Metal-independent variants of phosphoglycerate mutase promote resistance to nutritional immunity and retention of glycolysis during infection

The ability of Staphylococcus aureus and other pathogens to consume glucose is critical during infection. However, glucose consumption increases the cellular demand for manganese sensitizing S. aureus to host-imposed manganese starvation. The current investigations were undertaken to elucidate how S. aureus copes with the need to consume glucose when metal-limited by the host. A critical component of host defense is production of the manganese binding protein calprotectin. S. aureus has two variants of phosphoglycerate mutase, one of which is manganese-dependent, GpmI, and another that is manganese-independent, GpmA. Leveraging the ability to impose metal starvation in culture utilizing calprotectin revealed that the loss of GpmA, but not GpmI, sensitized S. aureus to manganese starvation. Metabolite feeding experiments revealed that the growth defect of GpmA when manganese-starved was due to a defect in glycolysis and not gluconeogenesis. Loss of GpmA reduces the ability of S. aureus to cause invasive disease in wild type mice. However, GpmA was dispensable in calprotectin-deficient mice, which have defects in manganese sequestration, indicating that this isozyme contributes to the ability of S. aureus to overcome manganese limitation during infection. Cumulatively, these observations suggest that expressing a metal-independent variant enables S. aureus to consume glucose while mitigating the negative impact that glycolysis has on the cellular demand for manganese. S. aureus is not the only bacterium that expresses manganese-dependent and -independent variants of phosphoglycerate mutase. Similar results were also observed in culture with Salmonella enterica serovar Typhimurium mutants lacking the metal-independent isozyme. These similar observations in both Gram-positive and Gram-negative pathogens suggest that expression of metal-independent glycolytic isozymes is a common strategy employed by bacteria to survive in metal-limited environments, such as the host.


Introduction
The preferred carbon source for many pathogens is glucose and disruption of glycolysis reduces the ability of many invaders to cause infection [1][2][3][4][5][6][7][8][9][10]. The primacy of glucose as an energy source is emphasized by catabolite repression, which prevents bacteria from utilizing other carbon sources when glucose is present [1][2][3][11][12][13]. The advantage that sugar consumption provides is highlighted by the increased sensitivity of individuals with diabetes, especially those with hyperglycemia, to infections by Staphylococcus aureus, Streptococcus pneumoniae, Mycobacterium tuberculosis, Escherichia coli, Klebsiella pneumoniae and Candida albicans [14][15][16][17]. At the same time, host defenses reduce the ability of pathogens to consume glycolytic substrates by limiting metal availability or via other mechanisms [18,19]. The spread of antibiotic-resistant isolates has led the Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) to call for the development of novel therapeutics to treat S. aureus and other pathogens [20,21]. Understanding how pathogens generate energy and preserve the activity of critical metabolic pathways despite the concerted efforts of the immune system has the potential to identify new opportunities for therapeutic intervention.
Transition metals such as iron (Fe), manganese (Mn) and zinc (Zn) are essential for life, as they play an important role in facilitating the structure and function of proteins [22]. The host takes advantage of this essentiality by restricting the availability of Fe, Mn and Zn during infection, a defense known as nutritional immunity [23][24][25][26][27][28]. The prototypic example of Mn and Zn restriction is the staphylococcal abscess, which is virtually devoid of these metals [25,29]. A key mediator of the host Mn-withholding response is the immune effector calprotectin (CP). A heterodimer of S100A8 and S100A9, CP possesses two transition metal binding sites that chelate Mn and Zn with nanomolar and femtomolar affinities, respectively [25,27,[30][31][32][33][34][35]. Although CP binds other metals, including Fe (II) and nickel (Ni) [36,37], the primary metals withheld from S. aureus are Mn and Zn. CP comprises~50% of the total protein in the neutrophil cytoplasm and CP concentrations can exceed 1 mg/ml at sites of infection [38,39]. Mice lacking CP have defects in Mn sequestration and are more sensitive to a range of bacterial and fungal pathogens, including S. aureus, Acinetobacter baumannii, K. pneumoniae and Aspergillus fumigatus [25-27, 40, 41].
During infection, nutritional immunity inactivates Mn-dependent bacterial processes such as the Mn-dependent superoxide dismutase possessed by S. aureus. This in turn renders S. aureus more sensitive to the oxidative burst of immune cells. While glycolysis is canonically believed to be a magnesium-dependent process [42], many bacteria possess Mn-dependent variants of glycolytic enzymes, including phosphoglycerate mutase, enolase and pyruvate kinase [43][44][45][46][47][48][49][50]. In fact, in a number of pathogens, including S. aureus and S. pneumoniae, the consumption of sugars is dependent on Mn availability [18,51]. At the same time, glycolysis is critically important for the ability of S. aureus to cause infection and mutations that reduce the activity of this pathway frequently result in virulence defects and sensitize the bacterium to other host defenses such as NO� produced by immune cells [4][5][6][7][8]. While the best characterized mechanism utilized by bacteria to resist nutritional immunity is the use of high-affinity metal transporters, S. aureus and other pathogens also possess transporter-independent adaptations that critically contribute to the ability of the bacteria to overcome nutritional immunity [18,29,[52][53][54][55]. However, mechanisms that would enable them to maximize the consumption of sugars via glycolysis when metal-starved have not been described.
Despite increasing the cellular demand for Mn and sensitizing S. aureus to Mn starvation, we recently observed that S. aureus prefers to consume glucose even when Mn-starved by CP [18,55]. This paradoxical use of glycolysis despite the associated increase in cellular demand for Mn likely occurs in part as the fermentation of sugars enables the bacterium to generate energy when exposed to NO� produced by activated immune cells [8]. At the same time, it highlights a challenge faced by S. aureus and other pathogens, where adapting to one host defense sensitizes the bacterium to another aspect of the immune response [18,55]. Many bacteria, including S. aureus, possess multiple copies of glycolytic enzymes. While the presence of multiple isozymes is frequently associated with glycolytic and gluconeogenic flux, we wondered if another reason for maintaining multiple isozymes might exist. Specifically, we hypothesized that multiple copies of glycolytic enzymes may help S. aureus and other pathogens mitigate the stress of consuming sugars when metal-starved. The current investigations revealed that expression of a second metal-independent variant of phosphoglycerate mutase enables S. aureus to maintain the ability to consume glucose when Mn-starved and critically contributes to resisting nutritional immunity during infection. Similar results were also observed with Salmonella enterica, suggesting that expression of metal-independent isozymes is a common strategy employed by bacteria to survive in metal-limited environments.

