Cellular Mn/Zn Ratio Influences Phosphoglucomutase Activity and Capsule Production in Streptococcus pneumoniae D39

ABSTRACT

Capsular polysaccharide (CPS) is a major virulence determinant for many human-pathogenic bacteria. Although the essential functional roles for CPS in bacterial virulence have been established, knowledge of how CPS production is regulated remains limited. Streptococcus pneumoniae (pneumococcus) CPS expression levels and overall thickness change in response to available oxygen and carbohydrate. These nutrients in addition to transition metal ions can vary significantly between host environmental niches and infection stage. Since the pneumococcus must modulate CPS expression among various host niches during disease progression, we examined the impact of the nutritional transition metal availability of manganese (Mn) and zinc (Zn) on CPS production. We demonstrate that increased Mn/Zn ratios increase CPS production via Mn-dependent activation of the phosphoglucomutase Pgm, an enzyme that functions at the branch point between glycolysis and the CPS biosynthetic pathway in a transcription-independent manner. Furthermore, we find that the downstream CPS protein CpsB, an Mn-dependent phosphatase, does not promote aberrant dephosphorylation of its target capsule-tyrosine kinase CpsD during Mn stress. Together, these data reveal a direct role for cellular Mn/Zn ratios in the regulation of CPS biosynthesis via the direct activation of Pgm. We propose a multilayer mechanism used by the pneumococcus in regulating CPS levels across various host niches.

IMPORTANCE Evolving evidence strongly indicates that maintenance of metal homeostasis is essential for establishing colonization and continued growth of bacterial pathogens in the vertebrate host. In this study, we demonstrate the impact of cellular manganese/zinc (Mn/Zn) ratios on bacterial capsular polysaccharide (CPS) production, an important virulence determinant of many human-pathogenic bacteria, including Streptococcus pneumoniae. We show that higher Mn/Zn ratios increase CPS production via the Mn-dependent activation of the phosphoglucomutase Pgm, an enzyme that functions at the branch point between glycolysis and the CPS biosynthetic pathway. The findings provide a direct role for Mn/Zn homeostasis in the regulation of CPS expression levels and further support the ability of metal cations to act as important cellular signaling mediators in bacteria.

INTRODUCTION

Streptococcus pneumoniae is an opportunistic bacterial pathogen that asymptomatically colonizes the nose and throat of healthy humans. To establish asymptomatic colonization or carriage, the pneumococcus must adopt a state of homeostasis with the host’s defense systems and the constant barrage of disruptive events caused by inhalation and other microbial competitors. Dysbiosis (microbial imbalance) or a change in host factors can lead to mucosal dysfunction, permitting the pneumococcus access to other parts of the respiratory tract (1–3). A key virulence determinant for the survival and pathogenesis of the pneumococcus in the host is the production of capsular polysaccharide (CPS) (4–8). Although studies have clearly established a functional role for CPS in bacterial virulence, knowledge of how CPS production is regulated remains limited.

CPS expression levels can vary in response to nutrient oxygen and carbohydrate availability (9–14). The overall concentrations of these nutrients, in addition to transition metal ions, vary significantly between host environmental niches and infection stages (15, 16). Transition metals such as zinc (Zn) and manganese (Mn) are among the required micronutrients that serve as cofactors in many bacterial proteins involved in critical cellular processes and defenses against the host. Extensive research has been aimed at identifying how changes in single metal ions affect bacterial fitness and pathogenesis, specifically regarding mismetallation (17–20). For example, Zn toxicity competitively inhibits Mn binding to the pneumococcal Mn importer PsaBCA and the Mn-sensing regulator PsaR (21–25) as well as the glycolytic enzymes phosphofructokinase and glyceraldehyde-3-phosphate dehydrogenase in Streptococcus pyogenes (16, 26), thereby decreasing bacterial fitness. Disruption of Mn/Zn equilibrium ratios can alter pneumococcal cell division by modulating the activity of the Mn-dependent protein phosphatase PhpP (27). Excess Mn can induce czcD expression, possibly by competitively inhibiting Zn binding to the pneumococcal Zn-sensing regulator SzcA, resulting in decreased bacterial virulence (28). As such, the concentration of one metal relative to another is pivotal to bacterial colonization and the bacterium’s ability to cause disease.

In S. pneumoniae, glycolysis is connected to the CPS biosynthetic pathway via the phosphoglucomutase Pgm. Pgm is a metal-dependent phosphotransferase that catalyzes the reversible isomerization of glucose 6-phosphate (G6P) to glucose 1-phosphate (G1P) (Fig. 1) (29). G1P is required to generate UDP-glucose, a sugar nucleotide precursor for CPS biosynthesis in most S. pneumoniae strains (30). In addition to CPS biosynthesis, Pgm function is associated with the production of other cellular exoproducts, such as lipopolysaccharide and teichoic acid (31–34). Bacterial fitness studies using Tn-seq and CRISPRi-based knockdown strategies suggest that pgm is likely essential for growth and pathogenesis in several pneumococcal strains examined (100, 101). As such, Pgm is an important virulence factor for many pathogenic bacteria, including S. pneumoniae (35).

FIG 1.

Schematic overview of capsular polysaccharide biosynthesis in S. pneumoniae. Intracellular Mn homeostasis is maintained by the combined activities of the Mn-specific importer PsaBCA and exporter MntE (27, 98, 99). Intracellular Mn can serve as a metal cofactor in a variety of metalloenzymes (in green), including the phosphotyrosine-protein phosphatase CpsB and the putative metal-dependent phosphoglucomutase Pgm. These enzymes, among many others, play a role in CPS biosynthesis.

