhibits agr signaling by binding to AgrA and blocking AgrA function.

To begin to address the specificity of OHM for agr inhibition, we used qPCR to evaluate the effects of OHM on transcription of a series of agr- and non-agr-regulated genes involved in virulence, the stress response, metabolism, and drug efflux and resistance (14, 25, 48, 49) (Table 1; see also Fig. S4 in the supplemental material). With respect to virulence genes, OHM treatment resulted in a slight yet nonsignificant increase in the transcription of spa, which encodes protein A and is negatively regulated by agr (14). In contrast, the expression of the enterotoxin gene set7 decreased with OHM in LAC but not LACΔagr, and OHM had no effect on the expression of the saeR component of the SaeRS virulence regulator. Likewise, the transcription of genes involved in the stress response (asp23, crtM, and clpB) was not altered by OHM, suggesting that OHM does not induce a general stress response in LAC under the conditions tested (50–54). Among the metabolism genes examined, OHM had no significant effect on the transcription of the genes involved in electron transport (atpG and sdhA). However, OHM treatment significantly decreased the transcription of murQ, an N-acetylmuramic acid 6-phosphate lyase, in both LAC and LACΔagr. Although this protein, which is involved in cell wall recycling, is dispensable for growth in E. coli (55–57), its contribution to the growth of Gram-positive pathogens is less clear. However, the absence of bactericidal or bacteriostatic effects with OHM treatment suggests that MurQ is not required for growth under the conditions tested. In addition, OHM treatment did not increase the transcription of genes examined for their potential to contribute to drug efflux or resistance. Therefore, although there are some non-agr effects, these results suggest that OHM is not a general inhibitor of transcription or energetics, nor is it a general inducer of drug efflux. Furthermore, together with the above demonstrations of (i) OHM-mediated agr inhibition in a whole-cell assay, (ii) OHM-mediated inhibition of AgrA_C binding to agr promoter DNA by both EMSA and bead-based assay, and (iii) the direct binding of OHM to AgrA_C shown by SPR (Fig. 2 and 3), these results are consistent with a mechanism whereby OHM predominantly functions as an inhibitor of agr activation.

TABLE 1.

Transcriptional analysis of the agr specificity of OHM

Focus	Gene	agr regulation/association	Fold change in gene expression (vehicle/OHM treatment)a
			LAC	P value	Δagrb	P value
Virulence	spa (SAUSA300_0113)	Neg	<2		ND
	set7 (SAUSA300_0396)	NAc	−2.84	0.0019	<2
	saeR (SAUSA300_0691)	Pos	<2		ND
Stress response	asp23 (SAUSA300_2142)	NA	<2		ND
	crtM (SAUSA300_2499)	NA	<2		ND
	clpB (SAUSA300_0877)	NA	<2		ND
Metabolism	atpG (SAUSA300_2059)	NA	<2		ND
	murQ (SAUSA300_0193)	Pos	−4.99	<0.0001	−8.551	<0.0001
	sdhA (SAUSA300_1047)	NA	<2		ND
Efflux/antibiotic resistance	norA (SAUSA300_0680)	NA	<2		ND
	mdrA (SAUSA300_2299)	NA	<2		ND
	NaMDR gened (SAUSA300_0335)	NA	−3.23	0.0015	−2.26	0.0076

^aValues are shown if ≥2-fold.

^bAssay performed with Δagr strain if ≥2-fold difference with OHM treatment of LAC. ND, not done.

^cNA, not applicable.

^dNaMDR, Na⁺-driven multidrug efflux pump.

ω-Hydroxyemodin attenuates S. aureus SSTI.

Invasive S. aureus SSTIs require agr-regulated virulence factors (14, 19, 58–60). Therefore, we assessed the efficacy of OHM in an established mouse model of S. aureus SSTI (42). Over the course of a three-day infection with the USA300 isolate LAC, a single 5-μg dose of OHM administered at the time of infection significantly inhibited abscess (Fig. 4A and B) and ulcer (dermonecrosis) formation (Fig. 4A and C), as well as morbidity at day one postinfection (assessed by weight loss) compared to that of the vehicle-treated controls (Fig. 4D). In contrast, no differences were observed between the OHM and vehicle-treated mice infected with LACΔagr (Fig. 4A and data not shown), demonstrating the specificity of OHM for disrupting agr signaling without directly impacting the host. Importantly, the single OHM treatment reduced the day three and day seven postinfection bacterial burden at the site of infection in LAC-infected (Fig. 4E) but not Δagr-infected mice (Fig. 4F), suggesting that mice were better able to combat the infection in the absence of agr signaling.

