Discovery of an antivirulence compound that targets the Staphylococcus aureus SaeRS two-component system to inhibit toxic shock syndrome toxin-1 production

Abstract

Menstrual toxic shock syndrome (mTSS) is a rare but severe disorder associated with the use of menstrual products such as high-absorbency tampons and is caused by Staphylococcus aureus strains that produce the toxic shock syndrome toxin-1 (TSST-1) superantigen. Herein, we screened a library of 3920 small bioactive molecules for the ability to inhibit transcription of the TSST-1 gene without inhibiting the growth of S. aureus. The dominant positive regulator of TSST-1 is the SaeRS two-component system (TCS), and we identified phenazopyridine hydrochloride (PP-HCl) that repressed the production of TSST-1 by inhibiting the kinase function of SaeS. PP-HCl competed with ATP for binding of the kinase SaeS leading to decreased phosphorylation of SaeR and reduced expression of TSST-1 as well as several other secreted virulence factors known to be regulated by SaeRS. PP-HCl targets the virulence of S. aureus, and it also decreases the impact of TSST-1 on human lymphocytes without affecting the healthy vaginal microbiota. Our findings demonstrate the promising potential of PP-HCl as a therapeutic strategy against mTSS.

Menstrual toxic shock syndrome (mTSS) is a serious bacterial toxin-mediated disease that became widely recognized in the early 1980s due to an epidemic of cases in the United States that was associated with the use of high-absorbency tampons (1). Following early investigations into this epidemic, it was determined that these women were vaginally colonized with Staphylococcus aureus that produced a unique toxin now known as toxic shock syndrome toxin-1 (TSST-1) (2, 3). TSST-1 functions as a bacterial superantigen, forcing the activation of numerous T cells that can progress to a cytokine storm that characterizes mTSS (4). Approximately 30% of the general human population is understood to be colonized by S. aureus within the nasal passages (5). However, many other body sites are also frequently colonized (6), and in healthy women using tampons, ∼30 to 40% were vaginally colonized with S. aureus, with ∼5% of strains producing TSST-1 (7, 8). The use of high-absorbency tampons is understood to have altered the vaginal environment to create conditions that are sensed by S. aureus which can result in the increased production of TSST-1 (9, 10, 11, 12). Although tampons have been re-designed to reduce the risk of mTSS, the importance of mTSS for public safety remains, and understanding the regulatory pathways that control TSST-1 production in conditions mimicking the vaginal tract may be key to the prevention of future cases of mTSS.

The incidence of mTSS in the United States is ∼0.5 to 1.0 per 100,000 population (13), which is far below the number of women who are vaginally colonized by TSST-1⁺ S. aureus. Thus, multiple factors are likely protective for the development of mTSS including the proper use of menstrual management products (14), the presence of neutralizing anti-TSST-1 antibodies (15), the endogenous vaginal microbiota (16, 17, 18, 19), and environmental signals that repress TSST-1 production including low levels of O₂ and CO₂ levels (20), acidic pH (21), and high levels of glucose (12). Furthermore, S. aureus colonization will also vary during the menstrual cycle as the environment is dynamic due to hormonal fluctuations (12, 22).

As an alternative to the use of antibiotics with bacteriostatic or bactericidal activity, genetic regulatory systems in S. aureus have been targeted for antivirulence therapeutics. For example, the accessory gene regulator (Agr) two-component system (TCS) is a well-studied quorum sensing system that is activated via the production of endogenous auto-inducing peptides (AIPs), and apart from the SaeRS TCS, Agr is the other major exotoxin regulator in S. aureus (23). Indeed, Agr indirectly promotes the expression of TSST-1 by blocking the production of the repressor of toxins (Rot) protein (24). Coagulase-negative staphylococcal (CoNS) species, as well as S. aureus, can produce AIP variants that can function to inhibit heterologous Agr systems. For example, interference with Agr signaling via inhibitory AIP molecules produced by Staphylococcus hominis could suppress skin damage and inflammation in a mouse model of S. aureus-induced atopic dermatitis (25) and reduce lesion sizes in a mouse model of dermonecrosis without altering bacterial counts (26). However, targeting the Agr system could also potentially evolve strains into a persistent agr-deficient state (27).

The GraXRS sensing system is another TCS that responds to cell-envelope stress including cationic antimicrobial peptides and low pH (28, 29). A small molecule screen designed to find inhibitors of the early step in wall teichoic acid production identified a GraR inhibitor, which blocked intracellular survival of S. aureus within macrophages and enhanced larvae survival in a Galleria mellonella infection model (30). Furthermore, the autolysis-related locus TCS (ArlRS) controls many phenotypes including adhesion, capsule production, and metal transport, primarily through the MgrA stand-alone transcription factor (31). Using a mgrA promoter reporter, small molecule inhibitors of ArlRS were identified that could also reduce skin infections, but did not alter viable S. aureus recovered from the lesions (32).

Recently, an inhibitor of SaeS was found using an α-hemolysin promoter screen that inhibited SaeRS-regulated virulence factors expression, but also appeared to affect Agr, and was able to constrain experimental invasive infections by S. aureus (33). A structure-based virtual screen also identified the non-steroidal anti-inflammatory drug Fenoprofen as a direct SaeR inhibitor that could attenuate S. aureus virulence in vitro and in vivo (34). Although these compounds have not yet reached therapeutic use in humans, there are now multiple examples by which antivirulence compounds demonstrate the potential of targeting S. aureus TCSs to disrupt virulence and potentially increase the efficiency of antibiotic therapy.