The pneumococcal Pgm protein belongs to the ubiquitously dispersed superfamily of phosphoglucomutase/phosphomannomutase (PGM/PMM) enzymes. Crystallographic structures of several PGM/PMM family enzymes have been determined. The majority of these proteins are found coordinating a single Mg(II) ion in an octahedral geometry using the conserved metal-binding motif DXDXDR (36–39). Detailed biochemical and kinetic studies of various PGM/PMM enzymes demonstrate that metal binding is promiscuous; all characterized PMM/PGM enzymes have the capacity to bind a variety of divalent metal ions, including Mg(II), Mn(II), and Zn(II) (37, 38, 40–42). Although enzyme activity is most robust with Mg(II), Mn(II) and other divalent metal ions can also activate PGM/PMM enzymes in vitro.

Metal ion ratios are also likely to regulate the activity of the Mn-dependent phosphotyrosine-protein phosphatase CpsB, which is known to modulate capsule production by dephosphorylating autophosphorylated CpsD, a cytoplasmic phosphotyrosine-protein kinase (Fig. 1) (43–47). Cycling of the phosphorylation state of CpsD or its protein homologues promotes the proper synthesis of CPS (43, 48–51). Deviations from this cycle or the loss of cpsB or cpsD significantly reduces CPS levels and decreases bacterial virulence and colonization of the host (8, 52–55).

In this study, we examine the impact of bioavailable Mn/Zn ratios on CPS production and focus on two metal-utilizing enzymes (Pgm and CpsB) known to participate in modulating CPS levels. We ultimately reveal a direct role for Mn/Zn homeostasis in the regulation of CPS biosynthesis via the Mn-dependent activation of Pgm and propose a multilayer mechanism used by S. pneumoniae in regulating CPS levels across various host niches.

RESULTS

Manganese bioavailability correlates positively with capsular polysaccharide production.

We previously showed that perturbations in Mn/Zn homeostasis can disrupt pneumococcal cell division (27). We discovered by microscopy and confirmed here that Mn-mediated toxicity seemingly increases capsule thickness compared to nontreated wild-type (WT) strain D39 (see Fig. S1 in the supplemental material). The mntE-null mutant (ΔmntE) strain deficient in exporting Mn also exhibits increased cell chaining, indicative of increased capsule production (56). We note that WT and ΔmntE cells maintain a 1:1 Mn/Zn ratio (∼100 ng/mg protein for each metal) during routine microaerophilic growth in brain heart infusion (BHI) rich medium. When grown with exogenous Mn, the Mn/Zn equilibrium ratio shifts toward Mn (2:1) for both WT and ΔmntE cells; WT cells increase Mn concentrations 2-fold to ∼200 ng/mg protein but maintain Zn at ∼100 ng/mg protein, while ΔmntE cells increase Mn concentrations at least 4-fold to ≥400 ng/mg protein and Zn concentrations 2-fold to ∼200 ng/mg protein (27). Comparably, a czcD-null mutant (ΔczcD) deficient in exporting Zn is found to shift the Mn/Zn equilibrium ratio 8-fold toward Zn (≥800 ng/mg protein), while Mn levels remain relatively unchanged when grown with exogenous Zn (57, 58). Using such established pneumococcal strains and growth conditions, we further examined the impact of the bioavailable Mn/Zn ratio on CPS expression levels.

A modified mucoviscosity assay to determine the percentage of remaining cells after low centrifugation force was first used to qualitatively assess CPS levels (59, 60) of the WT, ΔmntE, and ΔczcD strains in BHI medium supplemented with or without Mn and Zn. The fundamentals of this assay are based on the CPS adherence capacity, or lack thereof, to surfaces. A 2-fold increased adherence after low-speed centrifugation was observed for unencapsulated pneumococcal cells lacking cps genes cps2A′ through cps2H′ compared to WT cells (Fig. S2A). Adherence of the Δcps strain was also independent of the presence of Mn and Zn. We further found that nontreated WT and ΔmntE cells share similar mucoviscosity levels (Fig. 2A, black bars). In contrast, when grown with increasing Mn concentrations, both strains were significantly (up to 2-fold) more likely to remain in solution after low-speed centrifugation (Fig. 2A). The decreased ability of cells grown with exogenous Mn to adhere to the surface suggests that a high Mn/low Zn ratio leads cells to express more CPS than in nontreated cells and correlates with our initial microscopy observations. It also indirectly supports the idea that pneumococcal cells expressing thicker capsules have a diminished ability to adhere to and colonize host epithelial tissues.

FIG 2.

Excess cellular Mn increases CPS expression levels, while excess Zn reduces CPS. Exponentially growing S. pneumoniae cells were diluted into prewarmed BHI broth supplemented with metals as indicated and harvested at an OD₆₂₀ of ∼0.2. (A and B) Mucoviscosity of cells grown with increasing Mn (A) or Zn (B) concentrations. (C and D) Quantification of total CPS by assays of uronic acids from cells grown with increasing Mn (C) or Zn (D) concentrations. Data shown represent the means from at least three independent cultures ± standard errors of the means (SEM). ns, not significant; *, P ≤ 0.1; **, P ≤ 0.05; ***, P ≤ 0.01 (determined using Student’s unpaired t test).

Conversely, ΔczcD cells exhibiting a low Mn/high Zn ratio were found to consistently pellet well, resulting in significantly fewer cells in solution after low-speed centrifugation than for nontreated WT cells (Fig. 2B). The WT mucoviscosity level was not significantly altered by Zn treatment. These data suggest that low-Mn/high-Zn growth leads to reduced CPS levels produced by the pneumococcus, which likely exposes extracellular surface antigens, thereby increasing adherence to surfaces.