FIG 4.

ω-Hydroxyemodin limits abscess formation and dermonecrosis and promotes bacterial clearance in a mouse model of S. aureus SSTI. SKH1 mice were subcutaneously injected with 5 × 10⁷ to 7 × 10⁷ CFU of LAC or Δagr along with OHM (0.2 mg/kg) or vehicle control. (A) Representative images of abscesses and ulcers on day three postinfection (D3) (scale bar = 5 mm). Day 3 postinfection abscess (B), ulcer area (C), and weight loss of LAC-infected mice. (E) Day 3 and day 7 postinfection bacterial burden at the site of infection. The data shown are the mean ± SEM (LAC, day 3, n = 12 to 16 mice per group; day 7, n = 5 mice per group). (F) Δagr day 3 bacterial burden at the site of infection. n = 6 mice per group. ns, not significant. *, P < 0.05; **, P < 0.01 by Mann-Whitney U test.

We predicted that if OHM treatment supported host-mediated clearance by disrupting agr signaling, OHM-treated LAC, but not LACΔagr, would be more readily killed by innate immune cells in vitro compared to in the vehicle-treated controls. As predicted, OHM treatment of LAC, but not LACΔagr, resulted in significantly increased intracellular killing by both mouse macrophages (Fig. 5A) and human PMNs (Fig. 5B) compared to that in the vehicle-treated controls. This increased killing was not a result of OHM-mediated effects on opsonophagocytosis, as the total number of bacteria phagocytosed (see Fig. S5 in the supplemental material) and the percentage of bacteria phagocytosed relative to the total inoculum (data not shown) were equivalent, regardless of whether the bacteria were pretreated with vehicle or OHM. Furthermore, OHM treatment of LAC but not LACΔagr protected human PMNs from killing by secreted agr-regulated virulence factors. PMNs showed significantly increased survival in the presence of supernatant from OHM- versus vehicle-treated LAC (Fig. 5C). Together, these results demonstrate that OHM supports the host-mediated clearance of S. aureus by inhibiting agr-mediated virulence.

FIG 5.

ω-Hydroxyemodin supports immune cell killing of agr⁺ S. aureus. (A) Mouse macrophage (RAW 264.7) intracellular killing of bacteria pretreated with OHM or vehicle control. The data are shown as the mean ± SEM normalized to 100% after 1 h of incubation at an MOI of 1:1. n = 6 from two independent experiments performed in triplicate. (B) Human polymorphonuclear leukocyte (PMN) intracellular killing of bacteria pretreated with OHM or vehicle control. The data are the mean ± SEM presented as the percent survival (top) and log CFU reduction (bottom) compared to time zero. (C) Supernatant lysis of human PMNs, assessed by LDH release, after 2 h of incubation with sterile supernatant from overnight cultures grown in the presence of OHM or vehicle control. The data are the mean ± SEM, presented as percent PMN viability compared to 100% lysis by Triton X-100. (B and C) Experiments were performed in triplicate with PMNs from two separate donors. A representative donor experiment is shown. ns, not significant. **, P < 0.01; ***, P < 0.001; ****, P ≤ 0.0001 by Student's t test.

ω-Hydroxyemodin limits inflammation mediated by S. aureus QS.

S. aureus uses a variety of virulence factors, many of which are regulated by the agr system, to evade host clearance mechanisms. These virulence factors cause tissue damage and inflammation and facilitate invasive infection (61–65). Therefore, we postulated that the reduction in bacterial burden in the LAC-infected, OHM-treated mice would be associated with reduced tissue damage and reduced local inflammatory cytokine production compared to the vehicle-treated controls. Histological analysis of day three postinfection skin sections confirmed the overall reduction in abscess formation and ulceration in OHM-treated mice (Fig. 6A). Additionally, skin sections from the vehicle-treated mice displayed a disorganized architecture at both the epithelium-to-necrosis transition (Fig. 6A, left inset) and the abscess periphery (right inset) compared to sections from OHM-treated mice. Also as predicted, OHM treatment resulted in a local cytokine profile matching that of LACΔagr-infected mice on day three postinfection (Fig. 6B), with significant reductions in interleukin-1β (IL-1β), tumor necrosis factor alpha (TNF-α), and IL-6, but not the anti-inflammatory cytokine IL-10, compared to the vehicle-treated controls. LAC-infected mice treated with OHM also showed reduced transcription of il-1β, tnfα, and il-6 at 24 h postinfection compared to the levels in the vehicle-treated mice (Fig. 6C). Finally, the activation of the NLRP3 inflammasome and the subsequent release of IL-1β is induced by pore formation in the host cell membranes by Hla (66–68), and the passive transfer of Hla-neutralizing antibodies is sufficient to limit the secretion of IL-1β (69). Therefore, we predicted that OHM treatment would be associated with reduced Hla expression and nlrp3 transcription at the site of infection. As expected, the OHM-treated mice showed decreased local Hla expression and decreased transcription of nlrp3 compared to the vehicle-treated controls (Fig. 6D and E). Together, these data demonstrate that OHM inhibition of agr signaling limits host tissue damage and inflammation during S. aureus SSTI.

