Bacterial strains and plasmids used in this study are listed in Table 1. Bacteria were routinely grown at 37°C and 150 rpm in tryptic soy broth (TSB) purchased from Becton Dickinson (Heidelberg, Germany) with a culture to flask volume ratio of 1:10. As required, the medium was supplemented with 50 μg kanamycin per milliliter.

Mice challenged with a S. aureus ccpA mutant displayed a reduced bacterial load in liver tissue in an abscess model (Li et al., 2010), suggesting that CcpA increases infectivity of S. aureus. To expand upon the in vivo function of CcpA in facilitating S. aureus infections, the effect of CcpA on infectivity was tested with an isogenic S. aureus strain pair Newman/MST14 (Newman ΔccpA; Seidl et al., 2006) in three murine infection models: an acute pneumonia model (Hartmann et al., 2014), a subcutaneous footpad swelling model (Nippe et al., 2011), and a catheter-related infection model (Rupp et al., 1999). In contrast to the murine abscess model (Li et al., 2010), the infectivity of S. aureus wild-type strain Newman and the ΔccpA mutant were equivalent in all three animal models when using normoglycemic C57BL/6N mice (Figure 1). Specifically, the bacterial loads in lung tissue and BAL were similar in the pneumonia model (Figure 1A); the swelling of footpads displayed comparable kinetics and maxima (Figure 1B); and the edema formation rates and bacterial loads in the catheter lumen and surrounding tissues were indistinguishable in the catheter infection model (Figure 1C). Taken together, these data demonstrate that under normoglycemic conditions infectivity of S. aureus is independent of CcpA.

Strain Newman and the isogenic ccpA mutant caused equivalent infections in three mouse models, which differs from that reported by Li et al. (2010). To determine if this difference was due to the animal model or to differences in the type of ccpA mutation (i.e., deletion mutant vs. transposon mutant), the infectivity of strains Newman and the ΔccpA mutant in the murine abscess model was assessed (Li et al., 2010). Consistent with the findings of Li et al. (2010), a significant decrease (about 3-log) in the bacterial loads in liver was observed in mice challenged with the ΔccpA mutant strain MST14 as compared to mice infected with the wild-type strain Newman (Figure 2A). Mice challenged with a trans-complemented MST14 derivative harboring plasmid pCN34_ccpA displayed bacterial loads in liver comparable to wild-type infected animals (1.1 × 10⁸ ± 9.7 × 10⁷ CFU/g vs. 3.2 × 10⁷ ± 1.8 × 10⁷ CFU/g; P = 0.238). These data confirm the previous observations and demonstrate that decreased bacterial burden in liver of mice infected with ccpA mutants was due to the absence of CcpA. In addition, the data raise the question as to why CcpA is important in liver abscesses, but dispensable in other organs or anatomical sites?

In systemic infections, S. aureus elicits a strong pro-inflammatory response by triggering the production and release of interleukins (IL)-1β (IL-1β), IL-6, IL-8, IL-18, and tumor necrosis factor alpha (TNFα) in immune and non-immune cells (Cui et al., 2000; Feezor et al., 2003; Hessle et al., 2005; Muller-Anstett et al., 2010). In S. aureus ccpA mutant challenged mice, the decreased bacterial loads in the liver raised the question of whether the lack of CcpA might also affect the innate immune response of the infected mice. To test this hypothesis, the concentrations of IL-1β, IL-6, TNFα, and the chemokine keratinocyte chemo-attractant (KC, syn. CXCL-1) were determined in mouse liver homogenates 4 days post-infection (Figures 2B,C and Figure S1). As expected, the concentrations of IL-1β (Figure 2B) and KC (Figure 2C) in the liver homogenates of MST14 challenged mice were significantly decreased relative to wild-type infected mice. Similarly, the cytokines IL-6 and TNFα were decreased in MST14 challenged mice; however, these effects were not statistically significant (Figure S1). Upon challenge of mice with the ccpA complemented MST14 derivative (MST14 pCN34_ccpA), the cytokine concentrations in liver tissue were restored to levels similar to mice infected with the wild-type strain (Figure 2B). Taken together, these data suggest that ccpA-positive S. aureus provoke an increased pro-inflammatory response in infected liver tissue.

Recently, it was determined that the DNA binding activity of CcpA in S. aureus is controlled by the serine/threonine kinase Stk1 (Leiba et al., 2012). To test whether the Stk1-dependent phosphorylation of CcpA would affect the infectivity of S. aureus, C57BL/6N mice were challenged with a MST14 derivative trans-complemented with a ccpA allele having threonine to alanine mutations at positions 18 and 33 (MST14 pCN34_ccpA_Ala) and the bacterial load and cytokine profiles were determined. This CcpA_T18A/T33A variant can no longer be phosphorylated by Stk1, but retains the regulatory activity toward its target genes (Leiba et al., 2012). Interestingly, while the bacterial loads in livers of mice infected with strain MST14 complemented with the wild-type ccpA allele were similar to those of mice challenged with the wild-type strain, the infection progress differed in mice infected with MST14 pCN34_ccpA_Ala (Figure 2A). Specifically, mice challenged with MST14 pCN34_ccpA_Ala significantly increased the bacterial load in livers by a factor of ~1.8-log relative to the MST14 infected mice, but remained ~1.2- and ~0.7-log below the levels of the wild-type and MST14 pCN34_ccpA strains, respectively (Figure 2A). Similarly, infecting mice with strain MST14 pCN34_ccpA_Ala resulted in a significant increase in the IL-1β levels in liver relative to that observed in mice challenged with the wild-type or MST14 pCN34_ccpA strains (Figure 2B). In contrast, KC levels were indistinguishable between these three groups (Figure 2C). Taken together, under in vivo conditions the Stk1-dependent phosphorylation of CcpA dramatically alters the infection process.