To advance therapeutics that could inhibit the production of TSST-1, we developed a cell-based platform using a TSST-1 transcriptional luciferase reporter assay with S. aureus grown in a vaginal-mimicking medium to simulate the environmental conditions of mTSS. Using this platform, we screened a library of small bioactive molecules and identified that phenazopyridine hydrochloride (PP-HCl) demonstrates antivirulence activity against TSST-1 expression without growth inhibitory properties. This molecule was further characterized to decipher its anti-virulent mechanisms, and we demonstrate that PP-HCl functions through inhibition of kinase activity of the SaeRS TCS, the major positive transcriptional activator of TSST-1 (35).

Results

PP-HCl represses the production of TSST-1 without inhibiting S. aureus growth

We screened a library of 3920 bioactive molecules (Fig. 1), each at 10 μM, for the ability to repress the activity of the TSST-1 promoter (P_tst) using S. aureus MN8 harboring pAmilux::P_tst, a plasmid that indirectly measures P_tst activity using a luciferase (lux) reporter (Table 1). Compound screening was performed in vaginally-defined media (VDM) containing low amounts of glucose (700 μM) to mimic the vaginal environment during mTSS when CcpA-mediated repression of TSST-1 is relieved (12). From the initial screen, seventy molecules demonstrated low luminescence (less than 100 raw RLU) with growth equal to or greater than 95% of the control cultures (Fig. 1). Among these compounds, 18 were previously described to possess antimicrobial activity and were not investigated further (Fig. 1). From the remaining 52 hits, compounds were carried forward based on commercial availability and therapeutic potential, and further tested under low throughput conditions with titrations from 0 to 50 μM using the S. aureus P_tst reporter strain. In this context, PP-HCl showed no growth inhibition up to 50 μM (Fig. 2A) with robust repression of P_tst with 5 μM or higher concentrations of PP-HCl (Fig. 2, B and C). To corroborate the luciferase reporter experiments, TSST-1 protein levels from S. aureus MN8 supernatants were evaluated using an anti-TSST-1 Western Blot. A similar trend was observed with decreased TSST-1 produced when S. aureus was grown for 4 h with 1 μM PP-HCl and with only a faint band visible when grown in 5 μM, with no TSST-1 bands observed at a concentration in between 10 to 50 μM PP-HCl incubation (Fig. 2D). Moreover, by 18-h growth, 5 μM PP-HCl showed decreased TSST-1 production that was strongly repressed with 50 μM PP-HCl (Fig. S1). These data demonstrate that PP-HCl decreases tst promoter activity and the production of TSST-1 protein without inhibiting S. aureus growth and may represent a new potential anti-virulent compound for S. aureus.

Figure 1.

Screening strategy for anti-virulent against TSST-1.S. aureus MN8 reporter assay monitoring growth and luminescence production (expression of the promoter P_tst) was performed with 3920 bioactive molecules each tested at 10 μM concentration. The graph represents all compounds tested with the 100 RLU threshold represented by an additional line; PP-HCl is marked in red. Compounds with a raw luminescence equal to or less than 100 RLU were further selected (380 molecules). Compounds lacking antimicrobial activity with normalized OD₆₀₀ equal to or greater than 95% of the control were selected leading to 70 putative inhibitory molecules. An additional selection was made for the bioactive molecules with no known previous antimicrobial activity (both antibiotics and antimycotics) resulting in 52 molecules of interest from the initial screen.

Table 1.

Bacterial strains used in this study

Strain	Description	Source or reference
S. aureus
MN8	Prototypic menstrual TSS strain, tst+	(60)
MN8 ΔccpA	MN8 with deletion of ccpA gene	(12)
MN8 Δrot	MN8 with deletion of rot gene	(24)
MN8 Δagr	MN8 with the agr operon replaced with a tetR marker	(16)
MN8 ΔsrrAB	MN8 with deletion of srrAB genes	(39)
MN8 ΔsarA	MN8 with deletion of sarA gene	(35)
MN8 ΔsaeS	MN8 with deletion of saeS gene	(35)
MN8 (pAmilux::Ptst)	MN8 containing pAmilux::Ptst	(35)
MN8 ΔccpA (pAmilux::Ptst)	MN8 ΔccpA containing pAmilux::Ptst	(12)
MN8 Δrot (pAmilux::Ptst)	MN8 Δrot containing pAmilux::Ptst	(24)
MN8 Δagr (pAmilux::Ptst)	MN8 Δagr containing pAmilux::Ptst	(24)
MN8 ΔsrrAB (pAmilux::Ptst)	MN8 ΔsrrAB containing pAmilux::Ptst	(39)
MN8 ΔsarA (pAmilux::Ptst)	MN8 ΔsarA containing pAmilux::Ptst	(35)
MN8 ΔsaeS (pAmilux::Ptst)	MN8 ΔsaeS containing pAmilux::Ptst	(35)
MN8 ΔsaeS (pALC2073::saeQRS)	MN8 ΔsaeS containing pALC2073::saeQRS	(35)
MN8 ΔsaeS (pALC2073::saeQRSL18P)	MN8 ΔsaeS containing pALC2073::saeQRSL18P	This study
Escherichia coli
XL1 Blue	Cloning host	Stratagene
Lactobacillus species
Lactobacillus crispatus ATCC 33820	Representative strain of community state type I (CST I)	ATCC
Lactobacillus gasseri ATCC 33323	Representative strain of CST II	ATCC
Lactobacillus jensenii ATCC 25258	Representative strain of CST V	ATCC

Figure 2.