FIG 6.

ω-Hydroxyemodin limits pathology and expression of inflammatory cytokines during S. aureus SSTI. (A) Top, representative hematoxylin and eosin micrographs of 5-μm sagittal sections of day three post-LAC infection abscess tissue. The abscess area is demarcated by a fine dashed line, while the ulcer surface length is marked by the thick dash line. Bottom, magnification of the transition from normal epithelium to necrotic tissue (left) and organization of the abscess tissue (right). (B) Multiplex analysis of cytokines present in the abscess tissue on day three postinfection with LAC or Δagr. (C and E) Quantification of il-1β, tnfα, il-6 (C), and nlrp3 (E) relative to hprt in abscess tissue at 24 h post-LAC infection. (D) Western blot analysis and quantification of Hla at the site of infection (day 3) in OHM- versus vehicle-treated mice (n = 4 mice/group). The data are the mean ± SEM from infection data, as described in the legend to Fig. 4. RQ, relative quantification; *, P < 0.05; **, P < 0.01; ***, P < 0.001 by Student's t test.

DISCUSSION

Recently, the National Institutes of Allergy and Infectious Diseases reported on the current status and future directions for its antibacterial resistance program (70). Two of the strategic approaches highlighted were (i) antivirulence strategies to disarm bacteria to reduce pathogenesis and (ii) approaches to harness the host immune system to better combat infections. Here, we report that OHM, a natural product isolated from the fungus P. restrictum (27), addresses both goals by directly inhibiting S. aureus QS-dependent virulence while indirectly bolstering the host immune response against S. aureus infection. In a mouse model of S. aureus SSTI, OHM significantly decreases abscess and ulcer formation and promotes bacterial clearance. Importantly, OHM treatment reduces tissue damage and limits local proinflammatory cytokine production to levels seen in mice infected with the agr deletion mutant. Furthermore, OHM treatment enhances immune cell-mediated killing of S. aureus in an agr-dependent manner. Therefore, these data demonstrate that antivirulence strategies can limit disease by disarming the bacteria while concurrently reducing inflammation and promoting host innate defense. In addition, this is the first polyhydroxyanthraquinone described with in vivo efficacy against MRSA infection, adding to the expanse of natural products with the potential to promote human health and advance antibiotic stewardship.

Numerous reviews have addressed the potential role of antivirulence strategies, including the disruption of QS, in combating the antimicrobial resistance crisis (4, 6, 7, 71–73). We chose to focus on the disruption of agr QS as an antivirulence approach to S. aureus SSTIs due to their predominance in S. aureus disease manifestations, as well as the established contribution of agr in facilitating these infections (11–14, 19). For example, agr deletion mutants (Δagr) are less pathogenic and more readily cleared during SSTIs than are wild-type strains (14, 19), and host innate effectors that disrupt agr signaling limit disease in skin infection models (20–22). Importantly, the sterile supernatant from agr⁺ but not from agr-null S. aureus strains is sufficient to cause skin lesions similar to those in an active infection, definitively demonstrating the role of agr-regulated secreted virulence factors in skin pathogenesis (58). Here, we used OHM in a prophylactic administration model, similar to that previously reported for the administration of competing AIP or the passive transfer of monoclonal antibodies targeting AIP4 (24, 58), to demonstrate that small-molecule-mediated disruption of agr signaling in vivo results in an “agr-null-like” host inflammatory profile.

