Manganese Detoxification by MntE Is Critical for Resistance to Oxidative Stress and Virulence of Staphylococcus aureus

Manganese (Mn) is generally viewed as a critical nutrient that is beneficial to pathogenic bacteria due to its function as an enzymatic cofactor and its capability of acting as an antioxidant; yet paradoxically, high concentrations of this transition metal can be toxic. In this work, we demonstrate Staphylococcus aureus utilizes the cation diffusion facilitator (CDF) family protein MntE to alleviate Mn toxicity through efflux of excess Mn. Inactivation of mntE leads to a significant reduction in S. aureus resistance to oxidative stress and S. aureus-mediated mortality within a mouse model of systemic infection. These results highlight the importance of MntE-mediated Mn detoxification in intracellular Mn homeostasis, resistance to oxidative stress, and S. aureus virulence. Therefore, this establishes MntE as a potential target for development of anti-S. aureus therapeutics.

maintenance of optimal intracellular metal levels are paramount for pathogen survival and replication during infection. Recent evidence points to Mn as a key micronutrient that contributes to the pathogenesis of a variety of bacterial species (1,2). This includes the Gram-positive bacterium Staphylococcus aureus, which is responsible for a diverse set of devastating diseases, including skin and soft tissue infections, osteomyelitis, septicemia, pneumonia, and endocarditis (1,3).
The microbial requirement for nutrient metal is exploited by the vertebrate immune system, which produces metal-sequestering proteins that restrict divalent cations, including Mn, from bacterial pathogens in a process termed nutritional immunity (4,5). This is best exemplified by the antimicrobial effects of the host protein calprotectin, which binds Mn, Zn, and other divalent cations and sequesters these metals at sites of infection (6)(7)(8)(9). Calprotectin, a heterodimer of S100A8 and S100A9, constitutes the majority of the cytoplasmic proteome in neutrophils (10). However, the bioavailability of Mn within the host environment can vary widely. Host tissues inherently exhibit differences in Mn levels (11,12). In healthy human hosts, Mn concentrations in blood and serum typically range from 20 to 200 nM (13). Liver and kidneys (20 to 50 nmol/g) are Mn rich compared to other organs, such as the brain, heart, and lungs, which have Mn concentrations of Ͻ20 nmol/g (13). Moreover, increases in dietary Mn significantly enhance the Mn content of the liver and heart and can impact the outcome of infection (11,14). This suggests S. aureus must be capable of responding to altered Mn levels during infection.
As metals can neither be synthesized nor degraded, modulation of transport of Mn is an important process to maintain optimal intracellular Mn concentrations in S. aureus. S. aureus encodes two Mn acquisition systems: MntABC, an ATP-binding cassette Mn 2ϩ permease, and the NRAMP homolog MntH (15). Expression of these systems in response to Mn limitation is controlled by the metalloregulatory protein and member of the DtxR family MntR (15). MntR functions as a transcriptional repressor of mntABC, preventing expression under conditions of sufficient Mn, and as an activator of mntH, upregulating expression under low-Mn conditions (15). Genetic inactivation of both Mn uptake systems results in reduced bacterial burdens in the liver and kidneys, and inactivation of mntC alone reduces virulence in an S. aureus sepsis model of infection (11,16). Moreover, S. aureus contains genes (sodA and sodM) that encode two superoxide dismutase (SOD) enzymes that utilize Mn as a cofactor and are important contributors to the ability of S. aureus to survive neutrophil-mediated killing (9). SodA strictly requires Mn, whereas SodM utilizes either Mn or Fe, and MntABC-dependent Mn import significantly enhances SOD activity in vitro (16,17). Combined, these results underscore the importance of Mn for S. aureus pathogenesis and resistance to oxidative stress.
Despite their importance to cellular physiology, an overabundance of essential transition metals, including Mn, is toxic and can lead to cell death. Although the mechanisms are not entirely defined, metal intoxication is predicted to be a result of mismetallation of noncognate enzymes and transcriptional regulators, inappropriate signaling to other metal transporters that are regulated allosterically, and/or disruption of redox cycling (18). Recent studies in Streptococcus spp. and Neisseria meningitidis suggest Mn detoxification through Mn efflux is an important component of bacterial pathogenesis. Streptococcus pyogenes is more susceptible to killing by human neutrophils in vitro when the gene encoding the Mn efflux pump MntE, a member of the cation diffusion facilitator (CDF) protein family, is deleted (19). Moreover, mntE mutants of S. pyogenes, Streptococcus pneumoniae, and Streptococcus suis demonstrate reduced pathogenesis in murine infection models (19)(20)(21). Mn efflux mediated by MntX is also important for N. meningitidis survival in vivo (22). Thus, sensing and tight regulation of intracellular Mn levels through both import and efflux are critical for bacterial survival within the host environment.
How S. aureus detoxifies excess Mn and what the consequences are of Mn intoxication have yet to be defined. It is estimated that approximately 30 to 45% of all known enzymes require a metal cofactor; therefore, intracellular metal levels must be tightly controlled to ensure proteins are populated with the proper metal species (18). S. aureus is capable of tolerating millimolar concentrations of Mn, implying S. aureus encodes a Mn detoxification system; however, the mechanism by which S. aureus senses and responds to Mn stress is unknown. Moreover, it is unclear in what environmental niches Mn detoxification is most important for S. aureus survival and pathogenesis.