PP-HCl decreases tst promoter activity and TSST-1 production.A, growth of wild-type S. aureus MN8 containing pAmilux::P_tst was assessed by optical density at 600 nm over 18 h with parallel assessment of luminescence. Growth with various concentrations of PP-HCl up to 50 μM was similar to untreated S. aureus. Results are presented as the average optical density at 600 nm. B, luminescence from wild-type S. aureus MN8 containing pAmilux::P_tst (measured in RLU) decreased drastically by 5 μM of PP-HCl and was almost absent at 50 μM PP-HCl. Results are presented as the averaged RLU ± SD. C, relative expression of the tst promoter was calculated as the area under the curve of luminescence over the area under the curve of OD₆₀₀ and demonstrates the same tendency as shown in the luminescence curves. The experiment was repeated with 3 biological replicates and error bars represent SD. Ordinary one-way ANOVA was performed (∗∗∗∗p < 0.0001). D, TSST-1 production in the supernatants of wild-type S. aureus MN8 was evaluated by Western Blot at the same concentrations of PP-HCl as tested during the luciferase assay. Supernatants were harvested after a 4-h incubation in VDM, concentrated using trichloroacetic acid, and normalized for 12 OD₆₀₀ units. Shown are exoprotein profiles (top panels) and Western blot analysis (bottom panels) of TSST-1 for wild-type S. aureus MN8.

Effect of PP-HCl on vaginal lactobacilli and T cell activation

The vaginal environment is considered to be strongly influenced by a beneficial microbiota (22), and the vaginal microbiota in women is generally classified into five Community State Types (CSTs). CST-I, CST-II, and CST-V, dominated by Lactobacillus crispatus, Lactobacillus gasseri and Lactobacillus jensenii, respectively, are considered to provide a protective function (36); thus, an important property of PP-HCl for use as an anti-virulent therapeutic would be the lack of activity against beneficial microbiota members that may also play an important role in preventing mTSS (37). To evaluate this, we assessed the growth of 3 representative beneficial vaginal Lactobacillus species in VDM containing increasing concentrations of PP-HCl (Fig. 3, A–C). Bacterial growth was similar in the presence of PP-HCl, and although there was a small growth defect with L. crispatus when grown in the presence of 250 μM PP-HCl, the compound did not inhibit the growth of either L. gasseri or L. jensenii, further indicating that PP-HCl does not possess overt antibiotic properties against Gram-positive bacteria (Fig. 3, A–C). As acidification of the vaginal environment is considered a key property of healthy CSTs (22), we also assessed the production of lactic acid in the presence of PP-HCl from the three lactobacilli species. Lactic acid production was similar for both treated and untreated conditions demonstrating that representatives of a healthy vaginal microbiota are not drastically affected by the presence of the anti-virulent compound (Fig. 3D).

Figure 3.

PP-HCl does not disrupt representative healthy membersofthe vaginal microbiota. Growth and lactic acid production from the three dominant representatives of healthy or stable microbiota communities were assessed. A, L. crispatus (ATCC 33820), (B) L. gasseri (ATCC 33323), and (C) L. jensenii (ATCC 25258) growth was assessed in 60 mM glucose VDM at concentrations of PP-HCl ranging from 0 to 250 μM. All growth curves are represented as the average of at least triplicated experiments. D, lactic acid production by the lactobacilli at 0, 50, and 100 μM of PP-HCl was assessed and no differences were detectable within the groups. Ordinary one-way ANOVA was performed and p value over 0.05 were considered non-significant (ns). The results are presented as the mean of three replicates ± SD.

Given that PP-HCl is used as an analgesic for urinary tract infections (38), we next evaluated PP-HCl for activity on eukaryotic cells with a focus on immune cells. To test this, human peripheral blood mononuclear cells (PBMCs) were isolated from healthy donor blood and treated with either purified recombinant TSST-1, PP-HCl, or in combination. As expected, TSST-1 induced production of the T cell cytokine IL-2 whereas PP-HCl did not; however, in combination, the IL-2 response was reduced relative to TSST-1 alone (Fig. S2A). We also assessed cell viability and TSST-1 treatment showed decreased cell viability, likely due to activation of induced cell death, whereas PP-HCl did not affect cell viability. In combination, cell viability was also decreased similar to TSST-1 alone (Fig. S2B). We then assessed supernatants from wild-type S. aureus MN8, or the S. aureus MN8 with a deletion in the TCS SaeS gene (Table 1). SaeRS is a direct and positive regulator of the tst promoter and an in-frame deletion in the histidine kinase gene saeS results in a drastic reduction of TSST-1 production (35). Wild-type S. aureus MN8 supernatant induced the production of IL-2 which was significantly reduced when using the ΔsaeS mutant and was similar to the co-treatment of PBMCs with the wild-type S. aureus MN8 supernatant and PP-HCl (Fig. S2C). None of the supernatants, or PP-HCl alone, altered cell viability in between the various conditions tested (Fig. S2D). To evaluate if the repression of T cell activation by PP-HCl was specific to superantigens, we activated T cells using anti-CD3/CD28 bead in the presence or absence of PP-HCl. These experiments demonstrated that PP-HCl also reduced IL-2 production from T cells demonstrating that this was not a superantigen-dependent phenomenon (Fig. S2E).