The disruption of agr signaling during SSTIs results in reduced bacterial burden at various time points postinfection (19, 25, 58, 60), suggesting that host-mediated bacterial clearance is more effective in the absence of QS. In support of this, we have shown here and elsewhere (25) that agr inhibitors enhance some mechanisms of innate immune bacterial clearance against agr⁺ S. aureus. The increased bacterial clearance likely results from inhibiting the expression of the agr-regulated secretome, which includes virulence factors that target host immune cells. Among these, the phenol-soluble modulins (PSMs) and Hla, in particular, contribute to abscess formation, dermonecrosis, and bacterial burden during SSTIs (59, 60, 74, 75). These toxins target a variety of cell types to suppress both innate and adaptive immunity (76). For example, S. aureus utilizes Hla to survive inside neutrophils and macrophages (45, 77–80) and to induce programmed cell death in T cells, B cells, and monocytes (81). In addition, Hla activates the NLRP3 inflammasome in a variety of cells, resulting in the release of the inflammatory cytokine IL-1β (67, 68, 82). Although neutrophils and IL-1β are needed for the ultimate clearance of S. aureus SSTI (83, 84), limiting inflammation caused by bacterial toxins clearly benefits the host.

Our molecular modeling studies positioned OHM near R218 of S. aureus AgrA. This residue, which is strictly conserved across multiple staphylococcal species (46), is required for agr function and contributes to AgrA binding to agr promoter DNA (38, 47). Although the potential exists for OHM to drive selection for an alternative amino acid at residue 218, any such mutation would likely result in agr dysfunction. The selection for QS-deficient isolates is unlikely to be of significant benefit to the pathogen, as these isolates are severely attenuated, more readily cleared by host defenses, and less effective at initiating infection (19, 37, 60). However, while the residues directly involved in OHM binding have yet to be definitively demonstrated, the question of whether OHM can select for mutation of other residues in the predicted binding site that would prevent OHM function while retaining agr activity will require empirical determination.

Along with inhibiting S. aureus agr activation, OHM likewise inhibits agr signaling by S. epidermidis, an important member of the skin microbiome and also an opportunistic pathogen (85). S. epidermidis agr regulation drives the mechanisms of resistance to host innate defense (86), suggesting that OHM or related analogues might prove efficacious against S. epidermidis infections. However, in its role as a commensal, S. epidermidis appears to benefit the host by such means as competing with S. aureus for colonization and contributing to overall skin immunity (87–89). Additionally, AIP1 produced by S. epidermidis is cross-inhibitory to S. aureus agr types I to III in vitro (85, 90, 91). Therefore, it is unclear whether OHM-mediated perturbation of S. epidermidis QS, during treatment for S. aureus SSTI, might result in unwarranted complications, such as increased host-mediated clearance of S. epidermidis and/or the loss of potentially beneficial cross-inhibitory AIP. To address these possibilities, it will be important to experimentally determine the in vivo implications of disrupting QS in an S. epidermidis-colonized host on subsequent S. aureus skin infection.

As is the case with existing antimicrobials, QS inhibitors (QSIs) may not be a one-size-fits-all solution. Although SSTIs comprise the vast majority of S. aureus infections, this pathogen causes a variety of disease manifestations, including pneumonia, osteomyelitis, endocarditis, and bloodstream infections (BSI). This raises the question of whether the use of QSIs will be universally beneficial. The contribution of agr to S. aureus pathogenesis has largely been demonstrated in models of SSTIs and pneumonia (14, 19, 24, 92). In contrast, agr dysfunction has been associated with persistent bacteremia in hospitalized patients (83, 84), suggesting that QSIs would be best utilized to prevent S. aureus invasion prior to BSI. Likewise, the disruption of agr is associated with biofilm formation in vitro (85), potentially limiting the utility of QSIs for the treatment of infections involving implanted devices, as well as osteomyelitis and endocarditis. Whether OHM and other QSIs would contribute to staphylococcal biofilm formation in vivo will require investigation in appropriate animal models of infection. Overall, however, QSIs might be a critical part of the developing arsenal for combating antibiotic resistance either alone, as adjuncts to existing antibiotics, or along with potential vaccines or other approaches to augment host defense.

A substantial portion of the compounds used to fight infections have their origins in natural products (26). Despite clearly representing a wealth of bioactive molecules, natural products have received limited attention with respect to identifying specific inhibitors of S. aureus QS that are not bactericidal. To date, other natural products identified as agr inhibitors include (i) α-cyperone from the nut grass plant Cyperus rotundus (93), (ii) the fungal metabolite ambuic acid (94), and extracts from (iii) the goldenseal plant Hydrastis canadensis (95) and (iv) three Italian medicinal plants, Ballota nigra, Castanea sativa, and Sambucus ebulus (96). Among these, OHM is the first to demonstrate in vivo efficacy against S. aureus QS. Therefore, we predict that targeted testing of structurally diverse natural products will continue to reveal a broad range of antivirulence molecules with the potential to support innate host defense mechanisms and to positively contribute to antibiotic stewardship.