In this study, we sought to identify the mechanisms by which S. aureus maintains intracellular Mn homeostasis to prevent metal intoxication and the contribution of Mn detoxification to staphylococcal pathogenesis. We describe a predicted CDF family protein (NWMN_2316) we have designated MntE, which functions to detoxify S. aureus of excess Mn and maintain optimal intracellular Mn concentrations. Under conditions of excess exogenous Mn, upregulation of mntE is the primary transcriptional response, and this is dependent on the transcriptional regulator MntR. Without MntE, S. aureus cells accumulate intracellular Mn, leading to a reduced tolerance to oxidative stress and a reduction in S. aureus-mediated lethality during systemic infection. Together, the findings in this report reveal Mn detoxification by MntE is critical for maintenance of optimal intracellular Mn levels and full virulence of S. aureus.

RESULTS
Transcriptional response of S. aureus to excess exogenous Mn. To identify potential candidate genes involved in the detoxification of excess Mn and maintenance of intracellular Mn homeostasis, we investigated the transcriptional response of S. aureus cells exposed to 1 mM MnCl 2 compared to untreated cells by using RNA sequencing (Fig. 1). Total transcript abundance was unchanged for the majority of S. aureus genes (see Table S1 in the supplemental material). A single transcript, NWMN_2316, which is predicted to encode a 288-amino-acid membrane protein in the CDF family, exhibited a significant 3.3-fold increase in response to Mn treatment. Other transcripts significantly changed include genes encoding the well-characterized Mn ABC transporter, MntABC, which exhibited a 3.5-to 3.6-fold decrease in transcript abundance. As NWMN_2316 was the transcript with the highest positive fold change in response to Mn treatment and proteins of the CDF family function as cation efflux pumps, we hypothesized this gene is an important component of the Mn detoxification strategy of S. aureus. Based on these observations and the data described below, we renamed NWMN_2316 as mntE.
MntE and its genomic context are conserved across staphylococci and other Gram-positive pathogens. Orthologs of mntE were found in the genomes of all S. Impact of Mn Detoxification on S. aureus Homeostasis ® aureus strains and staphylococcus species available within the SEED Viewer database ( Fig. 2A). Moreover, the genomic context was also conserved, as genes downstream of mntE, including gpmA, which encodes the Mn-requiring enzyme 2,3-bisphosphoglycerate-dependent phosphoglycerate mutase, were found in all species examined. Genes involved in cystine transport (tcyABC) and cell wall remodeling (fmhA) were also conserved. In contrast, upstream regions were not conserved across staphylococcal species.
Protein sequence alignment and phylogenetic analysis of S. aureus MntE with other characterized bacterial Mn efflux proteins revealed this protein is closely related to MntE of Streptococcus spp. (Fig. 2B), sharing 41 to 80% sequence identity. Consistent with other known bacterial efflux pumps of the CDF family, the predicted secondary structure of MntE contains six putative transmembrane domains (Fig. 2C). Furthermore, residues previously identified as critical for Mn binding in Streptococcus pneumoniae MntE are conserved in the S. aureus CDF protein ( Fig. 2C; see Fig. S1 in the supplemental material) (23). Taken together, these data indicate MntE-mediated detoxification of Mn is a common homeostatic strategy among staphylococci and other bacterial species.
MntR is required for upregulation of mntE transcription in response to exogenous Mn. Due to the role of Mn as an essential nutrient and the inherent toxicity of excess metals, we hypothesized mntE expression is tightly regulated. MntR is a transcriptional regulator of the DtxR family and the only staphylococcal regulator known to control the expression of genes encoding the Mn importers MntABC and MntH (15). Members of the DtxR family can function as bi-regulators, and MntR is known to control expression of Mn transport systems in other bacteria (24)(25)(26). Therefore, we hypothesized MntR may also play a role in the regulation of S. aureus MntE. Expression of mntE was significantly increased in wild-type (WT) S. aureus compared to a ΔmntR mutant in response to 1 mM MnCl 2 compared to untreated controls, as measured by quantitative reverse transcriptase PCR (qRT-PCR) (Fig. 3A). Conversely, no change was detected in mntA transcription, which is upregulated under Mn-limited conditions, demonstrating 1 mM MnCl 2 was sufficient to suppress the S. aureus Mn starvation response. To confirm these results, we created a bioluminescent reporter plasmid with 350 bp upstream of the predicted translational start site of mntE driving the expression of a luciferase operon, and measured mntE reporter activity. Luminescence significantly increased in a dose-dependent response to increasing concentrations of MnCl 2 in WT strains, but not in the ΔmntR strain, compared to promoterless controls (Fig. 3B). Combined, these data indicate S. aureus senses increasing concentrations of Mn and suggest that MntR contributes to the transcriptional activation of mntE expression to alleviate Mn toxicity.