As mentioned, mice challenged with a ccpA mutant had decreased bacterial numbers in the liver (Figure 2A), but not in kidneys (Li et al., 2010). This raised the question as to why the effect of ccpA inactivation was specific to the liver. Notably, glucose concentrations in the liver are elevated in comparison to those found in blood and intestinal fluids of the same animals (Appelboom et al., 1959; Wals and Katz, 1993). These observations suggest that CcpA might contribute to the infectivity of S. aureus in hyperglycemic individuals. To address this hypothesis, the infectivity of strains Newman, MST14, and MST14 pCN34_ccpA was assessed in an acute pneumonia model using female NOD mice that spontaneously develop a type 1 diabetes between 15 and 30 weeks of age (Leiter, 2001). Consistent with the pneumonia model using C57BL/6N mice (Figure 1A), challenging normoglycemic NOD mice (blood glucose level ≤ 10 mM) with any of the three strains resulted in equivalent CFU counts in lung tissue homogenates (Figure 3A) and BAL (Figure 3B) 24 h post-infection. Similarly, the total eukaryotic immune cell counts in BAL were comparable between all groups (Figure 3C), indicating an analogical course of infection caused by the three S. aureus strains. In contrast, infection of age-matched diabetic NOD mice (blood glucose level ≥20 mM) with the ΔccpA mutant MST14 significantly increased the bacterial loads relative to the wild-type strain in lung tissue (~2.7-log; Figure 3A) and BALs (~2.5-log; Figure 3B). This was accompanied by a significant increase in total immune cells in BALs of the MST14 challenged mice (Figure 3C), indicating an increased severity of infection. Complementation of the ΔccpA mutant with plasmid pCN34_ccpA restored all virulence traits to wild-type levels, confirming that the alterations observed with MST14 were caused by inactivation of ccpA.

To determine if CcpA was important for extrapulmonary S. aureus infections in diabetic NOD mice, strain Newman and MST14 were evaluated in the footpad swelling model and catheter related infection models. The pCN34_ccpA harboring MST14 derivative was not included, as maintenance of the pT181-derived plasmid within the trans-complemented derivative was not assured over the entire time course of these in vivo models (Krute et al., 2016). Similar to our findings with the pneumonia model, the wild-type and ΔccpA mutant infected normoglycemic mice displayed comparable footpad swelling kinetics (Figure 4A), and both strains persisted in the infected foot tissue for a comparable number of 11 days p.i. (Figure 4B). In contrast, when hyperglycemic NOD mice were infected with strains Newman and MST14, clear differences in the footpad swelling kinetics were observed (Figure 4A). The footpads of mice infected with the wild-type strain Newman showed enhanced swelling rates 5 days p.i. that were significantly increased at days 8 and 9 when compared to the footpad swelling rates of the strain MST14 challenged mice. Consistent with increased footpad swelling, significantly greater numbers of bacteria (~7.1-fold) were recovered from the tissue homogenates of infected feet of mice challenged with wild-type S. aureus in comparison to mice infected with the ΔccpA mutant (Figure 4B).

Similar to the footpad swelling model, infection symptoms in mice infected with strain Newman and MST14 were equivalent in normoglycemic NOD mice using a catheter related infection model (Rupp et al., 1999; Figure 5). In contrast, infection of diabetic NOD mice with the wild-type strain Newman significantly increased edema (Figure 5A) and enhanced bacterial loads in the tissue surrounding the catheter (~1-log; Figure 5B) compared to mice infected with the ΔccpA mutant strain MST14. In total, these data suggest that CcpA is critically important for the pathogenesis of S. aureus in hyperglycemic individuals.

CcpA increases the transcription of the α-hemolysin encoding gene hla in a glucose-responsive manner (Seidl et al., 2006; Leiba et al., 2012), and this toxin is a major virulence factor of S. aureus (reviewed in Vandenesch et al., 2012). To determine whether CcpA-mediated changes in hla transcription translate into a greater α-hemolysin accumulation, the capacities of S. aureus strain Newman and its ccpA derivatives to lyse rabbit erythrocytes were determined. Rabbit erythrocytes are particularly sensitive to the pore-forming toxin α-hemolysin (Cooper et al., 1966). When S. aureus strains Newman, MST14, and MST14 harboring pCN34_ccpA were streaked onto RBA plates lacking glucose, all three strains produced small and comparable hemolytic zones surrounding the growing colonies (Figure 6A). In contrast, when the three strains were streaked onto RBA plates supplemented with 0.5% glucose (~28 mM), strains Newman and MST14 harboring pCN34_ccpA produced larger hemolytic areas than strain MST14 (Figure 6B), suggesting that CcpA augments α-hemolysin accumulation in a glucose-rich environment.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.