Expression of metal-independent glycolytic enzymes increases in response to host-imposed metal starvation
As an initial step to identify how S. aureus promotes retention of glycolysis during infection, the repertoire of glycolytic enzymes possessed by S. aureus was assessed. This analysis revealed that S. aureus possess two copies of glyceraldehyde-3-phosphate dehydrogenase, aldolase and phosphoglycerate mutase. Notably, one of the phosphoglycerate mutase isozymes, GpmI, is predicted to be Mn-dependent, while the other, GpmA, is Mn-independent, utilizing 2,3-bisphosphoglycerate as a catalytic cofactor [56,57]. GpmI is encoded by the glycolytic operon that contains gapR, gapA, pgk, tpiA and eno, whereas gpmA is not part of an operon and is expressed independently of other glycolytic enzymes (Fig 1A). The fact that gpmI is in a locus with many other glycolytic enzymes, while gpmA is in a separate location, suggests that GpmI is the primary phosphoglycerate mutase and that GpmA is the secondary enzyme. One possibility is that GpmI and GpmA have directional preferences, with GpmI being essential for glycolysis and GpmA being essential for gluconeogenesis, as has been suggested for the two isozymes of glyceraldehyde-3-phosphate dehydrogenase encoded by S. aureus [58]. However, it is also possible that the metal-independent variant promotes retention of glycolytic flux when metal-starved by the host. As an initial step in evaluating this latter idea, expression of gpmA and gpmI in wild type bacteria exposed to CP was assessed. While expression of gpmI did not change, gpmA levels increased~40-fold in response to CP (Fig 1B). Expression of gpmA was also induced in a S. aureus mutant lacking the Mn transporters MntABC and MntH (ΔmntCΔmntH) to a level comparable to that caused by CP (Fig 1C). In contrast, expression of gpmI did not change in the ΔmntCΔmntH mutant ( Fig 1D). Cumulatively, these observations suggest that GpmI is the primary phosphoglycerate mutase used by S. aureus and that GpmA may be important when the bacteria experience Mn limitation.