S. aureus requires mntE for Mn detoxification. Metal toxicity that results in growth arrest is thought to be a result of mismetallation of noncognate enzymes, the dysregulation of metal-dependent signaling, and/or the disruption of redox cycling (18). Therefore, we hypothesized inactivation of mntE would disrupt Mn detoxification and lead to increased sensitivity to exogenous Mn. Inactivation of mntE resulted in substantially decreased growth on solid media and in liquid cultures when supplemented with 0.5 or 1 mM MnCl 2 ( Fig. 4A and B). Due to our observation that MntR is required for activation of mntE transcription in response to excess Mn and MntR represses mntABC under sufficient-Mn conditions, we predicted inactivation of mntR would also result in increased sensitivity to Mn. ΔmntR and mntE::Tn ΔmntR mutants exhibited decreased growth on solid media and liquid cultures supplemented with MnCl 2 , with the mntE::Tn ΔmntR mutant experiencing the most attenuated growth in liquid cultures ( Fig. 4A and B). No differences in growth were observed for strains with Change in bioluminescence was determined for S. aureus strains following incubation for 3 h in medium containing increasing amounts of MnCl 2 . The data are the averages from three independent experiments, each in biological triplicate with the standard error of the mean shown. Statistical significance was determined using a two-way ANOVA with Dunnett's correction for multiple comparisons, comparing luminescence data for WT p.P mntE luxABCDE and ΔmntR p.P mntE luxABCDE against WT p.luxABCDE for each condition. ***, P Ͻ 0.001; ****, P Ͻ 0.0001. Impact of Mn Detoxification on S. aureus Homeostasis ® mntC and mntH inactivated compared to WT S. aureus, demonstrating the growth conditions contained sufficient Mn levels. Constitutive expression of mntE in trans completely restored growth of both ΔmntR and mntE::Tn mutants in the presence of 1 mM MnCl 2 , further supporting the hypothesis that mntE is required for Mn detoxification and MntR is required for transcriptional activation of mntE (Fig. 4C).
Many CDF proteins are capable of removing multiple cation species from the cytoplasm, including the well-characterized YiiP of Escherichia coli (Zn/Fe) and CzcD of Bacillus subtilis (Zn/Co/Cd) (27)(28)(29). To investigate the specificity of S. aureus MntE, we measured intracellular metal levels of S. aureus strains grown in rich media following an overnight outgrowth in rich media (metal replete) or media treated with EDTA (metal deplete) using inductively coupled plasma mass spectrometry (ICP-MS) (Fig. 5). Under both metal-replete and metal-deplete conditions, strains inactivated for mntE accumulated significantly higher concentrations of intracellular Mn compared to WT S. aureus; however, there was no difference in intracellular Mn levels between the ΔmntR and WT strains. This phenotype can be complemented when mntE is constitutively expressed in trans (see Fig. S2A in the supplemental material). Interestingly, inactivation of mntR led to significantly higher levels of intracellular Fe, which can be suppressed by inactivation of mntE. Significantly lower intracellular Fe levels were observed in mntE::Tn and ΔmntR mntE::Tn strains compared to WT S. aureus, implying an inverse relationship between intracellular Mn and Fe levels ( Fig. 5A and B). When mntE is constitutively expressed in trans, Fe levels return to WT levels in the mntE::Tn strain and the mntE::Tn ΔmntR phenotype mimics that of the ΔmntR strain (Fig. S2B). No differences were observed in the intracellular concentration of other metals, including Zn, copper (Cu), cobalt (Co), , and Cu (F) were measured using inductively coupled plasma mass spectrometry and normalized to intracellular sulfur for S. aureus strains grown under metal-replete (TSB) and metal-deplete (TSB plus 100 mM EDTA) conditions. The data are the average of the means from three independent experiments, each in biological triplicate with error bars representing standard error. *, P Ͻ 0.05, **, P Ͻ 0.01, and ****, P Ͻ 0.0001, calculated by two-way ANOVA with Dunnett's correction for multiple comparisons, comparing metal levels for each strain with WT S. aureus. n.s., not significant. Impact of Mn Detoxification on S. aureus Homeostasis ® and nickel (Ni) among strains, suggesting MntE efflux is preferential to Mn, and Mn efflux affects intracellular Fe levels ( Fig. 5C to F).