We next quantified the total amount of CPS produced by these pneumococcal cells by assaying uronic acid (a CPS component) (61). Deletion of cps genes abolishes CPS production in pneumococcal cells (Fig. S2B). Similar levels of uronic acid were produced by both nontreated WT and ΔmntE cells (Fig. 2C, black bars). The addition of 100 μM Mn to BHI medium significantly increased uronic acids by ∼30% for both the WT and ΔmntE strains (Fig. 2C). In contrast, increasing Zn supplementation resulted in significantly less uronic acid produced by ΔczcD cells only; Zn supplementation had no discernible effect on uronic acid levels produced by WT cells (Fig. 2D). Although elevated Zn concentrations in BHI medium have also been shown to decrease hyaluronic acid capsule in S. pyogenes (26), we cannot rule out that the slower growth exhibited by ΔczcD cells in the presence of high Zn concentrations (≥40 μM) may also indirectly contribute to reduced uronic acid production. These data are consistent with those observed microscopically and qualitatively in the mucoviscosity assays and thereby support a strong positive correlation between the cellular Mn/Zn bioavailability ratio and S. pneumoniae capsule production.

Manganese-dependent activation of CpsB does not promote aberrant tyrosine dephosphorylation of CpsD.

To investigate the mechanism by which bioavailable Mn/Zn ratios affect CPS production, we focused first on the Mn-dependent phosphotyrosine-protein phosphatase CpsB, which is known to modulate capsule production by dephosphorylating autophosphorylated CpsD kinase (44, 47) (Fig. 1). In culture, oxygen availability can stimulate CpsB phosphatase activity (13). The presence of oxygen also leads to increased total associated Mn levels in S. pneumoniae, changing from ∼10 ng Mn/mg protein during anaerobiosis to 100 ng Mn/mg protein during microaerophilic growth (27). Together, these data suggested to us that Mn homeostasis is coordinated with oxygen levels and that increased CpsB activity likely results from the influx of Mn during microaerobic growth of the pneumococcus. Since fluctuations in Mn/Zn homeostasis can affect gene expression as well as the activation of Mn-utilizing enzymes like CpsB, we decided to measure the effect of Mn/Zn ratios on cpsB expression and CpsB enzyme activity. No significant change was observed in cpsB transcript levels with high-Mn or -Zn treatments for all strains (Fig. 3A and B), thereby eliminating an alteration of cpsB gene expression as a source contributing to influencing CPS levels.

FIG 3.

The Mn/Zn ratio does not induce cpsB or pgm transcription. RNA was extracted and quantified from exponentially growing S. pneumoniae cells grown in BHI broth supplemented with metals as indicated. (A and B) cpsB transcript levels relative to those of WT untreated cells during Mn-replete growth (A) or Zn stress (B). (C and D) pgm transcript levels relative to those of WT untreated cells during Mn-replete growth (C) or Zn stress (D). Data shown represent the means from at least three cultures ± SEM. ns, not significant; **, P ≤ 0.05; ***, P ≤ 0.01 (determined using one-way ANOVA [unpaired]).

Based on previous biochemical and structural studies of CpsB from S. pneumoniae and other bacterial homologues, CpsB can bind up to three metal ions coordinated by conserved aspartate residues, with two metal ions (M1 and M2) linked by a deprotonated water molecule (44–46). The binding of the third metal ion (M3) remains controversial since it binds weakly (45, 46, 62). The predicted catalytic mechanisms of CpsB are also similar to those of other binuclear metal-dependent hydrolases (63–65). We confirmed in vitro that purified apo-CpsB binds two metal molar equivalents of metal in competition with a modest-affinity competitor chelator, mag-fura-2 (mf2) (Fig. S3A and B). The association equilibrium constant (K_Me) for metal-binding site 1 differs by 1 order of magnitude for Mn (log K_a1, 5.53 ± 0.32 M⁻¹) versus Zn (log K_a1, 6.53 ± 0.19 M⁻¹), while the site 2 affinity is several orders of magnitude higher for Zn (log K_a2, 5.82 ± 0.92 M⁻¹) than for Mn (log K_a2, ≤4.0 M⁻¹, an upper limit from the mf2 titrations) (see Table S1 in the supplemental material for the respective dissociation equilibrium constant [K_d] values in molar concentrations). These findings are consistent with predictions from the Irving-Williams series of metal complex stabilities (66) and with our previously proposed metal homeostasis model in which the variation in magnitude between the metal-binding sites alludes to site 2 serving as a regulatory site that dictates enzyme activity through proper assembly and geometry of the binuclear metal cluster (27). Like other Mn-dependent enzymes, bioavailable Mn/Zn ratios may likely modulate CpsB phosphatase activity, ultimately affecting CPS production.

To evaluate the direct effect of Mn/Zn ratios on CpsB phosphatase activity, we used a tyrosine phosphatase assay with the phosphopeptide DADE(pY)LIPQQG as the substrate for CpsB. Mn(II) titration into purified apo-CpsB protein led to an Mn-dependent increase in CpsB phosphatase activity, with 100 μM Mn being optimum under our conditions (Fig. 4A). Zn(II) was significantly less effective than Mn in activating CpsB, and no activity was observed when Mg(II) was provided (Fig. 4B). Titration of Zn(II) into Mn-bound CpsB results in a significant decrease in CpsB phosphatase activity (Fig. 4C). These data suggest that CpsB phosphatase activity is likely regulated by intracellular Mn/Zn ratios. We were unable to support these findings in vivo due to minimal enzyme activity following the removal of free phosphates in pneumococcal cell lysates. We postulate that the resin used to remove free phosphates may have caused metal exchange in CpsB despite the use of metal-free buffer components since the purified Mn-bound CpsB protein is also inactivated using this process.

FIG 4.