Activity of PP-HCl on TSST-1 transcription bypasses the main tst repressors

To decipher genetic mechanisms involved in PP-HCl-dependent repression of TSST-1, we evaluated various key S. aureus MN8 regulatory mutant strains containing pAmilux::P_tst in the presence or absence of PP-HCl (Table 1). As before, 50 μM PP-HCl had little impact on S. aureus growth (Fig. 4, A and C). The two known positive regulators of tst transcription are the Agr and Sae TCSs where SaeR acts directly on the tst promoter, while Agr relieves repression of tst by Rot (24, 35). As predicted, tst expression was dramatically reduced in the agr and saeRS mutants without PP-HCl treatment (Fig. 4, B and E). Key repressor systems for tst transcription in rich media include the TCS SrrAB (9, 39), and the cytoplasmic intermediatory regulators ccpA, rot, and sarA (12, 24, 35); however, in the tested conditions, only the ΔccpA mutation showed an increase in tst expression (Fig. 4E). Surprisingly, the other known repressors (SrrAB, Rot, and SarA), in VDM with low levels of glucose, had decreased tst expression suggesting they may fulfill an activator role in this specific environment instead, and that other levels of regulation are yet to be described in these mTSS environmental conditions (Fig. 4E). PP-HCl completely inhibited luminescence within all the various strains suggesting that it repressed tst expression in each of these mutant backgrounds (Fig. 4D). The lack of TSST-1 transcription from these latter regulators in the presence of PP-HCl suggests that the compound does not function by activating these repressors (Fig. 4E).

Figure 4.

PP-HCl overcomes known genetic regulators of TSST-1 to inhibit transcription of tst. Growth curves (OD_600nm) and luciferase assays (RLU) of S. aureus MN8 reporter strains including central TSST-1 regulatory mutants were performed without (A and B) or with 50 μM of PP-HCl (C and D). E, the relative expression of the tst promoter in each mutant exposed to both conditions is presented. The experiment was repeated with 3 biological replicates. Ordinary one-way ANOVA was performed for statistical analysis (∗∗∗∗p < 0.0001). The results are presented as the geometric mean ± geometric SD.

PP-HCl-dependent inhibition of multiple exotoxin genes requires the SaeRS TCS

To further evaluate potential regulatory circuits involved in the PP-HCl-mediated repression of TSST-1, we conducted transcriptional analyses by RNA sequencing (RNA-seq) comparing wild-type S. aureus MN8 grown for 4h in the presence or absence of 50 μM PP-HCl (Fig. 5A). This analysis demonstrated that the TSST-1 gene, as well as several exotoxin genes, were repressed in the presence of PP-HCl (Fig. 5A and Table S1). Interestingly, α-toxin, TSST-1, Sbi, SCIN and CHIPs are all known to be positively regulated by the SaeRS TCS (35, 40) and each of these genes, including the gene for the more recently characterized myeloperoxidase inhibitor SPIN (41), all contain canonical SaeR binding sequences upstream of the respective promoters (42). Although the α-hemolysin toxin gene in S. aureus MN8 is truncated (denoted as α-hemolysin∗), the transcript is produced and is highly repressed in the presence of PP-HCl. Although not statistically different, we also noted a trend for decreased transcription of genes encoding the gamma-hemolysin (hlgA, hlgB, hlgC), also known to be SaeRS controlled (42), as well as both saeR and saeS (Fig. 5A and Table S1), with the latter changes likely due to the known autoregulation of the sae locus (43). Of genes that could be involved in virulence, we also noted the upregulation of sdrE. These data suggested that repression of tst transcription by PP-HCl, and the noted exotoxins, may be mainly mediated through inhibition of the SaeRS TCS.

Figure 5.

PP-HCl affects the transcriptional regulation of a variety of virulence factors related with SaeRS regulation.A, the volcano plot represents RNA-seq analysis comparing untreated S. aureus MN8 with MN8 treated with 50 μM of PP-HCl. Colored dots represent noted virulence factors or regulators that were significantly downregulated (red), upregulated (green) or did not change significantly in the presence of PP-HCl. See Table S1 for the extensive analysis of transcripts. B, the volcano plot represents RNA-seq analysis comparing wild-type S. aureus MN8 with MN8 ΔsaeS. Colored dots represent noted virulence factors or regulators that were significantly down regulated (red) or did not change significantly in the MN8 ΔsaeS mutant. See Table S1 for the extensive analysis of transcripts. C, TSST-1 production in supernatants by Western blot from wild-type S. aureus MN8, MN8 ΔsaeS, or MN8ΔsaeS complemented with wild-type saeS (pALC2073::saeQRS) or with constitutively active saeS_L18P (pALC2073::saeQRS_L18P) exposed to 0 or 50 μM PP-HCl. Supernatants were harvested after an 18-h incubation in VDM, concentrated using trichloroacetic acid and normalized for 12 OD₆₀₀ units. Shown are exoprotein profiles (top panels) and Western blot analysis (bottom panels) for TSST-1.