Mn efflux is required for resistance to oxidative stress. Despite the ability of Mn to serve as an antioxidant, excess Mn is inherently toxic, possibly due to increased oxidative stress. Thus, we hypothesized the increase in intracellular Mn by direct loss of mntE, or by reducing mntE expression through inactivation of mntR, would alter the ability of S. aureus to resist oxidative stress. To test this hypothesis, we performed an oxidative pulse experiment in which S. aureus strains were exposed to either the oxidative stressor sodium hypochlorite (NaOCl) or phosphate-buffered saline (PBS) for 30 min, serially diluted, and then plated onto tryptic soy agar (TSA) (Fig. 6A). NaOCl produces hypochlorous acid (HOCl), an oxidant that is produced within the phagolysosome of macrophages and neutrophils as part of the respiratory burst used to kill invading pathogens. Inactivation of both mntE and mntR led to alterations in colony phenotype compared to WT or single mutant strains when cells were statically exposed to 25 M NaOCl, where colonies formed by the double mutants were smaller. This phenotype was enhanced in the mntE::Tn ΔmntR mutant following treatment with 50 M NaOCl. No differences in colony morphology were observed between strains treated with PBS. Additionally, we assessed the ability of S. aureus strains to grow in the presence of NaOCl (Fig. 6B). Inactivation of both mntE and mntR resulted in a severe growth defect compared to the WT, mntE::Tn, or ΔmntR strain following a 50 M NaOCl treatment. Therefore, we concluded maintenance of Mn homeostasis coordinated by mntE and mntR contributes to the resistance of S. aureus to NaOCl-induced oxidative stress. Mn efflux contributes to S. aureus-mediated lethality and pathogenesis. During infection, Mn is required for S. aureus survival and proliferation within the host; however, it is unclear what role Mn efflux serves in S. aureus-mediated pathogenesis and virulence. To investigate the contribution of mntE to S. aureus virulence, 6-week-old female BALB/cJ mice were retro-orbitally infected with S. aureus and monitored for 10 days (Fig. 7A). The mortality rate following systemic infection was significantly reduced when mice were infected with mntE::Tn or mntE::Tn ΔmntR strains, but not the ΔmntR strain, compared to WT-infected animals (Fig. 7A). To assess the contribution of Mn homeostasis to bacterial dissemination and survival, 6-week-old female BALB/cJ mice were retro-orbitally infected with S. aureus and monitored for 4 days postinfection, and bacteria were enumerated from the heart, kidneys, and liver. Despite the difference in mortality rate, few notable differences in organ bacterial burdens were observed among mice infected with S. aureus WT and mutant strains (Fig. 7B). There were significantly fewer mntE::Tn mutant bacteria recovered in the kidneys and significantly fewer ΔmntR bacteria recovered from the hearts; however, no other significant differences were observed among strains in those tissues. No differences were observed in the number of bacteria recovered from the livers of mice infected with the various strains. Moreover, increasing the length of the infection to 10 days did not reveal any significant differences in bacterial burdens between S. aureus strains in murine hearts, livers, or kidneys (see Fig. S3 in the supplemental material).
Inactivation of mntE reduces fitness of S. aureus strains in vivo. The loss of mntE results in reduced resistance to NaOCl-mediated oxidative stress; therefore we hypoth- Impact of Mn Detoxification on S. aureus Homeostasis ® esized the reduction in mortality observed in mice infected with the mntE::Tn mutant S. aureus strains could be a result of increased susceptibility to oxidative damage. mntE::Tn and mntE::Tn ΔmntR strains from heart and kidney homogenates plated onto TSA exhibited a similar irregular colony phenotype (Fig. 8A), as was observed when strains were treated with NaOCl (Fig. 6A). However, this phenotype is reversible as mntE::Tn and mntE::Tn ΔmntR colonies revert back to the WT colony morphology  following additional incubation or a single in vitro passage (see Fig. S4 in the supplemental material). Thus, we predicted that strains inactivated for mntE were weakened by in vivo growth and exposure to an oxidative stressor would further damage these strains. Unlike NaOCl, each strain exhibits equal sensitivity to the superoxide-producing agent paraquat when grown in vitro (Fig. 8A). Therefore, we hypothesized decreased fitness due to mntE inactivation and subsequent damage during in vivo passage would result in increased sensitivity to paraquat compared to the WT strain. To test this, we exposed bacteria from murine kidney and heart homogenates that were infected with WT, mntE::Tn, ΔmntR, and mntE::Tn ΔmntR strains to paraquat (Fig. 8B). All strains cultured directly from organ homogenates demonstrated similar growth kinetics in vehicle-treated media. Conversely, following in vivo passage, strains inactivated for mntE or both mntE and mntR displayed extended lag phases in vehicle-treated media alone, despite equal numbers of CFU serving as the experimental inoculum (see Fig. S5 in the supplemental material). Moreover, this phenotype was dramatically enhanced by the addition of paraquat. Taken together, these data suggest MntE is critical for fitness of S. aureus in vivo, and that S. aureus must efflux Mn in vivo to resist oxidative damage.

DISCUSSION
As a required nutrient for life, Mn acquisition is critical for bacterial survival within the vertebrate host. Accordingly, many species have evolved high-affinity transporters to enhance acquisition of Mn and promote survival within the host environment. S. aureus relies on the ABC-type MntABC and NRAMP MntH transporters to import Mn, which enable the pathogen to counteract the antimicrobial effects of calprotectinmediated nutritional immunity (11). Insufficient intracellular Mn can have wide-ranging consequences, affecting Mn-dependent enzymes involved in transcription, metabolism, and defense against oxidative stress (9,15,16,(30)(31)(32). In this regard, it is unsurprising that inactivation of these transporters reduces S. aureus organ burdens during a murine model of infection (11). However, the inherent toxicity of excess Mn coupled with previous work establishing the contribution of Mn efflux in streptococcal pathogenesis and the number of putative Mn-specific efflux proteins that were recently inferred using comparative bioinformatics in other bacterial pathogens underscores the importance of Mn detoxification to bacterial pathogens (19)(20)(21)33).
In this report, we identify a CDF protein we have named MntE that is required for S. aureus detoxification of Mn and is conserved among staphylococcal species (Fig. 2). MntE functions to alleviate excess Mn from the cytoplasm, and inactivation of mntE renders S. aureus susceptible to Mn toxicity and subsequent growth arrest ( Fig. 4 and  5). Upregulation of mntE and downregulation of mntABC expression are the predominant transcriptional response to the addition of exogenous 1 mM MnCl 2 , demonstrating MntE is the major Mn detoxification system under these conditions ( Fig. 1 and 2). The S. aureus DtxR protein MntR bidirectionally regulates the import and export of Mn by repressing mntABC and activating mntE expression in response to endogenous Mn (Fig. 2) (15). This implies S. aureus intracellular Mn levels are tightly controlled by modulating Mn import and export in response to environmental conditions. However, a strain inactivated for mntR does not accumulate intracellular Mn to the same extent as an mntE::Tn mutant (Fig. 5A). This could be explained by the basal transcription of mntE we observed in WT S. aureus grown in tryptic soy broth (TSB) (Table S1), but it is also possible mntE is regulated by additional mechanisms.