Expression of a metal-independent phosphoglycerate mutase promotes staphylococcal resistance to manganese starvation
To elucidate the respective contributions of the two staphylococcal phosphoglycerate mutases to resisting metal starvation, wild type as well as ΔgpmA and ΔgpmI mutants were evaluated for their ability to grow in the presence of CP. Loss of the metal-dependent isozyme, GpmI, in S. aureus Newman did not alter the sensitivity of S. aureus to CP. Conversely, loss of the metal- were assessed by qRT-PCR. Expression was compared to wild type bacteria grown in the absence of CP. � = p�0.05 relative to media alone by one-way ANOVA on log-transformed values with Tukey's multiple comparisons test. Error bars indicate SEM. n�3. independent isozyme profoundly sensitized S. aureus to CP (Fig 2A, S1A Fig). Expression of GpmA from a plasmid reversed the increased sensitivity of ΔgpmA to CP (Fig 2B, S1B Fig). Increased sensitivity to CP was also observed upon loss of GpmA in the community-acquired MRSA strain USA300 JE2 (Fig 2C, S1C Fig), which was also reversed by expression of GpmA from a plasmid (Fig 2D, S1D Fig). Together, these results demonstrate that loss of GpmA makes S. aureus more sensitive to CP-imposed metal starvation.
To determine if Mn or Zn restriction was responsible for the increased sensitivity of ΔgpmA to CP, we leveraged CP binding site mutants with altered metal-binding properties [27,32]. Similar to WT CP, when ΔgpmA was grown in the presence of the ΔS2 mutant, which can bind either Mn or Zn, it was more sensitive to CP treatment than wild type bacteria or ΔgpmI ( Fig 2E). However, in the presence of the ΔS1 mutant, which cannot bind Mn, the increased sensitivity of ΔgpmA was abrogated. These observations suggest that loss of GpmA impairs the ability of the bacteria to cope with CP-induced Mn limitation. To further test this idea, wild type S. aureus, ΔgpmA and ΔgpmI were grown in medium depleted of Mn and Zn (NRPMI). Similar to what was observed in the presence of CP, the ΔgpmA mutant had a severe growth defect compared to wild type bacteria and the ΔgpmI mutant ( Fig 2F, S1E Fig). The growth defect of the ΔgpmA mutant in this medium was reversed by the addition of Mn but not Zn. Collectively, these results demonstrate that GpmA is crucial for growth when S. aureus is Mn-starved.

Expression of a metal-independent isozyme facilitates retention of glycolysis when Mn-starved
The majority of enzymes in the glycolytic pathway, including phosphoglycerate mutase, can be used for flux in both directions [59,60]. To determine if the GpmA-dependent growth defect is associated with decreased glycolytic or gluconeogenic activity, wild type S. aureus, ΔgpmA and ΔgpmI were grown in the presence of CP in a defined medium supplemented with either glucose or Casamino acids as the sole energy source. In the presence of glucose, the ΔgpmA mutant was more sensitive to CP than wild type bacteria or ΔgpmI (Fig 3A, S2A Fig). In contrast, there was no difference between any of the strains when Casamino acids were provided ( Fig 3B, S2B Fig). Supplementation of the glucose-containing medium with sodium pyruvate, which bypasses the phosphoglycerate mutase step, also enabled the ΔgpmA mutant to grow as well as wild type S. aureus or the ΔgpmI mutant ( Fig 3C, S2C Fig). Together, these observations indicate that loss of GpmA reduces the ability of S. aureus to consume glucose when Mnstarved. To evaluate if this apparent defect in glycolysis extended to other substrates dependent on glycolysis for consumption, growth of the ΔgpmA and ΔgpmI mutants in defined medium supplemented with glycerol was assessed. In the presence of glycerol, which enters glycolysis upstream of phosphoglycerate mutase, as the sole carbon source the ΔgpmA mutant was more sensitive to CP than wild type bacteria or ΔgpmI (Fig 3D, S2D Fig). Combined, these results reveal that GpmA plays an important role in retaining the ability to consume glycolytic substrates when Mn-starved.