Given many of the repressed virulence factor genes are part of the Sae regulon, we next conducted RNA-seq experiments comparing wild-type S. aureus MN8 with the ΔsaeS mutant grown for 4h. Strikingly, transcriptional changes in the ΔsaeS mutant appeared to phenocopy many of the virulence factors that were repressed by PP-HCl in wildtype MN8 (Fig. 5B and Table S2), further suggesting that PP-HCl may be acting directly on the Sae TCS. To evaluate the involvement of SaeRS in the inhibition of virulence factor expression by PP-HCl, the RNA-seq experiment was repeated in the S. aureus MN8 ΔsaeS mutant in the presence or absence of PP-HCl. Transcription of the genes encoding each of the exotoxins was not differentially changed in the ΔsaeS genetic background between the two conditions (Fig.S3 and Table S3). These data further suggest the importance of the SaeRS pathway for PP-HCl to repress TSST-1 and the other exotoxins.

S. aureus Newman encodes a variant of SaeS with proline in position 18 instead of leucine, which results in constitutive kinase activity (44, 45). To evaluate the specificity of the TSST-1 repression through SaeS, we generated the SaeS_L18P genetic variant within the pAL2073::saeQRS complementation plasmid and treated the various complemented strains (Table 1) with 50 μM PP-HCl. As expected, PP-HCl decreased TSST-1 production from wildtype MN8 and MN8 ΔsaeS complemented with wildtype saeS (Fig. 5C). Although the constitutively active L18P variant resulted in less TSST-1 production compared with the wildtype saeS complemented strain, this clone demonstrated no obvious repression of TSST-1 in the presence of PP-HCl (Fig. 5C). These results suggest that PP-HCl repression of TSST-1 is mainly through inhibition of SaeS signaling.

PP-HCl inhibits phosphorylation of SaeS

SaeS is the sensor histidine kinase of the SaeRS TCS that relays the presence of polymorphonuclear leukocytes (PMN or neutrophil)-produced signals via phosphorylation to its response regulator SaeR, which subsequently binds target DNA to alter gene transcription (40, 43, 46). Preynat-Seauve et al. proposed that PP-HCl could inhibit human kinases by binding the ATP-binding site (47) and therefore we hypothesized that PP-HCl may bind the ATP-binding site of SaeS to inhibit phosphorylation of SaeR. To evaluate this, S. aureus MN8 was grown in the presence or absence of PP-HCl, and the phosphorylated state of SaeR was assessed. Relative to untreated wild-type S. aureus MN8 cells, PP-HCl inhibited phosphorylation of SaeR to an extent similar to that observed in the MN8 ΔsaeS deletion mutant; the inhibition was more prominent by 18 h (Fig. 6A). Using Sortase A (SrtA) as an internal normalization control, PP-HCl decreased the expression of SaeR at both 4 and 18 h of incubation. Again this was similar to the MN8 ΔsaeS strain (Fig. 6, B and C). PP-HCl also reduced phosphorylation of SaeR compared with untreated wildtype cells, whereas phosphorylated SaeR was not detectable in the MN8 ΔsaeS deletion mutant (Fig. 6, D and E). As these results indirectly suggested that PP-HCl may potentially interfere with the ATP-binding site of SaeS, a competition assay between ATP and PP-HCl in the presence of purified SaeS and SaeR was designed. Low levels of ATP were not sufficient to compete against PP-HCl (Fig. 6F, compare lanes 4 and 5) whereas increased ATP concentrations were able to counteract the inhibition of PP-HCl (Fig. 6F, compare lanes 5 and 7). We note that we attempted to measure direct binding using isothermal titration calorimetry, but the poor solubility of PP-HCl in aqueous solution prevented us from doing so. Taken together, these data demonstrate that PP-HCl can inhibit the phosphorylation of SaeR and further suggest that PP-HCl may compete for the ATP-binding site of SaeS.

Figure 6.

PP-HCl targets the SaeRS TCS through competition for the ATP-binding site of SaeS. Phosphorylated SaeR levels were assessed using Phos-tag Western Blot both at 4- or 18-h incubation of S. aureus MN8 strains in the various conditions. S. aureus MN8 ΔsaeS was used as a negative control for phosphorylation of SaeR. A, representative figure of the biologically triplicated experiment is presented. Molecular weight markers are indicated on the left. B–E, calculations of relative protein levels were analyzed against sortase A (SrtA). Total and relative phosphorylated SaeR levels are compared at 4- and 18-h incubation, respectively. Results are presented as the mean of SaeR levels ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparison test was performed (∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001). F, phosphorylation assay of recombinant SaeR by MBP-SaeS in the competition of PP-HCl with ATP tested in near Km ATP concentration (200 μM; indicated by +) or in excess ATP (10 mM; indicated by ++). Molecular weight markers are indicated on the left. The results are presented as the mean of relative phosphorylated SaeR ± SD. Ordinary one-way ANOVA with Tukey’s multiple comparison test was performed (∗∗∗∗p < 0.0001).