During infection, oxidative stress is an important challenge S. aureus must overcome to survive. The oxidants generated by neutrophils react with macromolecules to damage cellular lipids, proteins, and DNA. Mn-cofactored superoxide dismutases are critical for resisting phagocyte-produced superoxide, and low-molecular-weight Mn complexes have antioxidant properties that may also serve to detoxify reactive oxygen species (9,34). The results reported here indicate elevated intracellular Mn following inactivation of mntE also leads to enhanced susceptibility to oxidative stressors in S. aureus ( Fig. 6 and 8). Concordant with these results are observations in Streptococcus spp., where loss of MntE function results in reduced bacterial growth in the presence Impact of Mn Detoxification on S. aureus Homeostasis of hydrogen peroxide or the oxidizing sulfhydryl reagent diamide, for S. pyogenes and S. suis, respectively (19,21). Therefore, disruption of intracellular Mn homeostasis either through excess or depletion can reduce the fitness of bacterial pathogens during infection.
Interestingly, the degree of difference in susceptibility to oxidative stress between S. aureus strains was dependent on the chemical oxidant applied and growth conditions ( Fig. 6 and 8). When grown in a rich medium, strains inactivated for mntE and mntR exhibited reduced growth kinetics compared to WT strains when treated with NaOCl, yet no differences were observed between strains treated with paraquat. Paraquat generates superoxide, which, while not a potent oxidant itself, is a precursor to other damaging molecules, such as H 2 O 2 , and hydroxyl radicals (35,36). NaOCl forms HOCl, which can react with amino acids to produce oxidizing chloramines that can serve as oxidants themselves and damage other cellular macromolecules (37). One possible explanation could be differences in which macromolecules are targeted and the degree of cellular damage that is produced by the presence of each oxidant; however, the mechanisms by which H 2 O 2 , HOCl, and other oxidants induce bacterial killing are still not entirely agreed upon (37). Moreover, S. aureus strains grown in vivo would experience nutrient limitation and more extensive oxidative damage due to host immune defenses, compared to those grown in a rich medium. Thus, the more dramatic fitness defect observed in response to paraquat treatment in strains with inactivated mntE and mntR is likely due to a synergistic effect of combined stressors (Fig. 8). Together, these results emphasize the impact of intracellular manganese homeostasis on S. aureus fitness and stress responses.
One possible mechanism contributing to the fitness defect observed in S. aureus strains inactivated for mntE could be the disruption of the intracellular Mn/Fe ratio. These data demonstrate the function of MntE significantly impacts the cytosolic concentrations of Fe, where the mntE::Tn and ΔmntR mntE::Tn strains had reduced Fe levels compared to the WT and ΔmntR strains (Fig. 5B). Curiously, the ΔmntR mutant contained significantly higher quantities of intracellular Fe, which could be the result of MntABC-mediated transport of Fe. However, previous work in Bacillus anthracis suggests MntC is unable to bind Fe, and thus, direct evidence that MntABC transports Fe is lacking (38). Displacement of Fe by Mn in Fe-requiring proteins would liberate free Fe into the cytoplasm, promoting oxidative damage incurred by Fenton chemistry. Furthermore, Mn has been shown to occupy Fe binding sites of the ferric uptake regulator (Fur) in B. subtilis, aberrantly repressing the expression of Fe uptake systems (39,40). In appropriate repression of Fe acquisition systems due to increased intracellular Mn could explain the differences in cytosolic Fe levels observed between WT and mntEinactivated S. aureus strains (Fig. 5B). Mn concentrations also influence the activity of the Fur family transcriptional regulator PerR, which controls the expression of genes involved in detoxifying reactive oxygen species, as well as iron homeostasis (39,(41)(42)(43)(44). It is possible the increased intracellular Mn in an mntE::Tn mutant binds to PerR, repressing expression of the PerR regulon, including the catalase gene katA and fur (15,45). This would limit the ability of S. aureus to mount an appropriate response to oxidative stress as well as disrupt fur-dependent expression of iron acquisition systems. Additional investigation of the connection between intracellular Mn and Fe levels is needed to uncover the impact of Mn toxicity on metal-related systems.