Mn-dependent activation of the CpsB phosphatase does not promote aberrant dephosphorylation of CpsD kinase. (A) Mn-titrated activation of purified apo-CpsB protein. (B) Comparison of the activation abilities of apo-CpsB by Mn(II), Mg(II), and Zn(II) cations. (C) Impact on Mn-bound CpsB activity by Zn(II) titration. (D) Percentage of phosphorylated CpsD protein present in cell lysates quantified by a mobility shift assay. The means from at least three independent cultures or measurements for purified protein ± SEM are shown. ns, not significant; *, P ≤ 0.1; **, P ≤ 0.05; ***, P ≤ 0.01 (determined using one-way ANOVA [unpaired]).

Since CpsB phosphatase activity increases in the presence of Mn in vitro, and CpsB activity is inversely correlated with the phosphorylation of CpsD (13, 47), we used the percentage of phosphorylated CpsD-FLAG-tagged protein as an indirect indicator of CpsB activity during Mn-replete growth in pneumococcal cell lysates. We found that CpsD phosphorylation remained unchanged (Fig. 4D) despite an observable increase in CPS levels for Mn-treated ΔmntE cells (Fig. 2). Deletion of cpsB led to significantly (P ≤ 0.03) higher levels of detectable CpsD than in CpsB⁺ strains. The deletion of cpsB did not alter the overall percentage of phosphorylated CpsD (Fig. 4D and Fig. S4) but produced a slight growth advantage under all conditions examined (Fig. S5). It remains to be determined if the rate of CpsD autophosphorylation is much higher than the rate of dephosphorylation by Mn-bound CpsB in cell lysates. Given that Mn concentrations of ≥0.1 mM were needed to saturate CpsB activity in vitro, it is plausible that CpsB may be mismetallated in vivo since the relative ratios of Mn/Zn remain similar for both the WT and ΔmntE strains independent of the presence of exogenous Mn. The need for higher Mn levels in vitro suggests that an as-yet-unidentified metal chaperone might be involved in directing Mn to CpsB in vivo. In any case, these data suggest that increased CPS production during Mn-mediated toxicity may not completely result from hyperactivation of CpsB and that the mechanism of regulating CPS production may be more complex, involving multiple arms of the CPS biosynthetic pathway.

Bioavailable Mn levels regulate Pgm phosphotransferase activity.

To further investigate the role of Mn in CPS biosynthesis beyond CpsB, we surveyed the pneumococcal chromosome for additional pneumococcal enzymes involved in capsule biosynthesis or glycolysis whose activities might be influenced by transition metal ion fluctuations. Upon review, we identified SPD_1326 (referred to here as Pgm), a virulence factor annotated as a metal-utilizing phosphoglucomutase. Pgm catalyzes the reversible isomerization of G6P to G1P, a reaction that occurs at the branch point between glycolysis and CPS biosynthesis (Fig. 1). The product of Pgm, G1P, is required to generate UDP-glucose, a sugar precursor for CPS formation in S. pneumoniae. Since Mn or Zn treatments did not significantly increase pgm transcript levels (Fig. 3C and D) but rather decreased them for both WT and mutant strains, we focused our attention on examining the effects of Mn/Zn ratios on Pgm activity.

We first assessed the metal-binding affinity and stoichiometry of purified apo-Pgm protein. Based on structural and biophysical characterization of other PGM/PMM family proteins (36, 39), pneumococcal Pgm is predicted to bind at least a single metal ion. However, we find that pneumococcal apo-Pgm can bind up to two metal molar equivalents of Zn(II) or Mn(II) (Fig. 5), similarly to the binding tendencies observed for other Mn-utilizing proteins (21, 27) and that observed for CpsB as described above (Fig. S3). The K_Me values for both sites 1 and 2 differ by approximately 3 orders of magnitude for Zn (log K_a1, 8.63 ± 0.01 M⁻¹; log K_a2, 6.58 ± 0.01 M⁻¹) versus Mn (log K_a1, 5.48 ± 0.04 M⁻¹; log K_a2, ≤4.0 M⁻¹) (see Table S1 in the supplemental material for the respective K_d values). These data reveal that Zn binds more tightly to Pgm than Mn, a finding consistent with our metal homeostasis model for regulation or Mn-utilizing enzymes.

FIG 5.

Pneumococcal Pgm protein binds up to two molar equivalents of metal. (A) Titration of Mn(II) into reaction mixtures containing 9 μM (squares), 11 μM (triangles), and 20 μM (circles) apo-Pgm protein and 8, 9, or 13 μM mf2, respectively. (B) Titration of Zn(II) into reaction mixtures containing 20 μM (squares), 10 μM (triangles), and 10 μM (circles) apo-Pgm protein and 14, 12, or 11 μM mf2, respectively. Absorbances at 366 nm (filled symbols) and 325 nm (open symbols) were monitored as representations of the metal-free and metal-bound states of mf2. The continuous lines represent a nonlinear global least-squares fit employing a two-site mf2 competition model with the optimized binding parameters from multiple experiments compiled. AU, arbitrary units.

We next assessed the effects of Mn/Zn ratios on Pgm specific activity. Titration of Mn(II) into apo-Pgm resulted in an Mn-dependent increase in Pgm specific activity, in which maximum specific activity was approached at 1 mM Mn (Fig. 6A). In contrast to previous studies of other PGM/PMM family members (37, 38, 40–42, 67, 68), Mg-bound Pgm was only half as active as Mn-bound Pgm when 10 mM metal was present in the reaction mixture. At 1 mM Mg, Pgm specific activity was ∼20% of that of Mn-bound Pgm (Fig. 6A). No Pgm specific activity was observed with Zn(II) (Fig. 6A). Titration of Zn(II) into Mn-bound Pgm resulted in a steady decrease in Pgm specific activity (Fig. 6B), indicating that Zn(II) can outcompete Mn(II) for binding in Pgm. At a 1:1 Mn/Zn ratio, activity was reduced to ∼50% of that of Mn-only-treated Pgm. Zn inhibition is consistent with many other Mn-dependent proteins and PGM/PMM family members (37, 38, 42).