Discussion

Prevention of mTSS has been a concern since the epidemic in the early 1980s (11), and although the incidence of mTSS has remained relatively low, outcomes from mTSS cases can be devastating and life-threatening (13). Preventing the production of TSST-1 during menses is important to avoid mTSS and product components must be evaluated to promote tampon safety, including their biocompatibility and chemical safety, impact on the vaginal mucosa and microbiota, as well as effects on S. aureus growth and TSST-1 production (48). Nevertheless, multiple conditions can still occur during the menstrual cycle allowing for the overproduction of TSST-1 to result in mTSS. To expand approaches to further prevent mTSS, we developed a screen to discover antivirulence molecules that inhibit TSST-1 production from S. aureus (Fig. 1). This screen used the prototypical mTSS strain S. aureus MN8 (21) coupled with growth in an environment designed to mimic vaginal conditions that could lead to mTSS (12). Here, we report the discovery of PP-HCl as a potential new antivirulence molecule that represses exotoxin production in S. aureus by inhibiting kinase activity of the SaeRS TCS.

In this study, we mimicked the conditions of the vaginal tract during mTSS including increased oxygenation, near-neutral pH, and decreased glucose content. In these conditions, we noted differential activity of known regulators of the toxin TSST-1. Compared to prior experiments where the media were not adapted to the environmental cues related to mTSS, mutations in Rot, SarA, and SrrAB reduced overall tst expression. In standard laboratory media, these factors function as repressors (Fig. 4) (20, 24, 35). These data solidify our previous findings that CcpA is the main repressor within the vaginal niche and that decreasing glucose levels during menses explains the de-repression of TSST-1 within this short time frame (12). This also highlights the intricate nature of regulatory networks within a bacterium and the importance of mimicking the disease-prone environment when studying the expression of virulence factors.

PP-HCl is an azo dye and analgesic that is used for the treatment of symptomatic urinary tract infections (38) and this compound was associated with inhibition of phosphatidylinositol kinases involved in nociception, explaining its analgesic effects (47). Interestingly, we noted that PP-HCl also functioned to limit T cell activation although further experiments are necessary to understand the mechanisms behind this phenomenon. The cytoplasmic component of bacterial histidine kinases, including SaeS, generally contains two conserved domains: one domain is involved in both autophosphorylation and phospho-transfer to a cognate response regulator, whereas the second domain acts as the catalytic ATP-binding domain (49). Following the S. aureus RNA-seq experiments in the presence of PP-HCl that suggested a role for SaeRS in the exotoxin repressive phenotype (Fig. 5, A and B), we confirmed these findings by challenging MN8 strains expressing either wildtype or the constitutively active SaeS_L18P variant with the inhibitory molecule. When the ΔsaeS strain was complemented with constitutively active SaeS_L18P in the presence of PP-HCl, TSST-1 was no longer repressed, suggesting that SaeRS is the main system inhibited by PP-HCl (Fig. 5C). Furthermore, we found that phosphoryl transfer from SaeS to its cognate response regulator SaeR is reduced in the presence of PP-HCl both in S. aureus (Fig. 6, A–E) and an in vitro system using recombinant SaeS and SaeR proteins (Fig. 6F). Excess ATP was able to overcome inhibition of SaeS by PP-HCl suggesting the compound may interfere with ATP binding (Fig. 6F) although future studies are focused on understanding how PP-HCl inhibits this activity.

Although our screen targeted the inhibition of TSST-1 production and identified the SaeRS system as the major pathway targeted by PP-HCl, the transcriptional changes may not be exclusive to SaeRS-dependent gene regulation. From the RNA-seq experiments (Fig. 5A and Tables S1–S3) we focused on altered expression of known virulence factors, and although multiple exotoxins were repressed, we also noted increased expression in sdrE transcripts in the presence of PP-HCl (Fig. 5A) that were no longer induced in the MN8 ΔsaeS strain (Fig. 5B). However, comparing wild-type S. aureus MN8 with the ΔsaeS mutant in the absence of PP-HCl, sdrE transcripts were not markedly altered (Fig. 5B), and furthermore, we did not observe a canonical SaeR binding motif (42) upstream of the sdrE gene. SdrE is an S. aureus surface protein that functions to inhibit complement activation by high-affinity binding of the complement regulatory protein Factor H (50), and thus PP-HCl could potentially enhance virulence under certain situations. Nevertheless, most PP-HCl repressed virulence factors overlapped with reduced expression from the ΔsaeS mutant in the absence of PP-HCl (Fig. 5B), and these repressed virulence factors were not altered in the ΔsaeS mutant in the presence of PP-HCl (Fig. S3). Additional key TCS regulators of TSST-1 include both the agr and srrAB systems (24, 39); however, transcripts from each of these systems were not repressed in the presence of PP-HCl, and since autoregulation is a hallmark of TCSs, this further suggests they are not involved in the PP-HCl-dependent repression of TSST-1.

