Inhibition of bacterial superoxide defense

Nutritional immunity is the name given to the sequestration of essential metals by vertebrates to limit growth of invading pathogens such as Staphylococcus aureus. While iron sequestration is the canonical example of nutritional immunity, recent work has highlighted the importance of manganese and zinc sequestration by vertebrates in controlling both bacterial and fungal infections. Our laboratories are investigating the biochemical basis for the sequestration of essential metals and the impact of this nutrient deprivation on the physiology of S. aureus and other pathogens. We have shown that during staphylococcal infection the manganese and zinc binding protein calprotectin (CP) is a critical component of nutritional immunity. Furthermore, this neutrophil protein is required to render the staphylococcal abscess devoid of manganese. However, the structural features responsible for binding these metals were unknown as was the impact of this CP-induced nutrient deprivation on S. aureus. In our recently published manuscript, we showed using isothermal titration calorimetry and mutagenesis experiments that CP contains two transition metal binding sites, and is capable of binding two zinc ions or one manganese ion with nanomolar affinity. Moreover, CP was found to reduce staphylococcal superoxide dismutase activity and other antioxidant defenses resulting in the accumulation of intracellular superoxide and increased sensitivity to oxidative stress. The inhibition of superoxide defenses by CP renders S. aureus more sensitive to neutrophil-mediated killing while animal modeling suggests that manganese sequestration by the host during infection reduces staphylococcal superoxide dismutase activity. In total, these results suggest a two hit model for the antimicrobial activity of CP whereby this protein both inhibits bacterial growth and renders microbial invaders more sensitive to other immune effectors.

Vertebrates utilize numerous approaches to control invading pathogens. One particularly powerful approach is the withholding of essential nutrients such as metals that are required for growth and proliferation of the invading pathogen. This defense has been termed nutritional immunity and the importance of this strategy is exemplified by the multitude of mechanisms utilized by the host to prevent access to iron (Schaible and Kaufmann, Nat Rev Microbiol 2004; Weinberg, Biochim Biophys Acta 2009). In addition to withholding iron during infection, vertebrates also sequester the essential metals manganese and zinc (Corbin et al., Science 2008). Nutritional immunity based on manganese and zinc sequestration is a potent defense against invaders because these elements play critical structural and catalytic roles in numerous bacterial processes. This defense contributes to controlling a wide range of pathogens including Staphylococcus aureus (Corbin et al., Science 2008; Urban et al., PLoS pathog 2009). Increasing antibiotic resistance and the high rates of morbidity and mortality associated with infection has resulted in S. aureus becoming a pathogen of substantial medical concern and highlights the need for new therapeutics (Grundmann et al., Lancet 2006; Lowy, N Engl J Med 1998; Said-Salim et al., Infect Control Hosp Epidemiol 2003).

The neutrophil protein calprotectin (CP) is a key contributor to nutritional immunity. CP is a manganese and zinc binding protein that can be found at sites of infection at concentrations in excess of 1 mg/ml (Clohessy and Golden, Scand J Immunol 1995). Mice lacking CP fail to sequester manganese away from S. aureus abscesses, and suffer higher bacterial and fungal burdens following infection (Corbin et al., Science 2008; Urban et al., PLoS pathog 2009). These observations highlight the importance of CP to nutritional immunity and the control of infection. CP inhibits S. aureus growth in vitro and this inhibition is reversed by the addition of excess manganese or zinc (Corbin et al., Science 2008). However, the structural features responsible for chelating manganese and zinc as well as the staphylococcal processes disrupted by this host defense were unknown. To lay the groundwork for the creation of therapeutics that leverage nutritional immunity, we set out to address these gaps in our knowledge regarding the impact of CP on infection.

CP is a member of the S100 class of EF-hand calcium binding proteins, which have several unique characteristics including cell-specific expression patterns as well as diverse intracellular and extracellular functions. S100 proteins are associated with a wide range of processes including cell differentiation, growth, motility and host defense (Heizmann, Methods Mol Biol 2002). Although there is some variability in their sequences, each subunit contains two EF-hand calcium binding motifs and typically self-associates to form homodimers (Heizmann et al., Front Biosci 2002). Unlike most S100 proteins, CP is a heterodimer, comprised of S100A8 and S100A9, which is highly preferred over the corresponding homodimeric species (Hunter and Chazin, J Biol Chem 1998).