Although our data provide evidence that S. aureus experiences Mn intoxication in vivo, when and where this occurs during the course of an infection remains elusive. Mn levels among host populations can vary based on exposure and diet (14,46). Moreover, Mn levels are not uniform across individual host tissues, and these levels can significantly impact infection outcomes (11,14). Mice fed diets high in Mn accumulate Mn within the heart, leading to enhanced colonization of that tissue and increased mortality (14). S. aureus is capable of infecting nearly all body sites, suggesting that S. aureus is adapted to survive under a wide range of environmental Mn concentrations. Alternatively, Mn intoxication could be a consequence of the activity of the high-affinity importers MntABC and MntH. Upregulation of these systems in response to calprotectin-mediated Mn limitation could lead to a sudden influx of Mn. Therefore, it is possible MntE is required to buffer the abrupt change in the intracellular Mn concentration to prevent Mn toxicity, in a "binge and purge"-type model. Host-imposed Fe limitation could be another scenario, where Mn efflux is required to ensure a proper Mn/Fe ratio when Fe is restricted. Another possibility by which S. aureus experiences Mn intoxication could be within the phagolysosome. Recent data demonstrate metal intoxication is an antimicrobial strategy employed by phagocytes (47,48). Upon engulfment, Cu and Zn accumulate within the phagocytic vacuole to induce intoxication. However, evidence of exogenous Mn intoxication as a killing strategy has not been uncovered. Instead, it is generally accepted that the phagolysosome is a Mn-poor environment due to the activity of host NRAMP1 transporters that deplete Mn from the phagocytic vacuole (49)(50)(51). Additional work to uncover the environmental niches in which MntE is essential is needed.
Data reported in this article suggest that MntE is a critical contributor to S. aureus virulence. Notably, mortality was significantly reduced for mice infected with mntE::Tn mutants (Fig. 7A). One possible mechanism contributing to the observed virulence defect may involve oxidative damage. As mentioned above, MntE-dependent removal of Mn from the cytoplasm alters S. aureus resistance to oxidative stress, and this phenotype is further exaggerated in mutants with inactivated mntE isolated from murine organs (Fig. 8). The small-colony morphology observed in these mutants is consistent with colony phenotypes previously reported as a response to oxidative damage, suggesting that mntE::Tn mutants experience enhanced oxidative stress in vivo (52). Increased oxidative damage would lead to reduced fitness and a redirection of resources to essential pathways instead of the production of potent, but energycostly virulence factors. Additionally, as proposed above, mismetallation of Fur with Mn or a decrease in fur transcription by way of PerR-mediated repression could also alter virulence factor function. Previous work has determined that Fur positively regulates the Sae and Agr regulatory networks under low-iron conditions (53). Therefore, disruption of Fur-Fe sensing would have a considerable impact downstream on the expression of genes involved in immune evasion, biofilm formation, and oxidative stress response (53)(54)(55)(56)(57). This is further evidenced by the absence of several cytolysins and hemolysins under low-iron conditions and the significant virulence defect in a murine pneumonia model observed when fur is deleted (58). Additional work examining the impact of Mn homeostasis on virulence factor production would provide clarity into the mechanism by which Mn detoxification contributes to S. aureus-mediated lethality.

MATERIALS AND METHODS
Bacterial strains and growth conditions. Bacterial strains (Table 1), plasmids (see Table S2 in the supplemental material), and primers (Table S2) are listed in the specified tables. Unless noted, S. aureus strains were routinely cultured on tryptic soy agar (TSA) or tryptic soy broth (TSB) supplemented with 10 g/ml erythromycin or 10 g/ml chloramphenicol when needed. When used, MnCl 2 was prepared as a 1 M stock in sterile water and an equal volume of water was added for all conditions. E. coli strains were grown on lysogeny broth (LB) or LB agar (LBA), supplemented with 50 g/ml carbenicillin when necessary. All growth in liquid medium occurred in an Innova 44 incubator shaking at 180 rpm, unless otherwise noted. Fifteen-milliliter round-bottom polypropylene tubes with aeration lids at a 45°angle were used for all standard cultures of 5 ml. Unless otherwise noted, all chemicals were from Sigma. All molecular biology reagents were from New England Biolabs and were used according to the manufac- turer's instructions. Phusion 2ϫ Hi-fidelity master mix was used for all PCRs. For construction of plasmids, all constructs were confirmed by Sanger sequencing (GeneWiz). All plasmids were transformed by electroporation from E. coli into the S. aureus cloning intermediate strain RN4220. Plasmid extractions and purifications were performed using a GeneJet plasmid miniprep kit (Thermo Fisher). Following isolation from RN4220, plasmids were transduced into final S. aureus strains with bacteriophage 85. Briefly, recipient strains were grown for 15 h in TSB at 37°C with 180 rpm of shaking. One hundred microliters of bacteriophage 85 isolated from RN4220 donor strains was combined with 1 ml of a 15-h recipient strain culture and 1 ml of phage buffer containing 1 mM MgSO 4 , 4 mM CaCl 2 , 50 mM Tris-HCl 2 (pH 7.8), 0.59% NaCl, and 0.1% gelatin. Bacteriophage reaction mixtures were incubated for 20 min in a 37°C water bath, followed by the addition of 2 ml of 1% sodium citrate. Bacterial cells were pelleted, resuspended in 2 ml TSB containing 0.5% sodium citrate, and incubated in a 37°C water bath for 1 h. Bacterial cells were centrifuged, and the cell pellets were resuspended in TSB supplemented with 0.5% sodium citrate and then plated onto TSA plates containing 0.5% sodium citrate and 10 g/ml chloramphenicol to select for transduction of the plasmid. Transduction of transposon library alleles. Transduction of the mntE::Tn(NE418) transposon allele into Newman WT and ΔmntR strains was performed using bacteriophage 85 as described previously (59,60). Alleles were confirmed by colony PCR with primers CMG-176 and CMG-177, as well as by Sanger sequencing.