Anti-virulent molecules targeting sensing systems from S. aureus have already shown potential. As previously mentioned, TCSs including the Agr system, ArlRS, GraRS and also SaeRS, have been subjects of prior investigation. In the context of preventing risks of mTSS, we demonstrated previously that repression of TSST-1 is mainly due to glucose sensed through CcpA (12) and, conversely, that TSST-1 is predominantly activated by SaeRS (35). We consider SaeRS to be an exceptional target as its activation has been associated with the positive regulation of several adhesins and toxins that result in attenuated virulence in vivo (35, 51, 52, 53, 54, 55). Furthermore, the IsdB iron uptake system was downregulated in the presence of PP-HCl and this system was important for the colonization of S. aureus USA300 in the mouse vaginal tract (56). Another advantage for targeting SaeRS compared to the Agr TCS is that, to our knowledge, no strains with a non-functional SaeRS TCS have been isolated from patients (57). We hypothesize that such strains would become avirulent and less prone to colonize the vaginal niche due to reduced toxin and adhesin expression. Altogether, SaeRS may be a primary choice for targeting superantigen-specific diseases such as mTSS. A key advantage of targeting virulence rather than bacterial growth is to limit antibacterial off-target activity which may otherwise lead to expansion of antimicrobial resistance (58). Indeed, PP-HCl did not induce the killing of S. aureus (Fig. 2), or key representative species of vaginal lactobacilli (Fig. 3) which may also be advantageous for preventing mTSS (37). Thus, PP-HCl may represent a new lead compound and strategy to further develop safer menstrual products to prevent the occurrence of mTSS.

Experimental procedures

Ethics statement

Human blood from healthy donors was obtained in accordance with the human subject protocol HSREB 110859 approved by the London Health Sciences Centre (LHSC) research ethics board at the University of Western Ontario. These studies abide by the Declaration of Helsinki principles. Volunteers were recruited by passive advertising through the Department of Microbiology and Immunology at the University of Western Ontario and all volunteers gave a written informed consent before each sampling. Each sample was fully anonymized and no information regarding the identity of the donor was retained.

Bacterial strains and screening of bioactive molecules

The list of strains used for this study is found in Table 1. Routine growth of S. aureus MN8 and derivatives was done aerobically at 37 °C in tryptic soy broth (TSB) with shaking at 250 rpm supplemented with the appropriate antibiotics as needed. For experiments measuring the production of TSST-1, S. aureus strains were grown in a vaginally defined medium (VDM) (59) modified to contain 700 μM glucose to mimic conditions favorable for the production of TSST-1 (12). The luciferase screen used S. aureus MN8 harboring P_tst reporter in fusion with the lux operon (pAmilux::P_tst). Briefly, bacteria were grown and the OD₆₀₀ was adjusted to 0.01 in fresh media and distributed in 384-well microplates using the Tempest dispenser (Formulatrix). An expression control was the reporter strain in the media of assay and the repression control was the reporter strain in assay media containing 60 mM glucose known to repress TSST-1 (12). Small molecules were pre-inoculated (at 10 μM) before adding the bacterial inoculum and 3920 molecules were tested. Both luminescence and OD₆₀₀ were measured every hour for 18 h during incubation at 37 °C with continuous agitation. For each molecule, the point with the highest luminescence was compiled with its respective OD₆₀₀. Seventy initial compounds with normalized growth equal to or better than 95% of the expression control and with raw luminescence lower than 100 RLU were selected. Known antimicrobials in this list were removed as the goal of the study is to identify molecules with anti-virulent but not antimicrobial properties.

S. aureus MN8 containing pAmilux::P_tst were grown as above in 96-well plates and subsequently challenged with remaining compounds and titrated from 0 μM to 50 μM and both luminescence and OD₆₀₀ were measured every hour for 18 h in a Biotek Synergy H4 multimode plate reader. Relative luminescence units (RLUs) were calculated from the area under the curve of the luminescence over the area under the curve of the absorbance during the same time. For later assays, only 0 and 50 μM PP-HCl in 700 μM glucose VDM were used with the various S. aureus MN8 strains. Western immunoblots for TSST-1 production were performed as previously described (35). Briefly, each strain tested was grown in 700 μM glucose VDM in presence or absence of PP-HCl at 37°C with agitation (250 rpm) for 4 or 18 h. Each sample was normalized to 12 OD₆₀₀ units and precipitated using trichloroacetic acid (TCA), washed with cold acetone, and resuspended in 8M urea. Samples were migrated on 12% acrylamide SDS-PAGE, then either stained using Ready Blue (Sigma-Aldrich) or transferred on PVDF membrane for anti-TSST-1 Western Blot.

Community State Types (CSTs) growth and lactic acid production

Growth of lactobacilli was assessed as previously described (19). Representatives of stable/healthy community state types were grown overnight in De Man, Rogosa and Sharpe (MRS) media, then subcultured in 60 mM glucose VDM supplemented with or without 50 μM PP-HCl at a starting OD₆₀₀ of 0.05. Cultures were incubated for 20 h at 37 °C without agitation. Growth was assessed by OD₆₀₀ readings every hour in a Biotek Synergy H4 multimode plate reader. Lactic acid production was assessed for each representative strain after 24 h incubation in the same conditions as for the growth assessment using the Lactate-Glo Assay (Promega) following the manufacturer’s protocol.

IL-2 quantification and viability of PBMC

PBMCs were isolated with Ficoll Paque Plus following the manufacturer’s protocol and seeded for a final concentration of 1 × 10⁶ cells/ml in 96-well plates. Cells were either challenged for 18 h at 37 °C in 5% CO₂ with S. aureus MN8 or MN8 ΔsaeS filtered supernatants exposed to 0 and 50 μM PP-HCl, 100 ng/ml of purified recombinant TSST-1 or with 1 μl of Dynabeads human T-activator CD3/CD28 (Gibco) (24). IL-2 was measured by human IL-2 ELISA (Invitrogen) using the manufacturer’s instructions. Data at a supernatant dilution factor of 1/6250 were normalized over the control condition and plotted in biological triplicates. Results are presented as the percentage of IL-2 production compared to their respective controls.