FIG 6.

Pgm activity in vitro is most robust with Mn. (A) Activation of purified apo-Pgm with decreasing concentrations of different metal cations. (B) Titration of Zn into Mn-bound Pgm. Data shown represent the means from at least three independent measurements ± SEM. nd, not determined; ns, not significant; ***, P ≤ 0.001 (by one-way ANOVA [unpaired]).

Our in vitro findings demonstrating that Mn activates Pgm were recapitulated in vivo using pneumococcal cell lysates from strains grown in BHI medium. ΔmntE cell lysates exhibited a 4-fold-higher Pgm specific activity than WT lysates (Fig. 7A, black bars). The addition of Mn(II) after harvest to WT lysates significantly increased Pgm specific activity 10-fold, to levels similar to those of the ΔmntE mutant (Fig. 7A, gray versus black bars). A 3-fold increase in Pgm specific activity was observed for WT lysates incubated with Mg(II) after harvest. Robust Pgm specific activity was also observed when WT cells were grown in BHI medium supplemented with Mn (Fig. 7B, black bars). The addition of Mn(II) after harvest to both WT and ΔmntE lysates resulted in a 2-fold-higher Pgm specific activity than in untreated lysates (Fig. 7B, gray versus black bars). Zn treatments were not monitored for strains since WT nontreated lysates grown in BHI medium alone produce minuscule Pgm activity, and any further Zn treatment was below the limit of detection in our assay. These findings in combination with the relatively similar pgm transcript levels observed for both untreated WT and ΔmntE strains (Fig. 3C) suggest that significantly fewer Pgm proteins in WT cells are correctly bound with the cognate Mn ion during routine microaerophilic growth than in ΔmntE cells.

FIG 7.

Pgm activity is elevated with high Mn/Zn ratios in pneumococcal cells. Pgm activity was measured before (black) and after 10 mM Mn (gray) or Mg (white) was added postharvest to cell lysates prepared from S. pneumoniae cultures grown in BHI medium (A) or BHI medium supplemented with 200 μM Mn (B). Data shown represent the means from at least three independent cultures ± SEM. ns, not significant; **, P ≤ 0.05; ***, P ≤ 0.001 (determined using Student’s unpaired t test).

Since all PGM/PMM family members, including pneumococcal Pgm (Fig. 6A), are biochemically active with Mg(II) in vitro, and several solved structures show Pgm homologues bound with Mg(II), we decided to measure the aptitude for Mg(II) to activate pneumococcal Pgm in cell lysates. The addition of Mg(II) after harvest to lysates increased WT Pgm activity 3-fold, while no impact on specific activity was observed for ΔmntE cell lysates (Fig. 7A, gray versus black bars). Likewise, no further significant specific activity increase was seen for cell lysates prepared from cells grown in BHI medium supplemented with Mn (Fig. 7B, white versus black bars). These data strongly support Mn being the preferred activating cognate metal for pneumococcal Pgm. In fact, depending on the organism, incubation with Mn(II) can produce 5 to 30% of the enzyme activity reported for Mg(II) (37, 38, 40–42, 67–69). Moreover, both Mn(II) and Zn(II) can outcompete Mg(II) for binding in multiple PGM/PMM family enzymes, resulting in reduced enzyme activity (38, 42, 70). These data collectively suggest that PGM/PMM family enzyme activities, including pneumococcal Pgm, are likely susceptible to relative changes in cellular bioavailable metal ion ratios that result in downstream implications for the production of CPS and other exoproteins. As such, we conclude that cellular metal ratios, particularly Mn/Zn ratios, will likely act directly on Pgm function, thereby establishing a regulatory mechanism for Mn in CPS production by S. pneumoniae.

DISCUSSION

CPS is considered the most prominent pneumococcal virulence factor (6, 7, 71). Its expression level greatly influences pneumococcal carriage and disease progression (8). Since higher CPS levels are unfavorable for pneumococcal colonization but necessary for evasion of host defense systems (8, 72, 73), it is postulated that CPS expression must be precisely regulated. In this study, we provide novel insight into the mechanism of CPS regulation by identifying a role for cellular metal ions in pneumococcal CPS biosynthesis. Our results demonstrate that pneumococcal cells can adjust CPS production levels in response to the nutritional metal ion pool, more specifically to bioavailable Mn/Zn ratios.

We show that dysregulation of overall metal homeostasis directly affects the metallation status of pneumococcal Pgm, ultimately affecting the production of sugar precursors for CPS formation (Fig. 8). We propose that high Mn/low Zn cellular ratios induce a state of hyperactivation for Pgm resulting in G1P abundance and likely positive flux through the CPS biosynthesis pathway. In contrast, low Mn/high Zn cellular ratios lead to inactive Pgm reducing G1P concentrations, which restrains the further production of sugar precursors and limits CPS formation. These findings are consistent with those reported for other Mn-dependent enzymes (27, 74, 75) and provide additional support for our previously proposed model, which states that Mn metalloenzymes are metallated by default, allowing the pool of bioavailable metal ions to dictate enzyme specific activities (27).

FIG 8.

Schematic overview of the effects of cellular Mn/Zn ratios on CPS production by pneumococcus. Dysregulation of Mn/Zn homeostasis directly influences the metallation status of the Mn-utilizing phosphoglucomutase Pgm, thereby affecting sugar precursor production necessary for CPS formation. A high Mn/low Zn ratio induces a state of hyperactivation for Pgm, resulting in thicker CPS (left), while a low Mn/high Zn ratio leads to inactive Pgm, thereby restraining sugar precursor production and limiting CPS formation (right). The Mn/Zn ratio can also regulate CpsB phosphatase activity, but its exact role in modulating CPS biosynthesis is still unclear.