Conservation of genomic context of mntE. The genomic context of mntE among staphylococci was assessed using the SEED Viewer database (pubseed.theseed.org [accessed January 2018]) (61). A 16-kb region was compared across genomes within the SEED database to identify genes whose sequence and relative position are conserved. Nineteen species from the Staphylococcus genus were identified, and based on conserved patterns of gene ontology, the following 8 species with the most similarity were chosen to be represented graphically:  (65,66). Putative Mn binding amino acid residues in S. aureus MntE were identified based on sequence similarity to Mn binding sites previously identified in S. pneumoniae (23).
RNA-seq analysis. S. aureus WT strains were streaked onto TSA and incubated at 37°C for 24 h. Single colonies were selected in biological triplicate and used to inoculate 5 ml of TSB. Cultures were grown for 15 h at 37°C. Cultures were back-diluted 1:100 in 5 ml TSB and grown at 37°C until mid-log phase. Cultures were centrifuged, the supernatant was removed, and the cell pellets were resuspended in TSB. Equal volumes of the resuspended cultures were added to TSB or TSB supplemented with MnCl 2 (final concentration of 1 mM MnCl 2 ). Cells were incubated for 30 min at 37°C and centrifuged, and the remaining cell pellet was subjected to RNA extraction. RNA was isolated using Tri reagent and chloroform and precipitated with isopropanol. RNA was isolated using the Qiagen RNeasy minikit. Precipitated RNAs were treated with DNase I (Thermo Fisher) following the manufacturer's instructions. RNA sequencing was performed by the Vanderbilt Technologies for Advanced Genomics (VANTAGE) core using the Illumina HiSeq 3000 platform (Illumina). The integrity and concentration of total RNA were determined using an Agilent 2100 Bioanalyzer system in combination with an RNA 6000 Nano kit (Agilent). rRNA was depleted using the Ribo-Zero rRNA removal kit (for bacteria) (Epicentre), and paired-end cDNA libraries were prepared with a TruSeq RNA library prep kit v2 (Illumina). Data analysis for sequencing experiments was performed on the basis of the CLC Genomics workbench (version 11.0.1; Qiagen) using the S. aureus Newman genome annotations (NCBI GenBank accession no. NC_009641.1). Prior to analysis, rRNA reads were removed in order to account for variations in rRNA depletion procedure among samples. Standard settings were used for adapter and quality trimming, as well as transcriptome sequencing (RNA-seq) analysis. Expression values were calculated as RPKM (reads per kilobase per million mapped reads) (67), and a lower cutoff of 5 RPKM was introduced for subsequent analysis.
qRT-PCR. Strains were streaked onto TSA and grown for 15 h at 37°C. Single colonies were selected in biological triplicate and inoculated into 5 ml of TSB and grown at 37°C for 15 h. Overnight cultures were subcultured 1:100 into TSB supplemented with 1 mM MnCl 2 or an equal volume of sterile water and grown until mid-exponential phase. An equal volume of ice-cold acetone-ethanol was added. RNA was isolated using Tri reagent and chloroform and precipitated with isopropanol. Precipitated RNAs were treated with DNase I (Thermo Fisher) following the manufacturer's instructions. RNA was isolated using the Qiagen RNeasy minikit. cDNA was synthesized from 2 g RNA via incubation with Moloney murine leukemia virus (MMLV) reverse transcriptase (Thermo), using a random hexamer primer mixture. Quan-titative PCR (qPCR) was performed using Bio-Rad iQ SYBR green supermix (Thermo Fisher) following the manufacturer's instructions, using primers CMG151/152 for gyrA, LEJ277/278 for mntE, and LEJ271/2712 for mntA. Transcript abundance was quantified using the threshold cycle (ΔΔC T ) method after normalization to gyrA abundance.
Bioluminescent reporter assay. The bioluminescent reporter plasmid was constructed by amplifying 350 bp of the intergenic region upstream of the predicted mntE translational start site using primer set CMG 221/CMG 222 and cloning the fragment into the pXen-1.luxABCDE plasmid (p.luxABCDE) upstream of the lux operon. S. aureus WT and ΔmntR strains with pXen-1.luxABCDE or pXen-1.P mntE luxABCDE (p.P mntE luxABCDE) were streaked onto TSA-chloramphenicol and grown for 24 h at 37°C. Single colonies in biological triplicate were used to start 5-ml cultures of TSB-chloramphenicol and grown for 15 h at 37°C. One microliter of each culture was inoculated into 99 l of TSB-chloramphenicol supplemented with 0, 125, 250, or 500 M MnCl 2 in 96-well clear-bottom black-walled plates and incubated at 37°C with linear shaking for 3 h. Relative bioluminescence was measured in a BioTek Cytation5 spectrophotometer with BioTek Gen5 software.