The viability of PBMCs was assessed by incubating the challenged cells for an additional hour with the resazurin-based PrestoBlue HS Cell Viability Reagent (Invitrogen) following the manufacturer instructions and plates were read for fluorescence at excitation of 560 nm and emission of 590 nm and for absorbance of 570 nm in biological triplicates. Results are presented as the percentage viability compared to their respective controls.

RNA-seq experiments

RNA-seq was performed as described previously (12, 19). Briefly, S. aureus MN8 or MN8 ΔsaeS were grown with 0 or 50 μM PP-HCl in 700 μM glucose VDM for 4 h, RNA was extracted using RNeasy Plus Mini kit (QIAGEN) and subsequently treated with Turbo DNA-free Kit (Ambion). RNA-seq and its comparative analysis was performed by SeqCenter. Twelve million paired-end Illumina sequencing was performed and followed by analysis as previously described (12). Raw RNA data were deposited at NCBI under BioProject accession no. PRJNA1080564.

Modification of the leucine to proline at position 18 within SaeS

A DNA fragment from the plasmid pALC2073::saeQRS overlapping position 18 was amplified using Phusion DNA polymerase (Thermo-Fisher) with the primers SaeSL18P_For (5′-tggtcatgaagtccctatgcgtattaagga-3′) and SaeSL18P_Rev (5′- gctaaaatagttgaagttaatggtatactcgatacgacgc-3′). This DNA product was annealed to pALC2073::saeQRS to obtain the complete vector using inverse PCR with the modification from leucine to proline in saeS. The final products were digested with the restriction enzyme DpnI to remove residual unmodified plasmids and transformed into Escherichia coli XL1 Blue. The modification of saeS in the final vector was confirmed by DNA sequencing (Plasmidsaurus). pALC2073::saeQRS_L18P was transformed into MN8 ΔsaeS and compared with MN8 ΔsaeS containing the original pAL2073::saeQRS when exposed to 50 μM PP-HCl in 700 μM glucose VDM. Strains were grown for 18 h and extracellular proteins were precipitated using TCA and assessed by anti-TSST-1 Western Blot as described previously.

Determination of SaeR phosphorylation state

S. aureus MN8 was grown in 700 μM glucose VDM supplemented with or without 50 μM PP-HCl and incubated for 4 or 18 h at 37 °C with agitation. S. aureus MN8 ΔsaeS was grown in the same conditions as a control for SaeR phosphorylation. Separation of SaeR and SaeR∼P was performed as described previously using 12% polyacrylamide gels containing the acrylamide-pendant Phos-tag ligand (52). Briefly, whole cell extracts were obtained by resuspending cell pellets in cell extract buffer (20 mM Tris [pH 7.0], 1× Protease Inhibitor Cocktail Set I (Sigma-Aldrich)) and transferred to sterile screw cap tubes containing silica beads. Cells were homogenized at room temperature using a Precellys 24-bead beater (Bertin technologies) for 3 cycles of 6500 rpm, 30 s each, and centrifuged. Whole-cell extracts were normalized by protein concentration (A₂₈₀) to 100 μg and electrophoresed on Phos-tag gels with standard running buffer (0.1% [w/v] SDS, 25 mM Tris, 192 mM glycine) at 4°C under constant voltage (150 V) for 2 h. Gels were washed for 15 min with transfer buffer (25 mM Tris [pH 8.3], 192 mM glycine, 20% methanol) containing 1 mM EDTA followed by a second wash without EDTA to remove manganese ions. Proteins were then transferred to PVDF membranes (Cytiva) and incubated with polyclonal rabbit antibodies to SaeR (1.5:1000) for 1 h. Membranes were then washed with TBST and incubated with StarBright Blue 700 goat anti-rabbit IgG (1:3500; Bio-Rad) for 1 h. Membranes were washed in TBST and signals were visualized using an Amersham ImageQuant800. The densities of the SaeR∼P signal were determined by quantification with Multi Gauge software (FujiFilm). The data are representative of three independent experiments, and a representative image is shown.

In vitro kinase assays using recombinant protein were performed as previously described (52). To test the effect of PP-HCl on kinase activity in vitro, 50 μM PP-HCl was added to the standard reaction mixture (5 μM MBP-SaeS, 10 μM SaeR-His₆, 200 μM ATP, 1× TKM buffer) and incubated at 37 °C for 1 h. The reaction was stopped by the addition of 5× SDS loading buffer. Phosphorylated and unphosphorylated forms of SaeR were separated using 12% phos-tag acrylamide gels and visualized by Coomassie blue staining. The resulting gels were imaged using an Amersham ImageQuant800, and the levels of SaeR∼P were determined by quantification with Multi Gauge software.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 10. Ordinary one-way ANOVA was used without correction for multiple comparisons unless mentioned otherwise.

Data availability

All data are contained within the article except the RNA-seq datasets that were deposited at NCBI under BioProject accession no. PRJNA1080564.

Supporting information

This article contains supporting information.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Reviewed by members of the JBC Editorial Board. Edited by Chris Whitfield