Similar to characterized PGM/PMM family enzymes (37, 40, 76, 77), we find that pneumococcal Pgm can bind a variety of metal ions, but only Mn(II) leads to robust enzyme activity. This is not entirely surprising given the nature of the pneumococcus as an Mn-centric pathogen, where the Mn/Fe and Mn/Zn ratios are each ≥1. More interesting is the fact that we observed two metal ion binding events per Pgm monomer. This contrasts with solved crystal structures of PGM/PMM family members, which have been shown to be bound to a single Mg(II) ion. A multiprotein sequence alignment comparing pneumococcal Pgm to select PGM/PMM family homologues reveals both shared conserved and nonconserved regions (see Fig. S6 in the supplemental material) that could account for the different enzymological properties. The pneumococcal Pgm amino acid sequence harbors the single conserved metal-binding motif (DXDXDR) found among all PGM/PMM family enzymes. We note that other phosphotransferases can bind multiple metal ions through the coordination of residues in the proximity of the substrate active site. For example, the Mn-dependent phosphopentomutase DeoB binds two Mn(II) ions per monomer, through the coordination of aspartate and histidine residues (78, 79). Further investigation is under way to identify spatially nearby active-site residues or an additional allosteric metal-binding pocket that coordinates the second metal ion-binding event observed in this study. The activation difference observed by the metallation status with Mn(II) suggests that pneumococcal Pgm may constitute a novel bacterial enzyme subclass of the PGM/PMM superfamily.

Our data suggest that the pneumococcus has evolved a multilayered approach to regulate CPS biosynthesis and coordinate its production with cell division. Reports show that carbohydrates, oxygen, and metal ions can vary significantly between host environment niches (16, 80, 81). Although Mn import is induced by the presence of oxygen in the pneumococcus (27), other host factors, such as the secretion of calprotectin, may also contribute to the physicality of Mn/Zn bioavailability. Further complications arise from the fact that measurements are of total metals present in host tissues and do not represent the free bioavailable Mn concentration that pathogenic bacteria likely experience. Therefore, we currently cannot determine whether Mn/Zn availability is primary or secondary to carbohydrate acquisition by S. pneumoniae.

We note, however, that the ability of S. pneumoniae to acquire carbohydrates for the synthesis of G6P, the substrate for Pgm, is ensured by over 10 extracellular glycan hydrolases, and approximately 30% of all uptake transporters are dedicated to importing sugars (82–84). The scarcity of free carbohydrates in the nasopharynx may prevent CPS production and promote initial colonization, while carbohydrate abundance would lead to robust CPS levels (80, 81). Since we cannot assume that the overall trend of free carbohydrate and Mn/Zn availability is a fair representation during pneumococcal pathogenesis, we propose two possibilities for the regulation of CPS expression. First, free carbohydrate and Mn concentrations may trend similarly within various host niches, excluding blood serum, thereby functioning additively to regulate pneumococcal CPS expression. As the pneumococcus becomes more invasive, both sugar substrates and Mn ions become readily available, resulting in thicker CPS. Alternatively, free carbohydrate and Mn concentrations may not trend similarly, yet the pneumococcal regulatory mechanism allows the two to function synergistically with each other. This would ensure that CPS is produced at a minimal level to provide some protection, especially since nonencapsulated pneumococcal strains are avirulent in invasive models of murine infections (71, 73, 85, 86). This multilayered regulatory model also allows for the addition of other mechanisms, which we are currently unaware of but may be revealed through further investigations. More importantly, it may also explain the inconsistencies observed between various reports owing to the diversity of pneumococcal strains and growth conditions used in standard laboratory practice.

As for CpsB, our data are consistent with the literature in that pneumococcal CpsB binds up to two metal ions and that CpsB is most active with Mn in vitro (44, 45). Our in vivo data do not eliminate or support a role for CpsB in regulating CPS expression levels since the percentage of phosphorylated CpsD remained unchanged, even though we suspect Mn hyperactivation of CpsB when exogenous Mn is supplied. It remains to be determined if the rate of CpsD autophosphorylation is much higher than the rate of dephosphorylation by Mn-bound CpsB in pneumococcal cells. Given that Mn concentrations of ≥0.1 mM showed saturated CpsB activity in vitro (Fig. 4A), it is plausible that CpsB might be mismetallated in vivo since the relative ratios of Mn/Zn remain similar for both the WT and ΔmntE strains independent of the presence of exogenous Mn. On the other hand, perhaps CpsB activity may not be overly influential in CPS biosynthesis as previously believed but instead functions as a mediator of the conformational configuration cycling of CpsD between octamer and monomer arrangements, thereby indirectly modulating CPS biosynthesis. More studies are needed to improve the knowledge of the physiological role of the pneumococcal CpsB and CpsD phosphoregulatory system.

In summary, our findings provide a direct role for Mn in the regulation of CPS expression, ultimately connecting previous results by others under various growth conditions (9–14, 87, 88). We show that the specific activities of both Pgm and CpsB are dependent on the relative cellular metal ion pool, in which Mn/Zn ratios dictate the activity of these metalloenzymes. We propose a multilayer mechanism in which both carbohydrate and Mn/Zn availability collaboratively function to ensure the production of critical sugar precursors for CPS formation. This multilayer mechanism allows for the precise regulation of CPS biosynthesis for virulence in the various host niches by exposing necessary surface proteins during colonization of the nasopharynx and by protecting the cell body from the host immune response during disease progression.

MATERIALS AND METHODS

Bacterial strains and plasmid construction.