Mn toxicity growth assays. S. aureus strains were streaked onto TSA and grown for 24 h at 37°C. Single colonies in biological triplicate were used to start 5-ml cultures of TSB and grown at 37°C for 16 h. One microliter of each culture was added to 99 l of TSB supplemented with 0, 0.5, or 1 mM MnCl 2 in a 96-well flat-bottom plate. The plate was incubated at 37°C with linear shaking in a Biotek Epoch II plate reader. The optical density at 600 nm (OD 600 ) of each sample was measured every hour and analyzed using BioTek Gen5 software. Using a sterile wooden applicator, liquid cultures were also streaked onto TSA or TSA supplemented with 0.5 or 1 mM MnCl 2 . Plates were incubated for 24 h at 37°C. ICP-MS. S. aureus strains were streaked onto TSA and grown for 24 h at 37°C. Single colonies in biological triplicate were used to inoculate 5-ml cultures of TSB or TSB plus 100 M EDTA. Cultures were incubated for 15 h at 37°C. Liquid cultures were back-diluted 1:100 in 5 ml TSB and incubated at 37°C for 3 h. Cells were centrifuged, the supernatant decanted, and then the cells were washed in sterile Milli-Q water (Millipore Sigma) twice. The cell pellet was resuspended in 0.5 ml Milli-Q water, and 0.5 ml of the suspension was transferred to 15-ml metal-free conical tubes. Cell suspensions were digested by the addition of 0.5 ml optimum-grade metal-free nitric acid (Thermo Fisher Scientific). Elemental quantification of acid-digested liquid samples was performed using an Agilent 7700 inductively coupled plasma mass spectrometer (Agilent, Santa Clara, CA). The following settings were fixed for the analysis: cell entrance ϭ Ϫ40 V, cell exit ϭ Ϫ60 V, plate bias ϭ Ϫ60 V, OctP bias ϭ Ϫ18 V, and collision cell helium flow ϭ 4.5 ml/min. Optimal voltages for Extract 2, Omega Bias, Omega Lens, OctP RF, and Deflect were determined empirically before each sample set was analyzed. Element calibration curves were generated using the Aristar ICP standard mix (VWR, Radnor, PA). Samples were introduced by peristaltic pump with 0.5-mm internal diameter tubing through a MicroMist borosilicate glass nebulizer (Agilent). Samples were initially up taken at 0.5 rps for 30 s followed by 30 s at 0.1 rps to stabilize the signal. Samples were analyzed in Spectrum mode at 0.1 rps collecting three points across each peak and performing three replicates of 100 sweeps for each element analyzed. Data were acquired and analyzed using the Agilent Mass Hunter Workstation software version A.01.02.
HOCl killing of S. aureus. S. aureus stationary-phase overnight cultures were diluted 10,000-fold in biological triplicate into PBS containing 25, 12.5, 6.25, 3.125, or 0 M NaOCl (Clorox). Samples were mixed well and incubated statically at 25°C for 30 min. Each sample was diluted 10-fold into TSB and incubated with shaking for 24 h at 37°C on a BioTek Epoch II plate reader. The optical density at 600 nm of each sample was measured every 30 min. Each sample was also serially diluted 10 0 through 10 Ϫ5 in 10-fold increments in PBS, and 10 l of each dilution was spot plated on TSA. Plates were incubated for 18 h at 37°C, and CFU/ml were enumerated.
Murine models of infections. All animal experiments were reviewed and approved by the Vanderbilt University Institutional Animal Care and Use Committee. Procedures were performed according to the institutional policies, Animal Welfare Act, NIH guidelines, and American Veterinary Medical Association guidelines on euthanasia. S. aureus strains were streaked onto TSA and grown for 24 h at 37°C. For all experiments, 6-week-old female BALB/cJ mice were retro-orbitally infected with 1 ϫ 10 7 or 3 ϫ 10 7 S. aureus CFU in 100 l of phosphate-buffered saline as previously described (68). Following the inoculation, the infection was allowed to proceed for up to 10 days, and then the mice were humanely euthanized. For mice infected with 1 ϫ 10 7 CFU, hearts, kidneys, and livers were removed 4 or 10 days postinfection to determine organ bacterial burdens. Organs were homogenized in sterile PBS using a Bullet Blender tissue homogenizer (Next Advance), and serial dilutions were plated onto TSA to enumerate CFU. To determine mortality curves, mice were infected with 3 ϫ 10 7 CFU, and the infection was allowed to proceed for 10 days. Natural mortalities and moribund animals that were euthanized were documented as mortality events.
Paraquat sensitivity assays. S. aureus strains were streaked onto TSA and incubated at 37°C for 24 h. Single colonies were selected in biological triplicate and used to inoculate 5-ml cultures of TSB. Liquid cultures were incubated for 15 h at 37°C. TSB with 0.3 and 0.6 mM paraquat was prepped using a 1 M paraquat dichloride hydrate stock prepared fresh daily with sterile Milli-Q water (Millipore Sigma). An equal volume of sterile water was added to TSB as a negative control. Two microliters of each culture was added to 198 l of each TSB-paraquat treatment in a 96-well flat-bottom plate and incubated at 37°C with linear shaking in a BioTek Epoch II plate reader. The optical density at 600 nm of each sample was measured every 1 h and analyzed using BioTek Gen5 software. Growth curves using murine hearts and kidneys were prepared in the same manner as described above, with the exception of 2 l of organ homogenate serving as the inoculum. Impact of Mn Detoxification on S. aureus Homeostasis ® Statistical analysis. All data analysis and statistical tests were performed in GraphPad Prism 6 software. Replicate numbers and statistical tests for each experiment are listed in the figure captions.
Data availability. The data received from RNA sequencing experiments have been deposited into the NCBI Gene Expression Omnibus (accession no. GSE124285) and were consolidated into Table S2 in the supplemental material.