Loss of GpmA does not affect manganese transporter function
The expression of Mn transporters is critical for the ability of S. aureus to resist host-imposed Mn starvation [29]. To confirm that the enhanced sensitivity of the ΔgpmA mutant was not due to an unanticipated impact on Mn transporter activity, ΔmntCΔmntHΔgpmA and ΔmntCΔmntHΔgpmI strains were assessed for CP sensitivity. Loss of GpmA, but not GpmI, in the ΔmntCΔmntH background further increased the sensitivity of the transporter double mutant, suggesting that GpmA and the Mn transporters function independently to promote resistance to Mn starvation (Fig 4A and 4B, S4A-S4C Fig). Expression of GpmA from a plasmid reverted the CP sensitivity of ΔmntCΔmntHΔgpmA back to that of the ΔmntCΔmntH mutant ( Fig 4C). Cumulatively, these results suggest that loss of GpmA does not sensitize the bacteria to Mn starvation by reducing Mn transport but rather by a Mn transporter-independent mechanism.

GpmA is necessary for invasive S. aureus disease and resisting manganese starvation during infection
Culture-based experiments suggest that the Mn-independent activity of GpmA enhances the ability of S. aureus to maintain glycolytic flux when Mn-limited. To evaluate the contribution of the two phosphoglycerate mutases to S. aureus pathogenesis, wild type (C57BL/6) mice were retro-orbitally infected with wild type S. aureus, ΔgpmA, or ΔgpmI and the infection was allowed to proceed for 4 days. During the course of the infection mice infected with ΔgpmA lost significantly less weight than mice infected with wild type S. aureus or ΔgpmI ( Fig 5A). Interestingly, mice infected with ΔgpmI lost slightly, but significantly, less weight than mice infected with wild type bacteria. Consistent with the weight loss, the ΔgpmA mutant had significantly decreased bacterial burdens in the liver, heart, and kidneys when compared to wild type bacteria (Fig 5B and 5C) indicating that GpmA plays an important role in establishing systemic disease. While mice infected with ΔgpmI did not show a statistically significant decrease in bacterial burdens in any of the organs, bacterial burdens in the kidneys were slightly lower than in mice infected with wild type bacteria (Fig 5C). To evaluate if the importance of GpmA during infection is driven by host restriction of Mn availability, CP-deficient (C57BL/6 S100A9-/-) mice, which do not remove Mn from liver abscesses [25, 29], were infected with wild type bacteria, ΔgpmA or ΔgpmI. Relative to wild type mice, the CP-deficient mice infected with ΔgpmA had increased bacterial burdens, indicating that the importance of GpmA during infection is driven by host-imposed Mn limitation. Moreover, in CP-deficient mice there was no difference in bacterial burdens between wild type bacteria, ΔgpmA or ΔgpmI (Fig 5B), suggesting either phosphoglycerate mutase isozyme is sufficient when Mn is available. Cumulatively, these results indicate that GpmA contributes to staphylococcal infection by promoting retention of glycolytic activity when Mn-starved by the host.

A metal-independent phosphoglycerate mutase enables glycolysis when Salmonella is manganese-starved
The expression of metal-dependent and -independent variants of phosphoglycerate mutase is not unique to S. aureus, and many pathogenic bacteria including E. coli, Shigella flexneri, S. enterica, Pseudomonas aeruginosa, Listeria monocytogenes and others possess both forms [57]. However, the molecular rationale for retaining two phosphoglycerate mutase isozymes in these organisms remains unknown [57]. In light of our observation with S. aureus, we wondered if metal-independent variants of phosphoglycerate mutase broadly promote retention of glycolytic potential when experiencing Mn starvation. To test this idea, the ability of wild type S. enterica serovar Typhimurium and a ΔgpmA mutant to grow in rich medium in the presence of CP was assessed. Similar to S. aureus, the Salmonella ΔgpmA mutant was more sensitive to CP treatment than wild type bacteria (Fig 6A, S4A Fig). Use of the CP variants with altered metal-binding properties revealed that loss of GpmA impaired the ability of Salmonella to cope with CP-induced Mn limitation ( Fig 6B). Cumulatively, these observations establish that expression of a Mn-independent variant of phosphoglycerate mutase promotes the growth of Salmonella when metal-starved.
To test if the importance of GpmA to Salmonella growth when metal-starved is attributable to its ability to promote consumption of glycolytic substrates, bacteria were grown in the GpmA enhances resistance to CP-induced manganese starvation GpmA enhances resistance to CP-induced manganese starvation presence of CP and provided with either glucose or Casamino acids as a carbon source (Fig 6C  and 6D, S4B and S4C Fig). As in rich medium, loss of GpmA increased the sensitivity of Salmonella to CP when glucose was provided as the sole carbon source. However, the enhanced sensitivity of the Salmonella ΔgpmA mutant was ablated in the presence of Casamino acids, which do not require the glycolytic pathway for consumption. Additionally, providing sodium pyruvate as a carbon source reversed the increased sensitivity to CP, whereas glycerol did not (Fig 6E and 6F, S4D and S4E Fig). Combined, these results suggest that the expression of metal-independent versions of phosphoglycerate mutase are a common mechanism employed by pathogenic bacteria to resist Mn limitation.