Bacterial strains and plasmids used in this study are listed in Table S2 in the supplemental material. All S. pneumoniae mutants were derived from wild-type S. pneumoniae serotype 2 D39 strain IU1781. Strains ISU52, IU17951, and IU17953 were constructed using the Janus cassette allele replacement method and counter-antibiotic selection as previously described (57, 89, 90). Bacterial strains expressing CpsD-3×FLAG-tagged protein behind its native promoter were constructed using standard methods as previously described (28, 91). When required, the following antibiotics were added to the culture media: kanamycin at 250 μg ml⁻¹, streptomycin at 250 μg ml⁻¹, and erythromycin at 0.3 μg ml⁻¹. All constructs were PCR and sequence verified.

Plasmids pCpsB and pPgm were constructed by amplifying the cpsB (SPD_0316) and pgm (SPD_1326) open reading frames from S. pneumoniae D39 genomic DNA. The PCR products were cloned into the pHisCT or pHis-Parallel vector behind a T7 promoter using isothermal assembly, transformed into Escherichia coli DH5α, and selected on culture medium containing 100 μg ml⁻¹ ampicillin. The resulting plasmid constructs were sequence verified and transformed into E. coli BL21 for protein expression and purification.

Growth conditions and phase-contrast microscopy of S. pneumoniae.

Standard BHI broth was prepared with MilliQ water. All S. pneumoniae strains were routinely grown in BHI broth at 37°C in an atmosphere of 5% CO₂. Briefly, bacterial glycerol stocks were inoculated into BHI broth, serially diluted, and grown overnight. The next day, exponentially growing cultures were diluted to an optical density at 620 nm (OD₆₂₀) of 0.005 in prewarmed BHI broth supplemented with MnCl₂ or ZnSO₄ as indicated. Cellular growth was monitored over time as the OD₆₂₀.

For microscopic analysis, a Quellung reaction was performed by mixing equal volumes of the cell culture (OD₆₂₀ of 0.2) and rabbit anti-S. pneumoniae type 2 antibody (Abcam). Samples were examined immediately by phase‐contrast microscopy. Swollen, dark cell bodies surrounded by visible capsules were imaged.

RNA isolation and quantitative real-time PCR (RT-PCR).

RNA was extracted from S. pneumoniae cells after 4 h of growth, treated with DNase, and converted to cDNA using standard laboratory procedures as previously described (92). PCR amplification was carried out using the DNA primer pairs listed in Table S3. PCR amplification was performed in a mixture (20-μl final volume) containing 2× SYBR green quantitative PCR (qPCR) master mix (Agilent Technologies), 250 nM each primer, and a 1/24 dilution of cDNA. Cycling conditions were as follows: 95°C for 3 min and 40 cycles of 95°C/20 s and 59°C for 20 s. PCR outcomes were normalized to the gyrA gene, and relative transcription levels were calculated by comparison of the ratio of treated to nontreated cells (93).

Mucoviscosity assay.

Exponentially growing S. pneumoniae cells were harvested at an OD₆₂₀ of approximately 0.2 by centrifugation at 4°C. Cell pellets were resuspended in 1/10 or 1/40 of the original culture volume with ice-cold phosphate-buffered saline (PBS) (pH 7.4) containing 11.9 mM phosphate, 137 mM NaCl, and 2.7 mM KCl. Cell suspensions were diluted to an OD₆₂₀ of 1.0 in 1 ml ice-cold PBS and centrifuged for 3 min at 1,000 × g at 4°C (59, 60), and the turbidity was reassessed at 620 nm. The fraction of remaining cells in solution was calculated by dividing the final OD₆₂₀ by the initial OD₆₂₀.

Uronic acid assay.

Uronic acid was measured from exponentially growing S. pneumoniae cultures using a protocol modified slightly from the one previously described (61). Briefly, cells were harvested by centrifugation after approximately 4 h of growth in BHI medium supplemented with increasing concentrations of MnCl₂ or ZnSO₄. Cell pellets were suspended in 150 mM Tris-HCl (pH 7.0) containing 1 mM MgSO₄, incubated with 0.1% deoxycholate, and subjected to enzymatic digestion of the cell wall (100 U mutanolysin), nucleic acids (0.1 mg/ml DNase and 0.1 mg/ml RNase), and proteins (0.1 mg/ml proteinase K). The supernatant was collected after centrifugation, mixed with a 98% (vol/vol) sulfuric acid–12.5 mM tetraborate solution, boiled, cooled on ice, and mixed with 0.15% (wt/vol) 3-phenylphenol in 0.5% NaOH. The absorbance was read at 520 nm at room temperature within 5 min. Samples incubated with 0.5% NaOH served as the background, and measurements were subtracted out before being normalized to total protein determined using the DC protein assay (Bio-Rad).

Mobility shift assay for detection of phosphorylated CpsD.

Phos-tag SDS-PAGE followed by Western blotting was performed as previously described (94). Cultures expressing CpsD-3×FLAG protein were harvested after 4 h of growth in BHI medium with or without MnCl₂ by centrifugation, washed with cold 20 mM Tris-HCl (pH 7), suspended in 1/30 of the original culture volume with cold 20 mM Tris-HCl (pH 7) containing a Pierce protease and phosphatase inhibitor cocktail (Thermo Scientific), and lysed mechanically using a FastPrep homogenizer (MP Biomedicals). Total protein was determined using the DC protein assay (Bio-Rad). Cell lysates (10 μg total protein, determined to be in the linear range using serially diluted cell lysates) were resolved by Phos-tag SDS-PAGE at 4°C (95). TotalStainQ (Azure Biosystems) was used to further quantitate total protein loaded for normalization of downstream Western blots. CpsD-3×FLAG protein was detected by Western blotting using rabbit anti-FLAG primary antibody (Sigma-Aldrich) and goat anti-rabbit IgG secondary antibody conjugated to IR800dyeCW (Li-Cor) using infrared fluorescence. Protein bands in Western blots were imaged and quantitated using an Azure Biosystems imaging system (Azure 600) and AzureSpot software (Azure Biosystems).