Although it is clear that manganese and zinc sequestration by CP is important to controlling infection, the molecular basis of CP’s antimicrobial activity is not known. To address these issues we are using an integrated approach combining chemistry, biophysical and structural analysis, and microbiology. We began by measuring the affinities of CP for manganese and zinc using isothermal titration calorimetry (ITC). This analysis revealed that CP is capable of binding a single manganese ion or two zinc ions with nanomolar affinity, suggesting that CP is capable of exquisite discrimination of transition metals. Since there are no high-resolution structures of transition metal-bound CP, we constructed a homology model based on the high-resolution crystal structure of the Zn-S100A12 complex (Moroz et al., J Mol Biol 2009) to generate hypotheses regarding the residues involved in manganese and zinc binding. This model suggested that CP possesses two transition metal binding sites. One site utilizes residues H17 and H27 from S100A8 and H91 and H95 from S100A9, while the other consists of residues H83 and H87 from S100A8 as well as H20 and D30, from S100A9. This hypothesis was tested by generating a mutant CP (ΔZn/Mn), in which all seven histidines were mutated to asparagines and the aspartic acid to serine. ITC experiments revealed manganese and zinc binding were abrogated in this mutant. To ensure that the lack of binding was not the result of structural effects, NMR spectroscopy was used to establish that ΔZn/Mn retains the native global structure. The antimicrobial activity of ΔZn/Mn was assessed and consistent with the proposed mechanism of action; this variant is effectively inactive with a 50% inhibitory concentration (IC₅₀) that is nine times the concentration found within host tissues (Clohessy and Golden, Scand J Immunol 1995). These results conclusively demonstrated the importance of manganese and zinc binding for the antimicrobial activity of CP. A number of other S100 proteins are known to bind zinc or copper including S100A7, S100A12 and S100B (Gläser et al., Nat Immunol 2004; Moroz et al., Acta Crystallogr D Biol Crystallogr 2003; Ostendorp et al., Biochim Biophys Acta 2011). Of these, S100A7 is known to inhibit bacterial growth and S100A12 is expressed by neutrophils (Gläser et al., Nat Immunol 2004; Moroz et al., Acta Crystallogr D Biol Crystallogr 2003). These observations suggest that other S100 proteins besides CP may contribute to nutritional immunity and sequester a variety of metals away from invading pathogens.

Having identified residues that are required for transition metal binding by CP and generated a powerful reagent, we next sought to determine the impact of CP-mediated manganese and zinc sequestration on S. aureus pathogenesis. The sequestration of manganese and zinc by CP presumably inactivates metal-dependent staphylococcal processes, the loss of which results in reduced bacterial growth. While approximately 6% of bacterial proteins are predicted to utilize zinc or manganese (Andreini et al., J Proteome Res 2006; Papp-Wallace and Maguire, Annu Rev Microbiol 2006), few have been experimentally validated. Further confounding the identification of the bacterial processes disrupted by CP is the observation that a single metal-dependent protein may be capable of using multiple metals to generate biochemical activity (Sobota and Imlay, Proc Natl Acad Sci USA 2011). One set of staphylococcal processes for which the metal dependency is known is superoxide resistance. S. aureus possesses two Mn-dependent mechanisms for coping with superoxide stress. The first mechanism utilized by S. aureus to resist superoxide stress is canonical detoxification via two Mn-dependent superoxide dismutases (SOD) known as SodA and SodM (Clements et al., J Bacteriol 1999; Valderas and Hart, J Bacteriol 2001). These proteins convert superoxide to hydrogen peroxide, which is subsequently converted to water by catalase. The second mechanism utilized by S. aureus to resist superoxide stress is uncharacterized. However, this process is known to be Mn-dependent and SOD-independent (Horsburgh et al., Mol Microbiol 2002; Horsburgh et al., Trends Microbiol 2002). These two mechanisms combine to protect S. aureus from endogenous sources of superoxide stress, such as respiration and exogenous sources, such as the oxidative burst of neutrophils.