Discussion
The consumption of sugars via glycolysis and metals are important for pathogens during infection [4, 5, 7-10, 24, 61-66]. Sugars are the preferred carbon source for S. aureus and other pathogens and increased availability of glucose renders individuals more sensitive to infection [14][15][16][17]. At the same time, the host restricts the availability of metals, inactivating Mn-dependent bacterial processes, including glycolysis [18,55]. While expression of high-affinity metal transporters enhances the ability of S. aureus and other pathogens to maintain glycolytic flux and are critical for infection, they are insufficient to ensure that the pathogen obtains sufficient Mn to activate Mn-dependent superoxide dismutases [27, 53]. The current work provides yet another example of a critical Mn-dependent enzyme, GpmI, which is inactivated by nutritional immunity. At the same time, host defenses force S. aureus to ferment sugars via the glycolytic pathway, increasing the cellular demand for this metal [5,7,8]. The current work identified a critical role for a priori redundant metal-independent variants of phosphoglycerate mutases in both S. aureus and Salmonella in enabling glucose consumption and resisting nutritional immunity. The Mn-independent variant of phosphoglycerate mutase in S. aureus, which is encoded outside of the canonical glycolytic operon, is the primary isozyme used by the bacteria when Mn-starved and is critical for infection due to the host metal-withholding response. The conserved function of the metal-independent variant of phosphoglycerate mutase in both S. aureus and Salmonella suggests that this approach is likely a common strategy for preserving the ability to consume glucose while minimizing the cellular demand for Mn.
S. aureus and Salmonella are not the only pathogens that experience metal limitation during infection nor are they the only pathogens to require glucose consumption for disease [1, 5-10, 28, 65, 66]. It therefore seems likely that other successful pathogens have adaptations that allow them to consume glucose when metal-limited. Notably, in addition to S. aureus and Salmonella, many other bacterial pathogens, including S. flexneri, P. aeruginosa, and L. monocytogenes possess metal-dependent and -independent variants of phosphoglycerate mutase [57]. As CP is one of the most abundant proteins at sites of infection [38, 39], it seems likely that all of these pathogens experience Mn limitation during infection. Combined with the current observations, this suggest that using metal-independent variants of phosphoglycerate mutase to preserve glycolytic function in the host when Mn-starved may be a conserved strategy.
Intriguingly, a number of bacterial pathogens, including S. pneumoniae, Enterococcus faecalis, Haemophilus influenzae, Neisseria meningiditis, and M. tuberculosis possess only a metalindependent version of phosphoglycerate mutase [56,57]. Similar to other pathogens, all of these organisms would be expected to encounter Mn-limited environments within the host. This raises the possibility that possessing only the metal-independent variant of phosphoglycerate mutase represents further adaptation to life in a Mn-poor environment. At the same time, the majority of bacteria, archaea, protozoa and fungi, including pathogens, possesses both metal-dependent and -independent versions of phosphoglycerate mutase. However, others, such as Helicobacter pylori, possesses only the metal-dependent variant but also encounter CP during infection [67]. The biochemical differences between the metal-dependent and independent-variants of phosphoglycerate mutase extend beyond their cofactor [47,56,[68][69][70]. Notably, in S. aureus the metal-dependent variant is in the primary glycolytic operon.
Together, these observations suggest that the metal-dependent variants of phosphoglycerate mutase provide some biological advantage.