Expression and purification of S. pneumoniae CpsB and Pgm.

Overnight-grown E. coli BL21 cells harboring either pCpsB or pPgm were diluted 1/100 into Luria-Bertani (LB) broth containing 100 μg ml⁻¹ ampicillin and grown aerobically at 37°C. At an OD₆₀₀ of approximately 0.6, 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) was added to induce protein synthesis. Cells were harvested by centrifugation after approximately 20 h of growth at 18°C and lysed in low-imidazole (20 mM) buffer [25 mM Tris-HCl (pH 8.0), 200 mM NaCl, 3 mM tris(2-carboxyethyl)phosphine] by sonication on ice. Cell debris was removed by centrifugation. Target proteins were purified using a Ni(II)-nitrilotriacetic acid (NTA) affinity column (GE Healthcare) preequilibrated with low-imidazole buffer. Bound proteins were eluted over a concentration gradient with high-imidazole (500 mM) buffer. Pooled fractions of eluted proteins were incubated with tobacco etch virus protease at 4°C for 48 h to remove the histidine-rich tag. The buffer was exchanged back to low-imidazole buffer, and the protein solution was injected onto a Ni(II)-NTA affinity column. The protein flowthrough was collected, concentrated, and further purified on a HiLoad 16/600 Superdex 75 size exclusion column (GE Healthcare) preequilibrated with buffer containing 25 mM HEPES (pH 8.0), 200 mM NaCl, 3 mM dithiothreitol (DTT), and 0.5 mM EDTA. EDTA was removed by running the solution through a gravity flow PD-10 desalting column (GE Healthcare). Divalent metal cations were removed from the solution by dialyzing the protein against several periods of 1 liter of buffer containing 25 mM HEPES (pH 8.0), 200 mM NaCl, 5% glycerol, and 3 mM DTT with 10 g/liter Chelex-100 resin (Bio-Rad). The final protein purity and identity were confirmed by SDS-PAGE and mass spectrophotometry. Protein concentrations were calculated using extinction coefficients of 18,000 M⁻¹ cm⁻¹ for CpsB and 53,750 M⁻¹ cm⁻¹ for Pgm.

Metal-binding affinity and stoichiometry determinations.

Metal-binding experiment data for purified CpsB and Pgm proteins were used to calculate the stoichiometry and metal affinity. Briefly, purified apo-CpsB or apo-Pgm protein was diluted into 1 ml metal-free buffer (25 mM HEPES [pH 8.0], 100 mM NaCl) containing fixed concentrations of metal-free mag-fura-2 (mf2), a metal competition chelator. Mn(II) and Zn(II) were independently titrated into the protein-mf2 solution. The absorbances of the solution at 366 nm and 325 nm were monitored after a 2-min incubation at room temperature, which reported on the λ_max for the metal-free and metal-bound states of mf2 using a Hewlett-Packard model 8452A spectrophotometer (21). The mf2 binding curves were fitted to an appropriate one-step competition binding model using Dynafit [96; see reference 21 for a sample Dynafit script file used to analyze metal (Mn,Zn)-binding equilibria for CpsB and Pgm].

Tyrosine phosphatase assay.

The phosphotyrosine phosphatase activity of purified CpsB protein toward the chemically synthesized phosphopeptide DADE(pY)LIPQQG was measured using the tyrosine phosphatase assay system (Promega). The standard assay was carried out in a final volume of 50 μl metal-free buffer (25 mM HEPES [pH 8.0], 100 mM NaCl), 100 μM phosphopeptide substrate, metal cation, and 5 μg of apo-CpsB at 37°C for 30 min (44). Reactions were terminated by the addition of a molybdate dye-additive mixture and allowed to develop for 15 min at room temperature. The concentration of released P_i was determined using a standard curve prepared from known concentrations of P_i measured spectrophotometrically at 600 nm.

Phosphoglucomutase assay.

A standard tandem enzymatic assay coupling Pgm activity to glucose 6-phosphate dehydrogenase (G6-PDH) was used to allow NADP⁺ reduction to be monitored over time (Sigma-Aldrich). G6-PDH was supplied in excess to ensure that the secondary reaction in the phosphoglucomutase assay was not the rate-limiting step. The reduction of NADP⁺ was monitored spectrophotometrically at 340 nm. The standard assay for in vitro analysis was conducted at room temperature in a 600-μl reaction volume. Assay components included 1 μg/ml of apo-Pgm in 176 mM metal-free glycylglycine (Gly-Gly) buffer (pH 7.4), 5 mM glucose 1-phosphate (G1P), 670 μM NADP⁺, 390 μM glucose 1,6-bisphosphate, 43 mM l-cysteine, metal cation, and 0.3 U/ml of G6-PDH (36, 97).

Pgm activity in pneumococcal lysates was evaluated using methods similar to the ones described above for the in vitro analysis of purified Pgm protein. Exponentially growing S. pneumoniae cells were harvested after 4 h of growth by centrifugation, resuspended in 1/50 of the volume of the original culture with 250 mM metal-free Gly-Gly buffer (pH 7.4), and lysed by sonication. Cell debris was removed by centrifugation. Cell lysates were incubated with 5 mM G1P for 5 min at room temperature prior to the addition of other assay components as described above. Activities were normalized to total protein determined by the DC protein assay (Bio-Rad).

Statistical analysis.

All growths and assays were performed independently in at least triplicate on multiple days. When appropriate, statistical analysis was performed using Student unpaired t tests or one-way analysis of variance (ANOVA) (unpaired) using GraphPad Prism software. Accepted P values are indicated in the figure legends.