If CP-mediated metal sequestration inhibits staphylococcal oxidative stress defenses, we hypothesized that CP treatment would increase the sensitivity of S. aureus to superoxide generating compounds. CP increases the sensitivity of S. aureus to the superoxide generating compounds paraquat and xanthine/xanthine oxidase, while the addition of glutathione reverses the enhanced sensitivity of S. aureus to paraquat challenge following CP treatment. These data indicate that CP renders S. aureus more sensitive to superoxide stress but do not address whether the enhanced sensitivity is mediated by metal sequestration. To address this issue, the ability of CP to increase staphylococcal sensitivity to superoxide was examined in the presence of excess manganese or zinc. The increased sensitivity of S. aureus to superoxide stress is reversed by the addition of excess manganese or zinc. Additionally, in contrast to wild-type CP, the ΔZn/Mn mutant does not enhance the sensitivity of S. aureus to oxidative stress. Together, these data indicate that manganese and zinc sequestration by CP is necessary to increase the sensitivity of S. aureus to superoxide. In addition to S. aureus Newman, we observed that CP treatment also increases the sensitivity to superoxide of USA300, the predominant community-associated methicillin resistant S. aureus isolate in the United States (Klevens et al., JAMA 2007), and S. epidermidis.

The enhanced sensitivity of Staphylococci to superoxide upon CP treatment supported our hypothesis that CP disrupts SOD defense mechanisms. To test this, we began by assessing the impact of CP treatment on SOD activity. CP reduces staphylococcal SOD activity to near background levels. Surprisingly, CP also reduces SOD levels to near background when superoxide stress is added, the addition of which normally results in a 5-fold increase in SOD activity. The addition of excess manganese reverses the CP-mediated reduction in SOD activity and ΔZn/Mn does not alter SOD activity. These results indicate that CP potently inhibits S. aureus SOD activity via metal sequestration. To address whether CP treatment disrupts the Mn-dependent SOD-independent mechanism of superoxide defense, a strain lacking SOD activity, ΔsodAΔsodM (sodA:tet sodM:erm), was examined. While ΔsodAΔsodM is more sensitive to superoxide stress than wild-type, CP treatment further increases this sensitivity. As with wild-type, the CP-induced increase in sensitivity of ΔsodAΔsodM to superoxide is reversed by the addition of excess manganese. While the two Mn-dependent superoxide defense mechanisms are inactivated by CP, it is possible that loss of these systems does not adversely affect bacterial superoxide levels. To address this issue, intracellular superoxide levels in wild-type and ΔsodAΔsodM were examined. Upon CP treatment, both wild-type and ΔsodAΔsodM have elevated levels of intracellular superoxide. In total, these data suggest that CP inactivates Mn-dependent superoxide defenses in S. aureus, resulting in accumulation of superoxide. Additionally, the glutathione experiments indicate that CP-mediated reduction in S. aureus growth is not due to loss of superoxide defenses but inactivation of other bacterial processes. Elucidation of the essential processes that are disrupted by CP-mediated metal sequestration requires the identification of the S. aureus processes that are dependent on either manganese or zinc. Furthermore, as CP can inhibit bacterial processes and the staphylococcal abscess is virtually devoid of manganese, it is likely that S. aureus has developed specific mechanisms for overcoming this host defense. While these bacterial defense mechanisms remain unknown, they represent potential new targets for therapeutic intervention.