Phosphoglycerate mutase is not the only enzyme in the glycolytic pathway that contains both metal-dependent and -independent versions. Both metal-dependent and -independent variants of fructose bisphosphate aldolase exist, and many pathogenic bacteria, including S. aureus, Salmonella, Borrelia burgdorferi, S. pneumoniae, and K. pneumoniae contain both enzymes. The metal-dependent variant is classically thought to utilize Zn as a cofactor [71][72][73]. Similar to GpmA, which is upregulated in response to CP treatment, the metal-independent staphylococcal aldolase is also upregulated in the presence of CP [74]. As aldolase is the only staphylococcal enzyme in glycolysis predicted to use Zn, this could explain why CP-imposed Mn limitation but not Zn limitation inactivates glycolysis in S. aureus [18,55]. More broadly, it suggests that the metal-independent fructose bisphosphate aldolase may contribute to retaining glycolytic function when metal-starved by the host. In both Salmonella and B. burgdorferi, the Zn-independent aldolase has been suggested to provide a mechanism for the bacteria to consume glucose when experiencing nitrosative stress [19]. In the case of B. burgdorferi, nitrosative stress was suggested to damage the Zn-dependent isozyme [19]. At the same time, in S. aureus nitrosative stress has been observed to lead to a transcriptional response similar to that seen when Zn-starved [74]. Regardless of the stress that the metal-independent aldolase responds to, these observations suggest that metal-independent isozymes are likely to be generally important to the ability of bacteria to maintain glycolytic flux during infection.
The presence of two isozymes with differing metal utilization is not limited to glycolysis or S. aureus and Salmonella. In Bacillus subtilis, Fe limitation leads to induction of the Fe-sparing response, which includes increased expression of flavodoxin that can replace the iron-containing electron transfer protein ferredoxin [75]. Similarly, E. coli contains two ribonucleotide reductases, a primary Fe-dependent enzyme, and a Mn-dependent enzyme that is crucial for survival when bacteria experience superoxide stress or Fe limitation [76,77]. Similarly, in response to Zn limitation, B. subtilis will induce expression of an alternative Zn-independent folate biosynthesis enzyme, GTP cyclohydrolase-IB (GCYH-IB). Induction of GCYH-IB prevents Zn starvation from inducing a folate auxotrophy [78]. B. subtilis also possesses L31 � and L33 � , Zn-independent ribosomal proteins, which promote survival when Zn-starved [79,80]. Combined with the current observations, this suggests that replacing metal-dependent enzymes with metal-independent variants is a common strategy for surviving Fe, Zn and Mn limitation. Notably, the switch between utilizing the metal-dependent and -independent isozyme is frequently driven by a metal-sensing regulator [75,76,[78][79][80]. In S. aureus, loss of the Fe-sensing regulator represses the expression of gpmA, but it is not predicted to be regulated by the Mn-sensing regulator MntR [81]. While the mechanisms that control expression of GpmA are unknown, this suggests that metal-independent regulatory circuits also play an important role in coordinating a response to metal starvation. This idea is supported by the observation that ArlRS, which contributes to staphylococcal growth in Mn-poor environments, appears not to sense Mn directly but rather the impact that Mn starvation has on glycolytic flux (Parraga et. al. In Press).
The current investigations add metal-dependent and -independent phosphoglycerate mutases to the growing list of enzymes that carry out apparently redundant biochemical reactions, but whose distinct reaction mechanisms enable microbes to cause infection or survive in other stressful environments [5,6,19,75,76,[78][79][80]. Many of these false redundancies have been identified by studying how microbes respond to metal limitation [75][76][77][78][79][80]. However, the importance of other pseudo-redundant enzymes has been revealed by investigating how pathogens cope with other stresses experienced during infection, such as the contribution of a second staphylococcal lactic acid dehydrogenase to growth in the presence of nitric oxide and infection [5,6,19]. While diverse stresses have been examined, a common denominator in all of these studies is that they have pushed the microbes outside of their ideal environments. As the importance of physiology to pathogenesis continues to be revealed, the current and prior studies reveal the importance of considering the environment encountered within the host when evaluating the contribution of enzymes to metabolism and infection.

Ethics statement
All experiments involving animals were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Illinois Urbana-Champaign (IACUC license number 15059) and performed according to NIH guidelines, the Animal Welfare Act, and US Federal law.

Bacterial strains
Bacteria were routinely grown on tryptic soy agar (TSA) plates. For routine overnight cultures, bacteria were grown in 5 ml of tryptic soy broth (TSB) or in Chelex-treated RPMI plus 1% Casamino acids (NRPMI) supplemented with 1 mM MgCl 2 , 100 μM CaCl 2 and 1 μM FeCl 2 [29] in 15 ml conical tubes at 37˚C on a roller drum. As needed, 10 μg/ml of chloramaphenicol was added for plasmid maintenance. S. aureus strain Newman and its derivatives were used for all of the experiments, unless otherwise indicated. For experiments using USA300 JE2 and derivatives (USA300 JE2 gpmA:erm and USA300 JE2 gpmI:erm), strains were obtained from the Nebraska library [82]. gpmA:erm, gpmI:erm, ΔmntCΔmntH gpmA:erm and ΔmntCΔmntH gpmI:erm were generated by transducing the gpmA:erm and gpmI:erm alleles via F85 phage from USA300 JE2 gpmA:erm and USA300 JE2 gpmI:erm. As needed 500 μM MnCl 2 was added to agar plates to facilitate recovery of mutants. All constructs were confirmed to be hemolytic by growth on TSA blood agar plates. To generate constructs for complementation studies, the gpmA coding sequence was amplified with the indicated primers (Table 1) and cloned into the pOS1 vector under the control of the P lgt promoter [83].
S. enterica serovar Typhimurium strain 14028 was used for all Salmonella experiments. The deletion of gpmA by inserting a chloramphenicol cassette was carried out using lambda redmediated recombination as described previously using the indicated primers (Table 1) [84,85]. The insertion of the cassette was checked by PCR analysis and the construct was moved into a clean background by P22 transduction.

CP growth assays
CP growth assays were performed as described previously [27,32]. Briefly, overnight cultures (grown 16-18 h at 37˚C on a roller drum) were used directly and diluted 1:100 into 96-well round-bottom plates containing 100 μl of growth medium (38% TSB and 62% calprotectin buffer (20 mM Tris pH 7.5, 100 mM NaCl, 3 mM CaCl 2 , 10 mM β-mercaptoethanol)) in presence of varying concentrations of CP. Unless otherwise indicated, the bacteria were grown overnight in NRPMI supplemented with 1 mM MgCl 2 , 100 μM CaCl 2 and 1 μM FeCl 2 and directly inoculated 1:100 into the assay medium. For all assays, the bacteria were incubated with orbital shaking (180 RPM) at 37˚C and growth was measured by assessing optical density (OD 600 ) every 2 hours. Prior to measuring optical density, the 96-well plates were vortexed. For experiments utilizing defined medium, the bacteria were precultured overnight in TSB. The defined medium consisted of 1.3 g/L NaCl, 2.6 g/L NH 4 Cl, 5.

Expression analysis
To assess the expression of gpmA and gpmI, S. aureus was grown as for CP growth assays in complex medium in the presence and absence of 240 μg/ml of CP, with the exception that they were precultured overnight in TSB. Bacteria were harvested during log phase growth (OD 600 = 0.1), when the samples were collected, an equal volume of ice-cold 1:1 acetone-ethanol was then added to the cultures, and they were frozen at -80˚C until RNA extraction. RNA was extracted and cDNA was generated, as previously described [86][87][88]. Gene expression was assessed by quantitative reverse transcription-PCR (qRT-PCR) using the indicated primers (Table 1, [29]) and 16S was used as a normalizing control.

Animal infections
Mouse infections were performed as described previously [25,27]. Briefly, 9-week old wild type or CP-deficient (S100A9-/-) mice were infected retro-orbitally with approximately 5 x 10 6 CFU in 100 μl of sterile phosphate-buffered saline. Following injection, the infection was allowed to proceed for 96 h before the mice were sacrificed. Livers, hearts and kidneys were removed, the organs were homogenized, and bacterial burden was determined by plating serial dilutions.