To address if CP inhibition of superoxide defenses would be relevant during infection, we assessed the ability of S. aureus to resist neutrophil-mediated killing following CP treatment. We observed that CP treatment increases the sensitivity of both wild-type and ΔsodAΔsodM to neutrophil-mediated killing. To address the relative contribution of the SODs to staphylococcal virulence, C57BL/6 mice were infected with either ΔsodAΔsodM or wild-type bacteria. The ΔsodAΔsodM mutant has a significant reduction in virulence, manifested by reduced colony forming units compared with wild-type in the livers of infected animals. Given the importance of the SODs to virulence, we next asked if manganese sequestration by the host reduces SOD activity during infection. To address this question, C57BL/6 and CP-deficient (C57BL/6 S100A9^−/−) mice were infected with either wild-type bacteria or ΔsodAΔsodM. Consistent with our prior results, we observed an increase in bacteria in the livers of CP-deficient mice infected with wild-type S. aureus when compared with C57BL/6 mice (Corbin et al., Science 2008). However, a statistical increase in the number of bacteria in the livers of CP-deficient mice infected with ΔsodAΔsodM when compared with C57BL/6 mice was not observed. This result indicates that the increase in bacteria observed in CP-deficient mice infected with wild-type S. aureus is in part due to increased SOD activity. By extension, these results suggest that manganese sequestration by CP in wild-type mice inhibits staphylococcal SOD activity.

In total, our results suggest a two hit mechanism of action where CP-mediated metal sequestration inhibits both factors essential for bacterial growth as well as those that protect the bacterium from host defense factors such as the neutrophil oxidative burst (Fig. 1). The ability of CP to reduce bacterial SOD activity is likely not limited to Staphylococci, as a range of medically relevant pathogens express Mn-dependent or Cu/Zn-dependent SODs including Streptococcus pneumoniae, Salmonella Typhimurium, Yersinia enterocolitica and Neisseria meningitidis (Fang et al., Proc Natl Acad Sci USA 1999; Lynch and Kuramitsu, Microbes Infect 2000; Roggenkamp et al., Infect Immun 1997; Yesilkaya et al., Infect Immun 2000). Furthermore, work with Neisseria gonorrhoeae, which lacks a Mn-dependent SOD, suggests that Mn-dependent SOD-independent defenses may protect against oxidative stress during infection (Seib et al., J Infect Dis 2004; Tseng et al., Mol Microbiol 2001; Veyrier et al., PLoS Pathog 2011). This observation raises the possibility that CP inhibition of SOD-independent oxidative stress defenses may also contribute to the control of invading pathogens.

Figure 1. Model of how metal sequestration by calprotectin affects S. aureus. (A) In the absence of calprotectin (CP) and metal sequestration by the host, S. aureus is able to acquire sufficient Mn and Zn to supply superoxide dismutases (SOD) and essential metal-dependent proteins with the appropriate cofactor. (B) Calprotectin contributes to the creation of a metal deficient environment by binding Mn and Zn, which are subsequently removed from the abscess. The reduced availability of Mn and Zn inactivates SODs, which in turn renders S. aureus more sensitive to the oxidative burst of neutrophils. Additionally, the reduced metal availability within the abscess leads to decreased activity of unknown but essential Mn and Zn dependent staphylococcal processes. The decreased activity of these essential processes in turn results in reduced bacterial growth.

Our results underscore the importance of manganese and zinc sequestration to combating infection and nutritional immunity. Furthermore, they provide key insights into how CP binds transition metals as well as the bacterial processes disrupted by this host defense. Additional studies are required to define the full array of metal-dependent bacterial processes and to identify which of these are inactivated by CP. Moreover, the structural basis for the transition metal binding specificity of CP needs to be elucidated and related to other members of the S100 protein family to establish if they can contribute to host defense and nutritional immunity. Ultimately, investigations into these areas could lead to the design of novel therapeutics based on nutritional immunity that could serve as alternatives to the traditional antibiotic treatments that are rapidly becoming obsolete in the face of increasing antibiotic resistance.

Glossary

Abbreviations:

ITC

isothermal titration calorimetry

CP

calprotectin

SOD

superoxide dismutase

Kehl-Fie TE, Chitayat S, Hood MI, Damo S, Restrepo N, Garcia C, et al. Nutrient metal sequestration by calprotectin inhibits bacterial superoxide defense, enhancing neutrophil killing of Staphylococcus aureus. Cell Host Microbe. 2011;10:158–64. doi: 10.1016/j.chom.2011.07.004.